Mining areas in China exhibit considerable spatial overlap with favorable solar irradiation conditions, as larger mining footprints tend to have higher solar energy potential^[23]. When planning PV systems in these mining areas, we consider the geographical features of key latitude zones and topography in PV layout design, while incorporating adaptive designs tailored to different types of mining land^[24]. The data on abandoned open-pit mines in China were obtained from publicly available sources ^[23,25], including three types of damaged mining land (i.e., excavated land, occupied land, and subsided land), ecological restoration land, and building rooftops. The scenario settings adopted in this study are as follows:

Scenario 1 adopts a strategy to maximize standalone PV development by prioritizing all available land for PV array installation to meet the mines’ clean electricity demand. The land used for PV layout comprises three types: damaged mining land, ecological restoration land, and building rooftops.

Scenario 2 employs a hybrid development strategy combining PV with ecological restoration. This approach aims to meet clean energy demand while supporting ecological restoration. To meet the mines’ clean energy demands and underscore the role of abandoned mine sites in building a renewable energy system, the vegetation selected for ecological restoration consists of energy crops. These plants provide biomass for clean energy. In this scenario, PV systems and energy crops are spatially separated with no overlap. Based on the land allocation scheme, ecological restoration land focuses solely on cultivating energy crops without PV layout, while building rooftops and damaged mining land are used for PV installation without vegetation. Cultivated energy crops are used for bioenergy production, including biopower, ethanol, and biodiesel.

Scenario 3 implements a “PV+” coupling model that emphasizes the coordinated development of PV deployment and ecological restoration. Ecological restoration is the primary goal, while PV deployment provides synergistic benefits. In this scenario, PV systems and energy crops coexist on the same land used for ecological restoration and on damaged mining land. To minimize impacts on plant growth, PV panels are installed with a minimum clearance of 2.5 meters above the ground, and a vertical clearance of 3.2 m at the panel center. The spaces beneath and around the panels are efficiently used to cultivate energy crops. Meanwhile, building rooftops are dedicated exclusively to PV deployment, with no vegetation cover. This dual-purpose approach promotes sustainable mine development.

A 600 Wp monocrystalline silicon PV module with a service life of 25 years was selected and installed at a fixed tilt angle, considering module efficiency, technological maturity, and market share^[24,26]. The specific module dimensions are 2,172 mm × 1,303 mm × 35 mm^[26]. The PV array consists of 44 modules arranged in two vertical rows. The minimum spacing between the front and rear rows of the array is calculated using^[24]:

(1) $$ D=\cos \beta \times \frac{H}{\tan [\arcsin (\sin \varphi \sin \delta+\cos \varphi \cos \delta \cos \omega)]} $$

(2) $$ D=\cos \beta \times L $$

(3) $$ L=\frac{H}{\tan \alpha} $$

(4) $$ \alpha=\arcsin (\sin \varphi \sin \delta+\cos \varphi \cos \delta \cos \omega) $$

(5) $$ \sin \alpha=\sin \varphi \sin \delta+\cos \varphi \cos \delta \cos \omega $$

(6) $$ \sin \beta=\frac{\cos \delta \sin \omega}{\cos \alpha} $$

as illustrated in Figure 1. Here, D represents the minimum spacing between the front and rear arrays of the PV module, L denotes the ground projection length of solar radiation, α is the solar altitude angle at 9:00 AM on the local winter solstice, φ represents the local latitude, β is the sun azimuth angle, which is 0° at 9:00 AM on the local winter solstice, δ is the sun declination, set at 3.5° for the winter solstice, and ω is the hour angle, which is 43.24° at 9:00 AM.

PVsyst is a specialized simulation platform widely used in PV system design and performance analysis^[27]. In this study, PVsyst is used to determine the optimal installation tilt angle of the PV modules, total solar irradiation, and peak sunshine hours. Subsequently, the minimum static distance between the front and rear rows of PV panels is calculated, as detailed in Supplementary Table 1.

The PV system’s output power is calculated using^[24]:

(7) $$ P=W \times H \times \eta \times(1- ƍ) $$

(8) $$ W=N \times W_{s} $$

(9) $$ N=\frac{A}{A_{s}} \times N_{s} $$

where P is the annual power generation of the PV system, in kWh, W represents the installed capacity, in kWp, H denotes the annual effective operating hours, in h, η is the overall system efficiency coefficient, 0.82^[24], ƍ is the module’s annual degradation rate, based on the manufacturer’s measured data: it is 2% in the first year, then stabilizes at 0.45% per year thereafter^[24,26], N is the total number of PV panels, W_s is the capacity of a single polysilicon PV module, in kWp, A is the application area per plan, in m², A_s is the required space per PV module with correction for open-pit mines, in m², and N_s is the number of PV panels per module.

