ISSN 1003-8035 CN 11-2852/P
  • 中国科技核心期刊
  • CSCD收录期刊
  • Caj-cd规范获奖期刊
  • Scopus 收录期刊
  • DOAJ 收录期刊
  • GeoRef收录期刊
欢迎扫码关注“i环境微平台”

基于XGBoost模型的三峡库区燕山乡滑坡易发性评价与区划

吴宏阳, 周超, 梁鑫, 袁鹏程, 余蓝冰

吴宏阳,周超,梁鑫,等. 基于XGBoost模型的三峡库区燕山乡滑坡易发性评价与区划[J]. 中国地质灾害与防治学报,2023,34(5): 141-152. DOI: 10.16031/j.cnki.issn.1003-8035.202206020
引用本文: 吴宏阳,周超,梁鑫,等. 基于XGBoost模型的三峡库区燕山乡滑坡易发性评价与区划[J]. 中国地质灾害与防治学报,2023,34(5): 141-152. DOI: 10.16031/j.cnki.issn.1003-8035.202206020
WU Hongyang,ZHOU Chao,LIANG Xin,et al. Assessment of landslide susceptibility mapping based on XGBoost model: A case study of Yanshan Township[J]. The Chinese Journal of Geological Hazard and Control,2023,34(5): 141-152. DOI: 10.16031/j.cnki.issn.1003-8035.202206020
Citation: WU Hongyang,ZHOU Chao,LIANG Xin,et al. Assessment of landslide susceptibility mapping based on XGBoost model: A case study of Yanshan Township[J]. The Chinese Journal of Geological Hazard and Control,2023,34(5): 141-152. DOI: 10.16031/j.cnki.issn.1003-8035.202206020

基于XGBoost模型的三峡库区燕山乡滑坡易发性评价与区划

基金项目: 国家自然科学基金项目(42371094;41907253;41702330);湖北省重点研发计划项目(2021BCA219)
详细信息
    作者简介:

    吴宏阳(1997-),男,硕士研究生,主要从事地质灾害风险评价与系统开发。E-mail:wuhongyangpower@163.com

    通讯作者:

    周 超(1989-),男,副教授,博士,主要从事地质灾害监测预警与风险评价研究。E-mail:zhouchao@cug.edu.cn

  • 中图分类号: P642.22

Assessment of landslide susceptibility mapping based on XGBoost model: A case study of Yanshan Township

  • 摘要: 滑坡易发性评价是精细化滑坡灾害风险评价的基础。为了提升滑坡易发性评价模型的精度和稳健性,以三峡库区万州区燕山乡为例,选取工程地质岩组、堆积层厚度等九个影响因子构建滑坡易发性评价指标体系,应用信息量模型定量分析滑坡发育与指标之间的关系。在此基础上,随机选取70%/30%的滑坡样本作为训练/验证数据集,应用极致梯度提升模型(extreme gradient boosting, XGBoost)开展易发性评价。随后从模型预测精度和模型稳定性两方面将其与决策树模型(decision tree, DT)和梯度提升树模型(gradient boosting decision tree, GBDT)进行对比。结果表明:研究区堆积层滑坡主要受长江水系、堆积层厚度和工程地质岩组影响。XGBoost模型具有最高的准确率(94.3%)和预测精度(97.3%)。在模型稳定性验证中,平均预测精度最高(97.3%),优于DT(91.3%)和GBDT(95.7%),模型标准差和变异系数均为0.01,低于其余两种模型。XGBoost在区域滑坡易发性评价与制图中得到了可靠的结果,为滑坡灾害空间预测提供了新的技术支撑。
    Abstract: Landslide susceptibility assessment forms the foundation for precise evaluation of landslide risk. To enhance the accuracy and robustness of landslide susceptibility mapping, a state-of-art machine learning algorithm named the extreme gradient boosting model (XGBoost) was introduced to this study. Yanshan Town in Wanzhou district, Three Gorges reservoir, was chosen as a case study. Nine influencing factors, including engineering geological lithology and thickness of deposit layer, were selected to construct the landslide susceptibility evaluation index system. The relationship between landslide development and these indicators is quantitatively analyzed using the information value model. Subsequently, 70% of landslide samples were randomly assigned for training, while the remaining 30% were used for validation. The XGBoost model was then employed for landslide susceptibility mapping. The output were compared with those of the decision tree model (DT) and gradient boosting decision tree (GBDT) in terms of prediction accuracy and model stability. The findings revealed that distance to the Yangtze River, soil thickness, and lithology were the primary factors influencing landslide development. The XGBoost model demonstrated the highest average prediction accuracy (97.3%) in 100 repeated trials, surpassing the DT (91.3%) and GBDT models. Moreover, the XGBoost model exhibited superior robustness with a standard deviation and coefficient of variation of 0.01, lower than the other two models. It also achieved the highest accuracy (94.3%) and prediction accuracy (97.3%) in the validation process. The proposed XGBoost model serves as a reliable assessment method and yields optimal results in regional landslide susceptibility mapping.
  • 地面沉降又称地面下降或地陷,是指在自然环境和人类建设活动影响下,由于地下松散土层及岩层压缩固结,导致地表标高损失的一种地质现象[1]。造成该现象的自然因素和社会经济因素被认为是地面沉降的驱动力,主要包括水文地质条件、矿产资源开发及地下水开采状况等[23]。地面沉降形成原因复杂,防护与治理难度较高[46],对人民的生命财产安全造成较大的损害。为深入探索沉降机制及其变化规律,国内外学者对沉降监测方法和驱动力因素进行研究,并不断发展新的理论技术[710]

