Yuan Ke,Zhang Chen,Zhao Jianlin,et al.Comparative Analysis on Models for Predicting the Spatial Distribution of Soil Organic Carbon Density with Limited Samples[J].Research of Soil and Water Conservation,2024,31(05):173-181,191.[doi:10.13869/j.cnki.rswc.2024.05.041]
有限样本下土壤有机碳密度空间分布预测模型对比分析
- Title:
- Comparative Analysis on Models for Predicting the Spatial Distribution of Soil Organic Carbon Density with Limited Samples
- 文章编号:
- 1005-3409(2024)05-0173-09
- 分类号:
- S153.621
- 文献标志码:
- A
- 摘要:
- [目的]探讨不同机器学习模型在有限样本条件下预测表层土壤有机碳密度(SOCD)空间分布的精度和适用性,为黄土高原流域尺度碳库研究提供参考。[方法]基于延河子流域有限样本,对比多元线性逐步回归(SR)、随机森林(RF)、极端梯度提升(XGB)、支持向量机(SVM)这4种机器学习模型对表层土壤(0—20 cm)SOCD的预测精度和稳定性。[结果](1)在有限样本条件下,4种机器学习模型均可以较好地预测流域尺度SOCD空间分布,其中SVM模型精度最优,其50次预测的RMSE,R2,MAE平均值分别为0.74,0.43,0.64;(2)不同土地利用类型的SOCD均值估算结果大小一致且具有显著差异,均为灌木林>林地>草地>耕地,研究区总有机碳储量(0—20 cm)为2.39×106 t;(3)SOCD空间分布预测因子重要性评价结果表明地形因子、NDVImax、近红外波段地表反射率(B5)以及K-T变化中的Brightness因子对模型预测精度具有显著贡献。[结论]研究表明在有限样本条件下机器学习模型结合相关变量因子可有效应用于黄土高原流域尺度表层SOCD空间分布反演及碳库研究。
- Abstract:
- [Objective]The aim of this study is to explore the accuracy and applicability of different machine learning models for predicting the spatial distribution of surface soil organic carbon density(SOCD)with limited samples, which can provide a references for the study of watershed scale carbon pool in the Chinese Loess Plateau.[Methods]In this study, we compared the accuracy and stability of the predicted SOCD in topsoil(0—20 cm)by four machine learning models, namely Multiple Linear Stepwise Regression(SR), Random Forest(RF), Extreme Gradient Boosting(XGB)and Support Vector Machine(SVM), based on the limited measured samples in a sub-watershed of Yanhe River in the Chinese Loess Plateau. [Results](1)Under the condition of limited samples, all models successfully and appropriately predicte the spatial distribution of SOCD, among which the SVM model has the best model performance, and the average RMSE, R2 and MAE of 50 predictions is 0.74, 0.43 and 0.64, respectively.(2)The average SOCD of different land use types are consistent between measured and predicted values but shows significant difference among land use types. SOCD decreases in the order:shrubland>forestland>grassland>cropland. The total organic carbon of cultivated land in the study area is 2.39×106 t(0—20 cm).(3)The evaluation of feature importance shows that terrain factors, NDVImax, near-infrared surface reflectance(B5)and Brightness index have significant contributions to the accuracy of predictions. [Conclusion]Under the condition of limited samples, the machine learning model combined with controlling features can be effectively applied to the prediction of the spatial distribution of topsoil SOCD at the watershed scale in the Chinese Loess Plateau.
参考文献/References:
[1]Li X H, Ding J L, Liu J, et al. Digital mapping of soil organic carbon using sentinel series data:A case study of the ebinur lake watershed in Xinjiang[J]. Remote Sensing, 2021,13(4):769.
[2]马海丽.黄河源区土壤有机碳影响因子作用机制及模拟模型研究[D].兰州:兰州大学,2021.
Ma H L. Study on the Mechanism and Simulation Model of Soil Organic Carbon Influencing Factors in the Source Area of the Yellow River[D].Lanzhou:Lanzhou University, 2021.
[3]Piao S L, Fang J Y, Ciais P, et al. The carbon balance of terrestrial ecosystems in China[J]. Nature, 2009,458:1009-1013.
[4]Lamichhane S, Kumar L, Wilson B. Digital soil mapping algorithms and covariates for soil organic carbon mapping and their implications:A review[J]. Geoderma, 2019,352:395-413.
