Vegetation patterns at the landscape scale are shaped by myriad processes and historical events, and understanding the relative importance of these processes aids in predicting current and future plant distributions. To quantify the influence of different environmental and anthropogenic patterns on observed vegetation patterns, we used simultaneous autoregressive modeling to analyze data collected by the Carnegie Airborne Observatory over Santa Cruz Island (SCI; California, USA). SCI is a large continental island, and its limited suite of species and well documented land use history allowed us to consider many potential determinants of vegetation patterns, such as topography, substrate, and historical grazing intensity. As a metric of vegetation heterogeneity, we used the normalized difference vegetation index (NDVI) stratified into three vegetation height classes using LiDAR (short, medium, and tall). In the SAR models topography and substrate type were important controls, together explaining 8-15 % of the total variation in NDVI, but historical grazing and spatial autocorrelation were also key components of the models, together explaining 17-21 % of the variation in NDVI. Optimal spatial autocorrelation distances in the short and medium height vegetation models (600-700 m) were similar to the home range sizes of two crucial seed dispersers on the island-the island fox (Urocyon littoralis santacruzae) and the island scrub-jay (Aphelocoma insularis)-suggesting that these animals may be important drivers of the island's vegetation patterns. This study highlights the importance of dynamic processes like dispersal limitation and disturbance history in determining presentday vegetation patterns.