Linking synoptic weather and ocean light attenuation variability in the Gulf of Mexico: constructing a 65-year Kd-Index

Ocean light attenuation (Kd) is an important water quality variable impacting a variety of different benthic species, and by extension, affecting regional oceanic ecosystems. Prior to the satellite era, however, reliable measurements of Kd are non-existent, making it difficult to examine whether climate-related trends are present in the time series of this crucial variable. Utilizing the Spatial Synoptic Classification (SSC) of surface weather types and a reanalysis-based categorization of patterns of mean sea-level pressure (MSLP), this research uses non-linear autoregressive neural network models with external input (NARX) to simulate a daily, 65-year time series of a recently developed Kd-Index (KDI) for nine regions in the Gulf of Mexico. Results indicate modeled-KDI correlations with actual-KDI values – produced from combined MODIS and SeaWiFS satellite observations – approach 0.89 for some regions, while hit rates in modeling actual-KDI spike events (when KDI > 80th percentile) are over 70% for some regions. Long term trends in KDI vary in strength and direction between regions, with the much of the Gulf generally experiencing slightly decreasing KDI, while KDI along parts of the Florida Gulf Coast appears to be increasing since 1948. However, trends in KDI spike events appear to be more consistently increasing throughout much of the study domain. Many of the spike events correspond to historical tropical storm activity, represented by a specific MSLP circulation pattern and a transitional SSC surface weather type. Future work extending from this research will aim to forecast KDI spike events days to weeks in advance, and project the long-term impacts of climate change on this important oceanic variable.