ABSTRACT 
 
 
 
 
Title of Document: AN INVESTIGATION OF THE RELATIONSHIP 
BETWEEN THE LEVEL OF ANTIBIOTIC 
RESISTANCE OF ESCHERICHIA COLI IN NON-
 TIDAL WETLANDS AND COMMON WETLAND 
HEALTH FACTORS 
  
 Neil Agarwal, Sean Ahearn, Erik Dudziak, Sehba Khan, 
Daniel Marcin, Matthew Shofnos, Emily Skoda, 
Padmasini Venkatachari, Robert Vocke III 
  
Directed By: Dr. David Tilley, Ph.D. 
Environmental Science & Technology 
 
 
This report investigated the prevalence of antibiotic resistance among Escherichia coli in the 
water of 13 non-tidal mitigation wetlands in Maryland, and its relation to land use and 
wetland health.  At each site, land use, surface and sub-surface water samples, soil samples, 
and vegetation cover were collected.  From the water samples, individual colonies of E. 
coli were isolated and tested, using the disc diffusion method, for resistance to the antibiotics 
ampicillin, ciprofloxacin, erythromycin, sulfisoxazole, and tetracycline.  According to soils, 
vegetation and water quality improvement criteria the wetlands function like healthy 
wetlands.  The wetlands' E. coli exhibit resistance to all of the antibiotics tested, except for 
ciprofloxacin.  There were statistically significant relationships found between land use and 
antibiotic resistance, vegetation, soil and water chemistry.  Surprisingly, E. coli in wetlands 
with smaller stocks of carbon and nitrogen in their soil exhibited more resistance to 
tetracycline, possibly indicating that soil quality plays an important role in fostering or 
fighting antibiotic resistance. The work demonstrates that antibiotic resistance is present in 
Maryland's wetlands, but that its spread could be subdued by healthy wetlands. 
  
 
AN INVESTIGATION OF THE RELATIONSHIP BETWEEN THE LEVEL OF 
ANTIBIOTIC RESISTANCE OF ESCHERICHIA COLI IN NON-TIDAL 
WETLANDS AND COMMON WETLAND HEALTH FACTORS 
 
 
 
By 
 
 
Team CRABS 
(Chesapeake Revitalization and Alternative Bay Solutions) 
 
Neil Agarwal 
Sean Ahearn 
Erik Dudziak 
Sehba Khan 
Daniel Marcin 
Matthew Shofnos 
Emily Skoda 
Padmasini Venkatachari 
Robert Vocke III 
 
 
 
Thesis submitted in partial fulfillment of the requirements of the Gemstone Program 
University of Maryland, College Park 
2008 
 
 
 
 
 
Advisory Committee: 
Dr. David Tilley, Ph.D, Mentor 
Dr. Andrew Baldwin, Ph.D, Discussant 
Ms. Denise Clearwater, Discussant 
Dr. Sam Joseph, Ph.D, Discussant 
Mr. Todd Nichols, Discussant 
Mr. Peter Sharpe, Discussant 
 
  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
? Copyright by 
 
Team CRABS 
Neil Agarwal, Sean Ahearn, Erik Dudziak, Sehba Khan, Daniel Marcin,  
Matthew Shofnos, Emily Skoda, Padmasini Venkatachari, Robert Vocke III 
2008 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ii 
Acknowledgements 
 
Team CRABS would like to thank: 
 
Ms. Denise Clearwater and the Maryland Department of the Environment 
 
Dr. Andy Baldwin, Mr. Peter Sharpe, Dr. Sam Joseph, Mr. Todd Nichols 
 
The wetland site managers: The Nature Conservancy, the Maryland State Highway 
Administration, Allison Townshend, Bob Cantin, Laura Connelly, Lisa Gebhardt, 
Doug Samson, Peggy Schaumburg, Bob Shockley, Erin Wisnieski 
 
Ms. Inbal Levavi 
Our librarian Ms. Mariann Burright 
 
Our mentor Dr. David Tilley  
 
Dr. James Wallace and the entire Gemstone Staff for their encouragement, expertise, 
and enthusiasm. 
 
 
 
This research was supported by the Maryland Department of the Environment 
Wetlands and Waterways Program. 
 
 iii 
Table of Contents 
 
 
Acknowledgements....................................................................................................... ii 
Table of Contents......................................................................................................... iii 
List of Tables ................................................................................................................ v 
List of Figures .............................................................................................................. vi 
 
1. Introduction............................................................................................................... 1 
1.1 The CRABS Team .............................................................................................. 1 
1.2 The Status of Maryland?s Non-Tidal Wetlands .................................................. 2 
1.3 The Problem: Pharmaceutical Runoff................................................................. 6 
1.4 Objectives ......................................................................................................... 12 
1.5 Outline of Study................................................................................................ 13 
1.6 General Study Hypotheses................................................................................ 14 
 
2. Literature Review.................................................................................................... 15 
2.1 Wetlands ........................................................................................................... 15 
2.1.1 Mitigation Wetlands................................................................................... 15 
2.2 Wetland Assessment Protocols ......................................................................... 16 
2.2.1 Vegetation .................................................................................................. 16 
2.2.1.1 Classification Systems ........................................................................ 16 
2.2.1.2 History................................................................................................. 17 
2.2.1.3 Vegetation Sampling Methods............................................................ 20 
2.2.2 Water Quality............................................................................................. 22 
2.2.3 Soil ............................................................................................................. 28 
2.3 Antibiotics......................................................................................................... 31 
2.3.1 Methods of Analysis .................................................................................. 31 
2.3.2 Selection of Antibiotics.............................................................................. 35 
2.3.3 A Brief Background of Antibiotics Chosen for This Study....................... 35 
Erythromycin .................................................................................................. 35 
Ciprofloxacin .................................................................................................. 36 
Ampicillin ....................................................................................................... 36 
Sulfisoxazole................................................................................................... 37 
Tetracycline..................................................................................................... 38 
Vancomycin .................................................................................................... 38 
2.3.4 Prevalence of Antibiotic Resistance .......................................................... 39 
 
3. Methods................................................................................................................... 44 
3.1 Site Selection .................................................................................................... 44 
3.2 Methods of Data Collection .............................................................................. 45 
3.2.1 Land Use .................................................................................................... 45 
3.2.2 Field Work ................................................................................................. 51 
3.2.2.1 Plot Setup ............................................................................................ 51 
3.2.2.2 Vegetation Sampling........................................................................... 54 
 
 iv 
3.2.2.3 Soil Sampling...................................................................................... 58 
3.2.2.4 Water Sampling .................................................................................. 59 
3.2.3 Lab Work ................................................................................................... 61 
3.2.3.1 Chemical Analysis of Soils................................................................. 61 
3.2.3.2 Chemical Analysis of Water Samples................................................. 62 
3.2.3.3 Antibiotic Resistance Testing of Water Samples................................ 63 
3.3 Methods of Data Analysis................................................................................. 67 
 
4. Results & Discussion .............................................................................................. 68 
4.1 Land Use ........................................................................................................... 68 
4.2 Vegetation Sampling......................................................................................... 75 
4.2.1 Species Leaf Area Cover and Land Use .................................................... 76 
4.2.2 Wetland Vegetation Prevalence Index....................................................... 79 
4.2.3 Shannon Diversity Index............................................................................ 82 
4.3 Water Sampling ................................................................................................ 84 
4.3.1 Total Nitrogen............................................................................................ 85 
4.3.2 Organic Nitrogen ....................................................................................... 87 
4.3.3 Nitrite ......................................................................................................... 89 
4.3.4 Nitrate ........................................................................................................ 91 
4.3.5 Ammonia.................................................................................................... 93 
4.3.6 Phosphorous............................................................................................... 97 
4.3.7 Ammonia versus Total Phosphorous ....................................................... 101 
4.3.8 Limitation of Water Results..................................................................... 103 
4.4 Soil Sampling.................................................................................................. 103 
4.5 Antibiotic Resistance ...................................................................................... 107 
4.5.1 Overall Levels of Resistance Across Sites............................................... 107 
4.5.2 Multidrug Resistance ............................................................................... 111 
4.5.3 Antibiotic Resistance Stratified By Study Site ........................................ 115 
4.6 Statistical Analysis.......................................................................................... 117 
4.6.1 Relationships Amongst Wetland Health Factors ..................................... 118 
4.6.2 Relationships to Land Use ....................................................................... 119 
4.6.3 Relationships to Antibiotic Resistance .................................................... 121 
 
5. Conclusion ............................................................................................................ 130 
 
References................................................................................................................. 134 
 
Appendix A............................................................................................................... 145 
Appendix B ............................................................................................................... 147 
Appendix C ............................................................................................................... 149 
Appendix D............................................................................................................... 157 
Appendix E ............................................................................................................... 160 
 
 
 v 
List of Tables 
 
Table 1: Example of urban or built-up land level II categories .................................. 18 
Table 2: Subsytems of wetlands derived from the Cowardin method ........................ 19 
Table 3: Brief descriptions of the 13 wetland sited used in this study ....................... 45 
Table 4: The 29 different land use classifications, sorted by color ............................ 48 
Table 5: 29 land use classifications condensed into three classifications to better suit 
CRABS? study .................................................................................................... 49 
Table 6: CRABS? 13 wetland sites organized by land use classification ................... 50 
Table 7: Dimensions of each intensive depth ............................................................. 55 
Table 8: Leaf area cover classes ................................................................................. 57 
Table 9: Wetland indicator status categories for plant species ................................... 80 
Table 10: Top 15 species by leaf area cover over all sites sampled by CRABS, along 
with their indicator status.................................................................................... 80 
Table 11: Soil nutrient levels from each plot at the 13 sampled wetland sites......... 104 
Table 12: Average value and standard deviation for soil nutrient data .................... 105 
Table 13: Relationship between multidrug resistance and resistance to individual 
antibiotics.......................................................................................................... 113 
Table 14: Percentages of resistant E. coli isolated, stratified by wetland site .......... 116 
Table 15: Statistically significant relationships found amongst wetland health factors
 ........................................................................................................................... 118 
Table 16: Statistically significant relationships found between wetland health factors 
and land use factors........................................................................................... 119 
Table 17: Results of 124 t-tests showing all statistically significant relationships 
between antibiotic resistance and various wetland indicators .......................... 122 
Table 18: Soil characteristics and nutrient concentrations ....................................... 145 
Table 19: Surface water quality at each plot-site, (BDL-below detection limit)...... 147 
Table 20: Sub-surface water quality at each site-plot, (BDL-below detection limit)148 
Table 21: Complete list of plant species found at each wetland site ........................ 149 
Table 22: Wetland Site Information ......................................................................... 157 
Table 23: Raw Land Use Data for 1000m Buffer..................................................... 160 
Table 24: Raw Land Use Data for 2000m Buffer..................................................... 161 
Table 25: LDI, % AG, % Urban & % Natural for the 13 wetland sites ................... 162 
 
 
 vi 
List of Figures 
Figure 1: Map of the 13 Maryland wetland sites used in this study ........................... 44 
Figure 2: Screenshot from GISHydro2000, showing land use classifications by color, 
as well as 1000m and 2000m circular buffers around wetland sites................... 47 
Figure 3: Diagram of a normal NCVS plot................................................................. 51 
Figure 4: Schematic diagram showing truncation and renumbering of the standard 
NCVS plot........................................................................................................... 52 
Figure 5: Schematic diagram of a completed Team CRABS plot .............................. 53 
Figure 6: Method of setting up nested depths within intensive module ..................... 55 
Figure 7: Example of a Millipore m-ColiBlue plate................................................... 65 
Figure 8: LDI values from 1000m and 2000m circular buffers surrounding each of the 
13 study sites....................................................................................................... 68 
Figure 9: Percentage of land used for agricultural purposes in the areas surrounding 
each of the 13 study sites .................................................................................... 69 
Figure 10: Percentage of urban land in the area surrounding each of the 13 study sites
 ............................................................................................................................. 70 
Figure 11: Percentage of ?natural? land use in the area surrounding each of the 13 
study sites............................................................................................................ 70 
Figure 12: Top ten species by cover at agricultural sites, along with their cover at 
natural and urban sites ........................................................................................ 77 
Figure 13: Top ten species by cover at natural sites, along with their cover at 
agricultural and urban sites ................................................................................. 77 
Figure 14: Top ten species by cover at urban sites, along with their cover at 
agricultural and natural sites ............................................................................... 78 
Figure 15: Wetland vegetation prevalence index for all sites..................................... 81 
Figure 16: Shannon diversity index for all sites ......................................................... 83 
Figure 17: Shannon diversity index with the x-axis showing the sites sorted from 
more natural to less natural by the ratio of natural to urban land use................. 84 
Figure 18: Total nitrogen levels across all sites in 2006............................................. 87 
Figure 19: Organic nitrogen levels across all sites in 2006 ........................................ 89 
Figure 20: Sub-surface nitrite levels across all sites in 2006 and in 2007.................. 91 
Figure 21: Surface water nitrate levels for all sites in 2006 ....................................... 92 
Figure 22: Sub-surface and surface ammonia levels in 2007 ..................................... 94 
Figure 23: Sub-surface ammonia levels in 2006 and in 2007..................................... 95 
Figure 24: Ammonia level averages from the Florida study and from CRABS? 
Maryland study ................................................................................................... 97 
Figure 25: Sub-surface and surface total phosphorous levels in 2006........................ 99 
Figure 26: Sub-surface total phosphorous in 2006 and in 2007 ................................. 99 
Figure 27: Phosphorous levels from the Florida study and from CRABS? Maryland 
study.................................................................................................................. 101 
Figure 28: Comparison of ammonia and total phosphorous levels across all sites, 
averaged ............................................................................................................ 102 
Figure 29: Regression of ammonia level versus total phosphorous level................. 102 
Figure 30: Gleying and mottling intensity values from each of the 13 wetland sites106 
 
 vii 
Figure 31: Overall percentage of E. coli isolates resistant to each antibiotic tested, 
spanning all wetland sites as well as data from both 2006 and 2007 ............... 107 
Figure 32: Comparison of overall E. coli resistance results from this study to 
resistance results from several similar previous studies ................................... 109 
Figure 33: Multidrug resistance for E. coli isolates that were tested against all six 
antibiotics.......................................................................................................... 112 
Figure 34: Relationship between multidrug resistance and resistance to individual 
antibiotics.......................................................................................................... 114 
Figure 35: Bacterial resistance to tetracycline plotted against soil nutrient levels ... 127 
 
 1 
1. Introduction 
 
1.1 The CRABS Team 
The Gemstone program at the University of Maryland, College Park is a 
unique program that brings together talented undergraduate students of varying 
disciplines to work on multi-year interdisciplinary team research projects.  Team 
CRABS was formed in 2005, and composed of nine undergraduate students in the 
Gemstone program studying the health of wetlands and waterways throughout the 
state of Maryland.  This team performed its research under the auspices of Dr. David 
Tilley of the University of Maryland?s Department of Environmental Science and 
Technology.   
Team CRABS was borne of its members' common concern for environmental 
issues and the health of the Chesapeake Bay watershed.  In selecting a research topic, 
CRABS decided to focus on the Chesapeake Bay's wetlands; in light of growing 
evidence that wetlands act as ecological "filters" that reduce the flow of pollutants 
through a watershed, it has become increasingly clear that the health of wetlands is 
closely tied to that of the watershed as a whole.  Being that these research interests 
aligned with the priorities of the Maryland Department of the Environment (MDE), 
the team?s research was predominantly funded by a grant from the agency?s Wetlands 
and Waterways Program. 
 The Maryland Department of the Environment is the state of Maryland?s chief 
environmental agency.  Its mission is to protect and restore the quality of Maryland?s 
air, water, and land resources, among other endeavors.  Founded in 1987, the MDE 
works to enforce and regulate health standards, growth issues, and environmental 
 
 2 
emergencies throughout the state.  Accordingly, the MDE is required to comply with 
the Clean Water Act of 1972.  This act requires each state to ?assess the quality of 
their waters every two years and publish a list of those waters not meeting the water 
quality standards set for them? (U. S. Environmental Protection Agency, 1972). 
 
1.2 The Status of Maryland?s Non-Tidal Wetlands 
Wetlands are unique transitional ecosystems ? neither strictly aquatic nor 
terrestrial ? which are saturated by water often enough during their growing seasons 
to develop characteristic hydric soils and support specially-adapted hydrophytic 
vegetation.  In particular, a fundamental characteristic of wetland soils is that they are 
anaerobic; as a result of being inundated by water they are less able to receive oxygen 
from the atmosphere.  Wetlands are particularly worthy of attention because beyond 
their own inherent complexity they have a wide variety of benefits to both humans 
and the broader environment.   
One of the most evident benefits of wetlands is their role in protecting 
biological diversity.  They are home to innumerable species of plants, animals, and 
microbes, including a number of economically important species and over one-third 
of all threatened and endangered species in the United States (U.S. Environmental 
Protection Agency, 1995).  Wetlands are also important for their ability to reduce 
shoreline erosion and to lessen the impact of flooding and droughts.  Because of their 
transitional hydrology, wetlands can take up storm water during periods of high 
rainfall while slowly releasing it during periods of low rainfall.  Such hydrological 
stabilization is important for both humans and the ecosystem as a whole. 
 
 3 
Another function of wetlands, and perhaps the one most pertinent to the 
present study, is their ability to improve water quality.  Wetlands have a number of 
characteristics that allow them to naturally take up sediments and harmful chemicals: 
they reduce water velocity and allow for deposition of suspended sediments, their 
shallow water allows for a high rate of chemical exchange between ground and water, 
and their high productivity in combination with their diversity of biogeochemical 
processes facilitates the uptake and/or degradation of numerous chemicals (Mitsch & 
Gosselink, 2000).  In the scheme of an entire watershed, wetlands are often situated 
between developed areas and low-lying aquatic systems like rivers and lakes.  As 
such, water draining through a watershed can improve in quality as it passes through 
one or more wetlands before reaching a more traditional waterway.  Considerable 
scientific attention has been focused on this idea, and an entire subfield is now 
devoted to the use of wetlands to treat polluted water (Kadlec & Knight, 1996).  
 Unfortunately, the benefits of wetlands have not always been so well-
 recognized.  In the past, human destruction of wetlands has been a considerable 
problem, with the U.S. having lost more than 50% of its original wetland area in less 
than two centuries (Balcombe et al., 2005).  Even as the issue of water pollution 
began drawing concern towards the second half of the twentieth century, wetland 
protection often fell by the wayside.  Most notably, when first enacted in 1972, the 
United States? Clean Water Act became a significant step in protecting the water 
quality of the nation?s navigable waterways, but failed to include any mention of 
wetlands. 
 
 4 
 In 1975, the Supreme Court cases United States v. Holland and Natural 
Resources Defense Council v. Calloway reinterpreted the Clean Water Act?s 
definition of water bodies to include wetlands (Mitch & Gosselink, 2000).  As a 
result, under Section 404 of the act, anyone wanting to dredge or fill a wetland now 
had to obtain a permit from the U.S. Army Corps of Engineers.  This permit process 
requires that wetland losses are avoided or minimized when possible, and that any 
unavoidable losses of wetland function are compensated for through the creation or 
restoration of so-called mitigation wetlands (National Research Council, 2001).  
 The use of mitigation wetlands in the U.S. has become a major route by which 
wetland functions are preserved.  In 1989, President Bush officially adopted the goal 
of ?no net loss? as his administration?s policy towards wetland conservation (National 
Research Council, 2001).  In the wake of this policy, mitigation wetlands created 
under Section 404 were responsible for an estimated net gain of 500 square km of 
wetland area between 1993 and 1999 (Mitsch & Gosselink, 2000). 
 Despite this recent progress in the realm wetland preservation, a separate issue 
is that existing wetlands are still under stress from water pollution resulting from 
human activity.  The ability of wetlands to reduce water pollution has been discussed, 
but it should be noted that wetlands can also be impacted by the same pollutants they 
filter out.  For example, a study by Barber at al. (2006a) on the fate of pollutants in a 
treatment wetland revealed that many of the toxins filtered out by that wetland ended 
up accumulating in its fish and compromising their health. 
Nitrogen and phosphorous are two of the most known and studied 
environmental water pollutants.  As plant nutrients, they are extremely important for 
 
 5 
their use as agricultural fertilizers.  However, when released into the environment in 
high quantities, nitrogen and phosphorous can cause imbalances to aquatic 
ecosystems in a phenomenon referred to as eutrophication.  Though eutrophication is 
generally associated with deeper-water systems, it can also be detrimental to 
wetlands.  Indeed, it has been noted many times that nutrient overabundance in a 
wetland can result in the dominance of certain plant species at the expense of others, 
reducing overall wetland biodiversity (Bedford, Walbridge, & Aldous, 1999; Tilman 
& Lehman, 2001).  Isolatable point sources of nutrients and other pollutants (e.g. 
factory outflows, landfills, etc.) are generally regulated under applicable laws, but 
non-point source pollution originates as runoff from various ill-defined sources (e.g. 
lawns, farms, roadways, etc.) and is thus particularly difficult to account for and 
regulate. 
Because wetland ecosystems protect downstream waterways from water 
pollution and are simultaneously impacted by pollution, monitoring wetland water 
quality and other aspects of wetland health and function is of utmost importance.  
Mitigation wetlands require special attention; while the intended purpose of 
mitigation wetlands is to replace lost wetland function, research has shown that many 
either do not fulfill their permit requirements or are unable to fully match natural 
reference sites in terms of value and function (National Research Council, 2001).  
Furthermore, the impacts of human development are particularly pertinent to the 
study of mitigation wetlands, due to the fact that placement of mitigation sites is often 
based on cost and convenience as opposed to ecological optimality.  In other words, 
mitigation wetlands are often ?artificially? placed in developed areas where they may 
 
 6 
be more heavily affected by human activity.  One recent study found wetland plant 
diversity to be affected by surrounding land use (Houlahan, Keddy, Makkay, & 
Findlay, 2006).  In light of this discussion, it is likely that there exist many more such 
relationships between human activity and wetland function. 
 
1.3 The Problem: Pharmaceutical Runoff 
Pharmaceuticals and personal care products are all but ubiquitous in the 
modern world, and as their use among humans and farm animals continues, so does 
the potential for their entry into the surrounding environment.  In normal use, up to 
90% of pharmaceuticals pass through their host system non-metabolized, while even 
metabolized portions may continue to pose an environmental risk (Jones, Voulvoulis, 
& Lester, 2003).  Additionally, unused pharmaceuticals are often subject to improper 
disposal.  Dietrich, Webb, and Petry (2002) note that the same properties that give 
pharmaceuticals the ability to function in the body also make them potentially 
persistent and harmful to aquatic ecosystems: pharmaceuticals must inherently 
interact with living systems, they are often designed for stability within such systems, 
and their proper bodily transport often requires that they be water soluble. 
It was not until the 1990s that the harmful potential of pharmaceuticals began 
to gain more widespread scientific attention (Dietrich et al., 2002).  However, the 
development of analytical methods capable of detecting pharmaceuticals in the 
environment was, and is, often a limiting factor in the study of this topic (Daughton & 
Ternes, 1999; Bruchet et al., 2005).  Early studies were relatively sparse and 
generally focused on narrow aspects of the issue; many authors noted the need for a 
 
 7 
clearer picture of the general extent of pharmaceutical contamination.  In 2002, a 
landmark study from the United States Geological Survey (Kolpin et al., 2002) 
presented the first nationwide picture of the occurrence of pharmaceuticals in 
waterways ? it found organic wastewater contaminants in 80% of streams studied.  
Compounds detected in the study originated in agricultural, industrial, and residential 
sources, and included antibiotics, antidepressants, disinfectants, fragrances, the 
metabolites of detergents, prescription and nonprescription painkillers, and 
hormonally active steroids.  The study noted that though concentrations were 
generally low, many compounds do not have guidelines as to what represents a safe 
level. 
There are in fact numerous factors complicating the issue of what constitutes a 
?safe level.?  Many manmade compounds are not easily biodegradable and will 
accumulate in the environment regardless of their rate of release.  But even for 
pharmaceuticals that degrade with relative ease, guidelines on acute ecological 
toxicity only cover a portion of the picture.  Constant and continual release of 
degradable compounds can present a chronic exposure risk that is difficult to account 
for, since organisms may spend their entire lives exposed to a compound in minute 
concentrations.  Ecological harm may accumulate so slowly that it is difficult or 
impossible to discern its original source (Daughton & Ternes, 1999; Jones et al., 
2000; Halling-S?rensen et al., 1998).  Furthermore, simultaneous occurrence of 
multiple contaminants is hindering efforts to establish causal relationships to 
individual compounds and creating the potential for complex interactions between 
different contaminants (Kolpin et al., 2002; Jones et al., 2003; Boxall, Kolpin, 
 
 8 
Halling-S?rensen, & Tolls, 2003).  Finally, analytical techniques used in studies of 
pharmaceutical prevalence have generally required a preexisting knowledge of 
analyte compounds, and as such have been of limited use in elucidating the chemical 
changes the pharmaceuticals may undergo before and after their release into the 
environment.  Undetected metabolites may remain active, or may even be 
transformed back into their parent compounds soon after their release into the 
environment (Glassmeyer & Shoemaker, 2005; Ternes, 2001; Halling-S?rensen et al., 
1998).  Recent advancements in analytical chemistry show hope in overcoming this 
issue (Kosjek, Heath, Petrovi?, & Barcel?, 2007).   
Driven by the need to account for these complexities, research detailing the 
environmental fate and effects of pharmaceuticals has intensified in recent years.  For 
example, Carballa et al. (2005) studied the fate of thirteen pharmaceutical and 
personal care compounds in a sewage treatment plant in northwest Spain.  Their study 
revealed a significant presence of eight compounds, and found that upon passing 
through the treatment plant, concentrations of most compounds were reduced 
substantially but not fully.  Barber et al. (2006) measured for various organic and 
inorganic contaminants in the Boulder Creek watershed of Colorado, including 
twelve pharmaceutical compounds.  In general, they found that contaminant 
concentrations increased in areas of high population density and spiked immediately 
downstream of wastewater treatment plants, and noted a possible correlation to 
endocrine disruptions in native fish.  Cleuvers (2003) performed lab tests of the 
toxicity of ten pharmaceutical compounds on three different aquatic organisms.  
Measured levels of acute toxicity for individual compounds were much greater than 
 
 9 
their likely level of occurrence in the environment.  However, the study found the 
toxicities of certain mixtures to be greater than the sum of their parts, affirming the 
potential for additive interactions.  Liu and Williams (2007) studied the sunlight-
 induced degradation of several ?-blocker pharmaceuticals, and were able to explicitly 
identify multiple degradation products for the drug propranolol.  They noted the 
potential usefulness of this pathway as a propranolol attenuation mechanism, but did 
not discuss the potential effects of newly formed products.  Meanwhile, a study by 
Bedner and MacCrehan (2006) revealed that acetaminophen, the active ingredient in 
many painkillers, can be transformed into decidedly more toxic compounds during 
chlorination in wastewater treatment plants.  Runnalls, Hala, and Sumpter (2007) 
studied the effects of clofibric acid, a persistent metabolite of the cardiovascular drug 
clofibrate.  They found that concentrations of clofibric acid almost as low as those 
reported in the environment are capable of causing reproductive impairments to the 
fathead minnow over a three week period.  While it has been noted that human 
exposure to pharmaceutical compounds may occur when ground and surface water 
are used to produce drinking water (Pauwels & Verstraete, 2006), the EPA asserts 
that no studies thus far have revealed any human effects from the release of 
pharmaceuticals (U.S. Environmental Protection Agency, 2007).   
Among the varying studies of pharmaceuticals in the environment, antibiotics 
have been identified as an area of particular concern.  Since the year 1942, when 
Anne Miller became the first person to be successfully treated using penicillin 
(Oransky, 2002), literally hundreds of new antibiotic compounds have been 
discovered or developed (Neu, 1992).  Antibiotics are an extremely important class of 
 
 10 
pharmaceuticals ? their use encompasses both human and veterinary medicine, and 
extends to non-therapeutic uses such as growth promotion in farm animals.  The 
widespread use of antibiotics is reflected by the fact that it is very common for 
antibiotics to be detected in studies of the pharmaceuticals in the environment.  
However, an oft-cited drawback to the widespread use of antibiotics is that it creates a 
selective pressure that favors the development and spread of bacteria that are resistant 
to treatment by antibiotics.  In under a century, the spread of antibiotic resistant 
bacteria has become a major issue; countless studies have documented cases of 
bacteria becoming resistant to commonly used antibiotics, to the point where many of 
the most widely used antibiotics have lost much of their original effectiveness.  The 
overall trend in the spread of antibiotic resistance creates a vicious cycle that 
threatens the use of existing antibiotics while constantly necessitating the increased 
development and use of new ones. 
In itself, the issue of antibiotic resistance has been the subject of much 
research.  Antibiotic resistance has been a well-known concern to both the scientific 
and lay communities for many years, and since antibiotic resistance poses a 
recognized threat to human medicine in addition to the environment, its study is 
naturally of high priority.  However, research conducted to date has generally leaned 
towards studying human isolates from human locales ? it is only recently that 
scientists have begun to study how antibiotic resistance may operate in the open 
environment.  Yet environmental studies of antibiotic resistance are significant for at 
least two reasons: it is important not to overlook the spread of antibiotic resistance as 
a form of pollution that can adversely natural ecosystems, and it is important to 
 
 11 
realize that environmental locales may be sites for the exchange of antibiotic 
resistance genes, which can subsequently have adverse effects on human medicine. 
Indeed, recent research by Halda-Alija (2004) and Biyela et al. (2004) 
indicates that wetlands and waterways may retain, and possibly facilitate the spread 
of, antibiotic resistant pathogens.  Halda-Alija (2004) demonstrated the presence of 
several species of antibiotic resistant pathogenic bacteria in wetlands in Mississippi, 
and concluded that wetlands may retain genes for antibiotic resistance but not 
necessarily aid in their dissemination.  Biyela et al. (2004) found antibiotic resistant 
bacteria and their corresponding genetic elements in water samples taken from the 
Mhlathuze River in South Africa, and revealed a strong correlation between the 
resistance profiles of bacteria in the environment and bacteria in stool samples of 
nearby human diarrhea patients.  The authors wrote that ?the prevalence of antibiotic 
resistance was directly related to the frequency of antibiotic usage,? and went on to 
conclude that the Mhlathuze River ?not only played a role as a reservoir but also was 
a medium of spread and evolution of antibiotic resistance.? 
In wetlands, antibiotics and antibiotic resistance may also have subtle effects 
on the microbial communities that play important roles in many aspects of a 
wetland?s biogeochemistry.  Examples of non-target microbes that may potentially be 
affected by antibiotics include photosynthetic cyanobacteria, nitrifying and 
denitrifying bacteria, and decomposers residing in the wetland detritus layer (Jones et 
al., 2003; Maul, Schuler, Belden, Whiles, & Lydy, 2006).  Changes that begin at the 
level of bacteria may cascade through the food chain (Daughton & Ternes, 1999), 
 
 12 
creating disturbances to the wetland ecosystem that fall well outside the bounds of 
simple considerations of acute or chronic toxicity. 
The topic of antibiotic resistance in the environment is poorly understood and 
subject to myriad complexities.  Given the likelihood of continued increases in the 
use of antibiotics and impingement of human activity on the environment, it should 
be abundantly clear that further study of this issue is merited. 
 
1.4 Objectives 
In order to add to the current body of wetland research and to help fill the 
significant research gaps surrounding the unknown effects of pharmaceutical runoff, 
the primary objective of this project was to assess the overall impact of human 
activity, broadly defined in this paper as ?human factors,? on specific indicators of 
wetland health.  The fundamental research question guiding this project was:  
How do human factors affect the health and function of Maryland's 
wetlands? 
Though this research question initially appeared direct in nature, it quickly 
proved to be quite complicated and multi-faceted in its scope.   For this study, 
?human factors? were defined by two parameters: the land use surrounding each 
wetland site, and the prevalence of antibiotic resistance found at each site.  This 
definition helped further split the fundamental research question into the pursuit of 
two more specific objectives: 
 
 13 
1. To use common indicators of wetland health and function to 
determine if Maryland?s wetlands are functioning properly, and to 
understand how they are impacted by surrounding land use.  
2. To assess the prevalence of antibiotic resistance in Maryland?s 
wetlands, and to determine its relationship to other indicators of 
wetland health and function. 
 
