| Abstract: | This article evaluates the effect of moving window (MW) size on observed fragmentation spatial patterns and proposes a method to identify an effective MW size using Simpson's diversity index. To test the robustness of the proposed method, we demonstrate its use in six cities in the Phoenix metropolitan area that have substantial variation in land composition and configuration. Next we explore the effects of gradient observation scale and the role of scale in removing noise. We compare and discuss two popular approaches to measuring urban-to-rural fragmentation gradients—concentric ring- and transect-based approaches—highlighting the usefulness of each approach in an extensive and rapidly urbanizing region. This study provides a new method for selecting window size, offers insights on scale effects, and provides guidance on gradient scale selection to achieve the best representation of land fragmentation patterns for urban analysis. |
