Weibo Ma
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Weibo Ma has a total of 6 co-authors. Here are the top co-authors by number of shared publications:
Guie Li
School of Public Policy and Management China University of Mining and Technology Xuzhou China
1 shared publicationJianlan He
School of Public Policy and Management China University of Mining and Technology Xuzhou China
1 shared publication- Qingwu Yan
School of Public Policy & Management, China University of Mining and Technology, Xuzhou 221116, China
1 shared publication - Yiyun Chen
Wuhan University (Wuhan, China)
1 shared publication Zhaoqi Wu
School of Public Policy and Management China University of Mining and Technology Xuzhou China
1 shared publication- Zihao Wu
School of Public & Management, China University of Mining and Technology, Xuzhou 221116, China
1 shared publication
Latest Publications
- 1 July 2025
- Published by Wiley in Soil Use and Management
- Vol. 41 (3)
Mapping Soil Organic Carbon in Coal‐Grain Composite Areas: Integrating Mining and Reclamation Factors With CatBoost Modelling
Accurate mapping of soil organic carbon (SOC) in coal‐grain composite areas is crucial for evaluating soil quality and ensuring food safety. While previous studies recognised the impacts of mining activities and land reclamation on SOC accumulation, these critical factors were rarely incorporated into prediction models. Moreover, many key environmental variables are categorical, creating risks of high‐dimensional sparse data and overfitting problems. Therefore, this study aims to (i) assess the effectiveness of the CatBoost model in handling both categorical and continuous variables; (ii) determine the effects of mining, reclamation, agricultural and natural factors on SOC using the SHapley Additive exPlanations (SHAP) method; and (iii) generate spatial distribution and uncertainty maps of SOC. The study region is Pei County in China, a typical coal‐grain composite area. Results demonstrated CatBoost's superior accuracy and robustness against overfitting. Incorporating mining and reclamation factors improved prediction R 2 by 34.2%. The spatial distribution of SOC was primarily influenced by the source of reclaimed farmland. Notably, farmland reclaimed from mining wasteland and rural settlements exhibited significantly lower SOC levels compared to farmland reclaimed from subsided land and ordinary farmland. SHAP analysis further highlighted contrasting SOC–distance relationships: ordinary farmland SOC increased with distance from mining areas, while reclaimed farmland exhibited the inverse trends. These findings underscore the necessity of integrating mining and reclamation factors in SOC mapping within coal‐grain composite areas and validate CatBoost's efficacy in modelling systems dominated by categorical variables.
- 1 July 2025
- Published by Wiley in Soil Use and Management
- Vol. 41 (3)
Mapping Soil Organic Carbon in Coal‐Grain Composite Areas: Integrating Mining and Reclamation Factors With CatBoost Modelling
Accurate mapping of soil organic carbon (SOC) in coal‐grain composite areas is crucial for evaluating soil quality and ensuring food safety. While previous studies recognised the impacts of mining activities and land reclamation on SOC accumulation, these critical factors were rarely incorporated into prediction models. Moreover, many key environmental variables are categorical, creating risks of high‐dimensional sparse data and overfitting problems. Therefore, this study aims to (i) assess the effectiveness of the CatBoost model in handling both categorical and continuous variables; (ii) determine the effects of mining, reclamation, agricultural and natural factors on SOC using the SHapley Additive exPlanations (SHAP) method; and (iii) generate spatial distribution and uncertainty maps of SOC. The study region is Pei County in China, a typical coal‐grain composite area. Results demonstrated CatBoost's superior accuracy and robustness against overfitting. Incorporating mining and reclamation factors improved prediction R 2 by 34.2%. The spatial distribution of SOC was primarily influenced by the source of reclaimed farmland. Notably, farmland reclaimed from mining wasteland and rural settlements exhibited significantly lower SOC levels compared to farmland reclaimed from subsided land and ordinary farmland. SHAP analysis further highlighted contrasting SOC–distance relationships: ordinary farmland SOC increased with distance from mining areas, while reclaimed farmland exhibited the inverse trends. These findings underscore the necessity of integrating mining and reclamation factors in SOC mapping within coal‐grain composite areas and validate CatBoost's efficacy in modelling systems dominated by categorical variables.