Elsevier

Journal of Cleaner Production
清洁生产学报

Volume 436, 10 January 2024, 140647
第 436 卷，2024 年 1 月 10 日，140647
Journal of Cleaner Production

Understanding the impact of cultivated land-use changes on China's grain production potential and policy implications: A perspective of non-agriculturalization, non-grainization, and marginalization
解读耕地利用变化对中国粮食生产潜力的影响及政策启示：非农化、非粮化和边缘化的视角

环境科学与生态学TOPEI检索SCI升级版 环境科学与生态学1区
https://doi.org/10.1016/j.jclepro.2024.140647 Get rights and content 获取权利和内容

Abstract 抽象的

Cultivated land plays a pivotal role in ensuring national food security. The challenges of non-agriculture, non-grain production and the marginalization of cultivated land have become increasingly pronounced in the context of urbanization and industrialization in China, posing potential threats to long-term food security. This study employs a combination of remote sensing and statistical data to reveal the effects of non-agriculturalization, non-grainization, and the marginalization of cultivated land on grain production potential (GPP) at the grid scale. From 1992 to 2020, the total farmland area initially exhibited an ascending trend followed by a subsequent decline, resulting in a net decrease of approximately 0.84% in China. the northward shift in the center of gravity of cultivated land has accentuated the spatial disparity between grain production and consumption. Non-agriculturalization, non-grainization, and marginalization of cultivated land contributed 35.22%, 22.58%, and 42.20% to the GPP loss, respectively. GPP loss exhibited significant differences across the agroclimatic zones, with the combined loss in the Middle-lower Yangtze Plain and Huang-Huai-Hai Plain exceeding 60% of the national total. Cultivated land marginalization has emerged as the most impactful land-use change affecting GPP, leading to a substantial loss of approximately 11471.30 × 104 t. The affected area is predominantly concentrated in the southern region. The marginalization of cultivated land has not been effectively alleviated, and the expansion trend continues. Conservation endeavors targeting cultivated land should prioritize enhancing farmers' willingness to cultivate, curbing the extensive utilization of farmland, and mitigating the escalating trend of farmland marginalization.
耕地在保障国家粮食安全中发挥着举足轻重的作用。在我国城镇化、工业化背景下，非农、非粮生产和耕地边际化的挑战日益凸显，对长期粮食安全构成潜在威胁。本研究采用遥感与统计数据相结合的方法，揭示网格尺度下非农化、非粮化、耕地边际化对粮食生产潜力（GPP）的影响。 1992年至2020年，我国耕地总面积先上升后下降，净减少约0.84%。耕地重心北移，粮食生产与消费的空间差距更加突出。非农化、非粮化和耕地边际化分别对GPP损失造成35.22%、22.58%和42.20%。不同农业气候区GPP损失差异显着，长江中下游平原和黄淮海平原损失合计超过全国60%以上。耕地边际化已成为影响GPP 影响最大的土地利用变化，导致耕地损失约11471.30 × 10 4 t。受影响地区主要集中在南部地区。耕地边际化并未得到有效缓解，扩张趋势仍在持续。 耕地保护工作要重点提高农民耕种意愿，遏制耕地粗放利用，减缓耕地边际化趋势加剧。

