Stated preference analyses commonly impose strong and unrealistic assumptions in response to spatial welfare heterogeneity. These include spatial homogeneity or continuous distance decay. Despite their ubiquity in the valuation literature, global assumptions such as these have been increasingly abandoned by non-economics disciplines in favor of approaches that allow for spatial patchiness. This paper develops parallel methods to evaluate local patchiness and hot spots in stated preference welfare estimates, characterizing relevant patterns overlooked by traditional approaches. The analysis draws from a choice experiment addressing river restoration. Results demonstrate shortcomings in standard treatments of spatial heterogeneity and insights available through alternative methods.