Both real‐time and long‐term environmental data perform well in predicting shorebird distributions in managed habitat

Abstract

Highly mobile species, such as migratory birds, respond to seasonal and interannual variability in resource availability by moving to better habitats. Despite the recognized importance of resource thresholds, species‐distribution models typically rely on long‐term average habitat conditions, mostly because large‐extent, temporally resolved, environmental data are difficult to obtain. Recent advances in remote sensing make it possible to incorporate more frequent measurements of changing landscapes; however, there is often a cost in terms of model building and processing and the added value of such efforts is unknown. Our study tests whether incorporating real‐time environmental data increases the predictive ability of distribution models, relative to using long‐term average data. We developed and compared distribution models for shorebirds in California's Central Valley based on high temporal resolution (every 16 days), and 17‐year long‐term average surface water data. Using abundance‐weighted boosted regression trees, we modeled monthly shorebird occurrence as a function of surface water availability, crop type, wetland type, road density, temperature, and bird data source. Although modeling with both real‐time and long‐term average data provided good fit to withheld validation data (the area under the receiver operating characteristic curve, or AUC, averaged between 0.79 and 0.89 for all taxa), there were small differences in model performance. The best models incorporated long‐term average conditions and spatial pattern information for real‐time flooding (e.g., perimeter‐area ratio of real‐time water bodies). There was not a substantial difference in the performance of real‐time and long‐term average data models within time periods when real‐time surface water differed substantially from the long‐term average (specifically during drought years 2013–2016) and in intermittently flooded months or locations. Spatial predictions resulting from the models differed most in the southern region of the study area where there is lower water availability, fewer birds, and lower sampling density. Prediction uncertainty in the southern region of the study area highlights the need for increased sampling in this area. Because both sets of data performed similarly, the choice of which data to use may depend on the management context. Real‐time data may ultimately be best for guiding dynamic, adaptive conservation actions, whereas models based on long‐term averages may be more helpful for guiding permanent wetland protection and restoration.

Changes in habitat suitability for wintering dabbling ducks during dry conditions in the Central Valley of California

“…Wetlands are typically managed for greater depth to support a broader suite of waterfowl, sometimes to the detriment of shorebirds. Analyses using similar models for shorebirds found that agricultural habitat was more important, especially rice habitat for dunlin (Conlisk, Golet, et al, 2022).…”

“…The 2:1 ratio was based on the average increase in TSS when predicting to the northern region when the North versus South models were used (and when predicting for the south when the South vs. North models were used). Shorebird ensemble models built in this way were superior to models that included the entire study region (Conlisk, Golet, et al, 2022).…”

“…Finally, we included a set of models that did not match duck observations to the specific dates of Landsat tiles but instead used the long‐term average of water conditions to generate duck habitat predictions (mimicking the approach in Conlisk, Golet, et al, 2022). For long‐term average covariate models, we adopted the same procedure of overlaying land cover and flooding, but flooding was derived from the long‐term (2001–2017) monthly average of surface water.…”

“…With improved, readily available, frequently collected, remotely sensed data (Pettorelli et al, 2014), we can create models to evaluate species response to real‐time (or near real‐time) environmental conditions. Recent studies have built models that use Landsat‐derived surface water availability to describe shorebird occupancy during two‐week windows surrounding the acquisition of a Landsat satellite image (Conlisk, Golet, et al, 2022). These models have proven to be highly accurate for describing habitat seasonality in shorebirds in the Central Valley of California.…”

Data integration reveals dynamic and systematic patterns of breeding habitat use by a threatened shorebird

“…These predictions can be valuable for assessing habitat suitability or informing the conservation of rare species at locations where occupancy information is uncertain 2 . One approach commonly used in conservation planning is to produce static realizations of species' distributions by pooling spatial data across time, because there may not be enough occurrence data to build models of species' responses to fluctuating conditions 3 . However, ecological processes underlying species' distributions are complex, and averaging temporal variation in habitat suitability or occurrences of highly mobile species may lead to a loss in precision needed for adaptive conservation planning 2,4 .…”