1.5 Outline of Study 
Mitigation wetland sites were selected for sampling with assistance from the 
Maryland Department of the Environment.  
Objective 1 was accomplished through the use of standard, well-studied 
environmental sampling protocols that involved collecting vegetation, soil, and water 
samples from each wetland site.  The soil and water samples were analyzed for 
specific characteristics and chemical composition, and standardized techniques were 
used to quantify and qualify the vegetation found at each site.   This collected data 
was then compared to accepted measures of wetland health and to site land use 
designations determined through the use of a geographic information system (GIS).  
Significant correlations were then evaluated and used to provide an overall 
assessment of the functioning of each of the Maryland wetland sites tested.   
Objective 2, required a more novel research strategy due to the scarcity of 
published bacterial antibiotic resistance data for wetlands.  To complete this 
objective, water samples collected at each site were tested for the presence of 
Escherichia coli.  If E. coli was present, a disc diffusion method was used to 
 
 14 
determine the bacterial population?s degree of resistance to several different 
antibiotics.  A standardized procedure was created for these lab analyses in order to 
maintain consistency and validity across the scope of the study.  Comparing the 
resistance results to Objective 1?s determination of wetland health at each site 
allowed for correlations to be drawn between the level of antibiotic resistance and the 
functioning of the wetland. 
 
1.6 General Study Hypotheses 
? Wetlands heavily impacted by human development will exhibit higher levels 
of nutrients in soil and water and lower plant diversity than wetland in natural 
settings. 
? Wetlands heavily impacted by human development will exhibit higher levels 
of antibiotic resistance than more natural wetlands due to the higher volumes 
of pharmaceutical pollution entering them. 
 
 
 15 
2. Literature Review 
 
2.1 Wetlands 
2.1.1 Mitigation Wetlands 
Mitigation wetlands are artificially constructed to combat the consistent loss 
of wetland acreage.  To compensate for this consistent, unavoidable loss, the U.S.  
Army Corps of Engineers and the MDE issue permits to restore, enhance, or create 
wetlands (Brinson & Rheinhardt, 1996).  In order to do this, non-wetlands are 
converted to jurisdictional wetland status.  Typically, this requires excavating a 
depression that connects with the ground water table, or by hydraulically connecting 
the site to a source of ground water (Brinson & Rheinhardt, 1996). 
As can be seen in the most recent report from the United States Fish and 
Wildlife Service (USFWS), we are still losing significant acreage of the nation?s 
wetlands (United States Fish and Wildlife Service, 1997).  When the amendments 
were added to the Clean Water Act in 1977, a federal ?no net loss? policy regarding 
wetlands was endorsed.  The primary tool to combat wetland loss was the large-scale 
construction of mitigation wetlands.   
Working in conjunction with the MDE, thirteen mitigation wetland sites were 
chosen to complete this research due to the opportune level of access provided by the 
government and the mutually beneficial baseline data collection used to complete our 
research as well as contribute to the MDE?s body of data for Maryland?s waterways. 
 
 
 16 
2.2 Wetland Assessment Protocols 
2.2.1 Vegetation 
Wetland vegetation is a valuable indicator of a wetland?s function, and it is a 
primary factor considered when classifying wetlands.  Plant community structure is 
viewed as ?one of the best indicators of the factors that shape wetlands within their 
landscape? (Bedford, 1996).  Several studies have indicated that vegetation 
composition influences various factors, including, but not limited to, groundwater 
chemistry, erosion, and diversity of wildlife (Balcombe, 2005).  In addition, 
vegetation study assists in delineating the boundaries of the regions under study 
(Cowardin et al., 1979).  For example, under some classification systems, areas 
dominated by certain types of vegetation are not considered as part of what is defined 
as a ?wetland? (Cowardin et al., 1979).  In addition, plant composition provides a 
greater insight into the wetland under study, for it has been shown to influence 
various other characteristics of the wetland.  Finally, identification of vegetation does 
not require extensive knowledge, and the skills necessary to classify species can be 
acquired relatively easily. 
 
2.2.1.1 Classification Systems 
As noted by the EPA, several wetland classification systems have emerged in 
an attempt to ?reduce variability ? and enable more sensitivity in detecting 
differences between least impacted and impaired wetlands? (Fennessy, 2004).  These 
systems are based on numerous factors, including hydrology, landscape features, and 
plant composition.  The majority of techniques are more quantitatively based, while 
 
 17 
more experienced researchers may rely on more qualitative methods.  The most 
prominent systems include the Anderson Classification System, the Cowardin 
Classification System, and the General Wetland Vegetation Classification System 
(GWVCS) (Dieck & Robinson, 2004).  Each system was developed with specific 
goals in mind: the Anderson System consists of a two-tier hierarchy to be used with 
remote sensing systems; the Cowardin System is more detailed and classifies regions 
by focusing on ecologically similar habitats; and the GWVCS builds on the Cowardin 
system (Dieck & Robinson, 2004).  Another prominent technique, and the one 
employed in this study, is the North Carolina Vegetation Survey (NCVS).  Finally, a 
quantitative method discussed is the Transect Method. 
 
2.2.1.2 History 
The Anderson Classification System was first published in the early 1970s as 
a revision to an existing classification system presented in the U.S. Geological Survey 
Circular 671.  The Anderson system was designed specifically so that it would be 
agreeable to data from satellite, aircraft, and other remote sensing sources (Anderson, 
et al. 1976).  The goal was to provide a uniform method (at a basic level) to 
categorize land use and cover for federal and state purposes.  The Anderson System 
has a multilevel structure to accommodate different sensors that provide data at 
various resolutions depending on altitude and scale.  Level I classification is designed 
for use with the LANDSAT program, which is a joint coalition between NASA and 
the U.S. Geological Survey to gather satellite photography of the Earth.  Level II is 
for use with high-altitude data (40,000 feet or above), level III is for use with mid-
 
 18 
altitude data (10,000 ? 40,000 feet), and level IV is for use with low-altitude data 
(below 10,000 feet), but these generalizations are not limiting ?It is intended that 
these latter levels of categorization will be developed by the user groups themselves, 
so that their specific needs may be satisfied by the categories they introduce into the 
structure? (Anderson, et al., 1976).  For this reason, the system was left open-ended 
and detailed specifications were not given, thereby giving respective users flexibility 
in defining their own categories at higher levels of classification.  Data at levels I and 
II is probably more useful for individuals seeking information on a nationwide or 
statewide basis, while levels III and IV are more useful to those seeking information 
on a more local basis.   
Based on their survey, Anderson et al. proposed nine level I categories: urban 
or built-up land, agricultural land, rangeland, forest land, water, wetland, barren land, 
tundra, and perennial snow or ice.  These level I divisions were further divided for a 
total of 37 level II categories.  For example, urban or built-up land is divided as 
follows:  
 
Table 1: Example of urban or built-up land level II categories 
Level I Level II 
1 Urban or Built-up Land 
 
1.1 Residential 
 1.2 Commercial and Services 
 1.3 Industrial 
 1.4 Transportation, Communications, and Utilities 
 1.5 Industrial and Commercial Complexes 
 1.6 Mixed Urban or Built-up Land 
 1.7 Other Urban or Built-up Land 
 
 
 
 19 
Within this framework, users may create additional distinctions.  Residential, 
for instance, may be further classified in level III as single-family units, residential 
hotels, mobile home parks, etc. (Anderson, et al., 1976).   
Unlike the Anderson System, the remaining classification systems discussed 
do not such a broad, top-down approach.  At the highest level, the Cowardin method 
identifies five main categories that incorporate the various elements such as landscape 
and vegetation to characterize a wetland.  These categories are: Marine, Estuarine, 
Riverine, Lacustrine, and Palustrine.  In order to identify the plant communities 
themselves and better define a wetland, scientists have subdivided these categories as 
follows: 
 
Table 2: Subsytems of wetlands derived from the Cowardin method 
System Subsystem 
Marine Subtidal 
Intertidal 
Estuarine Subtidal 
Intertidal 
Riverine Tidal 
Lower Perennial 
Upper Perennial 
Intermittent 
Lacustrine Littoral 
Limnetic 
Palustrine None 
 
 
The subsystems are further divided based on substrate, soils, water chemistry, 
or vegetation.  An example is the Emergent Wetland, which can be found under the 
latter four systems.  Emergent wetlands are characterized by the presence of grasses, 
sedges, and other herbaceous species.  The plants are often perennial, and Emergent 
 
 20 
Wetlands can be found throughout the United States.  They are also known as 
marshes, meadows, and sloughs.  Further information can be found in Classification 
of Wetlands and Deepwater Habitats of the U.S. (Cowardin et al., 1979).   
The GWVCS builds on the Cowardin System and is the one endorsed by the 
U.S. Geological Survey in their General Classification Handbook for Floodplain 
Vegetation in Large River Systems (Dieck & Robinson, 2004) The GWVCS further 
extends the Cowardin method by identifying 31 major classes within six hydrologic 
domains.  These domains range from those in which water is constantly present to 
those where it is rarely present.  The process begins with obtaining aerial 
photographs, either color-infrared (CIR) or true-color, though the former is preferred.  
This is often done in late summer, for this is when aquatic vegetation is at the height 
of its growth.  Areas that appear debatable in the photographs are then visited and 
detailed observations are recorded.  The aerial photographs are subsequently analyzed 
using a stereoscope; the vegetation is categorized based on the 31 classes, and other 
factors such as density are determined.  In a final step, the results of this interpretive 
work are digitally processed using geographic mapping software in order to reference 
them with real world coordinates (Dieck & Robinson, 2004).   
 
2.2.1.3 Vegetation Sampling Methods 
Vegetation sampling methods have been developed to serve a variety of 
purposes, and selecting the correct method for a particular study is difficult.  Two 
major methods, the line intercept method and the cover/nested cover quadrants 
method, will be discussed here because of their merits in regard to this study.  
 
 21 
The line-intercept, or transect method is a quantitative means of assessing the 
vegetation of a wetland.  It involves setting up transects across a site and recording 
the occurrence of plant species along the transect line at predetermined intervals.  
Transect lines can range in length from millimeters to kilometers, depending on their 
intended application.  This method is particularly useful and applicable in sites where 
the vegetation is sparsely distributed, or in areas populated by tall trees.  It is a tedious 
method in areas where plants are small, interwoven, tussocky, or densely populated.  
The transect method can be easily adapted for use in small or large areas, which 
makes it attractive to many state and federal environmental agencies. In Maryland 
specifically, its use is required by the Maryland Department of the Environment when 
wetlands exceed 5 acres (Bonham 1989). 
The cover quadrants method is a well-established technique used when 
counting individual stems is impractical. It involves visually estimating leaf area 
cover with respect to a prescribed plot. Multiple graduated quadrants can be ?nested? 
to form layers of quadrants that give an idea of the prevalence of species, as well as 
total leaf area cover (Bonham 1989). An example of this survey method is the North 
Carolina Vegetation Survey.  Established in 1987, the main goal of the NCVS is to 
categorize the natural plant life of North Carolina and nearby states (Peet, et al., 
1997).  This will better enable researchers to interpret the interactions between plant 
life and the general environment and to monitor those relationships on a long-term 
basis.  The underlying idea behind the NCVS is to create a flexible method of 
vegetation classification to satisfy numerous purposes.  It employs a module, 
quadrant, based approach to determining the plot layout.  All measurements and 
 
 22 
observations are taken within plots consisting of multiple 10 x 10 m modules.  The 
number of modules utilized is left to the discretion of the user, thereby providing 
flexibility.  In each module, both cover and stem data are collected (Peet, et al., 
1997).  According to Peet et al., ?Percentage cover represents a crude estimate of the 
vertical projection of leaf area and other aboveground parts (not leaf area index) and 
is thus an index of a species' potential contribution to community production.? The 
presence of a species, on the other hand, is measured by the stem count.  In order to 
be considered within a module, the species must have at least one stem originating 
from the soil within the module.  Further measurement specifications are detailed in 
Section 3.2.2.1: Plot Setup.   
The NCVS method was chosen as the vegetation sampling method for this 
study.  The flexibility and reliability of the method were primary motivators in this 
decision, as well as the established familiarity of the field samplers with the method. 
 
2.2.2 Water Quality 
Every body of water has an individual pattern of physical and chemical 
characteristics determined largely by the climatic, geomorphologic and geochemical 
conditions prevailing in the drainage basin and the underlying aquifer (Chapman, 
2006).  Certain variables can provide a strong picture of water quality at a particular 
site and can act as key indicators to the overall health of a wetland.  Total quantities 
of dissolved solids indicate the condition of water bodies of a similar nature.  Mineral 
content, which is determined from the amount of total dissolved solids, is also an 
essential feature of water quality that results from the balance between dissolution 
 
 23 
and precipitation.  Minerals and nutrients are necessary for wetlands to thrive, but an 
excess of nutrients can prove harmful.  Excessive nutrient inputs from sewage 
effluent, agriculture, or internal loading caused by fish foraging and excretion can 
destroy wetland vegetation (Wersal, 2006).  Thus, the specific level of nutrients can 
indicate much about the health of a wetland.  Lastly, oxygen content influences the 
solubility of metals and the presence of oxygen is necessary for many forms of 
biological life.  Hence, dissolved oxygen is also a vital factor to consider when 
addressing the health of any water body.   
Hydrological processes are important factors in making determinations about 
vegetation communities, wildlife habitat, nutrient cycling, and other wetland 
functions (Mitsch & Gosselink, 2000).  However, obtaining accurate water quality 
data is typically both expensive and time-consuming; therefore, water quality data for 
wetlands is scarce and only known for a few scattered wetlands in any one area 
(Kusler, 1998).   If it is not feasible to assess wetlands through extensive sampling, 
then a quicker assessment method is needed.  One of the MDE?s long term goals is to 
use this research study?s data to develop an improved rapid assessment method.  This 
rapid assessment method would only incorporate data that is relatively easy and quick 
to obtain in order to assess the health of the wetland.   
Water bodies can be classified by their water quality characteristics.  It is 
often easy to identify a certain body of water as a wetland, but in some environments 
(especially where wetlands and upland areas converge) it can be difficult to 
distinguish the boundaries of a wetland.  Some bodies of water may exhibit unusual 
water quality data and would be difficult to classify, but in general, water quality data 
 
 24 
is an accurate indicator of water body type.  It is also possible to further classify a 
wetland as tidal, non-tidal, mitigation, etc.  Researchers at the University of Michigan 
have devised a hydro geomorphic (HGM) wetland assessment method that can assess 
the functional condition of a specific wetland on the basis of a range of physical 
conditions.  HGM groups wetlands into seven different wetland classes and provides 
an ecologically-sound means for classifying, assessing, and comparing wetland 
hydrodynamics and related functions (Merkey, 2006).   
Once a body of water is classified as a wetland, it can be compared to other 
existing wetlands.  Water quality data, such as mineral content, can be analyzed and 
compared to data from wetlands that are already known to be healthy or unhealthy.  
Certain existing standards, including beneficial use (e.g., drinking, swimming), 
numeric (e.g., allowable concentrations of pollutants) and narrative components (e.g., 
unacceptable surface conditions), have been developed by governmental and 
environmental agencies.  These can be used to judge water quality data and to draw 
comparisons between bodies and among restoration and purification plans.   
In selecting this project?s water quality variables, we placed a high importance 
on the levels of nitrogen compounds present.  Plants and micro-organisms are 
constantly converting inorganic nitrogen to organic forms and thus the cycling of 
nitrogen is necessary for all living organisms.  Inorganic nitrogen occurs in a range of 
oxidation states and based on the different levels of nitrogen compounds one can 
determine in which stage of the nitrogen cycle the environment is operating 
(Chapman, 1996).   
 
 25 
The nitrogen compound variables that we decided to monitor were: Ammonia 
(NH3), Nitrite (NO2-), Nitrate (NO3-) + Nitrite, and Total Nitrogen (N).  Since we 
did not have the capability to test for these compounds ourselves, we decided to send 
the water samples collected to Appalachian Lab for chemical analysis, where they 
could also be quickly tested for Total Phosphorous.  Ammonia occurs naturally in 
wetlands and results from the breakdown of nitrogenous matter in the water.  It is also 
sometimes discharged into the wetlands through industrial waste.  High 
concentrations of ammonia are toxic to aquatic life given the pH level.  Unpolluted 
waters contain small amounts of ammonia and ammonia compounds, usually <0.1 mg 
l-1 as nitrogen.  Total ammonia concentrations measured in surface waters are 
typically less than 0.2 mg l-1 N but may reach 2-3 mg l-1 N (Chapman, 1996).  
Higher concentrations are suggestive of some type of organic pollution, thus 
ammonia was selected as a variable to indicate organic pollution.   
Nitrate (NO3-) is the main form of combined nitrogen found in wetlands, and 
can be reduced to nitrite (NO2-) by the denitrification process.  Nitrate has many 
natural sources, including plant and animal debris, and is an essential nutrient for 
wetland plants.  The level of nitrate plus nitrite in surface water gives an indication of 
both the nutrient status and the level of organic pollution.  Because of this, a 
combined quantification of nitrate and nitrite are included in almost all water quality 
surveys, especially background monitoring programs such as ours.  When influenced 
by human activities, surface waters can have nitrate concentrations up to 5 mg l-1 
NO3--N, but often less than 1 mg l-1 NO3-N (Chapman, 1996).  Concentrations in 
excess of 5 mg l-1 NO3--N usually indicate pollution by human or animal waste or 
 
 26 
fertilizer run-off.  In cases of extreme pollution, concentrations may reach 200 mg l-1 
NO3--N (Chapman, 1996).  The World Health Organization (WHO) recommended 
that the maximum limit for NO3- in drinking water should be 50 mg l-1 (or 11.3 mg 
l-1 as NO3-N) (Chapman, ed., 1996), and classified waters with higher concentrations 
as representing a significant health risk.  Nitrite concentrations in freshwaters are 
usually very low by comparison, 0.001 mg l-1 NO2--N, and rarely higher than 1 mg l-
 1 NO2--N (Chapman, 1996).  Total nitrogen levels, while subject to seasonal 
fluctuations, are used as a general indicator of pollution. 
Phosphorous is also a necessary nutrient for living organisms.  It is usually the 
limiting nutrient for algal growth.  Artificial increases in phosphorous levels due to 
unnatural activity are the leading cause of eutrophication (Chapman, 1996).  The 
weathering of phosphorus-bearing rocks and the decomposition of organic matter are 
the principal sources of phosphorous.  Because phosphorous plays such an 
instrumental role in the biological cycle, it is also included in almost all basic water 
quality surveys and monitoring programs.  As stated earlier, the presence of high 
phosphorous concentrations are largely responsible for eutrophic conditions and, like 
nitrogen, can also indicate the presence of pollution.  One must maintain an accurate 
knowledge of phosphorous levels in order to sufficiently manage a wetland site.  
Although there can be considerable seasonal fluctuations, in most natural surface 
waters phosphorus ranges from 0.005 to0.020 mg l-1 PO43--P.  Concentrations as 
low as 0.001 mg l-1 PO43--P may be found in some pristine waters and as high as 
200 mg l-1 PO43--P in some enclosed saline waters.  Average groundwater levels are 
about 0.02 mg l-1 PO43--P (Chapman, 1996).   
 
 27 
Dissolved oxygen (DO) readings of surface water are also useful in evaluating 
the health of a wetland.  The organisms that are responsible for self-purification 
processes are dependent upon oxygen.   Dissolved oxygen readings depend on both 
temperature and the level of biological activity present.  DO percentages can and do 
change, both over a seasonal basis and on a daily basis.  In still waters, pockets of 
both high and low concentrations of DO can be found in close proximity to one 
another.   Since oxygen is involved in a majority of chemical and biological 
processes, DO is also a very common variable used in water quality assessments.  In 
fresh-waters dissolved oxygen (DO) at sea level ranges from 15 mg l-1 at 0? C to 8 
mg l-1 at 25? C.  Concentrations in unpolluted waters are usually close to, but less 
than, 10 mg l-1 (Chapman, 1996).  Concentrations below 5 mg l-1 may adversely 
affect the functioning and survival of biological communities and below 2 mg l-1 
may lead to the death of most fish.   
Finally, the temperature of the water must be taken into consideration when 
evaluating all other variables.  Water temperature, especially surface water 
temperature, can fluctuate greatly because it is influenced by so many factors, 
including season, time of day, and cloud cover.  Water temperature affects many 
chemical, physical, and biological processes and thus effects the concentrations of 
many other variables (Chapman, 1996).  As temperature increases, the speed of 
chemical reactions also increases and the solubility of gas in the water decreases.  
Water temperature influences plant performance, especially photosynthetic rates 
(Pilon & Santamaria 2002), however, extremely high water temperatures can reduced 
photosynthetic rates and have a negative effect on the ecosystem (Spencer, 1986).  
 
 28 
Also, higher temperatures can lead to increased respiration rates and thus increased 
oxygen consumption.  So, when evaluating water quality data, unusually low DO 
readings must always be evaluated with respect to temperature in order to accurately 
analyze its context.  Groundwater, such as those samples collected from lysimeters, 
maintain a relatively constant temperature somewhere close to the mean annual air 
temperature. 
 
2.2.3 Soil 
The Federal Manual for Identifying and Delineating Jurisdictional Wetlands 
was published in 1987 with the help from four agencies: the USFWS, the EPA, and 
the United States Army Corps of Engineers, and the Soil Conservation Service.  
According to this manual, in order for an ecosystem to be classified as a wetland it 
must meet three criteria: it must possess hydrophytic vegetation, wetland hydrology, 
and hydric soils, thus establishing the importance of soils in the classification of 
wetlands (Lilly, 1993).  In most ecosystems, soil plays an important role in overall 
function; however, a special condition exists in wetlands because the soil is 
frequently saturated with water.  It is not uncommon to find a layer of standing water 
at most wetland sites throughout the growing season (USDA, 1998).  As the ground 
becomes saturated with water, water molecules begin to fill pore space once occupied 
by air, thus depriving the soil of oxygen.  As water fills the gaps found between the 
soil peds, the rate of oxygen diffusion through the soil is greatly reduced.  The 
reduced rate of oxygen circulation in the soil results in anaerobic conditions.  It is 
these anaerobic conditions that are responsible for the presence of wetland vegetation.  
 
 29 
When there is a lack of oxygen, other chemicals must be used in the chemical 
transformation process, including: nitrogen, manganese, sulfur, and iron.  The 
transformation reactions of these chemicals enable wetlands to act as sinks and 
filtration mechanisms for nitrogen and sulfur pollution.   
An important equation in determining the existence of a hydric soil is the 
Nernst equation.  Nernst equation: EH = E0 + 2.3[RT/nF]log[{ox}/{red}], is the 
equation used to calculate the redox potential of a soil (Mausbach et al, 1994).  Redox 
potential, also referred to as oxidation-reduction potential, is a measurement used to 
quantify the electrochemical reduction of wetland soils.  With an abundance of 
oxygen in the soil, the redox potential should be between +400 and +700 mV; 
however in an anaerobic soil, such as those found in wetlands, the redox potential 
fluctuates between ?400 and +400 mV.  As oxygen becomes unavailable, organic 
substrates look to donate electrons to substances other than oxygen, thus lowering the 
redox potential value.  The first terminal acceptors of electrons in anaerobic soils are 
nitrates (NO3
 -) (Mausbach et al, 1994). 
Nitrates are often found in wetland soils and are reduced to nitrites (NO2
 -) and 
ultimately to N2O and N2.  Nitrogen is often identified as one of the most limiting 
nutrients of wetland soils.  Fertilizers used by humans add a great deal of unnatural 
nitrates to the nitrogen cycle.  If anaerobic soils were not able to process out this 
excess nitrate, then pollution would present an even bigger challenge.  Through the 
process of denitrification, excess nitrates in the soil lithosphere are converted into 
gaseous forms of nitrogen, specifically N2 and N2O.  Nitrogen often enters the soil in 
the form of nitrates or ammonium (NH4
 +), from either crop residues or fertilizers.  
 
 30 
When there is a deficiency of oxygen, symbiotic microorganisms in the soil convert 
the NO3
 - into nitrogen oxide and nitrogen gases, which are able to escape into the 
atmosphere (Mausbach et al, 1994).  Plants are also responsible for the uptake of a 
large amount of nitrogen from the soil in the form of nitrates.  Nitrates are a pollution 
concern because they easily leach from the soil and pollute our waterways (Reppert et 
al, 1979). 
Iron and manganese are also important nutrients for they are also reduced in 
chemical transformations when oxygen is absent.  Typically, iron is found in its 
oxidized state as Fe(OH)3.  Oxidized iron is easily identifiable because the soil often 
has a reddish or brown hue.  Soils in wetlands are often identified with a grayish color 
because the iron in these soils is in its reduced form (Fe(OH)2).  A depleted or 
reduced matrix in a soil profile occurs when iron is either removed or reduced by 
chemical transformations.  The reduced forms of both iron and manganese are 
mobile; therefore, they have a tendency to accumulate in pore spaces.  Reduced iron 
and manganese are stripped from the soil peds, leaving behind the dull gray color of 
the mineral matrix (Soil Survey Staff, 1999).   
Often during the growing season the water table fluctuates.  This fluctuation 
of the water table can be identified by the presence of substances known as mottles.  
Root channels and macropores often contain sources of oxygen in anaerobic soils.  
Root channels, often referred to as oxidized rhizospheres, have available oxygen 
because oxygen escapes from the roots of plants into the surrounding root channel.  
Once this oxygen contacts the soil, it is able to transform iron and manganese back 
 
 31 
into their oxidized states.  This condition can be identified by the presence of reddish 
soil surrounding the root channels of plants (Richardson, et al, 2000). 
A final terminal electron receptor in the electron chain is sulfate (SO4 
2-).  
Sulfate reduction in wetland soils is often caused by the presence of microorganisms.  
Bacteria are able to convert sulfates into hydrogen sulfide, by using sulfur as a 
terminal electron receptor in anaerobic respiration.  The release of this hydrogen 
sulfide gas is responsible for the rotten egg odor which is prevalent at wetland sites 
(Mausbach et al, 1994). 
Another important characteristic of wetland soils is the presence of a large 
layer of organic matter near the surface of the soil profile.  This organic matter, often 
found in the form of peat or muck, plays an important role as an energy source for the 
many chemical transformations which occur in wetland soils.  Dead plants replenish 
the soil with nutrients as they are broken down by the many microorganisms present 
in soil (Richardson, et al, 2000). 
 
2.3 Antibiotics 
2.3.1 Methods of Analysis 
Testing for the presence of antibiotics in the environment encompasses two 
major possibilities: direct analytical testing of drug concentrations and indirect 
microbiological testing of resistance patterns in samples of collected bacteria. 
Direct testing techniques usually involve the use of gas or liquid 
chromatography in combination with mass spectrometry.  Ahrer, Scherwenk, and 
Buchberger (2001) note that until 2001, analysis of chemicals in surface waters was 
 
 32 
mostly conducted using gas chromatography in combination with mass spectrometry.  
However, since then, additional research has been performed exploring other 
technologies, such as the use of capillary electrophoresis with mass spectrometry.  In 
particular, liquid chromatography with tandem mass spectrometry is becoming more 
prevalent because the method is highly specific, enabling better detection of 
compounds.  The technique also allows for the separation and delineation between 
ions of different compounds with the same molecular mass (Fatta et al., 2007).  
Such direct testing methods tend to require sophisticated equipment and an 
involved process of sample preparation.  Through a process known as derivatization, 
water must be pretreated in order to yield optimal results (Ahrer, Scherwenk, & 
Buchberger, 2001).  This procedure may require the use of highly toxic or 
carcinogenic compounds.  The traditional process flow includes filtration, solid phase 
extraction, and/or derivatization prior to the actual sample analysis via liquid or gas 
chromatography and mass spectrometry (Fatta et al., 2007).  
In comparison, microbiological testing of bacterial resistance to antibiotics 
tends to be significantly less involved and less expensive.  Based on the generally 
accepted assumption that levels of antibiotic resistance increase with increasing 
prevalence of antibiotics themselves, testing for antibiotic resistance amongst wetland 
bacteria is thus a more feasible, albeit indirect, method of testing for the impact of 
antibiotics in a wetland.  To our knowledge, all resistance testing techniques measure 
minimum inhibitory concentration (MIC), the lowest concentration of a drug that will 
prevent the growth of bacteria.  In the United States and much of the world, the 
authority on determining standard for bacterial resistance testing is the Clinical and 
 
 33 
Laboratory Standards Institute (CLSI).  Several methods exist to test for antimicrobial 
resistance, the most common of which are the disc diffusion and broth dilution tests.  
The disc diffusion technique is the solid equivalent of the broth dilution 
technique (discussed below) and is often preferred because it enables one to 
simultaneously test for multiple drugs (Lorian, 2005).  Perhaps the most well-
 established method is the Kirby-Bauer disc diffusion assay, developed in 1966.  This 
method involves either plating a ?lawn? of bacteria onto a selective medium, such as 
Mueller Hinton agar, or inoculating the medium with the culture itself, and then 
placing small paper discs with known concentrations of antimicrobials onto the plate.  
After a given period of incubation, usually 18?24 hours, there will likely be an area 
around the disc in which the presence of antibiotic has prevented bacterial growth.  
The diameter of this so-called ?zone of inhibition? is measured, and bacterial 
sensitivity or resistance to the antibiotic is determined by comparing this diameter to 
an established standard.  Smaller diameters indicate higher concentrations of 
antibiotic required to inhibit bacterial growth, and thus indicate higher levels of 
bacterial resistance (Bauer et al., 1966).  A similar method is the gradient method.  
Commercially known as Etest, this technique uses plastic strips preloaded with 
antibiotics at various concentrations.  Numerous strips are placed on a plate in a 
spoke-like fashion, and an ellipse-shaped clearing around the strip results after 
incubation.  The MIC is determined by the intersection of the ellipse with the test 
strip (AB BIODISK).  
Often, disc diffusion techniques yield poor categorizations for large 
macromolecules such as vancomycin.  Because these drugs have large molecular 
 
 34 
weights, they take longer to diffuse into agar growth media, resulting in smaller 
differences in the sizes of the zones of inhibition.  This makes it difficult to 
differentiate between resistant and susceptible bacteria.  To overcome this problem, 
prediffusion methods have been developed. For example, Neo-Sensitabs are dry 
crystalline antimicrobial tablets that enhance susceptibility profiles (Katz, Luperchio, 
& Thorne, in press).  
The broth dilution, or microdilution, test is another means of testing for 
antibacterial resistance.  Various dilutions of a given drug are prepared (usually a 
minimum of ten) and loaded into microwells, inoculated with bacteria, and incubated.  
Chemical indicators are usually added to aid in the visualization of bacterial growth. 
After incubation, the turbidity and MIC are determined (Hyman et al., 2002).  
Susceptibility is evaluated by comparing these measurements with established 
standards.  Additionally, commercial advances have made plates preloaded with 
antibiotics available.  
Finally, efforts have been made to automate the testing process.  One such 
method is the Cobas-Bact technique developed by Roche Diagnostics.  This technique 
evaluates antimicrobial susceptibility in less than five hours.  After several rounds of 
incubation and centrifugation, turbidity of the inoculum is measured, and the system 
computer determines resistance based on growth of the bacteria.  However, a study 
conducted by Murray, Niles, and Heeren (1987) found discrepancies when this 
method was compared to traditional disc diffusion and broth dilution techniques.  
In this study, we elected to use the Kirby-Bauer disc diffusion method. 
Previous studies (Drew et al., 1972; Dornbusch et al., 1975; Gaudreau & Gilbert, 
 
 35 
1997) have repeatedly shown that when performed correctly this method remains a 
simple and inexpensive, yet reliable, means of testing for antibiotic resistance.  The 
issue of poor susceptibility profiles for vancomycin, one of the antibiotics we 
selected, was not a problem for us, since we were using the drug as a negative 
control. In addition, we possessed the facilities and equipment required to use this 
technique. Conversely, we did not have access to the specialized equipment necessary 
for direct testing or broth dilution. Escherichia coli was selected as a model organism 
due to its prevalence in the environment and the ease of isolating and testing for this 
bacterium. 
 
