Eco-environment evolution and driving factors of the Loess Plateau based on the XGBoost-SHAP approach

doi:10.7522/j.issn.1000-694X.2026.00032

Journal of Desert Research ›› 2026, Vol. 46 ›› Issue (3): 175-186.DOI: 10.7522/j.issn.1000-694X.2026.00032

Eco-environment evolution and driving factors of the Loess Plateau based on the XGBoost-SHAP approach

Xiaopeng Li¹(), Kang Li¹, Jing Xu², Fugui Jia¹, Shuang Lei¹

^1.College of Agricultural，Forestry and Economic Management /, Lanzhou University of Finance and Economics，Lanzhou 730020，China
^2.School of Accounting, Lanzhou University of Finance and Economics，Lanzhou 730020，China

Received:2026-01-09 Revised:2026-03-02 Online:2026-05-20 Published:2026-06-11

Abstract

Abstract:

To characterize long-term changes in eco-environment quality on the Loess Plateau and to reveal the underlying drivers， we constructed a spatiotemporal series of eco-environment quality for 2000-2025 using the Remote Sensing Ecological Index （RSEI）. Class structure and transitions were analyzed， and Sen's slope estimation and the Mann-Kendall test were applied to identify change direction and significance. For mechanism interpretation， three representative years （2000， 2010， and 2020） were selected. Climate， topography， land use， and human-activity factors were integrated to develop XGBoost models， and SHAP was employed to decompose contributions and identify nonlinear responses. The results show that：（1） The multi-year mean RSEI was 0.449， with a distinct southeast-northwest gradient （higher in the southeast and lower in the northwest）； the class composition was dominated by the “moderate” and “poor” categories， while higher-quality classes exhibited banded or patchy clustering. （2） Eco-environment quality improved overall from 2000 to 2025， with the annual mean RSEI increasing from 0.395 to 0.474 and the areal proportion of low-quality classes decreasing from 53.86% to 35.40%， indicating a structural shift toward medium-to-high classes. （3） Five-year class transitions were dominated by persistence， with stable proportions of 65.99%-80.32%， and changes mainly occurred as gradual shifts between adjacent classes， although the relative strength of improvement and degradation varied among periods. （4） Water-related factors consistently played a dominant role， with annual precipitation and water deficit ranking highest in contribution， and key drivers exhibited threshold and segmented responses， including a transition of slope effects from weakly negative to positive at approximately 10°， a shift of precipitation effects from inhibiting to promoting at around 400 mm， and a shift of water-deficit effects from positive or near-zero to persistently negative at about 50 mm. These findings provide quantitative evidence and mechanistic insights to support long-term monitoring and assessment， zoned management， and optimization of ecological restoration projects on the Loess Plateau.

Key words: Loess Plateau, RSEI, spatiotemporal dynamics, class transition, XGBoost, SHAP, driving factors

CLC Number:

X826

Xiaopeng Li, Kang Li, Jing Xu, Fugui Jia, Shuang Lei. Eco-environment evolution and driving factors of the Loess Plateau based on the XGBoost-SHAP approach[J]. Journal of Desert Research, 2026, 46(3): 175-186.

Figures/Tables 11

Fig.1 Land use on the Loess Plateau

Table 1 Classification of change trends based on Sen+M-K methods

$β$	Z	变化趋势
$β$ >0	2.58<Z	极显著增加
	1.96<Z≤2.58	显著增加
	1.65<Z≤1.96	微显著增加
	Z≤1.65	不显著增加
$β$ =0		无变化
$β$ <0	Z>-1.65	不显著减少
	-1.96<Z≤-1.65	微显著减少
	-2.58<Z≤-1.96	显著减少
	Z≤-2.58	极显著减少

Table 1 Classification of change trends based on Sen+M-K methods

$β$	Z	变化趋势
$β$ >0	2.58<Z	极显著增加
	1.96<Z≤2.58	显著增加
	1.65<Z≤1.96	微显著增加
	Z≤1.65	不显著增加
$β$ =0		无变化
$β$ <0	Z>-1.65	不显著减少
	-1.96<Z≤-1.65	微显著减少
	-2.58<Z≤-1.96	显著减少
	Z≤-2.58	极显著减少

