Characterizing and predicting fine-scale thermal exposure at stopover mudflats used by interior migrating shorebirds

Abstract

Management and conservation efforts for shorebirds are generally conducted with a focus on habitat quality (e.g., forage availability) but often lack detail on the thermal conditions that can directly influence migratory success. Much of this discrepancy is due to a dearth of empirical data or modeled predictions of microclimate exposure along migration routes (i.e., stopover sites). Our objectives were to, 1) quantify thermal environments at interior mid-continental stopover mudflats during autumn migration and 2) compare the ability of modeled outputs (T_model) from mechanistic models and local weather station data (T_weather) to predict field measurements (T_mudflat). T_mudflat was collected at shorebird height (10 cm) on three mudflats routinely used by shorebirds during late summer and autumn 2024 in Kentucky, United States. Temporal variation in mudflat microclimate was substantial across diel and seasonal periods (range; 1.71 – 48.13°C), yet notably, spatial variation was low for any given hour. Consequently, shorebirds experienced limited thermal options on mudflats that routinely exceeded thermal stress thresholds for shorebirds (i.e., 74.25% and 20.61% of temperatures ≥ 25°C and ≥ 35°C). Thermal extremes were most pronounced for diurnal periods (0700 – 1900 hour) in late summer when early long-distance migrants are using mudflat stopover sites. T_weather (F_{1, 7,687} = 40,820, R² = 0.84, p < 0.05; RMSE = 3.41) outperformed T_model (F_{1, 7,687} = 18,020, R² = 0.70, p < 0.05; RMSE = 6.15) at predicting T_mudflat, suggesting that T_weather may offer managers and researchers a relatively simple and often publicly accessible proxy for thermal exposure on shorebirds (albeit not necessarily a substitute for in situ microclimate measurements). Our findings 1) should encourage researchers and managers to add a thermal context to their current assessments of habitat quality for shorebirds, and 2) offer a framework for studying thermal patterns across different habitat types, latitudinal gradients, and sampling methods.