    传统的地面沉降监测手段主要为水准测量和GNSS,这种局部单点测量的技术不仅成本高而且空间分辨率低,难以识别和监测大面积的地表形变[11]。时序合成孔径雷达干涉测量(time-series interferometric synthetic aperture radar,TS-InSAR)技术不仅周期短、精度高,还能够全天时、全天候地对大范围的地表形变进行监测,极大地弥补了传统监测手段的不足[12]。其中永久散射体雷达干涉[13](permanent scatterer interferometric synthetic aperture radar,PS-InSAR)和小基线集雷达干涉[14](small baseline subset interferometric synthetic aperture radar,SBAS-InSAR)最具代表性。PS-InSAR方法能够有效获取高相干目标(如建筑、桥梁及裸岩等)的时序形变信息,但在植被茂盛、稳定散射体稀少的区域,无法取得足量的稳定目标点,易使得形变解算结果产生偏差[15]。而SBAS-InSAR方法是利用慢失相关滤波相位像素点获取地表形变信息,该类点能够在短时段内保持较强的相干性,且普遍存在于自然界中(如草地、裸土等)。因此SBAS-InSAR方法比PS-InSAR方法更适用于大范围区域的形变监测[1617]

    皖北地区的地面沉降问题历来较为突出,亳州市作为安徽省重要的新兴产业基地,其城市地质灾害监测一直受到政府及相关管理部门广泛关注。以往对该区域地面沉降的研究更侧重于观测数据的处理方法以及成因的简单分析[1819],对于其驱动力的量化研究尚且不足。探求地面沉降的主要驱动因素,能够为地质灾害防治和城市建设提供科学指导。本文以亳州市为研究区,选取2021年10月至2022年10月共62景Sentinel-1数据,利用SBAS-InSAR技术对亳州市地面沉降进行监测,分析亳州市地面沉降的时空分布特征,并基于地理加权回归(geographically weighted regression,GWR)模型,从地质环境、水文地质条件、人类工程活动和经济发展状况等方面对亳州市地面沉降的空间分异进行分析,探究亳州市地面沉降的主要驱动因素。

    亳州市位于黄淮海平原南端,皖、豫两省交界,全市下辖一区三县,总面积约8522.58 km2图1)。亳州市地处中朝准地台的淮河台坳二级构造单元,主要发育有褶皱、断裂构造。该地区地形起伏较小,地势西北高、东南低,辖境与黄河决口扇形地相连,总体呈典型的黄淮堆积型地貌[20]。地层属华北地层大区徐淮地层分区,第四系覆盖区内大部分基岩,第四系及新近系松散地层厚度在800~1000 m。亳州市煤炭资源丰富,根据《亳州市矿产资源总体规划(2021—2025年)》,区内现有煤矿产地17处,主要分布在涡阳、蒙城等地,保有资源储量43.50亿吨,占全省煤炭资源储量17.17%。主要含煤层为石炭、二叠系地层,煤层埋深600~1000 m。

