LB 9-80 - Environmental & biotic processes maintained historically sharp vegetation boundaries in the Midwestern United States

As ecosystems become increasingly destabilized by anthropogenic activities such as climate change, urbanization, and deforestation, the need for informed predictions of vegetation’s response to environmental change is becoming critical. An understanding of the historical relationship between vegetation and the environment is essential to informing these predictions, but the ecological mechanisms of the vegetation-environment relationship remain contentious. While some assume a “one to one” interaction existed between vegetation and environmental conditions, such as climate, soil characteristics, and topography (i.e., one set of environmental conditions corresponds to one vegetation community), others argue that a complex network of both environmental and biotic processes shaped the vegetation distribution (i.e., one set of environmental conditions corresponds to multiple vegetation states). The goals of this study are to estimate the degree to which (1) environmental drivers and (2) biotic interactions explain the presence of tree taxa and ecosystems in reconstructed historical vegetation.To address these goals, a Generalized Joint Attribute Model (GJAM) was used to characterize the relationship between 12 environmental drivers and the presence of 15 taxa, reconstructed from the 19th Century Public Land Survey notes of Illinois and Indiana. We show that across more than 78,000 observations, environmental drivers explained 30.3% of the variation in historical ecosystem presence. Of the environmental drivers, taxa were more sensitive to climate drivers, such as temperature (F [95% credible interval] = 1.813 [1.576 to 2.054]) and precipitation (F = 1.712 [1.573 to 1.859]), and less sensitive to soil characteristics and topography, such as cation exchange capacity (F = 1.158 [1.100 to 1.214]) and SAGA Wetness Index (F = 0.198 [0.170 to 0.233]). Beyond that explained by the environment, negative biotic interactions also worked to maintain distinct prairie, savanna, and forest ecosystems. There was a strong negative residual correlation between No Tree and two savanna taxa Oak (R [95% credible interval] = -0.663 [-0.825 to -0.571]) and Hickory (R = -0.207 [-0.263 to -0.161]). Similarly, when only ecosystem presence/absence was modeled, prairie and savanna ecosystems exhibited strong negative interactions (R = -0.670 [-0.771 to -0.618]). As a result, this study suggests that both environmental conditions and biotic interactions should be considered when predicting vegetation’s response to changing conditions and developing future land management strategies.