Human activity continues to drive global biodiversity change and loss. Comprehensive biodiversity monitoring is critical to evaluating and informing conservation policy and management, and diverse data sources that can enhance the spatial and temporal coverage of conventional field monitoring are needed. In this dissertation, I explore the potential contributions of broad-scale, publicly accessible geospatial datasets to biodiversity monitoring in the Hawaiian Islands. This region supports exceptional levels of endemism but continues to experience significant native habitat loss. Non-native plant species outnumber the native flora, and ecosystems like tropical dry forest are unlikely to recover without active restoration.

Though citizen science data are abundant, they are often collected opportunistically, and potential biases must be understood before utilizing observations. I examined citizen science participation patterns from 2008-2021 using over 93,000 species observations from iNaturalist. The majority of observations were made by visitors to Hawaii, who were more likely to access remote locations and make research-grade observations. However, during the COVID-19 pandemic, visitor activity declined significantly, demonstrating the importance of sustained, local participation for consistent monitoring. I then evaluated the utility of iNaturalist in invasive plant monitoring, and found that non-native species represented a high proportion of iNaturalist plant observations. Comparison of iNaturalist and professional agency observations for four example invasive species showed that iNaturalist data were biased toward accessible, disturbed sites, and professional data toward less accessible, native-dominated sites. Habitat suitability models built with the two datasets often produced distinct results, whereas combining the data provided a more comprehensive estimate of invasive species habitat. Finally, I used a Landsat Normalized Difference Vegetation Index (NDVI, a proxy of vegetation productivity) time series to evaluate changes in dry forest from 1999-2022. Despite regional declines in rainfall, native and restored dry forest NDVI increased during this period. Previous, coarser-scale studies have reported negative NDVI trends in the region, but Landsat resolution or finer is better suited to capturing conditions in fragmented dry forests and monitoring progress at restoration sites. Together, these studies illustrate the value of utilizing and integrating multiple, complementary data sources to improve the breadth and continuity of biodiversity monitoring.