Raphael Kelly
Age 15 | Calgary, Alberta
Canada-Wide Science Fair 2019 Excellence Award: Intermediate Level Silver Medal | Pacific Institue of Mathematical Sciences Award | Alberta Wilderness Society Award
Introduction
There are many studies that have followed migratory bird species with different characteristics and conservation statuses. Flack et al. (2016) followed the white stork (Ciconia ciconia), which is a Central European and African species that breeds in wetlands, open farmland, and sometimes on buildings (Sullivan et al., 2009). Hancock et al. (1992) and del Hoyo et al. (1992) determined that the white stork prefers to winter in drier climates over intensely cold and wet, or densely vegetated areas such as forests (as cited in Birdlife International, 2016c). The white stork is classified as least concern (LC) in Europe by the International Union for Conservation of Nature (IUCN) (Birdlife International, 2016c), due to population increase; however, the white stork is listed as a threatened species in the European Union’s (EU) Birds Directive, due to the bird’s decline in many European countries throughout the twentieth century (European Commission, 2019c). Kölzsch and Bauer et al. (2014) and Shariatinajafabadi et al. (2014) followed the barnacle goose (Branta leucopsis), a West European migratory species that winters in farmland and grassy wetlands near the coast of Europe (Sullivan et al., 2009). The barnacle goose is also classified as LC in Europe by the IUCN (Birdlife International, 2018a), but is listed as a threatened species by the European Union due to “conflicts with agriculture interests” (European Commission, 2019a). Kölzsch and Müskens et al. (2016) followed the greater white-fronted goose (Anser albifrons), which is a palearctic bird that breeds near bodies of water, on tundra, or on land, and winters in open areas near water (European Commission, 2019b). The greater white-fronted goose is classified as LC by the IUCN (Birdlife International, 2016a). Dodge et al. (2014) followed the turkey vulture (Cathartes aura), the large new world scavenger species that breeds and winters throughout the Americas. The main habitats of the turkey vulture include former and current forests, shrublands, grasslands, and deserts (BirdLife International, 2018b). The IUCN classifies this bird as LC (BirdLife International, 2018b). It has a score of 5 out of 20 on the State of North America’s Birds report, meaning it is of low concern (North American Bird Conservation Initiative, 2016). DeLuca et al. (2015) followed the blackpoll warbler (Setophaga striata), known as a songbird that inhabits the boreal forests and shrublands (Birdlife International, 2018c) of North America and migrates to the tip of South America (Sullivan et al., 2009). It is classified as near threatened (NT) by the IUCN (Birdlife International, 2018c) and is endangered in states such as Pennsylvania (Gross, 2014), both due to a decrease in its population size. It has a score of 11 out of 20 on the State of North America’s Birds report (North American Bird Conservation Initiative, 2016). Kochert et al. (2011) and Fuller et al. (1998) followed the Swainson’s hawk (Buteo swainsoni), which is found throughout prairies and agricultural regions of Canada and the United States, and winters in Central and South America (Sullivan et al., 2009). Globally, the Swainson’s hawk is categorized as LC by the IUCN (BirdLife International, 2016b); however, the species is considered threatened throughout parts of North America by bodies such as the California Department of Fish and Wildlife due to habitat loss (Battistone et al., 2016). According to the American Bird Conservancy (n.d.), its decline throughout the 1990s is attributable to the use of the monocrotophos pesticide. It has a concern score of 12 out of 20 on the State of North America’s Birds report (North American Bird Conservation Initiative, 2016).
As human populations grow and more land is used to accommodate human needs, animals’ natural habitats are encroached upon. If a migratory bird species is only present in areas with low human impact, that may indicate that the species cannot handle environmental changes due to urbanization. Therefore, the primary objective of this study is to determine the degree to which the habitats of migratory bird species are impacted by human influences on the environment. A secondary objective is to determine what common land features are present in the habitats of migratory birds, and to discuss how this information can be used to help conserve these species.
