Satellite and other remote sensing data have a huge potential to aid in monitoring biodiversity and its loss. However, it is increasingly difficult to navigate growing streams of data and optimize selection of most relevant habitat variables. Our research provides insights for this challenge based on a project that links arthropod biodiversity in Hawaiian forests to landscape-scale habitat properties detectable with remote sensing. We analyzed diverse types of imaging products to determine 1) which variables are least correlated and most helpful in understanding habitat heterogeneity across sites, and 2) to what extent different levels of detail (spatial resolution) affects our habitat measurements for similar types of products. Our findings will help guide how we study habitats, choose the best indicators for ecosystem health, and understand the cause-and-effect relationships between habitats and biodiversity.