基于梯度提升树模型的坡耕地土壤水蚀模拟与分析

李潼亮; 赵梓鉴; 李斌斌; 张风宝; 郭正; 何琪琳; 何庆; 杨明义

doi:10.13870/j.cnki.stbcxb.2024.03.008

您当前的位置：

首页 >

文章列表页 >

基于梯度提升树模型的坡耕地土壤水蚀模拟与分析

更新时间：2025-03-12

- 基于梯度提升树模型的坡耕地土壤水蚀模拟与分析
- Simulation and Analysis of Hydraulic Erosion in Sloping Farmland Based on Gradient Boosting Decition Tree Mode
- 水土保持学报 2024年38卷第3期页码：54-63
- 作者机构：
  
  1. 西北农林科技大学水土保持科学与工程学院黄土高原土壤侵蚀与旱地农业国家重点实验室, 712100, 陕西杨凌
  2. 水利部水土保持监测中心,北京,100053
  3. 中国科学院水利部水土保持研究所, 陕西杨凌,712100
  4. 中国科学院大学,北京,100049
- 作者简介：
  
  [ "李潼亮(1999—),男,硕士研究生,主要从事土壤侵蚀预报与机理研究。E-mail:tlleeeee99@foxmail.com" ]
- 基金信息：
  
  国家自然科学基金项目(42077071,42177338,41830758);国家重点研发计划项目(2022YFF1300805);中央高校基本科研业务费专项资金项目(2023HHZX001)
- DOI：10.13870/j.cnki.stbcxb.2024.03.008
  中图分类号： S157.1
- 纸质出版：2024
- 稿件说明：
移动端阅览
李潼亮, 赵梓鉴, 李斌斌, 等. 基于梯度提升树模型的坡耕地土壤水蚀模拟与分析[J]. 水土保持学报, 2024,38(3):54-63.

LI Tongliang, ZHAO Zijian, LI Binbin, et al. Simulation and Analysis of Hydraulic Erosion in Sloping Farmland Based on Gradient Boosting Decition Tree Mode[J]. 2024, 38(3): 54-63.
李潼亮, 赵梓鉴, 李斌斌, 等. 基于梯度提升树模型的坡耕地土壤水蚀模拟与分析[J]. 水土保持学报, 2024,38(3):54-63. DOI： 10.13870/j.cnki.stbcxb.2024.03.008.

LI Tongliang, ZHAO Zijian, LI Binbin, et al. Simulation and Analysis of Hydraulic Erosion in Sloping Farmland Based on Gradient Boosting Decition Tree Mode[J]. 2024, 38(3): 54-63. DOI： 10.13870/j.cnki.stbcxb.2024.03.008.

摘要

[目的] 针对黄土高原坡耕地土壤侵蚀过程复杂、人为干扰强烈且难以量化的特点

利用机器学习定量解析主要影响因素对坡耕地土壤水蚀的作用与贡献

模拟分析坡耕地土壤水蚀特征并探究其机理

为坡耕地土壤侵蚀的预报提供基础支撑。[方法] 基于黄土高原子洲试验站坡耕地小区1959—1969年产流产沙观测数据

精细化表征其影响因子

运用梯度提升树模型对侵蚀量和径流深的变化及其影响因素的贡献进行分析。[结果] 数据集中次降雨侵蚀量(0~122.72 t/km2)、径流深(0.02~17.20 mm)、降雨历时(2~1 410 min)及平均雨强(0.02~4.63 mm)属强变异

变异系数均＞1

且多数变量呈右偏态;在相同训练集和测试集划分情况下

对侵蚀量模型预测精度(R2=0.81)略优于径流深模型(R2=0.80)

但侵蚀量模型的层数(8层)大于径流深模型(5层)