Hypothesis
It was hypothesized that most of the species of migratory birds studied would consistently show low (<40%) human impact values in their migratory locations. In addition, it was predicted that natural parks/reserves as well as water, grasslands, and trees would be common land features of the migratory locations.
Materials & Methods
This study uses the suggested resources and techniques derived from a Science Buddies article (Science Buddies Staff, 2019).
Animal Tracking
Geolocator tracking data available online for the six species was collected. Movebank, a free and open source research library containing numerous animal migration tracking datasets, was used to retrieve the studies (Kranstauber et al., 2011). Instead of choosing only one, all six tracking studies mentioned in the Science Buddies article (Flack et al., 2015; Griffin, 2014; Bildstein et al., 2014; DeLuca et al., n.d.; Kölzsch and Kruckenerg, 2016; Kochert et al., 2011, and Fuller et al., 1998) were analyzed using data from Movebank, allowing for cross-comparison of the results as well as greater detail and accuracy in the analysis. The data was downloaded and opened in the geographic information system (GIS) program Google Earth Pro, an open source software that has now been made free to the public.
Quantifying Characteristics of Urbanization
To analyze the human impact on individual habitats, this study used different categories related to urbanization. A table in the ScienceBuddies article included three categories with which urbanization can be analyzed: land use, buildings, and roads. Within each category are different characteristics that indicate the degree to which the environment has been urbanized (Science Buddies Staff, 2019). The categories and characteristics used for this project were developed from a preliminary look through of the Google Maps GIS data and the Science Buddies list. The characteristics for each category were ordered from being indicative of least to most human impact. Characteristics of the three categories are as follows:
1. Land Use: water/grasses/shrublands/forests, parks, plains, agricultural, urban
2. Buildings: none, scattered, moderate, intense
3. Roads: no roads, pathway, small road/many pathways, road, many roads.
This study was a descriptive analysis, deriving information from qualitative descriptions of bird habitats. However, to analyze and compare the large number of locations, the data needed to be quantified. A categorical approach to data collection and analysis, where every location would be assigned one specific category for each of the three variables, was considered. However, a numerical approach was chosen because it would allow for easier numerical analysis while accounting for locations that do not easily fall into one category. Numerical scales out of 15 were chosen because 15 was the smallest number that allows for adequate distinction between neighbouring descriptions. Every description for each of the variables was mapped out evenly across scales of 15. Within this study, any given location could be assigned a score for each variable. For instance, an agricultural location might receive a land use score of 8 to 13 (depending on how urbanized it is), while a location with no roads would receive a roads score of 0.
Modelling Human Impact
For ease of analysis, the three categories of urbanization were combined into a single variable that could describe the overall amount of human impact on a location. To combine these three variables into one summative assessment of the human impact at a given location, a model was prototyped and tested. It is a weighted average of the three variable scores put into percent form and is referred to as Hi (human impact value). The variable scores are weighted because while all three variable characteristics (i.e. land use, buildings, roads) indicate the amount of human impact on a location, some give more information than others. For instance, the land use variable gives more definitive information about exactly how a location is being used and the amount of human development in that location than the concentration of buildings and roads. In addition, because buildings protrude into the air, they are of more significance in a bird habitat than roads. Thus, in the human impact percentage value, the land use score makes up 45%, buildings 35%, and roads 20% of the final score. The derivation of the formula for determining human impact value is shown in Figure 1.
Figure 1. Derivation of the human impact value formula. Human impact value (Hi) is a weighted average of the Lu, B, and R variables. In step 1, the value is multiplied by 1/15 because each of the variables is scored out of 15. In step 2, the value is multiplied by 20/20 to turn the coefficients of the variables into integers.
With this model, any location could be scored using the three variables and subsequently assigned a human impact value. Since there is a higher quantity of characteristics describing low human impact (e.g. grasses, forest, parks, plains) than terms describing higher human impact (e.g. urban, agricultural, many roads), an even mapping of every characteristic on a linear scale resulted in a wider spread of the terms describing low human impact than the spread of the terms describing high human impact. Thus, it was determined that an Hi from about 0% to 40% will be considered low impact; 40% to 80% will be moderate impact; and +80% will be high impact.