Keywords 关键词

Land-use change
Food production
Cropland protection
China

土地利用变化
食品生产
耕地保护
中国

1. Introduction 一、简介

Poverty eradication and zero hunger are the two major Sustainable Development Goals of the United Nations. The recently published Global Report on Food Crises 2023 by the Global Food Crisis Network reveals that more than 250 million individuals confront acute hunger, a predicament exacerbated by economic shocks and regional conflicts. Notably, only about one-third of the global population's demand is met by local food crop production on a global scale (Kinnunen et al., 2020). Amidst the compounding challenges of extreme weather, the COVID-19 pandemic, and regional conflicts intensifying food insecurity (Laborde et al., 2020; Liu et al., 2021; Pereira and Oliveira, 2020), the imperative of ensuring regional food self-sufficiency becomes particularly crucial (Bene, 2020; Brink et al., 2023). As one of the most populous countries globally, China, with only 7.8% of the world's arable land (Chai et al., 2019), bears a significant responsibility for grain production. In 2021, global cereal production reached 3070.6 billion kg, with China's contribution amounting to 633.8 billion kg, equating to approximately 21% of the world's cereal production (FAO, 2023). An International Monetary Fund report clarified that China's bumper grain had a positive effect on ensuring global food security (Wu et al., 2023).
Urbanization is widely acknowledged as a threat to food security (Gu et al., 2019). Urban expansion is expected to result in a reduction of 1.8–2.4% of global arable land by 2030. The associated food loss from these cropland reductions is estimated to have accounted for 3–4% of worldwide crop production in 2000, primarily affecting regions in Asia and Africa (Bren d'Amour et al., 2017). Over the last few decades, China has witnessed an unprecedented surge in urbanization, exerting considerable pressure on agricultural spaces (Li et al., 2017; Xu et al., 2020) and giving rise to a series of challenges in the effective utilization of cultivated land. Primarily, urban development has directly precipitated extensive farmland loss in the vicinity of cities (Liu et al., 2019; Zhou et al., 2020), resulting in the depletion of approximately 3.2 × 104 km2 of farmland from 2003 to 2016 alone (Qiu, B. et al., 2020). In contrast, the relentless growth of China's population has contributed to a substantial reduction in per capita arable land, heightening the need for a consistent and secure food supply (Ge et al., 2018). As consumption patterns and nutritional awareness have evolved, Chinese residents now exhibit an increased demand for a diverse range of foods, including vegetables, fruits, meat, eggs, milk, and other foods, shifting away from an exclusive reliance on grains for energy (Gu et al., 2019; Zhu and Begho, 2022). Driven by lucrative returns and evolving consumption trends, a considerable number of farmlands in China have transitioned away from grain production (Su et al., 2020; Zhang, D. et al., 2023). In the process of economic development, the rural labor force has migrated to cities, resulting in the extensive use of farmland (Liu et al., 2016) and even abandonment (Ren et al., 2023; Yan et al., 2016). While China's food security is currently guaranteed, long-term challenges persist, including tightening of water and land resources constraints for food production, structural inadequacies in the food supply, and regional disparities in food production (Liu and Zhou, 2021). The potential losses in grain production due to these changes in farmland utilization must not be underestimated.
城市化被广泛认为是对粮食安全的威胁（ Gu et al., 2019 ）。到 2030 年，城市扩张预计将导致全球可耕地减少 1.8-2.4%。耕地减少造成的粮食损失估计占 2000 年全球农作物产量的 3-4%，主要影响以下地区：亚洲和非洲（ Bren d'Amour 等，2017 ）。过去几十年来，中国城市化进程空前迅猛，给农业空间带来了巨大的压力( Li et al., 2017 ; Xu et al., 2020 )，也给耕地有效利用带来了一系列挑战。土地。首先，城市发展直接导致城市周边农田大量流失( Liu et al., 2019 ; Zhou et al., 2020 )，仅2003年至2016年就导致约3.2×10 4 km 2的耕地减少。 （ Qiu，B. 等人，2020 ）。相比之下，中国人口的持续增长导致人均耕地大幅减少，增加了对持续、安全的粮食供应的需求（ Ge et al., 2018 ）。 随着消费模式和营养意识的发展，中国居民对蔬菜、水果、肉、蛋、奶等食品的需求日益增加，不再完全依赖谷物作为能源（ Gu等人，2019 ；朱和贝戈，2022 ）。在丰厚回报和不断变化的消费趋势的推动下，中国相当多的农田已经不再从事粮食生产（ Su et al., 2020 ； Zhang, D. et al., 2023 ）。在经济发展过程中，农村劳动力向城市转移，导致耕地的大量利用（刘等，2016 ）甚至废弃（任志强等，2023 ；严等，2016 ）。目前，中国的粮食安全有保障，但长期挑战依然存在，包括粮食生产的水土资源约束趋紧、粮食供应结构性不足、粮食生产地区差异等（刘和周，2021 ）。耕地利用变化造成的粮食生产潜在损失不容低估。
To tackle the challenges confronting food production amidst economic development, the Chinese government has enacted a series of policy measures, such as the “Notice on Resolutely Stopping the Non-agriculturalization of Cultivated Land” (State Council, 2020) and the “A Guideline on Preventing Non-Grain Use of Arable Land and Stabilizing Grain Production to Ensure Food Security” (State Council, 2020). While governmental endeavors have resulted in some advancements in the protection of cultivated land the aforementioned issues still exist widely. Therefore, there is an immediate imperative to elucidate the threats posed to grain production by these changes in farmland utilization (Liu and Zhou, 2021).
为应对经济发展中粮食生产面临的挑战，中国政府出台了一系列政策措施，例如《关于坚决停止耕地非农化的通知》（国务院，2020）和《关于坚决遏制耕地非农化的通知》（国务院，2020）防止耕地非粮利用稳定粮食生产确保粮食安全”（国务院，2020）。尽管政府的努力在耕地保护方面取得了一些进展，但上述问题仍然普遍存在。因此，迫切需要阐明农田利用变化对粮食生产造成的威胁（刘和周，2021 ）。
Scholars have extensively investigated the repercussions of changes in cultivated land use on grain production during economic development (Wang, X. et al., 2020; Zhong et al., 2022). The competition between urban expansion and agricultural use has consistently been a focal point in the research on cultivated land-use change (Liu, 2018). The prevailing belief is that rapid urbanization inevitably results in a permanent loss of high-quality farmland (Chen et al., 2022; Tan et al., 2005). Over the period from 1992 to 2015, urban expansion led to an annual decline in Chinese grain production by 1245 × 104 t per year, causing a concomitant 2% decrease in the grain self-sufficiency rate per year (He et al., 2017). Since 2000, the rapid advancement of remote sensing technology has opened avenues for monitoring cropping patterns, gradually drawing attention to the impact of changes in cropping intensity on grain production (Han et al., 2022; Qiu et al., 2015; Yan et al., 2019). Reports indicate that the transition from double-to single-cropping rice in Southern China resulted in a 2.6% reduction in grain yield (Jiang et al., 2019). Additionally, extensive field investigations have found that the phenomena of non-grain crop planting and land abandonment are becoming increasingly common (Peng et al., 2021; Qiu, T. et al., 2020; Ren et al., 2023); however, these studies often focus on specific regions and samples. Regional-scale monitoring of non-grain production and cultivated land abandonment has gradually gained attention in recent years (Han and Song, 2020; Su et al., 2019, 2020). For instance, Zhang, D. et al. (2023) revealed the spatiotemporal characteristics of non-grain production on cultivated land in the Guanzhong Region. Zhang, M. et al. (2023) utilized land-use data to depict cropland abandonment over the past three decades in China. Despite partial exploration of non-grainization and abandonment, their threats to food production have not been quantified, and the full extent of their impact on grain production has not been clarified. Moreover, these studies focused on specific aspects or types of farmland changes, lacking a comprehensive evaluation and systematic calculation of their overall effects on food security. Further clarification of the impact on grain production and spatial heterogeneity associated with various cultivated land-use changes during the economic development process can guide the formulation of farmland management strategies and the exploration of effective pathways for increasing food production.
学者们广泛研究了经济发展过程中耕地利用变化对粮食生产的影响（王X等，2020 ；钟等，2022 ）。城市扩张与农业利用之间的竞争一直是耕地利用变化研究的焦点（刘，2018 ）。人们普遍认为，快速的城市化不可避免地会导致优质农田的永久丧失（ Chen等，2022 ； Tan等，2005 ）。 1992年至2015年，城市扩张导致中国粮食产量每年减少1245×10 4 t，粮食自给率每年下降2%（何等，2017） ）。 2000年以来，遥感技术的快速发展为种植模式监测开辟了途径，逐渐引起人们对种植强度变化对粮食生产的影响的关注(韩等，2022 ；邱等，2015 ；严等，2015) ., 2019 ).有报告显示，中国南方地区从双季稻转向单季稻导致粮食产量下降 2.6%（ Jiang et al., 2019 ）。 此外，广泛的实地调查发现，非粮作物种植和荒地现象日益普遍(彭等，2021 ；邱涛等，2020 ；任志强等，2023 )；然而，这些研究通常集中于特定区域和样本。近年来，区域尺度的非粮生产和耕地撂荒监测逐渐受到关注(韩和宋，2020 ；苏等，2019，2020 ) 。例如， Zhang, D. 等人。 (2023)揭示了关中地区耕地非粮生产的时空特征。张，M.等人。 （2023）利用土地利用数据来描述中国过去三十年的耕地撂荒情况。尽管对非粮化和抛荒进行了部分探索，但其对粮食生产的威胁尚未量化，其对粮食生产的影响程度尚未明确。而且，这些研究侧重于农田变化的具体方面或类型，缺乏对其对粮食安全的总体影响进行全面评估和系统计算。 进一步明确经济发展过程中各种耕地利用变化对粮食生产的影响和空间异质性，可以指导农地管理策略的制定和粮食增产的有效路径探索。
Currently, China is still in the process of accelerating urbanization, and the aforementioned changes in cultivated land use may become more common. In light of the existing problems of cultivated land use during China's urbanization since 1990, this study delves into the impact of three specific aspects of cultivated land changes as non-agriculturalization, non-grainization, and marginalization on grain production potential (GPP) at the grid scale by combining remote-sensing and statistical data. The primary objectives of this study were to (1) elucidate the magnitude of the impacts of non-agriculturalization, non-grainization, and marginalization on the GPP, (2) unveil the spatial heterogeneity characterizing the effects of these cultivated land-use changes on GPP, and (3) discuss policy implications for cropland protection in light of the current situation.
当前，我国仍处于城镇化加速进程中，上述耕地利用变化可能会更加普遍。针对1990年以来我国城镇化进程中存在的耕地利用问题，本研究深入探讨了非农化、非粮化、边际化三个具体方面的耕地变化对我国城镇化进程中粮食生产潜力（GPP）的影响。通过结合遥感和统计数据来确定网格规模。本研究的主要目标是（1）阐明非农化、非粮食化和边缘化对 GPP 的影响程度，（2）揭示空间异质性，表征这些耕地利用变化对 GPP 的影响。 GPP，(3)根据当前形势讨论耕地保护的政策影响。

