ABSTRACT 
 
 
 
 
Title of Dissertation: TIME IS ENERGY: DRIVERS OF MIGRATORY 
AND NON-MIGRATORY MOVEMENTS BY 
SOARING BIRDS   
  
 
 Julie Marie Mallon, Doctor of Philosophy, 2020 
  
 
Dissertation directed by: Professor William F. Fagan 
Department of Biology 
 
 
 
Vultures are obligate scavengers and search over large areas for carrion, which 
is ephemeral. To profit from carrion, they are also obligate soarers that rely on 
the availability of environmental updrafts to subsidize flight. This restricts their 
flight spatially and temporally to where and when strong updrafts are available. 
In this dissertation, I investigate how Turkey Vultures (Cathartes aura) use 
stopovers to both avoid inclement weather and to replenish spent fuel reserves, 
as well as how the timing of movement activity differs according to flight mode. 
Using GPS-tracking data from four migratory Turkey Vulture populations, I 
evaluated how Turkey Vultures respond to changing weather conditions. 
During stopovers, movement activity was positively correlated with conditions 
that promote thermal development, suggesting not all stopovers are used for 
weather-avoidance. Turkey Vultures began stopovers immediately in response 
to deteriorating weather conditions but their departure from stopovers was 
delayed relative to improvements in weather, behavior that is consistent with 
an energy-minimization strategy.  
I estimated total energetic costs for each migration and identify probable 
refueling stopovers. Only long-distance Turkey Vulture migrants regularly stop 
to feed. Overall migratory costs are driven by migration duration, and therefore 
are lower in the spring when vultures migrate faster, which may contribute to 
seasonal differences in flight behavior. 
Last, I compared the non-migratory movements of 49 avian species to test for 
the influence of flight mode on the timing of movement activity. Terrestrial 
soaring birds began activity later and stopped activity earlier than other birds. 
This study demonstrates that flight mode influences temporal patterns of daily 
movement activity of birds. 
This dissertation enhances our understanding of how soaring birds cope with 
the temporal and spatial restrictions on their movements. During non-migration, 
soaring birds are active for longer proportions of the day to search for food over 
greater spatial areas than flapping birds. During migration, total energetic costs 
are driven more by migration duration than by behavior. Therefore, an energy 
minimization strategy for Turkey Vultures is also a time minimization strategy 
and vultures are expected to optimize their migration so that they spend the 
fewest days migrating as possible. 
 
 
 
 
 
 
TIME IS ENERGY: DRIVERS OF MIGRATORY AND NON-MIGRATORY 
MOVEMENTS BY SOARING BIRDS 
 
 
 
by 
 
Julie Marie Mallon 
 
 
 
 
Dissertation submitted to the Faculty of the Graduate School of the 
University of Maryland, College Park, in partial fulfillment 
of the requirements for the degree of 
Doctor of Philosophy 
2020 
 
 
 
 
 
 
 
Advisory Committee: 
 
Professor William F. Fagan, Chair 
Professor Derek A. Paley 
Professor Gerald S. Wilkinson 
Associate Professor Christopher Rowe 
Dr. Keith L. Bildstein 
 
 
 
 
 
 
 
? Copyright by 
Julie Marie Mallon 
2020 
 
 
 
Preface 
This dissertation contains three chapters, written in manuscript form, and 
appendices to chapters 1 and 3. This dissertation includes a previously 
published work (chapter 3).  
The reference for the published chapter: 
 
Chapter 3: Diurnal timing of nonmigratory movement by birds: the 
importance of foraging spatial scales. 
Previously published under: Julie M. Mallon, et al. Diurnal timing of 
nonmigratory movement by birds: the importance of foraging spatial scales. 
Journal of Avian Biology. doi: 10.1111/jav.02612. 
. 
 
 
 
` 
ii 
 
Dedication 
 
To Eric, Graham, John, Michael, Terry and Terry: for their thoughtful discourse 
on the physical limitations on avian flight. 
iii 
 
Acknowledgements 
 
I am grateful for all the people I have had the pleasure to know and work with 
over the last several years. They have all contributed to this dissertation in 
some way. In particular, I would like to acknowledge: 
My advisor, Bill Fagan, who provided me the time to nurture my exploratory 
analyses into fully-fledged and impactful research that contribute importantly to 
our understanding of avian movement ecology.  
My committee members past and present, Gerald Wilkinson, Derek Paley, 
Chris Rowe, Helen Bailey, and Keith Bildstein, for their guidance. My past 
mentors - Todd Katzner, Keith Bildstein, and Kelly Swing  -  for helping me find 
my study system and my role as a researcher. A special thanks to Bill, Eric 
Haag & the BISI office, and my committee for supporting my last-minute 
decision to defend.  
Past and present Fagan Lab members, especially Nicki Barbour, Anshuman 
Swain, Jake Weissman, Elie Gurarie, Luisa Diele-Viegas, Nina Attias, and 
Silvia Alvarez. 
iv 
 
Courtney Zambory, Anna Davis, Simone de Roode, Dani Escobar, Kevin 
Bennett, Danielle Adams, Edward Hurme, and fellow BEES students for their 
friendship and support.  
Marlee Tucker, who expertly guided me through our large, collaborative project. 
Marlee and Thomas Mueller for arranging for me to visit the Senckenberg 
Institute. David Kirk and my many collaborators for sharing their data and 
expertise. Laurie Goodrich, JF Therrien, and the rest of Hawk Mountain 
Sanctuary staff for arranging for me to stay at the residence while I wrote my 
dissertation.  
Melissa and Nicki for caring for my dog during my various travels. Thea, for 
3,500 walks and counting. My family for their encouragement and embracing 
my love of birds. 
v 
 
Table of Contents 
 
Preface ............................................................................................................ ii 
Dedication ...................................................................................................... iii 
Acknowledgements ........................................................................................ iv 
List of Tables .................................................................................................. ix 
List of Figures .................................................................................................. x 
List of Abbreviations ...................................................................................... xii 
Chapter 1 : Inclement weather prevents migratory progress for obligate 
soaring migrants .............................................................................................. 1 
1.1 Abstract .................................................................................................. 1 
1.2 Introduction ............................................................................................ 2 
1.3 Methods ................................................................................................. 6 
1.3.1 Study Species ................................................................................. 6 
1.3.2 Weather Predictors .......................................................................... 8 
1.3.3 Stopover Classification .................................................................... 9 
1.3.4 Analysis ......................................................................................... 10 
1.4 Results ................................................................................................. 12 
1.4.1 Description of Stopover Use .......................................................... 12 
1.4.2 Onset of Stopovers ........................................................................ 14 
1.4.3 Departure from Stopovers ............................................................. 16 
vi 
 
1.4.4 Movement Behavior and Weather ................................................. 20 
1.5 Discussion ............................................................................................ 20 
1.5.1 Weather Avoidant Stopovers ......................................................... 21 
1.5.1 Movement Behavior during Stopovers ........................................... 22 
1.5.1 Short Stopovers ............................................................................. 24 
1.6 Conclusion ........................................................................................... 25 
Chapter 2 : Time is energy: optimal migration of an obligate soaring migrant
 ...................................................................................................................... 26 
2.1 Abstract ................................................................................................ 26 
2.2 Introduction .......................................................................................... 27 
2.3 Methods ............................................................................................... 31 
2.3.1 Study Species ............................................................................... 31 
2.3.2 The physiological model ................................................................ 32 
2.3.3 Quantifying feeding stops .............................................................. 35 
2.4 Results ................................................................................................. 39 
2.4.1 Total migration costs ..................................................................... 39 
2.4.2 Migrations requiring feeding stops ................................................. 39 
2.4.3 Quantifying feeding stops .............................................................. 41 
2.4.4 Optimal temporal and spatial patterns of feeding .......................... 43 
2.5 Discussion ............................................................................................ 45 
2.6 Conclusion ........................................................................................... 50 
Chapter 3 : Diurnal timing of nonmigratory movement by birds: the importance 
of foraging spatial scales ............................................................................... 51 
vii 
 
3.1 Abstract ................................................................................................ 51 
3.2 Introduction .......................................................................................... 52 
3.3 Methods ............................................................................................... 55 
3.3.1 Data ............................................................................................... 55 
3.3.2 Movement Characteristics ............................................................. 57 
3.3.3 Morphological and Ecological Characteristics ............................... 60 
3.3.4 Analysis ......................................................................................... 62 
3.4 Results ................................................................................................. 64 
3.5 Discussion ............................................................................................ 69 
3.6 Acknowledgements .............................................................................. 73 
Appendix A: Supplementary Material Chapter 1 ............................................ 75 
Appendix B: Supplementary Material Chapter 3 ............................................ 80 
Bibliography ................................................................................................... 81 
 
 
 
viii 
 
List of Tables 
 
Table 1.1. Definitions and sources of weather variables used.. ...................... 8 
Table 2.1. Descriptions of the breeding grounds, body mass, and migration 
distance by population. .................................................................................. 31 
Table 2.2. Model estimates of the number of fall and spring migration that 
require feeding stops, by population.. ............................................................ 41 
Table 3.1. Temporal variables and their definitions. ...................................... 58 
Table 3.2. Ecological variables used in analysis.. ......................................... 60 
 
ix 
 
List of Figures 
 
Figure 1.1. Map of daily migration locations for 33 individuals across four 
populations and three races. ........................................................................... 7 
Figure 1.2. Mean distance between stopovers, per population.. ................... 12 
Figure 1.3. Examples of four stopovers, ranked from most tortuous (A) to most 
sedentary (D).. ............................................................................................... 14 
Figure 1.4. Average weather conditions relative to the start of identified 
stopovers (red line) for n = 539 stopovers ..................................................... 15 
Figure 1.5. Average weather conditions relative to the end of identified 
stopovers (red line) for n = 539 stopovers ..................................................... 17 
Figure 1.6. Average values for several weather variables as a function of mean 
proportion of activity during stopovers (a measure of tortuousness) ............. 18 
Figure 1.7. Average values for several weather variables by mean first 
passage time during stopovers with a radius of 250 m (a measure of 
sedentariness). .............................................................................................. 19 
Figure 2.1. Map of Turkey Vulture fall migration through Central America. 
Migrants from the central USA converge and follow the Gulf coast of Mexico to 
the Isthmus of Tehuantepec, where they converge with migrants from the 
western USA. ................................................................................................. 38 
Figure 2.2. Modeled results for percent body mass consumed during migration 
while fasting, with separate models for each population. ............................... 40 
Figure 2.3 The difference in days between successive feeding stops. ......... 42 
Figure 2.4. Distribution of probable feeding stopovers (red) and active 
migration points (grey) for western and central C. aura meridionalis populations
 ...................................................................................................................... 44 
x 
 
Figure 2.5. The effect of the timing of feeding on post-migratory body mass (g)
 ...................................................................................................................... 45 
Figure 3.1. The NMDS ordination indicates inter-specific similarities within 
temporal activity patterns, among all 49 species. .......................................... 65 
Figure 3.2. Dot plots of flapping, terrestrial (obligate and facultative), and 
pelagic soaring birds by (a) start of activity relative to sunrise and (b) end of 
activity relative to sunset, with units in hours ................................................. 66 
Figure 3.3. Foraging habitats by (a) active time, (b) mean speed, and (c) 
distance ......................................................................................................... 68 
  
xi 
 
List of Abbreviations 
 
GPS, global positioning system 
BMR, basal metabolic rate 
FPT, first passage time 
MR, metabolic rate (specific to behavior; kJ/h) 
??, Days fasted  
??, energy savings of 4% per day  
??????, metabolic expenditure for activity (kJ/day)  
??, daily body mass (g) 
??, total duration of behavior (h/day) 
??????, Daily energy expenditure (kJ/day)  
??, energetic equivalent of the catabolism of fat (39.7 kJ/g) 
????, daily body mass change (g)  
??????????????, starting body mass (g) 
xii 
 
Chapter 1: Inclement weather prevents migratory 
progress for obligate soaring migrants 
 
1.1 Abstract 
Migrating birds experience weather conditions that change with time, which 
affect their decision to stop or resume migration. Soaring migrants are 
especially sensitive to changing weather conditions because they rely on the 
availability of environmental updrafts to subsidize flight. We used satellite 
tracking data from four migratory Turkey Vulture (Cathartes aura) populations, 
paired with local weather data to determine how frequently the vultures use 
stopovers and to evaluate how they use stopovers to avoid inclement weather. 
We analyzed 174 migrations of 33 individuals from 2006 - 2019 and identified 
590 stopovers. Stopover duration ranged from 2 h to more than 11 days, with 
51% of stopovers lasting < 24 h. Stopover use increased with migration 
distance and was more frequent during fall than spring migration. Turkey 
Vultures began stopovers immediately in response to deteriorating weather 
conditions but their departure from stopovers was delayed relative to 
improvements in weather, behavior that is consistent with an energy-
minimization migration strategy. During stopovers, the proportion of activity 
was positively correlated with conditions that promote thermal development, 
1 
 
suggesting not all stopovers are used for weather-avoidance. As revealed by 
our results, birds are capable of responding rapidly to their environment; 
therefore, for studies interested in external drivers of stopover use it is essential 
that stopovers be identified at fine temporal scales. 
 
1.2 Introduction 
As birds migrate, they pass through a variety of habitats and experience 
variable environmental conditions. Locally, migrating birds experience weather 
conditions that change with time, which are known to influence their flight 
behavior, especially the likelihood of stopping and resuming active migration. 
Passerines avoid departing stopovers during precipitation (Smith and 
McWilliams 2014) and are more likely to resume active migration during low 
wind speeds and decreasing surface pressure (Matthews and Rodewald 2010). 
Migratory raptors also avoid migrating under poor weather conditions, where 
they are rarely observed migrating during cold fronts but are observed in peak 
numbers immediately following a passing cold front (Allen et al. 1996, Inzunza 
et al. 2005). Similarly, the likelihood of soaring birds stopping over increases 
on days with cloud cover and rain (Goodrich 2010), and when thermal updraft 
strength is weak (sensu Dodge et al. 2014). 
2 
 
Soaring birds, like condors (Poessel et al. 2018), are especially sensitive to 
changing weather conditions because they rely on the availability of 
environmental updrafts to subsidize flight. Although terrestrial birds can soar 
using orographic (Bohrer et al. 2012) or turbulent updrafts (Mallon et al. 2015) 
to subsidize flight, thermals (Dodge et al. 2014) are the most important type of 
updraft for most soaring migrants because thermals are widely distributed 
across the landscape (Pennycuick 1975), allow birds to reach altitudes 
necessary for fast cross-country soaring, and allow for straight, efficient flight 
paths (Duerr et al. 2012). Thermals (i.e., vortices of ascending hot air 
surrounded by descending cooler air) (Pennycuick 1975, Kerlinger 1989) are 
generated by differential heating of the earth?s surface. Thermals are, therefore, 
an uncertain resource that only form under appropriate weather conditions 
(Kerlinger 1989). Poor weather conditions slow or prevent the development of 
updrafts, thereby forcing birds to switch to energetically expensive flapping 
flight (Spaar and Bruderer 1997, Ferland-Raymond et al. 2005) or grounding 
them. 
Despite recognizing the relationship between migratory flight behavior and 
changing weather conditions, stopover use by large, soaring birds has been 
understudied. This is in part due to challenges around defining what constitutes 
a stopover - rule selection varies depending on the species ecology and desired 
behavior. Some of the definitions of stopovers include: ? 150 km/day (Kochert 
3 
 
et al. 2011), < 100 km/day (Hake et al. 2001, Crawford and Long 2017), < 50 
km/day of directed flight (Sergio et al. 2017), ? 25 km/day (Vansteelant et al. 
2015), >24 h in an area <30 km in diameter (Shiu et al. 2006), and spending 
more than 24 h in an area (Monti et al. 2018). These definitions pose two 
problems. First, all of these definitions consider the durations of stopovers to 
be ? 24 h, due in part to the assumption that stopovers are primarily used to 
refuel. If we define stopovers based on the movement behavior of individuals, 
however, stopovers occur whenever individuals do not engage in fast, directed 
flight but are otherwise expected to do so. Using this movement-based 
definition of stopovers, the duration of stopovers may be much shorter than 24 
h as local weather conditions can change rapidly and may only prevent birds 
from flying for short periods of time (i.e., hours). Second, identification at the 
daily scale may underestimate the duration of stopovers that are > 24 h. For 
example, birds may actively migrate for several hours before stopping over and 
exceed the distance threshold. The following day would be identified as the 
start of the stopover, several hours after the bird ceased directed flight. 
Alternatively, a bird that is refueling may travel 10?s of kilometers in a day 
searching for food and would exceed these distance thresholds although they 
are not engaged in directed flight. Clearly, standardized measures of 
movements that constitute what are and are not stopovers are warranted. 
4 
 
