Marginal land in China suitable for bioenergy crops under diverse socioeconomic and climate scenarios from 2020–2100

ABSTRACT

Maximizing the development of renewable energy plays a critical role in mitigating the climate crisis. Marginal land provides space for the development of biomass energy; however, it remains unclear how the amount and spatial distribution of marginal land that is suitable for energy crop development will change in the future. Here, we project energy marginal land changes in China following the shared socioeconomic pathway (SSP) and/or representative concentration path (RCP). We provide datasets of marginal land, agriculturally suitable land, and potentially suitable for energy crops under historical scenarios and six future scenarios (i.e. SSP1–1.9, SSP1–2.6, SSP4–3.4, SSP2–4.5, SSP4–6.0, and SSP3–7.0) for the period 2020–2100, with a spatial resolution of 5 km. Under the six scenarios, from 2020–2100, the area of suitable marginal land ranged from 1.90–16.28 (Jatropha curcas L.) to 37.37–73.97 (Panicum virgatum L.) (×10⁴ km²), depending on the choice of energy crops and climate scenario. Based on the growing suitability of eight important bioenergy crops—Ricinus communis L., Saccharum officinarum L., Pistacia chinensis Bunge, Panicum virgatum L., Jatropha curcas L., Miscanthus giganteus J., Manihot esculenta Crantz, and Sorghum bicolor Moench—our dataset can be used to identify suitable locations for specific energy crops. This new synthetic dataset could support the development of multiscenario-based solutions related to carbon neutrality, ecosystem services, and energy transition.

KEYWORDS:

• Marginal land
• climate change
• energy crops
• liquid fuel

1. Introduction

Land use/land cover in China has undergone dramatic changes in recent decades, which is directly related to a wide range of high-stake issues, such as biodiversity, greenhouse gas emissions, ecohydrological processes, biogeochemical cycles, and climate extremes (Chen et al., 2019; Duveiller et al., 2018; Luo et al., 2022; Sy & Quesada, 2020). Marginal land is an often-ignored but significant part of land-use/land-cover, defined as areas that are poorly suited to field crops due to low crop productivity caused by inherent edaphic or climatic limitations or that are vulnerable to erosion or other environmental risks when cultivated (Csikós & Tóth, 2023; Gelfand et al., 2013; Zhu et al., 2022). Pulighe et al. (2019) underscore the importance of evaluating the concept of marginal lands from the perspective of human resource utilization and ecosystem services. In recent years, the study of marginal lands has emerged, particularly in terms of their potential for biomass and bioenergy production (Edrisi et al., 2022). Case studies from India demonstrate significant promise for these lands in meeting energy demands (Edrisi & Abhilash, 2016). Concurrently, economic and feasibility analyses of emerging technologies such as hydrothermal carbonization, which produces bioenergy from sawdust, underscore their importance in a circular economy (Vallejo et al., 2023). Within the framework of the United Nations Decade on Ecosystem Restoration, performing interdisciplinary research requires accelerating land restoration as part of a global commitment to sustainable development (Edrisi & Abhilash, 2021). Studies on the suitability of land for energy crops under climate change and land-use transitions provide key insights (Cronin et al., 2020). In the realm of socioeconomic opportunities, the focus on the potential for bioenergy crops on marginal and degraded lands, along with their carbon mitigation effects, has become a pivotal interest for achieving sustainability (Cervelli et al., 2020; Edrisi et al., 2022; Haberzettl et al., 2021; Panoutsou & Chiaramonti, 2020). Collectively, these investigations reveal the critical role of marginal lands in energy production and sustainable land management, offering valuable guidance for future research and policy-making (Jiang et al., 2019). Milbrandt et al. (2014) estimated that marginal lands (including abandoned cropland, abandoned mine lands, and brownfield) in the 48 contiguous states of the United States cover an area of approximately 865,000 km², approximately 11% of the land area of the contiguous United States. Similarly, Cai et al. (2011) reported that China’s marginal land area ranged from 5.2 to 21.3 (×10⁴ km²), representing approximately 5.43% to 22.26% of China’s total land area. These lands are generally unused and difficult to cultivate and have low economic value and varied developmental potential (Milbrandt et al., 2014). The complicated changes associated with marginal land are subject to a variety of social and economic factors and may be further exacerbated or mitigated in the context of climate change (Csikós & Tóth, 2023; Kim et al., 2023; Vera et al., 2022).

