Season length influences breeding range dynamics of trumpeter swans Cygnus buccinator.
Investigating shifts in the range of migratory species can be particularly challenging because few monitoring schemes are conducted at spatial scales that cover a majority of a species' range, particularly for birds (Parmesan 2006). However, continued changes in climate will likely have many impacts on Arctic and subarctic bird populations (Seavy et al. 2008), and there is broad evidence that avian species' ranges are shifting northwards (La Sorte & Thompson 2007), most likely due to a longer growing season and longer ice-free periods for birds that use wetlands. However, for wetland-breeding birds some of the potential benefits of climate change may be offset through the loss or shrinking of wetlands across the landscape (e.g. Klein et al. 2005, Riordan et al. 2006). Reductions in size and numbers of wetlands would reduce overall available habitat, and changes in plant and invertebrate communities (Smol et al. 2005) could alter food availability.
We used a long-term data set on a northern breeding bird that relies on wetlands for nesting, the trumpeter swan Cygnus buccinator, to complete one of the first detailed examinations of the potential effects of climate change on the breeding range of an avian species. Because the trumpeter swan primarily uses boreal forest wetlands for reproduction and because this species requires a relatively long breeding period to successfully fledge young (145-150 days; Hansen et al. 1971), changes in the availability of wetlands induced by climate change could have a substantial impact on the entire population. This species has recently recovered from near extinction in the early 1900s with populations in Alaska increasing at a rate of nearly 6% annually since 1968 (Caithamer 2001, Conant et al. 2002, Schmidt et al. 2009a). Recent research has also suggested that the population may be expanding at a higher rate in more recent years in more northern areas (Schmidt et al. 2009a), where the breeding range now extends beyond the historic northern limit (Hansen et al. 1971). This expansion into new areas has resulted in recent observations of an overlap of the breeding distribution with that of the smaller tundra swan Cygnus columbianus in some areas. Historically, the length of season required for successful breeding (ca 30-45 days less for tundra swans) may have functionally separated these two species in space, but a warming climate may induce new forms of competition if range overlap increases. Recolonization of historic habitats by similar species, the whooper swan Cygnus cygnus and mute swan Cygnus olor, have not been found to displace other waterfowl species (Poysa & Sorjonen 2000, Gayet et al. 2011). However, the introduction of mute swans to areas where they have not occurred historically can have profound impacts on other waterfowl populations and wetland habitats (Petrie & Francis 2003). Because trumpeter swans have not historically occupied tundra habitats in large numbers, range increases may negatively affect populations of tundra swans and other species using tundra wetlands.
Hypotheses and predictions
Our primary hypothesis was that recolonization patterns in trumpeter swans, as well as overall range expansions, may be related to recent increases in the ice-free period observed across the northern hemisphere (Magnuson et al. 2000) and in Alaska in particular (Sagarin & Micheli 2001). Hansen et al. (1971) originally identified a theoretical northern limit to the trumpeter swan breeding range based on the annual number of ice-free days in a given season. Based on this, we hypothesized that habitat occupancy patterns were related to breeding season length (possibly in a quadratic fashion), and a positive relationship between season length, as indexed by the number of days above freezing, and the probability of habitat occupancy would provide evidence that trumpeter swans are colonizing areas in response to breeding season length. To address this season-length hypothesis, we examined habitat occupancy throughout Alaska during 1968-2005 relative to the average number of days above freezing (and days above [freezing.sup.2]) during the previous five years.
Based on previous research utilizing historic Alaskan trumpeter swan survey data (Schmidt et al. 2009a,b), we also hypothesized that occupancy would generally increase due to recolonization of historic breeding areas as abundance increased, and we needed to control for this natural recolonization process to isolate any effects of season length. Although the full extent of the historic range is unknown, we expected that annual trend and/or the latitude of the survey unit would act as surrogates for the recolonization process. We also expected that the recolonization process may be non-linear (latitude2) reflecting the tendency of a recovering population to expand outward from a core range, in this case, middle latitudes.