2.3.2 Selection of Antibiotics 
Six antibiotics were selected for evaluation in this study: erythromycin, 
ciprofloxacin, ampicillin, sulfisoxazole, tetracycline, and vancomycin.  Based on our 
review of previous literature, these antibiotics are among those that are most 
commonly tested.  In addition, each antibiotic is representative of a different class of 
drugs.  Vancomycin was selected as a negative control.  
 
2.3.3 A Brief Background of Antibiotics Chosen for This Study 
Erythromycin 
Erythromycin is a macrolide drug that is often used to treat acne, strep throat, 
syphilis, and other infections caused by bacteria.  With a range of activity that is 
slightly broader than that of penicillin, erythromycin is a common alternative for 
individuals who are allergic to penicillin.  The drug possesses bactericidal properties, 
 
 36 
meaning that it has the ability to kill bacteria versus simply inhibiting bacterial 
growth, but the exact mechanism of action is still not fully understood.  The 
prevailing theory is that erythromycin binds to a subunit of bacterial ribosome and 
inhibits protein synthesis. 
 
Ciprofloxacin 
Ciprofloxacin is a broad spectrum antibiotic that is regularly used to treat 
various infections.  It belongs to a group of bactericidal compounds known as 
fluoroquinolones.  Furthermore, ciprofloxacin affects both Gram-positive and Gram-
 negative bacteria by targeting the enzymes topoisomerase IV and DNA gyrase, 
respectively.  Both of these enzymes are essential for bacterial DNA replication.  By 
inhibiting DNA replication and transcription, ciprofloxacin leads to chromosomal 
breaks and eventual death of the cell.  Neither of the two enzymes is present in 
eukaryotic cells, which is why ciprofloxacin is safe for human use.   
 
Ampicillin 
Ampicillin belongs to the aminopenicillin family, within the broader class of 
drugs known as beta-lactam antibiotics.  Beta-lactam drugs are bactericidal and were 
initially thought to only affect Gram-positive bacteria.  However, recent 
developments indicate that they are effective against various strains of Gram-negative 
bacteria as well.  Ampicillin acts by interrupting synthesis of the peptidoglycan layer 
of the cell wall.  By competitively inhibiting the enzyme transpeptidase ? a penicillin-
 
 37 
binding protein (PBP) ? ampicillin disrupts the final step of peptidoglycan synthesis 
and eventually leads to cell lysis.   
Two main mechanisms of bacterial resistance toward beta-lactam drugs exist.  
This class of drugs is characterized by what is known as a beta-lactam ring, which 
plays a critical role in drug interactions with PBPs.  If this ring is not intact, it could 
lead to bacterial resistance of the drug.  Bacteria that are able to produce enzymes 
such as beta-lactamase and penicillinase are able to hydrolyze this ring and disrupt its 
structure and overall effectiveness.  Resistance also emerges if the PBPs are altered in 
some way, making it difficult for the drug to bind.  This is seen in infections such as 
methicillin-resistant Staphylococcus aureus (MRSA).   
 
Sulfisoxazole  
Sulfisoxazole is one of several sulfonamide-based drugs, or sulfa drugs.  
Though some of these drugs do not possess antibacterial properties, sulfisoxazole 
does have bacteriostatic activity, meaning that it can inhibit growth, against both 
Gram-positive and Gram-negative bacteria.  The compound acts as a competitive 
inhibitor of an enzyme critical to folate synthesis.  Folate is essential for DNA and 
RNA synthesis, and thus sulfisoxazole effectively hinders cell division.  The drug is 
only effective versus bacterial cells, because mammals do not produce folate; instead, 
it is a dietary requirement.   
 
 
 38 
Tetracycline 
Tetracyclines are broad spectrum antibiotics that are often used to treat acne 
and ulcers, in addition to having several dental applications.  This class of drugs 
works by binding to the bacterial ribosome, thereby inhibiting protein production and 
subsequent growth.  Mechanisms of resistance towards tetracyclines include 
inactivation via enzymes or through the production of proteins that effectively pump 
the drug out of the cell.  Another method is that of ribosomal protection, in which a 
resistance gene encodes a protein, which performs one of many functions in order to 
defend the ribosome, including blocking the binding of or dislodging already bound 
tetracycline.   
 
Vancomycin 
Vancomycin is primarily effective against Gram-positive bacteria, for it acts 
by inhibiting cell wall synthesis; the outer membrane of Gram-negative bacteria 
impedes the molecule?s entry into the cell, and therefore vancomycin is unable to 
impact the cell wall.  Over the years, the drug gained popularity due to the fact that 
staphylococci had difficulty gaining resistance towards it, as opposed to their rapid 
development of resistance towards penicillin.  However, vancomycin is for the most 
part used only as a drug of last resort due to its strength and nature of side effects.  In 
our study, vancomycin was chosen as a negative control due to its activity and 
potency.   
 
 
 39 
2.3.4 Prevalence of Antibiotic Resistance 
Antibiotic resistance of bacteria is a growing problem, particularly in the 
hospital setting.  There has also been concern regarding vulnerability of patients in 
related settings, such as nursing homes (Weiner et al., 1999).  According to the World 
Health Organization, up to 60 percent of nosocomial, or hospital-acquired, infections 
in the United States are caused by drug-resistant bacteria.  In addition, these microbes 
could potentially be resistant to as many as 10 different antibiotics (World Health 
Organization, 1996).  Part of the problem stems from the fact that while it could take 
up to decades to develop a drug, these same drugs may not be effective for as long a 
period of time due to the quick nature of resistance transfer amongst bacteria.  
Additionally, Clark, Patterson, and Lynch (2003) note that heightened use of broad 
spectrum antibiotics such as the beta-lactamases has contributed to multidrug 
resistance.  On the other end of the spectrum is the fact that sanitation and other 
asceptic standards must be maintained; this could be an issue in third-world nations. 
While it is generally accepted that there is an increasing prevalence of 
antibiotic-resistant bacteria in hospitals, it is difficult to quantify this increase because 
of the lack of consistency of trends in all hospitals (Fridkin et al, 2002).  A majority 
of the studies to date have focused on intensive care units (ICUs); due to the quick 
pace required in an ICU, there may not be adequate time to follow proper sanitation 
techniques, and there is likely an increased risk of spreading resistance (Fridkin & 
Gaynes, 1999).  Reports from the Center for Disease Control (CDC) reaffirm this 
notion and assert that patients receiving treatment in ICUs are at an increased risk for 
nosocomial infections including pneumonia, urinary tract infection, and other 
 
 40 
bloodstream infections (Fridkin & Gaynes, 1999).  Additionally, Vincent et al. (1995) 
found that over 20 percent of patients admitted to ICUs in Western European 
hospitals developed ICU-acquired infections.   
Ampicillin, tetracycline, and ciprofloxacin were among the major antibiotics 
that we encountered in our survey of prior studies of resistance in hospitals.  Estrada-
 Garc?a et al. (2005) analyzed isolates of E. coli from children in Mexico that had been 
hospitalized for diarrhea.  Among these diarrheogenic isolates, it was found that 73 
percent were resistant to ampicillin.  In another study, Oteo et al. (2005) surveyed 32 
Spanish hospitals and found that among the 7,098 invasive E. coli isolates, 59.9 
percent were resistant to ampicillin.  Conversely, resistance of E. coli or other gram-
 negative species to ciprofloxacin was relatively low in most of the studies reviewed, 
ranging from 19.3 percent (Oteo et al., 2005) to approximately 40 percent (Aksaray et 
al., 2000).  However, it has been documented that resistance towards the drug is 
steadily increasing.  A survey by Neuhauser et al. (2003) noted a decrease in the 
susceptibility of gram-negative bacteria towards ciprofloxacin from 86 percent in 
1994 to 76 percent in 2000 ? that is, an increase in resistance from 14 percent to 24 
percent. 
Moving from the hospital to the community setting, Bartoloni et al. (2006) 
examined the occurrence of antibiotic-resistant E. coli in the feces of children living 
in urban neighborhoods of Peru and Bolivia.  Once again, highest resistance was seen 
towards ampicillin (95 percent) and tetracycline (93 percent).  In addition, a random 
sample of these isolates was selected and tested for multiple drug resistance patterns 
(i.e. combinations of drugs).  Ninety percent of these isolates exhibited multidrug 
 
 41 
resistance, and two of the three most common combinations included ampicillin and 
tetracycline.  Estrada- Garc?a et al. (2005) reported slightly lower multiple resistance 
results ? approximately 58 percent of samples were resistant to three or more drugs.  
Nevertheless, these statistics highlight the importance of studying antibiotic resistance 
in conditions other than those found in hospitals. 
Some research has been conducted reviewing bacterial resistance patterns in 
the natural environment.  For example, Zuccato et al. (2000) note that drugs such as 
erythromycin usually have long half-lives and are not biodegradable, and as a result 
their presence in the environment often persists, sometimes for over a year.  In fact, 
erythromycin has been detected in surface waters more frequently than other 
antibiotics; however, these levels have been noted to be below those which would 
foster resistance bacteria (Summers, 2002).  
Widespread resistance towards ciprofloxacin has been seen, despite its label 
as a ?drug of last resort.?  Initially, practitioners were cautious about using the drug 
due to its broad spectrum activity and potent nature, but this is no longer the case.  
Additionally, analogues of the drug have been employed much more in agriculture 
and farming.  Until September 2005, enrofloxacin (also known as Baytril), another 
drug in the fluoroquinolone family, was commonly used in poultry to prevent illness 
and to boost growth.  Studies have found that in the ten year period following the 
drug?s introduction in 1995, enrofloxacin resulted in increased bacterial resistance to 
ciprofloxacin by approximately 21 percent (Truant, 2005).  In another study, 
antimicrobial resistance patterns of Neisseria gonorrhoeae were studied over a four 
year period in Korea.  The investigators found that ciprofloxacin resistant isolates 
 
 42 
increased dramatically after the recommendation to use the drug for therapeutic 
purposes.  Specifically, the percent of resistant isolates increased from just 1 percent 
in 1999 to 48.8 percent in 2002 (Yoo et al., 2004). 
The extensive use of ampicillin has lead to widespread resistance, particularly 
in the hospital setting as noted above.  However, resistance is prevalent in the 
environment as well.  A study conducted in 2002 that surveyed 15 U.S. rivers found 
that ampicillin was ineffective against up to 50 percent of the bacteria sampled (Ash 
et al., 2002).  In addition, no strong pattern could be found regarding the resistant 
isolates, as they were found in rural as well as urban areas (Raloff, 1999). 
Current research indicates that resistance to sulfisoxazole is not extremely 
prevalent.  In a study conducted by Sayah et al. (2005) investigating the antimicrobial 
resistance patterns of E. coli in various environments, only a 13.3 percent resistance 
among isolates was noted.  Similarly, McKeon et al. examined over 250 bacterial 
isolates (E. coli included) from rural water sources in West Virginia, and detected less 
than 10.0 percent resistance to sulfisoxazole (McKeon et al., 1995). 
Research shows that there is indeed a concern over the prevalence of 
tetracycline-resistant bacteria in the environment.  Tetracycline is a common drug 
used in farming and swine cultivation, and in general, links have been suggested 
between such use and drug-resistant infections in humans (Chee-Sanford et al., 2001).  
In the general environment, Sayah et al. (2005) found a 27.3 percent resistance to 
tetracycline among isolates, while McKeon et al. (1995) noted 32.3 percent 
resistance.  In addition, Chee-Sanford et al. (2001) conducted a study to assess the 
prevalence of tetracycline-resistant genes in lagoons and groundwater surrounding 
 
 43 
two swine facilities.  Published in 2001, their results confirmed that all classes of 
genes conferring ribosomal protection were present in bacterial isolates.  These 
results were significant, since groundwater is a major constituent of drinking water, 
suggesting that despite intense processing, resistant bacteria have the potential to 
make their way into our everyday water sources. 
Though vancomycin is an incredibly strong drug, resistance towards it has 
been an emerging problem, particularly in the hospital setting since it is used as a 
?last line of defense.?  Vancomycin resistant enterococci (VRE) were first confirmed 
in the mid-1980?s, and since then, multi-drug resistant VRE have also been seen 
(Rice, 2001).  In addition, VRE have been isolated from both hospital and residential 
wastewater environments (Harwood et al, 2001). 
Our high level survey seems to indicate that bacterial resistance seen towards 
antibiotics in the natural setting is comparable in that seen in the hospital 
environment. 
 
 
 44 
3. Methods 
 
3.1 Site Selection 
Team CRABS studied thirteen wetland sites around Maryland, selected in 
cooperation with MDE.  All wetlands were non-tidal, mitigation sites.  These sites 
were selected to be representative of the each type of mitigation wetland present 
around the state, including emergent, scrub-shrub, and forested wetlands, but they 
were also selected on the basis of convenience and feasibility of obtaining landowner 
permission to visit the sites.  Figure 1 is a map of Maryland showing the location of 
the thirteen sites, and Table 3 gives brief descriptions of these sites. 
 
Figure 1: Map of the 13 Maryland wetland sites used in this study 
 
 
 
 
 45 
Table 3: Brief descriptions of the 13 wetland sited used in this study 
Code Name Brief description 
AUD Aud Small densely vegetated and wet site near the St. Mary?s River, 
just downhill from a horse farm 
BEE Beehive Small dug out site in Howard County in a residential 
neighborhood.  Train tracks run on the edge. 
BRY Bryantown Large site located just off of Route 5 in Charles County in a 
mainly agricultural and residential area 
CAL Calvert Very small site located on the edge on Calvert County just off of 
Route 4 at the Route 260 overpass.  Very wet. 
CBL Cumberland Small site nestled in the Appalachian Mountain foothills, just 
north of Interstate 70 
CCW Waldorf Charles County site a few miles west of Waldorf, MD.  Small 
yet diverse, shows elements of forests, lakes, and plains 
HCP Herring Creek Park Site in West Ocean City dominated by phragmites, very wet, 
only site located in Coastal Bays Watershed 
INC Irvine Nature Center Former farmland in central Baltimore County, allowed to 
naturally transform into a wetland 
JLP Jackson Lane 
Preserve 
Large site near Maryland/Delaware border operated by The 
Nature Conservancy.  Fairly flat with several ponds. 
KIN Kinder Usually dry site within 5 miles of the Chesapeake Bay.  Heavily 
covered by low growing sedges and rushes.  
MRK Merkle WMA Site just off of the Patuxent River.  Constructed wetland which is 
part of a much larger wildlife management area.  
PSC Piscataway Stream 
Valley Park 
Located near Route 301, east of Route 5.  Large ponds in a 
formerly agricultural area. 
SHK Shockley Densely forested site in Snow Hill in Worcester County.  Areas 
classified as wetlands usually contain long, narrow ponds. 
 
 
3.2 Methods of Data Collection 
3.2.1 Land Use 
The very first step in our research was to analyze the land surrounding each of our 
thirteen sites.  Following the example of previous studies, it was important to begin 
with a land use study as it would help form the framework for the rest of our research.   
The standard tool used in studying land use is Geographical Information Systems, 
GIS.  According to Environmental Systems Research Institute (ERSI), ?A geographic 
 
 46 
information system (GIS) integrates hardware, software, and data for capturing, 
managing, analyzing, and displaying all forms of geographically referenced 
information.? (ERSI).  Therefore, before we conduct our study we had to determine 
what GIS software to use, what type of data to examine, and where we would draw 
this data from. 
 In performing a GIS study, a researcher could chose from a wide variety of a 
data types including: watershed data, soil series data, average rainfall, transportation 
systems, elevation, sewage transport, roadways, population density, land use 
classification, and more.  For purposes of our study we wanted to focus specifically 
on the land use surrounding our sites.  
The software we chose to use was a program entitled GISHydro2000.  
GISHydro2000 was developed by the University of Maryland?s Department of Civil 
and Environmental Engineering, in collaboration with the Maryland State Highway 
Administration in 1997.  The software and its databases are updated by Dr. Glen 
Moglen, a professor from the University of Maryland?s Civil and Environmental 
Engineering department.  The program combines a GIS platform (ArcView3.x) with a 
database including land use, soils, drainage areas, watershed, channel delineation, 
peak discharge estimates data.  Through using this program, we were able to extract 
relevant land use data from Maryland Office of Planning, 2002 into the ArcView3.x 
software and perform a data analysis using program tools. 
After identifying surrounding land use classifications as the object of analysis, 
we first had to establish buffer zones around each site.  We decided to examine the 
surrounding land use for both a 1000m and 2000m circular buffer around each site.  
 
 47 
To create the buffers we first had to identify a center point of reference for each 
wetland site.  These site reference points were identified with the help of Google 
Earth and GPS coordinates taken on site.  The reference points were chosen for their 
close proximity to the center of each wetland site.  The coordinates of the center 
points were recorded so they could later be imported into GISHyrdo2000.  After 
uploading the 13 reference points (corresponding to our thirteen sites) to 
GISHyrdo2000, we drew a 1000m circular buffer around each point with ArcView 
tools.  Later 2000m circular buffers were also drawn, as shown in Figure 2. These 
buffers were all saved on a shape file on the program database so they could be 
reopened for later analysis. 
 
Figure 2: Screenshot from GISHydro2000, showing land use classifications by color, 
as well as 1000m and 2000m circular buffers around wetland sites 
 
 
 
 48 
The next step was to import land use data to our ArcView program file 
through use of the GISHyrdo2000 database.  We chose to use land use data from the 
Maryland Office of Planning database (2002 version).  Using program tools, we 
separately calculated the land area of each buffer which pertained to every land use 
classification.  For example, for both the 1000m and 2000m buffers surrounding 
Irvine site, we obtained square area values for low residential property, industry, 
pasture, brush, cropland, etc.  The 29 different land use classifications can be seen in 
Table 4.  After calculating land use areas for all thirteen sites for both buffers, we 
exported this data into a two excel spreadsheets for further statistical analysis. 
 
Table 4: The 29 different land use classifications, sorted by color 
Land Use Classification 
Color 
Code 
 LDI 
Values 
Urban Build-up     
Residential - Low Density   20.51 
Residential - Medium 
Density   26.5 
Residential - High Density   29.5 
Commercial   30.57 
Industrial   32.2 
Institutional   30.57 
Extractive   35.51 
Open Urban Land   5.65 
Agriculture     
Cropland   4.99 
Pasture   2.09 
Orchards   6.45 
Row Crops   9.11 
Forest     
Deciduous     
Evergreen     
Mixed     
Brush     
Water     
Wetlands     
Barren Land     
Beaches     
Bare Exposed Rock     
 
 49 
Bare Ground     
Large Lot Agriculture     
Large Lot Forest     
Feeding Operations   20.21 
Agriculture Buildings   20.21 
      
 
 
In Microsoft excel, all land use classification categories were assigned to one 
of the following groups: agricultural, urban, natural.  The breakdown of classification 
was as follows: 
 
Table 5: 29 land use classifications condensed into three classifications to better suit 
CRABS? study 
Condensed Classification Original Classifications 
Agricultural Agriculture, Cropland, Pasture, Orchards, Row Crops, 
Large Lot Agriculture, Feeding Operations, Agriculture 
Buildings 
Urban Urban Build-up, Residential (Low, Medium, High), 
Commercial, Industrial, Institutional, Extractive 
Natural Forest, Deciduous, Evergreen, Mixed, Brush, Water, 
Wetlands, Barren Land, Beaches, Bare Exposed Rock, 
Bare Ground 
 
 
Following these group delineations, land use areas were aggregated in order to 
establish overall land use data for Agricultural, Urban, and Natural.  These values 
would later be used in calculating Landscape Development Intensity (LDI) values for 
all thirteen sites. 
Landscape Development Intensity (LDI) Index is a measure commonly used 
in analyzing the impact surrounding land use has on a site of study.  For our research, 
LDI was the most important factor in determining the impact land use had on the 
location of our 13 sites.  LDI is a weighted average calculation using both an index 
 
 50 
coefficient and a percentage.  Every land use is assigned a value (the coefficient).  
From our study using GISHydro2000 we were able to determine the percentage of 
each land use found in the buffers surrounding all thirteen sites.  LDI values were 
calculated for both the 1000m buffer and 2000m buffer.   
 
 Example of LDI Calculation: 
  
 LDI = (% Residential Low Density) (LDI Value) + (% Cropland) (LDI Value) + (%  
Evergreen) (LDI Value) + (% Mixed) (LDI Value) + (% Brush) (LDI Value) + (% Feeding 
Ops) (LDI value) 
 
LDI = (0.355) (20.51) + (0.134)(4.99) + (0.098) (0) + (0.245) (0) + (0.189) 
(20.21) 
 
 
In addition to calculating LDI values, the thirteen sites were further grouped 
along three broad land use classification categories: Agricultural, Urban, and Natural.  
Sites were grouped according by computing Agricultural:Urban ratios. 
 
Table 6: CRABS? 13 wetland sites organized by land use classification 
Site Classification Wetland Site 
Agricultural Irvine, Kinder, Jackson Lane, Bryantown 
Natural Beehive, Herring Creek, Calvert, Cumberland 
Urban Aud, Piscataway, Waldorf, Merkle, Shockley 
 
 
These land use groups would be used later in portraying results from other areas 
of our study. 
 
 
 
 51 
3.2.2 Field Work 
3.2.2.1 Plot Setup 
At each wetland site, three representative plots were chosen within the 
wetland boundary.  Field researchers picked the starting point with a variety of 
methods.  Sometimes, location was picked to ensure that plots would cover distinct 
areas of the wetlands.  Other times, plots were chosen so as to be sufficiently far apart 
from one another, when size was an issue.  Sometimes, plot location was chosen with 
random methods.  The method for plot setup was modified from the North Carolina 
Vegetation Survey (NCVS) (Peet et al., 1998).  In the NCVS method, schematically 
represented in Figure 3 below, plastic poles are inserted vertically into the ground and 
used to mark off a 20 meter by 50 meter rectangular plot boundary.  This plot was 
divided into ten 10 meter by 10 meter subplots, or ?modules,? numbered as shown in 
Figure 3.  Intensive vegetation sampling is performed in module numbers 2, 3, 8, and 
9. 
 
Figure 3: Diagram of a normal NCVS plot 
 
Circles represent plastic poles used to mark off the plot. 
 
 
 52 
In our modified plot setup method, a 20 meter by 30 meter rectangle was 
marked off with poles, and divided into six 10 meter by 10 meter subplots.  To 
maintain sequential module numbering and for ease of remembrance, the modules 
normally numbered 8, 9, and 10 in the NCVS method were respectively renumbered 
4, 5, and 6 (see Figure 4).  Intensive vegetation sampling was conducted in modules 2 
and 5 of the modified plot. 
 
Figure 4: Schematic diagram showing truncation and renumbering of the standard 
NCVS plot 
 
 
 
 
After the plot was set up, GPS coordinates were taken at the start pole, and an 
azimuth was used to obtain the directional bearing of the plot?s center axis line.  The 
center axis line is the line connecting the start and end poles (see Figure 5). 
We used a modified survey method to allow us to cover more ground with 
similar results.  Using the NCVS 50 by 20 plot size, we would have had four 
intensive modules, rather than two.  However, the team decided that our time could 
 
 53 
be more efficiently spent by surveying three 30 by 20 plots, which would have 
encompassed 6 intensive modules, rather than surveying two 50 by 20 plots, covering 
8 intensive modules.  We felt that gaining information for more plots spread out 
across each wetland outweighed the cost of losing intensive modules adjacent to 
already-sampled intensive modules. 
 
 
Figure 5: Schematic diagram of a completed Team CRABS plot 
 
Start pole and direction of center axis are noted. 
 
 
 54 
3.2.2.2 Vegetation Sampling 
Vegetation sampling was also adapted from the NCVS method (Peet, et al., 
1998) and focused primarily on measuring both presence of plant species and the leaf 
area cover of those species.  The overall procedure involved setting up a series of 
nested squares, called ?depths,? inside the two intensive modules (i.e. modules 2 and 
5).  Working from the smallest to the largest of these nested squares, new plants were 
identified and established as present.  Once presence was established for all plants, 
each species was classified into 1 of 10 different classes of leaf area cover. 
Specifically, the nested depths were set up as shown in and Table 7.  Because 
the smallest depth (depth 5) is 10,000 times smaller than the module as a whole 
(depth 1), this method has the advantage of balancing specificity and detail with 
overall efficiency. 
 
 
 55 
Figure 6: Method of setting up nested depths within intensive module 
 
The largest square in this figure represents the entire module, and is labeled ?depth 
1.?  Each successive depth from 2 through 5 is one-tenth the area of the previous 
depth.  Note that since the depths travel in the direction of two corners, there is only 
one depth 1, but two each of depths 2 through 5.  The bottom left corner is referred to 
as corner 2, and the top right corner is referred to as corner 4, for reasons described 
in (Peet, et al., 1998).  Compare this figure to figure 3d, and note that in both module 
2 and in module 5, the locations of corners 2 and 4 coincide with the placement of 
plastic poles; this was intentional. 
 
 
Table 7: Dimensions of each intensive depth 
Depth Dimensions (m x m) Area (m2) 
5 0.1 x 0.1 0.01 
4 0.32 x 0.32 0.1 
3 1 x 1 1 
2 3.16 x 3.16 10 
1 10 x 10 100 
0* --- --- 
Asterisk denotes the ?overhang? depth, and applies to plants that broke the vertical 
plane of the module but did not actually have a stem inside the module. 
 
 
 
 56 
 
Beginning at depth 5 of corner 2 of intensive module 2 and working outwards, 
all plants were identified and marked as being present in that depth.  After depth 5 of 
corner 2, depths were visited in the following order: depths 4, 3, and 2 of corner 2, 
followed by depths 5 through 2 of corner 4.  Finally, the researchers moved to depth 1 
(the entire module), and marked down any plants not previously found.  If a plant was 
found overhanging the vertical plane of any part of the module, but did not have a 
stem actually inside the module, it was noted as being in ?depth 0.? 
Note that for each corner, plants were only marked down as being in the first 
depth in which they were found, since presence in that depth automatically implied 
presence in all depths containing it.  For example, if a plant was found in depth 5 of 
corner 2, and again in depth 3 of corner 2, it did not need to be marked down twice.  
However, if a plant was found in depth 5 of corner 2 and again in depth 3 of corner 4, 
it did need to be marked down twice, since those two depths do not overlap. 
After this presence sampling was completed, each species was subjectively 
categorized into a leaf area cover class, based on a visual estimate of total percentage 
of the module covered by that species? leaves. Table 8 shows the 10 different leaf 
area cover classes. 
 
 
 57 
Table 8: Leaf area cover classes 
Cover Class Estimated Leaf Area Cover 
1 Trace 
2 0-1% 
3 1-2% 
4 2-5% 
5 5-10% 
6 10-25% 
7 25-50% 
8 50-75% 
9 75-95% 
* 95-100% 
 
 
 
After completing the presence class and cover class procedure for the first 
intensive module, the same procedure was repeated in the second intensive module 
(i.e. module 5).  At this point, presence and cover of plants in these two intensive 
modules were catalogued separately, but had not been considered together.  Thus, to 
consider the plot as a whole, a residual walkthrough was performed, in which all 4 
remaining modules of the entire plot were visited, and any plants not previously 
found were recorded.  Finally, estimates were made of leaf area cover of every 
species found in the plot, based on their leaf area cover of the plot as a whole. 
Since in general each site had three plots, this whole vegetation procedure was 
repeated two additional times at each wetland site.  All vegetation data was recorded 
on standard NCVS data sheets.  Plants were identified using Newcomb's Wildflower 
Guide (Newcomb, 1989), Peterson's Tree and Shrub Guide (Peterson, 1973), and 
Grasses (Brown, 1992), combined with the expertise of the investigators.  Plants that 
 
 58 
could not be identified on site were bagged, labeled, and taken back to a lab for 
identification with the help of University of Maryland botanists. 
 
3.2.2.3 Soil Sampling 
Inside each plot, two cylindrical soil cores were taken at the same time and 
location.  The first soil core was analyzed on site for basic physical characteristics, 
and the second soil core was sent to UMCES Appalachian Laboratory for chemical 
analysis of nutrient content, including total nitrogen, total carbon, carbon/nitrogen 
ratio, and total phosphorous. 
At each plot, the first (on-site) soil profile was extracted to a depth of 50 ? 5 
cm using a steel soil auger.  The soil was carefully transferred to a half-pipe (see 
Figure 3e), where it was laid out and divided into broad color horizons.  The top and 
bottom depths of each horizon were noted.  Individual horizons were qualitatively 
characterized based on organic content, hydric characteristics such as gleying and 
mottling, and soil color.  Soil color was evaluated using a Munsell soil color chart.  In 
addition to the handwritten field notes and drawings of the profile, photographs of 
each soil profile were taken. 
The second soil sample was obtained from within 0.5 meters of the location 
where the first soil sample was taken, using a steel soil auger.  This soil core was 
taken down to a depth where the soil remained uniform for at least 15 cm; this depth 
was noted.  The entire sample was then bagged, labeled, and stored at 4 ?C for 
transport back to a lab where it could be temporarily stored.  Stored soil samples were 
 
 59 
kept refrigerated, and sent to UMCES Appalachian Laboratory within fourteen days 
of original collection. 
 
3.2.2.4 Water Sampling 
Water samples were collected from within each plot for two different types of 
laboratory analysis: (1) chemical nutrient analysis and (2) determination of the 
antibiotic resistance profile of E. coli.  A large set of surface and subsurface water 
samples was collected for chemical nutrient analysis, and a smaller set consisting 
solely of surface water samples was collected for antibiotic resistance 
characterization.   
For the samples collected for nutrient analysis, a consistent number of samples 
was not obtained; the researchers simply attempted to collect both a surface and a 
subsurface water sample as often as possible from every plot.  However, due to 
hydrologic variations between sites this was not always possible; some sites simply 
lacked water.  Thus the number of samples varied between sites. 
Surface water for use in nutrient analysis was collected in non-sterile 
polypropylene bottles, labeled, and chilled to 4? C to be transported back to a base 
laboratory for temporary storage.  Subsurface water samples for use in nutrient 
analysis were collected by use of a SoilMoisture Corp. suction lysimeter, planted with 
its base at a depth of between 45 cm and 55 cm.  After 24 hours, the subsurface water 
that had been pulled into each lysimeter was collected in a non-sterile polypropylene 
bottle, labeled, and chilled to 4? C for transport back to the base lab for temporary 
storage.  All water samples were stored for up to one week in a freezer, before being 
 
 60 
sent to a lab for chemical nutrient analysis (see Section 3.2.3.2: Chemical Analysis of 
Water Samples).   
For the water samples collected for antibiotic resistance testing, a single 
surface sample was collected from within each plot, subject to availability of surface 
water at the time of collection.  In general, between one and three distinct surface 
water samples were collected from each of the thirteen wetland sites. 
Surface samples for E. coli antibiotic resistance analysis were collected using 
sterile NASCO WhirlPak bags.  Samples were then immediately chilled to 4? C to 
arrest bacterial growth, and within 24 hours they were cultured for E. coli and tested 
for antibiotic resistance in the base lab.   
At one site (Kinder), surface water was not available in any plot, so instead a 
sterilized suction lysimeter was used to collect a subsurface sample.  The lysimeter 
was sterilized by soaking it in 6% hydrogen peroxide for 30 minutes, rinsing the 
lysimeter tubing with 30% hydrogen peroxide, and then rinsing all parts of the 
lysimeter with sterile distilled water.  The lysimeter was sealed off and allowed to sit 
for at least 24 hours, so that any residual hydrogen peroxide could degrade.  The 
subsurface water collected in this lysimeter was transferred to a sterile WhirlPak bag, 
and thereafter treated in the same manner as other samples.  Note also that no water 
samples were ever obtained from Bryantown, surface or otherwise. 
Date, time, and location were recorded any time a water sample was taken for 
any purpose during the course of research.  When possible, pH and dissolved oxygen 
(DO) readings were also taken on-site at the same time.  However, due to equipment 
failures during experimentation, pH and DO data are incomplete. 
 
