生物炭施用下农田温室气体减排效应预测——基于XGBoost的Elastic Net模型

Prediction of the Mitigation Effect of Biochar on Greenhouse Gas Emissions from Farmland——An Elastic Net Model Based on XGBoost

摘要

生物炭对农田甲烷 (CH₄) 和氧化亚氮 (N₂O) 的减排效果受土壤性质、生物炭性质及作物种类等多因素协同影响，导致其综合效应难以精准量化。为此，本文通过系统文献筛选，构建了生物炭作用下农田 CH₄ 和 N₂O 排放的数据库，并耦合极端梯度提升 (XGBoost) 与 SHAP 方法，识别影响排放的关键特征及其交互效应。在此基础上，以土壤 pH 为一级分层依据，利用分类与回归树 (CART) 算法确定各 pH 层内土壤全氮 (TN) 的最优分割阈值，建立 pH−TN 二级分层下的弹性网络 (Elastic Net) 显性预测模型。结果表明：土壤 pH、生物炭碳含量、生物炭施用量及 TN 是关键影响因子，且土壤 pH 与 TN 在 pH<6.5 和 pH>7.5 区间内呈现显著交互效应。碱性土壤中，CH₄ 和 N₂O 排放的 TN 调控阈值分别为 0.95 g•kg⁻¹ 和 0.80 g•kg⁻¹；酸性土壤中则分别为 1.99 g•kg⁻¹ 和 1.86 g•kg⁻¹。基于上述分层构建的 Elastic Net 预测模型中，酸性高 TN 分层的 CH₄ 模型预测精度最高 (R²=0.71)，酸性低 TN 层的 N₂O 模型预测精度最高 (R²=0.45)。本研究所构建的 pH−TN 二级分层预测模型，可为不同土壤条件下生物炭的温室气体减排策略制定提供科学依据与参考。

Abstract

The mitigation effects of biochar on methane (CH₄) and nitrous oxide (N₂O) emissions from farmland are subject to considerable variation depending on soil properties, biochar characteristics, and crop type, making accurate assessment of emission-reduction efficacy in cropland systems inherently difficult. In this study, a database of CH₄ and N₂O emissions from farmland under biochar amendment was compiled through systematic literature screening. Key features and their interaction were identified using extreme gradient boosting (XGBoost) and shapley additive explanations (SHAP). Soil pH was adopted as the primary stratification criterion, and the classification and regression tree (CART) algorithm was applied to determine the optimal split threshold of soil total nitrogen (TN) within each pH stratum. On this basis, pH–TN two-level stratified elastic net explicit predictive models were constructed. The results indicate that soil pH, biochar carbon content, biochar application rate, and soil TN are the key influential features, with significant interaction effects between soil pH and TN observed in the pH<6.5 and pH>7.5 ranges. The TN regulatory thresholds for CH₄ and N₂O emissions are 0.95 g•kg⁻¹ and 0.80 g•kg⁻¹ in alkaline soils, and 1.99 g•kg⁻¹ and 1.86 g•kg⁻¹ in acidic soils, respectively. Among the stratified elastic net models, the acidic high-TN stratum model achieves the highest predictive accuracy for CH₄ emissions (R²=0.71), while the acidic low-TN stratum model performs accurately for N₂O emissions (R²=0.45). The pH–TN stratified predictive framework developed in this study provides a reference for assessing biochar-induced greenhouse gas mitigation effects and informing application strategies under varying soil conditions.

关键词

温室气体 / 弹性网络 / 生物炭 / 机器学习 / 分层模型 / SHAP 可解释性 / 极端梯度提升树

Key words

greenhouse gases / elastic net / biochar / machine learning / stratified model / SHAP interpretability / XGBoost

参考文献

[1] IPCC. Climate Change 2021: The Physical Science Basis. Working Group I contribution to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change[R]. 2021.

[2] 姜克隽, 冯升波. 走向《巴黎协定》温升目标: 已经在路上[J]. 气候变化研究进展, 2021, 17(1): 1-6.

[3] JIANG K J, FENG S B. Going to the mitigation targets in Paris Agreement: the world is on the road[J]. Climate Change Research, 2021, 17(1): 1-6.

[4] XIA L, CAO L, YANG Y, et al. Integrated biochar solutions can achieve carbon—neutral staple crop production[J]. Nature Food, 2023, 4(3): 236-246.

[5] 中华人民共和国生态环境部. 中华人民共和国气候变化第四次两年更新报告[R/OL]. (2024−12−31)[2026−04−23]. https://unfccc.int.

[6] 吴乐诗, 李冬琴, 黎华寿, 等. 生物炭影响农田温室气体排放的研究进展[J]. 生态科学, 2024, 43(1): 247-258.

[7] WU L S, LI D Q, LI H S, et al. Research progress on the effects of biochar on greenhouse gas emissions from farmland[J]. Ecological Science, 2024, 43(1): 247-258.