Bioenergy production on abandoned mine land is calculated using^[28]:

(10) $$ M_{bioenergy-j, k}=A_{i} \times p_{i, j} \times f_{j, k} $$

where M_{bioenergy-j,k} is the amount of type k bioenergy production from type j energy crop, in t or kWh, A_i is the area of type i abandoned mine land, in ha, p_i,j is the unit output of type j energy crop on type i abandoned mine land, in t ha^-1, and f_j,k is the conversion coefficient from type j energy crop to type k bioenergy, in t t^-1 or kWh t^-1.

The energy conservation potential from PV power generation is calculated using:

(11) $$ C o n_{p}=P \times r+P \times F C-P \times E C $$

where Con_p represents the energy conservation due to PV power generation, in MJ, P is the power generation of the PV power station, in kWh, r is the conversion coefficient of the calorific value of electricity, 3.6 MJ kWh^-1, FC is the energy consumption of coal-fired power production over its life-cycle, quantified as 11.27 MJ kWh^-1[29], and EC is the energy consumption of PV power generation over its life-cycle, quantified as 1.81080 MJ kWh^-1[30].

The energy conservation through bioenergy generation is calculated using^[31]:

(12) $$ \begin{array}{l} Con _{bioenergy -k}=M_{{bioenergy }} \times H_{k}+M_{{bioenergy }} \times H_{k} \times \frac{C_{L C A-f}}{H_{f}} \\ -M_{{bioenergy }} \times C_{L C A-k} \\ \end{array} $$

where Con_bioenergy-k is the energy conservation achieved by using type k bioenergy instead of traditional fossil fuels, in MJ, M_bioenrgy is the production of type k bioenergy, in kg, H_k is the calorific value of type k bioenergy, in MJ kg^-1 or MJ kWh^-1, H_f is the calorific value of traditional fossil fuel that can be replaced by type k bioenergy, in MJ kg^-1 or MJ kWh^-1, C_LCA-f is the energy consumption during the life-cycle of traditional fossil fuel production that can be replaced by type k biomass energy, 11.27 MJ kWh^-1 for coal-fired power^[29], 56.35 MJ kg^-1 for gasoline^[31], and 55.81 MJ kg^-1 for diesel^[31], and C_LCA-k is the energy consumption during the life-cycle of type k bioenergy production, detailed in Supplementary Table 2.

The CO₂ mitigation of PV power generation is calculated using:

(13) $$ R E G_{p}=P \times\left(f_{f}-f_{p}\right) $$

where REG_p represents the CO₂ mitigation potentials of PV power, in t. P is the power generation of the PV power station, in kWh, f_f denotes the CO₂ emission coefficient for fossil fuel power production, 0.8426 ×10^-3 t CO₂ kWh^-1[32], and f_p denotes the CO₂ emission coefficient for PV power generation over its life-cycle, 0.2924 ×10^-3 t CO₂ kWh^-1[33].

CO₂ mitigation from bioenergy production throughout the life-cycle is calculated using:

(14) $$ G R V_{k}=M_{{bioenergy } f-k} \times f_{k} $$

where GRV_k is the greenhouse gas (GHG) reduction value of type k bioenergy, in t, and f_k is the carbon reduction coefficient of type k bioenergy, reported as 0.9288 × 10^-3 t CO₂ kWh^-1 for biopower^[34], 1.87 t CO₂ t^-1 for ethanol^[35], and 2.34 t CO₂ t^-1 for biodiesel^[36].

A life-cycle cost-benefit model was used to conduct a comprehensive economic assessment of PV power and bioenergy generation. The Levelized Cost of Electricity (LCOE) is a core indicator for evaluating the economic feasibility of PV projects, reflecting the average cost per unit of electricity generated over the project’s entire life-cycle. The economic model includes four primary cost categories: initial capital investment, operation and maintenance expenditures, financing costs, and other costs. Among these, the total initial investment costs include PV modules, auxiliary equipment, grid connection, installation, and other relevant components. Due to the substantial upfront investment, investors adopt a principal repayment scheme to service the bank loans for renewable energy projects, as given in^[37]:

(15) $$ \text { LCOE }=\frac{C_{I}+\sum_{n=1}^{N} \frac{C_{ {O\&M }, n}+C_{ {Others }, n}}{(1+r)^{n}}}{\sum_{n=1}^{N} \frac{E_{n}}{(1+r)^{n}}} $$