 61 
3.2.3 Lab Work 
3.2.3.1 Chemical Analysis of Soils 
Sediment samples were analyzed for nutrient content at UMCES Appalachian 
Laboratory in Frostburg, MD.  Each sample was analyzed for total phosphorous via 
acid digestion, total available phosphorous via Mehlich III extraction, total nitrogen, 
and total carbon. 
The acid digestion method for extracting phosphorous from sediments is 
intended to detect soil phosphorous in its totality.  It involves subjecting sediment 
samples to a harsh acidic reagent, removing all forms of phosphorous and converting 
them to orthophosphate (Fishman, 1993).  Mehlich III extraction involves agitating 
sediment samples in the presence of a gentler reagent, causing only a portion of total 
phosphorous to be converted to orthophosphate.  The amount of phosphate extracted 
by the Mehlich III technique is intended to mimic the amount of soil phosphorous 
available to wetland biota (Tran & Simard, 1993). 
For both methods of extraction, orthophosphate concentration was then 
measured using an automated colorimetric technique.  Orthophosphate samples are 
treated with several acidic reagents, resulting in the formation of a blue complex 
which absorbs light at 880 nm.  A spectrophotometer is then used in combination 
with an automated Flow Injection Analysis system to measure light absorbance and 
thus determine concentration of orthophosphate (Clesceri, Greenberg, & Eaton, 1998, 
Method 4500-P G). 
Soil total carbon and soil total nitrogen were both measured using the Dumas 
combustion technique.  In this method, sediment samples are dried and pulverized, 
 
 62 
and then placed in a furnace where they are combusted at extremely high 
temperatures in the presence of pure oxygen (O2).  Combustion removes all carbon 
and nitrogen from the solid phase and converts it into gaseous combustion products.  
Resulting gases are carried away in inert helium gas and then separated using gas 
chromatography.  Finally, they are measured using a thermal conductivity detector 
(Bremner, 1996). 
 
3.2.3.2 Chemical Analysis of Water Samples 
Water samples were analyzed at UMCES Appalachian Laboratory for total 
phosphorous and four types of nitrogen: total nitrogen, nitrogen in the form of nitrite 
and nitrate (NO2
 - + NO3
 -), nitrogen in the form of nitrite (NO2
 -), and nitrogen in the 
form of ammonia (NH3). 
Total phosphorous was extracted from water samples using a manual 
digestion method similar to the acid digestion technique used in soil analysis.  
Addition of an acid reagent converts phosphorous compounds to orthophosphate.  
Further addition of reagents causes the formation of a blue complex, which is 
measured automatically using a spectrophotometer and a Flow Injection Analysis 
system (Clesceri et al., 1998, Method 4500-P H). 
Total nitrogen was extracted from water samples using a different manual 
digestion technique.  In this technique, the water sample is exposed to a persulfate 
solution at high temperature, which causes the conversion of all nitrogen compounds 
into nitrate.  The sample is subsequently treated with cadmium to reduce nitrate to 
nitrite, and further treated with sulfanilamide and N-(1-naphthyl)ethylenediamine 
 
 63 
dihydrochloride to yield a magenta dye absorbing at 540 nm.  The concentration of 
dye is then determined spetrophotometrically via Flow Injection Analysis (Clesceri et 
al., 1998, Method 4500-N C). 
To determine the concentration of nitrite + nitrate, the same procedure is 
performed without the initial persulfate digestion step.  As a result, only nitrate and 
nitrite are incorporated into the magenta dye during cadmium reduction and dye 
treatment.  Dye concentration is still measured spectrophotometrically (Clesceri et al., 
1998, Method 4500-NO3
 - I). 
Finally, to determine the concentration of nitrite alone, the cadmium reduction 
step is also omitted from the procedure.  As a result, only the nitrite that was initially 
present in the water sample reacts with the sulfanilamide and N-(1-
 naphthyl)ethylenediamine dihydrochloride to form the resulting magenta dye 
(Clesceri et al., 1998, Method 4500-NO3
 - I). 
Ammonia nitrogen was determined using a separate colorimetric technique.  
In the technique for ammonia, the water sample is mixed with several reagents, and 
the dissolved ammonia reacts with these reagents to form indophenol blue.  
Indophenol blue concentration is then measured spectrophotometrically with Flow 
Injection Analysis (U.S. Environmental Protection Agency, 1999, Method 350.1). 
 
3.2.3.3 Antibiotic Resistance Testing of Water Samples 
The methods for antibiotic resistance testing were adapted from a study 
entitled Microbial Source Tracking of Escherichia coli in a Constructed Wetland 
(Orosz-Coghlan, et al., 2006).  Most supplies were obtained from Fisher Scientific, 
 
 64 
unless noted. Three 10x serial dilutions were performed for each sample of wetland 
water, using a 0.9 percent saline solution to normalize volume, creating solutions of 
100%, 10%, 1%, and 0.1% of the original concentration.  E. coli was then isolated 
from these samples using m-ColiBlue24? broth culture media and following the 
procedures as instructed. Each of the four dilutions was filtered through a separate 
filter with 0.45nm pores small enough to capture bacteria.  In general, 50 milliliters 
were passed through the filter for each dilution; variations in volume (due to 
availability at sites) were noted and later factored into calculated colony counts. 
The filters now contained the bacteria trapped from the diluted solution.  Each 
filter was placed in a sterile Petri dish on top of an absorbent pad containing a 
uniform volume of m-ColiBlue24? broth culture media, and the entire dish was 
incubated at 37? C for 18 to 24 hours.  The basis for using this media is bacterial 
growth selectivity: m-ColiBlue24? only supports coliform bacteria.  E. coli grow as 
blue colonies on the plate, while non-E. Coli coliforms grow as red colonies.  Once 
the incubation period was complete, both total coliform and E. coli coliform counts 
were taken and standardized as colony forming units per milliliter (CFU/mL).   
 
 
 65 
Figure 7: Example of a Millipore m-ColiBlue plate 
 
Red (total coliform) and blue (E. coli) colonies are visible.  Image from 
www,millipore.com. 
 
 
A minimum of three individual blue E. coli colonies were randomly selected 
from among the four plates prepared for each site. Often, it was difficult to select an 
isolated E. coli colony that made no contact with another colony of any type; in other 
instances, no E. coli colonies were detected. The selected isolates were then streaked 
onto separate non-selective Tripticase Soy agar (TSA) plates in order to proliferate 
the bacteria.  The TSA plates were incubated at 37? C for 24 hours.  After 24 hours, a 
single colony on each plate was selected, and the BD BBL? Enterotube II test was 
used to confirm the identity of the sample. This procedure consists of 15 simultaneous 
biochemical tests that allow the identification of Gram-negative species. Enterotubes 
were incubated at 37? C for 24 hours. At the same time that the enterotube test was 
performed, a sterile toothpick was used to obtain a portion of the same colony and 
inoculate 5 milliliters of separate sterilized TS broth solution.  The broth solutions 
were cultured at 37? C for 24 hours inside a shaking incubator.  
E. coli antibiotic resistance was determined using the standard Kirby-Bauer 
method for testing antibiotic resistance (Wikler, 2006).  In accordance with this 
 
 66 
standard protocol, after 24 hours, each broth culture was uniformly plated onto two 
plates of Mueller Hinton agar. Ideally, turbidity of each culture should have been 
measured in order to comply with McFarland standards of bacterial suspensions. 
However, we did not have access to the equipment required to carry out this step; 
instead, after a visual check of turbidity for consistency among samples, we 
uniformly plated 2 milliliters of each broth culture. Since all cultures were growing in 
similar conditions for the same period of time, we believe that there would not have 
been a significant difference in the concentrations of the solutions. Due to the plating 
technique, after incubation, an effective ?lawn? of bacteria results, and individual 
colonies cannot be distinguished.  BBL Antibiotic Sensitivity discs with standard 
dosages were obtained: erythromycin, 15?g; ciprofloxacin, 5 ?g; ampicillin, 10 ?g; 
sulfisoxazole, 25 ?g; tetracycline, 30 ?g; and vancomycin, 30 ?g. The discs were then 
spaced at least 5 centimeters apart on the plates, three on each of the two plates.  
These two plates were then incubated for exactly 24 hours at 37? C. 
After 24 hours, zone of clearance of the bacterial lawn around each of the 
antibiotic discs was measured and recorded.  The total diameter of each circular 
clearance, including the disc itself, was recorded.  However, if there was no visible 
clearance around the disc, a value of 0 mm was recorded.  Evaluation of resistance for 
a particular antibiotic was based on standard zone diameters taken from the Clinical 
and Laboratory Standards Institute (Wikler et al., 2006).   
 
 
 67 
3.3 Methods of Data Analysis 
Upon completion of data collection, all data was imported into Microsoft 
Excel and the STATA Data and Analysis Package (Statacorp., ver. 9).  Two types of 
statistical tests were then performed to determine significant correlations between all 
the different measurements taken in this study: least-squares regression and t-tests.   
Least-squares regression is a statistical means for determining whether there 
exists a significant relationship between two variables.  Microsoft Excel was used to 
run a least-squares regression on every possible two-variable combination, resulting 
in the creation of a large correlation matrix containing a list of all statistically 
significant relationships and their associated two-tailed p-values.  A p-value is 
essentially a measure of statistical confidence, with lower p-values indicating greater 
confidence.  
T-tests were particularly appropriate for testing the relationship of antibiotic 
resistance to other variables.  A t-test is used to determine whether there is a 
statistically significant difference between the characteristics of two groups; in this 
study the two groups used for each antibiotic were wetland sites with resistant isolates 
and wetland sites with only susceptible isolates.  A t-test was run between 
resistance/susceptibility to each antibiotic and each of the other wetland variables, 
with the main result being a two-tailed p-value.  If this p-value was less than 0.05, its 
respective correlation was reported as being statistically significant.  If this p-value 
was between 0.05 and 0.10, its correlation was reported as being almost statistically 
significant. 
 
 
 68 
4. Results & Discussion 
 
4.1 Land Use 
Computerized analysis of the land use at and around a wetland site is an 
important initial step in understanding the wetland's most basic characteristics.  
Landscape Development Intensity (LDI) is a common index used to provide an 
approximate gauge of how much a wetland site is impacted by surrounding human 
development.  LDI combines individual land use types into one value by use of a 
weighted summation formula.  In this study, LDI was calculated twice for each 
wetland site: once within a 1000 meter circular buffer extending out from the center 
of the site, and once within a 2000 meter circular buffer.  LDI values for the thirteen 
sites are shown in Figure 8. 
 
Figure 8: LDI values from 1000m and 2000m circular buffers surrounding each of 
the 13 study sites 
Landscape Development Intensity Index
 0
 3
 6
 9
 12
 15
 18
 SHK MRK CCW BRY PSC JLP KIN CBL INC AUD CAL HCP BEE
 Wetland Site
 La
 nd
 sc
 ap
 e 
D
 ev
 el
 op
 m
 en
 t 
In
 te
 ns
 it
 y 1000 m Buffer
 2000 m Buffer
  
 
 
 69 
 
LDI is convenient in that it provides a unified measure of the impact of land 
development on a wetland, but at the same time it does not provide details as to the 
type of land use that causes this impact.  Thus, land use was further broken down into 
the three broad classifications of ?agricultural?, ?urban?, and ?natural?.  Figure 9 to 
Figure 11 show the percentage of land around the study sites falling into each of these 
three categories.  As with LDI, tabulation was performed within two circular buffers: 
one of radius 1000m and the other of radius 2000m. 
 
Figure 9: Percentage of land used for agricultural purposes in the areas surrounding 
each of the 13 study sites 
Agricultural Land Use Around Study Sites
 0%
 20%
 40%
 60%
 80%
 100%
 SHK MRK CCW BRY PSC JLP KIN CBL INC AUD CAL HCP BEE
 Wetland Site
 P
 er
 ce
 nt
  A
 gr
 ic
 ul
 tu
 ra
 l L
 an
 d 
U
 se 1000 m Buffer
 2000 m Buffer
  
Land use tabulation was performed within two circular buffers extending around the 
center of each site - the first buffer had a radius of 1000 meters and the second buffer 
had a radius of 2000 meters. 
 
 
 
 
 70 
Figure 10: Percentage of urban land in the area surrounding each of the 13 study sites 
Urban Land Use Around Study Sites
 0%
 20%
 40%
 60%
 80%
 100%
 SHK MRK CCW BRY PSC JLP KIN CBL INC AUD CAL HCP BEE
 Wetland Site
 P
 er
 ce
 nt
  U
 rb
 an
  L
 an
 d 
U
 se
 1000 m Buffer
 2000 m Buffer
  
Land use tabulation was performed within two circular buffers extending from the 
center of each site - the first buffer had a radius of 1000 meters and the second buffer 
had a radius of 2000 meters. 
 
Figure 11: Percentage of ?natural? land use in the area surrounding each of the 13 
study sites 
Natural Land Use Around Study Sites
 0%
 20%
 40%
 60%
 80%
 100%
 SHK MRK CCW BRY PSC JLP KIN CBL INC AUD CAL HCP BEE
 Wetland Site
 P
 er
 ce
 nt
  N
 at
 ur
 al
  L
 an
 d 
U
 se
 1000 m Buffer
 2000 m Buffer
  
Land use tabulation was performed within two circular buffers extending from the 
center of each site - the first buffer had a radius of 1000 meters and the second buffer 
had a radius of 2000 meters. 
 
 71 
Moving from left to right in Figure 8 to Figure 11, wetland sites are arranged 
in order of increasing LDI within their 1000m buffers.  The most visible trend is that 
urban land use also tends to increase from left to right, while natural land use tends to 
decrease in this direction.  Thus LDI is directly correlated to percentage urban land 
use and inversely correlated to natural land use; this trend is easily understood, given 
that LDI is a measurement of human impact.  Urban land use constitutes the highest 
level of human impact on an ecosystem, while natural land use constitutes the lowest 
level of human impact. 
Indeed Shockley site, which had the lowest LDI values for both the 1000 and 
2000m buffers (0.457 and 0.888 respectively), also had the highest percentage of 
natural land use.  More specifically, mixed and brush land use made up over 90% of 
the total buffer.  As expected, Shockley had very low levels of agricultural and urban 
land use.  In fact, no urban land use was identified until the buffer was expanded to 
2000m.  It can be concluded that Shockley is the wetland site least affected by human 
development. 
On the opposite end of the spectrum, Beehive was the site with the highest 
LDI values (LDI = 16.085 and 15.459).  As noted earlier, high LDI values are 
associated with either high percentage of urban or agricultural land use.   Beehive 
site, because of its location in Elkridge, MD, had the greatest percentage of urban 
land use.  Three of the most abundantly found land uses were residential, industrial, 
and institutional.  After visiting Beehive site, it is clear the site is located in a highly 
populate residential neighborhood. 
 
 72 
Herring Creek had the second highest LDI for the 1000m buffer (9.532), 
which was nearly twice that of average site LDI.  Unexpectedly, of the three main 
land use classifications, natural land was the most prevalent.  Both water and 
wetlands were found in the top five most commonly occurring land use 
classifications.  Herring Creek?s high LDI value can be explained by it?s a relatively 
high level of urban land use (38.484%).  Agricultural land use was almost 
nonexistent. 
With respect to the 1000m buffer, Calvert had the third highest LDI value 
(8.1834).  A large portion of CAL?s development intensity can be contributed to 
urban land use, and more specifically, low density residential.  The presence of 
residential lands can be attributed to the sites? close proximity to MD route four.  In 
addition, a large percentage (49%) of land use was characterized as natural habitat, 
most notably deciduous forest. 
Aud?s LDI index (4.524) was just slightly below the site average.  An 
overwhelming majority of the overlaying 1000m buffer (58%) was classified as 
natural habitat.  Natural habitat classifications found around AUD include mixed, 
wetlands, and water.  Significant levels of both agriculture and urban land uses were 
found as well (27.02% and 14.88% respectively).  The agriculture lands are linked to 
a strong presence of cropland and the urban lands are associated with low density 
residential housing surrounding the site. 
Irvine had the fifth highest LDI index of the 13 sites.  A large percentage of 
the surrounding land use was classified as Agriculture, and more specifically 
cropland.  Irvine was established on a former farm, and large fields still occupy a 
 
 73 
large portion of the site; therefore, a large percentage of agricultural land use is to be 
expected.  When the buffer was expanded, Irvine?s close proximity to Baltimore city 
suburbs is reflected in the increase in urban land use.   
Cumberland had a slightly lower LDI than the average (4.044), due to the 
large amount of natural landscape.  Around the Cumberland site, the five most 
prevalent land use classifications were Mixed, Pasture, Deciduous, Residential, and 
Open Urban.  Cumberland was located in the foothills of the Appalachian Mountains; 
therefore, there was very little human development.  This lack of human influence is 
reflected in the low LDI value. 
Kinder had a lower LDI than the average (3.7792).  In the 1000m buffer 
around the site, there was very little urban development (2.94%) because the site was 
located in a rural area of Anne Arundel County.  There was an abundance of 
agriculture land (60.42%) because Kinder was located in between two large farms.  
As the buffer surrounding Kinder was expanded to 2000m, the impact of these two 
farms on LDI was reduced.  At this site there was also a strong presence of natural 
habitat (36.64%).  The majority of natural land can be classified as deciduous forest. 
Jackson Lane has a LDI number well below the average (3.3854).  No urban 
land was found in the 1000m buffer around the site because Jackson Lane is located 
in a remote area of Caroline County.  Over 90% of the land use found in the site?s 
1000m buffer consists of either Cropland or Deciduous Forest.  This lack of urban 
development and high percentage of natural environment (44.80%) explain Jackson 
Lane?s low LDI value. 
 
 74 
Averages for LDI, Agriculture, Urban, and Natural were calculated as follows 
for the thirteen sites: 
 Average LDI (1000m Buffer) = 5.00 
 Average LDI (2000m Buffer) = 6.12 
 Average % Agriculture (1000m Buffer) = 28.7% 
 Average % Agriculture (2000m Buffer) = 26.5% 
 Average % Urban (1000m Buffer) = 16.2% 
 Average % Urban (2000m Buffer) = 20.7% 
 Average % Natural (1000m Buffer) = 55.2% 
 Average % Natural (2000m Buffer) = 52.8% 
As seen from the above averaged data is can be determined that as the buffer 
was expanded the LDI value increased (by 1.12).  In also examining the averaged 
data, percent urban land use also increased; however, both percent agriculture and 
percent urban decreased. 
The LDI values for both the 1000m buffer and 2000m buffer were compared 
to one another in order to identify significant changes.  A difference in LDI 
(LDI2000-LDI1000) was calculated to determine significant increases or decreases.  
In examining these LDI differences, it was determined that nine of the 13 sites saw 
their LDI values increase as the buffer zone was expanded.  In addition, three sites 
showed a significant increase in value of at least 2.5.  These sites included Irvine, 
Cumberland, and Waldorf.  A large increase in LDI demonstrates that the impact on 
the buffer area is amplified as the buffer is expanded.  After determining significant 
increases in LDI, it was important to investigate and determine the relative causes.   
 
 75 
Irvine observed the largest increase in LDI of the 13 sites.  In examining the 
other data, it is clear that as the buffer was increased, high density urban land was 
found.  A larger portion of the buffer consisted of more highly developed lands; thus, 
resulting in a large LDI value.  This increase in LDI and % Urban land use could 
indicate a possible area of concern for the site.  Waldorf site also had a notably higher 
LDI value for its 2000m buffer.  In examining other data concerning Waldorf, it was 
clear that this increase is due to the decrease of natural land and corresponding 
increase in urban development.  Natural land uses, such as Mixed Forest or Brush, 
have very little impact on sites; therefore they are valued as very low on the LDI 
Index.  Conversely, urban land uses, such as Residential or Industrial lands, have a 
much higher impact on the surrounding environment.  As was the case with Waldorf, 
more urban land corresponded directly to a large increase in LDI.  Finally, 
Cumberland saw an increase of LDI of 2.857.  This increase in LDI again 
corresponds directly with an increase in percent urban land use (5.799%) and a 
decrease in percent natural land use (3.538%).  Although the increase in percent 
urban land use is not very large, the LDI increase can be explained by the high level 
of development found in the 2000m buffer.  Urban land uses with an extremely high 
LDI coefficient are the cause for Cumberland?s LDI increasing. 
 
4.2 Vegetation Sampling 
The study of wetland plants is essential to fully understand a wetland?s 
function.  Wetlands are a sink for nutrients and wetland plants greatly influence water 
chemistry and the cycling of those nutrients.  They are primarily responsible for the 
 
 76 
water filtering and cleansing that has been widely observed (Gersberg et al, 1986).  
They are also the base of the food chain and even provide a habitat for other 
taxonomic groups (Cronk and Mitsch 1994).  Team CRABS sampled vegetation at 36 
separate plots across Maryland and identified 256 separate species. 
 
4.2.1 Species Leaf Area Cover and Land Use 
Figure 12 shows the dominant plant community in the agricultural sites and 
traces the leaf cover through the natural and urban sites.  The y-axis shows the per-
 plot average midpoint leaf area cover over all of the agricultural, natural, or urban 
sites.  That is, it shows what percent of the average plot would be covered by a given 
species.  The x-axis shows the dominant plant community by species.  Figure 13 and 
Figure 14 are identical to Figure 12, except they track the dominant plant community 
in the natural and urban sites, respectively. 
 
 
 77 
Figure 12: Top ten species by cover at agricultural sites, along with their cover at 
natural and urban sites 
 
Agricultural sites included Irvine Nature Center, Jackson Lane Preserve, Kinder Site, 
and Bryantown. 
 
Figure 13: Top ten species by cover at natural sites, along with their cover at 
agricultural and urban sites 
 
Natural sites include Shockley, MD 228 Site, Merkle Wildlife Refuge, Piscataway 
Stream Valley Park, and Aud. 
 
 78 
Figure 14: Top ten species by cover at urban sites, along with their cover at 
agricultural and natural sites 
Urban sites include the US 220 Site, Beehive Site, MD 4/ MD 260 Site,  and Herring 
Creek Park. 
 
 
There are some clear relationships shown between dominant land use type and 
the dominant plant community.  In Figure 12, the agricultural and urban dominant 
plant distributions are remarkably different.  Likewise, in Figure 13, the natural and 
urban plant community dominant cover distribution show clear differences.  Figure 
14 does not show a clear enough trend between any of the three land use types to 
form any conclusions.   
The natural and agricultural sites generally have more woody species than the 
urban sites, which are dominated by herbaceous plants.  There may be some factor 
 
 79 
such as frequent or intense flooding, or more nutrient loading, that prevent the woody 
species from taking a hold in the urban wetland environment.  
Qualitatively, these results suggest that land use has a great impact on the 
dominant vegetation seen in mitigation wetlands.  For example, if vegetation 
sampling turned up a dominant species distribution similar to the agricultural sites in 
Figure 12, especially with a high occurrence of Salix nigra, it would be reasonable to 
conclude that the surrounding land is primarily designated as agricultural.  Being 
more cautious, another conclusion could be that the surrounding land is probably not 
urbanized.  This information could also be used to guess at dominant plant 
community based on GIS land use analysis.  Similarly, Figure 13 indicates that an 
average species distribution similar to that of the natural sites makes it likely that the 
surrounding land is not urban. 
 
4.2.2 Wetland Vegetation Prevalence Index 
Table 9 states the five different indicator designations for wetland status, 
based on the vegetation present.  Table 10 lists the top 15 species by percent cover 
found in all plots that Team CRABS sampled, along with their wetland indicator 
status.   
 
 80 
Table 9: Wetland indicator status categories for plant species 
Wetland Indicator Status Probability of Occurrence 
in Wetlands (%) 
Probability of Occurrence 
in Non-Wetlands (%) 
Weight 
Obligate wetland (OBL) >99 <1 1 
Facultative wetland (FACW) 67-99 1-33 2 
Facultative (FAC) 24-66 34-66 3 
Facultative upland (FACU) 1-33 67-99 4 
Upland (UPL) <1 >99 5 
Table from Cronk and Fennessy,  2001.  
  
 
Table 10: Top 15 species by leaf area cover over all sites sampled by CRABS, along 
with their indicator status 
Species Indicator Status 
Juncus effusus FACW 
Salix nigra FACW 
Phragmites australis FACW 
Dichanthelium clandestinum FACU 
Liquidambar styraciflua FAC 
Leersia oryzoides OBL 
Polygonum sagittatum OBL 
Typha latifolia OBL 
Acer rubrum FAC 
Ludwigia palustris OBL 
Polygonum persicaria FACW 
Typha angustifolia OBL 
Juncus canadensis OBL 
Scirpus atrovirens OBL 
Betula nigra FACW 
This data is of interest when considering the overall state of wetlands. 
 
 
 
The indicators were developed to supplement wetland delineation methods 
employed by the Army Corps of Engineers (Cronk and Fennessy, 2001).  When 
evaluating each site, the top ten species by cover were found and assigned a weight 
according to their indicator status.  A weighted average of the indicator value, also 
called the prevalence index (Wentworth et al. 1988), was then found at each site 
according to the following formula: 
 
 81 
 
 
 
Wi is the indicator weight of species i, Cij is the midpoint cover of species i in 
plot j, and p is the total number of species in plot j.  The results are shown in Figure 
15, with each site?s value being the average of the weighted average of each plot in 
the site.  The y-axis ranges from 1 (site totally comprised of OBL species) to 5 (site 
totally comprised of UPL species).  Any site that scores <3.0 is considered a wetland 
site.   
 
 
Figure 15: Wetland vegetation prevalence index for all sites 
 
Any value below 3.0 is accepted as an indication of the presence of a wetland. 
 
 
 
? 
Weighted Average j =
 CijWi
 i=1
 p
 ?
 Cij
 i=1
 p
 ?
 
 82 
Only one of the sites, Kinder, had a prevalence index over 3.0.  This is an 
encouraging result as it indicates that the mitigation wetlands in Maryland are 
performing well.  The sites in Figure 15 are arranged according to their ratio of 
natural to urban land use, with most natural on the left.  Although trends related to 
land use and plant life were seen in other analysis, there appears to be no trend here.  
This makes sense because in most cases, the presence of water in a wetland is dictated 
first by topography, and then by land use and other factors.  Since the presence of 
wetland plants is primarily dictated by hydrology, prevalence index should not be 
related to land use.  However, in a case where a wetland has a large area of 
impervious surfaces, or some other mechanism that would increase runoff during rain 
storms in its watershed, the prevalence index may shift toward FACW species that 
favor more constant inundation.  
 
4.2.3 Shannon Diversity Index 
The Shannon diversity index (SDI) is a measure of the biodiversity in a plot.  
It is maximized by an even distribution of leaf area cover between species and it 
increases with increasing species.  It can be calculated according to the following 
formula (Shannon and Weaver, 1949): 
? 
H = ? pi log pi
 i=1
 S
 ?  
 
H is the diversity index, s is the number of species in a plot, and pi is the ratio 
of the leaf area cover of the species to the combined leaf area cover of all species in 
the plot.  This study used a log base of 2.  Figure 16 shows the diversity index of each 
site.   
 
 83 
Figure 16: Shannon diversity index for all sites 
 
The higher the index value, the more evenly distributed species there are in the 
sampled area. 
 
 
 
The sites are again arranged by land use ?naturalness? from left to right, with 
left being the sites with the largest percentage of surrounding natural land use.  Some 
of the most diverse sites (Kinder, Calvert) did not rate low on the prevalence index 
(i.e. did not have a lot of wetland plants).  The combination of upland and wetland 
ecosystem probably increased the number of species found at these sites, and thus 
increased the diversity index.   
Figure 17 is a scatter plot of the diversity index.  The x-axis is dimensionless, 
as it simply represents the ratio of natural to urban land use, with most natural on the 
left.  There is a noticeable trend from high diversity in natural environments to low 
diversity in urban environments.  This could be because wetlands in urban 
 
 84 
environments have a greater nutrient and pollution inflow, and thus the less pollution 
tolerant species may not be able grow.   
 
Figure 17: Shannon diversity index with the x-axis showing the sites sorted from 
more natural to less natural by the ratio of natural to urban land use 
 
The linear regression R2 value is shown in the upper right corner.  This trend is 
statistically significant at the 5% level.   
 
 
4.3 Water Sampling 
Surface and sub-surface water quality measurements were obtained for total 
nitrogen, organic nitrogen, nitrite, nitrate, ammonia, and total phosphorous. 
A study released by the Florida Department of Environmental Protection in 
2005 sampled local wetlands in order to find biological indicators for developing the 
Florida Wetland Condition Index (Brown & Reiss, 2005).  The wetlands were 
 
 85 
grouped into three categories: Agricultural, Urban, and Reference.  Because this study 
grouped their wetlands into land-use categories similar to the way we did, we decided 
that it was a valuable tool for comparing out water quality data to other wetlands.  We 
too classified our wetlands as agricultural and urban, but our third category was 
natural.  For the sake of comparison, we will treat the Florida study?s reference 
wetlands the same as our natural wetlands.   
  It also helped that the Florida study sampled for many of the same nutrients 
as we did.  While the wetlands in Florida are in a vastly different ecosystem than our 
wetlands, they can still be used as a tool for comparison.  Even if you look at our 
wetlands alone, they too come from different ecosystems.  We did not expect to find 
exactly similar water quality data, but we did hope that there would be some 
similarities.  We looked at other studies to compare our numbers to (such as studies 
conducted in North Carolina and Ohio), but the Florida study was the most similar to 
ours.  Also, it was beneficial that the Florida study was conducted in 2005 and that it 
was relatively new data. 
 
4.3.1 Total Nitrogen 
In 2006, the site with the lowest level of total nitrogen present in surface water 
samples was Calvert with a reading of 1.219 mg-N/L (Figure 18).  The highest level 
was found at Bryantown with a reading of 16.834 mg-N/L.  The second highest level 
was at Aud with a reading of 14.224 mg-N/L. 
In 2007, the site with the lowest level of total nitrogen in surface water 
samples was once again Calvert and the highest level was now found at Cumberland. 
 
 86 
In the sub-surface samples, the average had fallen from 3.252 mg-N/L of total 
nitrogen in 2006 to 0.2005 mg-N/L in 2007.  As with nitrite+nitrate, the site with the 
lowest level in 2006 (Piscataway with 1.101 mg-N/L) was higher than the highest 
level the following year (MRK with 1.0418 mg-N/L).  Total nitrogen levels were 
much higher in 2006 for reasons that are not known to us. 
The highest reading in 2006 was Bryantown with 7.438 mg-N/L.  Bryantown 
had unusually high levels of total nitrogen at the surface water level as well as sub-
 surface.   
Again, our correlation index showed that there was a possible correlation 
between the level of urban land use present and one of the constituent levels, this time 
2006 total nitrogen levels.  Figure 18 shows the total nitrogen levels for the 2006 sub-
 surface and surface water samples.   
 
 
 87 
Figure 18: Total nitrogen levels across all sites in 2006 
2006 Total Nitrogen Levels (Sub-Surface vs Surface) 
0.000
 2.000
 4.000
 6.000
 8.000
 10.000
 12.000
 14.000
 16.000
 18.000
 BEE HCP CAL AUD INC CBL KIN JLP PSC BRY CCW MRK SHK
 Sites
 S
 ite
  A
 ve
 ra
 ge
  (
 m
 g/
 L)
 Sub-
 Surface
 Levels
 SW
 levels
  
 
 As stated earlier the sites are shown in order with the LDI rankings for each 
site, with the highest on the left.  If the correlation held true, it would make sense that 
the sites with the higher LDI rankings (those on the left) would have higher total 
nitrogen levels.  Looking at the figure, this is not obvious to the naked eye.  In order 
to further examine this correlation, more samples are needed.   
 
4.3.2 Organic Nitrogen 
In 2006 surface water samples the site with the highest organic nitrogen 
average was Irvine with a reading of 21.542 mg-N/L.  The site with the lowest 
average was Piscataway with 1.19 mg-N/L. 
The 2006 surface water average for all sites was 5.396 mg-N/L. 
 