Material and methods
Statewide surveys for trumpeter swans were first conducted in Alaska in 1968, and in 1975, 1980, 1985,1990,1995,2000 and 2005 surveys covering all known breeding areas were conducted once during August and/or September (Conant et al. 1996, Caithamer 2001, Schmidt et al. 2009a). Individual survey units were 1:63,360 quadrangle maps (units) where trumpeter swans were suspected to occur and were surveyed from a single engine fixed-wing aircraft with the intent of counting every individual within each unit. Survey units were considered to be occupied in a given year if [greater than or equal to] 1 swan was sighted, and additional units were added through time (e.g. N = 176 in 1968 to N=780 in 2005; Fig. 1)in response to apparent range expansion statewide (Schmidt et al. 2009a). We were unable to estimate detection probability from the data available, but previous work concluded that detection probability was high and likely constant for individual swans (Schmidt et al. 2009b), minimizing the potential effects on occupancy modeling for our large sampling units. Tundra swans, which could not be separately identified from trumpeter swans, were present in a limited number of units at the western edge of the observed range of trumpeter swans, but we did not expect the limited scope of this situation and the extensive overlap of the two species within individual survey units occupied by tundra swans to influence the results for trumpeter swans. We based the identification of units to be added to the survey on annual statewide waterfowl surveys during the years between swan surveys, which provide good coverage of waterfowl habitat in the state. Surveys covered the entire known breeding range in Alaska each year (see Fig. 1).
We acquired temperature information for our survey area from weather stations (N = 14) around the state that had collected data for a majority of the period from 1964 to 2005 (see Fig. 1). We then calculated the average number of days with a minimum temperature > 0[degrees]C during the potential breeding season (April-October) of the survey year and the four years prior to the survey (Fig. 2). We assumed that this 5-year average would be representative of the conditions that may have affected occupancy during each corresponding 5-year survey. We assigned units a number of days above freezing for each survey period from the weather station nearest to the center point of each unit, although in some cases, topography (i.e. mountain ranges) was taken into account when selecting the appropriate station. Most groupings of survey units contained a weather station near the center of the group to provide a representative number of days above freezing (see Fig. 1).
We used Bayesian hierarchical occupancy models (Royle & Dorazio 2008) to investigate the effect of the number of days above freezing on the probability of occupancy for each survey unit during 1968-2005. Because our data were complicated by changes in survey extent, we used Markov Chain Monte Carlo (MCMC) methods to fit models describing variation in trumpeter swan occupancy through time relative to year, latitude and days above freezing. We standardized or scaled parameters to have a mean = 0 and SD = 1 to facilitate convergence. These methods allowed us to include all survey units (N = 780) in the analysis each year regardless of the year that they were added to the survey, and they also allowed us to investigate models containing random effects (Schmidt et al. 2009a). In particular, if a unit was not surveyed in a given year, we treated it as a missing value and it was estimated along with all of the other parameters. This permitted unequal numbers of samples from many of the units to be included in the analysis.
When constructing our model set, we assumed that there was a trend in occupancy through time because of the apparent increase in the species' range and the increasing trend in the overall population (Schmidt et al. 2009a). This assumption helped us to control for general population increases and identify any influence of changes in average temperature on occupancy. We suspected that differences in recolonization rates at various latitudes might help explain changes in occupancy probability that were not explained by the season-length covariates alone. We used the Deviance Information Criterion (DIC) to select among competing models containing combinations of these variables (Spiegelhalter et al. 2002) and then added a normally distributed random effect of survey year as an intercept adjustment to the top fixed-effects model to help account for any remaining annual heterogeneity in the data. The resulting model set contained seven models. DIC model selection procedures are analogous to those of Akaike's Information Criterion, and we considered models within 10 units of one another to be competitive.
[FIGURE 1 OMITTED]
We used R 2.4.0 (R Development Core Team 2006) and WinBUGS 1.4.1 (Spiegelhalter et al. 2004) to fit our set of candidate models. We generated three independent Markov chains up to 200,000 iterations in length, and convergence properties were assessed using the Gelman-Rubin diagnostic (Gelman & Rubin 1992, Brooks & Gelman 1998; Rhat < 1.1 = convergence) and a visual inspection of the chains. Initial iterations were discarded as burn-in, and estimates of the posterior distributions of the parameters were calculated from the remaining iterations. The number of iterations necessary for convergence varied depending on model complexity. We estimated goodness-of-fit (P) in a manner similar to that used by Link & Sauer (2002) and Schmidt et al. (2009a) which compares the fit of the model to the fit of a replicate data set generated during each update. By comparing the observed and replicate data, we were able to assess whether the data could have come from the proposed model. Good fit occurs when P = 0.5, and major failures of the model occur when P < 0.01 or P > 0.99 (Gelman et al. 2004:175). Parameter estimates indicate means and 95% Bayesian credible intervals.