Fig.2 Spatial distribution of mean RSEI on the Loess Plateau

Fig.3 Spatial distribution of RSEI at five-year intervals on the Loess Plateau

Fig.4 Area proportions of RSEI classes on the Loess Plateau

Fig.5 RSEI class transitions on the Loess Plateau in 2000-2025

Fig.6 Sen trend and M-K test results of ecological quality on the Loess Plateau

Table 2 Evaluation metrics of the XGBoost model

评价指标	2000年	2010年	2020年
R²	0.842	0.823	0.871
MAE	0.024	0.027	0.022
RMSE	0.031	0.034	0.029

Fig.7 SHAP-based ranking of RSEI drivers on the Loess Plateau

Fig.8 SHAP beeswarm plot of RSEI drivers on the Loess Plateau

Fig.9 Interaction effects among driving factors

References 33

[1]	石学瑾，张彪，郭家龙，等.黄土高原典型小流域土壤侵蚀时空演变［J］.地理学报，2024，79（7）：1787-1803.
[2]	Feng X， Fu B， Piao S，et al.Revegetation in China's Loess Plateau is approaching sustainable water resource limits［J］.Nature Climate Change，2016，6（11）：1019-1022.
[3]	张琨，吕一河，傅伯杰，等.黄土高原植被覆盖变化对生态系统服务影响及其阈值［J］.地理学报，2020，75（5）：949-960.
[4]	Zhang B， Tian L， Yang Y，et al.Revegetation does not decrease water yield in the Loess Plateau of China［J］.Geophysical Research Letters，2022，49（9）：e2022GL098025.
[5]	王明宇，吴成永，陈克龙.青海湖流域植被绿度时空变化及影响因素［J］.中国沙漠，2025，45（5）：289-300.
[6]	顾杨旸，赵文吉，吴舒祺.1980-2020年黄河流域气象干旱时空变化特征［J］.中国农业气象，2025，46（8）：1192-1205.
[7]	徐涵秋.城市遥感生态指数的创建及其应用［J］.生态学报，2013，33（24）：7853-7862.
[8]	Kang S， Jia X， Zhao Y，et al.Spatiotemporal variation and driving factors of ecological environment quality on the Loess Plateau in China from 2000 to 2020［J］.Remote Sensing，2024，16（24）：4778.
[9]	王卫国，谢欢，冯国庆，等.京津风沙源治理工程区生态环境质量及驱动力［J］.中国沙漠，2025，45（4）：139-152.
[10]	白萌，莫兴国，莫淑红，等.生态恢复背景下陕北黄土高原水碳平衡演变及驱动机制研究［J］.地理研究，2025，44（11）：3180-3200.
[11]	刘健松，蔡音飞，王雅茹，等.黄土高原典型矿区生态脆弱性时空特征与驱动分析［J］.水土保持研究，2025，32（4）：217-226.
[12]	Chen T， Guestrin C.XGBoost：A scalable tree boosting system［C］//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining，2016：785-794.
[13]	杨佳，斛如媛，赵海平，等.基于XGBoost-SHAP的三峡库区生态屏障区生态系统服务权衡及驱动力分析［J］.水土保持学报，2025，39（6）：127-138.
[14]	孟飞，钱鑫彤，田浩.黄河流域改进遥感生态指数的多模型对比分析［J］.生态学杂志，2025，44（11）：3830-3840.
[15]	季翔林，阎跃观，郭伟，等.耦合遥感生态指数模型的山西省及规划矿区生态环境评价［J］.煤田地质与勘探，2023，51（3）：103-112.
[16]	Lundberg S M， Erion G， Chen H，et al.From local explanations to global understanding with explainable AI for trees［J］.Nature machine intelligence，2020，2（1）：56-67.