Energy is an important material basis for the survival and development of human society, and fossil energy shortages and environmental pollution are the two major challenges faced by the global community (Du et al., 2022; Keleş, 2011; Qu et al., 2022). The remaining recoverable reserves of coal, oil, and natural gas per capita in China are only 58.6%, 7.7%, and 7.1% of the world average, respectively (Du, 2008). Moreover, China accounts for a large share of the world’s CO₂ emissions, with a total of 14 GtCO₂e in 2020 (United Nations Environment Programme, 2022). Countries worldwide are actively adjusting the structure of their energy industries and developing liquid biofuels as an important way to address the above problems (Fu et al., 2020). In particular, the National Energy Administration (NEA) of China released the “13th Five-Year Plan for Biomass Energy Development” in 2016, which clearly encourages the development of biomass and other biomass industries at the national level (NEA, 2016). In 2022, the National Development and Reform Commission (NDRC) issued the “14th Five-Year Plan for bioeconomy development”, which proposed actively developing bioenergy and promoting the transformation of fossil energy to green and low-carbon renewable energy (NDRC, 2022). Shao and Chu (2007) assessed the energy crop resources in China and found that China has 1554 species of oil crops, including 154 species with oil contents greater than 40% in their seeds and 30 species of shrubs or arbor crops with abundant biofuel components. At present, the most commonly used bioenergy crops can be generally classified into starch-producing crops, sugar-producing crops, lignocellulosic biomass crops, and oilseed crops for biodiesel production (Li et al., 2010). The development and use of energy crops can not only reduce the overdependence on oil for industrial development and avoid additional carbon emissions but also have a good environmental impact (Xu et al., 2013). The use of marginal land for energy crops has good environmental and economic prospects; however, the challenges involved are enormous, the most fundamental of which–and the first to be considered–is the identification of suitable marginal land, taking into account food availability and ecological protection (Wu et al., 2010). Csikós and Tóth (2023) emphasize the importance of ensuring sustainability in the process of cultivating energy crops, supporting nature-based socioeconomic development. While the utilization of marginal lands can provide opportunities for the energy industry, careful consideration is needed to balance the negative impacts of land management. Furthermore, the research by Baude et al. (2019) focused on highlighting the significance of considering ecosystem services when selecting marginal lands to ensure sustainable land use. With limited cultivated land resources in China, the development of energy crops must follow the idea of “not using the grain intended for human consumption and not occupying the land intended for grain production” (Jiang et al., 2014).

With such a large potential distribution of marginal land and its great potential for renewable energy development, the protection and use of marginal land has become a major concern for policy-makers (Hao et al., 2022; Wu et al., 2022). Due to the poor quality of the soil on marginal land, there is a risk of further degradation due to artificial or natural factors (Vera et al., 2022; Werling et al., 2014). Existing research on energy crops on marginal land has focused on cultivation techniques, production processes, potential spatial distribution, production potential, and the environmental impacts of energy crops (Bassam, 2013; Chen et al., 2018; Shao & Chu, 2007; Werling et al., 2014; Xu et al., 2022; Zhuang et al., 2011). It has been recognized that, through good conservation and management, marginal lands can also have a positive impact on ecological improvement, mitigating energy scarcity, and reducing carbon emissions (Muscat et al., 2020; Wang et al., 2021; Xu et al., 2022). As a result, many countries have committed to developing liquid biofuels by growing energy crops on marginal lands to reduce the environmental pollution caused by the heavy use of fossil energy (Fu et al., 2020; Jacot et al., 2021; Robertson et al., 2017; Wang et al., 2021). Although the benefits of growing energy crops on marginal lands are widely recognized, the resources available for energy-friendly marginal lands to date are limited (Cai et al., 2011; Fu et al., 2020; Qin et al., 2011). Most studies have provided estimates of marginal land in China, but none have specifically examined changes in marginal land for various energy crops under future climate and socioeconomic scenarios in China (Fu et al., 2014, 2020; Jiang et al., 2014; Li et al., 2010; Shao & Chu, 2007).