[5]刘艳芳,宋玉玲,郭龙,等.结合高光谱信息的土壤有机碳密度地统计模型[J].农业工程学报,2017,33(2):183-191.
Liu Y F, Song Y L, Guo L, et al. Geostatistical models of soil organic carbon density prediction based on soil hyperspectral reflectance[J]. Transactions of the Chinese Society of Agricultural Engineering, 2017,33(2):183-191.
[6]李妙宇.黄土高原生态系统碳储量现状及固碳潜力评估[D].北京:中国科学院大学(中国科学院教育部水土保持与生态环境研究中心),2021.
Li M Y. Carbon Storages and Carbon Sequestration Potentials of the Terrestrial Ecosystems on the Loess Plateau[D].Beijing:Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, 2021.
[7]丁亚鹏,张俊华,刘玉寒,等.基于GWR模型的伊河流域土壤有机碳空间分布特征及影响因素分析[J].生态学报,2021,41(12):4876-4885.
Ding Y P, Zhang J H, Liu Y H, et al. Spatial distribution characteristics and influencing factors of soil organic carbon in Yihe River Basin based on GWR model[J]. Acta Ecologica Sinica, 2021,41(12):4876-4885.
[8]魏宇宸,卢晓丽,朱昌达,等.基于地形与遥感辅助信息的小流域尺度高分辨率有机碳空间分布预测研究[J].土壤学报,2023,60(1):63-76.
Wei Y C, Lu X L, Zhu C D, et al. High-resolution digital mapping of soil organic carbon at small watershed scale using landform element classification and assisted remote sensing information[J]. Acta Pedologica Sinica, 2023,60(1):63-76.
[9]Odebiri O, Mutanga O, Odindi J. Deep learning-based national scale soil organic carbon mapping with Sentinel-3 data[J]. Geoderma, 2022,411:115695.
[10]Xu L, Yu G R, He N P, et al. Carbon storage in China's terrestrial ecosystems:A synthesis[J]. Scientific Reports, 2018,8:2806.
[11]Sothe C, Gonsamo A, Arabian J, et al. Large scale mapping of soil organic carbon concentration with 3D machine learning and satellite observations[J]. Geoderma, 2022,405:115402.
[12]杜雪,王海燕.中国森林土壤有机碳活性组分及其影响因素[J].世界林业研究,2022,35(1):76-81.
Du X, Wang H Y. Active components of forest soil organic carbon and its influencing factors in China[J]. World Forestry Research, 2022,35(1):76-81.
[13]赵青,刘爽,陈凯,等.武夷山自然保护区不同海拔甜槠天然林土壤有机碳变化特征及影响因素[J].生态学报,2021,41(13):5328-5339.
Zhao Q, Liu S, Chen K, et al. Change characteristics and influencing factors of soil organic carbon in Castanopsis eyrei natural forests at different altitudes in Wuyishan Nature Reserve[J]. Acta Ecologica Sinica, 2021,41(13):5328-5339.
[14]Breiman L. Random forests[J]. Machine Learning, 2001,45(1):5-32.
[15]韩杏杏,陈杰,王海洋,等.基于随机森林模型的耕地表层土壤有机质含量空间预测①:以河南省辉县市为例[J].土壤,2019,51(1):152-159.
Han X X, Chen J, Wang H Y, et al. Spatial prediction of SOM content in topsoil based on random forest algorithm:A case study of Huixian city, Henan Province[J]. Soils, 2019,51(1):152-159.
[16]Liaw A, Wiener M. Classification and regression by RandomForests[J]. R News, 2002,2(3):18-22.
[17]刘尊方,雷浩川,盛海彦.基于XGBoost模型的湟水流域耕地土壤养分遥感反演[J].干旱区地理,2023,46(10):1643-1653.
Liu Z F, Lei H C, Sheng H Y. Remote sensing inversion of soil nutrient on farmland in Huangshui River Basin based on XGBoost model[J]. Arid Land Geography, 2023,46(10):1643-1653.
[18]叶淼,朱琳,刘旭东,等.基于连续小波变换、SHAP和XGBoost的土壤有机质含量高光谱反演[J/OL].环境科学:1-19[2024-02-18].
Ye M, Zhu L, Liu X D, et al. Hyperspectral inversion of soil organic matter content based on continuous wavelet transform, SHAP, and XGBoost[J/OL]. Environmental Science:1-19[2024-02-18].
[19]傅贵,韩国强,逯峰,等.基于支持向量机回归的短时交通流预测模型[J].华南理工大学学报:自然科学版,2013,41(9):71-76.