The 5-year average number of days above freezing observed at each weather station generally decreased as latitude increased, although variation was high (see Figs. 1 and 2). Model selection results indicated support for a model containing quadratic relationships of both latitude and the number of days above freezing. Models containing both of these variables performed much better (DIC = 4,380) than reduced models containing the latitude (DIC = 4,402) or days above freezing (DIC = 4,460) covariates alone, and the inclusion of a random year effect improved the final model (DIC = 4,367; Table 1). Model fit was determined to be satisfactory based on the Bayesian P-value (P = 0.44 for the simplest model).
Individual logit-scale parameter estimates from the best approximating model in the model set were fairly imprecise (Table 2), indicating uncertainty in their magnitude and direction, likely due to correlation between the latitude and days above freezing variables. Despite uncertainty in individual estimates and annual fluctuations (Fig. 3), predicted probability of occupancy increased from approximately 0.3 to 0.7 at both low and high latitudes and from approximately 0.7 to 0.9 at mid-latitudes over the observed range of the number of days above freezing (Fig. 4). Estimates for the general trend through time tended to be positive (0.238; 95% CI: -0.018--0.476). Estimates from models including only latitude or days above freezing covariates were unambiguous in their directions. Latitude had negative linear (-0.279; 95% CI: -0.361-0.198) and quadratic (-0.393; 95% CI: -0.473-0.314) effects, and days above freezing had a positive linear (11.726; 95% CI: 7.162--15.94) and negative quadratic (-3.767; 95% CI: -5.246-2.170) effects on occupancy on the logit scale in these two simpler models, respectively.
[FIGURE 2 OMITTED]
Strong support for models with a relationship between habitat occupancy and both the average number of days above freezing and recolonization parameters supported our hypotheses that observed patterns of occupancy were related to breeding season length as well as the overall population recovery. These models provide evidence that swan habitat use is directly related to the average annual temperature conditions at a given location. Although we do not see strong trending in the temperature data used in our analysis, the relationships we observed indicate that range expansion has likely occurred in Alaska as a result of climate warming. This conclusion is based on other climate research conducted over a longer time frame indicating that the timing of spring breakup has been advancing (Keyser et al. 2000), and a phase shift in the Pacific Decadal Oscillation has resulted in generally warmer conditions in Alaska since 1976 (Hartmann & Wendler 2005). These changes could explain why more peripheral areas of the current trumpeter swan breeding range were not used until later years, after additional warming had occurred.
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
Support for a non-linear latitude effect is most likely explained by range edge effects. The population was concentrated at middle latitudes within our survey area during the first years of our survey. The negative quadratic effect of latitude could indicate that it took some time for colonization outside of these areas to occur, even though habitat was available. Waterfowl in general, and swans in particular, are highly philopatric (Anderson et al. 1992), so substantial time may be required for swans to colonize newly available habitat, which would lower the average occupancy probability. Juveniles are the most likely dispersers (Lindberg et al. 1998) and usually take 4-5 years to reach breeding age, and this lag could have slowed colonization even after habitats became suitable. Additionally, northern areas might not be as likely to be occupied because of variable weather conditions and the time needed for colonization to occur, and southern areas might also have suboptimal weather conditions during brood rearing (high rainfall levels and cool spring temperatures) making these areas less desirable for breeding. We suspect that weather patterns at range edges (e.g. cool and rainy in coastal areas) could counteract the positive effects of the number of days above freezing and help explain why occupancy is higher in the interior relative to the Arctic and coastal areas. Mountainous terrain also limits the number of suitable breeding wetlands in many areas of southern Alaska, explaining the lower occupancy rates in those areas.