Compared with alternative models, the representative concentration pathways (RCPs) and shared socioeconomic pathways (SSPs) based on the CMIP6 Land Use Model Intercomparison Project exhibit enhanced spatial resolution, offering a more accurate portrayal of the intricacies within the climate system. This improvement is manifested in a more faithful representation of crucial processes such as clouds, radiation, and the marine biosphere, resulting in more reliable climate simulations. The integration of new climate variables, along with an expanded array of scenarios and experiments, renders CMIP6 more versatile and applicable. This heightened flexibility enables a comprehensive assessment of model performance across various developmental trajectories and external forcings in the future. Consequently, CMIP6 provides us with a good research method and background for exploring how the world will change during the remainder of the 21st century (O’Neill et al., 2016, 2017; van Vuuren et al., 2011). Therefore, to facilitate further development in the field of biomass energy, this paper provides a marginal land resource distribution dataset for future suitable energy sources in China under six scenarios: SSP119, SSP126, SSP434, SSP245, SSP460, and SSP370. The main research involves the spatial extent and marginal land suitable for the cultivation of eight high-energy crops (i.e. Ricinus communis L., Saccharum officinarum L., Pistacia chinensis Bunge, Panicum virgatum L., Jatropha curcas L., Miscanthus giganteus J., Manihot esculenta Crantz, and Sorghum bicolor Moench) under different future scenarios for the production of adequate quantities of fuel ethanol and biodiesel. The goals of ecosystem restoration and sustainable land management advocated by the United Nations Decade of Ecosystem Restoration are also closely related to the development of energy crops in China. In this context, investigating China’s marginal land resources suitable for future energy crops not only offers key information for the biomass energy sector but also aligns with the objectives of the United Nations Decade of Ecosystem Restoration. We are expected to provide a scientific basis for future land use policy decisions and sustainable ecosystem management, promote the sustainable development of biomass energy to cope with global climate change and promote an environmentally friendly energy transition.

2. Methods

2.1. Overall framework

The SSP is a set of paths that describes future social and economic development and policy trends. Each path represents a possible future social development scenario. The representative concentration pathways (RCPs) outline potential routes for future atmospheric greenhouse gas (GHG) concentrations. These trajectories, which are employed to depict forthcoming GHG concentrations, encompass a variety of plausible emission outcomes. The RCP offers a systematic framework for the assessment and comprehension of potential climate scenarios, depending on varying levels of GHG emissions. It was developed through a structured process to characterize the diverse pathways of future atmospheric GHG concentrations. First, we used the Coupled Model Intercomparison Project Phase 6 (CMIP6) to process the raw meteorological data under the six SSP–RCP (shared socioeconomic pathway–representative concentration pathway) scenarios (see Table 1) (i.e. SSP1–1.9, SSP1–2.6, SSP4–3.4, SSP2–4.5, SSP4–6.0, and SSP3–7.0) into a GIS-operable data format according to suitable growth regimes for the different bioenergy crops. The scenarios are low to high radiative with forcings from low to high, with the high-emission SSP5–8.5 scenario (highly unlikely pathway) not developing energy crops for the time being. Second, we used a reclassification scheme to harmonize the land-use types and then utilized a well-validated dataset of predicted land-use changes under different SSP–RCP scenarios for the classification of plant functional types in China published by Liao et al. (2020) and freely downloaded from https://geosimulation.cn/China_PFT_SSP-RCP.html (see Table 2). Finally, we integrated comprehensive SSP–RCP scenario meteorological data according to the suitable growth regime for bioenergy crops (see Table 3) with land-use projection data, downscaling the CMIP6-based meteorological data to the same spatial resolution of 5 km. We analysed the future desirable energy marginal land for China in the near-, medium-, and long-term by considering the period 2020–2100 in five-year intervals. Figure 1 shows the methodological framework for determining the marginal land suitable for the considered energy crops in China under the different climate scenarios.