Stopover behavior is ideal to study using Turkey Vultures (Cathartes aura) 
because vultures are obligate soaring birds (Ruxton and Houston 2004) and 
cannot sustain themselves aloft using flapping flight (Rosser and George 
1986). Hence, Turkey Vultures should be more sensitive to changes in weather 
than other soaring migrants and stopover more frequently than other migratory 
soaring species. We anticipated some stopovers may be quite short if they 
were responding to passing weather fronts in the late afternoon or early 
morning that overlapped with their daily roosting patterns. Therefore, we chose 
to be highly conservative with our selection of stopovers and identified the start 
and end of stopovers at the hourly scale. We sought to determine how 
extensively Turkey Vultures depend on stopovers, and how the birds use 
stopovers. To this end, we used satellite tracking data from four migratory 
Turkey Vulture populations, paired with local weather data. As a first step, we 
identified stopovers using first passage time (FPT). We automated 
identification of stopovers by selecting the radius and FPT threshold based on 
the structure of the data, to avoid under- and over-selection of stopovers, and 
to avoid selection biases. We explore general patterns of stopover use by 
comparing the frequencies of stopover use between seasons and across 
populations. 
Next, we sought to determine if Turkey Vultures migrated with a time or energy 
minimization strategy, and to evaluate when Turkey Vultures began their 
5 
 
stopovers, relative to the onset of inclement weather. Birds that minimize 
energetic costs should fly as long as weather conditions allow for energy-
efficient soaring flight; therefore energy-minimizers would not switch to flapping 
flight and instead would stop as soon as weather conditions deteriorate. Birds 
that minimize time would fly as long as possible, switching to flapping flight and 
stopping after weather conditions deteriorate. Alternatively, birds may not be 
time or energy minimizers but instead may be risk-adverse and stop flying 
before weather conditions deteriorate. As obligate scavengers, Turkey Vultures 
have several behavioral and physiological adaptations to minimize energetic 
costs (Hatch 1970, Mallon et al. 2015, Grilli et al. 2017); therefore, we 
hypothesized that Turkey Vultures would use an energy minimization strategy. 
Last, we evaluated if average weather conditions affected movement activity 
during stopovers, or only the decision to stop and resume migration. 
 
1.3 Methods 
1.3.1 Study Species 
We used GPS-GSM tracking data collected between 2006 - 2019 from four 
migratory populations that represent three of the seven subspecies of Turkey 
Vulture (C. aura aura, C. aura meridionalis, and C. aura ruficollis). These 
populations range across most of the species? distribution ? from Canada to 
6 
 
southern South America (Figure 1.1). Tracking data were provided by Hawk 
Mountain Sanctuary (Pennsylvania, USA) and accessed via Movebank 
(Wikelski and Kays 2018). 
 
Figure 1.1. Map of daily migration locations for 33 individuals across four populations 
and three subspecies: Southwest USA (orange; C. aura aura), Central Canada 
(purple; C. aura meridionalis), Western Canada (dark green; C. aura meridionalis), 
Southern South America (light green; C. aura ruficollis). 
 
7 
 
1.3.2 Weather Predictors 
We accessed weather data using Movebank Env-data annotation (Dodge et al. 
2013). Several weather variables were selected (Table 1.1), which were 
considered important to the use of thermal or orographic soaring by Golden 
Eagles (Duerr et al. 2015). To evaluate the behavioral response of Turkey 
Vultures to changing weather conditions, we excluded any static variables (e.g. 
NDVI, landcover). 
Table 1.1. Definitions of weather variables used. For variables that have units, units 
are provided in parentheses. Weather data were sourced from European Centre for 
Medium-Range Weather Forecasts (ECMWF) or Movebank. 
Weather variable Source Definition 
Air temperature (C) ECMWF Air temperature 2 m above the ground 
Boundary height ECWMF The depth of air next to the earth?s surface which 
(m) is most affected by the resistance to the transfer 
of momentum, heat, or moisture across the 
surface. 
Downward ECMWF Amount of downward thermal (longwave) 
longwave thermal radiation (heat emitted by the earth?s surface and 
radiation (j m^-2) atmosphere) at the earth?s surface 
Downward ECMWF Amount of downward incident solar (shortwave) 
shortwave thermal radiation at the earth?s surface 
radiation (j m^-2) 
Latent heat flux    E  C MWF Exchange of latent heat with the surface through 
(j m^-2) turbulent diffusion (accumulated) 
Precipitation ECMWF The accumulated fraction of the model grid cell 
fraction that was covered by large-scale precipitation. 
Orographic updraft Movebank The velocity of upward air movement caused 
velocity (m/s) when rising terrain forces air to higher elevations 
8 
 
Sensible heat flux  ECMWF Exchange of heat with between the earth?s 
(j m^-2) surface and the atmosphere through turbulent air 
motion, excluding any heat transfer resulting 
from condensation or evaporation (accumulated) 
Thermal updraft Movebank Velocity of thermal updrafts (air that rises as it is 
velocity (m/s) heated by the sun near the earth?s surface) 
Total atmospheric ECMWF Total water in the entire atmospheric column 
water (kg m^-2) (water vapor + cloud water + cloud ice) 
Windspeed (m/s) ECMWF Calculated from wind u and v components 
 
1.3.3 Stopover Classification 
We used only migrations with regularly collected data (i.e., mean data interval 
<3 hours). We manually annotated migrations from continuous tracks using first 
passage time (FPT) (R package adehabitatLT; Calenge 2006). For the 
purposes of annotating migration and stopover locations, we interpolated any 
gaps within the movement trajectories to get hourly fixes. 
To find potential stopovers using FPT, we selected the radius that minimized 
log-variance of FPT (between 2500 and 6000 m). The radii we selected were 
based on patterns of activity and inactivity in our dataset and will vary among 
species. To find the start and end times of each stopover, we used a threshold 
relative to the structure of the data to reduce under-selection of stopovers from 
tracks with high FPT variance and over-selection of stopovers from tracks with 
low FPT variance. Stopovers selected from FPT were rejected if <25% of the 
data was during daylight hours or if the duration was <2 hours. To find non-
9 
 
stationary stopovers that were missed with this approach, we considered 
groups of >30 points within a 15km buffer to be a ?stopover?. To improve the 
precision of the start and end times of stopovers in our dataset, we removed 
the first and last observations if those speeds exceeded 95% of all the speeds 
during the stopover. If an individual ceased activity during normal roosting 
hours, we considered 0800 the following morning to be the start of the stopover. 
1.3.4 Analysis 
To determine if Turkey Vultures migrated with a time- or energy-minimization 
strategy, we looked for any evidence of lagged responses to changing weather 
variables. As not all stopovers were expected to be in response to weather, we 
sought to remove noise from the dataset by ranking stopovers by their 
proportion of activity. We used the 33% of stopovers characterized as the least 
active for these analyses. We subset each weather variable from 7 hours 
before to 7 hours after the start of each stopover. To facilitate comparison 
across populations with different climates and across years, we used the hourly 
change of each weather parameter as our response variables. For each hour, 
we then averaged the rate of change across all stopovers. For each variable, 
we used a generalized linear model with a loess smoother to visually inspect 
the average rate of change over time relative to the start of stopovers. We then 
determined if the start of stopovers lagged relative to the minima and maxima 
for each weather variable. To determine if the decision to depart from stopovers 
10 
 
lagged relative to changing environmental conditions, we repeated this analysis 
using the end of stopovers as the model predictor. We report the population-
specific responses to changing weather variables at the start and end of 
stopovers in Appendix A. 
To compare movement behavior during stopovers, we used FPT and 
proportion of activity as response variables. FPT at a small radius (250 m) is a 
measure of sedentariness while proportion of activity within a fixed radius is a 
measure of tortuousness. Proportion of activity was calculated during non-
roosting hours, and activity was defined as > 1 km/h. For predictors, we 
selected the top eight variables with the strongest link to the start or end of 
stopovers. To evaluate the effect of mean weather conditions during a stopover 
on movement behavior, we used generalized linear mixed models (GLMM) with 
individuals as the random effect. We repeated this analysis instead using linear 
regressions where we averaged the response and predictor values to the 
individual-level. We found no difference in the results between the GLMM and 
linear regression, so we only report the results of the linear regression here. All 
analyses were done using R version 4.0.2 (R Core Team 2020). 
11 
 
1.4 Results 
1.4.1 Description of Stopover Use 
We analyzed 174 migrations of 33 individuals from 2006 - 2019 and identified 
590 stopovers. Frequency of stopover use varied among population (Figure 
1.2). In spring, Southwest USA used 1.1 stopovers per migration (n = 44 total 
stopovers) with a mean distance between stopovers of 2690 ? 1022 km. 
Central Canada used 6.7 stopovers per migration (n = 128) with a mean 
distance between stopovers of 1409 ? 1295 km. Western Canada used 2.9 
stopovers per migration (n = 32) with a mean distance between stopovers of  
 
Figure 1.2. Mean distance between stopovers, per population. All populations except 
Western Canada, travel farther between stopovers in spring than fall. With the 
exception of 1 migration, no migrations traveled farther than 5000 km without stopping. 
12 
 
 
1371 ? 730 km. In austral spring, Southern South America used 2 stopovers 
per migration (n = 22) with a mean distance between stops of 1735 ? 771 km. 
In fall, all populations except Western Canada used stopovers more frequently 
(Figure 1.2). Southwest USA used 2.3 stopovers per migration (n = 103 total 
stopovers) with a mean distance between stopovers of 1993 ? 1157 km. 
Central Canada used 8.2 stopovers per migration (n = 188) with a mean 
distance between stops of 960 ? 507 km. Western Canada used 2 stopovers 
per migration (n = 22) with a mean distance between stops of 2187 ? 942 km. 
In austral fall, Southern South America used 3.4 stopovers per migration (n = 
51) with a mean distance between stopovers of 1382 ? 824 km. 
Movement behavior during stopovers varied from highly sedentary to highly 
tortuous (Figure 1.3). Total distances moved during stops ranged from 0 km to 
601 km. Duration of stopovers ranged from 2 hours to more than 11 days. Fully 
51% of stopovers (n=301) were < 24 h in duration. 58% of stopovers (n=342) 
started or ended on a day where birds migrated > 100 km. All stopovers > 24 
h (n=289) had at least one stopover day where the bird travelled a total distance 
<100 km. Most of these (n=260) had at least one stopover day where the bird 
travelled a total distance <25 km. 
 
13 
 
 
Figure 1.3. Examples of four stopovers, ranked from most tortuous (A) to most 
sedentary (D). More tortuous stopovers have a lower first passage time and a higher 
proportion of activity. More sedentary stopovers have a higher first passage time and 
a lower proportion of activity. 
 
1.4.2 Onset of Stopovers 
Stopovers typically began in the midafternoon (Supplemental Figure A.1). At 
the species-level, Turkey Vultures responded to changing weather variables 
immediately and showed no lag in response at the species-level. Several 
weather variables (Table 1.1) had peak rates of increase within one hour of the 
start of stopovers, including downward longwave thermal radiation, downward 
shortwave thermal radiation, precipitation fraction, and total atmospheric water 
(Figure 1.4). 
Latent heat flux and thermal updraft velocity had a maximum rate of decrease 
within one hour of the start of stopovers. There was no predictable response to 
change in surface pressure. Other weather variables showed no peak in the 
rate of change but a gradual change over time, including boundary height, 
sensible heat flux, and temperature. The peak rate of increase for wind speed 
14 
 
was more than one hour before the start of stopovers, suggesting either a 
lagged response by vultures, or that the decision to stop was correlated with 
the decrease in wind speed. Responses were similar across three of four 
populations; however, the Western Canada population showed a delayed 
response to several variables (Supplemental Figure A.2c). 
 
Figure 1.4. Average weather conditions relative to the start of identified stopovers (red 
line) for n = 539 stopovers. The labels on each plot indicate the name of the y-axis. 
The average hourly change in each weather variable is shown in blue and the error 
around this estimate is shown in gray. Peaks of several variables (i.e., downward 
longwave radiation, downward shortwave radiation, latent heat flux, precipitation 
fraction, thermal updraft velocity, and total atmospheric water) are within 1 hour of the 
start of stopover, indicating rapid response by Turkey Vultures to deteriorating weather 
conditions. Several other variables are declining at the start of stopover, i.e., boundary 
height, temperature, and wind speed. 
15 
 
1.4.3 Departure from Stopovers 
Stopovers typically ended in the early morning (Supplemental Figure A.1). At 
the species-level, Turkey Vultures showed a delayed response to changing 
weather variables when departing stopovers and resuming migration. Overall, 
the departure from stopovers was less predictable than the onset of stopovers, 
indicated by larger confidence intervals (Figure 1.5). Several weather variables 
had a maximum rate of increase more than one hour before the end of 
stopovers, including boundary height, orographic updraft velocity, surface 
pressure, temperature, thermal updraft velocity, and wind speed. 
Other weather variables that had a maximum rate of decrease within one hour 
of the end of stopovers, including downward longwave radiation, precipitation 
fraction, and sensible heat flux. There was no predictable response to change 
in orographic updraft velocity, latent heat flux, or total atmospheric water. Only 
downward shortwave thermal radiation had a peak rate of change after the end 
of the stopover. Responses were similar across three of four populations; 
however, the Western Canada population departed from stopovers during 
increasing rates of precipitation and downward longwave radiation 
(Supplemental Figure A.3c). 
 
16 
 
 
Figure 1.5. Average weather conditions relative to the end of identified stopovers (red 
line) for n = 539 stopovers. The labels on each plot indicate the name of the y-axis. 
The average hourly change in each weather variable is shown in blue and the error 
around this estimate is shown in gray. Peaks of several variables (i.e., boundary 
height, orographic updraft velocity, surface pressure, temperature, thermal updraft 
velocity, and wind speed) are > 3 hours before the end of stopover, indicating a 
delayed response by Turkey Vultures to improving weather conditions. Several other 
variables have little to no change at after the end of stopover, suggesting these 
variables negatively affect flight performance (i.e., downward longwave radiation, 
precipitation fraction, and total atmospheric water). 
 
17 
 
 
Figure 1.6. Average values for several weather variables as a function of mean 
proportion of activity during stopovers (a measure of tortuousness). All variables were 
significant except total atmospheric water. P value and R^2 are reported if p < 0.1. 
Several variables (i.e., boundary height, downward longwave radiation, downward 
shortwave radiation, temperature, and thermal velocity) were positively associated 
with greater amounts of activity during stopovers. Other variables (i.e., precipitation 
fraction and wind speed) were negatively associated with greater amounts of activity 
during stopovers, indicating that Turkey Vultures avoid moving during high levels of 
precipitation and high wind speeds. 
18 
 
 
Figure 1.7. Average values for several weather variables by mean first passage time 
during stopovers with a radius of 250 m (a measure of sedentariness). All variables 
were significant except total atmospheric water and thermal updraft velocity. P value 
and R^2 are reported if p < 0.1. Several variables (i.e., boundary height, downward 
longwave radiation, downward shortwave radiation, and temperature) were negatively 
associated with longer first passage times during stopovers. Other variables (i.e., 
precipitation fraction and wind speed) were positively associated with longer first 
passage times during stopovers, indicating that Turkey Vultures avoid moving during 
high levels of precipitation and high wind speeds. 
19 
 
1.4.4 Movement Behavior and Weather 
At the individual level, proportion of activity (Figure 1.6) and first passage time 
(Figure 1.7) were correlated with several weather variables associated with 
soaring. Birds had more tortuous stopovers at higher values of boundary 
height, downward longwave thermal radiation, downward shortwave thermal 
radiation, temperature, and thermal updraft velocity. Birds had less tortuous 
stopovers at higher values of precipitation and wind speed but had no 
relationship with total atmospheric water. 
Birds were more sedentary during stopovers when temperature, boundary 
height, downward longwave thermal radiation, and downward shortwave 
thermal radiation decreased (Figure 1.7). Birds were more sedentary at higher 
values of precipitation and wind speed but had no relationship with total 
atmospheric water or thermal updraft velocity. 
 