[8] SHRESTHA R K, JACINTHE P A, LAL R, et al. Biochar as a negative emission technology: a synthesis of field research on greenhouse gas emissions[J]. Journal of Environmental Quality, 2023, 52(4): 769-798.

[9] LIN F, WANG H, SHAGHALEH H, et al. Effects of biochar amendment on N₂O emissions from soils with different pH levels[J]. Atmosphere, 2024, 15(1): 68.

[10] 王紫君, 王鸿浩, 李金秋, 等. 椰糠生物炭对热区双季稻田 N₂O 和 CH₄ 排放的影响 [J]. 环境科学, 2021, 42(8): 3931-3942.

[11] WANG Z J, WANG H H, LI J Q, et al. Effects of coconut chaff biochar amendment on methane and nitrous oxide emissions from paddy fields in hot areas[J]. Environmental Science, 2021, 42(8): 3931-3942.

[12] BU F, NAN Q, LI W, et al. Meta—analysis for quantifying carbon sequestration and greenhouse gas emission in paddy soils one year after biochar application[J]. Agronomy, 2022, 12(12): 3065.

[13] BELMONTE B A, FLORES J G A, MESTIZO C L, et al. A rough set—based model for predicting soil greenhouse gases response to biochar[J]. Carbon Research, 2024, 3(1): 68.

[14] MUKHERJEE A, LAL R. Biochar impacts on soil physical properties and greenhouse gas emissions[J]. Agronomy, 2013, 3(2): 313-339.

[15] HUANG Y, TAO B, LAL R, et al. A global synthesis of biochar’s sustainability in climate—smart agriculture: evidence from field and laboratory experiments[J]. Renewable and Sustainable Energy Reviews, 2023, 172: 113042.

[16] WU Q, WANG J, HE Y, et al. Quantitative assessment and mitigation strategies of greenhouse gas emissions from rice fields in China: a data—driven approach based on machine learning and statistical modeling[J]. Computers and Electronics in Agriculture, 2023, 210: 107929.

[17] DHALIWAL J K, PANDAY D, ROBERTSON G P, et al. Machine learning reveals dynamic controls of soil nitrous oxide emissions from diverse long—term cropping systems[J]. Journal of Environmental Quality, 2025, 54(1): 132-146.

[18] LUNDBERG S M, LEE S I. A unified approach to interpreting model predictions[C]// Proceedings of the 31st International Conference on Neural Information Processing Systems. Red Hook, NY, US: Curran Associates Inc, 2017: 4768-4777.

[19] PONCE−BOBADILLA A V, SCHMITT V, MAIER C S, et al. Practical guide to SHAP analysis: explaining supervised machine learning model predictions in drug development[J]. Clinical and Translational Science, 2024, 17(11): e70056.

[20] KEBONYE N M, AGYEMAN P C, BINEY J K M. Optimized modelling of countrywide soil organic carbon levels via an interpretable decision tree[J]. Smart Agricultural Technology, 2023, 3: 100106.

[21] DU P, HUAI H, WU X, et al. Using XGBoost−SHAP for understanding the ecosystem services trade—off effects and driving mechanisms in ecologically fragile areas[J]. Frontiers in Plant Science, 2025, 16: 1552818.

[22] 侯泓锾, 吴萍, 刘鑫慧, 等. 土壤背景 N₂O 排放对土地利用、土壤性质与气候参数的响应: 基于全球的荟萃分析 [J]. 中国生态农业学报 (中英文), 2026, 34(2): 357-366.

[23] HOU H H, WU P, LIU X H, et al. Responses of soil background N₂O emissions to land use, soil properties, and climate parameters based on a global meta analysis [J]. Chinese Journal of Eco—Agriculture, 2026, 34(2): 357-366.

[24] WANG J, HAINING R, ZHANG T, et al. Statistical modeling of spatially stratified heterogeneous data[J]. Annals of the American Association of Geographers, 2024, 114(3): 499-519.

[25] NIU J, TANG H, ZHANG Y, et al. Optimization of crop yield prediction models based on stratified soil properties[J]. Smart Agricultural Technology, 2025, 12: 101466.

[26] ZOU H, HASTIE T. Regularization and variable selection via the elastic net[J]. Journal of the Royal Statistical Society Series B: Statistical Methodology, 2005, 67(2): 301-320.

[27] ACIR N. Predicting soil fertility in semi—arid agroecosystems using interpretable machine learning models: a sustainable approach for data—sparse regions[J]. Sustainability, 2025, 17: 7547.

[28] PARGENT F, PFISTERER F, THOMAS J, et al. Regularized target encoding outperforms traditional methods in supervised machine learning with high cardinality features[J]. Computational Statistics, 2022, 37: 2671-2692.