Figure 1. The methodological framework for generating marginal land suitability data in China under different scenarios from 2020 to 2100.

Table 1. The SSP and RCP scenario matrix and the coupled scenarios covered in this study.

Tier	Scenario	Description
Tier 1 (including a set of basic and mandatory climate simulations that incorporate various SSP-based greenhouse gas concentration scenarios and other climate forcings)	SSP3–7.0	A sustainable future for a region is envisioned, with regions achieving greater success in sustainability. (Successful sustainable development in the region, relatively weak coordination at the global level, some degree of climate action)
	SSP2–4.5	To predict a future with the emphasis on regionalism and incoherence in global climate action. (Emphasis on regionalism, limited international cooperation, inconsistent climate action)
	SSP1–2.6	To describe a future that is at the global average level; there is no overemphasis on sustainability and no special emphasis on economic growth. (Average levels of global cooperation, moderate population growth rates, balanced economic growth and social development)
Tier 2 (focusing on specific scientific questions, additional climate variables, higher resolutions or more complex earth system processes)	SSP4–6.0	A high-energy-demand future is depicted, with economic growth and energy demand continuing to increase. (High energy demand, emphasis on economic growth, some degree of climate action)
	SSP4–3.4	To paint a future in which global climate action is somewhat challenged and international competition may lead to less international cooperation. (Geopolitical tensions, climate action blocked, international cooperation limited)
	SSP1–1.9	To emphasize the future of global sustainable development and focus on the balance of environmental protection, social justice and economic prosperity. (Global collaboration, low population growth, socioeconomic equality, sustainable development.)

Table 2. Data characteristics and sources.

Dataset	Resolution	Data source	Timespan
Land use projections	5 km	https://geosimulation.cn/China_PFT_SSP-RCP.html	2020–2100
CMIP6 meteorological projections	50 km	https://esgfnode.llnl.gov/search/cmip6/	2020–2100
DEM	1 km	https://srtm.csi.cgiar.org/	–
Soil properties	1 km	https://www.fao.org/soils-portal/soil-survey/soil-maps-and-databases/harmonized-world-soil-database-v12/ru/	2014
Administrative boundaries	–	http://bzdt.ch.mnr.gov.cn/	–

Table 3. Suitable growth systems for different energy crops.

Factor	Specific parameter	Ricinus communis L.	Saccharum officinarum L.	Pistacia chinensis Bunge	Panicum virgatum L.	Miscanthus giganteus J.	Jatropha curcas L.	Manihot esculenta Crantz	Saccharum officinarum L.
Terrain	Slope (°)	<25 (Liu et al., 2015)	<25 (Peng et al., 2020)	<25 (Liu et al., 2015)	<25 (Nie et al., 2022)	<20 (Fu et al., 2022)	<25 (Liu et al., 2015)	<25 (Fu et al., 2020)	<25 (Peng et al., 2020)
	Altitude (m)	<500 (Wu et al., 2013)	–	600–2000 (Pei & Chen, 2005)	<3000 (Wang, 2018)	<2000 (Hu et al., 2008)	<1600 (Chang & Wang, 2018)	<2000 (Fu et al., 2020)	–
Soil	Organic matter content (%)	>1.5	>1.5 (Peng et al., 2020)	>1.5 (Lu, 2013)	>1.5 (Nie et al., 2022)	>1.5	>1.5 (Liu et al., 2015)	>1.5 (Fu et al., 2020)	>1.5 (Peng et al., 2020)
	pH (H2O)	>4.5 (Yan & Wang, 2007)	4.5–8.0 (Peng et al., 2020)	5.5–7.8 (Lu, 2013)	4.4–8 (Nie et al., 2022)	4.2–8.0 (Xu & Yang, 2009)	5.5–8 (Wang, 2018)	3.8–8.0	4.5–8.0 (Peng et al., 2020)
Temperature	Annual average temperature (℃)	>10 (Zhang et al., 2009)	13–28 (Peng et al., 2020)	>10 (Fu et al., 2009)	10.3–30(Lewandowski et al., 2003; Liu et al., 2009)	15–28 (Xu & Yang, 2009)	>15 (Lv & Yu, 2009)	>18 (Fu et al., 2020)	13–28 (Peng et al., 2020)
	≥10°C Cumulative temperature (℃)	>2300 (Dong et al., 2008)	>6500 (Li et al., 2012)	>3800 (Liu et al., 2015)	>3500 (Zhang et al., 2018)	>1800	6000–9500	>2000 (Fu et al., 2020)	>6500 (Li et al., 2012)
Rainfall	Annual average precipitation (mm)	>600 (Yan & Wang, 2007)	>800 (Cai, 2002)	>300 (Liu et al., 2008)	>150 (Nie et al., 2022)	>600 (Xie et al., 2011)	>200 (Mulpuri et al., 2019)	600–6000 (Fu et al., 2020; Liu et al., 2015)	>800 (Cai, 2002)