The lack of precision in our estimates suggests uncertainty in the direction and magnitude of the effects. However, this is not surprising considering the partial correlation between latitude and days above freezing. We initially considered the days above freezing variable without latitude in the models, but this simpler model was not supported by the model selection results. In fact, models containing both latitude and days above freezing lowered the DIC by > 20 units over models containing either of these variables individually, indicating that both were important predictors. The correlation between these two variables made it difficult to precisely estimate the effects of each individual covariate, but together they provided a better representation of the data than either covariate alone. Inclusion of only latitude or season length may lead to erroneous or misleading conclusions because both appear to be important predictors. Regardless, together they indicate a strong positive effect of season length on occupancy probability. The support for a random year effect indicates that other sources of annual variation (e.g. precipitation and water levels) may also influence habitat occupancy, and the inclusion of this term accounted for these unmeasured variables.
Nearly 40 years of survey data, combined with the fact that 75-80% of the trumpeter swans in North America breed in Alaska, provided us with a unique opportunity to examine the effects of breeding season length over most of the breeding range of this species. For trumpeter swans, it appears that any negative effects of wetland shrinkage have so far been outweighed by the positive effects of increased available habitat due to extended ice-free periods. It is possible that some wetlands are no longer suitable due to shrinkage, while some deeper types, such as ponds and lakes, may be becoming shallower and therefore more suitable for breeding. If this is the case, we would expect range expansion to continue if climate warming continues. Schmidt et al. (2009b) found some evidence that cygnet production at lower latitudes is slowing, and this may indicate that breeding habitats are becoming saturated or that climate effects such as wetland shrinkage are affecting habitat suitability. For this population in particular, availability of sufficient wintering habitat may ultimately limit population growth (Mitchell 1994).
Based on our findings for trumpeter swans, we would expect continued climate warming to affect populations of boreal forest-nesting species in the future. Habitat requirements for swans are very similar to those of many boreal forest-nesting species, suggesting that the effects of climate warming on wetland-breeding birds could be widespread. Avian distributions are shifting northwards (La Sorte & Thompson 2007) and abundances of some waterfowl species in the Prairie Pothole Region are positively related to precipitation (Forcey et al. 2007), implying that pond loss would have negative population effects (Sorenson et al. 1998). Some species might be expected to experience changes in breeding phenology if seasons lengthen (Winkler et al. 2002, Moller et al. 2010), although these changes may not always have positive impacts. An advancement of spring conditions could cause a mismatch between breeding and food source availability (Both & Visser 2001, Inouye et al. 2000) or substantial changes in habitat (Ward et al. 2005). Either of these could negatively affect populations through reductions in breeding habitat and reproductive success. Although the future impacts of climate change on individual avian species are unknown, long-term monitoring programs like the trumpeter swan survey provide the best opportunity to identify population-level effects and should be encouraged. Because climate warming is expected to be more pronounced in northern regions, concentration of additional monitoring efforts in boreal forest habitats may help to identify species at risk of decline.
Acknowledgements--we thank M. Bertram, C. Handel, B. Mattson, C. McIntyre, J. Schmutz and two anonymous reviewers for helpful comments on a previous version of this manuscript. B. Conant, D. Groves, J. King, E. Mallek and many others with the U.S. Fish and Wildlife Service Migratory Birds Management Division were instrumental in collecting and compiling the survey data. Support for the first author was provided by a State Wildlife Grant from the Alaska Department of Fish and Game, the Department of Biology and Wildlife, the Institute of Arctic Biology of the University of Alaska Fairbanks and Alaska EPSCoR.
Anderson, M.G., Rhymer, J.M. & Rohwer, F.C. 1992: Philopatry, dispersal, and the genetic structure of waterfowl populations.--In: Batt, B.D., Afton, A.D., Anderson, M.G., Ankney, C.D., Johnson, D.H., Kadlec, J.A. & Krapu, G.L. (Eds.); Ecology and management of breeding waterfowl. University of Minnesota Press Minneapolis, Minnesota, USA, pp. 365-395.
Both, C. & Visser, M.E. 2001: Adjustment to climate change is constrained by arrival date in a long-distance migrant bird.--Nature 411: 296-298.
Brooks, S. & Gelman, A. 1998: General methods for monitoring convergence of iterative simulations. Journal of Computational and Graphical Statistics 7: 434-455.
