Reclaiming the Appalachian Landscape: Understanding Surface Mine Reclamation Through Remote Sensing

Surface mining has been the dominant driver of land cover change in the Central Appalachian region for decades, fundamentally altering the topography and ecological trajectory of over one million acres of forest. While the Surface Mining Control and Reclamation Act (SMCRA) of 1977 mandated the restoration of mined land to an equal or better land use, the ecological integrity of these reclaimed landscapes remains uncertain, often characterized by "arrested succession" and the proliferation of invasive species. This dissertation integrates multi source remote sensing, time-series analysis, and causal inference modeling to comprehensively evaluate the status, history, and structural quality of post-mining land cover in Virginia and Tennessee. To establish a baseline of current reclamation status, we first developed a novel classification framework integrating Sentinel-2 spectral data, CCDC phenological metrics, and 3DEP LiDAR structure to map seven land cover classes, with a specific focus on the invasive shrub Autumn olive (Elaeagnus umbellata). The classification achieved area-weighted overall accuracies between 74.3% and 76.8%, revealing that Autumn olive occupies approximately 9.8% of mined land in Virginia, peaking in prevalence on sites 14 years post-disturbance. Expanding this analysis temporally, we reconstructed a 37-year history (1984–2021) of land cover dynamics, documenting a region-wide decline in barren land from 60% to roughly 14–22%. However, succession trajectories diverged significantly by state; while Tennessee mines transitioned largely to coniferous and herbaceous cover, Virginia mines experienced a progressive expansion of Autumn olive to 14% of the landscape by 2021. Spatially explicit ANN-CA-MC simulations project autumn olive presence will persist into the near future, and that historical expansion is driven by transitions from herbaceous and shrub/scrub cover. Finally, to assess the quality of successfully reforested mines, we utilized causal random forests and sample balancing to quantify the effect of mining on forest structure compared to non-mine disturbances. We found that while canopy cover on mines converges with reference forests after 15–20 years, mine forests in Virginia remain significantly stunted, averaging 2.28 m shorter than forests recovering from non mining disturbances. Furthermore, standard spectral recovery metrics derived from Landsat may not full capture these structural deficits, highlighting the necessity of structural data for reclamation monitoring. Collectively, these findings demonstrate that while surface mines have successfully revegetated, they frequently fail to restore structural attributes and desired land cover composition, often diverting into stable, invasive-dominated states that require active management intervention

Name thesis:degree_name

Doctor of Philosophy

Level thesis:degree_level

doctoral

Discipline thesis:degree_discipline

Forestry

Department dc:contributor.department

Forest Resources and Environmental Conservation

Grantor dc:publisher

Virginia Tech

Year dc:date.issued

2026

Author dc:creator

• Putnam, Daniel Jacob

Chairs dc:contributor.committeechair

• Wynne, Randolph H.
• Thomas, Valerie Anne

Committee members dc:contributor.committeemember

• Coulston, John W.
• Bessac, Julie

• cloud computing
• change detection
• machine learning

Statement dc:rights

• Creative Commons Attribution 4.0 International

Licence dc:rights.uri

• http://creativecommons.org/licenses/by/4.0/

Language dc:language.iso

en

Dc Identifier Other

vt_gsexam:45603

OAI identifier oai:identifier

oai:vtechworks.lib.vt.edu:10919/140762