COS 307-5 - Synthetic aperture radar and LiDAR-based topobathymetry improve predictions of an invasive aquatic plant during climatic extremes

Abstract: Invasive aquatic plants, among other pressures, pose a great threat to coastal wetlands. The ability to predict suitable habitat in uninvaded yet vulnerable wetlands remains a critical task to prevent further harm to these ecosystems. Remote sensing (both passive and active) and geospatial data are important tools to help anticipate where new invasions are likely to occur via species distribution models (SDMs). The objective of this study was to assess the efficacy of active remote sensing datasets (synthetic aperture radar and light detection & ranging; SAR & LiDAR) with other geospatial data in predicting the potential distribution of an invasive aquatic plant based on its biophysical habitat requirements and dispersal dynamics. We also took advantage of climatic extremes (water levels) during the study period to investigate how these predictions may change between years. We compiled a time series of 1,100 field records on the occurrence of Hydrocharis morsus-ranae (European frogbit; EFB) with eight remote sensing and geospatial layers as predictors to train random forest models and assess predictive capacity of habitat suitability in coastal wetlands in Michigan, USA. Overall, we found that the model containing all predictors yielded the highest mean overall accuracy of 84.5 % and an area under the receiver operating characteristic curve of 95.9 % for predicting EFB occurrences. The most important variables for predicting EFB occurrence were 1) Distance to the nearest public boat launch 2) fetch, a proxy for wave energy (in m), and 3) mean NDVI during the summer growing season. SAR and LiDAR data proxied key biophysical characteristics of EFB habitat (emergent vegetation and water depth) but in different ways. The probability of suitable habitat varied across the study region and the area of highly suitable habitat (pixels assigned > 60% probability) varied considerably between the years of highest and lowest water levels. Despite variations, the influence of EFB’s dispersal dynamics (boat launches) clearly outweighed other biophysical components for establishment during this early invasion stage. Overall, our findings suggest that a variety of remote sensing and geospatial data are required to accurately predict the distribution of this species with implications for early detection and rapid response management.