2. Data sources and methods
2 数据来源和方法

2.1. Study area 2.1.研究区

This study included 31 Chinese provinces and municipalities (excluding Hong Kong, Macao, and Taiwan). According to the agricultural production conditions and climate, the study area was divided into nine agricultural climate zones (Fig. 1): Northeast China Plain (NCP): includes Liaoning, Jilin, and Heilongjiang; North Arid and Semiarid Region (NASR): Gansu, Ningxia, Xinjiang, and Inner Mongolia; Huang-Huai-Hai Plain (HHHP): Beijing, Tianjin, Hebei, Shandong, and Henan; Loess Plateau (LP): Shaanxi and Shanxi; Qinghai-Tibet Plateau (QTP): Qinghai and Tibet; Middle-lower Yangtze Plain (MLYP): Shanghai, Jiangsu, Zhejiang, Anhui, Jiangxi, Hubei, and Hunan; Sichuan Basin (SCB): Chongqing and Sichuan; Yunnan-Guizhou Plateau (YGP): Guangxi, Guizhou, and Yunnan; Southern China (SC): Guangdong, Hainan, and Fujian.
本研究包括中国31个省市（不包括香港、澳门和台湾）。根据农业生产条件和气候，将研究区划分为9个农业气候带（图1 ）： 东北平原（NCP）：包括辽宁、吉林、黑龙江；北方干旱半干旱地区（NASR）：甘肃、宁夏、新疆、内蒙古；黄淮海平原（HHHP）：北京、天津、河北、山东、河南；黄土高原（LP）：陕西、山西；青藏高原（QTP）：青海和西藏；长江中下游平原（MLYP）：上海、江苏、浙江、安徽、江西、湖北、湖南；四川盆地（SCB）：重庆、四川；云贵高原（YGP）：广西、贵州、云南；华南地区 (SC)：广东、海南和福建。
Fig. 1
  1. Download: Download high-res image (538KB)
    下载：下载高分辨率图像 (538KB)
  2. Download: Download full-size image
    下载：下载全尺寸图像

Fig. 1. Distribution map of agricultural climate zones in China.
图1 .中国农业气候带分布图。

2.2. Data collection and preprocessing
2.2.数据收集和预处理

The data included the European Space Agency Climate Change Initiative land use/cover (ESACCI_LC) from 1992 to 2020, China's land-use/cover datasets (CLUDs) provided by the Resource and Environmental Science and Data Center, the MODIS Enhanced Vegetation Index (EVI), potential crop yield data, and statistical data (Table 1).
数据包括欧洲航天局气候变化倡议1992年至2020年土地利用/覆盖（ESACCI_LC）、资源环境科学数据中心提供的中国土地利用/覆盖数据集（CLUD）、 MODIS增强植被指数（EVI） ）、作物潜在产量数据和统计数据（表1 ）。

Table 1. Data description.
表 1 .数据说明。

Data type 数据类型Time range 时间范围Spatial resolution 空间分辨率Source 来源
ESACCI_LC1992–2020300 m 300米http://climate.esa.int/en/projects/land-cover
CLUDs CLUD1990, 1995, 2000, 2005, 2010, 2015, 20201 km 1公里https://www.resdc.cn/
EVI (MOD13Q1) 房屋（MOD13Q1）2000–2020250 m 250米https://ladsweb.modaps.eosdis.nasa.gov/
Potential crop yield 作物潜在产量20101 km 1公里https://www.resdc.cn/
Statistical data 统计数据1990–2020–China Statistical Yearbook and China Rural Statistical Yearbook
中国统计年鉴和中国农村统计年鉴
Agricultural climate zones
农业气候区		––https://www.resdc.cn/
ESACCI_LC has a spatial resolution of 300 m and time coverage from 1992 to 2020, making it the longest continuous land-use product available worldwide. The product was classified in accordance with the Food and Agriculture Organization of the United Nations (FAO) Land Cover Classification System (LCCS), which includes 36 land-use types.
ESACCI_LC 的空间分辨率为 300 m，时间范围为 1992 年至 2020 年，使其成为全球范围内最长的连续土地利用产品。该产品根据联合国粮食及农业组织 (FAO )土地覆盖分类系统 (LCCS) 进行分类，其中包括 36 种土地利用类型。
The CLUDs provide detailed land-use information in China every five years from 1990 to 2020. The dataset was constructed based on Landsat MSS, TM/ETM, and Landsat8 satellite remote sensing data using man–machine interactive visual interpretation. It consists of 6 level-1 classes (cropland, forest, grassland, water, built-up area, and barren land) and 25 level-2 classes.
CLUD从1990年到2020年每五年提供一次中国详细的土地利用信息。该数据集基于Landsat MSS、TM/ETM和Landsat8卫星遥感数据，采用人机交互式视觉解释构建。包括耕地、森林、草原、水域、建成区、荒地6个一级类和25个二级类。
EVI data from 2000 to 2020 were obtained from the MOD13 Q1 dataset with a spatial resolution of 250 m and a temporal resolution of 16 days. These data were used to construct an multiple crop index (MCI) (Supplementary Text 1) dataset covering 2000 to 2020. MODIS data began in February 2000, and data for January 2000 were missing. Therefore, data from January 2001 were used to fill this gap.
2000年至2020年的EVI数据来自MOD13 Q1数据集，空间分辨率为250 m，时间分辨率为16天。这些数据用于构建涵盖2000年至2020年的多种作物指数（MCI）（补充文本1）数据集。MODIS数据始于2000年2月，2000年1月的数据缺失。因此，使用 2001 年 1 月的数据来填补这一空白。
Statistical data on the total grain yield and the proportion of the sown area of grain crops for each province from 1990 to 2020 were obtained from the China Statistical Yearbook and the China Rural Statistical Yearbook.
1990年至2020年各省粮食总产量和粮食作物播种面积比重统计数据来源于《中国统计年鉴》和《中国农村统计年鉴》。

