Comparative spatial modeling of Bretziella fagacearum distribution and contributing factors in northern Wisconsin, U.S.

Abstract

Bretziella fagacearum (Bretz) Z.W. deBeer, Marinc., T.A. Duong, & M.J. Wingf., the ascomycete fungus causing the “oak wilt” disease, is considered a virulent threat to North American oak forests, but the influence of the physical environment on this pathosystem remains unclear, particularly at the forest scale. This study explored the influence of terrain and soil factors on B. fagacearum infections, applying discrete and continuous spatial models to investigate the question: besides proximity to other infections, which environmental factors influenced B. fagacearum incidence? Locations of infections were recorded from 586 confirmed B. fagacearum sites, identified from 2004 through 2021 in a 76 km2 area of deep, sandy glacial outwash in Chequamegon-Nicolet National Forest, northern Wisconsin. Public datasets derived from remote sensing were incorporated as covariates, describing terrain elevation (USGS 10-m DEM), soil physical and chemical properties (POLARIS), and forest composition (WiscLand2). Spatial models included Generalized Additive Models (GAM) and Neyman-Scott Cluster Process Models (CPM). Results indicated that spatial dependence and the distribution of oak forests were the most important drivers of B. fagacearum distribution in this area, with more minor influence from elevation, hill shade, and drainage patterns. Comparison between modeling approaches indicated that—at this scale and in this area—the most accurate models were those which included host distribution, spatial dependence, as well as quantitative terrain and soil descriptions. However, a close approximation could be attained using nonlinear models (GAMs) which incorporated only host distribution and spatial dependence.