Dynamic responses of ground beetles to climate change and habitat characteristics

Image credit: agu 2020

Abstract

Recent studies report declines in the abundances of terrestrial insects with active debate on the environmental drivers of change. Among the largest, most diverse, and economically important groups is the ground beetles in family Carabidae. Attributing cause to declines is challenging, because environmental variables could have effects that are nonlinear and interact with one another. Static regression models for distribution or abundance data, including species distribution models, cannot identify cause, because they do not track the responses (change in abundance) that follow changes in environment and the other species that are also responding to the environment. Time-series data for communities (rather than individual species) are needed, but they cannot be analyzed with traditional models that assume simple autoregressive structures and Gaussian data. We developed a dynamic, biophysical framework, general joint attribute modeling (gjam) for dynamic data (gjamTime), to synthesize ground beetle observations with airborne hyperspectral remote sensing from the National Ecology Observation Network (NEON). The fully probabilistic framework includes uncertainties for observations, model, and parameters. Environmental predictors also include NEON hyperspectral images, which provided unique advantages in characterizing habitat characteristics at a 1-meter spatial resolution. Analysis includes three elements. First, using static gjam we determined that high canopy nitrogen, high lignin, high vegetation index, and low canopy water content are associated with high abundances of many carabid species. Second, we fitted annual time series with climate anomalies using gjamTime. For most species we observed negative responses to both winter minimum temperature and summer moisture deficit, indicating that warm winters and severe droughts may contribute to Carabid declines. A third component, now underway, combines these results with changes in the plant community. The modeling framework holds potentials to understand dynamics of ground beetles involving climate and habitat change.

Date
Dec 11, 2020 —
Location
Online