2.3. Research methods 2.3.研究方法

Since the 1990s, rapid urbanization in China has significantly affected the utilization of cultivated land. Approximately 300 × 104 hm2 of high-quality cultivated land was occupied by construction land between 1996 and 2009 (Kong, 2014). Meanwhile, with the continuous migration of the rural labor force, phenomena such as the extensive use and abandonment of cultivated land have become increasingly obvious (Wang et al., 2020, Wang et al., 2020Wang, Y. et al., 2020; Xu et al., 2019). In response to the primary issues related to farmland utilization in China, and based on the modes of cultivated land-use conversion, we categorize these changes into three types (viz. non-agriculturalization, non-grainization, and marginalization) and systematically analyzed their impact on GPP (Fig. 2).
20世纪90年代以来，中国快速的城市化进程显着影响了耕地的利用。 1996年至2009年间，约300×10 4 hm 2优质耕地被建设用地占用（ Kong，2014 ）。同时，随着农村劳动力的不断迁移，耕地的粗放利用和废弃等现象也日益明显(王等，2020 ，王等，2020王Y.等，2020徐等人，2019 ）。针对我国耕地利用的主要问题，根据耕地非农化的模式，我们将其分为非农化、非粮化、边缘化三种类型，并系统分析了其影响。对GPP的影响（图2 ）。
Fig. 2
  1. Download: Download high-res image (790KB)
  2. Download: Download full-size image

Fig. 2. Flowchart for estimating the impact of non-agriculturalization, non-grainization, and marginalization of cultivated land on GPP.

2.3.1. Estimation method for the impact of non-agriculturalization on GPP

“Non-agriculturalization” of cultivated land refers to the conflict between the demand for urban construction and the protection of cultivated land, which is mainly reflected in the occupation of cultivated land by urban expansion. This study extracted the regions where construction land expansion occupied cultivated land based on CLUDs and estimated the loss of GPP for each pixel (van Vliet et al., 2017). The formula is as follows:(1)GPP1=∑i=1nSi∙MCIi∙Yiwhere GPP1 represents the loss of GPP caused by non-agriculturalization, Si is the area of the ith cultivated land converted into construction land pixels, MCIi is the mean value of the MCI of the ith pixel over multiple years, and Yi is the potential crop yield per unit area of the ith pixel; i = 1,2,3, …The purpose of using the multi-year average of the MCI was to exclude technological factors, such as fertilizers, seeds, agricultural machinery, and the effects of climate change on grain production.

2.3.2. Estimation method for the impact of non-grainization on GPP

“Non-grainization” of cultivated land mainly refers to the use of farmland for planting economic crops such as vegetables and fruits to replace food production (Liang et al., 2022; Xiao et al., 2015). Two aspects of nongrain use were also considered. First, the loss of GPP owing to the planting of cash crops was estimated by analyzing the changes in the proportion of sown areas for grain crops. Second, we estimated the loss of GPP due to the conversion of farmland into orchards (including fruit orchards, mulberry orchards, tea gardens, and other types of plantations) using land use transfer analysis. The formula is as follows:(2)G1=∑j=1mΔRj∙CSAj∙Yj(3)G2=∑i=1nAi∙MCIi∙Yi(4)GPP2=G1+G2where GPP2 represents the loss of GPP caused by non-grainization; G1 and G2 represent the loss of GPP caused by planting cash crops and converting farmland into orchards, respectively; ΔRj represents the change in the proportion of sown area for grain crops in the jth province; CSAj represents the annual average crop-sown area in the jth province; Yj represents the average value of the potential crop yield per unit area in the jth province; j = 1,2,3, …,31; and Ai represents the area of the ith pixel that has been converted into orchards. Due to the lack of reliable remote sensing data on grain-sown proportions, the loss of GPP caused by planting cash crops was estimated based on provincial administrative units.

2.3.3. Estimation method for the impact of marginalization on GPP

“Marginalization” of cultivated land can be divided into recessive and dominant. Recessive marginalization of cultivated land refers to the reduction in production factors and weakening of farmland utilization intensity, such as a decrease in MCI and fertilizer input. Dominant marginalization of cultivated land refers to changes in land use, such as the conversion of cultivated land to forest land or pasture, and abandonment is an extreme manifestation of farmland marginalization. In this section, we reveal the impact of the marginalization of cultivated land on GPP through the reduction in MCI and land abandonment (Guo et al., 2023; Jiang et al., 2019).
  • (1)
    Method for estimating the impact of MCI reduction on GPP
The traditional MCI, based on administrative divisions, was calculated using national agricultural census data (including grain-sown and cultivated land areas). However, different data sources have led to significant differences in China's cultivated land survey data since 1990. Additionally, the MCI calculated based on statistical data did not exclude fallow land, resulting in an underestimation of the calculated index. Therefore, in our study, a remote-sensing monitoring method was used to generate an annual MCI dataset. This method was based on a recent study by Qiu et al. (2022), which generated cropping intensity based on phenology-based mapping algorithms with pixel purity-based thresholds and effectively improved the mixed-pixel problem of MODIS images. This study provides an open-processing code that will not be repeated here. Because of the historic amounts of rain caused by the 2020 monsoon in China, the effective observations were lower (Qiu et al., 2022). Meanwhile, the EVI data used in our study started in 2000; therefore, we used the period from 2000 to 2019 for the calculation. The formula for estimating GPP loss due to a decrease in MCI is as follows:(5)G3=∑i=1n(MCI2000−MCI2019)∙CAi∙Yiwhere G3 represents the loss of GPP caused by the decrease in MCI; MCI2000 and MCI2019 represent the MCI values of the ith pixel in 2000 and 2019, respectively; CAi is the area of the ith pixel. We focused only on the loss of GPP caused by a decrease in MCI and part of the increase in MCI was excluded.
  • (2)
    Estimating the impact of farmland abandonment on GPP
Farmland abandonment refers to the natural succession of idle farmlands. Considering the characteristics of crop rotation in China, we adopted the definition of abandoned farmland from the International Workshop on Land Consolidation and Land Banking in 2011 as farmland that had not been cultivated for at least two consecutive years. Based on the ESACCI_LC dataset, we analyzed the land-use trajectory of each pixel and identified abandoned pixels that had been classified as non-farmland for at least two consecutive years. The year of the first conversion to non-farmland was defined as the start of abandonment and the duration of abandonment was recorded. Only farmland converted to natural cover was included and not farmland converted to impervious surfaces, water bodies, glaciers, or snow. Additionally, the vegetation succession sequence after farmland abandonment should include annual and biennial weeds, perennial grasses, woody plants, and successional species. Without human disturbance, succession from bare land to weeds takes approximately 1–3 years, whereas small shrubs require at least five years, and trees require at least ten years (Romero-Díaz et al., 2017). Only the natural abandonment of farmland was recorded here, while government-induced farmland abandonment that was converted into garden land or forest within two years is discussed in Section 2.3.2. Using the ESACCI_LC dataset from 1992 to 2020, the identified abandoned map was from 1993 to 2019. Based on this, the cumulative area of farmland abandoned over the past years was calculated depending on the year of abandonment initiation and duration. Combined with the MCI and potential crop yield data, the GPP loss caused by farmland abandonment was estimated using the following formula:(6)G4=∑i=1nALAi∙MCIi∙Yiwhere G4 represents the GPP loss caused by farmland abandonment and ALAi is the area of the ith abandoned pixel.