1.5 Discussion 
We found that Turkey Vultures frequently stop during migration to avoid flying 
during poor weather conditions, adding to the growing literature recognizing 
weather avoidance as an important driver of stopover use by migrating birds 
(Calvert et al. 2009, Arizaga et al. 2011, Clipp et al. 2020). Turkey Vultures 
ceased directed flight quickly in response to changes in several weather 
20 
 
variables, indicating Turkey Vultures migrate using an energy-minimization 
strategy. As obligate soaring birds, vultures are highly sensitive to changing 
weather conditions that affect the availability of updrafts. The variables they 
responded to are broadly associated with thermal updraft strength (Duerr et 
al. 2015, Poessel et al. 2018), which is the most important type of updraft for 
these migratory populations (Bohrer et al. 2012). The lack of response to 
orographic updraft velocity suggests that this was not an important updraft 
source for these populations, which largely avoid areas of high topographic 
relief in their migration routes. 
1.5.1 Weather Avoidant Stopovers 
Although Turkey Vultures responded rapidly to changing weather when 
choosing to stopover, they showed a delayed response when choosing to take-
off and resume migration. The most frequent time for take-off for vultures to 
resume migration was in the morning. At early hours, there is rapid change in 
variables such as temperature and thermal updraft velocity, which occur before 
thermals are fully formed. Therefore, this delayed response to changing 
conditions at the end of the stopover is likely an important part of their energy 
minimization strategy, as vultures wait until thermal updrafts have fully formed 
to avoid energetically expensive flapping flight (Harel et al. 2016a). 
Interestingly, Egyptian Vultures, which are also obligate soaring migrants, 
rarely stop for more than 24 h during migration (L?pez-L?pez et al. 2014, 
21 
 
Buechley et al. 2018). Compared to Turkey Vultures (Dodge et al. 2014), 
Egyptian Vultures migrate much shorter distances and migrate for fewer days 
(L?pez-L?pez et al. 2014). In addition to having shorter migrations, Egyptian 
Vultures also migrate through more arid landscapes than most of the 
populations in our dataset, which suggests Egyptian Vultures generally migrate 
during better soaring conditions than Turkey Vultures. 
We did not see a strong response to wind speed when Turkey Vultures began 
or departed the stopovers. Ospreys are also soaring migrants that frequently 
use stopovers and their decision to switch between active migration and 
stopover is not influenced by winds (Thorup et al. 2006). However, we did see 
that the proportion of activity during stopovers decreased with wind speed. In 
addition, first passage time during stopovers increased with wind speed. 
Although winds may not be a determining factor in the decision to actively 
migrate, vultures appear to be more likely to refuel during stopovers with only 
low wind speeds, which are correlated with conditions that promote the 
development of thermal updrafts (Bohrer et al. 2012). 
1.5.1 Movement Behavior during Stopovers 
Although we did not see support at the species-level for Turkey Vultures being 
risk-adverse and stopping before the peak rate of change in weather, the 
Western Canada population showed a delayed response to the peak rate of 
change in weather, suggesting some individuals may have used risky flight 
22 
 
behaviors (Horvitz et al. 2014). Across all populations, we found several 
stopovers where vultures were slow moving, which could support the aversion 
of a Turkey Vulture to stopping by attempting to make migratory progress while 
its progress was slowed by weak updrafts. 
Alternatively, such movements could represent migrants foraging en route. 
Several raptor species use a fly-and-forage strategy on migration, including 
Montagu?s Harriers Circus pygargus (Klaassen et al. 2017), Ospreys Pandion 
haliaetus (Strandberg and Alerstam 2007), and Eleonora?s Falcons Falco 
eleonorae (Mellone et al. 2013). Fly-and-forage results in lower daily migration 
distances (Klaassen et al. 2017) and in slower, more tortuous movements 
(Mellone et al. 2013), but allows migrants to continue in the direction of their 
goal. These movements are quite similar to Figure 1.2B, where Turkey Vultures 
continually move in the direction of their goal, but at a very slow pace. While 
vultures do not capture live prey like other fly-and-forage migrants, vultures 
search over large areas for carrion. Turkey Vultures specialize in flying at low 
altitudes over forested landscapes, searching for carrion odors (Mallon et al. 
2015). 
The strong linear relationship between movement activity during stopovers and 
weather conditions that promote the development of thermals further suggests 
not all stopovers are used for weather-avoidance. Although we are unable to 
comment on how many stopovers are used for weather-avoidance or refueling, 
23 
 
the use of stopovers clearly increased with migration distance. The Central 
Canada population migrates > 2000 km farther than the other populations, 
indicating a greater energetic burden that likely requires some refueling stops. 
Although Turkey Vultures are expected to complete most of their migrations 
while fasting (Bildstein 2006), this finding suggests that some vultures do stop 
to feed during migration. While weather-avoidance stopovers are important for 
an energy-minimization strategy, frequent stops may increase their energetic 
costs and require additional refueling stops. As climate change is expected to 
increase the frequency of storms (Webster et al. 2005), further investigation is 
required to fully understand how the energetic cost of migration is impacted by 
stopovers used for weather-avoidance. 
1.5.1 Short Stopovers 
Half of the stopovers we identified did not meet the < 100 km and > 24 h criteria 
commonly used by other studies. Interestingly, of the stopovers that met the < 
100 km criteria, 90% also met the < 25 km criteria. However, these stopovers 
do not necessarily align with the calendar day (i.e., 0000 - 2400 h). Only 4% of 
stopovers we identified started and ended during normal roosting hours 
(approx. 1700 - 0900 h), where the timing of movement activity would match 
with the normal daily activity patterns of diurnal migrants (Mallon et al. 2020). 
Stopovers identified at the daily scale, therefore, would have incorrectly 
identified the start and end times for 96% of stopovers. As revealed by our 
24 
 
results, birds are capable of responding rapidly to their environment; therefore, 
for studies interested in external drivers of stopover use it is essential that 
stopovers be identified at fine temporal scales. 
 
1.6 Conclusion 
This study is the first to attempt to determine the function of stopover use by 
migrant Turkey Vultures. We used a data-driven approach to identify short-
duration stopovers < 24 h, which are primarily used to avoid poor weather 
conditions. Movement behavior during stopovers was driven by local weather 
conditions, where individuals moved more frequently during conditions that 
promote thermal updraft development, likely in search of carrion. Future studies 
may apply our approach to other species to investigate stopover use in relation 
to environmental drivers. 
 
 
 
25 
 
Chapter 2: Time is energy: optimal migration of an 
obligate soaring migrant 
 
 
2.1 Abstract 
Migration is an energetically expensive endeavor that requires significant fuel 
stores to complete. It has long been suspected that Turkey Vultures (Cathartes 
aura) fast during migration due to competition for carrion and their inability to 
pursue prey. Here we show evidence confirming that not all Turkey Vultures 
need to feed during migration and that most go long distances without feeding. 
Using GPS-trajectories of 33 Turkey Vultures over 174 migrations, we 
evaluated daily migration costs based on time spent in three behaviors: active 
migration, stopover, or roosting. We estimated total energetic costs for each 
migration assuming the birds were fasting with reduced metabolic rates. We 
compared body mass stores required to fuel migrations for four populations, 
which differed in migration distances and body mass. We identified probable 
refueling stops for migrations where energetic costs exceeded 25% of their 
starting body mass. 
26 
 
Total migration costs were directly proportional to migration duration. Fall 
migration lasted on average 5 to 8.6 days longer than spring migration. For the 
longest distance migratory population, at least 95.2% of migrations required 
refueling stops whereas for the other migratory populations, only 25% of 
migrations required refueling stops. When controlling for migration duration, 
spring migrants complete their migration with greater body condition than fall 
migrants due to energy savings associated with fasting. Spring migrants delay 
feeding until after passing through the Mesoamerican migratory bottleneck, 
whereas fall migrants feed throughout the migration route. 
Our results indicate that only long-distance Turkey Vulture migrants regularly 
stop to feed, but spring migrants avoided feeding within the Mesoamerican 
passage bottleneck, as expected. Overall migratory costs were driven by 
migration duration, and therefore are lower in the spring when vultures migrate 
faster, which may contribute to seasonal differences in flight behavior. 
 
2.2 Introduction 
Seasonal migration is one of the most energetically expensive activities an 
organism can undertake. The energetics of migration by passerines and 
shorebirds has been well studied (Gudmundsson et al. 1991, G?mez et al. 
2017), indicating frequent use of stopovers to feed and replenish spent fat 
27 
 
reserves. For many bird species, the time and energy spent at stopovers 
exceeds the amount of time and energy spent in active flight (Hedenstr?m and 
Alerstam 1997, Wikelski et al. 2003). Frequent refueling allows birds to avoid 
carrying excess body mass and can be an important strategy that allows for 
minimizing energetic expenditure during migration (Alerstam 2001). Birds that 
anticipate changes in food abundance later along their migration route will alter 
their fuel deposition rates (Houston 1998), to minimize excessive time and 
energy costs associated with refueling. 
In contrast to passerines and shorebirds, refueling behavior by migrating 
raptors remains a knowledge gap concerning the theory of optimal migration 
(Alerstam 2011). While empirical data is lacking on the refueling behavior of 
many migrating raptor species, studies of a small number of species indicate 
that raptors can afford to use a variety of fueling strategies during migration. 
For example, Swainson?s Hawks (Buteo swainsoni) require frequent or long 
stopovers to complete their migration (Bechard et al. 2006, Kochert et al. 2011). 
Some species differ in feeding strategy between sexes (Alerstam et al. 2006) 
or among age classes (Restani 2000). Most raptor species cannot complete 
their migration while fasting; some are observed to regularly feed during 
migration (Goodrich 2010) and others are known to carry insufficient fat stores 
(Gorney and Yom-Tov 1994). Among migrating raptors in Israel, stopovers are 
not used en masse and individuals hunt opportunistically  (Yosef 1996). Other 
28 
 
short-distance migrants may not carry significant fat stores because they can 
afford to spend time at stopovers (Strandberg et al. 2009). 
The findings of these studies are broadly applicable to migrating raptors but are 
less useful for understanding the energetics of migrating vultures. Although 
also a member of the order Accipitriformes, vultures differ from other birds of 
prey in that they are obligate scavengers. Turkey Vultures (Cathartes aura) 
cannot pursue prey due to the lack of powerful flight muscles for pursuit flight 
and the lack of grasping talons to kill and carry prey. While they have 
occasionally been reported to take small, live prey (Platt and Rainwater 2009), 
they must otherwise find carrion, which is highly ephemeral. Vulture diets, 
therefore, prevent the broadscale use of stopovers (Bayly et al. 2018, Monti et 
al. 2018). Moreover, in some migratory bottlenecks, the quantity of carrion 
available in the environment is insufficient to sustain the vast numbers of other 
migrating and residential vultures (Bildstein 2006) as over 2 million vultures 
migrate through the Mesoamerican land corridor seasonally (Bildstein and 
Zalles 2001, Inzunza et al. 2010, Cabrera-Cruz and Villegas-Patraca 2016). 
Turkey Vultures have been assumed to not feed during migration (Bildstein 
2006) due to high competition for food during their migration. For example, 
vultures migrating from North America reach a bottleneck in the Isthmus of 
Tehuantepec, Mexico (hereafter the Mesoamerican migratory bottleneck), 
where migrants following each of Mexico?s coasts converge, en route to their 
29 
 
wintering grounds (Figure 2.1). This creates a highly competitive system as 
migrants overlap with birds from different flyways and with resident cathartid 
vultures. Furthermore, carrion is ephemeral by nature, and is haphazardly 
distributed in space and time throughout the landscape, making it difficult to 
find without deviating from the migration route. Migratory vultures may arrive at 
their wintering grounds in poor condition (Kirk and Gosler 1994), further 
suggesting that not all individuals feed during the journey. Observations of 
migratory vultures arriving at their wintering grounds in poor condition (Kirk and 
Gosler 1994) and a lack of observations of pellets at roosts (Smith 1980) further 
suggests that feeding is uncommon during migration. 
This presents a unique research opportunity: vultures are well-adapted to 
minimize energetic expenditure (Ruxton and Houston 2004), yet it is unknown 
how they cope with energetic costs during migration. Turkey Vultures frequently 
stop during migration, but the primary driver of such stops appears to be 
changing weather conditions (so-called ?fire escape? stopovers; Bayly et al. 
2018, Chapter 1). Here, we use mathematical and statistical models to evaluate 
if Turkey Vultures can successfully migrate while fasting. Considering the total 
energetic costs of migration, we identify probable stopovers used for feeding 
based on analyses of movement of four migratory vulture populations, three 
migrating from North America and one migrating within South America. We 
specifically test to see if vultures appear to feed within the Mesoamerican 
30 
 
migratory bottleneck. Last, we test if there are optimal stopover areas for 
migrating Turkey Vultures that would minimize their energetic costs. 
 
2.3 Methods 
2.3.1 Study Species 
Turkey Vultures (Cathartes aura) are habitat generalists and occur from 
southern Canada to the Falkland Islands (Bildstein 2006). There are several 
subspecies of Turkey Vulture, three of which (C. aura aura, C. aura 
meridionalis, C. aura ruficollis) are studied here. We studied four populations 
in all: each population?s body mass, subspecies, maximum migration distance, 
and region are shown in Table 2.1. 
 
Table 2.1. Descriptions of the breeding grounds, body mass, and migration distance 
by population. 
Population Subspecies Body Mass Maximum Migration (kg) Distance (km) 
Southwest USA C. aura aura 1.2 6427 
Central Canada C. aura meridionalis 1.5 8915 
Western Canada C. aura meridionalis 1.7 5959 
Southern South 
America C. aura ruficollis 1.7 5191 
 
31 
 
2.3.2 The physiological model 
Using heart rate data of a single migrating Turkey Vulture, we first estimated 
the relative metabolic costs of three behaviors: soaring, resting, and roosting. 
Using annotated Turkey Vulture GPS tracks, we then used time spent in each 
behavior per day to calculate daily energetic costs during migration. Vultures 
were assumed to fast during migration unless otherwise noted. 
From heart rate data to behavior-specific metabolic rates 
Diurnal migrants such as vultures roost at night and are typically active during 
daytime hours when environmental updrafts are available for soaring. During 
daytime hours, birds may either be soaring (i.e., actively migrating) or resting 
(i.e., stopped). To determine the costs associated with these three behavioral 
states (i.e., roosting, soaring, and resting) during migration, we used heart rate 
data from one focal bird during fall migration in 2003. Mandel et al. (2008) 
reports the methods used to capture and tag the focal bird. After being fitted 
with telemetry, the focal bird was then fed and released at 2003-11-06 18:30:00 
EST. Using heart rate data, we calculated the increase in metabolic rate over 
the expected metabolic rate at rest (Calder 1968, Arad and Bernstein 1988) 
and calculated the mean increase in heart rate by each behavior. We found no 
difference between active flight and daytime activity during stopovers, so we 
used 1.5 x BMR (Duriez et al. 2014) for both behaviors. Turkey Vultures drop 
their body temperature while roosting (Hatch 1970) and drop the body 
32 
 
temperature of their extremities when ambient temperature is outside their 
thermal neutral zone (Arad et al. 1989). To reflect this energy saving adaptation 
and avoid underestimating energetic costs across a variety of climatic regions, 
we chose to use 1.1 x BMR as our multiplier for roosting. 
Estimating total migration costs 
Typical fuel loads for migrating raptors are 4 - 18%, and body fat is higher in 
the spring than fall (DeLong and Hoffman 2004). However, these estimates 
may be low if hungry individuals have a greater likelihood of being trapped on 
migration (DeLong and Hoffman 2004). To be conservative, we allow for a body 
fat range up to 25%, considering that metabolism of fuel stores during migration 
involves a combination of fat and muscle tissues (Piersma 1990). We do not 
account for water loss in our model, as water is a product of the catabolism of 
lipids (Blem 1980) and soaring migrants are not shown to be water stressed 
during migration (Gorney and Yom-Tov 1994) 
The basal metabolic rate (BMR) for Turkey Vultures is 4.1 Watts/kg (Arad and 
Bernstein 1988). Metabolic costs of migration were calculated each day at 
hourly intervals using three behavioral states (????): actively migrating (1.5 x 
BMR), stopped over/resting (1.5 x BMR), and roosting (1.1 x BMR). We 
assumed that energy savings due to fasting were the same as for Old World 
vultures (Prinzinger et al. 2002). This assumption is reasonable because Old 
World and New World vultures have convergently evolved similar behaviors 
33 
 
and morphologies as obligate scavengers even though they are not closely 
related (Johnson et al. 2016). Both groups of birds must fast for days at a time 
and should have similar metabolic minimization strategies to compensate for 
irregular food intake. Days fasted (??) were calculated where day 0 (non-fast) 
occurred the day before departure. For days 1 - 7, an energy savings of 4% per 
day (??; Prinzinger et al. 2002) was subtracted from the hourly metabolic cost 
(where day 1 = 4%, day 2 = 8%, ? day 7 = 28%), but beyond this we assumed 
no further benefit from fasting. To account for fasting, metabolic rate specific to 
each behavior (????, kJ/h) was calculated as: 
????(???) = ???? ? ?????? ? (1?????)                                                      (1) 
where 
?? = ???: ????[0,7]7: ?? > 7          .         
For each day, metabolic expenditure for activity (??????, kJ/day) was calculated 
for each of the three behavioral states. ?????? was estimated from daily body 
mass (??), total duration of behavior (??, h/day), and ????. Unless otherwise 
stated, we assumed that the bird is fasting and not excreting feces or pellets. 
On day 1, we set ?? equal to the starting body mass in Table 2.1. After day 1, 
?? from the previous day is used (Eq. 5). 
?????? = ?? ? ??????(???)    .                                                             (2) 
34 
 