 88 
In 2006 sub-surface water samples the site with the highest organic nitrogen 
average was Bryantown with 3.324 mg-N/L.  The site with the lowest average was 
Herring Creek with a reading of 0.809 mg-N/L. 
The 2006 sub-surface water average for all sites was 1.573 mg-N/L. 
One reason that surface water sample averages could be so much higher is that 
the layer of organic matter mixes in with the surface water, while the lysimeters take 
water samples from a depth in the ground which is almost always below the organic 
matter. 
In 2007 there were no organic nitrogen readings for either surface or sub-
 surface water samples. 
Figure 19 shows the organic nitrogen levels for the 2006 sub-surface and 
surface water samples.   
 
 
 89 
Figure 19: Organic nitrogen levels across all sites in 2006 
2006 Organic Nitrogen Levels (Sub-Surface vs Surface) 
0.000
 5.000
 10.000
 15.000
 20.000
 25.000
 BEE HCP CAL AUD INC CBL KIN JLP PSC BRY CCW MRK SHK
 Sites
 S
 ite
  A
 ve
 ra
 ge
  (
 m
 g/
 L)
 Sub-
 Surface
 Levels
 SW
 levels
  
  
4.3.3 Nitrite 
In 2006 the site with the least amount of nitrite in surface water samples was 
Herring Creek with a reading of 0.004 mg-N/L.  The site with the highest amount of 
nitrate was Bryantown with 0.21 mg-/L with the next highest being Merkle with 
0.030 mg-N/L.   
In 2007 Merkle had a higher reading than the year before and Bryantown 
dropped drastically so that now Merkle was the highest with a reading of 0.0539 mg-
 N/L and Bryantown was second with 0.0154 mg-N/L.  Instead of Herring Creek 
being the lowest reading in 2007 three sites (Calvert, Beehive, and Irvine) measured 
below detection limits.  Herring Creek did have the lowest reading among those sites 
that were able to be detected, with a reading of 0.0010 mg-N/L. 
 
 90 
The average nitrite reading for surface water samples was 0.026 mg-N/L in 
2006, but that fell to 0.0131 mg-N/L in 2007.  A large part of this can be accounted 
for by an unusually high 2006 reading for Bryantown. 
There was an unusually high reading for nitrite levels in the sub-surface water 
samples as well.  In 2006, Waldorf registered 26.005 mg-N/L of nitrite in the sub-
 surface samples.  This was by far the highest reading for both surface and sub-surface 
samples.  The next highest site was Jackson Lane with 0.378 mg-N/L.  The lowest 
two readings were Beehive and Piscataway with 0.015 and 0.020 mg-N/L, 
respectively.   
In 2007 Waldorf dropped to a lower level with a reading of 0.2320 mg-N/L, 
and the highest reading for a site was now Bryantown with 0.5826 mg-N/L.  
Surprisingly, Beehive was now the second highest with a reading of 0.3513 mg-N/L, 
whereas it had had the lowest levels the year before.  The lowest readings for 2007 
were Shockley and Piscataway with 0.0326 and 0.0493 mg-N/L, respectively.   
The average level of nitrite in the 2006 sub-surface samples for all 13 sites 
was 2.096 mg-N/L, distorted by the unusually high Waldorf reading.  In 2007, the 
samples had much more parity, and the average was 0.1948 mg-N/L.  Overall, the 
2006 numbers were very much higher than 2007.  Figure 20 shows the difference 
between 2006 sub-surface nitrite and 2007 sub-surface nitrite.  Make sure to 
remember that the Waldorf reading was over 26 mg-N/L in 2006, thus it does not fit 
on the figure. 
 
 91 
Figure 20: Sub-surface nitrite levels across all sites in 2006 and in 2007 
2007 Nitrite Levels vs. 2006 Nitrite Levels (Sub-Surface)
 0.0000
 0.1000
 0.2000
 0.3000
 0.4000
 0.5000
 BEE HCP CAL AUD INC CBL KIN JLP PSC BRY CCW MRK SHK
 Site Averages
 m
 g/
 L 2007 Nitrite Avg
 2006 Nitrite Avg
  
 
4.3.4 Nitrate 
In 2006 surface water samples, the site with the highest nitrate average was 
Merkle with a reading of 0.862 mg-N/L.  The site with the lowest average was Kinder 
with 0.001 mg-N/L without considering the seven sites (Shockely, Jackson Lane, 
Piscataway, Irvine, Beehive, Bryantown, and Aud) that were below the detection 
limit. 
In 2006 sub-surface water samples, the site with the highest nitrate average 
was Bryantown with a reading of 3.871 mg-N/L.  The site with the lowest nitrate 
average was Piscataway with 0.061 mg-N/L unless you consider Waldorf, which had 
an average nitrate reading of -38.54 mg-N/L.  This distorted the entire 2006 sub-
 surface water nitrate average. 
 
 92 
The 2006 site average for sub-surface ammonia was -2.07 mg-N/L due to 
CCW?s negative reading. 
There were no 2007 readings for nitrate for either surface or sub-surface 
samples. 
Our correlation index showed that there was a possible correlation between 
2006 sub-surface levels of nitrate and the level of urban land use.  Figure 21 shows 
the 2006 nitrate levels for surface samples at each site.  Because of Waldorf?s 
negative reading, it is just 0.00 mg-N/L for the figure. 
 
Figure 21: Surface water nitrate levels for all sites in 2006 
2006 Surface Water Nitrate Levels
 0.152
 0.346
 0.541
 0.330 0.313
 0.669
 0.246
 3.176
 0.061
 3.871
 0.0000
 1.365
 0.543
 0.000
 0.500
 1.000
 1.500
 2.000
 2.500
 3.000
 3.500
 4.000
 4.500
 BEE HCP CAL AUD INC CBL KIN JLP PSC BRY CCW MRK SHK
 Sites
 N
 itr
 at
 e 
Le
 ve
 l (
 m
 g/
 L)
 Site Avg
  
 
The sites are shown in order with the LDI rankings for each site, with the 
highest on the left.  Because sites with higher LDI rankings typically are indicative of 
 
 93 
a higher level of urban land use, it would make sense that the sites with the higher 
LDI rankings (those on the left) should have had higher nitrate levels.  Instead, the 
site averages on the right were much higher (note that CCW had a negative reading, 
thus the 0.00 site average).  This suggests that more samples are needed to further 
examine the possible correlation between land use and sub-surface nitrate levels. 
 
4.3.5 Ammonia 
In 2006, the site with the lowest ammonia reading in surface water samples 
was Calvert, with 0.027 mg-N/L.  The second lowest reading came from Herring 
Creek 0.031 mg-N/L.  The highest reading was Aud with 1.725 mg-N/L. 
In 2007 the site with the lowest ammonia reading for surface water was 
Herring Creek, with 0.0223 mg-N/L.  The second lowest reading came from Calvert 
with 0.0233 mg-N/L (Calvert and Herring Creek flip-flopped from 2006 to 2007).  
The highest reading was once again Aud, with 2.0669 mg-N/L.  Both years the 
second highest ammonia reading for surface water came from Bryantown. 
Seeing that Aud is downhill from a horse farm, it is not surprising that it 
would have the highest level of ammonia in surface water for both years.  Unusually 
high ammonia levels can be linked with fertilizer usage. 
Regarding the sub-surface samples, the lowest reading in 2006 came from 
Bryantown with 0.217 mg-N/L.  The second lowest reading came from Piscataway 
with 0.303 mg-N/L.  The highest reading was from Shockley with 1.153 mg-N/L. 
In 2007, numbers changed drastically.  Instead of having the highest reading, 
Shockley now had the lowest reading among the thirteen sites with 0.0326 mg-N/L.  
 
 94 
Shockley was the site that was the most natural of all sites.  There was no urban 
percentage of land-use and it was classified as 91% natural.  The fact that such a 
drastic drop in ammonia level occurred over the course of a year in a wetland that was 
almost completely isolated is something that could be pursued further.  The second 
lowest reading was Piscataway with 0.0493 mg-N/L.  The site with the highest 
reading was now Bryantown with 0.5826 mg-N/L.  Bryantown had been the site with 
the lowest ammonia reading among sub-surface samples the year before.  The sites 
with the highest and lowest readings had completely flip-flopped from 2006 to 2007.  
Both years the site with the second highest reading was Beehive. 
Figure 22 shows a visual comparison between sub-surface and surface 
ammonia levels for 2007, and Figure 23 shows a comparison between 2006 and 2007 
sub-surface ammonia readings.   
 
Figure 22: Sub-surface and surface ammonia levels in 2007 
2007 Ammonia Levels (Sub-Surface vs. Surface)
 0.0000
 0.5000
 1.0000
 1.5000
 2.0000
 2.5000
 BEE HCP CAL AUD INC CBL KIN JLP PSC BRY CCW MRK SHK
 Site Average
 A
 m
 m
 on
 ia
  L
 ev
 el
  (m
 g/
 L)
 Sub-Surface levels
 SW levels
  
 
 
 95 
Figure 23: Sub-surface ammonia levels in 2006 and in 2007 
2007 Ammonia Levels vs. 2006 Ammonia Levels (Sub-Surface)
 0.0000
 0.2000
 0.4000
 0.6000
 0.8000
 1.0000
 1.2000
 1.4000
 5.101 2.658 1.865 2.094 1.898 6.603 1.101 7.438 1.859 4.478 2.605 nitrate
 lys
 2006
 Site Averages
 m
 g/
 L 2007 Ammonia Avg
 2006 Ammonia Avg
  
 
The sites are listed in order of LDI (Land Development Index) rankings.  This 
rates a wetland based on the surrounding land use present.  In general, relatively 
higher LDI rankings correspond with a higher percentage of urban land use.  The 
wetlands with the highest LDI ranking are on the left.  Notice that the sites with the 
highest LDI rankings have higher sub-surface ammonia levels than surface ammonia 
levels.  Overall, sub-surface and surface ammonia do seem to follow each other.  
Note that Kinder and Waldorf did not have surface water ammonia readings for 2007.      
In the Florida study that we are using as comparison, there were 30 reference 
wetlands tested for ammonia levels and the average came out to 0.15 mg/L (Brown & 
Reiss, 2005).  In wetlands that we classified as predominately natural (similar to the 
Florida studies classification of ?reference?), surface water samples had an ammonia 
 
 96 
level average of 0.484 mg-N/L in 2006 and 0.797 mg-N/L in 2007.  The fact that 
these numbers are higher than the Florida numbers can be attributed to the fact that 
most of our natural wetlands were not 100% natural.   
Because we only had 13 sites, we wanted to have an even allotment of natural, 
urban, and agricultural wetlands.  We had 5 natural wetlands, 4 urban, and 4 
agricultural.  Because of this, there were some minor discrepancies.  For example, 
CAL was 49% natural, 40% urban and 11% agricultural.  It had to be labeled urban 
for statistical purposes because there were 5 sites that were more than 49% natural.  A 
more complicated method of allotting each wetland a certain weight based on each 
percentage land-use could be developed in the future.   
There were 19 agricultural wetlands in the Florida study that had an ammonia 
level average of 0.33 mg-N/L (Brown & Reiss, 2005).  This less than our 2006 
ammonia average of 0.601 mg-N/L, but much closer to the 2007 ammonia average of 
0.4158 mg-N/L for surface water in wetlands that we also classified as predominately 
agricultural.   
Finally, there were 26 urban wetlands in the Florida study that had an 
ammonia level average of 0.19 mg-N/L (Brown & Reiss, 2005).  Our urban wetlands 
had a 2006 ammonia average of 0.212 mg-N/L and a 2007 ammonia average of only 
0.115 mg-N/L in surface water samples.  The large gap is because we only had one 
wetland (Beehive) that was predominately urban, and its water quality numbers 
differed across the board from 2006 to 2007.   
Figure 24 compares the surface water ammonia averages from the Florida 
study to our 2007 numbers.   
 
 97 
 
Figure 24: Ammonia level averages from the Florida study and from CRABS? 
Maryland study 
Florida vs. Maryland ammonia levels
 0
 0.1
 0.2
 0.3
 0.4
 0.5
 0.6
 reference/natural  agricultural  urban
 land-use
 am
 m
 on
 ia
  m
 g/
 L
 Maryland
 Florida
  
 
4.3.6 Phosphorous 
In 2006, the site with the lowest phosphorous reading in surface water 
samples was Shockley with 0.065 mg-N/L.  The second lowest reading came from 
Herring Creek with 0.094 mg-N/L.  The highest reading was Aud with 12.487 mg-
 N/L. 
In 2007, the site with the lowest reading for phosphorous in surface water 
samples was Calvert with 0.0192 mg-N/L.  The second lowest reading was from 
Herring Creek with 0.0228 mg-N/L.  The site with the highest reading was again Aud 
with 5.1056 mg-N/L. 
 
 98 
The average reading for all 13 sites for surface water phosphorous was 1.0331 
mg-N/L in 2006 and 2.369 mg-N/L in 2007, it more than doubled. 
For the sub-surface samples, in 2006 the lowest phosphorous reading was 
from Herring Creek with 0.042 mg-N/L.  The second lowest reading was from 
Shockley with 0.059 mg-N/L.  The highest reading was from Merkle with 0.953 mg-
 N/L. 
In 2007, the lowest reading was again from Herring Creek, but it had gone all 
the way down to 0.0086 mg-N/L.  The second lowest reading was now Calvert with 
0.0240 mg-N/L.  Those two sites, along with Piscataway, Shockley, and Irvine were 
all below the lowest 2006 reading.  The highest phosphorous reading in sub-surface 
water samples for 2007 was from Beehive with 0.3150 mg-N/L.  This is merely one 
third of the highest reading from 2006.  In fact, the 2007 sub-surface numbers were 
much lower than 2006 with an average of 0.282 mg-N/L of phosphorous in 2006 and 
0.0932 mg-N/L in 2007. 
Surface water phosphorous increased on average from 2006 to 2007 while 
sub-surface phosphorous decreased. 
Figure 25 visually compares sub-surface phosphorous levels to surface 
phosphorous levels for 2006 and Figure 26 compares total phosphorous levels in sub-
 surface samples from 2006 and 2007.   
 
 
 99 
Figure 25: Sub-surface and surface total phosphorous levels in 2006 
Sub-Surface Total P vs. Surface Total P (2006)
 0.000
 2.000
 4.000
 6.000
 8.000
 10.000
 12.000
 14.000
 BEE HCP CAL AUD INC CBL KIN JLP PSC BRY CCW MRK SHK
 Site Averages
 m
 g/
 L Sub-Surface Avg
 SW Avg
  
 
Figure 26: Sub-surface total phosphorous in 2006 and in 2007 
2007 Total P levels vs. 2006 Total P levels (Sub-Surface)
 0.0000
 0.2000
 0.4000
 0.6000
 0.8000
 1.0000
 1.2000
 BEE HCP CAL AUD INC CBL KIN JLP PSC BRY CCW MRK SHK
 Site Averages
 m
 g/
 L 2007 Avg
 2006 Avg
  
 
 100 
Notice that at almost all the sites, surface water levels were much higher than 
sub-surface levels.  At some sites, the surface water levels were significantly higher.   
Of 30 reference wetlands in the Florida study the average level of 
phosphorous was 0.08 mg-N/L (Brown & Reiss, 2005).  In wetlands that we 
classified as natural, surface water samples had an average phosphorous level of 
3.141 mg-N/L in 2006 and 2.1394 mg-N/L in 2007.  Here is one example where our 
numbers were inexplicably higher than those of the Florida studies.   
Out of 19 wetlands classified as agricultural in the Florida study, the average 
level of phosphorous was 0.81 mg-N/L (Brown & Reiss, 2005).  In wetlands that we 
classified as agricultural, surface water samples had an average phosphorous level of 
3.389 mg-N/L in 2006 and 0.5341 mg-N/L in 2007.  These numbers are much more 
on par with the Florida numbers, suggesting that we classified agricultural wetlands 
much like they did, but our idea of natural wetlands and their idea of reference 
wetlands may not be as similar. 
Lastly, the Florida study sampled phosphorous levels from 26 wetlands that 
they considered urban; the average level of phosphorous in these wetlands was 0.24 
mg/L (Brown & Reiss, 2005).  In wetlands that we classified as urban, surface water 
samples had an average phosphorous level of 0.384 mg-N/L in 2006 and only 0.301 
mg-N/L in 2007.  Again, this can be attributed to the fact that we had only one 
predominately urban wetland (Beehive) and for some reason the ammonia levels 
AND phosphorous levels dropped drastically from 2006 to 2007. 
Figure 27 compares the surface water phosphorous levels from the Florida 
study to our numbers.   
 
 101 
 
Figure 27: Phosphorous levels from the Florida study and from CRABS? Maryland 
study 
Florida vs. Maryland phosphorous levels
 0
 0.5
 1
 1.5
 2
 2.5
 reference/natural  agricultural  urban
 land-use
 ph
 os
 ph
 or
 ou
 s 
(m
 g/
 L)
 Florida numbers
 Maryland numbers
  
 
4.3.7 Ammonia versus Total Phosphorous 
Out of all the constituents that we tested for, the two that showed the most 
correlation were ammonia and total phosphorous.  In 2007 surface water samples, if 
the ammonia level went up, so did the total phosphorous, and vice versa.  Figure 28 
and Figure 29 below show the movement of the two variables.  There were no 
samples for either Kinder or Waldorf.   
 
 
 102 
Figure 28: Comparison of ammonia and total phosphorous levels across all sites, 
averaged 
Ammonia Levels vs. Total P levels (2007 SW)
 0.0000
 1.0000
 2.0000
 3.0000
 4.0000
 5.0000
 6.0000
 BEE HCP CAL AUD INC CBL JLP PSC BRY MRK SHK
 Site Averages
 m
 g/
 L Ammonia Level Avg
 Total P avg.
  
 
Figure 29: Regression of ammonia level versus total phosphorous level 
Ammonia Level vs. Total P Level (2007 SW)
 y = 2.5421x - 0.0983
 R2 = 0.8788
 -1.0000
 0.0000
 1.0000
 2.0000
 3.0000
 4.0000
 5.0000
 6.0000
 0.0000 0.5000 1.0000 1.5000 2.0000 2.5000
 Ammonia Level Averages (mg/L)
 To
 ta
 l P
  L
 ev
 el
  A
 ve
 ra
 ge
 s 
(m
 g/
 L)
 Site Averages
 Linear (Site Averages)
  
 
 103 
4.3.8 Limitation of Water Results 
Our study shows that in various aspects of water quality, values may differ 
greatly from year to year.  However, it is important to note that while extreme 
variations in nutrient levels may occur, this result may also be a figment of the 
study?s small sample size.  In order to gain confidence, these values should be taken 
repeatedly over time, as one-time samples can lead to a distorted conclusion.   
 
4.4 Soil Sampling 
Our soils data results were broken into two categories: data obtained through 
soil nutrient analysis via Applachacian Laboratory and data obtained through onsite 
collection.  The soil samples sent to Appalachian Laboratory were analyzed for the 
following nutrient levels: total carbon (TC), total nitrogen (TN), total phosphorus 
(TP), plant available phosphorus (PAP), and carbon:nitrogen ratio (CN).  Soil nutrient 
measurements were taken using the following unit: milligram of nutrient / kilogram of 
soil basis.  Soil nutrient data was obtained for every plot at every sampled wetland 
site (see Table 11).  An average value and standard deviation were calculated for all 
five nutrient levels across all sites (see Table 12).  The standard deviation values for 
all five nutrient levels were extremely high, revealing that there was a high degree of 
variability between soil samples.  In fact, standard deviation values exceeded the site 
averages for both total carbon levels and plant available phosphorus levels.  This high 
degree of variability can be explained either by the sensitivity of the laboratory 
testing equipment or by large differences in nutrient composition of the soil samples.  
 
 
 104 
Table 11: Soil nutrient levels from each plot at the 13 sampled wetland sites 
 Total Nitrogen Total Carbon C:N Ratio 
Total 
Phosphorus 
PA 
Phosphorus 
 mg-N/kg-soil mg-C/kg-soil mg-C / mg-N mg-P/kg-soil mg-P/kg-soil 
BEE 1 420.67 6449.0 15.33 324.00 3.06 
BEE 2 649.32 14324.4 22.06 445.15 6.57 
BEE 3 702.39 13695.6 19.50 257.64 13.14 
HCP 1 175.95 2026.5 11.52 118.57 17.95 
HCP 2 431.62 7541.0 17.47 161.40 8.23 
HCP 3 554.46 9442.7 17.03 314.27 9.00 
CAL 1 112.74 639.2 5.67 259.34 0.20 
AUD 1 732.13 10772.7 14.71 443.57 82.56 
AUD 2 462.67 6541.3 14.14 398.06 58.94 
AUD 3 1019.13 13627.9 13.37 449.39 88.57 
INC 1 586.31 11701.7 19.96 525.10 9.58 
INC 2 1562.28 33199.3 21.25 382.67 6.80 
INC 3 405.03 5101.7 12.60 402.22 0.99 
CBL 1 727.93 10957.2 15.05 413.12 1.41 
CBL 2 820.79 11371.5 13.85 560.29 1.03 
CBL 3 531.97 6721.5 12.64 501.24 1.06 
KIN 1 426.76 4387.2 10.28 635.38 14.23 
KIN 2 450.48 4496.4 9.98 826.65 34.11 
KIN 3 492.41 5118.2 10.39 490.92 5.71 
JLP 1 890.45 16981.6 19.07 330.38 30.45 
JLP 2 546.45 10125.9 18.53 264.72 6.67 
JLP 3 1495.17 30022.7 20.08 295.36 6.64 
PSC 1 335.55 5040.8 15.02 220.02 5.38 
PSC 2 157.33 1038.2 6.60 786.37 18.00 
PSC 3 200.68 3001.4 14.96 293.01 36.62 
BRY 1 688.47 8989.0 13.06 508.47 5.88 
BRY 2 611.91 7838.0 12.81 590.03 9.46 
BRY 3 343.42 4647.2 13.53 249.70 9.23 
CCW 1 628.75 9652.1 15.35 532.83 23.47 
CCW 2 658.51 11066.4 16.81 597.45 58.05 
CCW 3 729.83 13189.8 18.07 565.71 44.59 
MRK 1 466.31 10057.4 21.57 143.54 22.50 
MRK 2 694.15 15553.1 22.41 244.60 19.57 
MRK 3 687.67 12348.2 17.96 203.46 27.18 
SHK 1 129.38 4455.6 34.44 17.65 14.78 
SHK 2 3515.44 74851.8 21.29 307.58 7.29 
SHK 3 665.42 29836.8 44.84 59.83 8.97 
 
 
 
 105 
Table 12: Average value and standard deviation for soil nutrient data 
Soil Nutrient Analysis Data 
 Avg St. Deviation 
Total Nitrogen 
mg-N/kg-soil 667.836 572.546 
Total Carbon 
mg-C/kg-soil 12075.977 12962.285 
C:N Ratio 
mg-C / mg-N 16.843 7.004 
Total Phosphorus 
mg-P/kg-soil 381.614 187.212 
PA Phosphorus 
mg-P/kg-soil 19.402 21.891 
 
 
 For TN, the plots with the lowest levels were Calvert and Shockley 1 with 
values of 112.74 and 129.38 respectively.  In addition, Calvert site had values 
significantly lower than the average values for TC, CN, and PAP.  A recurring trend 
appeared to emerge: if a certain soil sample had atypically high or low values for one 
nutrient variable, than it was likely it had abnormally high or low values for another 
variable.  Another example of this drawn conclusion is Irvine 2, which had 
abnormally high values for both TN (1562.28) and TC (33,199.3).  In addition, even 
within the same site, nutrient levels between plots could be significantly different.  
For example, Shockley 1 had a very low TN level (129.38) and Shockley 2 had a very 
high TN level (3515.44). 
In addition to a nutrient level analysis as described above, other important soil 
characteristics were examined on site.  These characteristics included: mottling, 
gleying, organic matter thickness (OMT), and soil depth (SD).  The data obtained for 
OMT and SD was measured using a yard stick and the values were recorded in 
centimeters.  Mottling and gleying were analyzed using an intensity rating scheme 
(based on a scale of 1-10).  The dotted line on the graph below (Figure 30) 
 
 106 
demonstrates that any site with a value above three exhibits significant anaerobic 
soils.  The averages for both mottling (m=4.36) and gleying (g=4.67) were well above 
this statistically significant threshold.  This result indicates that as a whole, the 
wetland sites illustrated a significant presence of mottling and gleying.   
 
Figure 30: Gleying and mottling intensity values from each of the 13 wetland sites 
Gleying & Mottling (Site Averages)
 0.00
 2.00
 4.00
 6.00
 8.00
 10.00
 BE
 E
 HC
 P
 CA
 L
 AU
 D IN
 C
 CB
 L
 KI
 N
 JL
 P
 PS
 C
 BR
 Y
 CC
 W
 MR
 K
 SH
 K
 SITE
 R
 at
 in
 g Mottling
 Gleying
 Averages:
 Mottling = 4.36
 Gleying = 4.67
  
 
Although a very wide range of intensity values existed, nine sites 
demonstrated significant mottling and twelve sites demonstrated significant gleying 
values.  Furthermore, every site we sampled revealed some presence of gleying in the 
soil.  In addition, only one study site (Calvert) lacked any presence of mottling.  The 
purpose of the Figure 30 is to illustrate that all study sites had characteristics of 
anaerobic soils.  This signified that the wetland soils at the selected sites were 
functioning properly. 
 
 107 
4.5 Antibiotic Resistance 
4.5.1 Overall Levels of Resistance Across Sites 
A total of 33 isolates of Escherichia coli were obtained during the sampling 
period in 2006, and 35 isolates were obtained in 2007, creating a final total of 68 
colonies of E. coli isolated during the course of this study.  The resistance or 
susceptibility of each colony to six different antibiotics was then determined via disk 
diffusion.  Overall rates of resistance to these six antibiotics are shown in Figure 31. 
 
Figure 31: Overall percentage of E. coli isolates resistant to each antibiotic tested, 
spanning all wetland sites as well as data from both 2006 and 2007 
Overall Resistance to Each Antibiotic
 n = 67
 n = 66
 n = 62
 n = 65
 n = 31
 n = 58
 0%
 20%
 40%
 60%
 80%
 100%
 Erythromycin Ampicillin Sulfisoxazole Tetracycline Ciprofloxacin Vancomycin
 P
 e
 rc
 e
 n
 ta
 g
 e
  o
 f 
C
 o
 lo
 n
 ie
 s 
R
 e
 si
 st
 a
 n
 t 
 .
  
For each antibiotic, 'n' denotes the number of colonies tested against that 
antibiotic.  Vancomycin was used as a negative control. 
 
 
 
It is important to note that the number of colonies tested against each 
antibiotic, designated ?n?, varies in Figure 31.  This is due to experimental 
 
 108 
inconsistencies.  On occasion, the zone of inhibition around an individual antibiotic 
disc would be amorphous instead of circular, and thus its radius could not be 
measured, nor could the respective colony?s resistance to that antibiotic be 
determined.  Amorphous inhibition zones account for the minor variations in values 
of n.  However, a more significant inconsistency is that of erythromycin; because 
standardized erythromycin discs were not available in 2007, only isolates from 2006 
were tested against this antibiotic.  As a result, the n value for erythromycin (n = 31) 
is about half that of the other antibiotics. 
Given that the noted inconsistencies are, for the most part, minor, they do not 
greatly diminish the ability of our results to show overall trends in rates of E. coli 
resistance to the tested antibiotics.  Erythromycin exhibited the highest percentage of 
resistant colonies, at 94%, while ampicillin, sulfisoxazole, and tetracycline had 
successively lower rates of resistance: 72%, 61%, and 36% respectively.  No colonies 
were determined to be resistant to ciprofloxacin.  An 87% resistance rate was 
observed towards vancomycin, which was considered to be a negative control. 
Basic qualitative differences in resistance rates to the six antibiotics fall well 
within the realm of expectation.  The fact that E. coli would exhibit the most 
resistance to erythromycin and ampicillin makes sense; both of these compounds are 
naturally occurring, both of them were discovered many decades ago, and both are 
widely used due to their broad spectrum of antibiotic activity.  In contrast, 
ciprofloxacin is an entirely synthetic compound discovered within the past few 
decades, and is used less often than many other antibiotics; therefore, less resistance 
was expected.  However, overall resistance rates for all 5 experimental antibiotics (i.e. 
 
 109 
vancomycin excluded) appear to be higher than rates of resistance found in previous 
studies, as shown in Figure 32. 
 
Figure 32: Comparison of overall E. coli resistance results from this study to 
resistance results from several similar previous studies 
Comparison of Resistance Results to Previous Studies
 A
 sh et al., 2002
 S
 ayah et al., 2005
 S
 ayah et al., 2005
 A
 kinbow
 ale et al., 2006
 A
 kinbow
 ale et al., 2006
 Y
 oo et al., 2004
 M
 cK
 eon et al., 1995
 M
 cK
 eon et al., 1995
 A
 sh et al., 2002
 E
 dge &
  H
 ill, 2005
 0%
 20%
 40%
 60%
 80%
 100%
 Erythromycin Ampicillin Sulfisoxazole Tetracycline Ciprofloxacin
 P
 e
 rc
 e
 n
 ta
 g
 e
  o
 f 
C
 o
 lo
 n
 ie
 s 
R
 e
 si
 st
 a
 n
 t 
 .
 CRABS
 Previous Studies
  
Results cannot be strictly compared due to differing methodologies and species 
studied, but in overall resistance rates found in this study were similar to or higher 
than those found in previous studies. 
 
 
The rate of resistance seen towards erythromycin was significantly larger than 
rates of resistance found in previous research.  Prior studies of other Gram-negative 
species, such as Campylobacter jejuni, have indicated no resistance to erythromycin 
in water samples (Levesque, 2007).  A study conducted by Akinbowale et al. (2006), 
in which Gram-positive and Gram-negative isolates from aquacultural environments 
were tested for susceptibility to 19 drugs, noted only a 47.1 percent resistance to 
erythromycin.  However, previous research by Zuccato et al. (2000) indicates that 
 
 110 
erythromycin has a tendency to maintain its presence in the environment for long 
periods of time.  This implies that bacterial isolates have had ample time to develop 
resistance to the drug, which could help explain our results.   
The measured rate of resistance to ampicillin was also higher than rates seen 
in previous studies, though not drastically so.  A recent study conducted in Australia 
reported 54.8 percent of 100 Gram-negative and four Gram-positive isolates to be 
resistant to ampicillin (Akinbowale et al., 2006), while another study conducted in the 
United States found 50 percent of E. coli isolates to be ampicillin resistant (Ash et al., 
2002).  As noted, the high resistance rates we found in this study are realistic given 
the widespread nature of penicillin use.  But it should also be pointed out that ?-
 lactam compounds do not persist in the environment, and therefore the high rate of 
ampicillin resistance was observed in spite of the transient nature of ampicillin in 
waterways.  This would seem to corroborate the idea that over time, the constant 
release of a short-lived compound can mimic the one-time release of a long-lived 
compound in terms of its chronic effects. 
Tetracycline results were more in line with those of previous studies.  In a 
study conducted by Sayah et al. (2005), tetracycline resistance was found at a rate of 
27.3 percent across 1,286 isolates.  Our data fell within this range; we observed a 36 
percent resistance rate. 
The low levels of bacterial resistance to ciprofloxacin found in this study 
agreed with the results of other authors.  A 2005 study of surface waters and fecal 
pollution sources near Hamilton, Ontario found resistance to ciprofloxacin in less 
than 1 percent of 462 isolates (Edge & Hill, 2005).  Studies prior to this one also 
 
 111 
found resistance to ciprofloxacin of less than 1 percent in E. coli isolates from surface 
water (Ash et al., 2002). 
Vancomycin was used as a negative control in this study.  Because 
vancomycin specifically targets Gram-positive bacteria, whereas E. coli is Gram-
 negative, it was expected for E. coli to exhibit total resistance towards vancomycin 
irrespective of any kind of developed resistance.  The exact reasons for the 
occurrence of only 87 percent resistance to vancomycin are unknown, but one 
possibility may be a genetic mutation. 
 