3. Results

3.1. Spatiotemporal changes of cultivated land

The temporal and spatial changes in cultivated land in China from 1992 to 2020 are shown in Fig. 3. Overall, the cultivated land area first increased and then decreased. From 1992 to 2000, cultivated land increased by approximately 1.99% (a total of 552.58 × 104 hm2), but after 2000, the cultivated land area decreased continuously, accounting for approximately 2.82% (a total of 785.02 × 104 hm2) of the 1992 level. The changes in the total amount of cultivated land varied across the nine agricultural climate zones. The regions with an increase in cultivated land included NASR, NCP, MLYP, and SC. During the study period, the cultivated land in NCP and NASR continued to expand, and NASR cultivated land expanded the most, at approximately 273.17 × 104 hm2. However, both MLYP and SC showed an increasing trend before 2000 and a decreasing trend after 2000. The regions with decreased cultivated land included HHHP, LP, YGP, SCB, and QTP. In these regions, cultivated land experienced a sustained decline, with HHHP decreasing by approximately 8.26%, especially.
Fig. 3
  1. Download: Download high-res image (1011KB)
  2. Download: Download full-size image

Fig. 3. 1992–2020 change of cultivated land area in China and nine agricultural climate zones (ratio of annual area and multi-year averaged area).

To further clarify the spatiotemporal characteristics of farmland, grain production, and consumption, we drew maps of farmland, grain production, and the population center of gravity, as shown in Fig. 4. From 1992 to 2020, the cultivated land center of gravity moved overall in the northwest direction by approximately 18.94 km, with a westward movement of 9.3 km and a northward movement of 16.5 km. Meanwhile, the grain production center of gravity moved northeastward by 252.57 km. Overall, there was a significant northward trend in both the farmland and food production centers of gravity. In contrast, the population center of gravity showed a southeastward movement, with a total shift of approximately 50.22 km, indicating a worsening of the dual disconnect phenomenon in the spatial distribution of China's food production and consumption.
Fig. 4
  1. Download: Download high-res image (353KB)
  2. Download: Download full-size image

Fig. 4. Spatial shift of cultivated land (a), grain production (b), and population (c) gravity center from 1990 to 2020.

3.2. Influence of non-agriculturalization on GPP

From 1990 to 2020, the non-agriculturalized area in China reached 1512.73 × 104 hm2, with a non-agriculturalization rate of approximately 8.54%. Most non-agriculturalized areas were distributed in the eastern region, particularly HHHP and MLYP, which accounted for 63% of the national non-agriculturalized area (Fig. 5). From the perspective of different periods, the non-agriculturalization area of farmland was the largest from 2010 to 2020, reaching 1432.68 × 104 hm2, followed by that in 1990–2000, which was 1052.47 × 104 hm2, and the smallest was from 2000 to 2010, which was 186.66 × 104 hm2. In addition, regardless of the period, non-agriculturalization in HHHP and MLYP was the most severe, with proportions exceeding 60%.
Fig. 5
  1. Download: Download high-res image (2MB)
  2. Download: Download full-size image

Fig. 5. Area of non-agricultural farmland from 1990 to 2020.

Combined with the remote sensing monitoring results of the MCI over the years and potential crop yield data, the loss of GPP caused by the non-agriculturalization of cultivated land was estimated, and a 10 km × 10 km grid was utilized to calculate the amount of GPP loss within 100 km2, as shown in Fig. 6. Non-agriculturalization of cultivated land in China resulted in a total GPP loss of 9573.19 × 104 t from 1990 to 2020. The spatial distribution was similar to that of nonagricultural farmlands, mainly in the eastern region. There were significant differences in the GPP losses in each agricultural zone. The largest losses in GPP of 4223.71 × 104 t and 3256.08 × 104 t were observed in HHHP and MLYP, respectively, accounting for over 78% of the total losses in the country. In contrast, economic development in the western region was relatively slow and GPP losses resulting from urban construction were relatively low. Specifically, in the QTP region, the potential loss was only 7.22 × 104 t.
Fig. 6
  1. Download: Download high-res image (989KB)
  2. Download: Download full-size image

Fig. 6. Loss of GPP due to non-agriculturalization of cultivated land in each 10 km × 10 km grid from 1990 to 2020.

3.3. Influence of non-grainization on GPP

3.3.1. Loss of GPP due to the planting of cash crops

Fig. 7 shows the changes in the proportion of sown areas for grain crops in China and the nine agricultural zones from 1990 to 2020. During this period, the proportion of the sown area for grain crops in China first decreased and then increased, continuously declining from 76.50% in 1990 to 65.22% in 2003, and then fluctuating and rebounding to 69.7% in 2020. Overall, the proportion of the sown area for grain crops decreased by approximately 6.8%, showing a clear trend for farmland non-grain. Among the agricultural zones, only NCP showed a fluctuating upward trend in the proportion of the sown area for grain crops, increasing by approximately 3.73%, whereas the other eight zones experienced different degrees of decline. The proportion of sown area for grain crops in SC, QTP, and YGP decreased by over 20%, whereas those of NASR and SCB decreased by over 15%. Most zones experienced a continuous decline in the proportion of sown area for grain crops during the study period, especially in the southern zones, such as SCB, YGP, and SC. In a few zones, such as NCP, HHHP, and MLYP, the proportion of sown area for grain crops rebounded after 2005, while the declining rates of the sowing ratios in other zones also slowed, indicating the initial effectiveness of the Chinese government's protection measures for farmland since 2005, while the phenomenon of non-grain cultivation on farmland persisted.
Fig. 7
  1. Download: Download high-res image (796KB)
  2. Download: Download full-size image

Fig. 7. Proportion of sown area for grain crops in China and nine agricultural climate zones from 1990 to 2020.