Daily energy expenditure (??????, kJ/day) was the sum of ?????? across the three 
behavioral states: soaring, resting, and roosting, yielding 
?????? = ??????? ??????        (3) 
Catabolism of fat yields 39.7 kJ/g (Ricklefs 1974), which is the energetic 
equivalent of body mass change (??). We calculated daily body mass change 
(????, g) from daily energy expenditure (??????) and ??, yielding 
???? = ??????/??       .                                                               (4) 
Body mass (??) was calculated as the starting body mass (??????????????, Table 2.1) 
minus the sum of daily body mass changes (????), specifically 
?? = ???? ????????????????=1 ??????     .                                                           (5) 
When simulating a feeding stop, birds feed on the first day of stopover where 
d is set to 0 and 200 g is added to the body mass (??). Fasting begins the 
following day, where d=1. 
2.3.3 Quantifying feeding stops 
To identify which migrations required feeding stops, we determined how many 
days a Turkey Vulture could migrate using less than 25% of its body mass. We 
also estimated how many additional days of migration could be afforded per 
feeding day, assuming they assimilated 200 g. For each migration, we 
determined if the total duration was longer than these thresholds. When the 
35 
 
total duration was longer, we calculated how many feeding stops would be 
required to complete the migration with no less than 75% of its starting body 
mass. 
Using stopovers identified from Chapter 1, we assigned likelihood scores that 
each stopover was used for feeding, depending on its tortuosity and duration. 
Tortuousness is associated with the search and capturing of prey (Fauchald 
and Tveraa 2003) and was positively associated with weather conditions that 
favored soaring flight, indicating highly tortuous stops were not used for 
weather avoidance (Chapter 1). We also assumed longer durations were more 
likely to be associated with the search and handling of food items. We made 
this assumption based on the rapid response by Turkey Vultures to changing 
weather conditions (Chapter 1), suggesting the duration of weather-avoidance 
stopovers are as short as possible.  
To calculate likelihood scores, we first calculated tortuosity of each stopover as 
the ratio of total distance of all locations within a stopover to straight line 
distance between the first and last stopover locations. We then increased the 
likelihood of stopovers if their durations with increasing total duration. 
Stopovers that largely overlapped with roosting and < 24 h were given the 
lowest likelihood scores.  
To identify how many feeding stops were necessary for an individual to 
complete its migration successfully, we first calculated the total amount of body 
36 
 
mass consumed for each migration, assuming the bird had fasted. Then we 
calculated the number of feeding stopovers needed to complete their migration 
while losing no more than 25% of their initial body mass. For migrations with 
several stopovers, the stopover with the highest likelihood score was selected 
as the feeding stopover. 
To estimate the number of feeding stops used, we modeled the minimum 
number of feeding stopovers that would allow all individuals to successfully 
migrate while consuming < 25% of their starting body mass. To evaluate the 
robustness of our estimates, we compared the number of days between 
successive feeding stops to the number of days it took Turkey Vultures to use 
25% of their body mass. We assumed the last day before migration and the 
first day post-migration were also days when vultures fed. 
Optimal timing of feeding 
We assumed that a vulture assimilated 200 g of mass from each refueling stop. 
In the wild, C. aura meridionalis consumes an estimated 228.5 g of carrion each 
day, based on their field metabolic rate (Grilli et al. 2019). Vultures regularly go 
several days without food (Hatch 1970), and therefore likely eat larger meals 
than their daily intake requirement when they find carrion. At least 80% of the 
mass of carrion is available for assimilation (Barton and Houston 1993) and 
Turkey Vultures have a metabolic efficiency of > 87% (Tabaka et al. 1996), 
suggesting 70% or more of the mass of carrion consumed can be assimilated 
37 
 
into tissue. However, it is unknown whether fasting would negatively affect this 
assimilation efficiency. 
We then explored if there were optimal times along the migratory route for 
refueling stopovers to be used. We simulated 30 different 30-day migrations, 
with one feeding stopover on every migration day. Each day consisted of 13 h  
 
Figure 2.1. Map of Turkey Vulture fall migration through Central America. Migrants 
from the central USA converge and follow the Gulf coast of Mexico to the Isthmus of 
Tehuantepec, where they converge with migrants from the western USA. Migrants 
then follow the narrow Mesoamerican migratory corridor to their overwintering grounds 
in northern South America. 
of roosting and 11 h of either active migration or resting. We report the results 
as final body mass on day 30. 
38 
 
To estimate the spatial distribution of feeding stops used, we assigned the 
stopovers with the highest likelihood scores as feeding stops. The number of 
feeding stops per migration was determined from migration duration. Due to 
the Mesoamerican migratory bottleneck (Figure 2.1), we expected North 
American Turkey Vulture populations to feed north of 30N latitude. During fall 
migration, vultures were expected to feed before reaching the bottleneck but 
during spring migration, vultures would feed after passing the bottleneck.  
 
2.4 Results 
2.4.1 Total migration costs 
We calculated the energetic costs for 94 fall and 80 spring migrations for 33 
birds. Total migration costs had a positive, linear relationship with migration 
duration (? = 0.8581, p < 0.001), and was not affected by population or starting 
body mass (Figure 2.2). Differences in proportion of body mass required among 
populations were due to relative differences in maximum migration distances 
(Table 2.1). 
2.4.2 Migrations requiring feeding stops 
We predicted that fasting Turkey Vultures used 25% of their body mass in 23 
days. Each feeding day that assimilated 200 g of mass afforded a vulture an 
extra 12 days of migration before reaching the 25% limit. 95.2% of the long- 
39 
 
 
Figure 2.2. Modeled results for percent body mass consumed during migration while 
fasting, with separate models for each population. Results from n = 174 fully completed 
migrations between 2003 - 2019. Percent body mass consumed, regardless of starting 
body mass, had a positive, linear relationship with migration duration. The dashed line 
indicates the 25% body mass threshold. 97.0%% of the Central Canada population?s 
migrations (green) exceed this threshold and would require feeding stops, whereas 
most migrations from the other three populations are below this threshold and no more 
than 29.9%% would require feeding stops. Differences in slopes between populations 
are attributable to decreasing rate of body mass loss over time. The rug plot shows 
the distribution of migration durations for each population. 
 
distance Central Canada population?s migrations exceeded the 25% body 
mass threshold (Figure 2.2) and required at least one feeding stop. Most 
migrations from the other three populations were below this threshold and only 
25.0% required at least one feeding stop. 87.9% of the Central Canada 
population required 2 or more feeding stops, compared to only 8.5% migrations 
40 
 
from the other three populations. Only 4.8% of the Central Canada population?s 
migrations and 75.0% of the other three population?s migrations did not require 
any feeding stops. The proportion of migrations that required feeding stops was 
greater overall in fall than spring (fall: 51.1%; spring: 31.2%). However, only 
short-distance populations required more feeding stops during fall than spring 
(Table 2.2). 
 
Table 2.2. Model estimates of the number of fall and spring migration that require 
feeding stops, by population. For each population, the total number of migrations and 
stopovers are reported by season. We also report the minimum number of non-fasting 
migrations and the predicted number of feeding stops used. 
Season Population Total Non-fasting Total Feeding Migrations Migrations Stops Stops 
Southwest USA 45 16 103 20 
 Central Canada 23 22 188 58 Fall
Western Canada 11 4 22 6 
Southern South America 15 6 51 11 
Southwest USA 39 3 44 3 
Spring Central Canada 19 18 128 36 
Western Canada 11 2 32 2 
Southern South America 11 2 22 2 
 
2.4.3 Quantifying feeding stops 
Of 590 stopovers, 23% were estimated to be used for feeding. These occurred 
during 42% of migrations (n = 174; Table 2.2). Feeding stops represented a 
marginally greater proportion of all stopovers during fall (26%; n=364) than 
spring (19%; n=226). Like the proportion of migrations that required feeding 
41 
 
stops, only short-distance populations required more feeding stops during fall 
than spring (Table 2.2). 
 
 
Figure 2.3 The difference in days between successive feeding stops. We assumed 
that the day before the onset of migration and the first day after migration were also 
feeding stops. Our estimates of which stops are used for feeding, using the 25% body 
mass rule, were very robust with 84% of feeding stops occurring < 23 days since the 
last feeding stop, indicating we underestimated the frequency of feeding stops 16% of 
the time. 
 
If our estimate of 23 days correctly represents the capacity of most Turkey 
Vultures to withstand fasting, then our estimates are robust as 84% of the 
feeding stops occurred < 23 days since the last feeding stop (Figure 2.3). For 
42 
 
16% of stops, we either underestimated the number of stops or incorrectly 
identified which stops were used for feeding. 
2.4.4 Optimal temporal and spatial patterns of feeding 
The spatial distribution of feeding stops used by C. aura meridionalis, differed 
slightly by season (Figure 2.4). Fall migrants infrequently stopped between 20N 
- 30N, while spring migrants completely avoided using feeding stops between 
20N - 30N. Spring migrants stopped to feed above 30N more often than fall 
migrants. The Western Canada population was less likely to stop to feed than 
the Central Canada population, regardless of season.  
The birds we predicted to finish migration with the best body condition and 
greatest total body mass were the birds that stopped as late as possible (Figure 
2.5). We predicted these birds finished migration with 3.4% greater final body 
mass than the birds that fed during the first days of migration. 
43 
 
 
 
Figure 2.4. Distribution of probable feeding stopovers (red) and active migration points 
(grey) for western and central C. aura meridionalis populations. The majority of feeding 
stopovers occur between 30 - 55 N latitude, regardless of the direction of migration. 
Feeding stops occur at any latitude during fall migration but near-exclusively above 30 
N during spring migration. There was no difference in the distribution of active 
migration points between seasons. The central population, which migrates farther, fed 
more extensively than the western population. 
44 
 
 
Figure 2.5. The effect of the timing of feeding on post-migratory body mass (g). We 
considered a 30 day migration and simulated the effect that a single 48 hour feeding 
stopover had on final body mass (g). The bird started at 1500 g body mass on day 0 
and used the same relative amounts of energy each day (i.e., 11 hours of activity and 
13 hours of roosting). During the stopover, the bird assimilated 200 g of body mass. 
Energy savings of 4% per day due to fasting (up to 28%) were applied from the start 
of migration and reset at the start of the stopover. Final body mass was lowest when 
stopovers occurred on days 5 - 10 and increased each day thereafter. Final body mass 
was greatest when the stopover was nearest to the end of the migration (i.e., day 28 - 
30).  
 
2.5 Discussion 
Our research reveals a unique optimal migration strategy, not seen in other 
migrants. Vultures minimize energy expenditure as they  require the use of 
soaring flight to reduce search costs so that they can profit from carrion (Ruxton 
and Houston 2004) and the durations of their flapping bouts are limited 
(Ferland-Raymond et al. 2005, Mallon et al. 2015). This results in total 
energetic costs that are driven more by migration duration than by behavior. 
45 
 
Therefore, an energy minimization strategy for Turkey Vultures is also a time 
minimization strategy. While vultures cannot sustain themselves aloft using 
active (i.e. flapping) flight, they can optimize their migration so that they spend 
the fewest days migrating as possible. 
Although the literature has considered Turkey Vultures to be fasting migrants, 
our results indicate this is only true for some individuals. The expected number 
of feeding stops is directly related to migration duration, and is higher for longer 
distance migrants (i.e., Central Canada population). Shorter distance migrants 
rarely need to feed more than once, if at all. However, our results did confirm 
that even for long-distance migrants that regularly feed, feeding stops are 
infrequent within the passage bottleneck and feeding is largely restricted to the 
continental United States, regardless of season. 
For Turkey Vultures that feed during migration, when they feed matters. Birds 
with the lowest energetic costs were birds that fed later. The difference in 
energetic costs is due to both the metabolic savings vultures have while fasting 
(Prinzinger et al. 2002) and that metabolic costs increase proportionally with 
body mass. If birds were able to optimize the timing of their feeding stops during 
both fall and spring migrations, we would find no difference in total energetic 
cost between seasons. However, due to avoidance of feeding within the 
Mesoamerican migratory bottleneck, the timing of feeding stops during 
migration results in a seasonal difference in energetic costs. Spring migrants 
46 
 
optimize their feeding stops by refueling after the Mesoamerican migratory 
bottleneck and maximize the energy savings associated with fasting. Fall 
migrants, in contrast, refuel before reaching the Mesoamerican migratory 
bottleneck and lose the metabolic savings due to fasting, when their body mass 
is relatively high. Fall migrants, therefore, complete their migration with lower 
body mass than spring migrants. 
Seasonal differences in the duration of migration, like that observed in our 
dataset (Chapter 1), is a well-known phenomena among migrating birds 
(Nilsson et al. 2013, Schmaljohann 2018). Non-breeding golden eagles have a 
shorter migration duration, only flying when conditions were optimal in order to 
minimize energy expenditure (Duerr et al. 2015). This suggests that conditions 
in spring are more favorable for soaring flight than during fall migration. In fall, 
they experience lower levels of downward solar radiation (Duerr et al. 2015), 
which generates weaker thermals and limits migration speed, contributing to 
longer and more energetically expensive fall migrations. Seasonal differences 
in migration speed of Turkey Vultures has been linked to longer migration days 
and greater thermal updraft strength in spring than in fall (Mellone et al. 2012, 
Dodge et al. 2014). 
The shorter duration of spring migration, combined with the directional effect 
on energetic gains associated with feeding, may result in large energetic 
savings for spring migrants. Consequently, spring migrants should arrive at 
47 
 
their breeding grounds with greater body condition, in preparation for the 
breeding season. Alternatively, empirical investigation of the energetic costs of 
migrations may reveal that Turkey Vultures use less energetically efficient flight 
behaviors, such as more flapping or more risky flight behaviors (Harel et al. 
2016a) to arrive at breeding grounds faster (Duerr et al. 2015). Although Turkey 
Vultures cannot maintain flight altitude using flapping flight due to a lack fast-
twitch muscle fibers (Rosser and George 1986), they still flap for take-off and 
landing, and may flap while soaring if the updrafts are weak (Ferland-Raymond 
et al. 2005), which would increase energetic costs. 
Without empirical data on where and when vultures feed, our estimates were 
largely restricted to identifying which migrations required feeding stops and how 
much carrion they need to consume. Our models serve as minimum estimates 
because vultures should feed opportunistically if a carcass were found along 
the migration route. We also expect that vultures will seek food before 
consuming all of their fuel stores, so some may stop to feed even though they 
could complete their migration while fasting. Our estimate of fasting for 23 days 
exceeds the currently known longest period of fasting a vulture has survived 
(16 days; Hatch 1970). However, the study had a limited sample size and used 
residential Turkey Vultures from the C. aura septentrionalis population, which 
may not fast for extended periods of time. The author also conducted the 
experimental fasting in January outside of any phenological cues that may 
48 
 
affect metabolism and while the vultures were exposed to temperatures outside 
their thermal neutral zone (Arad et al. 1989). Therefore, we expect vultures to 
be able to withstand longer periods of fasting than previously reported.  
Considering the number of days between feeding stops, 84% of the identified 
feeding stops agreed with our 23 day fasting limit. The other 16% of stops were 
either incorrectly identified, leading to large temporal gaps between 
consecutive feeding stops, or we underestimated the number of feeding stops 
used. Future studies of the movements of New World vultures, combined with 
stomach temperature data and accelerometry will help identify which stops are 
used for feeding with greater certainty and fill in gaps in our current 
understanding of how frequently New World vultures feed. As New World and 
Old World vultures are not closely related, it is important to study the effect of 
fasting on the metabolism of New World vultures. Documenting this energy 
efficient adaptation will confirm our findings that predict there is an effect of 
season and migration direction on the optimal migration strategies of Turkey 
Vultures. 
Further investigation using empirical data on energetic expenditure may reveal 
other interesting optimization strategies, such as energy minimization outside 
of the passage bottleneck and time minimization within the passage bottleneck 
(sensu Efrat et al. 2019). Similar energetic patterns may be found in other 
migrating vultures that also rely on ephemerally available carrion. Egyptian 
49 
 
Vultures (Neophron percnopterus) migrate through migratory bottlenecks 
(Buechley et al. 2018) where competition for carrion is expected to be high, if 
they feed during migration. Both Egyptian and Griffon Vultures (Gyps fulvus) 
cross ecological barriers and may have better flight performance (Harel et al. 
2016a) and greater energetic gain by feeding after crossing. 
 