表明侵蚀机理相较径流机理更为复杂;通过梯度提升树模型与SHAP算法对自变量重要性进行排序发现

影响侵蚀模型与径流模型的自变量重要性不同。[结论] 受特征提取的限制

在侵蚀量与径流深较小时预测结果不理想

未来研究应当通过引入更多自变量组合方式寻找更多相关变量以提高对侵蚀事件的预测。产流和产沙的主要影响因素存在差异

降水本身特征对产流过程起主要作用

侵蚀产沙过程中主要受到降水与地形相关自变量的共同影响。基于数据驱动

为揭示黄土高原坡耕地侵蚀机理提供参考

并为区域坡耕地土壤侵蚀防治提供科学依据。

Abstract

[Objective] This article employs machine learning to quantitatively analyze soil water erosion in Loess Plateau slope farmland

addressing its complexity and quantification challenges due to human interference. We aim to simulate erosion characteristics

explore its mechanisms

and support erosion prediction. [Methods] Using 1959-1969 data from Zizhou Experimental Station

we characterized the influencing factors and analyzed erosion and runoff depth changes with a gradient boosting decision tree. [Results] The dataset showed significant variability in secondary rainfall erosion (0~122.72 t/km2)

runoff depth (0.02~17.20 mm)

rainfall duration (2~1 410 min)

and average intensity (0.02~4.63 mm)

often right-skewed. The erosion model (R2=0.81) slightly outperformed the runoff depth model (R2=0.80)

despite its greater complexity (8 layers vs. 5). Using the gradient boosting tree model and SHAP algorithm

we found differing key factors for erosion and runoff. [Conclusion] Limitations in feature extraction lead to less accurate predictions for small erosion and runoff depths. Future research should explore more independent variable combinations to enhance predictions. Main influencing factors differ between runoff and sediment production. Precipitation mainly influences runoff

while erosion and sediment production depend on precipitation and terrain-related variables. In summary

this data-driven study illuminates slope farmland erosion mechanisms on the Loess Plateau

providing a scientific basis for erosion control in the region.

关键词

Keywords

references

NGUYEN K A, CHEN W, LIN B S, et al. Using machine learning-based algorithms to analyze erosion rates of a Watershed in Northern Taiwan[J].Sustainability,2020,12(5):e2022.

江忠善, 郑粉莉. 坡面水蚀预报模型研究[J].水土保持学报,2004,18(1):66-69. JIANG Z S, ZHENG F L. Water erosion prediction model at hillslope scale[J].Journal of Soil Water Conservation,2004,18(1):66-69.

张玉斌,郑粉莉,贾媛媛.WEPP模型概述[J].水土保持研究,2004,11(4):146-149. ZHANG Y B, ZHENG F L, JIA Y Y. Overview of WEPP model[J]. Research of Soil and Water Conservation,2004,11(4):146-149.

李凤,吴长文.RUSLE侵蚀模型及其应用[J].水土保持研究,1997,4(1):109-112. LI F, WU C W. RUSLE erosion model and its application[J]. Research of Soil and Water Conservation,1997,4(1):109-112.

郑海金,汤崇军,汪邦稳,等.基于人工神经网络的红壤坡面水土流失预测模型[J].亚热带水土保持,2009,21(4):10-13,33. ZHENG H J, TANG C J, WANG B W, et al. Predicting model of soil and water loss based on manpower neural network on red-soil slope[J].Subtropical Soil and Water Conservation,2009,21(4):10-13,33.

陈劭锋,刘全友,陆中臣,等.黄土高原多沙粗沙区侵蚀产沙的多维临界[J].生态学报,2007,27(8):3277-3285. CHEN S F, LIU Q Y, LU Z C, et al. The multi-dimensional thresholds of sediment yield in the area with abundant and coarse sediment on the Loess Plateau[J].Acta Ecologica Sinica,2007,27(8):3277-3285.

洪伟,吴承祯.闽东南土壤流失人工神经网络预报研究[J].土壤侵蚀与水土保持学报,1997,11(3):53-58. HONG W, WU C Z. Study on artificial neural network prediction of soil loss in southeast Fujian[J].Journal of Soil and Water Conservation,1997,11(3):53-58.

彭清娥,曹叔尤,刘兴年,等.坡面产沙BP神经网络模型研究[J].水土保持学报,2002,16(3):79-82. PENG Q E, CAO S Y, LIU X N, et al. Research on BP neural network model for slope sediment yield[J].Journal of Soil and Water Conservation,2002,16(3):79-82.

赵西宁,王万忠,吴普特,等.坡面入渗的人工神经网络模型研究[J].农业工程学报,2004,20(3):48-50. ZHAO X N, WANG W Z, WU P T, et al. Artificial neural network model for soil infiltration in slope farmland[J].Transactions of the Chinese Society of Agricultural Engineering,2004,20(3):48-50.

