Integrated modeling predicts shifts in waterbird population dynamics under climate change

Abstract

Climate change has been identified as one of the most important drivers of wildlife population dynamics. The in‐depth knowledge of the complex relationships between climate and population sizes through density dependent demographic processes is important for understanding and predicting population shifts under climate change, which requires integrated population models (IPMs) that unify the analyses of demography and abundance data. In this study we developed an IPM based on Gaussian approximation to dynamic N‐mixture models for large scale population data. We then analyzed four decades (1972–2013) of mallard Anas platyrhynchos breeding population survey, band‐recovery and climate data covering a large spatial extent from North American prairies through boreal habitat to Alaska. We aimed to test the hypothesis that climate change will cause shifts in population dynamics if climatic effects on demographic parameters that have substantial contribution to population growth vary spatially. More specifically, we examined the spatial variation of climatic effects on density dependent population demography, identified the key demographic parameters that are influential to population growth, and forecasted population responses to climate change. Our results revealed that recruitment, which explained more variance of population growth than survival, was sensitive to the temporal variation of precipitation in the southern portion of the study area but not in the north. Survival, by contrast, was insensitive to climatic variation. We then forecasted a decrease in mallard breeding population density in the south and an increase in the northwestern portion of the study area, indicating potential shifts in population dynamics under future climate change. Our results implied that different strategies need to be considered across regions to conserve waterfowl populations in the face of climate change. Our modelling approach can be adapted for other species and thus has wide application to understanding and predicting population dynamics in the presence of global change.

Microhabitat nest‐site selection by ducks in the boreal forest

“…All of these variables are different in the boreal forest compared to the prairies (i.e., Prairie Pothole Region and the Canadian Prairie‐Parklands), where most studies of North American ducks have been conducted (Keith , Greenwood et al , Drever et al , Baldassarre ). The importance of the boreal forest for breeding ducks is likely to increase under most climate‐change scenarios that predict distributional shifts north for many species of birds with increasingly warm and dry prairie climates (Drever et al , Cumming et al , Stralberg et al , Zhao et al ). As such, additional studies are needed to improve our understanding of the nesting ecology of boreal ducks.…”

Integrating broad‐scale data to assess demographic and climatic contributions to population change in a declining songbird

“…For the initial time step ( t = 1992), numbers of survivors and recruits were determined mainly by the count data and (weakly informative; see below) prior distributions. We modeled s m,t and g m,t at subsequent time steps based on Gaussian approximations of binomial and Poisson distributions, respectively (Zhao et al, ):…”

“…Integrated population models (IPMs) provide a formal framework for jointly modeling independent count and CMR data (Besbeas, Freeman, Morgan, & Catchpole, ; Hostetler, Sillett, & Marra, ; Schaub & Abadi, ) along with climate predictors of demographic rates (Zipkin & Saunders, ). With one recent exception (Zhao, Boomer, & Royle, ), IPMs that have included climate covariates of demographic rates have been limited to population studies across relatively small spatial extents (Woodworth, Wheelwright, Newman, Schaub, & Norris, ) or have modeled multiple local populations independently (Weegman, Arnold, Dawson, Winkler, & Clark, ). Data from long‐running national and continental scale avian monitoring programs (Dunn et al, ; Gregory et al, ; Robinson, Julliard, & Saracco, ; Sauer & Link, ) present a unique opportunity for extending IPMs to broad‐scale applications.…”

“…However, development and application of IPMs for broad‐scale bird‐monitoring data are challenging because of a variety of sampling issues, including mismatches in sizes and spatio‐temporal distribution of sampling areas between monitoring data sets and properly accounting for observation error associated with multi‐site, multi‐observer studies (Ahrestani, Saracco, Sauer, Royle, & Pardieck, ; Robinson, Morrison, & Baillie, ). Additionally, IPMs developed for local scale studies based on binomial and Poisson population processes may not be appropriate at regional scales where population responses represent averages of local studies (Zhao et al, ).…”

Monitoring boreal avian populations: how can we estimate trends and trajectories from noisy data?

“…By drawing strength from multiple sources of information about population status, IPMs provide a potentially powerful framework for separating the signal from the noise in fluctuations of boreal bird populations (Besbeas et al 2002, Schaub and Abadi 2011. IPMs are extremely flexible and can accommodate missing data, hierarchical effects, and spatially and temporally structured data with relative ease (Schaub et al 2007, Chandler and Clark 2014, Ahrestani et al 2017, Zhao et al 2019. When demographic data are available for linked breeding and winter populations, full life-cycle IPMs can help link population-level processes operating across the annual cycle (Rushing et al 2017).…”