2.6 Conclusion 
We concluded that many Turkey Vultures must feed during migration. Birds that 
migrate greater distances generally need to feed more often, but some 
individuals from the Central Canada population were able to complete their 
migrations quickly enough to avoid stopping to feed, especially in the Spring. 
Migration speed (Nilsson et al. 2013), therefore, is the most important factor to 
determine energetic costs. Consequently, season (Schmaljohann 2018), 
conditions experienced during migration (Duerr et al. 2015) and experience 
(Mueller et al. 2013) are all major contributing factors that influence energetic 
costs of soaring migrants.   
50 
 
Chapter 3: Diurnal timing of nonmigratory movement 
by birds: the importance of foraging spatial scales 
 
Previously published under: Julie M. Mallon, et al. 2020. Diurnal timing of 
nonmigratory movement by birds: the importance of foraging spatial scales. 
Journal of Avian Biology doi: 10.1111/jav.02612. 
 
3.1 Abstract 
Timing of activity can reveal an organism?s efforts to optimize foraging either 
by minimizing energy loss through passive movement or by maximizing 
energetic gain through foraging. Here, we assess whether signals of either of 
these strategies are detectable in the timing of activity of daily, local movements 
by birds. We compare the similarities of timing of movement activity among 
species using six temporal variables: start of activity relative to sunrise, end of 
activity relative to sunset, relative speed at midday, number of movement 
bouts, bout duration, and proportion of active daytime hours. We test for the 
influence of flight mode and foraging habitat on the timing of movement activity 
across avian guilds. We used 64570 days of GPS movement data collected 
between 2002 and 2019 for local (non-migratory) movements of 991 birds from 
51 
 
49 species, representing 14 orders. Dissimilarity among daily activity patterns 
was best explained by flight mode. Terrestrial soaring birds began activity later 
and stopped activity earlier than pelagic soaring or flapping birds. Broad-scale 
foraging habitat explained less of the clustering patterns because of divergent 
timing of active periods of pelagic surface and diving foragers. Among pelagic 
birds, surface foragers were active throughout the day while diving foragers 
matched their active hours more closely to daylight hours. Pelagic surface 
foragers also had the greatest daily foraging distances, which was consistent 
with their daytime activity patterns. This study demonstrates that flight mode 
and foraging habitat influence temporal patterns of daily movement activity of 
birds. 
3.2 Introduction 
An animal?s movement behavior is heavily influenced by its evolutionary 
history, which affects movement capacity and behavior (Norberg and Norberg 
1988, Tobalske 2001). An animal?s movement path is based, in part, on the 
distribution of resources (Fryxell et al. 2004), which is determined by their 
environment. These interact when animals forage, as they need to traverse the 
landscape according to their movement capacities to locate resources 
distributed non-randomly in the environment (Suryan et al. 2008). To maximize 
energetic gains from foraging, the timing of an animal?s foraging movements is 
expected to correspond to either the temporal availability of its resources 
52 
 
(Rydell et al. 1996, Lang et al. 2018) or the quantity and quality of resources 
required (Jetz et al. 2004, Ramesh et al. 2015, Cid et al. 2020). Alternatively, 
animals can reduce their energy expenditure by timing their foraging activity 
when their movements are most energetically efficient (Chapman et al. 2011, 
Shepard et al. 2013) via behavioral thermoregulation (Matern et al. 2000) and 
passive movement (Krupczynski and Schuster 2008). Both strategies are used 
by animals to forage optimally (Stephens and Krebs 1986), but these strategies 
have yet to be evaluated together within any group of animals. 
Birds are distinct from other vertebrates because most birds are volant and 
most fly actively (i.e., by flapping) while a smaller number fly passively (i.e., by 
soaring). Soaring birds save energy by using updrafts (Baudinette and 
Schmidt-Nielsen 1974) to move across the landscape. One tradeoff faced by 
terrestrial soaring birds is that the availability of updrafts is skewed towards 
daylight hours (Pennycuick 1978).  Switching to flapping flight can further 
extend the activity of soaring birds (Stark and Liechti 1993, Harel et al. 2016b) 
as flapping flight is self-powered and can therefore be used in a broader suite 
of conditions.  
When animals can be flexible in the timing of their movements, their activity is 
expected to be driven more by ecological interactions and the need to acquire 
resources. These needs can manifest as temporal matching between 
consumers and their resources. For example, Black?legged Kittiwakes (Rissa 
53 
 
tridactyla) time their foraging concurrently with tidal cycles, when prey are most 
accessible (Irons 1998). Alternatively, the amount of movement activity may be 
due to resource quality. When high quality food items are available, animals 
can spend more time resting as their energetic needs are met more quickly (Saj 
et al. 1999, Fleischer Jr et al. 2003, M?nard et al. 2013). Despite long lasting 
interest in the factors that shape animal activity times, it is still poorly 
understood how internal traits and external conditions jointly shape the timing 
of movement across avian species. 
Using daily movement activity data from a wide range of avian species, we 
tested for broad-scale differences in the temporal patterns by flight mode and 
foraging habitat. Temporal patterns do not only describe when individuals are 
moving, but they also convey information about the behaviors driving those 
movements (Pasquaretta et al. 2020). Therefore, temporal patterns of 
movement activity are best described using a suite of variables. First, we 
evaluated the similarity of temporal patterns among species using multivariate 
analyses and test for signals of foraging habitat and flight mode among clusters 
of species in ordinal space. Due to geographic and dietary segregation, we 
expected to find the greatest differences in multivariate space to be between 
birds from terrestrial and pelagic foraging habitats.  
Second, we hypothesized that the timing of daily movement activity is more 
restricted for species that soar, because the flight performance of soaring birds 
54 
 
varies within a day (Mellone et al. 2012) due to temporal variation in availability 
of environmentally derived updrafts (Spiegel et al. 2013). We predicted start 
and end times of movement activities would differ between flight modes. 
Flapping birds are unrestricted in their capacity to move and therefore can be 
active before sunrise and after sunset; in contrast, we expected terrestrial 
soaring birds to be limited to daylight hours. Soaring flight is most beneficial for 
large-bodied birds (Hedenstr?m 1993), which are often raptorial (Schoener 
1968); consequently, the use of soaring flight covaries with trophic level and 
morphology (Viscor and Fuster 1987, Baliga et al. 2019). We also predicted 
that pelagic soaring birds would be less temporally restricted than non-soaring 
birds as dynamic soaring is not driven directly by solar energy but by wind and 
wave energy (Pennycuick 1982). 
3.3 Methods 
3.3.1 Data 
We compiled GPS tracking data for 49 bird species whose movements were 
studied between 2002 and 2019. We obtained data from Movebank 
(www.movebank.org; Wikelski and Kays 2018) or through direct contributions 
by co-authors (Supplementary material Appendix B.1). For quality control, we 
removed anomalous locations with speeds greater than 80 kmh-1 for flapping 
species and locations with speeds greater than 100 kmh-1 for soaring species. 
All speeds were calculated as the speed between points. We calculated UTM 
55 
 
zones from coordinates. To evaluate the timing of movement relative to local 
sunrise and sunset, all timestamps were converted from GMT to local time. 
Our dataset included movements from 49 species (Supplementary material 
Appendix B.1). These species represent 14 orders: Accipitriformes, 
Anseriformes, Bucerotiformes, Charadriiformes, Ciconiiformes, Falconiformes, 
Gruiformes, Otidiformes, Passeriformes, Pelecaniformes, Phaethontiformes, 
Phoenicopteriformes, Procellariiformes, and Suliformes. Most (n=46) species 
were non-Passeriformes, and all Passeriformes were from the same genus 
(Corvus). 
We analyzed movement data at the daily scale. Most of the data were sampled 
at hourly time intervals, so we subsampled high resolution data to an hourly 
scale with location intervals ? 57 min (mean time between locations: 79.5 ? 
31.1 min). To accurately assess active and inactive states while maximizing 
number of sampling days, we excluded sampling intervals ? 180 min. We did 
not interpolate missing points. 
Days included in the analysis had a minimum of eight locations per day. We 
selected eight-hour minimum time periods to represent the daily scale because 
many telemetry units do not sample continuously and, instead, cycle on and off 
to save battery life. To avoid any potential bias in movements due to handling 
during tagging, we excluded the first day of tracking for all studies. We included 
56 
 
species with at least 20 days of data. Full sampling information is provided in 
Supplementary material (Appendix B.1). 
Due to known intra-specific differences that occur in association with migration 
(Cagnacci et al. 2016), our analyses explore non-migratory daily foraging 
movements. To compare local, foraging movements of birds, we standardized 
the data to include only non-migratory movements by excluding migrations from 
individuals with range shifts > 500 km. We intentionally selected a high 
threshold to avoid removing exploratory and foraging movements by individuals 
that did not migrate in partially migratory populations. 
3.3.2 Movement Characteristics 
Measurement errors due to error in calculations of latitude and longitude by 
global positioning system (GPS) are inherent in movement tracking studies 
(Frair et al. 2010) and can inflate estimates of movement activity. After 
comparing the distributions of location errors across species, we characterized 
locations as either ?active? or ?inactive? according to their mean speed. Species 
with a mean speed < 9 kmh-1 had an activity threshold of 50 mh-1. This threshold 
was conservative relative to the distributions of mean location errors across 
most of the species (Supplementary material Appendix B.2). Species with a 
mean speed > 9 kmh-1 had an activity threshold of 300 mh-1.   
57 
 
These different thresholds allowed us to identify active versus inactive periods 
for terrestrial and pelagic birds, which forage at different spatial scales 
(Schoener 1968, Oppel et al. 2018). To confirm our results were not sensitive 
to spatial scale, we compared our results using a smaller threshold (25 mh-1) 
and found no difference in the change in activity levels (Supplementary material 
Appendix B.2). To determine if the sampling frequency affected the activity 
patterns of any groups of species, we compared our results to a 20 min 
sampling scheme. The differences between these two sampling schemes were 
linear; our results are therefore robust to temporal sampling  (Supplementary 
material Appendix B.2). 
Table 3.1. Temporal variables and their definitions. 
Variable Definition 
Sunrise Activity The time difference between first activity and sunrise 
Sunset Activity The time difference between last activity and sunset 
Relative Speed at Speed at solar noon relative to mean speed 
Midday 
Number of Movement Number of groups with 1+ consecutive, active hours 
Bouts 
Activity Duration The length of time between non-active locations 
Proportion of Daytime Number of daytime active locations / total number of 
Activity daytime locations 
 
Based on these daily, active hours, we summarized temporal characteristics of 
daily movements using six variables, defined in Table 3.1. The objective of 
these measures was not to reliably estimate species averages for these 
temporal variables, but to provide standard, relative measures that could allow 
58 
 
for multispecies comparisons. We included the timing of activity relative to 
sunrise and sunset to understand the relationship between activity and light 
availability, while accounting for variation in latitudes and time of year across 
datasets. Several sampling regimes were set to collect data between sunrise 
and sunset, which limit our interpretations. However, these intervals were 
selected by experts on the focal species? biology, so we do not expect that the 
true mean start and end times of activity would differ strongly from our results. 
We list species with limited sampling periods (i.e., mean start or end of 
sampling time were within the hour of local sunrise and sunset) in 
Supplementary material Appendix B.2. The distributions of the timing of 
movement activity for each species are reported in Supplementary material 
Appendix B.3. 
To determine if movements were clustered in time or dispersed throughout the 
day, we defined number of movement bouts as the number of groups of 
consecutive active hours. We used the duration of these groups of consecutive 
active hours to represent activity duration. To determine how active species are 
at midday, for each day we calculated relative speed, which is the speed at 
solar noon divided by their speed averaged across all active bouts. Last, to 
compare activity among species, we calculated the proportion of time birds 
were active during the day, which was the proportion of hours between sunrise 
and sunset where the individual exceeded the speed threshold. We calculated 
59 
 
this metric using the number of hours during daytime, rather than hours during 
the full day, because species with limited sampling periods would have 
artificially high activity levels. We first calculated each temporal variable at the 
daily scale and then found the mean of each temporal variable at the species 
level (Supplementary material Appendix B.1). 
3.3.3 Morphological and Ecological Characteristics 
Ecological characteristic data were taken from the Elton 1.0 database (Wilman 
et al. 2016), which broadly describes the feeding ecology of all extant bird 
species in terms of the percent contribution of diet items and of different 
foraging habitats. We combined variables that were redundant for the species 
in our dataset; Table 3.2 lists the variables used and how they were derived. 
Table 3.2. Ecological variables used in analysis. Data sourced from Elton database 
(Wilman et al. 2016). Where variables are combined, sample sizes are indicated in 
parentheses. Final sample size used in analyses are in column N. 
Variable Category N Definition 
Foraging Above Ground 7 midcanopy (6) + canopy (3) + aerial (6) 
Habitat 
 Ground 34 ground (34) + understory (5) 
 Water (other) 29 freshwater or non-obligate pelagic species 
that forage below (4) + around surf (18) 
 Pelagic 7 pelagic specialist that forage around surf 
Surface 
 Pelagic Diver 8 pelagic specialist that forage below surf 
    
Diet Herbivore 17 plant (17) + seed (14) 
 Frugivore 6 fruit 
60 
 
 Carnivore 33 endotherms (18) + ectotherms (14) + 
unknown (4) 
 Piscivore 25 fish 
 Invertivore 32 invertebrates 
 Scavenger 20 carrion 
    
Flight Mode Pelagic 7 pelagic birds that soar >20% of the time 
Soaring 
 Obligate 8 terrestrial birds that cannot sustain flapping 
Soaring flight 
 Facultative 14 terrestrial birds that soar >20% of the time 
Soaring 
 Flapping 20 birds that flap >80% of the time 
 