    图  1  研究区范围
    Figure  1.  Study area scope

    亳州市地下水类型可划分为松散岩类孔隙水、碳酸盐岩类裂隙溶洞水和基岩裂隙水三种类型。按照含水层的埋藏条件,可进一步划分为浅层孔隙含水层组(50 m以浅)、中深层隙含水层组(50~165 m)、深层孔隙含水层组(165~660 m)、超深层孔隙含水层组(660~900 m)。根据《2021年亳州市水资源公报》,亳州市地下水资源总量约为15.67×108 m3,浅层地下水供水量4.59×108 m3,中深层地下水供水量为1.47×108 m3。全市域内已形成9个超采区,总面积约980.9 km2,其中浅层地下水超采区开采量约0.4×108 m3/a,中深层地下水超采区开采量约1.38×108 m3/a。

    Sentinel-1卫星是欧洲航天局发射的地球观测卫星,重访周期为12天,具有干涉宽幅(IW)、超宽幅(EW)、波(WV)和带状图(SM)四种工作模式。本文选取2021年10月至2022年10月间,共62景升轨Sentinel-1干涉宽幅(IW)模式的SLC影像用于形变监测,数据的基本参数见表1

    表  1  Sentinel-1卫星数据参数表
    Table  1.  Parameters of Sentinel-1 satellite data
    参数数值监测日期
    轨道高度/km7002021-10-02、2021-10-14、2021-10-26、
    2021-11-07、2021-11-19、2021-12-01、
    2021-12-13、2022-01-06、2022-01-18、
    2022-01-30、2022-02-11、2022-02-23、
    2022-03-07、2022-03-19、2022-03-31、
    2022-04-12、2022-04-24、2022-05-06、
    2022-05-18、2022-05-30、2022-06-11、
    2022-06-23、2022-07-05、2022-07-17、
    2022-07-29、2022-08-10、2022-08-22、
    2022-09-03、2022-09-15、2022-09-27、
    2022-10-09
    重访周期/d12
    入射角/(°)29~46
    分辨率/m5×20
    幅宽/m250
    极化方式VV
    轨道号142,101 / 142,106
    下载: 导出CSV 
    | 显示表格

    SRTM数据由美国国家航空航天局(NASA)和美国国家地理空间情报局(NGA)生产并面向全球用户免费发布,该数据覆盖了全球约五分之四的陆地表面,分辨率为30 m,高程精度为±16 m。本文中用于去除干涉测量过程中由地形起伏因素导致的地形相位。

    驱动力因子的选择主要依据研究区地质环境、水文地质条件、人类工程活动和经济发展状况四方面的综合影响,共选取8个指标,分别为松散层厚度、中深层地下水埋深、深层地下水埋深、中深层水位变幅、深层水位变幅、道路密度、人口密度和单位面积GDP。

    将覆盖研究区的N+1幅影像进行配准后,参照一定的阈值组成M个干涉对[21],则有:

    $$ \frac{{N + 1}}{2} \leqslant M \leqslant \frac{{N(N + 1)}}{2}$$ (1)

    假设第ii∈1, 2, ···, M)个干涉对的主辅影像获取时间为tatb(其中tbta之后),并且其干涉相位中除去形变相位的部分已被剔除,则该干涉对的相位可以表示为:

    $$ \varDelta \varphi _j^{({t_a},{t_b})} = {\varphi ^{{t_b}}} - {\varphi ^{{t_a}}} $$ (2)

    那么M个干涉对的形变相位可以表示为如下矩阵形式:

    $$ \varDelta \varphi ={\left[\varDelta {\varphi }_{1},\varDelta {\varphi }_{2},\varDelta {\varphi }_{3},\cdots ,\varDelta {\varphi }_{M}\right]}^{T} $$ (3)

    每个干涉都可以产生一个观测方程,结合式(2)(3),可以组成具有M个观测方程的方程组,其中有N个待求未知数,矩阵形式方程组如下:

    $$ A\varphi = \varDelta \varphi $$ (4)

    式中:A——M×N的系数矩阵。

    若矩阵A的秩r(A)大于N,则可以通过最小二乘法求解式(4),公式如下:

    $$ \varphi = {\left( {{A^T}A} \right)^{ - 1}}{A^T}\varDelta \varphi $$ (5)

    但在实际计算中,r(A)通常小于N,无法求得ATA矩阵的逆矩阵,此时则对矩阵A进行奇异值分解,分解形式如下:

    $$ A = US{V^T} $$ (6)

    式中:U——M×M阶正交矩阵,由AAT的特征向量组成;

    S——M阶对角矩阵;

    V——由ATA的特征向量组成的N×M阶正交矩阵。

    $$ {A^ + } = V{S^ + }{U^T} $$ (7)
    $$ \varphi = {A^ + }\varDelta \varphi $$ (8)