(16) $$ C_{I}=W_{P V} \times C_{W} $$

(17) $$ C_{O \& M, n}=W_{P V} \times P_{O \& M, n} $$

(18) $$ C_{L, n}=R_{{Load }} \times l \times C_{I} \times\left[1-\frac{n-1}{T}\right], n=1,2, \ldots, T $$

(19) $$ C_{{Others }}=C_{I} \times R_{ {Others }} $$

where C_I is the initial investment cost of the PV system, CNY, C_O&M,n is the annual operation and maintenance costs, CNY, C_Others is the other annual related cost, including taxes, fees, and degradation related expenses, CNY, E_n is the electricity generation in the n-th year, in kWh, r is the discount rate, 5%, n is the operation period of the project, 25 years, W_PV is the installed capacity of the PV system, C_W is the unit initial investment cost, 2.77 CNY W^-1[38], P_O&M,n is the operation and maintenance cost per unit, 0.046 CNY W^-1[38], C_L,n is the annual loan interest payment, CNY, R_Loan is the ratio of loans to initial investment, 30%, T is the loan repayment period, 15 years, l is the long-term loan interest rate, 3.5%^[39], and R_Others is the ratio of other costs to initial investment, 10%^[40].

Revenue is derived from the sale of electricity, ethanol, and biodiesel. Electricity revenue is calculated based on the policy principle of “self-consumption priority with surplus electricity fed into the grid”^[41]. The prices of electricity, ethanol, and biodiesel are based on market data. The specific values are determined using:

(20) $$ C_{k}=C_{u, k} \times E_{k} $$

(21) $$ R_{k}=P_{u, k} \times E_{k}+E_{{self }} \times P_{ {self }}+E_{ {grid }} \times P_{ {grid }} $$

(22) $$ B_{k}=R_{k}-C_{k} $$

where C_k is the total life-cycle cost of the type k of energy, CNY, and C_u,_k is the unit cost of the type k of renewable energy, 0.3018 CNY kWh^-1 for PV power (LCOE), 0.475 CNY kWh^-1 for biopower^[42], 9,000 CNY t^-1 for ethanol^[43], and 4,500 CNY t^-1 for biodiesel^[44]. E_k is the type k of energy production, in t or kWh, R_k is the revenue from the type k of renewable energy, CNY, P_u,k is the unit price of the type k of bioenergy, 0.648 CNY kWh^-1 for biopower^[45], 6,500 CNY t^-1 for ethanol^[46], and 8,000 CNY t^-1 for biodiesel^[47], E_self is the self-consumed power generation, in kWh, P_self is the self-consumption electricity price, 0.643 CNY kWh^-1[48], E_grid is the grid-fed power generation, in kWh, P_grid is the feed-in tariff (capped at the benchmark price of desulfurized coal-fired power), 0.326 CNY kWh^-1[49], and B_k represents the net benefit from energy production, CNY.

For comparative analysis, the levelized cost per unit area (LCOA) and energy output per unit area (EOA) were defined for abandoned mine lands, which are expressed as:

(23) $$ L C O A=\frac{C_{k}}{A_{k}} $$

(24) $$ E O A=\frac{E_{k} \times H_{k} \times 10^{-9}}{A_{k}} $$

where LCOA is the energy production cost per unit area, CNY km^-2, C_k is the total life-cycle cost of the type k of energy, CNY, A_k is the land area occupied by raw materials for producing type k energy, in km², EOA is the energy production per unit area, in PJ km^-2, E_k is the k type of energy production, in t or kWh, and H_k is the calorific value of type k renewable energy, in MJ t^-1 or MJ kWh^-1.

Data on abandoned open-pit mine lands across China were obtained from published remote sensing-based monitoring datasets and historical mine verification records^[23,25]. The ecological reclamation plan involves three categories of energy crops: cellulose-, carbohydrates-, and oil-producing crops. Biomass productivity for these energy crops on reclaimed mine lands was estimated based on their yields from marginal lands, with a limiting correction factor of 0.8 applied to account for low soil fertility and potential heavy metal contamination at mine sites^[24]. The spatial distribution of these energy crops, their unit yield on abandoned mine lands, and corresponding bioenergy conversion coefficients are presented in Supplementary Table 2. The data sources for other parameters are described in the section on parameter accounting formulas. All figures were created using Microsoft Excel 2021 and QGIS Desktop 3.44.0.