The effects of planting cash crops on the GPP in each region were estimated based on changes in the proportion of sown areas for grain crops (Fig. 8). Approximately 5744.48 × 104 t of GPP was lost due to the planting of cash crops nationwide, which was mainly distributed in the south. The GPP loss in YGP was the largest, at approximately 1461.88 × 104 t. Additionally, during the study period, the proportion of the sown area for grain crops in NCP increased, which correspondingly increased the GPP by approximately 515.27 × 104 t.
Fig. 8
  1. Download: Download high-res image (407KB)
  2. Download: Download full-size image

Fig. 8. Changes in GPP due to the planting of cash crops from 1990 to 2020.

3.3.2. Loss of GPP due to conversion to orchards

Fig. 9 shows the area of cultivated land transformed into orchards in China and the nine agricultural zones from 1990 to 2020. From 1990 to 2020, approximately 107.21 × 104 hm2 of farmland was transformed into orchards nationwide, with more areas being transformed in the south than in the north. The transformation area of different agricultural zones varied greatly, among which SC was the largest, reaching 30.19 × 104 hm2, while QTP only had 1.66 × 104 hm2. From the perspective of different periods, the transformation of cultivated land into orchards was most significant from 2010 to 2020, with 121.95 × 104 hm2 of farmland being used for the production of economic crops, such as fruit trees and tea, directly leading to a reduction in GPP. The period from 2000 to 2010 saw the smallest transformed areas, with only approximately 24.54 × 104 hm2. During this decade, LP had the largest area of transformation, accounting for 43.03% of the entire country. From 1990 to 2000, the area of transformation was approximately 101.91 × 104 hm2. During these three periods, the transformation area in SC was relatively large, especially in 1990–2000 and 2010–2020, far exceeding that in other regions. Regardless of the period, the area of transformation in QTP was the smallest.
Fig. 9
  1. Download: Download high-res image (347KB)
  2. Download: Download full-size image

Fig. 9. Area of cultivated land transformed into orchards in China and nine agricultural climate zones from 1990 to 2020 (unit: 104 hm2).

Further calculation results showed that the conversion of cultivated land to orchards resulted in a GPP loss of 394.55 × 104 t, which was mainly distributed in the southern and eastern regions (Fig. 10). West China is dominated by mountains, deserts, and plateaus with little arable land. It is not the main area of fruit, tea, and other crop production; therefore, few areas of cultivated land were transformed into orchards. SC and MLYP had the highest number of transformation areas and the largest GPP loss (99.70 × 104 t and 92.98 × 104 t, respectively), accounting for 25% and 24% of the national total, respectively. These areas have good water and heat conditions and are suitable for fruit trees and other cash crops. These areas are also the center of China's economic development and the main population-gathering areas, which expands the consumption market for economic crops and encourages local farmers to engage in non-grain production.
Fig. 10
  1. Download: Download high-res image (769KB)
  2. Download: Download full-size image

Fig. 10. Loss of GPP due to conversion to orchards in each 10 km × 10 km grid from 1990 to 2020.

3.4. Influence of marginalization on GPP

3.4.1. Loss of GPP due to MCI reduction

From 2000 to 2019, there was a significant spatial change in cropping intensity in China. As shown in Fig. 11, changes in cropping intensity mainly occurred during the conversion between single and double cropping. Approximately 1912.39 × 104 hm2 of double-cropping land was converted to single-cropping land, which was mainly distributed in MLYP and YGP.
Fig. 11
  1. Download: Download high-res image (2MB)
  2. Download: Download full-size image

Fig. 11. Spatial distribution of cropping intensity conversion from 2000 to 2019.

The loss of GPP caused by MCI reduction was calculated pixel-by-pixel based on the spatiotemporal variation in MCI and potential crop yield data, and the results are shown in Fig. 12. The GPP loss caused by MCI reduction is common nationwide, and the impact of MCI reduction on GPP in various agricultural regions is significant. Overall, GPP loss caused by MCI reduction was approximately 10655.34 × 104 t in China. In terms of spatial distribution, the GPP loss gradually increased from west to east. The GPP loss caused by the decline in MCI was significantly different among agricultural zones, and MLYP had the largest potential loss at 4505.92 × 104 t, accounting for approximately 42% of the country's total. The weakening of cropping intensity in China is believed to be mainly concentrated in the southern region, especially in the conversion of double-to single-cropping rice in the southern region, which has put great pressure on grain production (Jiang et al., 2019). Similar results were also observed in this study, the total GPP loss caused by MCI reduction in the southern region is greater than that in the northern region. However, a strong threat of declining MCI was also observed in the northern plain, such as in HHHP. As a major grain-producing area in China, HHHP bears significant responsibility for national food security, and the decline in cropping intensity in the north deserves more attention.
Fig. 12
  1. Download: Download high-res image (1022KB)
  2. Download: Download full-size image

Fig. 12. Loss of GPP due to MCI reduction in each 10 km × 10 km grid from 2000 to 2019.

3.4.2. Loss of GPP due to farmland abandonment

  • (1)
    Overall farmland abandonment trends
We constructed land-use trajectories for each pixel base year of the land-use data and constructed an abandoned farmland map of China from 1993 to 2019. From the annual increase in abandoned farmland, a total of 1063.88 × 104 hm2 of farmland was abandoned from 1993 to 2019, accounting for approximately 25% of the total farmland area in 2020 (Fig. 13a). The abandoned farmland of newly added farmland showed a trend of fluctuating increase, with the highest amount of newly abandoned farmland occurring in 2016 at approximately 177.66 × 104 hm2, and the abandonment rate was 4.18%. Previous studies have widely acknowledged that there is a certain level of error in the remote sensing-based identification of farmland abandonment time, and acceptable results for fallow identification can be obtained by relaxing the identification error to 2–3 years (Yin et al., 2018, 2020). Therefore, to further understand the temporal trend of farmland abandonment, we aggregated newly abandoned data every year into three-year intervals. For example, abandoned pixels that occurred in 1993, 1994, or 1995 were aggregated into the category “1993–1995”. The newly abandoned farmland area in China exhibited an N-shaped change based on the interannual changes in abandoned farmland after aggregation. The newly abandoned farmland area remained low during 1993–1995 and 1996–1998 at 47.50 × 104 hm2 and 48.16 × 104 hm2, respectively. Subsequently, it increased to 200.94 × 104 hm2 in 1999–2001, reaching its first peak of farmland abandonment. In the following decade, the expansion of farmland abandonment was somewhat controlled, gradually declining to 50.00 × 104 hm2 in 2011–2013, and the new farmland abandonment reached a new peak during 2017–2019, at approximately 248.05 × 104 hm2.
Fig. 13
  1. Download: Download high-res image (487KB)
  2. Download: Download full-size image

Fig. 13. Interannual changes of newly abandoned farmland area (a) and accumulated farmland abandoned land area (b).