Foraging habitats were collapsed to five levels: above ground, ground, 
freshwater, pelagic surface, and pelagic diver. Similarly, several diet variables 
were collapsed to six levels: herbivore, frugivore, carnivore, piscivore, 
invertivore, and scavenger. 
Flight mode was described as either flapping or soaring. Although many 
species may occasionally be observed soaring, we included only species that 
soar regularly. We further subdivided soaring into obligate, facultative, and 
pelagic soaring. All other species were categorized as flapping. 
We gathered morphometric data for three variables: body mass (kg), wing span 
(m), and wing area (m2). Where wing area values were missing, but wing span 
was known, we calculated wing area using aspect ratio (wingspan2/wingarea) 
61 
 
from a closely related species. Then, using known wingspan and estimated 
aspect ratio, we were able to derive wing area and relative wing loading 
(Pennycuick 2008). Wing spans were unknown for two species (Anas 
poecilorhyncha and Grus nigricollis), which we excluded from the analyses of 
morphological characteristics. We controlled for the effect of body size by using 
relative wing loading (bodymass2/3/wingarea; Norberg and Norberg 1988). We 
used only relative wing loading and aspect ratio in our analyses. All species? 
morphological data and sources, as well as ecological character data, are 
provided in the Supplementary material (Appendix B.4). 
3.3.4 Analysis 
To determine which guilds were most similar in the timing of movement activity, 
we quantified dissimilarity across the suite of temporal variables (listed in Table 
3.1) using non-metric multidimensional scaling (NMDS). NMDS is a distance-
based ordination that maximizes rank order correlation, which is suitable for 
non-parametric data. Accipitriformes and Anseriformes were over-represented 
in our dataset, making our dataset phylogenetically uneven.  
To correct for this, we bootstrapped our NMDS analysis by randomly 
subsampling four species (the median size of other orders with multiple 
individuals) within each order, iterated 100 times. For each iteration, we then 
tested for any significant diet, foraging, flight, and morphological correlates of 
the NMDS (Table 3.2). Our final analysis included only variables that were 
62 
 
significant predictors for at least 20% of subsampled datasets. This allowed us 
to exclude any predictors that would have been significant only due to the 
skewedness of our dataset. As many behaviors and adaptations have 
coevolved, we also report any highly correlated predictors. 
To test our flight mode hypothesis, we used one-way ANOVAs followed by 
TukeyHSD post-hoc tests. We excluded one species that was an outlier with 
regards to daytime movement activity, Cory?s shearwater (Calonectris 
diomedea), as our estimates of activity duration exceeded those of known 
activity budgets (Ramos et al. 2019). To explore the drivers of clustering among 
foraging habitats in ordinal space, we compared the distributions of active 
hours among foraging groups. To assess if the differences in activity level are 
due to physiological limitations of flight speed, we included a post-hoc analysis 
of the mean daily net squared displacement, a measurement of daily foraging 
distance, according to foraging habitat. Due to insufficient sample size, 
terrestrial above ground foragers were excluded from this analysis. We report 
summary statistics as mean and standard deviation. We performed analyses 
using R version 3.6.3 (2020-02-29) (R Core Team 2020); a list of R packages 
used can be found in the Supplementary material (Appendix B.5). 
 
63 
 
3.4 Results 
We summarized 64570 days of movement data for 991 birds. For three 
species, movement data came from fewer than three individuals 
(Supplementary material Appendix B.1). Wing spans ranged from 0.71 to 2.81 
m and body mass ranged from 0.44 to 9.87 kg, a range that includes the body 
masses of 28% of all volant non-Passeriformes. 
We found several continuous covariates related to the clustering of species 
according to the suite of temporal variables. Variation along NMDS1 was 
largely attributed to terrestrial ground foragers and pelagic surface foragers 
(Figure 3.1a). Terrestrial ground foragers were largely comprised of 
Accipitriformes and were therefore correlated with obligate and facultative 
soaring birds, scavenging (r=0.63), and carnivorous diets (r=0.71). Pelagic 
surface foragers were comprised of Procellariiformes and some Suliformes 
(i.e., Frigatebirds), which were positively correlated with high aspect ratio wings 
(r = 0.76), pelagic soaring, and invertivores. However, there was no separation 
between terrestrial and pelagic foragers in ordinal space. The greatest 
separation between foraging groups was between pelagic diving and pelagic 
surface foragers (Figure 3.1b). Variation along NMDS2 was largely attributed 
to flapping versus soaring flight (Figure 3.1c). In our dataset, body size was 
related to flight mode (soaring 3.65  ? 2.76 kg, flapping 1.20  ? 0.60 kg). 
64 
 
 
Figure 3.1. The NMDS ordination indicates inter-specific similarities within temporal 
activity patterns, among all 49 species. (a) NMDS annotated with environmental fit 
loadings (included if significant at p-value <0.05). Environmental fit loadings were 
bootstrapped to correct for an uneven sample across phylogeny. Ground foragers 
were correlated with carnivorous diets (r =0.71) and scavenging diets (r =0.63) and 
both were omitted from the environmental loadings for clarity. Pelagic surface foragers 
were correlated with high aspect ratio wings (r =0.76), which was removed for clarity. 
NMDS annotated by (b) pelagic foraging habitats and (c) flight mode. Ellipses 
represent 90% confidence interval around the centroid of each group. (b) There is little 
overlap between the pelagic foraging groups, indicating that pelagic divers (purple) 
have different activity patterns than pelagic surface foragers (green). Terrestrial 
foragers (grey) had high overlap with pelagic foragers, indicating little differences 
between terrestrial and pelagic foragers, overall. (c) There is little overlap between 
flight modes, indicating that soaring species (light green) have different activity 
patterns than flapping species (blue). Stress value is 0.15. 
 
Soaring birds had higher relative speeds than flapping birds at midday (soaring: 
0.901 ? 0.232; flapping: 0.568  ? 0.211; F = 26.28, df = 1, p <0.001). Obligate 
soaring birds began activity later than flapping birds (obligate  
65 
 
 
Figure 3.2. Dot plots of flapping, terrestrial (obligate and facultative), and pelagic 
soaring birds by (a) start of activity relative to sunrise and (b) end of activity relative to 
sunset, with units in hours. (a) Terrestrial soaring birds began activity after sunrise, 
with obligate soaring birds beginning activity later than facultative soaring birds. (b) 
Terrestrial soaring birds ceased activity before or at sunset, with obligate soaring birds 
stopping activity earlier than facultative soaring birds. Sunrise and sunset times were 
similar for facultative soaring and flapping birds. 
 
soaring: 3.250  ? 1.035 h; flapping: 0.750  ? 1.943 h; F = 14.542, df = 3, p < 
0.001; TukeyHSD p=0.017; Figure 3.2a). Similarly, obligate soaring birds 
stopped activity earlier than did flapping birds (obligate soaring: -1.286  ? 0.881 
66 
 
h; flapping: 0.850  ? 1.755 h; F = 6.777, df = 3, p < 0.001; TukeyHSD p= 0.018; 
Figure 3.2b). The same pattern was observed for pelagic soaring birds. 
Obligate soaring birds began activity later than pelagic soaring birds (pelagic 
soaring: -3.143  ? 3.532 h; TukeyHSD p < 0.001; Figure 3.2a) and obligate 
soaring birds stopped activity earlier than pelagic soaring birds (pelagic 
soaring: 2.429  ? 2.37 h; TukeyHSD p <0.001; Figure 3.2b). Post-hoc tests did 
not reveal significant differences in the start or end times between obligate and 
facultative soaring birds (sunrise p = 0.159, sunset p = 0.224), but obligate 
soaring birds were active for a shorter range of hours in the day (Figure 3.2).  
Activity distributions differed by foraging habitats (Figure 3.3a). Pelagic surface 
foragers were active most continuously throughout the day and terrestrial 
ground foragers had the narrowest range of active hours. Differences in the 
activity patterns of pelagic surface foragers and diving foragers corresponded 
to differences in daily foraging distances (Figure 3.3c). Daily maximum net 
squared displacement was greatest among pelagic surface foragers, indicating 
they travelled the furthest within a day of any foraging group (F = 3.373, df = 3, 
p = 0.027). These differences were not due to differences in mean flight speed 
(Figure 3.3b). Pelagic foragers had greater mean flight speeds than terrestrial 
foragers, but this was partially an artifact of our methods requiring different 
activity thresholds. 
67 
 
 
Figure 3.3. Foraging habitats by (a) active time, (b) mean speed, and (c) distance. For 
all plots, terrestrial above ground foragers were excluded due to small sample size (n 
= 2). (a) Distributions of active (black) and inactive (grey) hours by foraging habitat. 
Pelagic surface foragers were active a greater proportion of the day than pelagic diving 
and terrestrial foragers, whose activity was more clustered during midday. (b) Mean 
speed between points. There is no difference in maximum speeds among pelagic 
foraging habitats. (c) Post-hoc analyses of log-transformed squared net displacement 
(in meters) of daily foraging trips according to foraging habitat. Pelagic surface 
foragers travel farther than pelagic divers on daily foraging trips, suggesting 
differences in their activity levels is driven by their respective foraging distances. 
 
68 
 
3.5 Discussion 
In this study we have combined a rich GPS tracking data set, spanning over 
several species and guilds, and used a multispecies comparative approach to 
test for intrinsic factors that shape the timing of activity by birds. We found 
broad-scale differences in the timing of avian daily movement activity between 
flight modes, supporting our hypothesis. Movements of Accipitriformes, which 
represent the largest proportion of soaring birds in our broad dataset, were 
largely restricted to daytime hours. This effect was even stronger among 
obligate soaring birds (i.e. Old World and New World vultures). Soaring species 
were further differentiated from flapping species by higher relative speeds at 
midday. These findings were not surprising as updrafts are stronger around 
midday than in the morning or late afternoon, supporting previous research 
suggesting their activity is more strongly linked to the temporal availability of 
updrafts (Mandel and Bildstein 2007, Bildstein et al. 2009, Nathan et al. 2012, 
Sur et al. 2017) than to their spatial availability (Mallon et al. 2015). 
Flapping species were characterized by a lower percent of activity during the 
day. This suggests either flapping species are less active than soaring species, 
or they are similarly active, but on different spatial scales. Flapping species 
were largely represented by Anseriformes (i.e., ducks and geese) and 
Pelecaniformes (i.e., herons), which forage locally (e.g. < 1 km) by walking, 
stalking, dabbling, or grazing. These species generally feed on abundant or 
69 
 
localized resources (i.e., herbivores and granivores), and therefore spend 
greater amounts of time foraging within a given area (Mueller and Fagan 2008).  
Other species either face less temporal predictability of resources or have 
greater spatial heterogeneity of resources and are more mobile as a 
consequence (Mueller and Fagan 2008). This is true of soaring species, many 
of which use a fly-and-forage strategy where birds spend substantial time in 
flight searching for food over large spatial scales (e.g. 10?s of kms; Ruxton and 
Houston 2004). This is consistent with other findings concerning foraging space 
use: large-bodied birds, which tend to feed on high-quality resources and 
forage over large spatial scales (Schoener 1968), travel farther in 
homogeneous environments than heterogeneous environments (Tucker et al. 
2019). Among mammals, trophic level is correlated with home range size (Jetz 
et al. 2004), which is positively correlated with activity levels (Cid et al. 2020), 
suggesting a positive relationship between space use and activity levels over 
large scales.  
Like our results for terrestrial species, the temporal patterns we observed of 
pelagic species are a consequence of the spatial scale they forage over. While 
flight mode is related to the same morphological adaptations that allow pelagic 
species to specialize as surface or diving foragers (Ashmole 1971), we argue 
instead that the differences in timing among pelagic birds are not due to flight 
mode but to foraging behavior. In our dataset, pelagic surface foragers were 
70 
 
comprised of Suliformes (i.e. boobies and frigatebirds) and Procellariiformes 
(i.e. albatrosses and shearwaters), which forage over different spatial scales 
(Oppel et al. 2018). Although in other colonies, Suliformes respond to 
intraspecific competition by traveling further from the colony to forage (Oppel 
et al. 2015), Suliformes in our dataset forage closer to their colonies relative to 
the Procellariiformes, which frequently forage in open ocean. This difference in 
space use also likely drives the observed differences in the temporal patterns 
of their movement activity. To travel further, but at similar flight speeds, 
Procellariiformes have longer foraging trips that often extend overnight. This 
resulted in Suliformes appearing to be relatively less active, as their foraging 
trips in our dataset were always < 24 h. The predominantly diurnal activities of 
Suliformes contributed greatly to the overlap in temporal activity patterns 
between pelagic and terrestrial foragers. There was better contrast among 
pelagic birds when comparing foraging groups, as frigatebirds are not diving 
foragers like other Suliformes, but are surface foragers that behave more like 
Procellariiformes. Frigatebirds in our dataset did move at night but are diurnal 
foragers that sleep on the wing (Rattenborg et al. 2016). This, in part, explains 
why the differences in start times between pelagic surface and diving foragers 
were more distinct than between Procellariiformes and Suliformes alone. 
At least for some species, the relative significance of flight mode and foraging 
habitat may not be clear cut. The timing of their movements may not be driven 
71 
 
by food availability, but instead by foraging restrictions. For example, in arid 
climates, some birds reduce their activity during midday as a means of 
behavioral thermoregulation (Silva et al. 2015, Gudka et al. 2019). Likewise, 
visually orienting species are limited by the availability of light. As such, 
although fruits and seeds are available at all hours, Passerines begin activity 
at dawn when there is sufficient light to detect their food resources (Roth and 
Lima 2007). Temporal segregation of foraging can also be driven by pressures 
to avoid predators or kleptoparasites (Baglione and Canestrari 2009), such as 
frigatebirds. Such adaptive behavior is thought to have contributed to the 
evolution of nocturnal foraging behaviors by some pelagic species (Hailman 
1964). 
Interpretation of our results is limited as we compiled our dataset from several 
different studies, which were biased towards larger, data-rich species that can 
support the weight of telemetry units. Also, sampling schemes across studies 
were uneven in terms of inter-location frequency and effort; this required us to 
use data averaged at the species level. If our data could be resolved on the 
scales specific to each guild, rather than standardized across species, we might 
have identified other ecological variables, such as diet, as important drivers of 
movement activity. Nevertheless, our approach provided standardized activity 
metrics for 49 bird species, which allowed us to compare intrinsic drivers of 
movement activity across a diversity of avian guilds.  
72 
 
Although our analyses were restricted to temporal attributes of movement, the 
relationship between physiological limitations on flight speed and activity 
duration lead us to hypothesize that the spatial scales animals forage over is 
an important driver of the timing of movement activity. Our results show that 
animals have predictable, intrinsic patterns to the timing of local movements 
that make up the large-scale behaviors we are interested in studying.  
Recognizing that spatial scale indirectly influences the timing of movement 
activity, future studies that focus on the spatial attributes of animal movement 
should consider the temporal attributes of movement as well. For example, 
studying spatial and temporal patterns in concert may reveal intraspecific 
differences due to personality influences on movement behavior (Spiegel et al. 
2017, Hertel et al. 2019). With the development of smaller, high-resolution 
tracking devices, future research may apply analyses such as ours to the full 
diversity of birds, filling gaps of our knowledge on granivorous, frugivorous and 
insectivorous species (e.g., passerines, shorebirds, swifts, etc.), which may 
reveal interesting new phylogenetic or allometric predictors of movement. 
3.6 Acknowledgements 
We would like to acknowledge Jose Manuel Lopez Vazquez, Steve Kirkland, 
Joseph Brandt, and ?a?an H. ?ekercio?lu for assistance in accessing data. 
We would also like to acknowledge those that have contributed data: Missy 
Braham, Annette Broderick, Adam Duerr, Sharon Poessel, and Brendan 
73 
 
Godley. We would like to thank Jesper Madsen, Justin Calabrese, and Ren? 
Janssen for their helpful comments. We would also like to acknowledge all of 
those that helped during data collection. 
 
  
74 
 
Appendix A: Supplementary Material Chapter 1 
 
 
Supplemental Figure A.1. Density plots of the timing of the starts and ends of 
stopovers. Stopovers started most frequently around 1500 h and ended most 
frequently around 1000 h. Typical roosting times begin and end approximately at 1700 
h and 0900 h, respectively. Due to some gaps in the data, some stopovers appeared 
to end before 0700 h, during normal roosting hours. 
 