2.2. Procedure

The marginal land dataset for suitable energy crops in China under different climate scenarios for 2020–2100 includes the CMIP6 meteorological projection dataset, the Chinese land-use change projection dataset under different SSP–RCP scenarios, soil data, and other datasets such as digital elevation model (DEM) data. These datasets utilized are considered as the base data and are generated through data format conversion, reclassification and data merging. The steps of the study are as follows:

Step 1: Meteorological projection data processing. Daily meteorological data from Earth System Model version 4.1 (GFDL-ESM4) in CMIP6 for different scenarios from 2020–2100 were converted from NetCDF file format to TIF format using MATLAB (R 2016a) software. The data were reduced to a resolution of 5 km using cubic convolutional interpolation. Based on the suitable growth conditions of the different bioenergy crops (Table 3), the ranges of the mean annual temperature, annual cumulative temperature and annual precipitation were extracted for the respective crops.

Step 2: Land-use projection data processing. The Department of Science and Education of the Ministry of Agriculture of China defines marginal land for energy crops as idle winter land and energy-suitable wasteland that can be used to grow energy crops (where energy-suitable wasteland refers to open forest, natural grassland, shrubland, and unused land that is suitable for the cultivation of energy crops). By comprehensively considering the constraints of arable land protection policy, ecological protection, and the development constraints of energy crops on a large scale and combining the characteristics of land resource types suitable for energy crop development, 10 land-use types suitable for reclaiming and growing energy crops were extracted from the study of Liao et al. (2020). These land-use types included bare; needleleaf deciduous tree, boreal; broadleaf deciduous tree, tropical; broadleaf deciduous tree, temperate; broadleaf deciduous tree, boreal; broadleaf deciduous shrub, temperate; broadleaf deciduous shrub, boreal; C₃ arctic grass; C₃ nonarctic grass; and C₄ grass. In this product, the grid value represents the proportion of a land type, and the total proportion of all land types in the grid is 100%. Each land use type has a grid value of 0–1, representing the likelihood of that land use type. Therefore, if the proportion of these 10 land use types in a grid is equal to or higher than 0.5, the grid will be identified as potential marginal land for energy crops.

Step 3: Soil and topographic data processing. Soil pH and soil organic matter content data were obtained from the World Soil Database v1.21. The DEM data were derived from the Space Shuttle Radar Topography Mission. The DEM data and soil property data were used to determine the appropriate altitude, slope, soil organic matter, and soil pH data for the different bioenergy crops according to their suitable growth systems (Table 2).

Step 4: Based on the classification results obtained through the above three steps, the meteorological projections, land-use projections, and topographic and soil data of different years for different scenarios were overlaid.