From the perspective of the cumulative area, the total area of abandoned farmland increased from 1993 to 2020 (Fig. 13b). Large-scale abandonment of farmland was relatively rare in 1993, and the small size of abandoned patches resulted in them being overlooked in remote-sensing image recognition at a scale of 300 m. Statistics show that the total area of abandoned arable land in 1993 was 0. However, the total area of abandoned farmland began to gradually increase from 1994, growing from 47.29 × 104 hm2 to 1034.01 × 104 hm2 in 2020. The annual change in the cumulative area of abandoned farmland can be divided into three stages: the first stage is from 1993 to 2002, when the total area of abandoned land increased swiftly, with a growth rate of 41.82 × 104 hm2/yr; the second stage is from 2002 to 2013, during which the total abandoned area increased slowly at an annual average rate of 22.23 × 104 hm2/yr; and the third stage is from 2013 to 2020, during which the total abandoned area increased rapidly at a rate of 59.00 × 104 hm2/yr.
  • (2)
    Durations of farming cessation between 1993 and 2019
We determined the duration of farmland abandonment in China and nine agricultural zones (Fig. 14). Cessation of agricultural activity for 2–5 years was the most common form of abandonment, accounting for approximately 39.13% of the total, followed by 16–20 years (accounting for 23.9%), and farmland abandoned for more than 20 years (19.32%). The duration of farmland abandonment had different characteristics in different regions, but overall, abandonment for 2–5 years was the most common in all agricultural zones, except in NASR and QTP.
Fig. 14
  1. Download: Download high-res image (288KB)
  2. Download: Download full-size image

Fig. 14. Farming cessation durations between 1993 and 2019.

  • (3)
    Spatiotemporal distribution of farmland abandonment
We constructed annual maps of abandoned farmland to identify spatiotemporal distribution patterns and calculated the area of abandoned farmland within 10 km grids for three sub-periods (i.e., 1993–2000, 2000–2010, and 2010–2019) (Fig. 15). The total amount and spatial extent of abandoned farmland gradually increased over the three sub-periods. Significant spatial heterogeneity was observed in the distribution of abandoned farmland during the three periods. From 1993 to 2000, abandoned farmland was mainly located on both sides of the Hu Huanyong Line (Supplementary Text 2), with more abandoned farmland in the east than in the west. From 2000 to 2010, abandoned farmland expanded outward from previously abandoned areas, and the amount of abandoned farmland increased in both the east and west. During this period, the amount of abandoned farmland on both sides of the Hu Huanyong Line was approximately equal. From 2010 to 2019, the phenomenon of abandoned farmland became more widespread, with a southern shift in the pattern of abandoned farmland, particularly in YGP, where there was a significant increase in the amount of abandoned farmland.
Fig. 15
  1. Download: Download high-res image (1MB)
  2. Download: Download full-size image

Fig. 15. Area of abandoned farmland within a 10 km × 10 km grid between 1993 and 2019.

  • (4)
    Effects of farmland abandonment on GPP
Farmland abandonment directly led to a decrease in the grain-sown area and GPP. Based on the identification results of abandoned farmland, combined with the MCI and potential crop yield, we estimated the GPP loss caused by abandoned farmland (Fig. 16). Farmland abandonment led to a decline of 815.96 × 104 t in the national GPP from 1992 to 2019, which was mainly distributed in the southern, central, and northeastern regions. Especially in YGP, farmland abandonment significantly impacted grain production, directly leading to a GPP loss of 183.18 × 104 t. QTP was least affected by farmland abandonment, with a GPP loss of 4.92% of the country.
Fig. 16
  1. Download: Download high-res image (1MB)
  2. Download: Download full-size image

Fig. 16. Loss of GPP due to farmland abandonment in each 10 km × 10 km grid from 1993 to 2020.

4. Discussion

  • (1)
    Main changes in cultivated land use affecting China's GPP
In recent years, propelled by the expansion of farmland in northern China, advancements in crop mechanization, crop optimization, and enhanced planting techniques (Zhong et al., 2022), China has achieved an equilibrium between grain supply and demand (Xia et al., 2022). Nonetheless, the potential ramifications of cultivated land use changes on the GPP warrant attention. Urbanization, at the expense of cultivated land, poses a substantial challenge to grain production (Bren d'Amour et al., 2017). Estimates indicate that the global non-agriculturalization of cultivated land from 2000 to 2040 could result in an approximate loss of 65 million tons of crop production, with China contributing the highest proportion (van Vliet et al., 2017). In China, two-thirds of urbanization directly resulted in the loss of cultivated land, of which 37% was of high quality (Qiu, B. et al., 2020). Our findings also confirmed that non-agriculturalization of farmland still dominated the impact on grain production during 1990–2020, leading to a GPP loss of approximately 9573.19 × 104 t (Fig. 17). It is noteworthy that the decline in cropping intensity due to extensive cultivated land management gradually had a pronounced impact on GPP. On a national scale and across most agricultural zones, the GPP loss caused by the decline in MCI surpassed the impact of non-agriculturalization of cultivated land. This phenomenon was more prevalent in southern China, such as in MLYP and YGP. In the context of global warming, the accumulated temperature contours of wheat, rice, and corn in northern China have shifted northward (Yang et al., 2015). Climate change has expanded suitable cultivation areas and increased grain yield per unit in northern China (Ha et al., 2021), thereby facilitating the expansion of farmland. However, global warming has also introduced potential hazards, such as heat stress, cold waves, and frost (Lesk et al., 2016). Conversely, the cropping intensity of cultivated land in southern China has seen a widespread decrease, with approximately 85.42% of cultivated land experiencing a decline during 2000–2015 (Yan et al., 2019). Scholars attribute farmland marginalization to the expansion of urban areas and the growth of external agricultural labor (Wang, Y. et al., 2020; Yan et al., 2019). The decline in GPP primarily occurs in the southern regions due to these factors.
Fig. 17
  1. Download: Download high-res image (277KB)
  2. Download: Download full-size image

Fig. 17. Changes in GPP caused by cultivated land-use change in China and nine agricultural zones.