Data Collection
Each geolocator tracking study that was downloaded was opened individually in Google Earth Pro. Every species was narrowed down to five individuals/tracks to analyze. Tracks that were chosen were those whose data did not have errors (in the Google Earth Pro representation) or end prematurely and that best represented the general path of the/each population that was being looked at. Track ID, migration length, and other identifications were collected. Next, the breeding, wintering, and stopover locations of each track were identified. For each location, the image was adjusted to a position of about 500 m above ground level (see Figure 2). From there, each location was scored based on the three variables. The results were collected in a spreadsheet and in a notebook. A total of 105 locations were analyzed with three qualitative data points each. Written descriptions were created, and the entire process was documented with images.
Figure 2. An image of a white stork breeding location taken from Google Earth (2017) at 53°17’34.21” N, 10°48’00.52” E. Breeding location found using data from Flack et al. (2015).
Results
During the exploratory phase, averages, standard deviations, and ranges for subsets of all the scores were calculated. The data and descriptions collected were reviewed together. These results are shown in Table 1. Graphs showing the scores for each bird in the study are shown in Figures 3 and 4.
The average Hi for each species in descending order was: 56% ± 36% (turkey vulture); 41% ± 22% (white stork); 40% ± 19.5% (greater white-fronted goose); 37% ± 19% (barnacle goose); 34% ± 17.5% (blackpoll warbler); and 29% ± 22% (Swainson’s hawk). The turkey vulture, white stork, and greater white-fronted goose all had average Hi values which mostly fall into the moderate impact category (40% to 80%). The barnacle goose, blackpoll warbler, and Swainson’s hawk all had average Hi values which mostly fall into the low impact category (0% to 40%). The final average Hi for every species was 40% with a range from ± 22.5% and a variance of 39%. Across every species, the average Hi value for the breeding locations was 33% ± 17.5%, 43% ± 21% for the stopover locations, and 43% ± 29% for the wintering locations. These averages mostly fall in the low-moderate impact range. The average of the buildings (B) scores assigned to each location was about 22.33% lower than the land use (Lu) and roads (R) variable scores. The five most common characteristics in the written descriptions of each location were, from most common to least: agricultural, forested, wetlands, plains, and grassy.
Table 1. A summary data table showing the statistical measures of mean, variance, and range for each set of recorded values by habitat location and by species. Variances of Hi values are multiplied x100 for ease of analysis.
Figure 3. Quantitative measures of human impact at migratory bird habitats by species, habitat and variable. Box graphs summarizing the data collected by species and habitat location.
Figure 4. A summary box graph showing the averages and spreads of recorded Hi values across species and habitat locations. summary box graph showing the averages and spreads of recorded Hi values across species and habitat locations.
Discussion
Primary Objective
Qualitative data about the relative degrees of human impact on the habitats of six migratory bird species were collected and quantified using numerical methods. Certain land features about the habitats were also documented. The primary objective of this study was to determine the degree to which the habitats of migratory species are impacted by human influences on the environment. There was a moderate variance and a wide range of Hi values among the whole data set. This demonstrates that the amount of human impact suitable for migratory birds varies among species and that some species are much more adaptable to urbanization than others.
The average Hi of the white stork’s breeding locations was 30% higher than its wintering locations. This is reflected in the EU’s label of the white stork as threatened in Europe (European Commission, 2019c)—since human contact with the species is more present in its European breeding locations than its African wintering habitats. The opposite is true for the barnacle goose whose average wintering location Hi was 31% higher than its breeding location. The data collected showed that the barnacle goose winters near wetlands and farmlands throughout western Europe, yet these wintering habitats have “conflicts with agriculture interests” as mentioned by the EU (European Commis- sion, 2019a). The same is true for the greater white-fronted goose whose average Hi score was 38% higher in its wintering locations than its breeding habitats. Since both species winter in highly populated areas of Europe, there is an increased opportunity for human interference.