段军彪,景旭,上官周平.基于遗传算法的BP网络在小流域侵蚀量预测中的应用[J].西北农业学报,2008,17(2):317-320. DUAN J B, JING X, SHANGGUAN Z P. Application of predicting soil erosion in the small watershed based on BP network model by genetic algorithm[J].Acta Agriculturae Boreali-Occidentalis Sinica,2008,17(2):317-320.

黄俊,金平伟,向家平,等.基于BP网络和回归分析的红壤可蚀性预测[J].中国水土保持科学,2015,13(3):8-15. HUANG J, JIN P W, XIANG J P, et al. Prediction of red soil erodibility based on BP neural network and regression analysis[J].Science of Soil and Water Conservation,2015,13(3):8-15.

李世欣,温建,邵孝侯,等.偏最小二乘法与人工神经网络耦合的小流域产沙模型[J].河海大学学报(自然科学版),2010,38(2):149-153. LI S X, WEN J, SHAO X H, et al. Sediment yield model for small watersheds based on coupling of partial least square regression and artificial neural network[J].Journal of Hohai University (Natural Sciences),2010,38(2):149-153.

王权威,唐莉.基于ABC-BP的土壤侵蚀量预报模型研究[J].水力发电,2017,43(9):1-4,44. WANG Q W, TANG L. Study on soil erosion prediction model based on artificial bee colony algorithm and BP neural network[J].Water Power,2017,43(9):1-4,44.

胡亚萍,董贝贝,姜宏立,等.SVR与BP神经网络对小流域次降雨侵蚀产沙预测结果的比较[J].北方环境,2013,25(1):114-117. HU Y P, DONG B B, JIANG H L, et al. Comparison of the results of the small watershed rainfall erosion and sediment yield predicted by SVR and BP nerve network[J].Northern Environment,2013,25(1):114-117.

DE LA ROSA D, MAYOL F, MORENO J A, et al. An expert system/neural network model (ImpelERO) for evaluating agricultural soil erosion in andalucia region, southern Spain[J].Agriculture, Ecosystems and Environment,1999,73(3):211-226.

LICZNAR P, NEARING M A. Artificial neural networks of soil erosion and runoff prediction at the plot scale[J].Catena,2003,51(2):89-114.

ARABAMERI A, SANTOSH M, SAHA S, et al. Spatial prediction of shallow landslide: Application of novel rotational forest-based reduced error pruning tree[J].Geomatics, Natural Hazards and Risk,2021,12(1):1343-1370.

ARABAMERI A, PRADHAN B, REZAEI K. Spatial prediction of gully erosion using ALOS PALSAR data and ensemble bivariate and data mining models[J].Geosciences Journal,2019,23(4):669-686.

ARABAMERI A, PRADHAN B, REZAEI K, et al. Spatial modelling of gully erosion using evidential belief function, logistic regression, and a new ensemble of evidential belief function-logistic regression algorithm[J].Land Degradation and Development,2018,29(11):4035-4049.

HAO J Q, LIN Y, REN G X, et al. Comprehensive benefit evaluation of conservation tillage based on BP neural network in the Loess Plateau[J].Soil and Tillage Research,2021,205:e104784.

周飞燕,金林鹏,董军.卷积神经网络研究综述[J].计算机学报,2017,40(6):1229-1251. ZHOU F Y, JIN L P, DONG J. Review of convolutional neural network[J].Chinese Journal of Computers,2017,40(6):1229-1251.

孟然,沈蔚,栾奎峰,等.基于梯度提升决策树算法的水深反演研究[J].海洋湖沼通报,2023,45(1):45-50. MENG R, SHEN W, LUAN K F, et al. Water depth retrieval based on gradient boosting decision tree algorithm[J].Transactions of Oceanology and Limnology,2023,45(1):45-50.

ELITH J, LEATHWICK J R, HASTIE T. A working guide to boosted regression trees[J].Journal of Animal Ecology,2008,77(4):802-813.

曹玉茹,高洋洋.基于SHAP值惩罚特征的集成分类方法研究[J].统计与决策,2023,39(6):21-26. CAO Y R, GAO Y Y. An ensemble classification method based on SHAP value to penalize features[J].Statistics and Decision,2023,39(6):21-26.