Caithamer, D.F. 2001: Trumpeter swan population status, 2000.--Migratory Bird Management, U.S. Fish and Wildlife Service, Laurel, Maryland, USA, 14 pp.
Conant, B., Hodges, J.I., Groves, D.J., Cain, S.L. & King, J.G. 1996: An atlas of the distribution of trumpeter swans in Alaska and instructions for the use of an archival system.--Migratory Bird Management, U.S. Fish and Wildlife Service, Juneau, Alaska, USA, 31 pp.
Conant, B., Hodges, J.I., Groves, D.J. & King, J.G. 2002: Census of trumpeter swans on Alaskan Nesting Habitats, 1968-2000.--Waterbirds 25: 3-7.
Forcey, G.M., Linz, G.M., Thogmartin, W.E. & Bleier, W.J. 2007: Influence of land use and climate on wetland breeding birds in the Prairie Pothole region of Canada. Canadian Journal of Zoology 85: 421-436.
Gayet, G., Guillemain, M., Mesleard, F., Fritz, H., Vaux, V. & Broyer, J. 2011: Are mute swans (Cygnus olor) really limiting fishpond use by waterbirds in the Dombes, eastern France.--Journal of Ornithology 152: 45-53.
Gelman, A., Carlin, J.B., Stern, H.S. & Rubin, D.B. 2004: Bayesian data analysis.--Chapman and Hall/CRC, Boca Raton, Florida, USA, 668 pp.
Gelman, A. & Rubin, D.B. 1992: Inference from iterative simulation using multiple sequences.--Statistical Science 7: 457-511.
Hansen, H.A., Shepherd, P.E.K., King, J.G. & Troyer, W.A. 1971: The trumpeter swan in Alaska.--Wildlife Monographs 26: 3-83.
Hartmann, B. & Wendler, G. 2005: The significance of the 1976 Pacific climate shift in the climatology of Alaska. Journal of Climate 18: 4824-4839.
Hinzman, L.D., Bettez, N.D., Bolton, W.R., Chapin, F.S., III, Dyurgerov, M.B., Fastie, C.L., Griffith, B., Hollister, R.D., Hope, A., Huntington, H.P., Jensen, A.M., Jia, G.J., Jorgenson, T., Kane, D.L., Klein, D.R., Kofinas, G., Lynch, A.H., Lloyd, A.H., McGuire, A.D., Nelson, F.E., Oechel, W.C., Osterkamp, T.E., Racine, C.H., Romanovsky, V.E., Stone, R.S., Stow, D.A., Sturm, M., Tweedie, C.E., Vourlitis, G.L., Walker, M.D., Walker, D.A., Webber, P.J., Welker, J.M., Winker, K.S. & Yoshikawa, K. 2005: Evidence and implications of recent climate change in northern Alaska and other arctic regions.--Climatic Change 72: 251-298.
Inouye, D.W., Barr, B., Armitage, K.B. & Inouye, B.D. 2000: Climate change is affecting altitudinal migrants and hibernating species.--Proceedings of the National Academy of Sciences of the United States of America 97: 1630-1633.
Keyser, A.R., Kimball, J.S., Nemani, R.R. & Running, S.W. 2000: Simulating the effects of climate change on the carbon balance of North American high-latitude forests.--Global Change Biology 6: 185-195.
Klein, E., Berg, E.E. & Dial, R. 2005: Wetland drying and succession across the Kenai Peninsula lowlands, south-central Alaska.--Canadian Journal of Forest Research 35: 1931-1941.
La Sorte, F.A. & Thompson, F.R., III. 2007: Poleward shifts in winter ranges of North American birds. Ecology 88: 1803-1812.
Lindberg, M.S., Sedinger, J.S., Derksen, D.V. & Rockwell, R.F. 1998: Natal and breeding philopatry in a black brant, Branta bernicula nigricans, metapopulation. Ecology 79: 1893-1904.
Link, W.A. & Sauer, J.R. 2002: A hierarchical analysis of population change with application to cerulean warblers.--Ecology 83: 2832-2840.