    式中:UT——U的转置矩阵;

    A+S+——矩阵A、矩阵S的广义逆矩阵。

    将求解出时序形变量φ,除以形变所对应的时间间隔,即可求解出对应的形变速率。

    SBAS-InSAR技术路线如图2所示,获取地面形变信息的流程主要包括两部分:数据预处理和SBAS-InSAR工作流。本研究使用ENVI平台的SARscape对Sentinel-1数据进行处理,SARscape是由sarmap公司开发的一款专业的雷达影像处理软件,已被广泛应用于处理ERS-1/2、RADARSAT-1/2、ENVISAT ASAR、ALOS PALSAR以及 Sentinel-1(哨兵)等一系列星载雷达数据[2225]

    图  2  SBAS-InSAR技术路线
    Figure  2.  SBAS-InSAR technical workflow

    Sentinel-1数据预处理步骤如下:①将数据导入为SARscape的标准格式;②对同一时期两景SAR影像进行镶嵌;③按照研究区范围对数据进行裁剪。对预处理后的影像进行SBAS-InSAR处理,主要流程包括:①对输入数据以最优的组合方式配对;②配对后的像对进行干涉处理;③利用控制点对所有数据重去平;④去除大气相位并估算形变速率;⑤地理编码,将形变结果投影到地理坐标系上。

    空间关系具有异质性和非平稳性规律,为了对空间数据进行精确局部描述,Fotheringham基于局部光滑的思想提出了地理加权回归模型(geographical weighted regression , GWR)[26]。GWR实质上是一种空间变系数回归模型[27],可以根据空间数据的位置信息生成对应的局部回归系数,从而对变量的局部空间关系与空间异质性进行合理的解释[28]

    运用GWR模型进行回归分析时,考虑到因子间的多重共线性问题会影响模型的可靠性,因此本文首先计算各因子的方差膨胀因子(VIF)。结果显示(表2),各因子的VIF均处于0到10之间[29],表明因子间不存在多重共线性。

    表  2  模型多重共线性检验
    Table  2.  Model multicollinearity test
    因子VIF因子VIF
    中深层地下水埋深1.234526松散层厚度1.519002
    中深层水位变幅1.625721人口密度1.116396
    深层水位变幅1.681352道路密度1.053348
    深层地下水埋深2.087465单位面积GDP2.481104
    下载: 导出CSV 
    | 显示表格

    在此基础上,对亳州市地面沉降建立GWR模型如下:

    $$ {y_i} = \sum\limits_{i = 0}^p {{\beta _j}} ({u_i},{v_i}){x_{ij}} + {\varepsilon _i}\quad i = 1,2,\cdots, n $$ (9)

    式中:yi——响应变量;

    βj(ui, vi)——第i个样本点在(ui, vi)处的第j个回归 参数;

    xij——影响因素;

    εi——随机误差项。

    采用赤池信息准则最优带宽策略,构建不同运行模式下的活动强度GWR模型[27],结果如表3所示。地面沉降GWR模型的可决系数(R2)为0.394583,表明自变量与因变量之间具有相关性。校正可决系数(adjusted R2)是0.373125,说明可解释因变量在模型中具有较高比例。

    表  3  2022年地面沉降GWR回归模型参数
    Table  3.  Ground subsidence GWR regression model parameters for 2022
    监测年份带宽赤池信息准则可决系数校正可决系数
    2022年82411850.6575450.3945830.373125
    下载: 导出CSV 
    | 显示表格

    为验证本文中研究区地面沉降数据的可靠性,选取谯城区周围3个同期水准点测量值与SBAS-InSAR监测结果进行对比。结果(表4)显示,SBAS-InSAR结果与水准测量值的误差在1 mm以内,说明监测结果具有较高的可信度。

    表  4  SBAS-InSAR监测结果与水准数据对比
    Table  4.  Comparison between SBAS-InSAR monitoring results and leveling data
    点名实测形变量/mmSBAS-InSAR监测的形变量/mm差值/mm
    BJ0133.830.83
    BJ02−1−0.54−0.46
    BXJ08−4−3.78−0.22
    下载: 导出CSV 
    | 显示表格

    监测结果与实测数据间存在一定误差,主要是因为SAR影像在干涉过程中受到大气延迟、地形起伏和失相干等多种因素影响产生的误差。此外,水准测量获取的是单个监测点的高程变化,而SBAS-InSAR结果则是一个单元格网(面状)的平均形变量,二者不一定完全对应。

