NERRS banner

System-wide Monitoring Program

Synthesis of the Water Quality Data from 1995 to 2000
General Introduction

The System-wide Monitoring Program (SWMP) of the National Estuarine Research Reserve (NERR) system collects water quality observations at 30-minute intervals at numerous locations throughout the United States. Initiated at two sites in North Carolina in 1994, the NERR SWMP has grown substantially, with 51 active sites at 22 Reserves nationwide at the end of 2000. Given the diverse biogeography and physiography of these sites, the NERR SWMP provides an ideal platform for attempting to characterize the short-term variability and long-term changes in estuarine systems, as well as to document the response of these systems to episodic events such as tropical systems and drought conditions.

A recent report characterized the short-term variability in dissolved oxygen, water temperature, salinity, and depth at 44 sites from 22 Reserves between 1996-1998 (Wenner et al. 2001). The report provided a thorough descriptive examination of water quality observations at each site, net ecosystem metabolic analyses based on dissolved oxygen observations, and characterization of the influence of natural cycles (light-dark, tidal) on water quality. Furthermore, the report provided a critical first analysis of the link between watershed and water body characteristics and subsequent water quality.

In an effort to improve certain aspects of the analytical methods used previously, to increase the length of the time series for trends assessment, and to ensure that all archived data collected by the NERR SWMP since 1995 had been analyzed, a second synthesis project was initiated in fall 2001. Rather than focus on detailed examination of data for each site participating in the NERR SWMP, the intent of the current synthesis project was to examine data at the Reserve and System-level, with particular emphasis on assessing trends in water quality between 1995-2000. To this end, quantification of qualitative attribute data collected in the previous synthesis, and quantification of new attribute data not available in the previous synthesis were primary objectives of the current study. Metabolic analyses for all sites and years were also included in the current study. In addition, several modifications were made to the statistical analyses, the most notably of which was the use of General Additive Models (GAM), rather than Harmonic Models, to characterize the periodic and cyclical nature of water quality observations for individual site deployments. The impetus of the change was to use models that allow the patterns in the data to emerge, rather than forcing the data to fit specific patterns. Lastly, investigation of the effects of tropical systems, discussed briefly in the previous study, was included in this report.

The current report consists of six independent, but inter-related, chapters. The first chapter characterizes seasonal and regional trends in sampling between 1995-2000. A critical first step in our analysis was to evaluate and compare data sets to determine what, if any, sampling bias existed. Disproportionate sampling among sites, regions, seasons and years can complicate interpretation of observed differences in parameters at these levels. This chapter also examines the effects of deployment duration on dissolved oxygen readings. Drift in dissolved oxygen (% sat) due to increased fouling of DO probes with increased deployment duration represents a potential bias in determination of the frequency and duration of hypoxic events. In order to determine if potential drift occurred, and to assess trends in hypoxia and supersaturation at varying time intervals over the entire data set, we re-evaluated potential drift at varying deployment time intervals. The second chapter characterizes seasonal and inter-annual trends for specific water quality variables. This chapter also includes statistical tests for deployment-level differences in DO extremes, salinity, pH, and turbidity among Reserves with respect to site, season, and year.

Chapters 3 through 6 represent applied approaches to characterizing NERR SWMP water quality data. The intent of these chapters is to attempt to document the dominant processes responsible for influencing distributions of water quality parameters and to characterize the response of water quality parameters to both natural cycles and episodic events. Chapter 3 examines the relationship between the watershed, land-use practices within the watershed, and climate in influencing water chemistry at NERR SWMP sites. Several multivariate analyses were used to classify sites and determine general relationships regarding land-water interactions. Trends in production, respiration, and net ecosystem metabolism at 42 sites are provided in chapter 4. Improved techniques for modeling short (diel, tidal) and intermediate (seasonal, inter-annual) periodicity in water quality variables are presented in chapter 5. Chapter 6 examines changes in water quality at NERR SWMP sites associated with the passage of 21 tropical systems by one or more NERRs between 1995-2000.

Data analyzed in this report were collected prior to the publication of the previous synthesis work (Wenner et al. 2001); thus, sampling and design recommendations provided in the previous synthesis could not incorporated into the System Wide Monitoring Program prior to our current analysis. As such, those recommendations are not reiterated here; however, three additional recommendations are suggested as a result of the current synthesis. First, although improved ancillary attribute data were collected for this report, several desired attribute data (SAV/Shellfish Bed cover, freshwater flow, tidal excursion) were unobtainable, were incomplete, or were collected too far away from NERR sites to be meaningful. Furthermore, attribute data that we were successful in obtaining were collected from multiple sources and took over six months to compile for the 51 sites at 22 NERRs. As a result of our experiences, we recommend that each Reserve attempt to maintain current and relevant attribute data for their sites (see Chapter 3), as well as attempt to locate the desired attribute data that could not be included in this report. Second, nutrient data sampling was too sporadic and incomplete to provide meaningful interpretation of data that could be related to metabolic processes (see Chapter 4). We therefore recommend modification of sampling protocols to include increased sampling at more pertinent locations (i.e., along a salinity gradient that encompasses the sources of nutrients and the dilution of nutrients). Finally, because the objective of this synthesis was to provide a more thorough “big picture” analysis, the details of the site-level processes were not included. Specifically, over 33,000 graphs for deployment-level periodicity analyses were created for the 1995-2000 NERR SWMP dataset, but only examples of these analyses were included in this report. It is our sincere hope that the efforts to model periodicity at each of these sites will be further enhanced by knowledgeable interpretation by experts at each Reserve.