Magnuson, J.J., Robertson, D.M., Benson, B.J., Wynne, R.H., Livingstone, D.M., Arai, T., Assel, R.A., Barry, R.G., Card, V., Kuusisto, E., Granin, N.G., Prowse, T.D., Stewart, K.M. & Vuglinski, V.S. 2000: Historical trends in lake and river ice in the Northern Hemisphere. Science 289: 1743-1746.
Mitchell, C.D. 1994: Trumpeter swan (Cygnus buccinator) In: Poole, A. & Gill, F. (Eds.); The birds of North America, No. 105. The Academy of Natural Sciences, Philadelphia and The American Ornithologists' Union, Washington D.C., USA, pp. 1-24.
Moller, A.P., Flensted-Jensen, E., Klarborg, K., Mardal, W. & Nielsen, J.T. 2010: Climate change affects the duration of the reproductive season in birds.--Journal of Animal Ecology 79: 777-784.
Parmesan, C. 2006: Ecological and evolutionary responses to recent climate change--Annual Review of Ecology, Evolution, and Systematics 37: 637-669.
Parmesan, C. 2007: Influences of species, latitudes, and methodologies on estimates of phenological response to global warming.--Global Change Biology 13: 1860-1872.
Parmesan, C. & Yohe, G. 2003: A globally coherent fingerprint of climate change impacts across natural systems.--Nature 421: 37-42.
Petrie, S.A. & Francis, C.M. 2003: Rapid increase in the lower Great Lakes population of feral mute swans: a review and recommendation.--Wildlife Society Bulletin 31: 407-416.
Poysa, H. & Sorjonen, J. 2000: Recolonization of breeding waterfowl communities by the whooper swan: vacant niches available.--Ecography 23: 342-348.
R Development Core Team 2006: R: A language and environment for statistical computing.--R foundation for Statistical Computing, Vienna, Austria. Available at: http://www.r-project.org (Last accessed on 22 August 2011).
Riordan, B., Verbyla, D. & McGuire, A.D. 2006: Shrinking ponds in subarctic Alaska based on 1950-2002 remotely sensed images.--Journal of Geophysical Research 111: G4002, DOI: 10.1029/2005JG000150.
Root, T.L., Price, J.T., Hall, K.R., Schneider, S.H., Rosenweig, C. & Pounds, J.A. 2003: Fingerprints of global warming on wild animals and plants.--Nature 42: 57-60.
Royle, J.A. & Dorazio, R.M. 2008: Hierarchical modeling and inference in ecology: the analysis of data from populations, metapopulations and communities.--Academic Press, London, UK, 444 pp.
Sagarin, R. & Micheli, F. 2001: Climate change in nontraditional data sets.--Science 294: 811.
Schmidt, J.S., Lindberg, M.S., Johnson, D.S., Conant, B. & King, J. 2009a: Evidence of trumpeter swan population growth using Bayesian hierarchical models.--Journal of Wildlife Management 73: 720-727.
Schmidt, J.S., Lindberg, M.S., Johnson, D.S. & Schmutz, J.A. 2009b: Environmental and human influences on trumpeter swan habitat occupancy in Alaska.--Condor 111:266-275.
Seavy, N.E., Dybala, K.E. & Snyder, M.A. 2008: Climate models and ornithology.--Auk 125: 1-10.
Serreze, M.C., Walsh, J.E., Chapin, F.S., III, Osterkamp, T., Dyurgerov, M., Romanovsky, V., Oechel, W.C., Morison, J., Zhang, T. & Barry, R.G. 2000: Observational evidence of recent change in the northern high-latitude environment.--Climate Change 46: 159-207.
Smol, J.P., Wolfe, A.P., Birks, H.J.B., Douglas, M.S.V., Jones, V.J., Korhola, A., Pienitz, R., Ruhland, K., Sorvari, S., Antoniades, D., Brooks, S.J., Fallu, M., Hughes, M., Keatley, B.E., Laing, T.E., Michelutti, N., Nazarova, L., Nyman, M., Paterson, A.M., Perren, B., Quinian, R., Rautio, M., Saulnier-Talbot, E., Siitonen, S., Solovieva, N. & Weckstrom, J. 2005: Climate-driven regime shifts in the biological communities of arctic lakes.--Proceedings of the National Academy of Science of the United States of America 102: 4397-4402.
