HR: 08:05h
AN: H31I-01    [Abstracts]
TI: A Brief History of the Collection amd Analysis of Data From Large Fixed-Station Monitoring Networks
AU: * Smith, R A
EM: rsmith1@usgs.gov
AF: U S Geological Survey, MS 413, Sunrise Valley Drive, Reston, VA 20192 United States
AU: Alexander, R B
EM: ralex@usgs.gov
AF: U S Geological Survey, MS 413, Sunrise Valley Drive, Reston, VA 20192 United States
AU: Schwarz, G E
EM: gschwarz@usgs.gov
AF: U S Geological Survey, MS 413, Sunrise Valley Drive, Reston, VA 20192 United States
AB: Since the early 20th century, a broad combination of state and federal fixed-station monitoring programs has served as the principle source of data for water quality assessments in the United States. Due to past limitations of data management and analysis, use of these data in studies of long-term water quality trends at regional and national scales did not begin until 1970, although examinations of water quality time-series from smaller numbers of sites were used in more localized assessments prior to that time. With the emergence in the early 1980s of statistical procedures specifically-designed for trend analysis of water-quality data, and rapid dissemination in the mid-1980s of GIS technology capable of integrating trend observations with spatially- and temporally-referenced explanatory variables, large-scale interpretation of water quality trend patterns met with notable successes. Also, development of more accurate statistical techniques for contaminant load estimation during this period, through correlation with continuous flow records, has proved valuable in a variety of water quality studies, especially those relating to coastal zone management. Beginning in the early 1990s, interpretation of large-scale water quality phenomena took a major step forward by moving from simple correlations between fixed-station water quality and explanatory data, to more complex analyses of the mass balance linking contaminant source-supply rates with in-stream flux measurements based on fixed-site monitoring records. The primary modeling technique used in these studies has been SPARROW (SPAtially-Referenced Regression On Watershed attributes), in which a detailed digital map of stream segments and their contributing watersheds is the spatial reference frame for all variables in the model. In recent calibrations, nutrient models based on hundreds of monitoring stations now include statistically-significant coefficients for more than 20 watershed attributes including hydrologic, climatic, soil, land-cover, and contaminant-source variables. These models have found use in a wide variety of water resource management applications. Model coefficients compare favorably with independently derived values from published process studies. The availability of long-term fixed-station water quality records has been key to the ability to observe a mass balance between contaminant sources and in-stream flux because, at large spatial scales, the balance is not contemporaneous, and temporal averaging or temporal referencing of observational data is required. Ironically, as the ability to extract useful information from fixed station data has increased in recent years, the collection rate of water quality data from large-scale fixed-station water quality monitoring networks has declined.
UR: http://water.usgs.gov/nawqa/sparrow/
DE: 1847 Modeling
DE: 1848 Monitoring networks
DE: 6339 System design
SC: Hydrology [H]
MN: Fall Meeting 2005