Fu G, Han G Q, Lu F, et al. Short-term traffic flow forecasting model based on support vector machine regression[J]. Journal of South China University of Technology(Natural Science Edition), 2013,41(9):71-76.
[20]沈飞龙.基于机器学习与光谱信息的土壤铁氧化物估算模型研究[D].武汉:华中农业大学,2022.
Shen F L. A Model for Soil Iron Oxides Estimation Based on Machine Learning and Spectral Information[D].Wuhan:Huazhong Agricultural University, 2022.
[21]Zhao J L, Vanmaercke M, Chen L Q, et al. Vegetation cover and topography rather than human disturbance control gully density and sediment production on the Chinese Loess Plateau[J]. Geomorphology, 2016,274:92-105.
[22]丁倩,张弛.基于地理探测器的中国陆地生态系统土壤有机碳空间异质性影响因子分析[J].生态环境学报,2021,30(1):19-28.
Ding Q, Zhang C. Influential factors analysis for spatial heterogeneity of soil organic carbon in Chinese terrestrial ecosystem with geographical detector[J]. Ecology and Environmental Sciences, 2021,30(1):19-28.
[23]国佳欣,朱青,赵小敏,等.不同土地利用类型下土壤有机碳含量的高光谱反演[J].应用生态学报,2020,31(3):863-871.
Guo J X, Zhu Q, Zhao X M, et al. Hyper-spectral inversion of soil organic carbon content under different land use types[J]. Chinese Journal of Applied Ecology, 2020,31(3):863-871.
[24]李龙,秦富仓,姜丽娜,等.土地利用方式和地形对半干旱区土壤有机碳含量的影响[J].土壤,2019,51(2):406-412.
Li L, Qin F C, Jiang L N, et al. Effects of land use type and terrain on soil organic carbon(SOC)content in semi-arid region[J]. Soils, 2019,51(2):406-412.
[25]唐海涛,孟祥添,苏循新,等.基于CARS算法的不同类型土壤有机质高光谱预测[J].农业工程学报,2021,37(2):105-113.
Tang H T, Meng X T, Su X X, et al. Hyperspectral prediction on soil organic matter of different types using CARS algorithm[J]. Transactions of the Chinese Society of Agricultural Engineering, 2021,37(2):105-113.
[26]王富华,黄容,高明,等.生物质炭与秸秆配施对紫色土团聚体中有机碳含量的影响[J].土壤学报,2019,56(4):929-939.
Wang F H, Huang R, Gao M, et al. Effect of combined application of biochar and straw on organic carbon content in purple soil aggregates[J]. Acta Pedologica Sinica, 2019,56(4):929-939.
[27]张祎,李鹏,肖列,等.黄土高原丘陵区地形和土地利用对土壤有机碳的影响[J].土壤学报,2019,56(5):1140-1150.
Zhang Y, Li P, Xiao L, et al. Effects of topography and land use on soil organic carbon in hilly region of Loess Plateau[J]. Acta Pedologica Sinica, 2019,56(5):1140-1150.
[28]郭治兴,袁宇志,郭颖,等.基于地形因子的土壤有机碳最优估算模型[J].土壤学报,2017,54(2):331-343.
Guo Z X, Yuan Y Z, Guo Y, et al. Optimal estimation model of soil organic carbon based on the terrain factor[J]. Acta Pedologica Sinica, 2017,54(2):331-343.
[29]王绍强,周成虎.中国陆地土壤有机碳库的估算[J].地理研究,1999,18(4):349-356.
Wang S Q, Zhou C H. Estimating soil carbon reservior of terrestrial ecosystem in China[J]. Geographical Research, 1999,18(4):349-356.
[30]周国模,姜培坤.不同植被恢复对侵蚀型红壤活性碳库的影响[J].水土保持学报,2004,18(6):68-70,83.
Zhou G M, Jiang P K. Changes in active organic carbon of erosion red soil by vegetation recovery[J]. Journal of Soil Water Conservation, 2004,18(6):68-70,83.
相似文献/References:
[1]沈艳,傅瓦利,蓝家程,等.岩溶山地不同土地利用方式土壤颗粒有机碳和矿物结合态有机碳的分布特征[J].水土保持研究,2012,19(06):1.
SHEN Yan,FU Wa-li,LAN Jia-cheng,et al.Distribution Characteristics of Soil Particulate Organic Carbon and Mineral-associated Organic Carbon of Different Land Use in Karst Mountain[J].Research of Soil and Water Conservation,2012,19(05):1.