[29] KRZYWINSKI M, ALTMAN N. Classification and regression trees[J]. Nature Methods, 2017, 14(8): 757-758.

[30] XU X, LI T, YUE Q, et al. Geographical differences in the effect of biochar on crop yield and greenhouse gas emissions—a global simulation based on a machine learning model[J]. Current Research in Environmental Sustainability, 2024, 7: 100239.

[31] ACHTNICH C, BAK F, CONRAD R. Competition for electron donors among nitrate reducers, ferric iron reducers, sulfate reducers, and methanogens in anoxic paddy soil[J]. Biology and Fertility of Soils, 1995, 19(1): 65-72.

[32] BODELIER P L E, LAANBROEK H J. Nitrogen as a regulatory factor of methane oxidation in soils and sediments[J]. FEMS Microbiology Ecology, 2004, 47(3): 265-277.

[33] LE MER J, ROGER P. Production, oxidation, emission and consumption of methane by soils: a review[J]. European Journal of Soil Biology, 2001, 37(1): 25-50.

[34] HE D, DONG Z, ZHU B. An optimal global biochar application strategy based on matching biochar and soil properties to reduce global cropland greenhouse gas emissions: findings from a global meta—analysis and density functional theory calculation[J]. Biochar, 2024, 6(1): 92.

[35] CAYUELA M L, VAN ZWIETEN L, SINGH B P, et al. Biochar's role in mitigating soil nitrous oxide emissions: a review and meta—analysis[J]. Agriculture, Ecosystems & Environment, 2014, 191: 5-16.

[36] BUTTERBACH—BAHL K, BAGGS E M, DANNENMANN M, et al. Nitrous oxide emissions from soils: how well do we understand the processes and their controls?[J]. Philosophical Transactions of the Royal Society B: Biological Sciences, 2013, 368(1621): 20130122.

[37] BAI J, DE ALMEIDA MOREIRA B R, BAI Y, et al. Assessing biochar’s impact on greenhouse gas emissions, microbial biomass, and enzyme activities in agricultural soils through meta—analysis and machine learning[J]. Science of the Total Environment, 2025, 963: 178541.

[38] LONGEPIERRE M, WIDMER F, KELLER T, et al. Limited resilience of the soil microbiome to mechanical compaction within four growing seasons of agricultural management[J]. ISME Communications, 2021, 1(1): 44.

[39] GUO J H, LIU X J, ZHANG Y, et al. Significant acidification in major Chinese croplands[J]. Science, 2010, 327(5968): 1008-1010.

[40] ZHANG Q, XIAO J, XUE J, et al. Quantifying the effects of biochar application on greenhouse gas emissions from agricultural soils: a global meta—analysis[J]. Sustainability, 2020, 12(8): 3436.

[41] JI C, JIN Y, LI C, et al. Variation in soil methane release or uptake responses to biochar amendment: a separate meta—analysis[J]. Ecosystems, 2018, 21(8): 1692-1705.

[42] NAN Q, HU S, QIN Y, et al. Methane oxidation activity inhibition via high amount aged biochar application in paddy soil[J]. Science of the Total Environment, 2021, 796: 149050.

[43] GUAN S, QI Z, LI S, et al. Effects of rice root development and rhizosphere soil on methane emission in paddy fields[J]. Plants, 2024, 13: 3223.

[44] ZHENG Y, CAI Y, JIA Z. Role of methanotrophic communities in atmospheric methane oxidation in paddy soils[J]. Frontiers in Microbiology, 2024, 15: 1481044.

[45] FENG Y, XU Y, YU Y, et al. Mechanisms of biochar decreasing methane emission from Chinese paddy soils[J]. Soil Biology and Biochemistry, 2012, 46: 80-88.

[46] MENG X, ZHENG E, HOU D, et al. The effect of biochar types on carbon cycles in farmland soils: a meta analysis[J]. Science of the Total Environment, 2024, 930: 172623.

[47] GUO P, DIAO J, ZUO T, et al. Heterogeneous responses of soil nitrous oxide emissions to nitrogen addition: a global meta—analysis[J]. Catena, 2025, 254: 108923.

[48] DAS S, MOHAPATRA A, SAHU K, et al. Nitrogen dynamics as a function of soil types, compaction, and moisture[J]. Plos One, 2024, 19(4): e0301296.

[49] 蒋顺, 武敬, 陈波, 等. 复合钝化剂对砷污染土壤钝化修复及微生物群落结构的影响[J]. 安徽工业大学学报 (自然科学版), 2024, 41(6): 655-666.

[50] JIANG S, WU J, CHEN B, et al. Effect of composite passivator on arsenic contaminated soil passivation remediation and microbial community structure[J]. Journal of Anhui University of Technology (Natural Science), 2024, 41(6): 655-666.

基金资助

安徽省教育厅青年骨干教师境内访学研修资助项目(JNFX2025007)