3. Data records

The constructed dataset has a spatial resolution of 5 km and a time step of 5 years and covers the six SSP–RCP scenarios from 2020–2100 (Table 1). All the data are stored in the commonly used geo-tiff format with the WGS-84 coordinate system. Within each geo-tiff file, there are different socioeconomic and climatic scenarios: SSP3–7.0, SSP2–4.5, SSP1–2.6, SSP4–6.0, SSP4–3.4, and SSP1–1.91. Within each geo-tiff file, there are eight energy crops, namely, Ricinus communis L., Saccharum officinarum L., Pistacia chinensis Bunge, Panicum virgatum L., Jatropha curcas L., Miscanthus giganteus J., Manihot esculenta Crantz, and Sorghum bicolor Moench. For file naming and structure, all files with the same SSP–RCP scenario were grouped into the same folder, named “SSP-RCP”, and the same crops were included in the same folder, named with the corresponding crop name; for example, the file for willow millet data stored under SSP1–RCP1.9 was named “SSP1-RCP19_Panicum virgatum”, where “SSP” and “RCP” denote the SSP and RCP scenarios, respectively, and each geo-tiff file is named “Year”.tif, with “Year” denoting the year of the data.

4. Results

We first identified the potential amount and spatial distribution of marginal land in China in different socioeconomic and climatic contexts (Figure 2). Then, on the basis of known marginal land, according to the constraints of slope < 25°, organic matter content > 1.5%, annual precipitation >160 mm, and cumulative temperature > 2000°C agricultural suitability (Fu et al., 2020; Liu et al., 2015; Ministry of Natural Resources, 2020; Wang et al., 2009), we determined the potential amount and spatial pattern of marginal land suitable for agriculture in China (Figure 3). Finally, we obtained a spatially explicit dataset of the marginal land distribution for suitable energy crops in China under different climate scenarios for the period 2020–2100, allowing for the identification of potentially suitable marginal land after excluding existing natural reserves (Figure 4). The fluctuations in the total marginal land area generated after 2050 are shown in Figure 2. In the SSP1–1.9 scenario, sustainable development is a key objective. Governments are likely to implement land planning and management policies to ensure sustainable land use, such as limiting unwarranted urban sprawl, encouraging sustainable agricultural development, and protecting natural ecosystems. However, climate change may still have an impact on land use, e.g. changes in precipitation and temperature may affect agricultural production and land suitability. This leads to fluctuations in the marginal land area, and the agricultural suitability land and willow suitability land in Figures 3 and 4 are extracted from the marginal land and therefore will have the close fluctuating trend as marginal land does.

Figure 2. Potential marginal land distribution in China under the SSP1–1.9 scenario, 2020–2100.

Figure 3. Potential marginal land distribution suitable for agriculture in China under the SSP1–1.9 scenario, 2020–2100.

Figure 4. Potential marginal land distribution suitable for the energy crops Panicum virgatum L. in China under the SSP1–1.9 scenario from 2020–2100.

Figure 5 shows the change in suitable area for the different energy crops from 2020 to 2100 under the six considered climate scenarios. The range of suitability of different energy crops under the six scenarios is shown in Table 4.

Figure 5. Comparison of the temporal trends of land areas suitable for eight bioenergy crops under the six considered scenarios during 2020–2100.

Table 4. Range of suitable acreage for different energy crops under six climate scenarios from 2020 to 2100.

Energy crops	Range of suitable planting area (×10^4 km^2)
Ricinus communis L.	8.18–23.14
Saccharum officinarum L.	4.90–33.53
Pistacia chinensis Bunge	5.23–23.45
Panicum virgatum L.	37.37–73.97
Miscanthus giganteus J.	10.90–48.76
Jatropha curcas L.	1.90–16.28
Manihot esculenta Crantz	4.65–26.78
Sorghum bicolor Moench	29.71–69.70

Overall, the eight bioenergy crops do not appear to vary much in terms of their own suitable area under different scenarios until 2050; moreover, after 2050, the impact of different climate scenarios on the suitability of bioenergy crops will increase, and the differences in suitable regions will expand. Furthermore, the eight bioenergy crops were promoted under the SSP3–7.0, SSP4–6.0, and SSP2–4.5 scenarios, with the greatest increase occurring under SSP3–7.0; in particular, the suitable area for each bioenergy crop increased rapidly after 2060. We presume that SSP3–7.0 itself represents a relatively high level of social vulnerability and a relatively high level of forcing, accompanied by substantial land-use change (especially global woodland cover reduction). The development of all eight bioenergy crops was restricted under the SSP1–1.9, SSP1–2.6, and SSP4–3.4 scenarios. Among them, Ricinus communis L., Panicum virgatum L., and Sorghum bicolor Moench were the most affected by the SSP4–3.4 scenario, with suitable areas almost reaching their minimum values in different years under different scenarios. Moreover, Saccharum officinarum L., Pistacia chinensis Bunge, Jatropha curcas L., Miscanthus giganteus J., and Manihot esculenta Crantz were the most negatively affected under SSP1–1.9, and almost all of these plants exhibited minimum suitable areas in all years of all scenarios. Additionally, the eight bioenergy crops were less affected under SSP1–2.6, with little fluctuation in the suitable area between the starting and ending years.