Sorenson, L.G., Goldberg, R., Root, T.L. & Anderson, M.G. 1998: Potential effects of global warming on waterfowl populations breeding in the northern Great Plains.--Climatic Change 40: 343-369.
Spiegelhalter, D.J., Best, N.G., Carlin, B.P. & van der Linde, A. 2002: Bayesian measures of model complexity and fit.--Journal of the Royal Statistical Society B 64: 583-639.
Spiegelhalter, D.J., Thomas, A., Best, N.G. & Lunn, D. 2004: WinBUGS Version 1.4.1. User manual.--Medical Research Council, Biostatistics Unit, Cambridge, UK, 60 pp.
Stafford, J.M., Wendler, G. & Curtis, J. 2000: Temperature and precipitation of Alaska: 50 year trend analysis. Theoretical and Applied Climatology 67: 33-44.
Ward, D.H., Reed, A., Sedinger, J.S., Black, J.M., Derksen, D.V. & Castelli, P.M. 2005: North American brant: effects of changes in habitat and climate on population dynamics.--Global Change Biology 11: 869-880.
Wendler, G. & Shulski, M. 2008: A century of climate change for Fairbanks, Alaska.--Arctic 62: 295-300.
Winkler, D.W., Dunn, P.O. & McCulloch, C.E. 2002: Predicting the effects of climate change on avian life-history traits.--Proceedings of the National Academy of Science of the United States of America 99:13595-13599.
Joshua H. Schmidt * & Mark S. Lindberg, Department of Biology and Wildlife and Institute of Arctic Biology, University of Alaska Fairbanks, Fairbanks, Alaska, 99775, USA--e-mail addresses: email@example.com (Joshua H. Schmidt); firstname.lastname@example.org (Mark S. Lindberg)
Devin S. Johnson, National Marine Mammal Laboratory, Alaska Fisheries Science Center, NOAA, Seattle, Washington, 98115, USA--e-mail: email@example.com
David L. Verbyla, Department of Forest Sciences, University of Alaska Fairbanks, Fairbanks, Alaska, 99775, USA e-mail: firstname.lastname@example.org
* Present address: Central Alaska Network, National Park Service, 4175 Geist Road, Fairbanks, Alaska, 99709, USA
Corresponding author: Joshua H. Schmidt
Received 14 January 2011, accepted 25 June 2011
Associate Editor: Kjell Einar Erikstad
Table 1. Comparison of model selection results based on Deviance Information Criterion (DIC) for models describing occupancy probabilities for trumpeter swans. Trend = trend across years, lat = latitude, days = 5-year average number of days above freezing, e(year) = random effect of year and [K.sub.est] = estimated number of parameters. DIC = 4,367 for the best model. [K.sub. [DELTA] Model est] DIC trend+lat+[lat.sup.2]+days+[days.sup.2]+e(year) 10 0 trend+lat+[lat.sup.2]+days+[days.sup.2] 6 13 trend+lat+[lat.sup.2] 4 35 trend+days+[days.sup.2]+e(year) 8 79 trend+days+[days.sup.2] 4 93 trend+days 3 124 trend+lat 3 126 Table 2. Logit-scale parameter estimates from the best approximating model describing occupancy probability for trumpeter swans. Trend = trend across years, lat = latitude, days = 5-year average number of days above freezing, [tau.sub.e(year)] is the precision of the random year effect, and LCI and UCI refer to the upper and lower bounds of the 95% credible intervals around the mean parameter estimates. Parameter Mean LCI UCI intercept -0.858 -4.129 1.867 trend 0.238 -0.018 0.476 lat -0.062 -0.175 0.042 lat2 -0.470 -0.569 -0.369 days 1.609 -2.272 6.294 days2 0.012 -1.705 1.425 taue(year) 0.073 0.011 0.359
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Original article|
|Author:||Schmidt, Joshua H.; Lindberg, Mark S.; Johnson, Devin S.; Verbyla, David L.|
|Date:||Dec 1, 2011|
|Previous Article:||A new method to determine bird and bat fatality at wind energy turbines from carcass searches.|
|Next Article:||Emigration rates and population turnover of teal Anas cvecca in two major wetlands of western Europe.|