Building an Automated Integrated Observing System to Detect Sea Surface Temperature Anomaly Events in the Florida Keys

Satellite-derived sea surface temperature (SST) images have had limited applications in near-shore and coastal environments due to inadequate spatial resolution, incorrect geo-correction, or cloud contamination. We have developed a practical approach to remove these errors using Advanced Very High Resolution Radiometer (AVHRR) and MODerate-resolution Imaging Spectro radiometer (MODIS) 1-km resolution data. The objective was to improve the accuracy of SST anomaly estimates in the Florida Keys and to provide the best quality (in particular, high temporal and spatial resolutions) SST data products for this region. After manual navigation of over 47000 AVHRR images (1993-2005), we implemented a cloud-filtering technique that differs from previously published image processing methods. The filter used a 12-year climatology and 3-day running SST statistics to flag cloud-contaminated pixels. Comparison with concurrent (+/- 0.5 h) data from the SEAKEYS in situ stations in the Florida Keys showed near-zero bias errors (<0.05 degrees C) in the weekly anomaly for SST anomalies between -3 degrees C and 3 degrees C, with standard deviations