4.5.2 Multidrug Resistance 
Multidrug resistance refers to the situation in which a single bacterial colony 
simultaneously exhibits resistance to multiple antibiotics.  The vast majority of E. coli 
isolates from this study were multidrug resistant.   
Figure 33 shows multidrug resistance counts for the 23 colonies that were 
tested against all six antibiotics.  All but one colony was resistant to more than one 
antibiotic, and more than half of the isolates (63 percent) were resistant to three or 
more antibiotics.  No isolate was resistant to all six of the antibiotics tested; this 
would have been impossible, given that no E. coli isolates were found to be resistant 
to ciprofloxacin. 
 
 
 112 
Figure 33: Multidrug resistance for E. coli isolates that were tested against all six 
antibiotics 
Multidrug Resistance
 0
 2
 4
 6
 8
 1 2 3 4 5 6
 Number of Antibiotics
 N
 u
 m
 b
 e
 r 
o
 f 
S
 a
 m
 p
 le
 s 
R
 e
 si
 st
 a
 n
 t 
(n
  =
  2
 3
 )
  
Of the 23 E. coli isolates tested against all six antibiotics used in this study, 
the overwhelming majority showed simultaneous resistance to more than one 
antibiotic.  Twenty-two colonies were resistant at least two antibiotics, 21 colonies 
were resistant to three or more antibiotics, and 13 colonies were resistant to four or 
more antibiotics. 
 
 
With respect to multi-drug resistance, our results indicate greater resistance 
than do those of previous studies.  A survey of bacterial resistance of 250 isolates 
from rural water sources in West Virginia conducted by McKeon et al. (1995) noted 
that approximately 78 percent of all isolates exhibited resistance to multiple 
antibiotics.  However, the study included not only E. coli, but also strains of C. 
freundii and Enterobacter cloacae.  In fact, E. coli isolates exhibited the lowest rate 
of multidrug resistance of all the bacterial species studied (only 14 percent).   
 
 113 
Prior studies have investigated the mechanisms of multidrug resistance at a 
molecular level.  Studies of protein interactions in E. coli indicate that there may be a 
link between multi-drug resistance and the protein MsbA, which is involved in ATP 
production (Woebking et al., 2005).  While such molecular analysis is beyond the 
scope of this study, it may be possible to make certain inferences about the genetic 
basis of antibiotic resistance.  For instance, it is interesting to note that resistance to 
certain antibiotics was correlated to higher chances of multidrug resistance: 
approximately 91.30 percent of isolates resistant to erythromycin were resistant to 
three or more antibiotics, while only 47.93 percent of isolates resistant to 
sulfisoxazole were resistant to 3 or more antibiotics.  This may indicate a genetic 
linkage of phenotypes for antibiotic resistance, in that the same 
chromosomes/plasmids may carry the traits for resistance to multiple antibiotics.  
More detailed results of this correlation can be found in Table 13 and Figure 34. 
 
Table 13: Relationship between multidrug resistance and resistance to individual 
antibiotics 
 
  Resistant to? 
If resistant to? No resistance** 1 class 2 classes 3+ classes 
Percent Resistant 
to 3+ classes 
Erythromycin 1 0 1 21 91.30% 
Tetracycline 10 0 0 13 56.52% 
Ampicillin 4 1 0 18 78.26% 
Sulfisoxazole 12 0 0 11 47.83% 
      
* NOTE: Vancomycin and Ciprofloxacin were included when analyzing multidrug resistance, but 
each drug was not specifically evaluated since vancomycin was a control drug and no resistance 
was seen towards ciprofloxacin 
** Implies no resistance to respective drug, therefore multidrug resistance with respect to that drug 
was not analyzed for given isolate 
 
 114 
Figure 34: Relationship between multidrug resistance and resistance to individual 
antibiotics 
Multidrug Resistance Given Resistance to One 
Antibiotic
 0.00%
 20.00%
 40.00%
 60.00%
 80.00%
 100.00%
 Erythromycin Tetracycline Ampicillin Sulfisoxazole
 Antibiotic
 P
 er
 ce
 nt
  R
 es
 is
 ta
 nc
 e
  
 
A recent antibiotic resistance study was conducted on bacterial isolates 
originating from catfish at three fish farms in Vietnam.  Results indicated that 
numerous multidrug resistance profiles included ampicillin (Sarter et al., 2007).  Our 
findings agree with such results ? we found that colonies of E. coli resistant to 
ampicillin were highly likely (78.26 percent) to be multidrug resistant.  
Meanwhile, Sayah et al. (2005) noted in their study that isolates resistant to 
multiple drugs were more likely to exhibit this multidrug resistance if a combination 
of agents including tetracycline was used.  Our results did not indicate a strong 
correlation between resistance to tetracycline and multidrug resistance.  Given our 
limited sample size, it is difficult to make a conclusive argument about this point.  
But the study by Sayah et al. (2005) consisted of isolates from wildlife, surface water, 
and farm environments, among others, and the difference in origins of isolates and 
environmental factors may help explain why our results are somewhat different.  In 
 
 115 
addition, Sayah et al. (2005) explain, ?the multidrug resistance exhibited by E. coli in 
this study could have been the result of independent, simultaneous development of 
resistance to different agents or could have been the result of co-selection of 
resistance determinants.?  This assertion could very well pertain to our samples as 
well.   
 
4.5.3 Antibiotic Resistance Stratified By Study Site 
Antibiotic resistance was also stratified by wetland site.  Percent resistance for 
each antibiotic at each site was calculated, and then an overall ?Site Average 
Resistance? was calculated by taking the percent resistance for all antibiotics at a 
particular site and averaging them together ? see Table 14 below.  Irvine Nature 
Center had the lowest overall level of antibiotic resistance at 30.3%, while the 
Herring Creek Park and Shockley sites had the highest overall resistance at 75.00 
percent.   
 
 
 116 
Table 14: Percentages of resistant E. coli isolated, stratified by wetland site 
Antibiotic Resistance Percentages by Wetland Site 
 Ery. Amp. Sulf. Tet. Cip. Van. Site Average 
Resistance 
Irvine 0 36 73 0 0 73 30.3 
Kinder 0 67 83 17 0 83 41.7 
Bryantown 0 75 75 50 0 100 50.0 
Jackson Lane 67 67 67 17 0 83 50.0 
Waldorf 100 50 0 50 0 100 50.0 
Beehive 100 75 50 25 0 75 54.2 
Piscataway 100 43 33 57 0 100 55.6 
Aud 100 100 33 33 0 100 61.1 
Cumberland 100 100 78 33 0 78 64.8 
Merkle 100 100 40 56 0 100 66.0 
Calvert 100 100 50 50 0 100 66.7 
Herring Creek 100 100 100 100 0 50 75.0 
Shockley 100 100 50 100 0 100 75.0 
The rightmost column, "Site Average Resistance," is an overall average of the 
resistance rates to individual antibiotics at a particular site, and provides an 
indication of that site's overall level of antibiotic resistance.  Sites are arranged in 
order of increasing Site Average Resistance. 
 
Characterizing antibiotic resistance by wetland site is an interesting endeavor, 
but at the same it is extremely important not to place serious weight on the results in 
Table 14.  Stratifying the resistance results of no more than 68 isolates across 13 
wetland sites means the percentages listed for each site in this table are based on a 
miniscule sample size.  There is little to no statistical significance.  To account for 
this limitation when trying to correlate antibiotic resistance to other aspects of 
wetland health and function in this study, all of the resistance data should be 
combined into a single statistical pool and analyzed together ? such an undertaking is 
the subject of the following section.  For future studies, the number of colonies 
isolated and characterized must be significantly larger. 
 
 
 117 
4.6 Statistical Analysis 
We used statistical analysis to determine links between antibiotic resistance 
and other wetland characteristics.  We ran a series of t tests in STATA 9.2 Data 
Analysis and Statistical Software.  We added six dummy variables, one for each 
antibiotic, and assigned values of ?1? for resistant and ?0? for not resistant.  When 
partitioned with respect to ciprofloxacin resistance, every sample was contained in the 
?not resistant? category, so running a t test was impossible.  Results concerning 
erythromycin should also be considered with some reservation, since all but two 
samples showed resistance to erythromycin.  We ran t tests to see if separating our 
samples in this way led to statistically significant differences in wetland 
characteristics.  Results show some links between antibiotic resistance and measures 
of water quality, soil quality, and surrounding land use, but no links with measures of 
vegetation diversity 
 
 
 118 
4.6.1 Relationships Amongst Wetland Health Factors 
 
Table 15: Statistically significant relationships found amongst wetland health factors 
Wetland Health Factor Wetland Health Factor P ? value Direction 
Organic Matter Thickness Gleying 0.024 + 
Organic Matter Thickness Surface NH4 0.003 + 
Organic Matter Thickness Surface Total N <0.001 + 
Organic Matter Thickness Surface Organic N 0.012 + 
Organic Matter Thickness Subsurface NO2 0.001 + 
Soil Depth Gleying 0.006 + 
Soil Depth Soil Available P 0.044 - 
Mottleing Gleying <0.001 + 
Mottleing Soil Total C 0.035 - 
Mottleing Soil C-N ratio 0.029 - 
Mottleing Soil Total P 0.044 + 
Soil C-N ratio Soil Total P <0.001 - 
Soil Total N Soil Total C <0.001 + 
Soil Available P Surface NH4 0.034 + 
Soil Available P Surface Total P <0.001 + 
Surface NH4 Surface NO2 0.016 + 
Surface NH4 Surface Total N <0.001 + 
Surface NO2 Surface Total N 0.008 + 
Surface NO2 Surface NO3 <0.001 + 
Surface NO2 Surface Organic N 0.013 + 
Surface NO3 Subsurface NO2 + NO3 <0.001 + 
Surface NO3 Subsurface Total N 0.007 + 
Surface NO3 Subsurface NO2 0.006 + 
Subsurface NO2 + NO3 Subsurface Total N <0.001 + 
Subsurface NO2 + NO3 Subsurface Organic N <0.001 + 
Subsurface NO3 Subsurface Organic N <0.001 + 
?Surface? refers to data from surface water samples, ?subsurface? refers to data 
from subsurface water samples, and ?soil? refers to data from soil samples. 
 
 
 
 Organic matter thickness has a positive significant relationship to surface NH4 
surface total nitrogen, surface organic nitrogen, and subsurface NO2.  An increase in 
the nitrogen levels in a wetland cause an increase in plant and animal abundance.  
This in turn can cause the organic matter thickness to increase.  
 Some of the very significant relationships  (p < 0.001) are fairly obvious and 
intuitively understandable. Mottleing and gleying are both indicators of wetland soils, 
so one does not usually appear without the other in a true wet environment.  Soil total 
 
 119 
nitrogen and soil total carbon are also closely related.  Wetlands tend to retain carbon 
and a more productive wetland (with more total nitrogen) will be producing more 
organic matter to eventually add to the soil total carbon.  The other very significant 
relationships (Surface NH4 to surface total nitrogen, surface NO2 to surface NO3, and 
subsurface NO2 + NO3 to subsurface total nitrogen) all involve nitrogen levels and are 
positively correlated as would be expected from a well functioning wetland.  
 
4.6.2 Relationships to Land Use 
 
Table 16: Statistically significant relationships found between wetland health factors 
and land use factors 
Wetland Health Factor Land Use Factor P - value Direction 
Organic Matter Thickness Agriculture ? 1000 m buffer 0.002 + 
Organic Matter Thickness Agriculture ? 2000 m buffer 0.056 + 
Mottleing Agriculture ? 1000 m buffer 0.048 + 
Soil C-N ratio Nature ? 1000 m buffer 0.004 + 
Soil Total P Agriculture ? 1000 m buffer 0.023 + 
Soil Total P Natural ? 1000 m buffer 0.041 - 
Surface NH4 Agriculture ? 1000 m buffer 0.014 + 
Surface Total N Agriculture ? 1000 m buffer 0.018 + 
Surface Organic N Agriculture ? 1000 m buffer 0.001 + 
Subsurface NO2 + NO3 Agriculture ? 2000 m buffer 0.003 + 
Subsurface NO3 Agriculture ? 2000 m buffer 0.005 + 
?Agriculture? and ?Nature? refer respectively to the amount of agricultural and 
natural land use found in a respective buffer zone. 
 
 
 The significant relationships found between land use factors and wetland 
health factors are detailed in Table 16.  All but two of the significant relationships we 
found between land use factors and wetland health factors involved agricultural land 
use.   The two outliers involved natural land use.  Interestingly, there were no 
significant relationships found between the Land Development Index (LDI) and any 
wetland health factors.  This indicates that the LDI is not a good predictor of soil or 
 
 120 
water nutrient levels in a mitigation wetland when used with a 1000 or 2000 meter 
radius land use buffer.  
 Both the 1000 meter and 2000 meter agricultural land use showed a 
significant positive relationship to organic matter thickness.  Agriculture is a large 
contributor to the nitrogen influx a wetland sees and nitrogen levels directly influence 
the productivity of a wetland.  As mentioned previously, a more productive wetland 
would accumulate more organic matter, and thus the soil organic matter thickness 
would increase over time.  This result is further verified by the fact that it was 
repeated for both size buffers and thus the chance of it being a false positive (a 
problem for studies with small sample sizes) is significantly reduced.  A confounding 
variable that was not examined by this study, but influences organic matter thickness 
substantially is the age of the mitigation wetland in question, i.e. an older wetland 
should have a thicker organic layer.  
 The assertion that agricultural land use increases nitrogen content in a wetland 
is bourn out by other significant relationships found in this study.  There are 
significant positive relationships between agricultural land use on either the 1000 
meter or 2000 meter level and surface total nitrogen, surface organic nitrogen, surface 
ammonia, subsurface nitrate + nitrite, subsurface total nitrogen, subsurface nitrite, and 
subsurface organic nitrogen. 
 An interesting but also expected result is the relationship between soil total 
phosphorous and natural and agricultural land use.  Agricultural land use increases 
soil total phosphorous, probably because of runoff from livestock yards and some 
 
 121 
fertilizers.  We would then expect to see a decrease in soil total phosphorous when a 
wetland is set in a more natural setting and the data bears this assumption out.  
 
4.6.3 Relationships to Antibiotic Resistance 
T-tests were performed to compare bacterial resistance data to other wetland 
measurements, within the categories of land use, soils, water quality, and vegetation.  
Figure 18 through 20 and 23 through 25 in the appendices provide lists of all 
variables included in statistical analysis.  The variable for surface water nitrate levels 
was omitted from this analysis due to insufficient data.  Furthermore, due to the 
abundance of plant species found in this study, vegetation data was not included on a 
plant-by-plant basis ? instead, Shannon Diversity Index (described in Section 4.2.3: 
Shannon Diversity Index) was used to consolidate all vegetation data into a single 
variable representing overall plant diversity.  Note also that the antibiotics 
vancomycin and ciprofloxacin were not included in this statistical analysis.  
Vancomycin was not included because it was a negative control.  Ciprofloxacin was 
not included because there were not enough isolates in the ciprofloxacin-resistant 
group to mathematically allow for a t-test to be performed. 
 Because there were four antibiotics on which t-tests were run (tetracycline, 
sulfisoxazole, erythromycin, and ampicillin) and 31 variables to run them against, a 
total of 124 t-tests were performed for this portion of data analysis.  Table 17 shows 
the results of all such tests. 
 
 
 122 
Table 17: Results of 124 t-tests showing all statistically significant relationships 
between antibiotic resistance and various wetland indicators 
  Bacterial Resistance to Antibiotics 
  Ampicillin Sulfisoxazole Erythromycin Tetracycline 
Soil Organic Matter Thickness   0.0682  
Soil Depth  0.0768   
Soil Mottling  0.0773   
Soil Gleying     
Soil Total N   0.0104 0.0579 
Soil Total C   0.0237 0.0981 
Soil C:N Ratio     
Soil Total P     
Soil Available P     
Surface Water Total N  0.0766   
Surface Water Ammonia  0.0705   
Surface Water Organic N   0.0042 0.0290 
Surface Water Nitrite     
Surface Water Nitrite + Nitrate   0.0382  
Surface Water Total P     
Subsurface Water Total N     
Subsurface Water Ammonia  0.0436   
Subsurface Water Organic N     
Subsurface Water Nitrite     
Subsurface Water Nitrite + 
Nitrate   0.0302  
Subsurface Water Nitrate   0.0236  
Subsurface Water Total P  0.0522   
Plant Shannon Diversity Index     
Urban Land Use (1000 m)     
Urban Land Use (2000 m)     
Agricultural Land Use (1000 m)   0.0071 0.0754 
Agricultural Land Use (2000 m)   0.0691  
Natural Land Use (1000 m)   0.0423 0.0138 
Natural Land Use (2000 m)   0.0361 0.0046 
Landscape Development 
Intensity (1000 m)     
L
 an
 d 
U
 se
 , V
 eg
 et
 ati
 on
 , W
 at
 er
 , a
 nd
  S
 oil 
M
 ea
 su
 re
 m
 en
 ts
  
Landscape Development 
Intensity (2000 m)    0.0640 
Shaded boxes represent significant (p < 0.05) or almost significant (0.05 < p < 0.10) 
correlations, with two-tailed p-values listed to four decimal places.  Ciprofloxacin 
was excluded from this analysis due to insufficient data in its resistant group.  
Abbreviations: C = carbon, N = nitrogen, P = phosphorous. 
 
 
 123 
 Ampicillin was most notable for its lack of statistically significant 
relationships to other measurements.  While erythromycin, sulfisoxazole, and 
tetracycline each had six or more significant (p < 0.05) or almost significant (0.05 < p 
< 0.10) relationships to other variables, ampicillin had none.  The lack of statistically 
significant relationships for ampicillin may be due to the small sample size of isolates 
tested ? especially since ampicillin had only three isolates in its resistant group.  If the 
lack of relationships is due to something more than sample size considerations, a 
speculative conclusion might be that bacterial resistance genes to ?-lactam antibiotics 
have now become so widespread that they are essentially present everywhere, 
irrespective of other environmental factors.  Indeed, ?-lactams have been in 
widespread use for decades, and ampicillin had the second highest rate of resistance 
out of the five non-control antibiotics in this study. 
 Sulfisoxazole had correlations to several soil and water quality measurements, 
but no correlations to land use and vegetation metrics.  Resistance to sulfisoxazole 
had an almost significant relationship (0.05 < p < 0.10) with soil depth and soil 
mottling.  As described in Section 4.6.1: Relationships Amongst Wetland Health 
Factors, soil depth and mottling are themselves strongly correlated with each other, so 
here they should be considered together.  Resistant isolates tended to occur in 
wetlands with deeper soil layers and higher levels of mottling, both of which tend to 
be traits of properly functioning, anaerobic wetlands.   
Sulfisoxazole resistance also had correlations to several water quality 
measurements, which were somewhat stronger than its correlations to soil 
measurements.  Resistant isolates tended to be found at sites with higher levels of 
 
 124 
surface water ammonia and surface water total nitrogen (0.05 < p < 0.10), lower 
levels of subsurface water ammonia (p < 0.05), and lower levels of subsurface water 
total phosphorous (0.05 < p < 0.10).  Without further study, it would be difficult to 
say whether such relationships are the result of bacterial resistance affecting wetland 
biogeochemical cycling, the result of wetland biogeochemistry affecting sulfisoxazole 
resistance, or whether there is some lurking cause (such as the input of runoff 
containing both nutrients and pharmaceuticals) affecting both factors simultaneously.  
It goes without saying that any causal scheme is certain to be complex. 
 Erythromycin resistance had significant correlations to several nitrite (NO2
 -) 
and nitrate (NO3
 -) water quality measurements.  Resistance tended to occur in 
wetlands with lower subsurface water nitrate levels (p < 0.05), lower subsurface water 
nitrite + nitrate levels (p < 0.05), and lower surface water nitrite + nitrate levels (p < 
0.05).  Nitrite and nitrate ions are produced during the process of nitrification, in 
which ammonium nitrogen is oxidized in the upper aerobic layer of a wetland.  
Therefore, the correspondence between resistance and lower levels of nitrite and 
nitrate may be an indication that resistance tends to occur in more anaerobic wetlands, 
in which the balance of nitrogen is towards more reduced forms like molecular 
nitrogen (N2 ? produced from nitrate in the anaerobic process of denitrification) and 
ammonia.  It is interesting to note that this aligns with the result for sulfisoxazole 
resistance, which tended to be found in more anaerobic soils. 
 Nitrification occurs via the actions of Gram-negative microbes such as 
Nitrosomonas sp. and Nitrobacter sp. (Gomez, Mendez, & Lema, 1996).  Since 
erythromycin has actions against Gram-negative bacteria (Costanzo, Murby, & Bates, 
 
 125 
2005), it is possible that nitrification is correlated to erythromycin resistance by some 
microbial-level link.  One possibility is that the presence of erythromycin has some 
direct effect on the functions or populations of these microbes.  A study by Costanzo 
et al. (2005) found that high doses of erythromycin were capable of reducing the rate 
of bacterial denitrification in aquatic sediments.  That experiment, however, tested 
only the short-term effects of a high concentration (1000 ?g/L) of antibiotic, and did 
not consider the long-term effects of lower concentrations, nor did it attempt to 
correlate these results to the occurrence of bacterial resistance.  A second possibility 
for a microbial link between erythromycin resistance and nitrification is that 
nitrifying bacteria, in addition to their functions in nitrogen cycling, also have some 
sort of function in attenuating the spread of bacterial resistance.  In this study, lower 
levels of erythromycin resistance corresponded to higher levels of nitrite and nitrate, 
and higher levels of these two molecules would intuitively correspond to properly 
functioning populations of nitrifying bacteria.  Thus, the proper functioning of 
nitrifying bacteria might aid in reducing the spread of bacterial resistance. 
 The most striking relationships found in this data analysis were those shared 
by erythromycin and tetracycline; for both antibiotics, resistance was correlated to 
natural and agricultural surrounding land use, organic nitrogen levels in surface 
water, and total nitrogen and carbon levels in soil.  As shown in Table 17, the fact that 
land use was measured within both 1000 m and 2000 m buffers means that there were 
a total of eight possible relationships between erythromycin/tetracycline resistance 
and agricultural/natural land use.  Seven out of eight of these relationships turned out 
to be statistically significant (p < 0.05) or almost significant (0.05 < p < 0.10).  Prior 
 
 126 
to the execution of this study it was hypothesized that since humans are a major cause 
of the spread of antibiotic resistance, bacterial resistance would tend to occur at 
wetland sites that were more impacted by human development.  Thus it was expected 
that natural land use would be associated with bacterial susceptibility, while urban 
and agricultural land use would be associated with bacterial resistance.  Instead, the 
trends observed in this study were the opposite of those hypothesized; erythromycin 
and tetracycline resistance were found in wetlands with lower agricultural land use 
and higher natural land use.  It may be that the relationship between land use and 
bacterial resistance is more complicated than initially anticipated, or it may be that 
land use does not play a role as significant as we previously thought. 
Resistance to erythromycin and tetracycline was also associated with higher 
levels of soil total nitrogen and soil total carbon, as well as higher levels of surface 
water organic nitrogen.  A visual depiction of the relationship between bacterial 
tetracycline resistance and soil total nitrogen & carbon is shown in Figure 35.  
Statistical significance values for erythromycin relationships were stronger than those 
for tetracycline, but tetracycline was selected for this graph because the relatively 
even split between tetracycline-susceptible and tetracycline-resistant isolates made for 
easier visualization. 
 
 
 
 127 
Figure 35: Bacterial resistance to tetracycline plotted against soil nutrient levels 
Tetracycline Resistance vs. Soil Nutrient Levels
 0
 10000
 20000
 30000
 40000
 0 500 1000 1500 2000
 Soil Total Nitrogen (mg/kg)
 S
 oi
 l T
 ot
 al
  C
 ar
 bo
 n 
(m
 g/
 kg
 )
 Resistant Susceptible
  
Each of the 29 points represents a different plot at which bacterial resistance data 
was collected.  Plots in which at least one bacterial colony was resistant are marked 
with grey squares, while plots in which all bacterial colonies were susceptible are 
marked with outlined white triangles.  Soil total nitrogen and carbon levels are on the 
axes and thus increase upwards and to the right. 
 
 
 
In Figure 35, sites exhibiting tetracycline resistance tend to be clustered 
towards the bottom left, where there are lower levels of soil total nitrogen and carbon.  
Meanwhile, sites with susceptible isolates tend more towards the upper right.  The 
interesting point here is that organic nitrogen, total nitrogen, and total carbon are all 
indicators of productive wetland systems.  Therefore, sites with resistant isolates 
tended to be less productive, while sites with susceptible isolates tended to be more 
productive.  The relationship held true for both tetracycline and erythromycin ? more 
productively functioning wetland systems were associated with less bacterial 
resistance. 
 
 128 
Assuming that productive wetlands have larger, healthier, and more diverse 
natural microbial populations, it might be possible that wetland microbes are actually 
responsible for mitigating the spread of antibiotic resistance.  Perhaps natural bacteria 
are part of some biogeochemical process that results in the degradation of antibiotics, 
reducing the spread of antibiotic resistance.  Or perhaps simple population dynamics 
are at work, dictating that it is more difficult for bacterial resistance genes to spread 
and take hold in the presumably larger bacterial populations of productive wetland 
systems.  Any causational relationship between wetland productivity and reduced 
bacterial resistance could also help to explain the seemingly counterintuitive results 
of the agricultural/natural land use correlations to tetracycline and erythromycin: 
increased agricultural land use around a wetland (and the associated loss of natural 
land) would be associated with increased nutrients from agricultural runoff, which 
could cause increases in wetland productivity and an associated decrease in antibiotic 
resistance. 
 Of course, the present study was an observational study and thus it is 
impossible to determine actual causes behind the relationships we found.  Further 
studies must be performed to gain a better understanding of the biogeochemical 
mechanisms underlying our results.  Controlled microcosm studies might be 
particularly useful.  For example, wetland microcosms could be designed in which 
nutrients such as nitrogen and phosphorous are varied while different antibiotics are 
simultaneously fed into the system at low concentrations.  Resulting microbial 
processes could then be monitored by a variety of methods: microbial population 
could measured by culturing water or sediment samples and obtaining colony counts, 
 
 129 
rates of resistance to different antibiotics could be measured as was done in the 
present study, time decay of antibiotics themselves could be monitored via analytical 
measurements, and bacterial transfer of resistance genes and associated nucleic acid 
plasmids could be studied via biochemical methods.  The presence of a control 
system would help to isolate specific causal factors. 
 The inability to determine causal relationships should not diminish the 
importance of the numerous correlations found in this study.  Definitive trends were 
noted in the results, and they are especially important due to the relative paucity of 
previous research in this area. 
 
 130 
5. Conclusion 
 
The first important conclusion reached by this study is that Maryland?s 
mitigation wetlands on the whole appear to be functioning properly.  At the 13 
mitigation sites we visited, field data in the categories of water quality, soils, and 
vegetation tended to be indicative of productive wetland systems.  For example, 12 of 
13 wetland sites had a wetland vegetation prevalence index indicative of a wetland 
ecosystem.  Every site had a least some degree of mottling and/or gleying, indicative 
of an anaerobic environment.  Finally, nutrient levels in water sample were generally 
on par with those found in a similar study of wetlands in Florida.  The proper 
functioning of the sites in this study was particularly important, because as mitigation 
wetlands they were designed to replace lost wetlands, and if they didn?t function as 
wetlands then it would be impossible for them to serve their regulatory purpose. 
 Certain sites had anomalous characteristics.  For example, the Bryantown and 
Aud sites had unusually high levels of surface water ammonia, surface water total 
phosphorous, and surface water total nitrogen.  Because this trend encompassed 
several different measurements and multiple sampling dates, it can be reasonably 
assumed that it is the result of some unique characteristic of these two sites and not 
the result of anomalous water samples.  Another example of an unusual site was the 
Kinder site, which was unique for its high wetland vegetation prevalence index, 
indicating that it was the only site in this study that could not be considered a wetland 
based on its vegetation.   
The interesting point about Kinder is that a simple glance at it potentially 
foretold the same information learned through intensive vegetation sampling; upon 
 
 131 
our first visit to the site we immediately noted that it had no standing water and was 
instead dominated by dry vegetation and ticks.  Proper hydrology is one of the most 
fundamental prerequisites to a functioning wetland and without sufficient ground 
saturation, Kinder simply could not support the necessary hydrophytic wetland 
vegetation.  Mitigation sites must be designed with this consideration in mind, for it 
would be wasteful to invest time and money attempting to create a wetland site 
without ensuring the proper hydrology to support it. 
 The second major conclusion of this study is that land use has significant 
relationships to field indicators of wetland health.  The vast majority of statistically 
significant correlations found in this study were between agricultural land use and 
nutrient levels.  Furthermore, the directions of the correlations we found agreed with 
our hypothesis; increasing agricultural land use, i.e. increasing human impact, 
correlated to increases in surface water total nitrogen, surface water organic nitrogen, 
and surface water ammonia, among other variables. 
 Land use correlations are intellectually interesting for their ability help 
elucidate the potential impacts of human development on wetland health, but they 
also serve a more practical purpose.  In the future, deeper integration of land use 
studies with field measurements of wetland health will most likely help to make 
wetland monitoring strategies more efficient.  Characterization of wetland sites based 
on surrounding land use can be made without visiting the sites themselves, so that in 
the future wetland managers can use existing correlations to help determine which 
sites need might need more in-depth attention based solely on their land use 
 
 132 
characteristics.  Such a system would allow limited resources to be used to greater 
ends in the goal of environmental protection. 
 A third conclusion of this study is that antibiotic resistance is present in 
Maryland wetlands.  Antibiotic resistance was found at all thirteen sites studied, with 
overall levels that were equal to or higher than levels previously reported in the 
environment.  Trends for individual antibiotics were similar to those reported in 
previous studies, with high rates of resistance to erythromycin and ampicillin, 
intermediate rates of resistance to sulfisoxazole and tetracycline, and a very low rate 
of resistance to ciprofloxacin. 
 The widespread prevalence of antibiotic resistance in Maryland?s wetlands is 
interesting because historically, antibiotic resistance has been studied only in more 
traditional locales like hospitals.  Meanwhile, the release of antibiotics and other 
pharmaceuticals into the environment has been largely overlooked.  However, this 
and other studies demonstrating the presence of high rates of antibiotic resistance in 
the environment reveal that the role of antibiotics in the natural world merits closer 
examination.  The fact that environmental antibiotic resistance rates are similar to 
hospital resistance rates suggests that genes for bacterial resistance are present 
everywhere, and that the effects of antibiotic resistance cannot be escaped in any 
locale. A final conclusion arising from this study is that the occurrence of antibiotic 
resistance has a complicated relationship to wetland health and function. In general, 
more productive wetlands tended to have lower rates of antibiotic resistance.  It seems 
that while wetlands are sites where antibiotic resistance occurs, they may also have 
the ability to ?filter? out antibiotic resistance in the same way that they filter out other 
 
 133 
pollutants.  This result is a especially exciting because it suggests at a previously 
unstudied function for wetlands. 
 A limiting factor of this study was its sample size, which made it difficult to 
stratify data and to find statistically significant relationships.  Beyond that, in a 
system as complex as a wetland, it is difficult to determine causal relationships ? 
microcosm studies would be particularly helpful in the future.  The present study was 
an attempt, through an observational methodology, to fill in a significant gap in the 
state of current scientific knowledge, and it revealed a number of open questions for 
future research on the ever-changing relationship between humans and wetlands. 
 
 
 134 
References 
 
AB BIODISK. (n.d.). Etest: novel antimicrobial gradient technique. Retrieved  
April 20, 2008, from http://www.abbiodisk.com/ 
 
Ahrer, W., Scherwenk, E., & Buchberger, W. (2001). Determination of drug  
residues in water by the combination of liquid chromatography or capillary  
electrophoresis with elexctrospray mass spectrometry. Journal of  
Chromatography, 910, 69 ? 78.  
 
Akinbowale, O.L., Peng, H., & Barton, M.D. (2006). Antimicrobial resistance in  
bacteria isolated from aquaculture sources in Australia. Journal of Applied 
Microbiology, 100, 1103 ? 1113.  
 