The blackpoll warbler and Swainson’s hawk are both species of the Americas with moderate conservation concerns. The North America’s Birds report gave the species concern statuses of 11 and 12, respectively (North American Bird Conservation Initiative, 2016). These two species had the lowest average Hi scores (34% and 28%, respectively), possibly due to protective measures taken because of their conservation statuses. With the exception of certain agricultural areas and communities present in the wintering locations of the blackpoll warbler, the results showed that these species stayed in land with little urbanization, such as forests and parks.
The turkey vulture had the highest and widest range of Hi scores out of all the species, possibly due to high adaptability to “changing landcover, food sources and human proximity” (Campbell, 2015). The turkey vulture was the only bird studied that is not considered threatened or endangered by any conservation group addressed for this study. The images collected for the turkey vulture showed that it is present in highly populated cities, unlike the other species studied. The fact that its population size is stable (BirdLife International, 2018) indicates that the species is able to exist largely unaffected by urbanization.
Secondary Objective
The secondary objective is to determine what common land features are present in the habitats of migratory birds, and to discuss how this information can be used to help conserve these species. Among every species, certain land use features were characteristic. Mainly, these were agricultural areas, trees, and plains. In addition to these, the waterfowl species were present in locations with bodies of water (e.g. lakes) present. For almost every species, the lowest scores came from the buildings scores. This demonstrates that while more land based human impact in the form of agricultural development and the creation of roads is tolerable, buildings are not for many of these birds. This may be because while roads and agricultural fields do not project too far into the air, buildings provide direct interference to the flying paths of these birds. Loss et al. (2014) suggest that in the United States alone, anywhere between 365 and 988 million birds die annually as a result of collisions with tall buildings.
Sources of Error and Further Research
This study is a descriptive analysis meaning that the biggest possible source of error is human bias in the collection of the observational data. This flaw was partially combatted by developing and using the same criteria for ranking every location across species; however, human bias cannot be completely escaped. Another limitation is that although multiple tracking studies were used, the six studies represent less than 0.004% of approximately 1817 different migratory species around the world (Rolland et al., 2014). Thus, this data is not representative of migratory birds as a whole.
However, this study provides insight into the breeding, wintering, and stopover locations as well as corridors of the six migratory bird species studied. It also demonstrates one method of collecting and analyzing qualitative data numerically. For instance, a future study may use pre-determined land cover data to categorize and compare the features present in the habitats of different species.
Additionally, this project pinpoints many factors that need to be considered when developing conservation strategies for migratory species. For instance, because such a large percentage of the habitats, stopovers, and corridors of the migratory birds studied are agricultural, an increased use of certain pesticides, may have a negative effect on these populations of migratory birds. Neonicotinoids, a group of pesticides, have already been shown to have negative biological and behavioral effects on songbirds, with the potential to decrease the reproductive success and migration survival of these populations (Eng et al., 2019). By monitoring the use of pesticides, humans can decrease their impact on these birds.
A possible revision of this analysis would be to include more studies in order to get a clearer picture of what these results would be for a larger percentage of migratory bird species. Similarly, more studies involving one species or homologous species may help develop specific conservation plans for threatened species. More specific data could be collected to develop stronger conclusions. Alternatively, other variables which may present interesting trends (e.g. temperature, wind direction, etc.) could be studied to analyze other dimensions of the migratory patterns of bird species. A different method of collecting similar data that would include less human intervention could be the use a supervised machine learning application to classify satellite image data into categories such as agricultural or urban.
Conclusion
The hypothesis for this study is partially accepted. The average human impact on the locations of migratory birds are within the low-moderate range (Hi ≤ 60%). However, certain species (like the turkey vulture and Swainson’s hawk) are more adaptable to human development than others showing that not all migratory birds require low human impact in their habitats. Additionally, Hi values greater than 40% were consistent among the species, as were values less than 40%.