    通过SBAS-InSAR处理,得到2021年10月至2022年10月内亳州市地表形变速率(图3)。亳州市整体沉降速率为5~30 mm/a,平均沉降速率为5.7 mm/a。地面沉降主要分布于谯城区东北部、涡阳县城、利辛县城以及蒙城县的部分地区;沉降最严重区域位于涡阳县公吉寺镇以北,受煤矿开采影响,沉降速率幅值达到84.3 mm/a;谯城区东北侧的地面沉降幅值为25.8 mm/a;在利辛县城及蒙城县的部分地区内,大多数区域地面沉降速率幅度小于10 mm/a水平,局部区域地面沉降幅度达到20 mm/a水平。

    图  3  亳州市2021年10月至2022年10月形变速率分布图
    Figure  3.  Distribution map of the subsiding rate of Bozhou from October 2021 to October 2022

    监测时段内沿雷达视线向(Line of Sight, LOS)的时序累计形变量如图4所示。在涡阳县中部、谯城区东北部、利辛县西部以及蒙城县中部,均监测到明显形变,形变量随时间推移逐渐增大。截至观测结束,涡阳县受煤矿开采影响区域,累积沉降量幅值达到83.4 mm,其余地区最大累计沉降量为27.3 mm;亳州市大部分区域地表累计沉降量处于5~30 mm水平,平均累计沉降量为7.3 mm左右。

    图  4  亳州市2021年10月至2022年10月时序累计形变量图
    Figure  4.  Time-series accumulated deformation map in Bozhou City from October 2021 to October 2022

    为了能够有效地掌握建模数据的分布情况,本文使用最小值、中值、最大值及平均值对模型运算结果进行叙述性统计。各建模变量拟合系数如表5所示,当系数为正时,自变量与因变量呈正相关关系;当系数为负时,自变量与因变量呈负相关关系,且拟合系数的绝对值越大,相关性越强。因此,各因素对地面沉降的贡献度排序依次为深层水位变幅、中深层水位变幅、中深层地下水埋深、深层地下水埋深、单位面积GDP、松散层厚度、道路密度、人口密度。

    表  5  模型运算结果叙述性统计
    Table  5.  Descriptive statistics of model calculation results
    变量最小值中值最大值平均值
    深层水位变幅-1.4870.9387.7693.141
    中深层水位变幅-1.4820.6022.6740.596
    中深层地下水埋深-0.747-0.3110.065-0.341
    深层地下水埋深-0.293-0.0500.085-0.104
    单位面积GDP-0.0030.0000.001-0.001
    松散层厚度-0.0140.0000.013-0.0005
    道路密度-0.0000.0000.0010.0005
    人口密度-0.0010.0000.0010.000
    下载: 导出CSV 
    | 显示表格

    亳州市煤矿资源主要分布在涡阳、蒙城两县,其中涡阳县现有矿产地13处,是区内主要的采煤区。结合驱动力因子回归结果与煤炭实际开采情况可知,采煤区沉降受地下水抽取与煤矿开采共同影响,而煤矿开采是沉降严重区域形变的主导因素。驱动力因子回归系数显示,采煤区地面沉降与中深层水位变幅,见图5(c)、深层水位变幅,见图5(d)呈显著正相关,说明地下水水位变化对地面沉降具有一定贡献。而该地区沉降最严重区域位于涡阳县公吉寺镇以北的信湖煤矿,最大沉降达83.4 mm。信湖煤矿于2021年9月16日正式投产,随着采矿活动的进行,矿区沉降速率持续加快。煤炭被采出后,形成采空区,随着采空区范围不断扩张,采空区上部覆岩和周围岩体的应力平衡遭到破坏,覆岩受到的重力作用逐渐增加,当压力超过临界值后,煤层顶板及周围岩体发生弯曲、断裂和垮落,导致整个上覆岩层的变形和移动,最终在地表形成大范围塌陷坑[30]

    图  5  各因子对地面沉降影响的回归系数图
    Figure  5.  Regression coefficients of different factors influencing ground subsidence