Aksaray, S., Dokuzo?uz, B., G?vener, E., Y?cesoy, M., Yulu?, N., Kocag?z, S.,  
& et al. (2000). Surveillance of antimicrobial resistance among gram-negative 
isolates from intensive care units in eight hospitals in Turkey. Journal of 
Antimicrobial Chemotherapy, 45, 695 ? 699.  
 
Albert, D.A., & Minc, L.D. (2004). Plants as regional indicators of Great Lakes  
coastal wetland health.  Aquatic Ecosystem Health & Management, 7(2), 233 
? 247. 
 
Andersen, S.R., & Sandaa, R.A. (1994).  Distribution of tetracycline resistance  
determinants among gram-negative bacteria isolated from polluted and 
unpolluted marine sediments. Applied and Environmental. Microbiology. 60, 
908 ? 912.  
 
Anderson, J.R., Hardy, E.E., Roach, J.T., & Witmer, R.E. (1976). A land use and  
land cover classification system for use with remote sensor data. Washington, 
D.C.: U.S. Government Printing Office.   
 
Ash, R.J., Mauck, B., & Morgan, M. (2002). Antibiotic resistance of gram- 
negative bacteria in rivers, United States. Emerging Infectious Diseases, 8, 
713 ? 716.  
 
Balcombe, C., Anderson, J., Fortney, R., Rentch, J., Grafton, W., & Kordek, W.  
(2005). A comparison of plant communities in mitigation and reference 
wetlands in the mid-Appalachians. Wetlands, 25 (1), 130 ? 142. 
 
Barber, L.B., Keefe, S.H., Antweiler, R.C., Taylor, H.E., & Wass, R.D. (2006).   
Accumulation of contaminants in fish from wastewater treatment wetlands.  
Environmental Science and Technology, 40(2), 603-611. 
 
 
 
 
 135 
Barber, L.B., Murphy, S.F., Verplank, P.L., Sandstrom, M.W., Taylor, H.E., &  
Furlong, E.T.  (2006). Chemical loading into surface water along a 
hydrological, biogeochemical, and land use gradient: a holistic watershed 
approach. Environmental Science and Technology, 40(2), 475-486. 
 
Bartoloni, A., Pallecchi, L., Benedetti, M., Fernandez, C., Vallejos, Y., Guzman,  
E., & et al. (2006). Multidrug-resistant commensal Escherichia coli in 
children, Peru and Bolivia. Emerging Infectious Diseases, 12, 907 ? 913.  
 
Bauer, A.W., Kirby, W.M., Sherris, J.C., & Turck, M. (1966). Antibiotic  
susceptibility testing by a standardized single disc method. American Journal 
of Clinical Pathology, 45, 493 ? 496.  
 
Bedner, M., & MacCrehan, W.A. (2006). Transformation of acetaminophen by  
chlorination produces the toxicants 1,4-benzoquinone and N-acetyl-p-
 benzoquinone imine. Environmental Science and Technology. 40(2), 516 ? 
522. 
 
Bedford, B. L., Walbridge, M. R., & Aldous, A. (1999). Patterns in nutrient 
availability and plant diversity of temperate North American wetlands. 
Ecology, 80(7), 2151-2169. 
 
Biyela, P.T., Lin, J., & Bezuidenhout. (2004). The role of aquatic ecosystems as  
reservoirs of antibiotic resistant bacteria and antibiotic resistance genes. Water 
Science & Technology, 50(1), 45 ? 50. 
 
Bonham, C. D. (1989). Measurements for Terrestrial Vegetation. New York: Wiley.  
 
Booth, A.M., Hagedorn, C., Graves, A.K., Hagedorn, S.C., & Mentz, K.H.  
(2003). Sources of fecal pollution in Virginia?s Blackwater River.  Journal of  
Environmental Engineering, 129(6), 547 ? 552. 
 
Boxall, A.B.A., Kolpin, D.W., Halling-S?rensen, B., & Tolls, J.  (2003).  Are  
veterinary medicines causing environmental risks?  Environmental Science &  
Technology, 37(15), 286A-294A. 
 
Bremner, J.M. (1996). Nitrogen-Total. In J. M. Bigham et al. (Eds.), Methods of soil 
analysis, part 3: chemical methods (pp. 1085-1121). Madison: Soil Science 
Society of America and American Society of Agronomy. 
 
Brinson, M. & Rheinhardt, R. (1996). The role of reference wetlands in functional  
assessment and mitigation. Ecological Applications, 6, 69 ? 76. 
 
Brown, L.. (1992). Grasses: An identification guide. Boston: Houghton Mifflin.  
 
 
 136 
Brown, M.T. & Reiss, K.C. (2005). The Florida wetland condition index  
(FWCI): Developing biological indicators for isolated depressional forested 
wetlands. Gainesville: The University of Florida. 
 
Bruchet, A., Hochereau, C., Picard, C., Decottignies, V., Rodrigues, J.M., &  
Janex-Habibi, M.L. (2005). Analysis of drugs and personal care products in 
French source and drinking waters: the analytical challenge and examples of 
application.  Water Science & Technology, 52(8), 53 ? 61. 
 
Burkett, V. & Kusler, J. (2000). Climate change: Potential impacts and  
interactions in wetlands of the United States. Journal of the American Water 
Resources Association, 36(2), 313 ? 320. 
 
Carballa, M., Omil, F., Lema, J.M., Llompart, M., Garcia, C., Rodriguez, I., & et  
al. (2005). Behaviour of pharmaceuticals and personal care products in a 
sewage treatment plant of northwest Spain. Water Science & Technology, 
52(8), 29- 35. 
 
Chapman, D. (1996). Water quality assessments: A guide to use of biota,  
sediments and water in environmental monitoring. London: Spon Press. 
 
Chee-Sanford, J.C., Aminov, R.I., Krapac, I.J., Garrigues-Jeanjean, N., & Mackie,  
R.I. (2001). Occurrence and diversity of tetracycline resistance genes in  
lagoons and groundwater underlying two swine production facilities. Applied 
and Environmental Microbiology, 67, 1494 ? 1502.  
 
Clark, N.M., Patterson, J., & Lynch J.P. 3rd. (2003). Antimicrobial resistance  
among gram-negative organisms in the intensive care unit. Current Opinion in 
Critical Care, 9? 413 ? 423. 
 
Clesceri, L. S., Greenberg, A. E., & Eaton, A. D. (1998).  Standard methods for the 
examination of water and wastewater (20th ed.).  Washington, D.C.: 
American Public Health Association. 
 
Cleuvers, M. (2003). Aquatic ecotoxicity of pharmaceuticals including the  
assessment of combination effects. Toxicology Letters, 142(3), 185-194 
 
Cohen, M.J., Lane, C.R., Reiss, K.C., Surdick, J.A., Bardi, E., & Brown, M.T.  
(2005). Vegetation based classification trees for rapid assessment of isolated 
wetland condition. Ecological Indicators, 5(3), 189 ? 206. 
 
Costanzo, S.D., Murby, J., & Bates, J. (2005). Ecosystem response to antibiotics  
entering the aquatic environment. Marine Pollution Bulletin, 51(1-4), 218- 
223. 
 
 
 137 
Cowardin, L.M., Carter, V., Golet, F.C., & LaRoe, E.T. (1979). Classification of  
 wetlands and deepwater habitats of the United States. Washington, D.C.:  
   U.S.Fish and Wildlife Service and U.S. Department of the Interior.   
 
Cronk, J.K., & Fennessy, M. S. (2001). Wetland plants: Biology and ecology. 
 Washington, D.C.: Lewis. 
 
Cronk, J.K. & Mitsch, W.J. (1994). Aquatic metabolism in four newly constructed  
freshwater wetlands with different hydrologic inputs. Ecological Engineering, 
3, 449 ? 468. 
 
Daughton, C.G., & Ternes, T.A.  (1999).  Pharmaceuticals and personal care  
products in the environment: agents of subtle change?  Environmental Health 
Perspectives, 107(Supplement 6), 907-938. 
 
DeLuca, W.V., Studds, C.E., Rockwood, L.L., & Marra, P.P. (2004). Influence of  
land use on the integrity of marsh bird communities of Chesapeake Bay, USA.  
Wetlands, 24, 837 ? 847. 
 
Dieck, J. J., & Robinson, L.R.. (2004). General classification handbook for  
floodplain vegetation in large river systems. Washington, D.C: U.S. 
Geological Survey. 
 
Dietrich, D.R., Webb, S.F., & Petry, T.  (2002).  Hot spot pollutants:  
pharmaceuticals in the environment.  Toxicology Letters, 131(1-2), 1-4. 
 
Dillow, J.J.A. (1996). Technique for estimating maginiture and frequency of peak 
flow in Maryland (USGS water resources investigations report). Towson, 
MD: U.S. Geological Survey.  
 
Dornbusch K., Nord, C.E., & Olsson, B. (1975).  Antibiotic susceptibility testing  
of anaerobic bacteria by the standardized disc diffusion method with special  
reference to bacteroides fragilis.  Scandinavian Journal of Infectious Diseases, 
7(1), 59 ? 66. 
 
Drew, W.L., Barry, A.L., O?Toole, R., & Sherris, J.C. (1972). Reliability of the  
Kirby-Bauer disc diffusion method for detecting methicillin-resistant strains  
of Staphylococcus aureus, 24(2), 240 ? 247. 
 
Edge, T.A., & Hill, S. (2005). Occurrence of antibiotic resistance in Escherichia  
coli from surface waters and fecal pollution sources near Hamilton, Ontario. 
Canadian Journal of Microbiology, 51, 501 ? 505.  
 
Environmental Systems Research Institute. (n.d.) The guide to geographical 
information systems. Retreived, April 28, 2008, from 
http://www.gis.com/whatisgis 
 
 138 
Estrada-Garc?a, T., Cerna, J.F., Paheco-Gil, L., Vel?zquez, R.F., Ochoa, T.J.,  
Torres, J., & et al. (2005). Drug-resistant diarrheogenic Escherichia coli,  
Mexico. Emerging Infectious Diseases, 11, 1306 ? 1308.  
 
Fatta, D., Nikolaou, A., Achilleos, A., & Meri?, S. (2007). Analytical methods for  
tracing pharmaceutical residues in water and wastewater. Trends in Analytical 
Chemistry? 26? 515 ? 533. 
 
Fennessy, M.S., Jacobs, A.D., & Kentula, M.E. (2004). Review of rapid methods  
for assessing wetland condition. Washington, D.C.: U.S. Environmental 
Protection Agency. 
 
Findlay, S.E.G., Kiviat, E., Niedler, W.C., & Blair, E.A. (2002). Functional  
assessment of a reference wetland set as a tool for science, management and 
restoration. Aquatic Sciences, 64(2), 107 ? 117. 
 
Fishman, M. J. (1993). Methods of analysis by the U.S. Geological Survey National 
Water Quality Laboratory: determination of inorganic and organic 
constituents in water and fluvial sediments (USGS Open-File Report 93-125).  
Denver: U.S. Geological Survey. 
 
Fridkin, S.K., & Gaynes, R.P. (1999). Antimicrobial resistance in intensive care  
units. Clinics in Chest Medicine, 20, 303 ? 316.  
 
Fridkin, S.K., Hill, H.A., Volkova, N.V., Edwards, J.R., Lawton, R.M., Gaynes,  
R.P., & et al. (2002). Temporal changes in prevalence of antimicrobial 
resistance in 23 U.S. hospitals. Emerging Infectious Diseases, 8, 697 ? 701.  
 
Gaudreau, C., & Gilbert, H. (1997). Comparison of disc diffusion and agar  
dilution methods for antibiotic susceptibility testing of Campylobacter jejuni 
subsp. Jejuni and Campylobacter coli. Journal of Antimicrobial 
Chemotherapy, 39, 707 ? 712. 
 
Gersberg, R.M., Elkins, B.V., Lyon, S.R., & Goldman, C.R. (1986). Role of  
aquatic plants in waste water treatment by artificial wetlands. Water  
Resources, 20, 363 ? 368. 
 
Glassmeyer, S.T., & Shoemaker, J.A. (2005). Effects of chlorination on the  
persistence of pharmaceuticals in the environment. Bulletin of Environmental 
Contamination and Toxicology, 74(1), 24-31 
 
Gomez, J., Mendez, R. & Lema, J.M. (1996). The effect of antibiotics on  
nitrification processes: Batch assays. Applied Biochemistry and 
Biotechnology, 57/58, 869-876. 
 
 
 139 
Goni-Urriza, M., Capdepuy, M., Arpin, C., Raymond, N., Caumette, P., &  
Quentin, C. (2000). Impact of an urban effluent on antibiotic resistance of 
riverine Enterobacteriaceae and Aeromonas spp. Applied and Environmental 
Microbiology,66, 125 ? 132. 
 
Gutrich, J. & Hitzhusen, F. (2004). Assessing the substitutability of mitigation  
wetlands for natural sites: Estimating restoration lag costs of wetland 
mitigation. Ecological Economics, 48(4), 409 ? 424. 
 
Halda-Alija, L. (2004). Incidence of antibiotic-resistant Klebsiella pneumoniae  
and Enterobacter species in freshwater wetlands. Applied Microbiology, 39, 
445 ? 450. 
 
Hale, S.S., Paul, J.F., & Heltshe, J.F. (2004). Watershed landscape indicators of  
estuarine benthic condition. Estuaries, 27(2), 283 ? 295. 
 
Halling-S?rensen, B., Nors Nielsen, S., Lanzky, P.F., Ingerslev, F., Holten  
L?tzh?ft, H.C., & J?rgensen, S.E. (1998). Occurence, fate, and effects of  
pharmaceutical substances in the environment - a review. Chemosphere,  
36(2), 357-393. 
 
Harwood, V.J., Brownell, M., Perusek, W., & Whitlock, J.E. (2001).  
Vancomycin-resistant Enterococcus spp. isolated from wastewater and 
chicken feces in the United States. Applied and Environmental Microbiology, 
67, 4930 ? 4933.  
 
Houlahan, J. E., Keddy, P. A., Makkay, K., & Findlay, C. S. (2006). The effects of 
adjacent land use on wetland species richness and community composition. 
Wetlands, 26(1), 79-96. 
 
Huang, C.H., Renew, J.E., Smeby, K.L., Pinkston, K.E., & Sedlak, D.L. (2001).   
Assessment of potential antibiotic contaminants in water and preliminary  
occurrence analysis. Water Resources Update, 120, 30 ? 40. 
 
Hyman, J.M., Thorpe, T.C., Matsumara, P.M., Jeffrey, S.R., Maresch, M.J., &  
Barron, W.G. (2002). U.S. Patent No. 6416969. Washington, D.C.: U.S. 
Patent and Trademark Office. 
 
Jones, O.A.H., Voulvoulis, N., & Lester, J.N.  (2003).  Potential impact of  
pharmaceuticals on environmental health.  Bulletin of the World Health 
Organization, 81(10), 768-769. 
 
Kadlec, R. H., & Knight, R. L. (1996).  Treatment Wetlands.  Boca Raton: Lewis-
 CRC. 
 
 
 140 
Katz, B.D., Luperchio, S.A., & Thorne, G.M. (in press). Detection of daptomycin- 
nonsusceptible strains using the Neo-Sensitab? prediffusion method. 
Diagnostic Microbiology and Infectious Disease. 
 
King, R.S., Beaman, J.R., Whigham, D.F., Hines, A.H., Baker, M. E., & Weller,  
D.E.  (2004). Watershed land use is strongly linked to PCBs in white perch in 
Chesapeake Bay subestuaries. Environmental Science and Technology, 38, 
6546 ? 6552. 
 
Kolpin, D.W., Furlong E.T., Thurman, E.M., Zaugg, S.D., Barber, L.B., &  
Buxton, H.T.  (2002). Pharmaceuticals, hormones, and other organic  
wastewater contaminants in U.S. streams, 1999-2000: A national  
reconnaissance. Environmental Science and Technology, 36(6), 1202 ? 1211. 
 
Kosjek, T., Heath, E., Petrovi?, M., & Barcel?, D. (2007). Mass spectrometry for  
identifying pharmaceutical biotransformation products in the environment.  
Trends in Analytical Chemistry, 26(11), 1076-1085 
 
Kusler, J. A., & Kentula, M. E. (1990). Wetland creation and restoration: The  
status of the science. Washington, D.C.: Island Press. 
 
Levesque, S. (2007). Comparison of antimicrobial resistance of Campylobacter  
jejuni isolated from humans, chickens, raw milk, and environmental water in 
Quebec. Journal of Food Protection, 70, 729 ? 735.  
 
Lilly, J.P. (1993). North Carolina Cooperative Extension Service Publication. 
 
Liu, Q.-T., & Williams, H.E. (2007). Kinetics and degradation products for direct  
photolysis of ?-blockers in water. Environmental Science & Technology,  
41(3), 803-810. 
 
Lorian, V. (Ed.). (2005). Antibiotics in laboratory medicine. New York:  
Lippincott Williams & Wilkins. 
 
Maul, J.D., Schuler, L.J., Belden, J.B., Whiles, M.R., & Lydy,  
M.J. (2006). Effects of the antibiotic ciprofloxacin on stream microbial  
communities and detritivorous macroinvertebrates. Environmental Toxicology 
and Chemistry, 25(6), 1598-1606. 
 
Mausbach, M. J. & Richardson, J.L. (1994). Biogeochemical processes  
in hydric soils. Current Topics in Wetland Biogeochemistry, 1, 68 ?  
127.  
 
McKeon, D.M., Calabrese, J.P., & Bissonnette, G.K. (1995). Antibiotic resistant  
gram-negative bacteria in rural groundwater supplies. Water Research, 29, 
1902 ? 1908.  
 
 141 
Merkey, D.H. (2006). Characterization of wetland hydrodynamics using HGM  
and subclassification methods in southeastern Michigan, USA. Wetlands, 
26(2), 358 ? 367. 
 
Merkley, M., Rader, R.B., McArthur, J.V., Eggett, D. (2004). Bacteria as  
bioindicators in wetlands: Bioassessment in the Bonneville basin of Utah, 
USA. Wetlands, 24(3), 600 ? 607. 
 
Mitsch, W. J., & Gosselink, J.G. (2000).  Wetlands (3rd ed.).  New York: Wiley. 
 
Moglen, G.E. & Beighley, R.E. (2000). Using geographical information systems 
(GIS) to determine the extent of gaged streams in a region. Journal of 
Hydrologic Engineering, 5(2), 190 ? 196.  
 
Moglen, G.E., & Casey, M.J. (1998). A perspective on the use of GIS on hydrologic 
and environmental analysis in Maryland. Infrastructure, 3(4), 15 ? 25.  
 
Munsell Color. (1975). Munsell soil color charts. Munsell color, Kallmorgan  
Corporation, Baltimore, MD. 
 
Murray, P.R., Niles, A.C., & Heeren, R.L. (1987). Comparison of a highly  
automated 5-h susceptibility testing system, the Cobas-Bact, with two  
reference methods: Kirby-Bauer disc diffusion and broth microdilution.  
Journal of Clinical Microbiology, 25, 2372 ? 2377. 
 
National Research Council. (2001). Compensating for wetland losses under the Clean 
Water Act. Washington, D.C.: National Academy Press. 
 
Neu, H.C. (1992). The crisis in antibiotic resistance. Science, 257(5073), 1064- 
1073. 
 
Neuhauser, M.M., Weinstein, R.A., Rydman, R., Danziger, L.H., Karam, G., &  
Quinn, J.P. (2003). Antibiotic resistance among gram-negative bacilli in U.S.  
intensive care units. The Journal of the American Medical Association, 289,  
885 ? 888.  
 
Newcomb, L. (1989). Newcomb?s wildflower guide. Boston: Little, Brown and  
Company. 
 
Oransky, I. (2002). Orvan hess. The Lancet, 360, 1179. 
 
Orosz-Coghlan, P.A., Rusin, P.A., Karpiscak, M.M., & Gerba, C.P. (2006).  
Microbial source tracking of Escherichia Coli in a constructed wetland. Water 
Environment Research. 78(3), 227 ? 232. 
 
 
 142 
Oteo, J., L?zaro, E., de Abajo, F.J., Baquero, F., Campos, J., & Spanish members  
of EARSS. (2005). Antimicrobial-resistant invasive Escherichia coli, Spain. 
Emerging Infectious Diseases, 11, 546 ? 553.  
 
Pauwels, B., Verstraete, W.  (2006).  The treatment of hospital wastewater: an  
appraisal. Journal of Water and Health, 4(4), 405-416 
 
Peet, R.K., Wentworth, T.R., Duncan, R.P., & White, P.S. (1998). The North  
Carolina vegetation survey protocol: A flexible, multipurpose method for 
recording vegetation composition and structure. Castanea, 63, 262 ? 274.  
 
Peterson, R. (1973). Trees and shrubs. Boston: Houghton Mifflin. 
 
Pilon, J. & Santamar?a, L. (2002). Clonal variation in morphological and  
physiological responses to irradiance and photoperiod for the angiosperm 
Potamogeton pectinatus. Journal of Ecology, 90, 859 ? 870. 
 
Ragan, R.M. (1991). A geographic information system to support statewide 
hydrologic and non-point source pollution modeling (FHWA/MD-91/02). 
College Park, MD: Maryland State Highway Administration.  
 
Raloff, J. (1999). Waterways carry antibiotic resistance. Science News, 155, 356.  
 
Reed, P.B. (1988). National list of plants that occur in wetlands: 1988  
Washington. St. Petersburg, FL: U.S. Fish and Wildlife Service. 
 
Reppert, R.T., Sigleo, W., Stakhiv, F., Messuram, L., & Meyers, C. (1979).  
Wetland values: Concepts and methods for wetlands evaluation. Ft. Belvoir,  
VA: Institute for Water Resources and U.S. Army of Engineers. 
 
Richardson, J.L. & Vepraskas, M.J. (eds.). (2000). Wetland soils: Their genesis,  
morphology, hydrology, landscapes and classification. Boca Raton: CRC  
Press. 
 
Runnalls, T.J., Hala, D.N., & Sumpter, J.P.  (2007).  Preliminary studies into the  
effects of the human pharmaceutical Clofibric acid on sperm parameters in  
adult Fathead minnow.  Aquatic Toxicology, 84(1), 111-118 
 
Salvesen, D. (1990). Wetlands: Mitigating and regulating development Impacts.  
Washington, D.C.: Urban Land Institute. 
 
Sarter, S., Nguyen, H.N.K., Hung, L.T., Lazard, J., & Montet, D. (2007).  
Antibiotic resistance in gram-negative bacteria isolated from farmed catfish.  
Food Control, 18, 1391 ? 1396.  
 
 
 143 
Sayah, R.S., Kaneene, J.B., Johnson, Y., & Miller, R. (2005). Patterns of  
antimicrobial resistance observed in Escherichia coli isolates obtained from 
domestic- and wild-animal decal samples, human septage, and surface water. 
Applied and Environmental Microbiology, 71, 1394 ? 1404.  
 
Simon, N.S. (n.d.) Antibiotics and biogeochemical cycling of nutrients in  
Chesapeake Bay sediments. Retrieved March 10, 2006, from  
http://chesapeake.usgs.gov/studiesprojects/Simon_antibiotics_fnl.pdf 
 
Spencer, W. & Bowes, G. (1986). Photosynthesis and growth of water hyacinth  
under CO2 enrichment. Plant Physiology. 82, 528 ? 533.  
 
Stevens, M.L. & Bursick, R. (1990). Washington state hydric soils guidebook.  
Olympia, WA: Washington Department of Ecology. 
 
Summers, A.O. (2002). Generally overlooked fundamentals of bacterial genetics  
and ecology. Clinical Infectious Diseases, 34, S85 ? S92.  
 
Sutherland, W. J. (2006). Ecological census techniques: A handbook. New York:  
Cambridge University Press. 
 
Ternes, T.A. (2001). Analytical methods for the determination of pharmaceuticals  
in aqueous environmental samples. Trends in Analytical Chemistry, 20(8),  
419-434 
 
Tilman, D., & Lehman, C. (2001). Human-caused environmental change: Impacts on 
plant diversity and evolution. Proceedings of the National Academy of 
Sciences, 98(10), 5433-5440. 
 
Tran, T. S., & Simard, R. R. (1993). Mehlich III-Extractable Elements. In M. R. 
Carter (Ed.), Soil sampling and methods of analysis (pp. 43-49). Ann Arbor: 
Canadian Society of Soil Science, Lewis Publishers. 
 
Truant, P. (2005). Agricultural use of antibiotics prompts human health concerns:  
FDA bans controversial poultry drug. Nations Health, 35. Retrieved February 
26, 2008, from http://www.medscape.com/viewarticle/520518.  
 
U.S. Department of Agriculture. (2003). Field indicators of hydric soils in the  
United States: Guide for identifying and delineating hydric soils, version 5.01. 
Retrieved February 29, 2008, from  
ftp://ftp-fc.sc.egov.usda.gov/NSSC/Hydric_Soils/FieldIndicators_v5_01.pdf  
 
U.S. Department of Agriculture, Natural Resources Conservation Service.  
(1999). Soil taxonomy: A basic system of soil classification for making and 
interpreting soil surveys. Washington, D.C.: U.S. Government Printing Office 
 
 
 144 
U.S. Department of Agriculture, Soil Conservation Service. (1991). Hydric  
soils of the United States. Washington, D.C.: U.S. Government Printing 
Office.  
 
U.S. Environmental Protection Agency. (1995).  America's wetlands: our vital link 
between land and water (EPA No. 843-K-95-001).  Office of Water, Office of 
Wetlands, Oceans and Watersheds. 
 
U. S. Environmental Protection Agency. (1972). Clean water act. Retrieved  
April 28, 2008, from 
http://www.epa.gov/compliance/civil/cwa/cwaenfstatreq.html  
 
U.S. Environmental Protection Agency. (1999).  Methods and guidance for analysis 
of water (EPA No. 821-C-99-004). Washington, D.C.: Office of Water, U.S. 
Environmental Protection Agency. 
 
U.S. Environmental Protection Agency. (2007).  Pharmaceuticals and  
personal care products (PPCPs). Retrieved December 22, 2007, from 
http://www.epa.gov/ppcp/ 
 
U. S. Fish and Wildlife Service. (1997). Status and trends of the nations  
wetlands: 1997 report. Retrieved April 19, 2007, from  
www.nwi.fws.gov/bha/SandT/download/SandTPaper.pdf 
 
The University of Arizona. (2000). Pharmaceuticals in our water supplies.  
Arizona Water Resource, 9(1), Retrieved from  
http://cals.arizona.edu/AZWATER/awr/july00/feature1.htm 
 
Vepraskas, M. J. (1994). Redoximorphic Features for Identifying Aquic  
Conditions. Tech. Bulletin 301. North Carolina Ag.Research Service, North 
Carolina State Univ., Raleigh 
 
Vincent, J.L., Bihari, D.J., Suter, P.M., Bruining, J., White, M.H., Nicolas- 
Chanoin, M., & et al. (1995). The prevalence of nosocomial infection in 
intensive care units in Europe. The Journal of the American Medical 
Association, 274, 639 ? 644.  
 
Weaver, W. & Shannon, C.E. (1949). The mathematical theory of communication. 
Urbana, Illinois: University of Illinois.  
 
Wentworth, T.R., Johnson, G.P., & Kologiski, R.L. (1988). Designation of wetlands 
by weighted averages of vegetation data: A preliminary evaluation. Water 
Resources Bulletin. 24, 389 ? 396. 
 