The most common type of land type present in the migratory habitats of all species studied was agricultural. This may indicate a need for special attention to be paid to the types of chemicals that are used in agricultural areas in order to best minimize the effect of human urbanization on migratory birds as a whole. Forested, wet, plain, and grassy areas were also common. The buildings scores were consistently the lowest among the variables recorded, which may imply that buildings negatively impact bird populations. Conversely, the roads scores were the highest implying that roads do not have a large impact on these birds.
Animal tracking can be done more easily using big data. This can serve to improve existing ecological networks. Online bird observation databases such as eBird.org can be used by both locals of bird habitats and researchers. Locals can easily gather information for researchers to analyze and improve the eco networks that drive their conservation methods. It is important that everyday people also know about this research so that they can be aware of the impact they can have on the world and attempt to make changes to the benefit and prosperity of all creatures on Earth.
Acknowledgements
This research would not have gotten as far as it did without the help of Ms. Shoults and Ms. Ferreira. I am also very grateful to Shannon, my editor, and, most of all, to my mom, Maria.
Raphael Kelly
My name is Raphael and I am currently a 15-year-old grade 10 student. I am a lover of mathematics and philosophy. Aside from academics, I also participate in basketball and track and field. In my free time, I like to solve math problems, read books, and make music. In the short term, I hope to learn more about data science and applied mathematics. In the long term, I hope to get involved in academia and research. My interests are connected through a value for creativity, pursuit of excellence, and belief in the importance of authentic humanity.
Works Cited
American Bird Conservancy. (n.d.). Swainson’s Hawk. American Bird Conser- vancy. Retrieved April 27, 2020, from https://abcbirds.org/bird/swain- sons-hawk/
Battistone, C., Marr, J., Gardner, T., Gifford, D. (2016, April 11). Status Review of Swainson’s Hawk in California. California Department of Fish and Wild- life, Wildlife and Fisheries Division, Nongame Wildlife Program. Retrieved from https://nrm.dfg.ca.gov/FileHandler.ashx?DocumentID=133622&in- line
Bildstein, K., Barber, D., & Bechard, M. J. (2014). Data from: Environmental drivers of variability in the movement ecology of turkey vultures (Cathartes aura) in North and South America. Movebank Data Repository. https://doi. org/doi:10.5441/001/1.46ft1k05
BirdLife International. (2016a). Anser albifrons. The IUCN Red List of Threat- ened Species 2016: e.T22679881A59950850. https://dx.doi.org/10.2305/ IUCN.UK.2016-3.RLTS.T22679881A85980652.en. Downloaded on 27 April 2020.
BirdLife International. (2016b). Buteo swainsoni. The IUCN Red List of Threat- ened Species 2016: e.T22695903A93533217. https://dx.doi.org/10.2305/ IUCN.UK.2016-3.RLTS.T22695903A93533217.en. Downloaded on 27 April 2020.
BirdLife International. (2016c). Ciconia ciconia. The IUCN Red List of Threatened Species 2016: e.T22697691A86248677. https://dx.doi.org/10.2305/ IUCN.UK.2016-3.RLTS.T22697691A86248677.en. Downloaded on 27 April 2020.
BirdLife International. (2018a). Branta leucopsis. The IUCN Red List of Threat- ened Species 2018: e.T22679943A131909954. https://dx.doi.org/10.2305/ IUCN.UK.2018-2.RLTS.T22679943A131909954.en. Downloaded on 27 April 2020.
BirdLife International. (2018b). Cathartes aura. The IUCN Red List of Threat- ened Species 2018: e.T22697627A131941613. https://dx.doi.org/10.2305/ IUCN.UK.2018-2.RLTS.T22697627A131941613.en. Downloaded on 27 April 2020.
BirdLife International. (2018c). Setophaga striata. The IUCN Red List of Threat- ened Species 2018: e.T22721737A131459482. https://dx.doi.org/10.2305/ IUCN.UK.2018-2.RLTS.T22721737A131459482.en. Downloaded on 27 April 2020.
Campbell, M. (2015). The adaptation of the turkey vulture (Cathartes aura, Lin- naeus 1758) to changing circumstances and ecologies. [Abstract]. [Simon Fraser University]. Research Gate.
