Spatial and Temporal Considerations in the Primary Mountaintop Removal and Valley Fill Mining Region of West Virginia

EPA Grant Number: F13E30934
Title: Spatial and Temporal Considerations in the Primary Mountaintop Removal and Valley Fill Mining Region of West Virginia
Investigators: Merriam, Eric Richard
Institution: West Virginia University
EPA Project Officer: Lee, Sonja
Project Period: August 18, 2014 through August 18, 2016
Project Amount: $84,000
RFA: STAR Graduate Fellowships (2013) RFA Text |  Recipients Lists
Research Category: Academic Fellowships , Fellowship - Aquatic Ecology


Successful management of aquatic resources within the mountaintop removal–valley fill (MTR-VF) mining region of central Appalachia will require the ability to predict ecosystem responses to current and future land use development across large spatial and temporal scales. The overall goal of this research is to develop a process to facilitate management and regulatory decisions in the face of current and future land use change within this dynamic region. Through a regional analysis of the primary MTR-VF province of West Virginia, the specific objectives of this research are to (1) compare the predictive success of spatial models constructed at different scales and (2) field test the ability of spatially derived models to predict changes in stream conditions over time.


In-stream data will consist of invertebrate, water chemistry and physical habitat samples collected from 160 sites during the summers of 2010 and 2011. Land cover and use data will consist of cumulative natural (e.g., coal geology) and anthropogenic (e.g., mining-related land cover, residential structures, NPDES permits) landscape characteristics for all 1:24,000 segment-level watersheds. A combination of multiple regression and boosted regression tree analyses will be used to construct models predicting aquatic condition from landscape attributes across multiple spatial scales (i.e., watershed and regional). Temporal uncertainty of the spatially derived models will be assessed through forecasting. A future landscape characterization will be constructed using 2013 NAIP aerial orthoimagery, and the spatially-derived models will predict future conditions. In-stream data collected from a network of sites during the summer of 2013 will be used to compare predicted to observed in-stream conditions.

Expected Results:

The hypothesis for objective 1 (i.e., comparing predictive success of spatial models constructed at different scales) is that the proper scale for model construction is the watershed scale. More specifically, the inferred cause-effect relationships will vary with spatial scale and more accurately reflect true relationships at smaller spatial scales. Consequently, predictive performance will increase with decreasing spatial scale. The working hypothesis for objective 2 (i.e., testing the ability of spatial models to predict changes over time) is that spatial patterns relating landscape attributes to in-stream conditions will allow accurate predictions associated with temporal landscape change. Strong relationships should exist between observed and predicted in-stream conditions. However, watershed-scale models will offer the best temporal predictions owing to more accurate underlying cause-effect relationships.

Potential to Further Environmental/Human Health Protection

This project will result in an integrated process that can be used to inform management (e.g., restoration) and regulatory (e.g., mine permitting and mitigation) decisions within the MTR-VF region based on current and expected future conditions. Because the MTR-VF region represents the headwaters of the Mississippi River basin, which supports one of the highest levels of biodiversity within North America, it will be critical to protect and manage these resources now and in the future. Moreover, decreased human health within the study region has been associated with degraded aquatic condition, suggesting that preserving and managing ecosystem integrity will simultaneously benefit the health of individuals living in mining-affected areas.

Supplemental Keywords:

land use, mountaintop mining, watershed models

Progress and Final Reports:

  • 2015
  • Final