Over the 25-year operational life-cycle of PV systems, annual PV power generation in abandoned open-pit mines in China shows a gradual decline [Figure 2]. Under Scenario 1, the designated PV deployment area of 9,136.48 km² can accommodate 28.70 million polysilicon PV modules, with a total installed capacity of 757.71 GWp. Considering solar resource availability and various losses, the annual power generation is projected to decrease from 872.75 TWh yr^-1 in the first year to 764.97 TWh yr^-1 in the 25th year, with an average annual power output of 812.64 TWh yr^-1 [Figure 2 and Table 1]. In Scenario 2, 12.91 million polysilicon PV modules are installed across 5,936.67 km² of abandoned open-pit mine land, yielding an installed capacity of 340.82 GWp, with an estimated average annual power generation of 368.12 TWh yr^-1 [Figure 2]. The remaining 3,199.81 km² of land is allocated to the cultivation of energy crops for bioenergy production. The bioenergy potential varies substantially depending on conversion approaches and planting scale. Specifically, annual biomass yields are 3.75-18.17 Mt yr^-1 for cellulose energy crops, 0.35-0.66 Mt yr^-1 for oil energy crops, and 1.62-9.60 Mt yr^-1 for carbohydrate energy crops. The corresponding bioenergy potential equals 3.75-18.17 TWh yr^-1 of biopower from cellulose feedstocks, 0.35-0.66 Mt yr^-1 of biodiesel from oil feedstocks, and 1.62-2.94 Mt yr^-1 of ethanol from carbohydrate feedstocks [Table 1]. In Scenario 3, 16.50 million polysilicon PV modules are deployed over 7,536.58 km² of abandoned open-pit mine land, achieving an installed capacity of 435.61 GWp and estimated average annual power generation of 468.49 TWh yr^-1 [Figure 2]. The annual biomass yields of energy crops reach 8.51-41.23 Mt yr^-1 for the cellulose type, 1.04-1.50 Mt yr^-1 for the oil type, and 3.67-21.78 Mt yr^-1 for the carbohydrate type. This corresponds to a bioenergy output of 8.51-41.23 TWh yr^-1 for biopower, 0.78-1.50 Mt yr^-1 for biodiesel, and 3.67-6.68 Mt yr^-1 for ethanol [Table 1].

The overall renewable energy production potential across the three scenarios varies with research objectives and design configurations, ranging from 1.34 EJ yr^-1 to 2.93 EJ yr^-1, following the order: Scenario 1 > Scenario 3 > Scenario 2. Under the established modeling framework, PV power accounts for more than 94% of the total renewable energy potential.

In Scenario 1, PV power could offset approximately 2.93 EJ yr^-1 of coal-fired power. Throughout the full life-cycle, PV systems require an annual energy input of 1.47 EJ yr^-1 while delivering a net energy saving of 10.61 EJ yr^-1 compared to coal-fired power. In Scenario 2, the overall energy substitution potential ranges from 1.34 EJ to 1.40 EJ, accounting for 45.76%-47.99% of the potential observed in Scenario 1. Although bioenergy potential fluctuates substantially under different utilization strategies, PV power maintains a dominant contribution exceeding 97%, stabilizing the total substitution capacity in Scenario 2. After considering the energy consumption associated with renewable energy production (0.67-0.72 EJ), the net energy conservation for Scenario 2 ranges from 4.82 EJ to 5.02 EJ. Scenario 3 exhibits a higher energy substitution potential (1.72-1.87 EJ) and greater energy conservation (6.14-6.61 EJ), making a 26.76%-40.04% increase compared with Scenario 2 [Table 1]. The share of PV capacity is positively correlated with overall system potential. Scenario 1 achieves the highest substitution efficiency through maximal PV deployment. Scenario 2 is constrained by inherent low energy density and variability of bioenergy resources, thereby reducing the total system potential. In contrast, the “PV+” coupling mode in Scenario 3 effectively offsets the disadvantages of biomass feedstocks’ low energy density.

The total carbon mitigation potential from substituting conventional fossil fuels varies substantially across scenarios, ranging from 283.25 Mt CO₂ to 613.79 Mt CO₂ [Table 1]. Specifically, Scenario 1 shows the highest mitigation potential at 613.79 Mt CO₂, driven primarily by large-scale PV deployment displacing coal-fired power, resulting in an annual carbon reduction of 755.30 kg CO₂ per MWh. Scenario 3 lies in the intermediate range, with carbon reductions between 355.68 and 392.14 Mt CO₂, where PV power accounts for over 94% of the total mitigation. Scenario 2 shows the lowest carbon reduction potential, ranging from 278.85 to 294.92 Mt CO₂. Within Scenarios 2 and 3, carbon reductions vary by bioenergy pathway, ranging from 10.18 Mt to 23.10 Mt CO₂ for biopower, 1.18 Mt to 2.68 Mt CO₂ for biodiesel, and 4.27 Mt to 9.68 Mt CO₂ for ethanol. From a per-unit-area perspective, PV power delivers far greater carbon mitigation potential (611.63 t CO₂ ha^-1) compared with bioenergy alternatives such as biopower (102.96 t CO₂ ha^-1), ethanol (61.99 t CO₂ ha^-1), and biodiesel (17.13 t CO₂ ha^-1) [Figure 3]. Specifically, PV power’s mitigation potential is approximately 5.94 times that of biopower, 9.87 times that of ethanol, and 35.71 times that of biodiesel. This highlights that PV power substitution delivers considerably stronger per-unit-area carbon mitigation than all evaluated bioenergy options.