Figure 6 shows boxplots indicating the suitable areas for the eight bioenergy crops under the different climate scenarios. Among the six scenarios, Jatropha curcas L. had the smallest range, and Panicum virgatum L. had the largest range. Compared to other five climate scenarios, SSP3–7.0 scenario exhibited the greatest fluctuations for the suitable area of all eight bioenergy crops. The median values of suitable area for Panicum virgatum L., Sorghum bicolor Moench, and Miscanthus giganteus J. were greater than those for the other crops due to their specific growth suitability; in other words, when promoting cultivation in different climate scenarios, the suitability of crops should be used as one of the references when selecting crops in some cases. The suitability of Ricinus communis L., Saccharum officinarum L., and Pistacia chinensis Bunge differed under the different climate scenarios. Ricinus communis L. and Pistacia chinensis Bunge had higher median values of suitable areas under future SSP1–1.9, SSP1–2.6, SSP2–4.5, and SSP3–7.0 climate scenarios. Under the SSP4–3.4 and SSP4–6.0 scenarios, the median values of the suitable areas for Ricinus communis L., Saccharum officinarum L., and Pistacia chinensis Bunge were similar. Moreover, the suitable areas for both Jatropha curcas L. and Manihot esculenta Crantz were low under the different climatic scenarios, which provided us with options for selecting energy crop varieties for cultivation and development.

Figure 6. Boxplots of the areas suitable for energy crops under different scenarios.

5. Discussion

5.1. Comparison to existing datasets of agricultural suitability of land

Although different datasets of agricultural suitability are difficult to compare due to different assumptions about land classes, different underlying data, differences in spatial resolution and thresholds, we can compare our dataset with the dataset of Schneider et al. (2022) estimates of global inventories of suitable, arable, and usable cropland under different scenarios and policies. Schneider et al. (2022) identified land ranges for 23 crops, first-generation bioenergy crops, and second-generation bioenergy crop suitability for the periods 2010–2039, 2040–2069, and 2070–2099. The inventory land use classification of available cropland from Schneider et al. (2022) excludes human impervious surfaces, forests, wetlands, and protected areas, assumptions most similar to our definition of marginal land. A comparison of the agricultural suitability of land in different periods under the RCP2.6 scenario with our data at different periods revealed that the distribution of our agricultural suitability locations overlapped well with that of Schneider et al. (2022) (distribution above moderate suitability); however, there was a large difference in area, and an important reason for this difference is that we excluded existing cropland, whereas Schneider et al. (2022) included existing cropland. By comparing their RCP2.6 scenario data focused on the suitability of land for the second-generation energy crop Panicum virgatum L. with our data from different periods, we found a significant overlap between the distribution of suitable land for Panicum virgatum L. in central and southern China and distribution indicating equal suitability in dataset provided by the Schneider et al., however with some differences in northern China and Xinjiang in northwestern China. In addition to the differences in the original agricultural suitability land, there are also differences in the energy crop suitability evaluation system between our study and the research conducted by Schneider et al. (2022). The choice of baseline and crop growth suitability thresholds may create uncertainty. According to Schneider et al. (2022) by 2100, under RCP2.6, the global extent of potential energy crops will increase by 5% compared to that in the historical period, with the largest increases in North America, Asia, and Russia at + 11% and + 12%, respectively, an increase largely due to global warming causing the agricultural frontier to move northwards. In our study in China, the range of potential energy crops increased by 12.2% overall compared to that in 2020.

