Spatiotemporal Evolution and Driving Analysis of Carbon Storage in Hebei Province Based on InVEST-CatBoost-MGWR Models

Zhengyan Zhang; Hui  Chao

doi:10.54097/h7d8f271

Authors

Zhengyan Zhang Department of Surveying and Mapping Engineering, North China University of Science and Technology, Tangshan, Hebei 063210, China
Hui Chao Department of Surveying and Mapping Engineering, North China University of Science and Technology, Tangshan, Hebei 063210, China

DOI:

https://doi.org/10.54097/h7d8f271

Keywords:

Carbon storage, multi-model coupling, spatial heterogeneity, spatiotemporal dynamics.

Abstract

Land use and cover change significantly affects regional carbon cycles. From a multi-model coupling perspective, this study focuses on the mechanisms by which land use change influences carbon storage in Hebei Province. The InVEST model was employed to quantify the spatiotemporal dynamics of carbon storage from 2000 to 2025. The CatBoost machine learning algorithm was used to identify key driving factors, whose contributions were interpreted through SHAP values. Additionally, the multiscale geographically weighted regression (MGWR) model was applied to reveal the spatially non-stationary relationships between these factors and carbon storage. The results indicate that carbon storage in Hebei Province experienced a continuous decline from 2000 to 2025, decreasing from 1.887 billion tons to 1.841 billion tons,a total reduction of 46 million tons, or 2.44%. Notably, the decline accelerated markedly in the later period: the reduction was only 0.23% from 2000 to 2015 but reached 2.23% from 2015 to 2025. NDVI was the most important driving factor during 2000–2015, whereas slope emerged as the dominant factor by 2025, reflecting the increasing control of topographic conditions over carbon storage. The influence of each factor exhibited significant spatial heterogeneity. NDVI had the strongest impact in the Taihang Mountain area, while population density showed a significant negative effect in the urban agglomerations of the central Hebei Plain. These findings provide a scientific basis for low-carbon territorial spatial planning and the realization of carbon peak and carbon neutrality goals in Hebei Province.

Downloads

Download data is not yet available.

References

[1] Zhang, L., Yang, J. W., Cao, C. J., et al. (2026). Land use change and its carbon sequestration capacity in the Yellow River Basin (Henan section). Yellow River, 48(3), 107–112+118. (In Chinese)

[2] Jin, M. Z. (2020). Carbon storage function and value assessment of alpine grassland ecosystem in Maduo County, Sanjiangyuan region [Doctoral dissertation]. Lanzhou University, China. (In Chinese)

[3] Yu, Z. X., & Liu, Y. (2025). Spatiotemporal evolution and multi scenario prediction of carbon storage in the Xiangjiang River Basin based on PLUS and InVEST models. Journal of Agricultural Resources and Environment, (2025), 1–15. (In Chinese)

[4] Lou, J. Z. (2025). Carbon storage change and prediction in the karst region of Southwest China based on the PLUS InVEST model. Southern Agricultural Machinery, 56(7), 67–70. (In Chinese)

[5] Fan, Y., Dou, P. F., Tian, Y. F., et al. (2025). Ecological protection and restoration zoning at the county level in the low mountain and hilly areas of Fujian based on the supply demand relationship of ecosystem services. Chinese Journal of Ecology, (2025), 1–13. (In Chinese)

[6] Obateru, O. R., Okhimamhe, A. A., Fashae, A. O., et al. (2025). Assessing the status of ecosystem regulating services in the urbanising Rainforest and Guinea savanna ecological regions of Nigeria using InVEST models. Urban Climate, 61, 102410.

[7] Dat, T. P., Naoto, Y., Trang, T. T. N., et al. (2021). Improvement of mangrove soil carbon stocks estimation in North Vietnam using Sentinel 2 data and machine learning approach. GIScience & Remote Sensing, 58(1), 68–87.

[8] Hu, Z. H., Wang, L. F., Shang, R. G., et al. (2025). Changes and driving factors of tree layer carbon storage in old growth forests in Taiyanghe Provincial Nature Reserve, Yunnan. Chinese Journal of Applied Ecology, (2025), 1–11. (In Chinese)

[9] Haukenes, L. V., Asplund, J., Nybakken, L., et al. (2025). Disentangling drivers of organic layer and charcoal carbon stocks in boreal pine and spruce forests with different fire histories. Forest Ecosystems, 14, 100334.

[10] Hao, Z. X., Ma, J. K., Wang, A., et al. (2025). Changes and influencing factors of habitat quality in the Shanxi section of the Yellow River Basin based on the InVEST MGWR model. Chinese Journal of Applied Ecology, (2025), 1–12. (In Chinese)

[11] Zhang, J. Y., Yang, X. B., & Meng, X. S. (2025). Spatiotemporal distribution characteristics of forest carbon storage in the Beijing Tianjin Hebei region. Journal of Northeast Forestry University, 53(7), 75–83. (In Chinese)

[12] Huang, C. H., Yang, J., & Zhang, W. J. (2013). Research progress of ecosystem service function assessment models. Chinese Journal of Ecology, 32(12), 3360–3367. (In Chinese)

[13] Zhang, P. (2022). Multi scale spatiotemporal distribution of land use and ecosystem services based on InVEST and PLUS models [Doctoral dissertation]. Hebei University of Engineering, China. (In Chinese)