Regional potential for renewable energy production on abandoned open-pit mines varies substantially across provinces. In Scenario 1, Inner Mongolia ranks highest, with 369.14 PJ of PV power generation, followed by Xinjiang (210.49 PJ), Shandong (209.88 PJ), and Hebei (200.69 PJ). In contrast, the municipalities of Shanghai and Beijing exhibit the lowest output, accounting for only 0.6%-3.7% of Inner Mongolia’s generation. Scenario 2 shows a similar regional pattern, with Inner Mongolia leading at 217.35 PJ, followed by Xinjiang (104.37 PJ), Guangxi (93.80 PJ), and Shandong (74.44 PJ). This consistent trend underscores the dominant influence of PV power generation on regional renewable energy potential. Scenario 3 adopts an integrated “PV+” coupling model, which further reinforces Inner Mongolia’s leading position (245.24 PJ), followed by Xinjiang (133.00 PJ), Guangxi (97.81 PJ), and Yunnan (95.81 PJ) [Figure 4]. This scenario demonstrates that integrating energy crops can significantly enhance renewable energy potential in key regions.

This study reveals that the potential for PV power generation varies across provinces, influenced by the distribution and characteristics of abandoned mine lands [Figure 5]. In Scenario 1, Inner Mongolia, Xinjiang, Shandong, and Hebei have the highest PV power generation potential due to their extensive areas of abandoned open-pit mines. When focusing exclusively on damaged mining land, Guangxi emerges as the leading region for PV potential under Scenario 2. With the inclusion of ecological restoration land, Yunnan’s PV power potential is further enhanced. Across all scenarios, Inner Mongolia, Xinjiang, Shandong, Guangxi, Hebei, and Yunnan consistently demonstrate strong potential for renewable energy production, energy conservation, and carbon emission reduction. Therefore, these provinces with abundant abandoned mines should be prioritized for integrated ecological restoration and renewable energy development. Such an approach would not only support provincial and national carbon-neutrality and emissions-reduction targets but also serve as a benchmark for sustainable development in the mining sector.

Renewable energy development, particularly through PV power generation and bioenergy production on abandoned mine lands, displays varying energy conservation potential across provinces. Provinces with higher renewable energy generation also show the greatest energy conservation potential. Specifically, Inner Mongolia and Xinjiang exhibit the greatest potential, saving 1,339.02 PJ and 763.54 PJ under Scenario 1. These values are approximately 1.7 and 2.0 times higher than those under Scenarios 2 (784.87 PJ and 374.33 PJ) and 3 (878.32 PJ and 471.13 PJ), respectively [Figure 6]. These two provinces also lead in carbon mitigation, with reductions of 77.45 Mt CO₂ in Scenario 1, 45.38 Mt CO₂ in Scenario 2, and 50.76 Mt CO₂ in Scenario 3 [Figure 7].

In terms of cost per unit area [Table 2], the order from highest to lowest is Scenario 1 > Scenario 3 > Scenario 2. This pattern is attributed to the longer operational life-cycle of PV systems, which results in higher maintenance and operating costs. However, energy output per unit area follows the same trend, indicating that despite the higher costs, PV systems deliver substantially greater energy generation. Consequently, Scenarios 1 and 3 achieve higher energy returns than Scenario 2 despite their greater costs. From a benefit-cost ratio perspective, Scenario 1 is the most efficient configuration, whereas Scenario 2 is the least economically viable. Biomass energy systems are typically constrained by land requirements and crop growth cycles. In addition, multi-stage conversion processes introduce efficiency losses, resulting in lower stability and lower efficiency for continuous energy supply compared with PV systems as well as weaker economic performance. In contrast, the “PV+” coupling model optimizes land use by reducing revegetation costs while simultaneously providing ecological and power generation benefits. These results demonstrate that the proposed “PV+” scenarios not only align with national energy and ecological restoration strategies but also are economically feasible.