Modeling Slough Crayfish Populations in Response to Hydrologic Variability

Abstract

Understanding and predicting animal population dynamics as a function of hydrologic variation is critical for the management of freshwater ecosystems. Crayfish are important fauna supported by the hydro-dynamic freshwater wetlands of the Everglades. We modeled the complex relationships between slough crayfish (Procambarus fallax) population densities and hydrologic conditions using a spatially and temporally extensive 21-year dataset. We applied linear mixed–effect models, a classification and regression tree (CART), and random forest (RF) algorithms to develop predictions and eco-hydrologic interpretations. The random forest model demonstrated the greatest predictive ability (R2 = 0.56) followed by linear mixed–effect models (R2 = 0.45) and the regression tree (R2 = 0.29). Primary effects of hydrologic terms were similar among models, but the RF model identified important nonlinear and threshold relationships. Lower average depths (appr. 30–60 cm) over the year prior to the sample, in conjunction with relatively long periods of inundation, and moderate recent depths, were associated with greater crayfish densities. The three methods revealed consistent understanding of crayfish eco-hydrologic relations and provide insight for natural resource management.