Patricia F. Donovan

Spatial and temporal relationships among watershed mining, water quality, and freshwater mussel status in an eastern USA river

The Powell River of southwestern Virginia and northeastern Tennessee, USA, drains a watershed with extensive coal surface mining, and it hosts exceptional biological richness, including at-risk species of freshwater mussels, downstream of mining-disturbed watershed areas. We investigated spatial and temporal patterns of watershed mining disturbance; their relationship to water quality change in the section of the river that connects mining areas to mussel habitat; and relationships of mining-related water constituents to measures of recent and past mussel status. Freshwater mussels in the Powell River have experienced significant declines over the past 3.5 decades. Over that same period, surface coal mining has influenced the watershed. Water-monitoring data collected by state and federal agencies demonstrate that dissolved solids and associated constituents that are commonly influenced by Appalachian mining (specific conductance, pH, hardness and sulfates) have experienced increasing temporal trends from the 1960s through ~2008; but, of those constituents, only dissolved solids concentrations are available widely within the Powell River since ~2008. Dissolved solids concentrations have stabilized in recent years. Dissolved solids, specific conductance, pH, and sulfates also exhibited spatial patterns that are consistent with dilution of mining influence with increasing distance from mined areas. Freshwater mussel status indicators are correlated negatively with dissolved solids concentrations, spatially and temporally, but the direct causal mechanisms responsible for mussel declines remain unknown.

Using public domain data to aid in field identification of hydric soils

Hydric soil field identification is a common activity for natural resource professionals and planners, but it can be time consuming and labor intensive. This study used Soil Survey Geographic Database (SSURGO), National Wetlands Inventory (NWI), National Land Cover Data (NLCD), and other public domain data to make digital hydric soil predictive maps of two study areas in western Virginia. Soil scientists used the predictive maps as guides to conduct hydric soil field surveys and compared the results to delineations of SSURGO map units dominated by hydric soils and NWI and NLCD wetlands. At Stuarts Draft, 15% of the 1296-ha study area was composed of hydric soils compared with 14% estimated by SSURGO. At Blacksburg, 3% of the 828-ha study area was composed of hydric soils compared with 1% estimated by SSURGO. Both NWI and NLCD estimated 1% wetlands at each area. Locational correspondence was higher between the field survey and SSURGO than between the field survey and the NWI and NLCD wetlands at both study areas. The predictive maps were useful because the SSURGO delineations were closely aligned with field survey delineations, had <2% false negative identifications compared with >13% for NWI and NLCD at Stuarts Draft, and had ≤ 2% false positive identifications. Overlaying NWI and NLCD onto SSURGO polygons resulted in ≤ 1% improvement of predictive map utility, but all indicators of hydric soils were useful in narrowing the specific location of hydric soils within large SSURGO delineations.