  
75 
 
 (a) Southwest USA 
 
(b) Central Canada 
 
 
  
76 
 
(c) Western Canada 
 
(d) Southern South America 
 
Supplemental Figure A.2. Average weather conditions relative to the start of 
identified stopovers (red line), by population. The labels on each plot indicate the name 
of the y-axis. (a) Southwest USA, n = 138. (b) Central Canada, n = 283. (c) Western 
Canada, n = 49. (d) southern South America, n = 69. Differences in responses across 
populations due to a combination of unequal sample sizes, differing weather variable 
interactions associated with local climates, and stopovers used for feeding. 
77 
 
  
(a) Southwest USA 
 
(b) Central Canada 
 
 
  
78 
 
(c) Western Canada 
 
(d) Southern South America 
 
Supplemental Figure A.3. Average weather conditions relative to the end of identified 
stopovers (red line), by population. The labels on each plot indicate the name of the y-
axis. (a) Southwest USA, n = 138. (b) Central Canada, n = 283. (c) Western Canada, 
n = 49. (d) southern South America, n = 69. Differences in responses across 
populations due to a combination of unequal sample sizes, differing weather variable 
interactions associated with local climates, and stopovers used for feeding. 
79 
 
Appendix B: Supplementary Material Chapter 3 
 
All supplemental materials for chapter 3 are available at  
doi: 10.1111/jav.02612 
(current url: https://www.avianbiology.org/appendix/jav-02612). 
 
80 
 
Bibliography 
 
Alerstam, T. 2001. Detours in bird migration. - Journal of Theoretical Biology 
209: 319?331. 
Alerstam, T. 2011. Optimal bird migration revisited. - Journal of Ornithology 
152: 5?23. 
Alerstam, T., Hake, M. and Kjell?n, N. 2006. Temporal and spatial patterns of 
repeated migratory journeys by ospreys. - Animal Behaviour 71: 555?566. 
Allen, P. E., Goodrich, L. J. and Bildstein, K. L. 1996. Within-and among-year 
effects of cold fronts on migrating raptors at Hawk Mountain, Pennsylvania, 
1934-1991. - The Auk: 329?338. 
Arad, Z. and Bernstein, M. H. 1988. Temperature regulation in turkey 
vultures. - Condor: 913?919. 
Arad, Z., Midtg?rd, U. and Bernstein, M. H. 1989. Thermoregulation in turkey 
vultures: Vascular anatomy, arteriovenous heat exchange, and behavior. - 
The Condor 91: 505?514. 
Arizaga, J., Mendiburu, A., Andueza, M., Fontanilles, P., Fourcade, J.-M. and 
Urbina-Tobias, P. 2011. Deteriorating weather conditions predict the use of 
81 
 
suboptimal stopover sites by aquatic warblers Acrocephalus paludicola. - Acta 
Ornithologica 46: 202?206. 
Ashmole, N. P. 1971. Seabird ecology and the marine environment. - Avian 
biology 1: 223?286. 
Baglione, V. and Canestrari, D. 2009. Kleptoparasitism and temporal 
segregation of sympatric corvids foraging in a refuse dump. - The Auk 126: 
566?578. 
Baliga, V., Szabo, I. and Altshuler, D. 2019. Range of motion in the avian 
wing is strongly associated with flight behavior and body mass. - Science 
Advances 5: eaaw6670. 
Barton, N. W. and Houston, D. C. 1993. A comparison of digestive efficiency 
in birds of prey. - Ibis 135: 363?371. 
Baudinette, R. and Schmidt-Nielsen, K. 1974. Energy cost of gliding flight in 
herring gulls. - Nature 248: 83?84. 
Bayly, N. J., Rosenberg, K. V., Easton, W. E., Gomez, C., Carlisle, J., Ewert, 
D. N., Drake, A. and Goodrich, L. 2018. Major stopover regions and migratory 
bottlenecks for Nearctic-Neotropical landbirds within the Neotropics: A review. 
- Bird Conservation International 28: 1?26. 
82 
 
Bechard, M. J., Sarasola, J. H. and Woodbridge, B. 2006. A re-evaluation of 
evidence raises questions about the fasting migration hypothesis for 
Swainson?s hawk (Buteo swainsoni). - Hornero 21: 65?72. 
Bildstein, K. L. 2006. Migrating raptors of the world: Their ecology & 
conservation. - Cornell University Press. 
Bildstein, K. and Zalles, J. 2001. Raptor migration along the Mesoamerican 
land corridor. - Hawkwatching in the Americas. Hawk Migration Association of 
North America, North Wales, Pennsylvania: 119?141. 
Bildstein, K. L., Bechard, M. J., Farmer, C. and Newcomb, L. 2009. Narrow 
sea crossings present major obstacles to migrating griffon vultures Gyps 
fulvus. - Ibis 151: 382?391. 
Blem, C. R. 1980. The energetics of migration. - Animal migration, orientation, 
and navigation: 175?224. 
Bohrer, G., Brandes, D., Mandel, J. T., Bildstein, K. L., Miller, T. A., Lanzone, 
M., Katzner, T., Maisonneuve, C. and Tremblay, J. A. 2012. Estimating 
updraft velocity components over large spatial scales: Contrasting migration 
strategies of golden eagles and turkey vultures. - Ecology Letters 15: 96?103. 
Buechley, E. R., Oppel, S., Beatty, W. S., Nikolov, S. C., Dobrev, V., 
Arkumarev, V., Saravia, V., Bougain, C., Bounas, A., Kret, E. and others 
2018. Identifying critical migratory bottlenecks and high-use areas for an 
83 
 
endangered migratory soaring bird across three continents. - Journal of Avian 
Biology 49: e01629. 
Cabrera-Cruz, S. A. and Villegas-Patraca, R. 2016. Response of migrating 
raptors to an increasing number of wind farms. - Journal of Applied Ecology 
53: 1667?1675. 
Cagnacci, F., Focardi, S., Ghisla, A., Van Moorter, B., Merrill, E. H., Gurarie, 
E., Heurich, M., Mysterud, A., Linnell, J., Panzacchi, M. and others 2016. How 
many routes lead to migration? Comparison of methods to assess and 
characterize migratory movements. - Journal of Animal Ecology 85: 54?68. 
Calder, W. A. 1968. Respiratory and heart rates of birds at rest. - The Condor 
70: 358?365. 
Calenge, C. 2006. The package adehabitat for the R software: Tool for the 
analysis of space and habitat use by animals. - Ecological Modelling 197: 
1035. 
Calvert, A. M., Taylor, P. D. and Walde, S. 2009. Cross-scale environmental 
influences on migratory stopover behaviour. - Global change biology 15: 744?
759. 
Chapman, J. W., Klaassen, R. H., Drake, V. A., Fossette, S., Hays, G. C., 
Metcalfe, J. D., Reynolds, A. M., Reynolds, D. R. and Alerstam, T. 2011. 
84 
 
Animal orientation strategies for movement in flows. - Current Biology 21: 
R861?R870. 
Cid, B., Carbone, C., Fernandez, F. A., Jansen, P. A., Rowcliffe, J. M., 
O?Brien, T., Akampurira, E., Bitariho, R., Espinosa, S., Gajapersad, K. and 
others 2020. On the scaling of activity in tropical forest mammals. - Oikos 
129: 668?676. 
Clipp, H. L., Cohen, E. B., Smolinsky, J. A., Horton, K. G., Farnsworth, A. and 
Buler, J. J. 2020. Broad-scale weather patterns encountered during flight 
influence landbird stopover distributions. - Remote Sensing 12: 565. 
Crawford, R. E. and Long, J. A. 2017. Habitat preferences of juvenile Scottish 
ospreys Pandion haliaetus at stopover and wintering sites. - Ringing & 
Migration 32: 1?18. 
DeLong, J. and Hoffman, S. 2004. Fat stores of migrant sharp-shinned and 
Cooper?s hawks in New Mexico. - Journal of Raptor Research 38: 163?168. 
Dodge, S., Bohrer, G., Weinzierl, R., Davidson, S. C., Kays, R., Douglas, D., 
Cruz, S., Han, J., Brandes, D. and Wikelski, M. 2013. The environmental-data 
automated track annotation (Env-data) system: Linking animal tracks with 
environmental data. - Movement Ecology 1: 3. 
Dodge, S., Bohrer, G., Bildstein, K., Davidson, S. C., Weinzierl, R., Bechard, 
M. J., Barber, D., Kays, R., Brandes, D. and Han, J. 2014. Environmental 
85 
 
drivers of variability in the movement ecology of turkey vultures (Cathartes 
aura) in North and South America. - Phil. Trans. R. Soc. B 369: 20130195. 
Duerr, A. E., Miller, T. A., Lanzone, M., Brandes, D., Cooper, J., O?Malley, K., 
Maisonneuve, C., Tremblay, J. and Katzner, T. 2012. Testing an emerging 
paradigm in migration ecology shows surprising differences in efficiency 
between flight modes. - PLoS One 7: e35548. 
Duerr, A. E., Miller, T. A., Lanzone, M., Brandes, D., Cooper, J., O?Malley, K., 
Maisonneuve, C., Tremblay, J. A. and Katzner, T. 2015. Flight response of 
slope-soaring birds to seasonal variation in thermal generation. - Functional 
Ecology 29: 779?790. 
Duriez, O., Kato, A., Tromp, C., Dell?Omo, G., Vyssotski, A. L., Sarrazin, F. 
and Ropert-Coudert, Y. 2014. How cheap is soaring flight in raptors? A 
preliminary investigation in freely-flying vultures. - PLoS One 9: e84887. 
Efrat, R., Hatzofe, O. and Nathan, R. 2019. Landscape-dependent time 
versus energy optimizations in pelicans migrating through a large ecological 
barrier. - Functional Ecology 33: 2161?2171. 
Fauchald, P. and Tveraa, T. 2003. Using first-passage time in the analysis of 
area-restricted search and habitat selection. - Ecology 84: 282?288. 
86 
 
Ferland-Raymond, B., Bachand, M., Thomas, D. and Bildstein, K. 2005. 
Flapping rates of migrating and foraging turkey vultures Cathartes aura in 
Costa Rica. - Vulture News 53: 4?9. 
Fleischer Jr, A. L., Bowman, R. and Woolfenden, G. E. 2003. Variation in 
foraging behavior, diet, and time of breeding of Florida scrub-jays in suburban 
and wildland habitats. - The Condor 105: 515?527. 
Frair, J. L., Fieberg, J., Hebblewhite, M., Cagnacci, F., DeCesare, N. J. and 
Pedrotti, L. 2010. Resolving issues of imprecise and habitat-biased locations 
in ecological analyses using GPS telemetry data. - Philosophical Transactions 
of the Royal Society B: Biological Sciences 365: 2187?2200. 
Fryxell, J. M., Wilmshurst, J. F. and Sinclair, A. R. 2004. Predictive models of 
movement by Serengeti grazers. - Ecology 85: 2429?2435. 
G?mez, C., Bayly, N. J., Norris, D. R., Mackenzie, S. A., Rosenberg, K. V., 
Taylor, P. D., Hobson, K. A. and Cadena, C. D. 2017. Fuel loads acquired at 
a stopover site influence the pace of intercontinental migration in a boreal 
songbird. - Scientific Reports 7: 1?11. 
Goodrich, L. J. 2010. Stopover ecology of autumn-migrating raptors in the 
central Appalachians. 
87 
 
Gorney, E. and Yom-Tov, Y. 1994. Fat, hydration condition, and moult of 
steppe buzzards Buteo buteo vulpinus on spring migration. - Ibis 136: 185?
192. 
Grilli, M. G., Lambertucci, S. A., Therrien, J.-F. and Bildstein, K. L. 2017. 
Wing size but not wing shape is related to migratory behavior in a soaring 
bird. - Journal of Avian Biology 48: 669?678. 
Grilli, M. G., Bildstein, K. L. and Lambertucci, S. A. 2019. Nature?s clean-up 
crew: Quantifying ecosystem services offered by a migratory avian scavenger 
on a continental scale. - Ecosystem Services 39: 100990. 
Gudka, M., Santos, C. D., Dolman, P. M., Abad-G?mez, J. M. and Silva, J. P. 
2019. Feeling the heat: Elevated temperature affects male display activity of a 
lekking grassland bird. - PLoS One 14: e0221999. 
Gudmundsson, G. A., Lindstr?m, ?. and Alerstam, T. 1991. Optimal fat loads 
and long-distance flights by migrating knots Calidris canutus, sanderlings C. 
alba and turnstones Arenaria interpres. - Ibis 133: 140?152. 
Hailman, J. P. 1964. The Galapagos swallow-tailed gull is nocturnal. - The 
Wilson Bulletin: 347?354. 
Hake, M., Kjell?n, N. and Alerstam, T. 2001. Satellite tracking of Swedish 
ospreys Pandion haliaetus: Autumn migration routes and orientation. - 
Journal of Avian Biology 32: 47?56. 
88 
 
Harel, R., Duriez, O., Spiegel, O., Fluhr, J., Horvitz, N., Getz, W. M., Bouten, 
W., Sarrazin, F., Hatzofe, O. and Nathan, R. 2016a. Decision-making by a 
soaring bird: Time, energy and risk considerations at different spatio-temporal 
scales. - Philosophical Transactions of the Royal Society B: Biological 
Sciences 371: 20150397. 
Harel, R., Horvitz, N. and Nathan, R. 2016b. Adult vultures outperform 
juveniles in challenging thermal soaring conditions. - Scientific Reports 6: 
27865. 
Hatch, D. E. 1970. Energy conserving and heat dissipating mechanisms of 
the turkey vulture. - The Auk: 111?124. 
Hedenstr?m, A. 1993. Migration by soaring or flapping flight in birds: The 
relative importance of energy cost and speed. - Philosophical Transactions of 
the Royal Society of London B: Biological Sciences 342: 353?361. 
Hedenstr?m, A. and Alerstam, T. 1997. Optimum fuel loads in migratory birds: 
Distinguishing between time and energy minimization. - Journal of Theoretical 
Biology 189: 227?234. 
Hertel, A. G., Leclerc, M., Warren, D., Pelletier, F., Zedrosser, A. and Mueller, 
T. 2019. Don?t poke the bear: Using tracking data to quantify behavioural 
syndromes in elusive wildlife. - Animal Behaviour 147: 91?104. 
89 
 
Horvitz, N., Sapir, N., Liechti, F., Avissar, R., Mahrer, I. and Nathan, R. 2014. 
The gliding speed of migrating birds: Slow and safe or fast and risky? - 
Ecology Letters 17: 670?679. 
Houston, A. I. 1998. Models of optimal avian migration: State, time and 
predation. - Journal of Avian Biology: 395?404. 
Inzunza, E. R., Hoffman, S. W. and Goodrich, L. J. 2005. Stopover ecology of 
neotropical migrants in central Veracruz, Mexico. - In: Ralph, C. John; Rich, 
Terrell D., Editors 2005. Bird Conservation Implementation and Integration in 
the Americas: Proceedings of the Third International Partners in Flight 
Conference. 2002 March 20-24; Asilomar, California, Volume 2 gen. Tech. 
Rep. PSW-gtr-191. Albany, CA: US Dept. of Agriculture, Forest Service, 
Pacific Southwest Research Station: P. 657-672 191 
Inzunza, E. R., Goodrich, L. J. and Hoffman, S. W. 2010. North American 
population estimates of waterbirds, vultures and hawks from migration counts 
in Veracruz, Mexico. - Bird Conservation International 20: 124?133. 
Irons, D. B. 1998. Foraging area fidelity of individual seabirds in relation to 
tidal cycles and flock feeding. - Ecology 79: 647?655. 
Jetz, W., Carbone, C., Fulford, J. and Brown, J. H. 2004. The scaling of 
animal space use. - Science 306: 266?268. 
90 
 
