Evaluation of remotely sensed prediction and forecast models for Vibrio parahaemolyticus in the Chesapeake Bay

Over the last decade, an increase of gastrointestinal illness due to Vibrio parahaemolyticus in the consumption of raw shellfish has been reported in multiple regions around the United States. Studies mainly attribute this increase to rising sea surface temperatures and prolonged warm seasons in the mid-latitudes. Historically, temperature has been the main environmental determinant used to predict V. parahaemolyticus concentrations in shellfish and surface water. However, studies using in situ sampling campaigns have shown that additional water quality parameters can be useful in predicting the bacterium. While the time and cost of obtaining in situ samples throughout the Chesapeake Bay at regular time intervals can exceed available resources, satellite remote sensing has the potential to provide predictions at higher temporal and spatial resolutions. This study uses satellite ocean color remote sensing and sea surface temperature (SST) from the Moderate Resolution Imaging Spectroradiometer (MODIS) to investigate the utility of remotely sensed information for Vibrio parahaemolyticus predictions in the Chesapeake Bay and whether additional remotely sensed information can improve predictions over conventional SST-based models. We find that the addition of remotely sensed salinity, total suspended solids, and chlorophyll-a generally improves presence and abundance predictions compared to SST-only models. Models using remote sensing reflectances and SST also show potential for V. parahaemolyticus predictions, which could bypass the intermediary step of deriving water quality products from reflectances. Remotely sensed ocean color products and SST from one week prior to in situ V. parahaemolyticus measurements are evaluated and shown to be useful in bacterium predictions, which could provide lead-time for management decisions. The forecast models using ocean color products in addition to SST showed improvement over SST-only forecast models. The results of this study suggest that remote sensing can be a valuable tool to aid in higher resolution V. parahaemolyticus predictions and forecasts in the Chesapeake Bay, particularly when multiple environmental predictors are employed. However, the complexities of using remotely sensed data for ecological modeling applications and evaluating model performance also highlight the need for more research in this area.