Assessing the Nation's Coastal Waters....Better
Citation:
Kiddon, J., L. Harwell, B. Kreakie, Walt Nelson, AND H. Sullivan. Assessing the Nation's Coastal Waters....Better. Coastal & Estuarine Research Federation (CERF) 24th Biennial Conference, Providence, Rhode Island, November 06 - 10, 2017.
Impact/Purpose:
USEPA ORD has expended huge efforts since the 1990s developing methods to effectively assess the nation’s coastal condition. Currently, those methods are employed by USEPA’s Office of Water to assess coastal waters in the future via the National Aquatic Resource Surveys (NARS). Efforts continue to refine the methods of assessing and reporting on the data collected. This presentation addresses two aspects of the refinement exercise, 1) informing the audience of the existence of a dataset of water quality data recently assembled from over 13,000 sites, and 2) highlighting several results derived from analysis of the dataset. The specific results highlighted are: 1) the survey data provide useful characterization of condition at 26 subregions—a refinement over the five regions currently reported upon, and 2) analysis of the data by Random Forest modelling technique (a machine learning approach) reveals robust, independent relationships between chlorophyll and total nitrogen, total phosphorus, and subregion—relationships of concern to estuarine scientists interested in assessing and remediating impaired water quality in estuaries and coastal waters. The impact of these findings are twofold. Generally, the results highlight the utility of long-term surveys to national, state, and regional entities responsible for controlling water quality impairment. And specifically, the results contribute to the understanding of estuarine water quality processes.
Description:
The USEPA has been assessing estuarine and coastal condition in the United States since 1999 via the National Coastal Assessment (NCA) and National Aquatic Resources Surveys (NARS) programs. Approximately 1500 randomly selected coastal sites were surveyed annually during summers from 1999 to 2006, and in 2010 and 2015. Consistent sampling and analysis methods were employed in all studies to evaluate the condition of the water column, sediments, benthic community, and contaminants in fish tissue. Comparable NARS assessments will continue to be conducted at five year intervals. All water quality data described here are publically available. In this presentation, we use the extensive water quality data from more than 13,000 sites to characterize the magnitude and variability of water quality indicators spatially in four nation-scale regions and 26 subregions, and temporally over four time intervals. We also examine relationships between chlorophyll and nutrient concentrations (total and dissolved nitrogen and phosphorus components) by region and subregion. Random Forest modelling, a machine learning approach, reveals particularly robust, independent relationships between chlorophyll and total nitrogen (TN), total phosphorus (TP), and subregion. Finally, we report on ongoing efforts to set TN and TP thresholds that are appropriate for reporting coastal WQ condition at national and large-regional scales. These results highlight the utility of information collected in surveys such as the NCA and NARS programs, which feature uniform assessments over broad spatial extents but at relatively infrequent time intervals.