[14] Luo, L., Wu, L. K., Wang, Q. N., et al. (2025). Spatiotemporal evolution and prediction of mangrove wetland carbon storage in Hainan Island based on the PLUS InVEST model. Science of Soil and Water Conservation, (2025), 1–14. (In Chinese)

[15] Yue, S. J. (2024). Characteristics of ecosystem service function changes and future multi scenario simulation in the Beijing Tianjin Hebei region based on PLUS and InVEST models [Master’s thesis]. Henan Agricultural University, China. (In Chinese)

[16] Li, L. Y., Chen, L., Han, M. J., et al. (2024). Spatiotemporal evolution and multi scenario simulation of ecosystem carbon storage in Hebei Province. Chinese Journal of Applied Ecology, 35(12), 3247–3256. (In Chinese)

[17] Li, J. P., Xia, S. X., Yu, X. B., et al. (2020). Carbon storage of terrestrial ecosystems in Hebei Province based on the InVEST model. Journal of Ecology and Rural Environment, 36(7), 854–861. (In Chinese)

[18] Niu, X. G., Zhu, X. L., Liu, M. Y., et al. (2024). Spatiotemporal evolution characteristics and driving factors of ecosystem carbon storage and habitat quality in Hebei Province. Bulletin of Soil and Water Conservation, 44(6), 353–365. (In Chinese)

[19] Liu, F. Y. (2020). Application of root cause localization for anomalous faults based on CatBoost model [Doctoral dissertation]. Lanzhou University, China. (In Chinese)

[20] Shi, J. (2023). Application of BO CatBoost in earthquake prediction in the Western region [Master’s thesis]. Guangxi Normal University, China. (In Chinese)

[21] Du, Y. X., Hu, Z. J., Chen, W. N., et al. (2021). Transient stability assessment method of power system based on improved CatBoost. Electric Power Automation Equipment, 41(12), 115–122. (In Chinese)

[22] Guo, W., Huang, Z., Zhou, M. Z., et al. (2026). Prediction of inrush current of no load transformer based on CatBoost algorithm. Electric Drive, 56(2), 34–42. (In Chinese)

[23] Lai, J., Qi, S., Chen, J., et al. (2025). Exploring the spatiotemporal variation of carbon storage on Hainan Island and its driving factors: Insights from InVEST, FLUS models, and machine learning. Ecological Indicators, 172, 113236.

[24] Yi, F. L., Chen, D. R., Yang, H., et al. (2024). Classification of mild cognitive impairment using XGBoost SHAP interpretable machine learning framework. Chinese Journal of Health Statistics, 41(3), 423–429. (In Chinese)

[25] Ma, Z. L., Han, M., Kong, X. L., et al. (2026). Interpretable driving analysis of carbon storage in terrestrial ecosystems of the Yellow River Delta based on SHAP XGBoost model. Environmental Science, (2026), 1–28. (In Chinese)

[26] Huang, T., Peng, M. M., Wan, C. F., et al. (2026). Intelligent prediction of fatigue life of corroded steel wires in bridge cables based on SSA LSTM SHAP. Engineering Mechanics, (2026), 1–7. (In Chinese)

[27] Li, W. L., Li, F. R., Zhen, Z., et al. (2025). Estimation of forest carbon storage in Liaoning Province based on multiscale geographically weighted regression model. Journal of Central South University of Forestry and Technology, 45(9), 130–140. (In Chinese)

[28] Liu, Y., Tian, L. H., & Shi, D. R. (2025). Spatiotemporal variation characteristics of carbon storage and its driving factors in the Qinghai Lake Basin. Qinghai Agricultural and Forestry Science and Technology, (3), 1–7+19. (In Chinese)

[29] Chu, X. L. (2025). Spatiotemporal evolution characteristics and multi scenario prediction of carbon storage in the Tumen River Basin under land use change [Doctoral dissertation]. University of Jinan, China. (In Chinese)

[30] Zhang, Y. S. (2024). Ecosystem service changes, trade offs/synergies, and driving factors in Eastern Tibet [Doctoral dissertation]. Liaoning Technical University, China. (In Chinese)

[31] Zhang, Q. D. (2024). Spatiotemporal evolution and prediction analysis of ecosystem carbon storage in the Danjiangkou Reservoir area [Doctoral dissertation]. Henan University, China. (In Chinese)

[32] Zhou, M. J. (2025). Changes in distribution and carbon storage of mangroves in Hainan Island over the past 30 years [Doctoral dissertation]. Hainan Tropical Ocean University, China. (In Chinese)

[33] Liu, F., Wei, J. S., Zhou, M., et al. (2016). Remote sensing models of carbon storage in shrublands of Armeniaca sibirica and Caragana microphylla based on Landsat 8 OLI data. Forest Resources Management, (1), 112–117. (In Chinese)

[34] Shen, G. Y., & Zhang, M. Z. (2015). Estimation of spatial distribution of forest carbon density at multi scale regions based on sequential Gaussian cosimulation. Journal of Southwest Forestry University, 35(2), 55–62. (In Chinese)

[35] Xin, H., Lei, W., Jie, H., et al. (2023). Restorative perception of urban streets: Interpretation using deep learning and MGWR models. Frontiers in Public Health, 11, 1141630.

[36] Li, K. K., Zhou, Z. H., & Wang, Z. (2024). Socioeconomic spatial driving factors of soil erosion in China: An analysis based on the multiscale geographically weighted regression model. Journal of Huazhong Agricultural University, 43(6), 29–38. (In Chinese)