DeLuca, W. V., Woodworth, B. K., Rimmer, C. C., Marra, P. P., Taylor, P. D., Mc- Farland, K. P., Mackenzie, S. A., & Norris, D. R. (n.d.) Data from: Trans- oceanic migration by a 12 g songbird. Movebank Data Repository. https:// doi.org/doi:10.5441/001/1.jb182ng4
DeLuca, W.V., Woodworth, B. K., Rimmer, C. C., Marra, P. P., Taylor, P. D., Mc- Farland, K. P., Mackenzie, S. A., & Norris, D. R. (2015) Transoceanic mi- gration by a 12 g songbird. Biology Letters 11(4). https://doi.org/10.1098/ rsbl.2014.1045
Dodge, S., Bohrer, G., Bildstein, K., Davidson, S.C., Weinzierl, R., Bechard, M.J., Barber, D., Kays, R., Brandes, D., Han, J., & Wikelski, M. (2014). Envi- ronmental drivers of variability in the movement ecology of turkey vultures (Cathartes aura) in North and South America. Philosophical Transactions of the Royal Society B, 369(1643). https://doi.org/10.1098/rstb.2013.0195
Dorst, J. P. (2019, August 6). Migration. In Encyclopædia Britannica. Retrieved April 2020, from https://www.britannica.com/science/migration-animal
Eng, M. L., Stutchbur, B. J. M., & Morrissey, C. A. (2019). A neonicotinoid in- secticide reduces fueling and delays migration in songbirds [Abstract]. Sci- ence, 365(6458), 1177-1180. https://doi.org/10.1126/science.aaw9419
European Commission. (2019, July 8a). Branta leocupsis. Retrieved April 2020, from https://ec.europa.eu/environment/nature/conservation/wildbirds/ threatened/b/branta_leucopsis_en.htm
European Commission. (2019, July 8b). White-Fronted Goose (Greenland Sub- species). Retrieved April 2020, from https://ec.europa.eu/environment/ nature/conservation/wildbirds/threatened/a/anser_albifrons_flavirostris_ en.htm
European Commission. (2019, July 8c). White Stork. Retrieved April 2020, from https://ec.europa.eu/environment/nature/conservation/wildbirds/threat- ened/c/ciconia_ciconia_en.htm
Flack, A., Fiedler, W., Blas, J., Pokrovski, I., Mitropolsky, B., Kaatz, M., Agh- ababyan, K., Khachatryan, A., Fakriadis, I., Makrigianni, E., Jerzak, L., Shamin, M., Shamina, C., Azafzaf, H., Feltrup-Azafzaf, C., Mokotjomela, T. M., & Wikelski, M. (2015). Data from: Costs of migratory decisions: A comparison across eight white stork populations. Movebank Data Reposi- tory. https://doi.org/doi:10.5441/001/1.78152p3q
Flack, A., Fiedler, W., Blas, J., Pokrovski, I., Kaatz, M., Mitropolsky, M., Agh- ababyan, K., Fakriadis, I., Makrigianni, E., Jerzak, L., Azafzaf, H., Fel- trup-Azafzaf, C., Rotics, S., Mokotjomela, T. M., Nathan, R., & Wikel- ski, M. (2016), Costs of migratory decisions: A comparison across eight white stork populations. Science Advances 2(1). https://doi.org/10.1126/ sciadv.1500931
Fuller, M. R., Seegar, W. S., & Schueck, L. S. (1998). Routes and travel rates of migrating peregrine falcons Falco peregrinus and Swainson’s hawks Buteo swainsoni in the western hemisphere. Journal of Avian Biology 29, 433- 440. https://doi.org/10.2307/3677162
Google Earth V 7.3.3.7673 (64-bit). (2017). [White Stork Breeding Location]. Retrieved May 14, 2020, from https://earth.google.com
Griffin, L. (2014). Data from: Forecasting spring from afar? Timing of migra- tion and predictability of phenology along different migration routes of an avian herbivore [Svalbard data]. Movebank Data Repository. doi:10.5441/001/1.5k6b1364