[17]	Xia J， Zhang G， Ma S，et al.Spatial heterogeneity of driving factors in multi-vegetation indices RSEI based on the XGBoost-SHAP model：a case study of the Jinsha River Basin，Yunnan［J］.Land，2025，14（5）：925.
[18]	许银，袁江，符素华.黄土高原切沟汇水区植被恢复模式对土壤饱和导水率的影响及机制［J］.水土保持学报，2025，39（5）：138-146.
[19]	潘占冬，王轶夫，王可月，等.基于改进型遥感生态指数的将乐县生态质量评估及驱动分析［J］.环境科学，2025，46（10）：6419-6431.
[20]	李晓鹏，李康，雷双，等.基于TVDI的新疆地区干旱时空变化特征［J］.干旱区研究，2025，42（4）：589-599.
[21]	Lundberg S M， Lee S I.A unified approach to interpreting model predictions［C］//Advances in Neural Information Processing Systems，2017，30：4765-4774.
[22]	Liu M H， Luo X L， Qi L A，et al.Simulation of the spatiotemporal distribution of PM_2.5 concentration based on GTWR-XGBoost two stage model：a case study of Chengdu Chongqing economic circle［J］.Atmosphere，2023，14（1）：115.
[23]	陈宜欣，宁晓刚，张翰超，等.遥感生态指数（RSEI）模型及应用综述［J］.自然资源遥感，2024，36（3）：28-40.
[24]	李双双，段生勇，胡佳岚，等.黄土高原植被变化主导空间模态及其影响因素［J］.地理学报，2024，79（7）：1768-1786.
[25]	Zhang J， Yang G， Yang L，et al.Dynamic monitoring of environmental quality in the Loess Plateau from 2000 to 2020 using the Google Earth Engine platform and the remote sensing ecological index［J］.Remote Sensing，2022，14（20）：5094.
[26]	李慧，张少伟，赵阳.黄河流域黄土高原生态质量时空演变格局及其驱动因素［J］.水土保持研究，2024，31（4）：233-240.
[27]	王经宇，杨丽萍，王美，等.陕北黄土高原生态环境质量时空变化监测及驱动力分析［J］.环境科学，2025，46（7）：4522-4533.
[28]	Wu S， Yue Z， Wang W.Three-dimensional identification and characterization of drought events in the Loess Plateau［J］.Remote Sensing，2025，17（17）：3049.
[29]	Zhou X， Liu G， Wu Y，et al.XGBoost-SHAP based multi-scale evaluation and nonlinear response of river health in Guangdong Province［J］.Ecological Indicators，2025，178：114138.
[30]	Xiong Y， Sun Z， Shen H，et al.Spatiotemporal dynamics of annual precipitation and future projections of China's 400 mm isohyet［J］.Remote Sensing，2025，17（17）：3078.
[31]	吴向莹，游智，杨曾慧子，等.长江中下游矿冶流域生态系统服务价值时空演变与多情景模拟［J］.环境科学与技术，2025，48（S1）：179-193.
[32]	郑金涛，金晓媚，李晴，等.基于地理探测器的张承地区表层土壤湿度变化的驱动因素分析［J］.水文地质工程地质，2024，51（2）：204-214.
[33]	雷可欣，张怀清，邱汉清，等.国家公园生态系统服务时空变化特征及其影响因素［J］.生态学报，2025，45（23）：11442-11462.

Eco-environment evolution and driving factors of the Loess Plateau based on the XGBoost-SHAP approach