  • (2)
    Implications for cultivated land protection
According to the characteristics of cultivated land-use change and its impact on the GPP, it can be categorized as irreversible or reversible (Lu et al., 2019). Non-agriculturalization of farmland predominantly occurs when cultivated land is repurposed for buildings and infrastructure, often undergoing hardening treatments. This portion of farmland proves arduous to reclaim for agricultural cultivation, leading to a permanent loss of GPP (Yang and Li, 2000). While advancing rapid economic development, achieving a harmonious equilibrium between urbanization and farmland preservation is paramount. The government ought to employ high-precision and high-frequency remote sensing technology to strengthen the monitoring of cropland loss hotspots, furnish timely insights into farmland occupation, and effectively guide patterns of urban expansion.
The impact of farmland non-grainization on grain production has become increasingly apparent in southern China, prompting greater caution in the implementation of agricultural industrialization policies. In recent years, driven by the pursuit of higher agricultural product value, local governments have vigorously developed specialty agriculture and weakened the grain production function of cultivated land, resulting in hidden losses in grain production (Su et al., 2020). To mitigate the repercussions of non-grain production and ensure stability in grain production, the Chinese government introduced the “Guideline on Preventing Non-Grain Use of Arable Land and Stabilizing Grain Production to Ensure Food Security” in 2020, highlighting the imperative nature of preventing non-grainization. Moreover, the government should extensively harness technologies, including remote sensing and the Internet of Things, at the macro level to precisely ascertain the extent and distribution of non-grainization. This approach facilitates the scientific regulation of both the production quantity and spatial distribution of non-grain crops.
The marginalization of cultivated land has emerged as a pivotal issue in China's land utilization, significantly impacting food production. Unfortunately, few related policies have been implemented to curb the marginalization of farmland, and this phenomenon is gradually worsening. In recent years, the Chinese government has committed to land consolidation projects (Yin et al., 2022). According to statistics, land consolidation projects augmented farmland by 276 × 104 hm2 and 184 × 104 hm2 during 2001–2010 and 2011–2015, respectively (Zhou and Cao, 2020). Nevertheless, the supplementary farmland exhibited suboptimal quality and was progressively abandoned (Chen et al., 2023; Lu et al., 2024). Both the reduction in cropping intensity and farmland abandonment constitute recoverable land-use changes and fully restoring their GPP can yield more benefits than land consolidation projects. In other words, preventing further marginalization of farmland and activating extensive or even unused cultivated land is an urgent problem to be solved in current farmland protection efforts.
  • (3)
    Availability of identification results for farmland abandonment
This study attempted to estimate the impact of farmland abandonment on grain production in China, and the results showed that farmland abandonment was widespread in all agricultural zones, with a national farmland abandonment rate of approximately 25% from 1993 to 2019. A recent study by Zhang, M. et al. (2023) indicated that the abandonment rate of arable land in China between 1992 and 2015 was 18.59%. This study used the definition of abandoned farmland by the Food and Agriculture Organization as farmland that has not been used for agricultural production for more than five consecutive years, whereas the definition used in our study was farmland that had not been farmed for two consecutive years. Because of the longer study period and the difference in the definition of abandoned farmland, the abandonment rate identified in our study was higher. Notably, our identification process for abandoned farmland considered the vegetation succession sequence, enhancing accuracy in pinpointing naturally abandoned cultivated areas and bolstering the reliability of our results.
  • (4)
    Limitations and uncertainties
Model uncertainties cause deviations in the estimation results. Firstly, the primary data used in our study were derived from remote sensing monitoring, and the scale effect and mixed-pixel problem inherent in remote sensing pixels are challenging to circumvent. For example, in identifying the non-agriculturalization of cultivated land, the model failed to detect cultivated land occupied by small-scale construction, resulting in an underestimation of the calculated outcomes. Nevertheless, the estimation framework proposed in this study serves as a valuable reference for more nuanced regional-scale research. Secondly, we only discuss the effects of the MCI change and farmland abandonment on the GPP in the process of cultivated land marginalization, with no consideration given to alterations in agricultural production inputs.

5. Conclusion

Based on the practical problems of cultivated land use in China over the past 30 years, a quantitative analysis at the grid scale was conducted to assess the impact of three primary cultivated land-use changes (viz. non-agriculturalization, non-grainization, and marginalization) on GPP. From 1992 to 2020, the total farmland area initially exhibited an ascending trend followed by a subsequent decline, resulting in a net decrease of approximately 0.84% in China. the northward shift in the center of gravity of cultivated land has accentuated the spatial disparity between grain production and consumption. The GPP loss attributable to non-agriculturalization, non-grainization, and marginalization of farmland constituted 35.22 %, 22.58 %, and 42.20%, respectively. These losses demonstrated substantial variability across the different agricultural climate zones, with the combined loss in the MLYP and HHHP surpassing 60% of the national total. Farmland marginalization has become the most significant farmland utilization change affecting the GPP, resulting in a GPP loss of approximately 11471.30 × 104 t, primarily concentrated in the southern region. Farmland marginalization has not been effectively alleviated and continues to expand. Subsequent farmland conservation efforts should prioritize preventing the extensive utilization of farmland and curbing the expansion of marginalization. Shifting the focus from large-scale cultivation of new farmland to enhancing the efficiency of existing cultivated land usage promises greater returns.
This study tailored a pixel-scale estimation framework to evaluate potential grain production losses resulting from diverse cultivated land-use changes amid economic development. It systematically unveiled the impact and spatial heterogeneity of non-agriculturalization, non-grainization, and marginalization of cultivated land on GPP. The findings provided a reference for exploring effective strategies to enhance grain production. Furthermore, our research framework holds promise for application in other regions or countries experiencing rapid urbanization. To augment the depth of future research, a broader spectrum of cultivated land-use changes should be included in the estimation framework.

CRediT authorship contribution statement

Dan Lu: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing. Zhanpeng Wang: Data curation, Formal analysis. Kangchuan Su: Investigation, Methodology. Yajuan Zhou: Data curation. Xinxin Li: Investigation. Aiwen Lin: Methodology, Project administration, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article:Download: Download Word document (363KB)

Multimedia component 1.

Data availability

Data will be made available on request.

References

Cited by (13)

  • How does coal mining affect land use in townships? A perspective of land use transition

    2024, Journal of Cleaner Production

  • Spatial zoning and effect evaluation of county high-standard farmland siting delineation for sustainable cultivated land use in China: A case study in Dali, Shaanxi

    2024, Ecological Indicators

  • Spatio-temporal pattern evolution and regulatory zoning of suitability for farmland scale utilization in China based on multi-source data

    2024, Ecological Indicators

  • Unraveling patterns, causes, and nature-based remediation strategy for non-grain production on farmland in hilly regions

    2024, Environmental Research

  • Spatiotemporal Distribution and Driving Mechanisms of Cropland Long-Term Stability in China from 1990 to 2018

    2024, Land

  • Impacts of Cropland Utilization Patterns on the Sustainable Use Efficiency of Cropland Based on the Human–Land Perspective

    2024, Land
View all citing articles on Scopus
View Abstract
© 2024 Elsevier Ltd. All rights reserved.