GC51B-0407:
Development of a Water Clarity Index for the Southeastern U.S. As a Climate Indicator

Friday, 19 December 2014
Scott C Sheridan1, Chuanmin Hu2, Cameron C Lee1, Brian Barnes2, Doug Pirhalla3, Varis Ransibrahmanakul3 and Karsten A Shein4, (1)Kent State University Kent Campus, Kent, OH, United States, (2)University of South Florida, St. Petersburg, FL, United States, (3)National Ocean Service, Silver Spring, MD, United States, (4)NOAA NCDC, Asheville, NC, United States
Abstract:
A common index of water quality is water clarity, which can be estimated by measuring the diffuse attenuation coefficient for downwelling irradiance (Kd). Kd estimates the availability of light to marine organisms at various depths. Marine habitats, including such species as coral and seagrass, can be negatively affected by extreme episodes of sediment suspension, where water clarity is reduced and little light penetrates. Evidence of increased stress on coastal ecosystems exists, partially due to climate change, yet a systematic analysis of extreme events and trends is difficult due to limited data.

To address this concern, we have developed as a potential climate indicator a Kd-Index for nine regions along the US coast of the Gulf of Mexico, in which Kd values have been standardized over time and space to allow for a more holistic assessment of climate drivers and their trends. Variability in the Kd-Index is then assessed with regard to occurrences of surface weather types (using the Spatial Synoptic Classification), a synoptic climatology of mean sea-level-pressure patterns across the region, along with heavy precipitation events.

Kd can be estimated from MODIS and SeaWiFS observations from 1997 to date; an earlier period of satellite observations from 1978-86 is also available. A non-linear autoregressive neural network model with external input (NARX) is used to develop the historical relationship between Kd-Index and atmospheric conditions, and then this model is used to simulate a full time series from 1948 to 2013. The modeled data set is strongly correlated with observations, with correlations above 0.8 for many regions. Hit rates of extreme Kd-Index values – those which would most likely be associated with a negative environmental impact – exceed 70% in some regions. Across the full data set, long term trends vary slightly across regions but are generally small. Trends in extreme events appear to be more consistently increasing across the domain.