Gross, D. (2014, August 19). Blackpoll Warbler. Pennsylvania Game Commission https://www.pgc.pa.gov/Wildlife/EndangeredandThreatened/Pages/Black- pollWarbler.aspx
Kochert, M. N., Fuller, M. R., Schueck, L. S., Bond, L., Bechard, M. J., Wood- bridge, B., Holroyd G. L., Martell, M. S., & Banasch, U. (2011) Migra- tion patterns, use of stopover areas, and austral summer movements of Swainson’s hawks. The Condor, 113(1), 89-106. https://doi.org/10.1525/ cond.2011.090243
Kölzsch, A., Bauer, S., de Boer, R., Griffin, L., Cabot, D., Exo, K., van der Jeugd, H.P., & Nolet, B.A. (2014). Forecasting spring from afar? Timing of mi- gration and predictability of phenology along different migration routes of an avian herbivore. Journal of Animal Ecology, 84(1), 272-283. https://doi. org/10.1111/1365-2656.12281
Kölzsch, A., Kruckenberg, H., Glazov, P., Müskens, G. J. D. M., & Wikelski, M. (2016). Data from: Towards a new understanding of migration tim- ing: slower spring than autumn migration in geese reflects different de- cision rules for stopover use and departure. Movebank Data Repository. doi:10.5441/001/1.31c2v92f
Kölzsch, A., Müskens, G.J.D.M., Kruckenberg, H., Glazov, P., Weinzierl, R., No- let, B.A., & Wikelski, M. (2016) Towards a new understanding of migra- tion timing: slower spring than autumn migration in geese reflects different decision rules for stopover use and departure. Oikos, 125(10), 1496-1507. https://doi.org/10.1111/oik.03121
Kranstauber, B., Cameron, A., Weinzierl, R., Fountain, T., Tilak, S., Wikelski, M., & Kays, R. (2011). The Movebank data model for animal tracking. Envi- ronmental Modelling & Software 26(6): 834–835. https://doi.org/10.1016/j. envsoft.2010.12.005
Loss, S. R., Will, T., Loss, S. S., & Marra, P. P. (2014). Bird–building collisions in the United States: Estimates of annual mortality and species vulnerability. The Condor, 116(1), 8–23. https://doi.org/10.1650/CONDOR-13-090.1
Migration Corridors. (n.d.). All About Watersheds. Retrieved April 30, 2020, from https://allaboutwatersheds.org/library/kyw-poster-files-and-links/mi- gration-corridors
North American Bird Conservation Initiative. (2016). Species Assessment Sum- mary and Watch List. State of North America’s Birds 2016. https://www. stateofthebirds.org/2016/resources/species-assessments/
Rolland, J., Jiguet, F., Jønsson, K. A., Condamine, F. L., & Morlon, H. (2014). Set- tling down of seasonal migrants promotes bird diversification. Proceedings of the Royal Society B, 281(1784). https://doi.org/10.1098/rspb.2014.0473
Science Buddies Staff. (2019, February 5). Here Today, Gone Tomorrow: Sav- ing Migratory Animals. Science Buddies. https://www.sciencebuddies.org/ science-fair-projects/project-ideas/EnvSci_p057/environmental-science/ saving-migratory-animals
Shariatinajafabadi, M., Wang, T., Skidmore, A. K., Toxopeus, A. G., Kölzsch, A., Nolet, B. A., Exo, K-M., Griffin, L., Stahl, J., & Cabot, D. (2014) Migratory herbivorous waterfowl track satellite-derived green wave index. PLoS ONE 9(9): e108331. https://doi.org/10.1371/journal.pone.0108331
Sullivan, B. L., Wood, C. L., Iliff, M. J., Bonney, R. E., Fink, D., & Kelling S. (2009). eBird: a citizen-based bird observation network in the biological sciences. Biological Conservation 142: 2282-2292.