[2]陈媛媛,周运超.灰质白云岩土壤有机碳的团聚体保护[J].水土保持研究,2012,19(06):82.
CHEN Yuan-yuan,ZHOU Yun-chao.Protection of Soil Organic Carbon in the Calcite Dolomite Aggregate[J].Research of Soil and Water Conservation,2012,19(05):82.
[3]裴会敏,许明祥,李强,等.侵蚀条件下土壤有机碳流失研究进展[J].水土保持研究,2012,19(06):269.
PEI Hui-min,XU Ming-xiang,LI Qiang,et al.Advances in Soil Organic Carbon Losses under Erosion[J].Research of Soil and Water Conservation,2012,19(05):269.
[4]李玉进,王百群.施用柳枝稷茎和叶对土壤有机碳与微生物量碳的影响及其分解特征[J].水土保持研究,2012,19(05):78.
LI Yu-jin,WANG Bai-qun.Effect of Amendment of Leaf and Stem of Switchgrass(Panicum virgatum) on Soil Organic Carbon and Microbial Biomass Carbon as well as the Decomposition Characteristic of the Leaf and Stem[J].Research of Soil and Water Conservation,2012,19(05):78.
[5]赵发珠,韩新辉,杨改河,等.黄土丘陵区不同退耕还林地土壤有机碳、氮密度变化特征[J].水土保持研究,2012,19(04):43.
ZHAO Fa-zhu,HAN Xin-hui,YANG Gai-he,et al.Change Characteristics of Density of Soil Organic Carbon and Nitrogen under Land Shifted into Forestland in Hilly Loess Region[J].Research of Soil and Water Conservation,2012,19(05):43.
[6]杨洋,王百群,李玉进.苜蓿对旱地土壤有机碳氮变化的驱动作用[J].水土保持研究,2012,19(03):78.
YANG Yang,WANG Bai-qun,LI Yu-jin.Driving Effect of Alfalfa (Medicago sativa L.) on Dynamics of Soil Organic Carbon and Nitrogen in Dryland[J].Research of Soil and Water Conservation,2012,19(05):78.
[7]沙晋明,陈鹏程,陈松林.土壤有机质光谱响应特性研究[J].水土保持研究,2003,10(02):21.
SHA Jinming,CHEN Pengcheng,CHEN Songlin.Characteristics Analysis of Soil Spectrum Response Resulted From Organic Material[J].Research of Soil and Water Conservation,2003,10(05):21.
[8]贾松伟,贺秀斌,陈云明,等.黄土丘陵区土壤侵蚀对土壤有机碳流失的影响研究[J].水土保持研究,2004,11(04):88.
JIA Song-wei,HE Xiu-bin,CHEN Yun-ming,et al.Effect of Soil Erosion on Soil Organic Carbon Loss on the Loess Hilly Areas[J].Research of Soil and Water Conservation,2004,11(05):88.
[9]徐华君.阿尔泰山区土壤有机碳氧化稳定性的初步比较分析[J].水土保持研究,2007,14(06):27.
XU Hua-jun.Comparative Analysis of the Soil Organic Carbon Oxidation Stability in Altai Mountain, Xinjiang[J].Research of Soil and Water Conservation,2007,14(05):27.
[10]范胜龙,黄炎和,林金石.表征土壤有机碳区域分布的优化空间插值模型研究——以福建省龙海市为例[J].水土保持研究,2011,18(06):1.
FAN Sheng-long,HUANG Yan-he,LIN Jin-shi.The Optimized Interpolation Models and Its Relationship with Soil Sampling Density on Detecting Spatial Variability of Farmland Soil Organic Carbon —A Case Study in Longhai City, Fujian Province[J].Research of Soil and Water Conservation,2011,18(05):1.
备注/Memo
收稿日期:2023-11-03 修回日期:2023-11-17
资助项目:陕西林业科技创新重点专项(SXLK2023-02-15); 国家自然科学基金(41907048); 中央高校基本科研费专项资金(300102260206)
第一作者:袁可(1998—),男,四川省南部人,硕士研究生,研究方向为土壤有机碳遥感反演与制图。E-mail:2021126063@chd.edu.cn
通信作者:赵建林(1988—),男,陕西省镇巴人,副教授,主要从事土壤侵蚀与区域生态评价工作。E-mail:jianlin.zhao@chd.edu.cn