Estimating solar power capacity on abandoned coal mines

This article is part of the Global Coal Mine Tracker, a project of Global Energy Monitor.

Download full dataset

Report an error

Sub-articles:

Global coal mines Coal mines in China Coal mines in India Coal mines in the United States Coal mines in Australia Coal mines in Russia Coal mines in Indonesia

Related-articles:

Largest coal mines in the world Coal mine impacts from Covid-19 Environmental issues of coal Coal mining Coal reserves Coal terminals

The opportunity to site solar energy projects on former coal mining lands is gaining increasing attention as a strategy to support renewable energy deployment while repurposing degraded industrial landscapes.

Global Energy Monitor (GEM) estimated solar photovoltaic (PV) capacity (MWac) on the footprints of abandoned surface coal mines for a briefing, "Bright Side of the Mine: Solar's Opportunity to Reclaim Coal's Footprint," published on June 18, 2025.

Introduction

GEM's approach relies primarily on the land area available at former mining sites, incorporating key parameters from it's Global Coal Mine Tracker and TransitionZero's Solar Capacity framework to create a scalable, replicable estimation based on spatial footprint alone.

Dataset and site selection criteria

GEM's initial methodology drew on the May 2025 release of the Global Coal Mine Tracker. From this dataset, 882 recently abandoned surface mines were identified, along with a broader set of 3,753 operating coal mines. For the mine size, GEM relies on reported mine permit boundaries when available, based on company reports or government permitting documents, and mapped degraded acreage based on satellite imagery when that information was otherwise unavailable.

To identify those mines with realistic potential for solar repurposing, the analysis filters for surface mines that have either closed since 2020 or are expected to close by 2030. Closure years are determined based on reported mine life, reserves-to-production ratios, or official retirement schedules. These criteria ensure that selected sites are either idle or likely to become idle within five years, offering a practical timeframe for redevelopment.

Equation

The potential solar capacity in alternating current is calculated using the following formula:

Capacity MWac = A × I × η × GCR × ILR

Where

• (A) is area of the mine site
• (I) is assumed solar irradiance
• (η) is panel efficiency
• (GCR) is the ground coverage ratio
• (ILR) is the inverter loading ratio

GEM pulled each of these variables from authoritative or benchmarked sources.

Table 1: Equation Variables and sources

Variable	Description	Value/Assumption	Source
A	Coal Mine Area	Derived from spatial footprint	Global Energy Monitor (GEM)
I	Nominal Solar Irradiance	1 kW/m²	TransitionZero
η	Panel Efficiency	19%	National Renewable Energy Laboratory (NREL)
GCR	Ground Coverage Ratio	30%	TransitionZero (20–80% range)
ILR	Inverter Loading Ratio	87%	TransitionZero

Benchmarking

To validate and contextualize the estimated solar capacities, GEM employed two benchmarking exercises. The first leverages updated land-use assumptions published by National Renewable Energy Laboratory (NREL), for solar power plants in the United States. In 2013, NREL published "Land-Use Requirements for Solar Power Plants in the United States," which includes direct and total area assumptions for small (1-20 MW) and large (>20 MW) scale solar power plants. In 2025, NREL used updated assumptions in another report titled "Land of Opportunity: Potential for Renewable Energy on Federal Lands." The updated assumptions do not differentiate between small and large-scale projects. We incorporated this updated assumption (5.75 acres/MWDC for the total land area) in our estimates as a benchmark. These assumptions provide a useful check against the raw area-to-capacity conversion used in the primary estimate.

The second benchmarking method uses TransitionZero’s global solar project dataset. With Irradiance, Panel Efficiency, and Inverter Loading Ratio fixed and the Mine Footprint given, we can estimate the Ground Coverage Ratio. To do this, the nearest 100 TransitionZero polygons to any given abandoned coal mine were identified, regardless of national boundaries. The idea behind this benchmarking was to explore the distribution of the area/capacity ratio for the closest TransitionZero polygons. The second step was to estimate a hypothetical solar power plant with the same footprint as the abandoned coal mine. Accordingly, the TransitionZero benchmark values were calculated by using the area/capacity ratio in each TransitionZero polygon and incorporating each area/capacity ratio and the actual coal mine footprint to calculate 100 benchmark points.

This process enabled the derivation of confidence intervals and comparative ranges around the GEM estimate.

Confidence intervals

Confidence intervals were constructed around the estimated MWac value using the TransitionZero benchmark sample. Although this approach assumes that the area-to-capacity ratios from nearby projects are normally distributed, visual inspections of a sample of 30 mines support this approximation. The standard deviation and margin of error were calculated using a sample size of 100, and the confidence interval is defined using a 95% confidence level. This provides a statistically informed range within which the true solar capacity is likely to fall, thereby increasing the robustness of the final results.

Output fields and interpretation

The output of this methodology includes the following values for each abandoned surface mine:

• Estimated MWac: The primary GEM capacity estimate using the equation above
• Benchmark NREL MWac: A secondary estimate based on NREL’s land-use assumptions
• Benchmark TZ 95 CI Low and Benchmark TZ 95CI High: The lower and upper bounds of the 95% confidence interval derived from TransitionZero benchmarking

Because of the inherent uncertainties in site conditions and solar design, the preferred reporting format for public or policy outputs is to present ranges rather than point estimates. For example: The closed Duralie Coal Mine in Australia has the potential to host a solar power plant with an estimated capacity ranging from 192.9 MW to 203.8 MW. When reporting at a national or regional level, these ranges can be summed to reflect aggregate capacity bands.

Limitations and Areas for Future Improvement

The current methodology made no adjustment to account for mine pools or vertical shading from pit depth. Future iterations plan to incorporate remote sensing data, legal records, zoning documents, and/or mine engineering records to better constrain usable land. Additionally, more granular GCR values based on system size or topography could increase accuracy. Improvements to the sampling of TransitionZero benchmarks and the statistical methodology for confidence interval generation are also under consideration, as is the inclusion of benchmarking against actual mine-to-solar conversions.