 
 
 
 145 
Appendix A 
 
Table 18: Soil characteristics and nutrient concentrations 
Site & 
Plot 
Number 
Organic 
Matter 
Layer 
thickness 
Water table 
(cm below 
surface) 
Sulfur 
smell 
Presence 
of mottles 
Presence of 
Gleying 
Total 
Nitrogen 
mg/kg dw 
Total 
Carbon 
mg/kg 
dw 
C:N 
Ratio 
Total 
Phosphor
 us mg/kg 
dw 
Plant 
Available 
Phosphorus  
mg/kg dw 
 cm          
           
INC 1 15 Surface 
Water 
--- 6-7* 8-9* 586.31 11701.7 19.96 525.10 9.58 
INC 2 15 15 --- 5* 8-9* 1562.28 33199.3 21.25 382.67 6.80 
INC 3 1 > 25 cm --- 5* 6-7* 405.03 5101.7 12.60 402.22 0.99 
BEE 1 3 > 25 cm --- none, 
orange 
1-2* 420.67 6449.0 15.33 324.00 3.06 
BEE 2 10 > 25 cm --- 3-4* 4-5* 649.32 14324.4 22.06 445.15 6.57 
BEE 3 6 10 --- 0 0 702.39 13695.6 19.50 257.64 13.14 
CCW 1 1 > 25 cm --- 2-3* 0 628.75 9652.1 15.35 532.83 23.47 
CCW 2 4 > 25 cm --- 7-8* 8-10* 658.51 11066.4 16.81 597.45 58.05 
CCW 3 3 > 25 cm --- 2-3* 2-3* 729.83 13189.8 18.07 565.71 44.59 
KIN 1 10 Surface 
Water 
6* 8-9* 7-9* 426.76 4387.2 10.28 635.38 14.23 
KIN 2 8 5 3* 7* 7* 450.48 4496.4 9.98 826.65 34.11 
KIN 3 7 Surface 
Water 
--- 8-10* 8-10* 492.41 5118.2 10.39 490.92 5.71 
CAL 1 2 Surface 
Water 
8 0 0 112.74 639.2 5.67 259.34 0.20 
CAL 2           
BRY 1 5 Surface 
Water 
0 8* 6* 688.47 8989.0 13.06 508.47 5.88 
BRY 2 9 Surface 
Water 
0 7-8* 6-7* 611.91 7838.0 12.81 590.03 9.46 
BRY 3 10 Surface 
Water 
3* 8-9* 9-10* 343.42 4647.2 13.53 249.70 9.23 
MRK 1 10 Surface 
Water 
7* 4* 3* 466.31 10057.4 21.57 143.54 22.50 
MRK 2 11 Surface 
Water 
6 2.50 2.5 694.15 15553.1 22.41 244.60 19.57 
MRK 3 11 Surface 
Water 
--- 0 2.5 687.67 12348.2 17.96 203.46 27.18 
PSC 
1(1A) 
5 Surface 
Water 
2 5* 5* 335.55 5040.8 15.02 220.02 5.38 
PSC 
2(1B) 
9  2 4-5* 4-5* 157.33 1038.2 6.60 786.37 18.00 
PSC 3 8 Surface 
Water 
7 2 2 200.68 3001.4 14.96 293.01 36.62 
SHK 1 4 10 0 0 0 129.38 4455.6 34.44 17.65 14.78 
SHK 2 8 3 1 0 6--8 3515.44 74851.8 21.29 307.58 7.29 
SHK 3 5 > 25 cm 1 2 9 665.42 29836.8 44.84 59.83 8.97 
 
 146 
JLP 1 6 Surface 
Water 
 8 6--9 890.45 16981.6 19.07 330.38 30.45 
JLP 2 8 Surface 
Water 
8 7 6 546.45 10125.9 18.53 264.72 6.67 
JLP 3 12 Surface 
Water 
7 0 0 1495.17 30022.7 20.08 295.36 6.64 
HCP 1 4 > 25 cm 7 9 7 175.95 2026.5 11.52 118.57 17.95 
HCP 2 5 > 25 cm 6 8 58 431.62 7541.0 17.47 161.40 8.23 
HCP 3 8 > 25 cm 7 8 7 554.46 9442.7 17.03 314.27 9.00 
CBL 1 4 > 25 cm 0 6 0 727.93 10957.2 15.05 413.12 1.41 
CBL 2 3 25 0 6 0 820.79 11371.5 13.85 560.29 1.03 
CBL 3 4 > 25 cm 2 7 9 531.97 6721.5 12.64 501.24 1.06 
AUD 1 7 Surface 
Water 
6 3 9--10 732.13 10772.7 14.71 443.57 82.56 
AUD 2 2 > 25 cm 0 5   462.67 6541.3 14.14 398.06 58.94 
AUD 3 9 6 5 2 1--2 1019.13 13627.9 13.37 449.39 88.57 
 
 
 
 147 
Appendix B 
 
Table 19: Surface water quality at each plot-site, (BDL-below detection limit) 
Site & Plot 
Number 
Ammonia-N Nitrite-N Total Nitrogen Total 
Phosphorus 
Nitrate Org-N 
 mg-N/L mg-N/L mg-N/L mg-P/L mg-N/L mg-N/L 
       
INC 1       
INC 2 1.078 0.0178 22.62 3.13 B.D.L. 21.54 
INC 3       
BEE 1 0.015 B.D.L. 0.80 0.06 B.D.L. 0.79 
BEE 2 0.282 0.0034 2.06 0.65 B.D.L. 1.77 
BEE 3       
CCW 1 0.027 0.0037 1.64 0.29 B.D.L. 1.61 
CCW 2 0.210 0.0120 6.76 1.54 0.0016 6.54 
CCW 3       
KIN 1       
KIN 2 0.556 0.0097 3.52 0.57 B.D.L. 2.96 
KIN 3 0.518 0.0150 3.46 0.81 0.0013 2.93 
CAL 1 0.052 0.0080 2.25 0.35 0.0882 2.10 
CAL 2 0.012 B.D.L. 0.80 0.12 0.0036 0.79 
BRY 1       
BRY 2       
BRY 3 1.077 0.2104 16.83 10.81 B.D.L. 15.63 
MRK 1 0.200 0.0100 12.20 0.92 B.D.L. 12.00 
MRK 2 0.324 0.0145 6.68 0.93 0.0355 6.31 
MRK 3 0.058 0.0652 3.94 0.10 1.6875 2.13 
PSC 1(1A) 0.073 0.0043 1.17 0.15 B.D.L. 1.10 
PSC 2(1B) 0.148 0.0077 1.66 0.36 B.D.L. 1.50 
PSC 3 0.067 0.0019 1.04 0.42 B.D.L. 0.97 
SHK 1 0.019 0.0035 1.76 0.10 B.D.L. 1.73 
SHK 2 0.012 B.D.L. 0.87 0.04 B.D.L. 0.86 
SHK 3 0.022 0.0042 1.86 0.10 B.D.L. 1.83 
JLP 1 0.087 0.0170 3.19 0.24 B.D.L. 3.09 
JLP 2 0.040 0.0088 2.76 0.35 B.D.L. 2.71 
JLP 3 0.108 0.0106 5.20 0.58 B.D.L. 5.08 
HCP 1 0.037 0.0033 0.96 0.05 0.0058 0.92 
HCP 2 0.034 0.0061 1.02 0.07 B.D.L. 0.97 
HCP 3 0.049 0.0023 5.72 0.38 0.0038 5.67 
CBL 1 0.131 B.D.L. 3.30 0.36 0.0034 3.17 
CBL 2 0.085 0.0022 1.09 0.11 0.0098 0.99 
CBL 3 0.173 0.0038 2.13 0.11 0.0071 1.94 
AUD 1 0.448 0.0057 2.72 48.02 B.D.L. 2.27 
AUD 2       
AUD 3 1.217 0.0077 8.78 6.87 B.D.L. 7.56 
 
 148 
Table 20: Sub-surface water quality at each site-plot, (BDL-below detection limit) 
 
Site & Plot 
Number 
 
Ammonia-N 
 
Nitrite-N 
 
Total Nitrogen 
 
Total 
Phosphorus 
 
Nitrate 
 
Org-N 
 mg-N/L mg-N/L mg-N/L mg-P/L mg-N/L mg-N/L 
       
INC 1 0.8052 1.5876 6.6254 0.221 0.7815 3.4511 
INC 2 0.6754 0.5340 2.5244 0.1824 0.3002 1.0148 
INC 3       
BEE 1 1.2102 0.0275 2.3982 0.0794 0.2295 0.9311 
BEE 2 1.2097 0.0261 2.4055 0.06435 0.3168 0.8530 
BEE 3 1.0866 0.0086 4.2689 0.11495 0.0959 3.0779 
CCW 1 1.1854 0.0078 2.4796 0.3289 0.0008 1.2856 
CCW 2 0.5188 0.0145 2.4718 1.5194 0.0992 1.8393 
CCW 3       
KIN 1 0.2334 B.D.L. 0.8209 0.0189 0.3187 0.2688 
KIN 2 0.832 0.145 2.658 0.144 0.5061 1.1751 
KIN 3 0.918 0.007 2.000 0.150 0.1008 0.9747 
CAL 1 0.4258 0.1815 3.2039 0.2244 0.4459 2.1508 
CAL 2 0.7596 0.0240 7.9725 0.9396 3.8298 3.3592 
BRY 1       
BRY 2 0.2170 0.0258 7.4376 0.4057 3.8705 3.3243 
BRY 3       
MRK 1 0.4521 0.0869 7.6985 3.2433 1.0557 6.1038 
MRK 2 0.2320 B.D.L. 2.6389 0.0302 0.0427 2.3642 
MRK 3 0.8928 0.0379 7.7793 0.1222 4.0134 2.8352 
PSC1(1A 0.1771 B.D.L. 0.8494 0.0935 0.0162 0.6561 
PSC 2(1B)      
PSC 3 1.3270 0.0712 3.7608 0.145 0.1589 2.2037 
SHK 1 3.3764 0.0364 5.4467 0.0672 0.4626 1.5713 
SHK 2 0.0858 0.1733 3.1557 0.1279 1.4133 1.4833 
SHK 3       
JLP 1 0.9034 0.0035 18.7948 0.9219 13.0906 4.7973 
JLP 2 1.5558 0.0037 4.3501 0.74 0.1147 2.6759 
JLP 3 0.4520 1.1238 13.0007 0.7274 6.7333 4.6916 
HCP 1 0.3998 0.0088 3.8972 0.0655 1.4686 2.0200 
HCP 2 0.4125 0.0594 2.6293 0.0712 0.5963 1.5611 
HCP 3 4.0430 0.0755 5.0700 0.0894 0.1387 0.8128 
CBL 1       
CBL 2 0.7249 0.1662 4.6532 0.1377 2.2312 1.5309 
CBL 3 2.0149 0.0293 3.2970 0.0492 0.2544 0.9984 
AUD 1 0.3886 0.0656 3.4781 0.3338 1.0229 2.0010 
AUD 2 0.8058 0.0042 1.5356 0.5072 0.0893 0.6363 
AUD 3 1.4069 0.0107 4.3200 0.0906 0.2696 2.6328 
 
 149 
Appendix C 
 
Table 21: Complete list of plant species found at each wetland site 
 
Aud  
Acer rubrum 
 Agrostis sp. 
 Alisma subcordatum 
 Apocynum cannabinum 
 Aster sp. 
 Aster sp. #2 
 Baccharis halimifolia 
 Baptisia tinctoria 
 Betula nigra 
 Bidens connata 
 Bidens sp. 
 Carex lurida 
 Carex scoparia 
 Carex sp. 
 Cephalanthus occidentalis 
 Cyperus strigosus 
 Dichanthelium clandestinum 
 Diospyros virginiana 
 Eleocharis acicularis 
 Erigeron annuus 
 Eupatorium perfoliatum 
 Eupatorium sp. 
 Fraxinus pennsylvanica 
 Galium tinctorium 
 Hypericum mutilum 
 Impatiens capensis 
 Juncus canadensis 
 Juncus effusus 
 Juncus tenuis 
 Leersia oryzoides 
 Lespedeza sp. 
 Lespedeza sp. #2 
 Liquidambar styraciflua 
 Liriodendron tulipifera 
 Ludwigia palustris 
 Lysimachia sp. 
 Mikania scandens 
 Mikania scandens 
 Panicum sp. 
 Parthenocissus quinquefolia 
 Phragmites australis 
 Pinus sp. 
 Platanus occidentalis 
 Polygonum persicaria 
 Polygonum sagittatum 
 Quercus rubra 
 Rosa palustris 
 Sagittaria latifolia 
 Salix nigra 
 Schoenoplectus 
tabernaemontani 
 Toxicodendron radicans 
 Typha angustifolia 
 Typha latifolia 
 Vaccinium sp. 
 Viburnum recognitum 
 
 
Beehive  
Acer rubrum 
 Asclepias incarnata 
 Asclepias sp. 
 Asclepias sp. #2 
 Aster puniceus 
 Bidens sp. 
 Brasenia schreberi 
 Carex lurida 
 Carex scoparia 
 Carex sp. 
 Carex stipata 
 Dichanthelium clandestinum 
 Eleocharis quadrangulata 
 Eupatorium perfoliatum 
 Fraxinus sp. 
 Galium tinctorium 
 Impatiens capensis 
 Juncus canadensis 
 Juncus effusus 
 Leersia oryzoides 
 Ludwigia palustris 
 Lycopus americanus 
 
 150 
 Mikania scandens 
 Panicum sp. 
 Peltandra virginica 
 Pilea pumila 
 Platanus occidentalis 
 Polygonum sagittatum 
 Prunus sp. 
 Rubus sp. 
 Salix nigra 
 Schoenoplectus 
tabernaemontani 
 Sparganium erectum 
 Stachys sp. 
 Triadenum virginicum 
 Typha angustifolia 
 Typha latifolia 
 Typha X glauca 
 Ulmus sp. 
 Vicia cracca 
 Wisteria frutescens 
 
Bryantown  
Acer rubrum 
 Agrimonia parviflora 
 Agrostis gigantea 
 Asclepias incarnata 
 Asclepias sp. 
 Aster sp. 
 Betula nigra 
 Bidens frondosa 
 Boehmeria cylindrica 
 Carex lurida 
 Carex scoparia 
 Carex vulpinoidea 
 Cephalanthus occidentalis 
 Cuscuta gronovii 
 Cyperus sp. 
 Dichanthelium clandestinum 
 Diospyros virginiana 
 Eleocharis obtusa 
 Fraxinus pennsylvanica 
 Galium tinctorium 
 Hypericum mutilum 
 Impatiens capensis 
 Juncus canadensis 
 Juncus effusus 
 Juncus tenuis 
 Leersia oryzoides 
 Liquidambar styraciflua 
 Lycopus americanus 
 Mikania scandens 
 Nyssa sylvatica 
 Onoclea sensibilis 
 Oxalis sp. 
 Parthenocissus quinquefolia 
 Phalaris arundinacea 
 Poa sp. 
 Polygonum arifolium 
 Polygonum persicaria 
 Polygonum sagittatum 
 Rosa multiflora 
 Rosa palustris 
 Rubus allegheniensis 
 Rumex crispus 
 Salix nigra 
 Sambucus canadensis 
 Smilax rotundifolia 
 Solidago sp. 
 Spartina cynosuroides 
 Stachys sp. 
 Styrax grandifolius 
 Toxicodendron radicans 
 Verbena hastata 
 Verbena sp. 
 Wisteria frutescens 
 
Calvert  
Acer rubrum 
 Agrostis gigantea 
 Asclepias incarnata 
 Baccharis halimifolia 
 Betula nigra 
 Boehmeria cylindrica 
 Carex crinita 
 Carex lupulina 
 Carex lurida 
 Carex scoparia 
 Carex sp. 
 Cladium mariscoides 
 Dichanthelium clandestinum 
 Eleocharis obtusa 
 Eupatorium perfoliatum 
 
 151 
 Fraxinus pennsylvanica 
 Galium tinctorium 
 Juncus canadensis 
 Juncus effusus 
 Juncus tenuis 
 Leersia oryzoides 
 Liquidambar styraciflua 
 Lithospermum sp. 
 Lonicera japonica 
 Lycopus uniflorus 
 Mikania scandens 
 Oxalis stricta 
 Phalaris arundinacea 
 Polygonum arifolium 
 Polygonum aviculare 
 Polygonum sagittatum 
 Polygonum sp. 
 Populus heterophylla 
 Robinia pseudoacacia 
 Rosa palustris 
 Rubus allegheniensis 
 Salix nigra 
 Scirpus atrovirens 
 Scirpus cyperinus 
 Scirpus cyperinus 
 Solidago sp. 
 Toxicodendron radicans 
 Triadenum virginicum 
 Typha latifolia 
 
Cumberland  
 Acer rubrum 
 Agrimonia sp. 
 Agrostis gigantea 
 Agrostis sp. 
 Allium vineale 
 Apocynum cannabinum 
 Aster sp. 
 Aster sp. #2 
 Bidens frondosa 
 Bidens laevis 
 Calystegia sepium 
 Carex crinita 
 Carex lupulina 
 Carex lurida 
 Carex scoparia 
 Carex squarrosa 
 Carex stipata 
 Carex vulpinoidea 
 Centaurea biebersteinii 
 Cephalanthus occidentalis 
 Coronilla varia 
 Dichanthelium clandestinum 
 Eleocharis acicularis 
 Equisetum arvense 
 Erigeron annuus 
 Eupatorium perfoliatum 
 Eupatorium sp. 
 Fraxinus pennsylvanica 
 Galium tinctorium 
 Impatiens capensis 
 Iris sp. 
 Juncus acuminatus 
 Juncus canadensis 
 Juncus effusus 
 Juncus sp. 
 Juncus tenuis 
 Ludwigia alternifolia 
 Lycopus americanus 
 Mimulus ringens 
 Osmunda cinnamomea 
 Panicum sp. 
 Plantago major 
 Poaceae --hairy ligule 
 Polygonum persicaria 
 Polygonum sagittatum 
 Polygonum sp. 
 Rosa palustris 
 Rudbeckia hirta var. 
pulcherrima 
 Rudbeckia sp. 
 Rumex crispus 
 Schoenoplectus 
tabernaemontani 
 Sisyrinchium angustifolium 
 Solanum carolinense 
 Solidago sp. 
 Solidago sp. 
 Solidago sp. #2 
 Tilia americana 
 Toxicodendron radicans 
 Trifolium repens 
 
 152 
 Typha angustifolia 
 Typha latifolia 
 Ulmus rubra 
 Verbena hastate 
 
Waldorf  
Acer rubrum 
 Achillea millefolium 
 Agrostis sp. 
 Allium vineale 
 Asclepias verticillata 
 Aster sp. 
 Betula nigra 
 Bidens laevis 
 Carex frankii 
 Carex lupulina 
 Carex lurida 
 Carex scoparia 
 Carex sp. #2 
 Carex vulpinoidea 
 Cephalanthus occidentalis 
 Cephalanthus occidentalis 
 Dichanthelium clandestinum 
 Eleocharis obtusa 
 Eleocharis sp. 
 Eupatorium sp. 
 Fraxinus pennsylvanica 
 Galium tinctorium 
 Iris sp. 
 Juncus canadensis 
 Juncus effusus 
 Juncus tenuis 
 Leersia oryzoides 
 Lespedeza virginica 
 Liquidambar styraciflua 
 Lycopus americanus 
 Lysimachia sp. 
 Mentha arvensis 
 Mikania scandens 
 Oligoneuron album 
 Oxalis sp. 
 Panicum virgatum 
 Panicum virgatum 
 Parthenocissus quinquefolia 
 Phalaris arundinacea 
 Poa sp. 
 Polygonum hydropiper 
 Polygonum persicaria 
 Polygonum sagittatum 
 Potentilla recta 
 Quercus bicolor 
 Rorippa nasturtium-aquaticum 
 Rosa multiflora 
 Rumex obtusifolius 
 Rumex sp. 
 Salix nigra 
 Scirpus sp. 
 Setaria viridis 
 Solidago sp. 
 Spartina cynosuroides 
 Specularia perfoliata 
 Toxicodendron radicans 
 Typha angustifolia 
 Typha latifolia 
 Veronica sp. 
 
Herring Creek Nature Park 
 Acer rubrum 
 Agrostis gigantea 
 Apocynum cannabinum 
 Asclepias incarnata 
 Aster sp. 
 Baccharis halimifolia 
 Campsis radicans 
 Carex sp. 
 Cephalanthus occidentalis 
 Dichanthelium clandestinum 
 Festuca sp. 
 Galactia regularis 
 Ilex opaca 
 Juncus canadensis 
 Juncus effusus 
 Juncus longii 
 Juncus sp. 
 Juncus tenuis 
 Juniperus communis 
 Liquidambar styraciflua 
 Mikania scandens 
 Myrica cerifera 
 Phragmites australis 
 Pinus pungens 
 Pinus taeda 
 
 153 
 Poaceae --hairy ligule 
 Rosa palustris 
 Rumex crispus 
 Scirpus americanus 
 Taxodium distichum 
 
Irvine Nature Center  
Acer rubrum 
 Agrimonia parviflora 
 Agrostis gigantea 
 Agrostis sp. 
 Allium vineale 
 Asclepias incarnata 
 Betula sp. 
 Bidens coronata 
 Blephilia hirsuta 
 Boehmeria cylindrica 
 Brassicaceae sp. 
 Carex crinita 
 Carex laxiflora 
 Carex lurida 
 Carex scoparia 
 Carex stipata 
 Carex vulpinoidea 
 Celastrus scandens 
 Chenopodium album 
 Crataegus phaenopyrum 
 Eleocharis obtusa 
 Epigaea repens 
 Fragaria virginiana 
 Fraxinus pennsylvanica 
 Galium aparine 
 Galium tinctorium 
 Glyceria striata 
 Impatiens capensis 
 Juncus effusus 
 Juncus sp. 
 Juncus tenuis 
 Leersia oryzoides 
 Lindera benzoin 
 Lycopus uniflorus 
 Mentha spicata 
 onethera sp. 
 Oxalis sp. 
 Parthenocissus quinquefolia 
 Phalaris arundinacea 
 Polygonum arifolium 
 Polygonum pensylvanicum 
 Polygonum persicaria 
 Polygonum sagittatum 
 Polystichum acrostichoides 
 Quercus rubra 
 Ranunculus sp. 
 Rosa multiflora 
 Rubus idaeus 
 Rubus sp. 
 Rumex sp. 
 Scirpus atrovirens 
 Sisyrinchium montanum 
 Solidago sp. 
 Stachys sp. #1 (recognized) 
 Toxicodendron radicans 
 Trifolium hybridum 
 Trillium sp. 
 Typha latifolia 
 Ulmus americana 
 Viburnum recognitum 
 Viburnum sp. 
 
Jackson Lane Preserve  
Acer rubrum 
 Allium vineale 
 Aster sp. 
 Betula nigra 
 Bidens aristosa 
 Bidens connata 
 Bidens frondosa 
 Carex comosa 
 Carex scoparia 
 Carex sp. 
 Clethra alnifolia 
 Cornus amomum 
 Echinochloa crus-galli 
 Eleocharis acicularis 
 Eupatorium sp. 
 Fraxinus pennsylvanica 
 Hieracium sp. 
 Hypericum mutilum 
 Iris sp. 
 Juncus canadensis 
 Juncus effusus 
 Juncus marginatus 
 
 154 
 Juncus sp. 
 Juncus tenuis 
 Liquidambar styraciflua 
 Lonicera japonica 
 Mikania scandens 
 Oxalis stricta 
 Polygonum persicaria 
 Quercus alba 
 Quercus bicolor 
 Quercus lyrata 
 Quercus rubra 
 Rhexia mariana 
 Rhexia mariana 
 Rorippa islandica 
 Rubus allegheniensis 
 Rumex crispus 
 Salix nigra 
 Solidago sp. 
 Xanthium sp. 
 
Kinder 
 Agrostis gigantea 
 Agrostis sp. 
 Allium vineale 
 Amaranthus L. sp. 
 Andropogon gerardii 
 Asclepias incarnata 
 Aster sp. 
 Betula nigra 
 Carex lupulina 
 Carex lurida 
 Carex scoparia 
 Carex squarrosa 
 Carex stipata 
 Carex vulpinoidea 
 Cephalanthus occidentalis 
 Cornus amomum 
 Dactylis glomerata 
 Dichanthelium acuminatum 
 Dulichium arundinaceum 
 Festuca sp. 
 Fraxinus pennsylvanica 
 Galium tinctorium 
 Juncus effusus 
 Juncus tenuis 
 Leersia oryzoides 
 Limnobium spongia 
 Liquidambar styraciflua 
 Lolium pratense 
 Oxalis sp. 
 Phalaris arundinacea 
 Phleum pratense 
 Platanus occidentalis 
 Poa sp. 
 Polygonum arifolium 
 Polygonum hydropiper 
 Polygonum persicaria 
 Quercus bicolor 
 Quercus palustris 
 Quercus rubra 
 Rubus allegheniensis 
 Rumex crispus 
 Salix nigra 
 Scirpus atrovirens 
 Solanum carolinense 
 Viola sp. 
 
Merkle Wildlife Refuge 
 Acer rubrum 
 Agrostis gigantea 
 Asclepias incarnata 
 Asclepias sp. 
 Aster puniceus 
 Aster sp. 
 Betula nigra 
 Boehmeria cylindrica 
 Carex lurida 
 Carex vulpinoidea 
 Chamaecyparis thyoides 
 Crataegus iracunda 
 Cuscuta gronovii 
 Cyperus esculentus 
 Dichanthelium clandestinum 
 Distichlis spicata 
 Eleocharis obtusa 
 Erigeron annuus 
 Eupatorium perfoliatum 
 Fern sp. 
 Fraxinus pennsylvanica 
 Fraxinus sp. 
 Galium sp. 
 Galium tinctorium 
 
 155 
 Hydrangea arborescens 
 Hypericum mutilum 
 Impatiens capensis 
 Iris sp. 
 Juncus canadensis 
 Juncus effusus 
 Juncus marginatus 
 Juncus sp. 
 Juncus tenuis 
 Juniperus communis 
 Leersia oryzoides 
 Lespedeza violacea 
 Liquidambar styraciflua 
 Liriodendron tulipifera 
 Lycopus uniflorus 
 Mentha arvensis 
 Mentha spicata 
 Microstegium vimineum 
 Mikania scandens 
 Nyssa sylvatica 
 Onoclea sensibilis 
 Parthenocissus quinquefolia 
 Phalaris arundinacea 
 Pinus taeda 
 Poa sp. 
 Poaceae --hairy ligule 
 Polygonum persicaria 
 Polygonum sagittatum 
 Quercus rubra 
 Rhexia mariana 
 Rosa palustris 
 Rubus allegheniensis 
 Rumex crispus 
 Salix nigra 
 Sambucus canadensis 
 Scirpus cyperinus 
 Sisyrinchium angustifolium 
 Smilax rotundifolia 
 Solanum carolinense 
 Solidago sp. 
 Spartina cynosuroides 
 Toxicodendron radicans 
 Typha angustifolia 
 Typha latifolia 
 Vaccinium corymbosum 
 Verbena sp. 
 Veronica serpyllifolia 
 Veronica sp. 
 Viburnum recognitum 
 Wisteria frutescens 
 
Piscataway Stream Valley Park  
Acer rubrum 
 Achillea millefolium 
 Agrostis gigantea 
 Asclepias incarnata 
 Aster sp. 
 Betula nigra 
 Bidens connata 
 Carex lupulina 
 Carex lurida 
 Carex scoparia 
 Carex sp. 
 Carex vulpinoidea 
 Cephalanthus occidentalis 
 Cornus amomum 
 Cyperus strigosus 
 Dichanthelium clandestinum 
 Echinochloa walteri 
 Erigeron annuus 
 Eupatorium perfoliatum 
 Fraxinus pennsylvanica 
 Galium tinctorium 
 Impatiens capensis 
 Juncus canadensis 
 Juncus effusus 
 Juncus sp. 
 Juncus tenuis 
 Leersia oryzoides 
 Liquidambar styraciflua 
 Liriodendron tulipifera 
 Lycopus uniflorus 
 Mikania scandens 
 Osmunda cinnamomea 
 Parthenocissus quinquefolia 
 Phragmites australis 
 Platanus occidentalis 
 Polygonum hydropiperoides 
 Polygonum persicaria 
 Polygonum punctatum 
 Quercus alba 
 Quercus phellos 
 
 156 
 Rosa palustris 
 Rubus allegheniensis 
 Rumex crispus 
 Salix nigra 
 Scirpus cyperinus 
 Solidago sp. 
 Stachys sp. 
 Tilia americana 
 Toxicodendron radicans 
 Triadenum virginicum 
 Typha latifolia 
 
Shockley  
Acer rubrum 
 Aster sp. 
 Aster sp. 
 Cladium mariscoides 
 Clethra alnifolia 
 Cyperus sp. 
 Cytisus scoparius 
 Dichanthelium clandestinum 
 Dichanthelium sp. 
 Echinochloa muricata 
 Eleocharis acicularis 
 Eleocharis obtusa 
 Erigeron annuus 
 Ilex opaca 
 Iris sp. 
 Juncus canadensis 
 Juncus effusus 
 Juncus marginatus 
 Juncus sp. 
 Juncus sp. 
 Juncus tenuis 
 Liquidambar styraciflua 
 Lonicera sempervirens 
 Ludwigia palustris 
 Magnolia virginiana 
 Microstegium vimineum 
 Osmunda cinnamomea 
 Pinus serotina 
 Polygonum persicaria 
 Quercus alba 
 Rhexia mariana 
 Rhynchospora alba 
 Rosa multiflora 
 Rubus allegheniensis 
 Salix nigra 
 Sisyrinchium sp. 
 Smilax laurifolia 
 Smilax rotundifolia 
 Solidago sp. 
 Stachys sp. 
 Toxicodendron radicans 
 Trifolium sp. 
 Vaccinium sp. 
 Viburnum s
 
 157 
 
Appendix D 
 
Table 22: Wetland Site Information 
 
The following are site descriptions provided by MDE or MD SHA for the wetland 
assessment project. 
 
Irvine Nature Center (INC) is located on the former site of a farm in northwest 
Baltimore County.  Large, clear fields still occupy a major part of the property, while 
deciduous forests surround them.  Plot 1 lies in an open field with several species of 
wetland vegetation, and plots two and three are in forests on opposite corners of the 
field containing plot 1. 
 
Beehive (BEE) is in a fairly developed residential and industrial area in Elkridge.  A 
railroad runs on the south side of the site.  The site lies in the center of a 
neighborhood, with a road on one side and houses on another.  The site lies 
approximately 10 feet lower than the surrounding area.  The majority of the site holds 
standing water, with parts slightly above water level and others up to a foot below 
water.  Vegetation seldom grows above 10 feet in the site. 
 
SHA Description: Beehive Site - is a 2.4 acre site created to partially mitigate for 
non-tidal wetland impacts associated with the MD 100 project.  The site is located on 
the east side of a tributary to Shallow Run, north of Loudon Avenue and east of Smith 
Avenue in Howard County.  Shallow Run is a tributary to Deep Run.  The site 
consists of 2.4 acres PEM creation.  Construction of the site was completed in Spring 
1995.  The site is accessible directly from Loudon Avenue.  Based on SHA-GIS data, 
land use within the vicinity of the mitigation site is a mix of institutional, industrial, 
low and medium density residential, and forested. 
 
Waldorf (CCW) is an approximately 10 acre site with its southern border on MD 
228.  Most of the site is dry land.  In the two southern corners, there are small lakes.  
Plot 1 is located just outside one of these lakes.  In the middle of the site, there is a 
patch of forest; other than this, trees in the site grow no more than 10 feet tall.  Much 
of the southwest corner of the site, where plot 2 is located, is covered by plants 
growing less than 1 inch tall.  Plot 3, in the north of the site, is mostly covered with 
short trees, growing in dry ground. 
 
SHA Description: MD 228 Site - is a 12.4 acre site created to partially mitigate for 
non-tidal wetland impacts associated with the MD 228 project.  The site is located on 
the on the north side of MD 228, approximately one mile east of Bealle Hill Road in 
Charles County.  The site consists of 12.4 acres of PFO/PEM creation.  Construction 
of the site was completed in Fall 1995.  The site is accessible directly from MD 228 
Westbound.  Based on SHA-GIS data, land use within the vicinity of the mitigation 
site is a mix of forested and low-density residential. 
 
 
 158 
 
Calvert (CAL) is a small site located at the side of MD 4.  Most of the site is 
underwater.  Tree trunks were laid horizontally across the shallower of the two lakes 
on site.  The shallower lake hosts a variety of wetland vegetation as well as our only 
plot on the site. 
 
SHA Description: MD 4/MD 260 Site - is a 1.10-acre site created to mitigate for 
forested wetland impacts associated with the construction of the MD 4/260 
interchange. This site is located within the floodplain of Lyons Creek south of the 
newly constructed MD 4/260 interchange and west of MD 4 in Calvert County, 
Maryland.  Construction of the site was completed in Spring 2003.  The site consists 
of 0.82 acre PFO creation, 0.38 acre PEM creation, and 0.16 acre bare ground.  The 
bare ground area consists of highly acidic soils (acid sulfate soils?).  SHA is currently 
investigating various acid-tolerant vegetation for future planting.  The site is 
accessible directly from MD 4 Southbound.  Based on SHA-GIS data, land use within 
the vicinity of the mitigation site is primarily forested.  
 
Kinder (KIN) is a dry site located in a fairly rural area in Anne Arundel County.  The 
site is located between two farms.  The majority of the vegetation is low growing 
sedges and rushes.  The surrounding area is forested on one side and farmland on the 
other three. 
 
SHA Description: Kinder Site - is a 10.89-acre site created to mitigate for forested 
wetland impacts associated with improvements to MD 468.  This site is located 
within a former pasture west of MD 468 and south of Sudley Road in Anne Arundel 
County, Maryland.  Construction of the site was completed in Spring 2003.  The site 
consists of 9.0 acres of PFO restoration/enhancement and 1.89 acres meadow/dry 
forest creation.  This area is fenced/gated.  However, the chain connecting the gate is 
slack enough to allow the students to "shimmy" though.  Outside of the fenced area is 
a 18.81 acre PFO preservation area.  Based on SHA-GIS data, land use within the 
vicinity of the mitigation site is primarily agricultural and low density residential.  
 
Bryantown (BRY) is a densely vegetated site in Charles County.  Nearly 100% of 
the site is covered by some sort of plant life.  The site is located in a very agricultural 
area, with a farm on its west side.  The northern part of the site is heavily forested, 
with tree trunks growing within a foot of each other.  Standing water exists on most 
of the area, and a creek runs through the site. 
 
Merkle Wildlife Management Area (MRK) hosts a 9 acre site on the banks of the 
Patuxent River.  A majority of the site is covered with vegetation and standing water.  
The surrounding area is relatively undeveloped compared to other sites.  Trees 
approximately 15 feet tall dominate each of our 3 plots. 
 
Piscataway Stream Valley Park (PSC) is located near MD 301, a recycling center, 
and a police firing range.  Despite all of this, the area is still fairly undeveloped.  Most 
of the plot areas we selected were submerged.  Willow is very common in sites 1 and 
2, while trees are rare in plot 3. 
 
 159 
 
 
Shockley (SHK) is a diversely vegetated site near Snow Hill on the eastern shore.  
Much of the area is densely forested.  We chose plots in slightly depressed, 
submerged areas with no overhanging trees.  The site overall is usually left alone to 
let nature run its course; some years ago, a fire burned down a significant part of the 
forests, but the areas are left to grow back in naturally. 
 
Herring Creek Park (HCP) is located just a few miles from downtown Ocean City.  
It is surrounded by neighborhoods and is less than a mile from busy Route 50.  Our 
plots are located off the sidewalk of the park in two submerged areas dominated by 
phragmites australis. 
 
Jackson Lane Preserve (JLP) is a TNC wetland located on former farmland less 
than a mile from the Delaware line in Caroline County.  The entire surrounding area 
is farmland.  The area is mostly full of low growing vegetation with forest lining the 
border of the property.  Several ponds are located throughout the wetland; we chose 
our three plots in these ponds. 
 
Cumberland (CBL) is located in the foothills of the Appalachian Mountains near 
Cumberland, MD.  The area is not densely populated at all.  Some but very few 
people live within sight of the wetland, which is just down a hill to the side of US 
220.  Half of the site closer to the highway is covered in water; the other half is dry 
ground.  Trees are very rare in the site.  The site is also very close to I-70. 
  
Aud (AUD) is located on Flat Iron Road, very close to St. Mary?s River in St. Mary?s 
County.  Most notably, the wetland is downhill from a horse farm.  The site consists 
of two slightly depressed wetlands containing usual wetland vegetation (typha 
latifolia, mikania scandens, etc.).  Standing water is only prevalent in one of the three 
plots.