Soils and forage quality as predictors of white-tailed deer Odocoileus virginianus morphometrics

摘要

Biologists tasked with managing cervids could benefit from models predicting physical characteristics. Differences in white-tailed deer Odocoileus virginianus morphometrics across soil resource areas in Mississippi, USA, provide opportunity to test the predictive capacity of soil chemical and forage quality variables. Using principal components analysis (PCA), we modeled variation in body mass and antler score of >= 1.5-year-old male deer against seven soil chemical variables and 12 forage quality variables to elucidate potential nutritional factors corresponding with physical variation among 21 deer populations. We developed separate sets of models at the levels of state and soil resource area (Delta, Thin Loess and Lower Coastal Plain) and compared statewide models with general linear models (GLM) that related deer morphometrics to nominal classification variables representing the three soil resource areas. PCA distinguished a gradient of increasing soil fertility and forage quality that explained 58% of body mass and 52% of antler score variation statewide. However, the GLM using soil resource area as the explanatory variable explained 78 and 61%, respectively, indicating that management models should use soil resource area to designate areas with broadly similar nutritional planes. Within soil resource areas, the region with the greatest soil fertility and forage quality (Delta) did not model successfully for either body mass or antler score. The Thin Loess was successfully modeled for antler score, but only the Lower Coastal Plain, which had the lowest level of soil and forage quality, was successfully modeled for both morphometric variables. The Delta may have represented an area with habitat quality sufficiently high to render small variations non-influential. In contrast, the generally poorer soil quality of Thin Loess and Lower Coastal Plain soils and forages may act as a limiting factor on physical expression, which allowed some response to relatively small fluctuations in range quality. The potential utility of soil and forage metrics within soil resource areas to estimate deer physical qualities appears to be primarily for fine-tuning estimates largely determined by factors such as density and land use.