Johnson, J. A., Brown, J. W., Fuchs, J. and Mindell, D. P. 2016. Multi-locus 
phylogenetic inference among New World vultures (Aves: Cathartidae). - 
Molecular Phylogenetics and Evolution 105: 193?199. 
Kerlinger, P. 1989. Flight strategies of migrating hawks. - University of 
Chicago Press. 
Kirk, D. A. and Gosler, A. G. 1994. Body condition varies with migration and 
competition in migrant and resident South American vultures. - The Auk: 933?
944. 
Klaassen, R. H., Schlaich, A. E., Bouten, W. and Koks, B. J. 2017. Migrating 
Montagu?s harriers frequently interrupt daily flights in both Europe and Africa. 
- Journal of Avian Biology 48: 180?190. 
Kochert, M. N., Fuller, M. R., Schueck, L. S., Bond, L., Bechard, M. J., 
Woodbridge, B., Holroyd, G. L., Martell, M. S. and Banasch, U. 2011. 
Migration patterns, use of stopover areas, and austral summer movements of 
Swainson?s hawks. - The Condor 113: 89?106. 
Krupczynski, P. and Schuster, S. 2008. Fruit-catching fish tune their fast 
starts to compensate for drift. - Current Biology 18: 1961?1965. 
Lang, S. D., Mann, R. P. and Farine, D. R. 2018. Temporal activity patterns of 
predators and prey across broad geographic scales. - Behavioral Ecology 30: 
172?180. 
91 
 
L?pez-L?pez, P., Garc??a-Ripoll?s, C. and Urios, V. 2014. Individual 
repeatability in timing and spatial flexibility of migration routes of trans-
Saharan migratory raptors. - Current Zoology 60: 642?652. 
Mallon, J. M., Bildstein, K. L. and Katzner, T. E. 2015. In-flight turbulence 
benefits soaring birds. - The Auk 133: 79?85. 
Mallon, J. M., Tucker, M., Beard, A., Bierregard Jr, R. O., Bildstein, K. L., 
B?hning-Gaese, K., Brzorad, J. N., Buechley, E. R., Bustamante, J., 
Carrapato, C., and others. 2020. Diurnal timing of nonmigratory movement by 
birds: The importance of foraging spatial scales. - Journal of Avian Biology: 
e02612. 
Mandel, J. T. and Bildstein, K. L. 2007. Turkey vultures use anthropogenic 
thermals to extend their daily activity period. - The Wilson Journal of 
Ornithology: 102?105. 
Mandel, J., Bildstein, K., Bohrer, G. and Winkler, D. 2008. Movement ecology 
of migration in turkey vultures. - Proceedings of the National Academy of 
Sciences 105: 19102?19107. 
Matern, S. A., Cech, J. J. and Hopkins, T. E. 2000. Diel movements of bat 
rays, Myliobatis californica, in Tomales Bay, California: Evidence for 
behavioral thermoregulation? - Environmental Biology of Fishes 58: 173?182. 
92 
 
Matthews, S. N. and Rodewald, P. G. 2010. Urban forest patches and 
stopover duration of migratory Swainson?s thrushes. - The Condor 112: 96?
104. 
Mellone, U., Klaassen, R. H., Garc??a-Ripoll?es, C., Limi?ana, R., L?opez-
L?opez, P., Pav?on, D., Strandberg, R., Urios, V., Vardakis, M. and Alerstam, 
T. 2012. Interspecific comparison of the performance of soaring migrants in 
relation to morphology, meteorological conditions and migration strategies. - 
PLoS One 7: e39833. 
Mellone, U., L?pez-L?pez, P., Limi?ana, R., Piasevoli, G. and Urios, V. 2013. 
The trans-equatorial loop migration system of Eleonora?s falcon: Differences 
in migration patterns between age classes, regions and seasons. - Journal of 
Avian Biology 44: 417?426. 
M?nard, N., Motsch, P., Delahaye, A., Saintvanne, A., Le Flohic, G., Dup?, 
S., Vallet, D., Qarro, M. and Pierre, J.-S. 2013. Effect of habitat quality on the 
ecological behaviour of a temperate-living primate: Time-budget adjustments. 
- Primates 54: 217?228. 
Monti, F., Gr?millet, D., Sforzi, A., Dominici, J. M., Bagur, R. T., Navarro, A. 
M., Fusani, L., Klaassen, R. H., Alerstam, T. and Duriez, O. 2018. Migration 
distance affects stopover use but not travel speed: Contrasting patterns 
93 
 
between long-and short-distance migrating ospreys. - Journal of Avian 
Biology 49: e01839. 
Mueller, T. and Fagan, W. F. 2008. Search and navigation in dynamic 
environments?from individual behaviors to population distributions. - Oikos 
117: 654?664. 
Mueller, T., O?Hara, R. B., Converse, S. J., Urbanek, R. P. and Fagan, W. F. 
2013. Social learning of migratory performance. - Science 341: 999?1002. 
Nathan, R., Spiegel, O., Fortmann-Roe, S., Harel, R., Wikelski, M. and Getz, 
W. M. 2012. Using tri-axial acceleration data to identify behavioral modes of 
free-ranging animals: General concepts and tools illustrated for griffon 
vultures. - Journal of Experimental Biology 215: 986?996. 
Nilsson, C., Klaassen, R. H. and Alerstam, T. 2013. Differences in speed and 
duration of bird migration between spring and autumn. - The American 
Naturalist 181: 837?845. 
Norberg, U. and Norberg, R. 1988. Ecomorphology of flight and tree-trunk 
climbing in birds (H Ouellet, Ed.). ? Proceedings of the International 
Ornithological Congress 19: 2271?2282. 
Oppel, S., Beard, A., Fox, D., Mackley, E., Leat, E., Henry, L., Clingham, E., 
Fowler, N., Sim, J., Sommerfeld, J. and others 2015. Foraging distribution of 
94 
 
a tropical seabird supports Ashmole?s hypothesis of population regulation. - 
Behavioral Ecology and Sociobiology 69: 915?926. 
Oppel, S., Bolton, M., Carneiro, A. P., Dias, M. P., Green, J. A., Masello, J. F., 
Phillips, R. A., Owen, E., Quillfeldt, P., Beard, A. and others 2018. Spatial 
scales of marine conservation management for breeding seabirds. - Marine 
Policy 98: 37?46. 
Pasquaretta, C., Dubois, T., Gomez-Moracho, T., Delepoulle, V. P., Le Loc?h, 
G., Heeb, P. and Lihoreau, M. 2020. Analysis of temporal patterns in animal 
movement networks. ? Methods in Ecology and Evolution 00: 1?13. 
Pennycuick, C. 1975. Mechanics of flight. - Avian Biology 5: 1?75. 
Pennycuick, C. 1978. Fifteen testable predictions about bird flight. - Oikos 30: 
165?176. 
Pennycuick, C. J. 1982. The flight of petrels and albatrosses 
(Procellariiformes), observed in South Georgia and its vicinity. - Philosophical 
Transactions of the Royal Society of London. B, Biological Sciences 300: 75?
106. 
Pennycuick, C. 2008. Modelling the flying bird. - Elsevier. 
Piersma, T. 1990. Pre-migratory ?fattening? usually involves more than the 
deposition of fat alone. - Ringing & Migration 11: 113?115. 
95 
 
Platt, S. G. and Rainwater, T. R. 2009. Noteworthy observations of foraging 
turkey vultures. - The Wilson Journal of Ornithology: 839?841. 
Poessel, S. A., Brandt, J., Miller, T. A. and Katzner, T. E. 2018. 
Meteorological and environmental variables affect flight behaviour and 
decision-making of an obligate soaring bird, the California condor Gymnogyps 
californianus. - Ibis 160: 36?53. 
Prinzinger, R., Nagel, B., Bahat, O., B?gel, R., Karl, E., Weihs, D. and 
Walzer, C. 2002. Energy metabolism and body temperature in the griffon 
vulture (Gyps fulvus) with comparative data on the hooded vulture 
(Necrosyrtes monachus) and the white-backed vulture (Gyps africanus). - 
Journal f?r Ornithologie 143: 456?467. 
Ramesh, T., Kalle, R., Sankar, K. and Qureshi, Q. 2015. Role of body size in 
activity budgets of mammals in the western ghats of India. - Journal of 
Tropical Ecology 31: 315?323. 
Ramos, R., Morera-Pujol, V., Cruz-Flores, M., L?pez-Souto, S., Brothers, M. 
and Gonz?lez-Sol??s, J. 2019. A geolocator-tagged fledgling provides first 
evidence on juvenile movements of Cory?s shearwater Calonectris borealis. - 
Bird Study 66: 283?288. 
96 
 
Rattenborg, N. C., Voirin, B., Cruz, S. M., Tisdale, R., Dell?Omo, G., Lipp, H.-
P., Wikelski, M. and Vyssotski, A. L. 2016. Evidence that birds sleep in mid-
flight. - Nature communications 7: 1?9. 
R Core Team 2020. R: A language and environment for statistical computing. 
- R Foundation for Statistical Computing. 
Restani, M. 2000. Age-specific stopover behavior of migrant bald eagles. - 
The Wilson Journal of Ornithology 112: 28?34. 
Ricklefs, R. 1974. The energetics of reproduction in birds. - Avian energetics 
15: 152?297. 
Rosser, B. W. and George, J. C. 1986. Slow muscle fibres in the pectoralis of 
the turkey vulture. - Zoologischer Anzeiger 217: 252?258. 
Roth, T. C. and Lima, S. L. 2007. The predatory behavior of wintering 
Accipiter hawks: Temporal patterns in activity of predators and prey. - 
Oecologia 152: 169?178. 
Ruxton, G. D. and Houston, D. C. 2004. Obligate vertebrate scavengers must 
be large soaring fliers. - Journal of Theoretical Biology 228: 431?436. 
Rydell, J., Entwistle, A. and Racey, P. A. 1996. Timing of foraging flights of 
three species of bats in relation to insect activity and predation risk. - Oikos: 
243?252. 
97 
 
Saj, T., Sicotte, P. and Paterson, J. D. 1999. Influence of human food 
consumption on the time budget of vervets. - International Journal of 
Primatology 20: 977?994. 
Schmaljohann, H. 2018. Proximate mechanisms affecting seasonal 
differences in migration speed of avian species. - Scientific Reports 8: 1?10. 
Schoener, T. W. 1968. Sizes of feeding territories among birds. - Ecology 49: 
123?141. 
Sergio, F., Tanferna, A., De Stephanis, R., Jim?nez, L. L., Blas, J. and 
Hiraldo, F. 2017. Migration by breeders and floaters of a long-lived raptor: 
Implications for recruitment and territory quality. - Animal Behaviour 131: 59?
72. 
Shepard, E. L., Wilson, R. P., Rees, W. G., Grundy, E., Lambertucci, S. A. 
and Vosper, S. B. 2013. Energy landscapes shape animal movement 
ecology. - The American Naturalist 182: 298?312. 
Shiu, H.-J., Tokita, K.-i., Morishita, E., Hiraoka, E., Wu, Y., Nakamura, H. and 
Higuchi, H. 2006. Route and site fidelity of two migratory raptors: Grey-faced 
buzzards Butastur indicus and honey-buzzards Pernis apivorus. - 
Ornithological Science 5: 151?156. 
98 
 
Silva, J. P., Catry, I., Palmeirim, J. M. and Moreira, F. 2015. Freezing heat: 
Thermally imposed constraints on the daily activity patterns of a free-ranging 
grassland bird. - Ecosphere 6: 1?13. 
Smith, N. G. 1980. Hawk and vulture migrations in the Neotropics. Pages 51-
65 in Keast, A. and Morton, E. S. Migrant birds in the Neotropics: Ecology, 
behavior, distribution, and conservation: A symposium held at the 
Conservation and Research Center, National Zoological Park, Smithsonian 
Institution, October 27-29, 1977. - The Symposia of the National Zoological 
Park (USA). 
Smith, A. D. and McWilliams, S. R. 2014. What to do when stopping over: 
behavioral decisions of a migrating songbird during stopover are dictated by 
initial change in their body condition and mediated by key environmental 
conditions. - Behavioral Ecology 25: 1423?1435. 
Spaar, R. and Bruderer, B. 1997. Migration by flapping or soaring: Flight 
strategies of marsh, Montagu?s and pallid harriers in southern Israel. - The 
Condor 99: 458?469. 
Spiegel, O., Getz, W. M. and Nathan, R. 2013. Factors influencing foraging 
search efficiency: Why do scarce lappet-faced vultures outperform ubiquitous 
white-backed vultures? - The American Naturalist 181: E102?E115. 
99 
 
Spiegel, O., Leu, S. T., Bull, C. M. and Sih, A. 2017. What?s your move? 
Movement as a link between personality and spatial dynamics in animal 
populations. - Ecology letters 20: 3?18. 
Stark, H. and Liechti, F. 1993. Do levant sparrowhawks Accipiter brevipes 
also migrate at night? - Ibis 135: 233?236. 
Stephens, D. W. and Krebs, J. R. 1986. Foraging theory. - Princeton 
University Press. 
Strandberg, R. and Alerstam, T. 2007. The strategy of fly-and-forage 
migration, illustrated for the osprey (Pandion haliaetus). - Behavioral Ecology 
and Sociobiology 61: 1865?1875. 
Strandberg, R., Alerstam, T., Hake, M. and Kjell?n, N. 2009. Short-distance 
migration of the common buzzard Buteo buteo recorded by satellite tracking. - 
Ibis 151: 200?206. 
Sur, M., Suffredini, T., Wessells, S. M., Bloom, P. H., Lanzone, M., 
Blackshire, S., Sridhar, S. and Katzner, T. 2017. Improved supervised 
classification of accelerometry data to distinguish behaviors of soaring birds. - 
PLoS One 12: e0174785. 
Suryan, R. M., Anderson, D. J., Shaffer, S. A., Roby, D. D., Tremblay, Y., 
Costa, D. P., Sievert, P. R., Sato, F., Ozaki, K., Balogh, G. R. and others 
100 
 
2008. Wind, waves, and wing loading: Morphological specialization may limit 
range expansion of endangered albatrosses. - PLoS One 3: e4016. 
Tabaka, C. S., Ullrey, D. E., Sikarskie, J. G., DeBar, S. R. and Ku, P. K. 1996. 
Diet, cast composition, and energy and nutrient intake of red-tailed hawks 
(Buteo jamaicensis), great horned owls (Bubo virginianus), and turkey 
vultures (Cathartes aura). - Journal of Zoo and Wildlife Medicine: 187?196. 
Thorup, K., Alerstam, T., Hake, M. and Kjell?n, N. 2006. Traveling or stopping 
of migrating birds in relation to wind: An illustration for the osprey. - 
Behavioral Ecology 17: 497?502. 
Tobalske, B. W. 2001. Morphology, velocity, and intermittent flight in birds. - 
American Zoologist 41: 177?187. 
Tucker, M. A., Alexandrou, O., Bierregaard Jr, R. O., Bildstein, K. L., B?hning-
Gaese, K., Bracis, C., Brzorad, J. N., Buechley, E. R., Cabot, D., Calabrese, 
J. M. and others 2019. Large birds travel farther in homogeneous 
environments. - Global Ecology and Biogeography 28: 576?587. 
Vansteelant, W., Bouten, W., Klaassen, R., Koks, B., Schlaich, A., Van 
Diermen, J., Van Loon, E. and Shamoun-Baranes, J. 2015. Regional and 
seasonal flight speeds of soaring migrants and the role of weather conditions 
at hourly and daily scales. - Journal of Avian Biology 46: 25?39. 
101 
 
Viscor, G. and Fuster, J. 1987. Relationships between morphological 
parameters in birds with different flying habits. - Comparative Biochemistry 
and Physiology Part A: Physiology 87: 231?249. 
Webster, P. J., Holland, G. J., Curry, J. A. and Chang, H.-R. 2005. Changes 
in tropical cyclone number, duration, and intensity in a warming environment. 
- Science 309: 1844?1846. 
Wikelski, M. and Kays, R. 2018. Movebank: Archive, analysis and sharing of 
animal movement data. - World Wide Web electronic publication. 
Wikelski, M., Tarlow, E. M., Raim, A., Diehl, R. H., Larkin, R. P. and Visser, 
G. H. 2003. Costs of migration in free-flying songbirds. - Nature 423: 704?
704. 
Wilman, H., Belmaker, J., Simpson, J., de la Rosa, C., Rivadeneira, M. M. 
and Jetz, W. 2016. EltonTraits 1.0: Species-level foraging attributes of the 
world?s birds and mammals. 
Yosef, R. 1996. Raptors feeding on migration at Eilat, Israel: opportunistic 
behavior or migratory strategy? - Journal of Raptor Research 30:242-245. 
 
 
 
102