水利部黄河水利委员会. 黄河流域子洲径流实验站水文实验资料数据集(1959—1969年)[S]. 地球系统科学数据共享平台-黄土高原科学数据共享平, 2012 Yellow River Conservancy Commission of the Ministry of Water Resources. Dataset of hydrological experiments at Zizhou station in the Yellow River Basin (1959—1969)[S]. Data Sharing Infrastructure of Earth System Science_Data Sharing Infrastructure of Loess Plateau, 2012.

张龙齐,贾国栋,吕相融,等.黄土高原典型地区不同植被覆盖下坡面土壤侵蚀阈值研究[J].水土保持学报,2023,37(2):187-198. ZHANG L Q, JIA G D, LÜ X R, et al. Research of soil erosion thresholds on the lower slopes of different vegetation cover in typical areas of Loess Plateau[J].Journal of Soil and Water Conservation,2023,37(2):187-198.

王尧,蔡运龙,潘懋.贵州省乌江流域土壤侵蚀模拟:基于GIS、RUSLE和ANN技术的研究[J].中国地质,2014,41(5):1735-1747. WANG Y, CAI Y L, PAN M. Soil erosion simulation of the Wujiang river basin in Guizhou province based on GIS, RUSLE and ANN[J].Geology in China,2014,41(5):1735-1747.

QIU M L, VAN DE VOORDE T, LI T, et al. Spatiotemporal variation of agroecosystem service trade-offs and its driving factors across different climate zones[J].Ecological Indicators,2021,130:e108154.

HAMILTON J. Game theory: Analysis of conflict, by myerson, R. B., Cambridge: Harvard university press[J].Managerial and Decision Economics,1992,13(4):369.

EBRAHIMI-KHUSFI Z, TAGHIZADEH-MEHRJARDI R, ROUSTAEI F, et al. Determining the contribution of environmental factors in controlling dust pollution during cold and warm months of western Iran using different data mining algorithms and game theory[J].Ecological Indicators,2021,132:e108287.

吴发启,范文波.土壤结皮对降雨入渗和产流产沙的影响[J].中国水土保持科学,2005,3(2):97-101. WU F Q, FAN W B. Effects of soil encrustation on rainfall infiltration, runoff and sediment generation[J].Science of Soil and Water Conservation,2005,3(2):97-101.

何绍浪,李凤英,何小武.水蚀预报中降雨侵蚀力研究进展[J].水土保持通报,2018,38(2):262-270. HE S L, LI F Y, HE X W. Research progress of rainfall erosivity for water erosion prediction[J].Bulletin of Soil and Water Conservation,2018,38(2):262-270.

李小辉,贾本有,范子武,等.典型作物对水土流失影响的小区试验研究[J].水利水电技术,2019,50(2):95-100. LI X H, JIA B Y, FAN Z W, et al. Runoff plot observation study on the influences of typical crops on water and soil erosion[J].Water Resources and Hydropower Engineering,2019,50(2):95-100.

杨晓芬,吴发启,马波,等.黄土坡耕地玉米作物的防蚀作用研究[J].西北农林科技大学学报(自然科学版),2012,40(10):97-102. YANG X F, WU F Q, MA B, et al. Studies on anti-erosion effects of the maize in loess sloping fields[J].Journal of Northwest A&F University (Natural Science Edition),2012,40(10):97-102.

陈科兵,吴发启,姚冲.黄土高原南部地区人工模拟暴雨条件下不同坡度谷子坡耕地产流产沙过程[J].水土保持学报,2021,35(3):90-95,103. CHEN K B, WU F Q, YAO C. Runoff and sediment yield processes in millet cultivated land with different slopes under artificial simulated rainstorm in the southern Loess Plateau[J].Journal of Soil and Water Conservation,2021,35(3):90-95,103.

崔钦凯,刘俊娥,陈浩,等.草被覆盖、雨强和坡度对黄土坡面径流含沙量的影响[J].水土保持学报,2023,37(5):40-47. CUI Q K, LIU J E, CHEN H, et al The effects of grass cover, rain intensity, and slope on the sediment content of Loess Slope runoff[J].Journal of Soil and Water Conservation,2023,37(5):40-47.

浏览量

4270

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

川西高山峡谷区不同种植模式坡面微地形变化及其对细沟发育的响应

等高反坡阶对坡耕地土壤质量的改良效果评价

等高犁耕对黄土坡耕地坡形与坡度的影响

不同耕作方式对豫西旱区坡耕地水肥利用与流失的影响

连续14年黑土坡耕地秸秆覆盖免耕水土保持效应