5.2. Uncertainty and future research directions

In this study, we used a spatial resolution of 5 km for the land use prediction data. However, we must note that resolution limitations may produce certain biases in the determination of suitable planting boundaries for energy crops. In future studies, raw high-precision climate prediction data should be used and bias correction procedures should be implemented to improve the accuracy of the results and ensure a more accurate description of the range of suitability for energy crops. In addition, we recognize the need to consider the effects of rasterization when integrating factors such as socio-economic data to enhance a comprehensive assessment of suitable conditions for energy crops. We will work to delve more deeply into these aspects to ensure that our research is more comprehensive and robust. Due to the limitations of the research topic and data processing methods, we focused mainly on changes in marginal land suitability ranges for energy crops in this paper. However, this preliminary work provides the necessary background and data support for our future in-depth exploration. It is worthy to further investigate the mechanisms affecting energy crop utilization pathways to gain a more comprehensive understanding of the impact of different crop yields and fuel conversion rates on land use allocation decisions. This will further refine our research and provide additional insights and directions for future studies.

6. Conclusion

Our results demonstrate that China has a wide distribution of marginal land with unrivalled potential for the development of renewable biomass energy crops without competing with agriculture for land. Under the different climate–economic scenarios, the area of energy crop distribution increased more rapidly under the SSP3–7.0 scenario, followed by those under the SSP4–6.0 and SSP2–4.5 scenarios. Among the different energy crops, Panicum virgatum L., Sorghum bicolor Moench, and Miscanthus giganteus J. exhibited the greatest potential for expansion in terms of suitability. By mapping the potential spatial distribution of energy crops in marginal lands, we also highlight the importance of developing energy crops in marginal lands for ecological enhancement, the mitigation of energy scarcity, and the reduction of carbon emissions.

Competing interests

The authors declare no competing interests.

Data availability statement

The marginal land data presented here are freely available for download from Figshare (http://doi.org/10.6084/m9.figshare.23909754.v1), and the marginal land suitable for agriculture can also be downloaded from http://doi.org/10.6084/m9.figshare.23909724.v1. The marginal land suitable for eight energy crops in China under the six considered scenarios during 2020–2100 is available in the Science Data Bank at https://www.scidb.cn/en/detail?dataSetId=f056cfac1ab94be3b48ed29e8bd304c4.

Additional information

Funding

The work was supported by the National Natural Science Foundation of China [41971250]; the Postdoctoral Fellowship Program of CPSF [GZC20232621].

Notes on contributors

Jingying Fu

Jingying Fu is a professor and doctoral supervisor at the Institute of Geographic Sciences and Natural Resources Research (IGSNRR) of the Chinese Academy of Sciences (CAS), a member of the Youth Innovation Promotion Association of CAS, and a member of the “Bingwei” Outstanding Young Talent Programme of IGSNRR. Her research interests focus on the sustainable use of resources supported by space information technology and the spatial and temporal modelling and evaluation of renewable energy resource potentials and benefits.

Qiang Gao

Qiang Gao is a master’s degree candidate in resources and environment at the Institute of Geographic Sciences and Natural Resources Research of the Chinese Academy of Sciences. His primary research focus is the distributional extent of marginal land suitable for growing energy crops in China.

Gang Lin

Gang Lin is an associate professor at the Institute of Geographic Sciences and Natural Resources Research of the Chinese Academy of Sciences. His research interest focuses on the utilization of remote sensing to assess the suitability of marginal lands for cultivating energy crops.

Dong Jiang

Dong Jiang is currently a professor at the Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, and the director of Key Laboratory for Resources Use and Environmental Remediation (RUER). His current research interests focus on the monitoring and modelling of resources and environment, natural resources and sustainable development, and geographical big data analysis and artificial intelligence.

Yanan Zhao

Yanan Zhao received a Ph.D. degree from Ningxia University, China, in 2022. He is currently a postdoctoral fellow at Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences. His research interests include grassland ecology and geographical big data analysis in marginal land.

Shuang Lu

Shuang Lu is currently pursuing a Master’s degree at China University of Mining and Technology (Beijing) while also serving as a visiting graduate student at the Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences.