We combined observations of bobcats (Lynx rufus) from bowhunters with remotely‐sensed data to build models that describe habitat and relative abundance of this species in the agricultural landscape of Iowa, USA. We calculated landscape composition and configuration from publicly available land cover, census, road, hydrologic, and elevation data. We used multiple regression models to examine county‐level associations between several explanatory variables and relative abundance of bobcats reported by surveyed bowhunters in each county. The most influential explanatory variables in the models were metrics associated with the presence of grassland, including Conservation Reserve, along with configuration of this perennial habitat with forests, although human population density and abundance of eastern cottontails (Sylvilagus floridanus) also correlated with abundance of bobcats. Validation of predictions against 3 years of independent data provided confidence in the models, with 66% of predictions within 1 bobcat/1,000 hunter‐hours and 95% within 5 bobcats/1,000 hunter‐hours of observed values. Once we accounted for landscape differences, no residual spatial trend was evident, despite relatively recent bobcat recolonization of Iowa. Models suggested that future range expansion of the bobcat population may be possible in some northern Iowa counties where habitat composition is similar to counties in southern Iowa where bobcats are abundant. Results from the county‐level model have been useful to the Iowa Department of Natural Resources in evaluating the expansion of this once rare species and for delineating harvest opportunities. © 2011 The Wildlife Society.

Bobcats (Lynx rufus), once common in the prairie-woodland mosaic of the Midwest, were largely extirpated from the Corn Belt region by 1900. In the 1990's, sightings of bobcats in Iowa began to increase, and they are now abundant in southern Iowa. With the dramatic expansion of rowcrop agriculture resulting in loss of habitat, wildlife managers do not know whether bobcats will again be widespread throughout Iowa. In order to predict where bobcats will eventually repopulate the state, I attempted to identify important variables that correlate with current bobcat distribution and to build models that predict the relative abundance and occurrence of bobcats. I used the programs ArcGIS and FRAGSTATS to calculate landscape composition and configuration from publicly available sources including landcover, census, road, hydrologic, and elevation data. I constructed classification and regression tree (CART) models to identify important variables for predicting bobcat distribution in Iowa. I built linear regression models of bobcat relative abundance at the county resolution. Models were based on bobcat sightings from the Iowa Department of Natural Resources Bowhunter Observation Survey. I also built logistic regression models of bobcat occurrence at the finer sub-watershed resolution. Sub-watersheds were classified by presence or absence based on locations from reported bobcat sightings, live captures, and carcass recoveries. In all models, both probability of bobcat presence and bobcat relative abundance were consistently influenced by the quantity and configuration of perennial grassland across Iowa. None of the models revealed favorable habitat outside of areas known to be occupied by bobcats in Iowa, suggesting that bobcats are already occupying areas of favorable habitat and will not substantially disperse to other parts of the state. ix These results have practical implications for wildlife conservationists regarding expected bobcat habitat use and distribution as the species becomes more abundant in the agricultural landscape of the Midwest.