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 11

References 33

Related Articles 15

Recommended Articles

Metrics

Comments

[1]	Xiaofeng Zuo, Yajuan Wei, Hejun Zuo, Haibing Wang, Min Yan, Xinglong Xie. Biomass distribution and carbon sink function of Nitraria tangutorum nebkhas in different climate zones [J]. Journal of Desert Research, 2026, 46(2): 86-94.
[2]	Jing Xu, Bin Yang, Hui Feng. The coupling pattern between new-type urbanization and economic resilience in Gansu Province and their driving factors [J]. Journal of Desert Research, 2026, 46(2): 263-276.
[3]	Bo Han, Jing Zhang, Xiaoguang Wang, Guangwei Huo, Peiying Liao, Zixiao Yang, Yitong Tao, Yunna Wu. Dynamic characteristics and driving forces of net primary productivity of vegetation in Horqin Sandy Land in 2001-2020 [J]. Journal of Desert Research, 2026, 46(2): 348-357.
[4]	Caixia Li, Tao Ren, Yongheng Li. Ecological sensitivity analysis in typical Loess Plateau gully region： a case study of Qingyang， Gansu [J]. Journal of Desert Research, 2025, 45(6): 166-175.
[5]	Lutong Wan, Xiaodan Liu, Hanqing Guo, Wanlai Xue, Kexuan Liu, Jin Li, Lili Wang, Honghao Liu. Multi-scale characteristics and driving factors of trade-off / synergy of ecosystem services in the ecological conservation area of western Beijing [J]. Journal of Desert Research, 2025, 45(6): 206-219.
[6]	Guanghui Shen, Jianying Zhang, Yue Han, Long Jia. Spatial distribution pattern and driving factors of Tenebrionidae in the Badain Jaran Desert margin [J]. Journal of Desert Research, 2025, 45(6): 323-333.
[7]	Pengfei Zhang, Jiaqi Han, Zhenjiang Zhao, Yanyan Dai, Geng Liu, Jinlong Chao, Lei Zhang. Coupling characteristics and the driving factors of agricultural water poverty and food production vulnerability in Fenhe River basin， China [J]. Journal of Desert Research, 2025, 45(5): 13-23.
[8]	Siqi Li, Haibing Wang, Xiaofei Yuan, Huricha Ao, Haoqin Yang. Temporal and spatial evolution of landscape ecological risks and driving forces in the Mu Us Sandy Land [J]. Journal of Desert Research, 2025, 45(5): 230-240.
[9]	Shurui Yang, Tian Yang, Lu Zhang, Dinghai Zhang, Haidi Qi. Moisture influencing factors on semi-fixed sand dunes in southeast edge of Tengger Desert [J]. Journal of Desert Research, 2025, 45(5): 328-337.
[10]	Huilin Zhang, Weiguo Wang, Yilan Bo, Zizhen Jin. Spatiotemporal dynamics of wind erosion prevention and sand fixation service in the Horqin Sandy Land [J]. Journal of Desert Research, 2025, 45(4): 227-240.
[11]	Xiulei Wang, Bo Zhang, Qing Chang, Xiangbo Sun, Xingyao Li, Bo Wu, Jia Li. Infrared camera traps for surveying mammal and bird diversity and activity rhythm in Zhongwei Shapotou National Nature Reserve， Ningxia [J]. Journal of Desert Research, 2025, 45(3): 162-174.
[12]	Liwen Yang, Xiaojun Li, Haotian Yang, Zhishan Zhang, Zhuoqun Shi, Xiao Qin, Dandan Hong, Dayong Wang. Quantitative assessment on the supply-demand balance of carbon sequestration service in artificial sand-binding vegetation ecosystem at the southeastern edge of the Tengger Desert [J]. Journal of Desert Research, 2025, 45(3): 222-232.
[13]	Xinyang Song, Mengyu Mu, Chenguang Zhao, Jiazheng Wang, Yanjun Huo, Changming Zhao, Chao Guan. Close-to-*nature restoration technological pattern for degraded Caragana korshinskii* stands in the Loess Plateau** [J]. Journal of Desert Research, 2025, 45(3): 283-290.
[14]	Shiyao Zhong, Chuanhua Li, Pengfei Qiao. Vegetation abrupt changes and attribution in the arid and semi-arid regions of Northwest China under aridity gradients from 2000 to 2020 [J]. Journal of Desert Research, 2025, 45(2): 275-283.
[15]	Yanyan Dai, Yiting Jiang, Pengfei Zhang, Geng Liu, Lei Zhang. Coupling coordination and driving factors between water resources and planting structure in Shanxi， China [J]. Journal of Desert Research, 2024, 44(6): 37-47.