You are here:
Spatial Patterns in Stream Condition Index from Statewide Stream Surveys in Virginia
Citation:
McManus, M. Spatial Patterns in Stream Condition Index from Statewide Stream Surveys in Virginia. 13th National Monitoring Conference, Virginia Beach, VA, April 24 - 28, 2023.
Impact/Purpose:
The assessment of rivers and streams by environmental agencies is often done through spatial surveys that estimate a stream condition index (SCI) based on aquatic macroinvertebrates. The spatial patterning of SCI among monitoring stations, such as SCI values are more similar for stations near each other than further away, can potentially indicate over what scale, or spatial extent, the effect of stressors on SCI occur. I studies the spatial patterning in SCI using data collected by the Virginia Department of Environmental Quality from their stream surveys of 11 basins from 2001-2018. At three basins, (James, Potomac, and Roanoke) the expected geostatistical pattern of similar SCI values at very near distances and more divergent values with increasing distances between stations occurred. Three basins (Chowan, New, and Rappahannock) showed a slight geostatistical pattern, and the five other basins (Big Sandy, Clinch-Powell, Holston, Shenandoah, and York) did not show any such pattern. Understanding why this variation in SCI spatial patterning occurs can help identify stressors in basins.
Description:
The benefits of spatially balanced probabilistic surveys are twofold. First, from a design perspective, one obtains a distribution of sample sites with a similar layout relative to the target population. Second, from an analytical perspective, the more spatially patterned response observed among the sample sites then the local neighborhood variance estimate produces smaller confidence limits around means and percentiles compared to other variance estimators. I investigated the spatial patterning in stream condition index (SCI) using data collected by Virginia Department of Environmental Quality from their stream surveys of 11 basins from 2001-2018. My hypothesis was that basins with small interquartile ranges (IQRs), estimated with the local neighborhood variance, would have more spatial structure, that is an identifiable partial sill and range in semivariogram plots. Whereas basins with larger IQR's, which should show just a horizontal smear of semivariances points. Three basins, (James, Potomac, and Roanoke) showed the expected geostatistical pattern of smallest SCI semivariances at very near distances and larger semivariances with increasing distances between stations. The semivariogram plots of three basins (Chowan, New, and Rappahannock) showed a slight geostatistical pattern, and the five other basins (Big Sandy, Clinch-Powell, Holston, Shenandoah, and York) did not show any such pattern. The hypothesis I proposed was not supported by the spatial data analysis.