    非采煤沉降区主要位于谯城区东北部与利辛县西部,累计沉降量幅值为27.3 mm。驱动力因子回归系数显示,地下水状况与非采煤区地面沉降相关性较强,中深层水位变幅,见图5(c)和深层水位变幅,见图5(d)与地面沉降呈现正相关。自20世纪80年代以来,亳州市城市经济持续发展,人口迅速增长,对地下水资源的需求量也逐年增加。据有关资料估测[20],亳州市目前地下水日开采量在3.5×105 m3左右,深层地下水开采量在1.9×105 m3。对深层地下水的过度开采,诱使承压水头持续降低,降落漏斗面积不断扩大。当水头压力差作用于下伏黏性土层时,黏性土层的中低压缩性,会导致其越流或者压密释水,引起自身测压水头下降,使土体被纵向压缩;而砂性含水层受水头减小的影响,会释放出一部分储存的水,使得含水层内部的应力状态发生变化。原本承压水头支撑的上覆载荷被转移至含水层砂砾间,致使砂砾间的有效压力增加,含水层被垂直压缩。在地表上监测到的沉降量,即为降落漏斗范围内黏性土层与含水砂性土层的压缩量之和[31]

    松散层厚度与地面沉降的关系如图5(e)所示,在亳州市中部和西南部回归系数为负,对地面沉降起抑制作用;在北部和南部回归系数为正,对地面沉降起促进作用。亳州市地处淮北平原,地下发育有第四系和新近系松散沉积物[20],构成了地面沉降的物质基础。城市公共设施建设快速发展,城市建筑物的荷载不断增加,导致松散层被压实,进而引发地面沉降[22];另一方面,松散层中不同土层持水性具有的很大差异,过度开采地下水,使得土层颗粒间的有效应力增大,孔隙体积被压缩,导致地面沉降。

    本文利用2021年10月至2022年10月期间62景Sentinel-1卫星SAR影像,采用SBAS-InSAR技术获取了亳州市该时段内的地面形变速率及累计形变量,并对地面沉降的时空格局以及驱动力因素进行了分析,结果如下:

    (1)亳州市全域地面基本稳定,但局部地区存在明显的地面沉降现象。2021年10月至2022年10月期间,亳州市地面沉降最严重区域位于涡阳县公吉寺镇以北,幅值为84.3 mm/a,主要受煤矿开采影响所致。其他因采水导致地面沉降最大速率为25.8 mm/a,位于谯城区东北侧。

    (2)监测时段内,涡阳县中部、谯城区东北部、利辛县西部以及蒙城县中部均监测到明显形变,并且形变量随时间推移逐渐增大,亳州市整体平均累计沉降量为7.3 mm左右,采煤沉降区累计沉降量幅值为83.4 mm,非采煤沉降区累计沉降量幅值为27.3 mm。

    (3)各驱动力因素对地面沉降的贡献度排序为深层水位变幅、中深层水位变幅、中深层地下水埋深、深层地下水埋深、单位面积GDP、松散层厚度、道路密度、人口密度。

  • 图  1   决策树模型流程图

    Figure  1.   Flowchart of decision tree model

    图  2   梯度提升树模型流程图

    Figure  2.   Flowchart of gradient boosting decision tree model

    图  3   极致梯度提升模型流程图

    Figure  3.   Flowchart of extreme gradient boosting model

    图  4   研究区位置及滑坡分布

    Figure  4.   Location of the study area and distribution of landslides

    图  5   典型滑坡全貌图

    Figure  5.   Overview of typical landslide

    图  6   指标相关性

    Figure  6.   The correlation plot of Indicator factors

    图  7   研究区易发性评价指标图

    Figure  7.   Indicator plot for landslide susceptibility assessment in the study area

    图  8   参数与预测精度关系曲线

    Figure  8.   Relationship curve between parameters and prediction accuracy

    图  9   滑坡易发性分级图

    Figure  9.   Landslide susceptibility classification map

    图  10   各易发区灾害点比例

    Figure  10.   Proportion of disaster points in different susceptibility zones

    图  11   模型 ROC曲线图

    Figure  11.   ROC curves of the different models

    图  12   抽样次数与预测精度关系曲线

    Figure  12.   The correlation curve between sampling times and prediction accuracy

    表  1   各因素状态信息量表

    Table  1   The weighted information values of each factor state

    指标分级信息量指标分级信息量指标分级信息量
    坡度/(°)0~91.28工程地质岩组
    砂岩夹泥岩、砂岩1.58斜坡结构
    顺向伏倾坡、顺向飘倾坡0.75
    9~210.75砂泥岩互层0.59顺斜坡−0.99
    21~33−0.83泥岩夹砂岩、泥岩1.46横向坡、逆斜坡、逆向坡−0.96
    33~45−3.28页岩夹灰岩、灰岩−1.79斜坡形态内向凸形坡0.31
    >45−9.32距长江距离/m
    0-4002.96直线凸形坡、外向凸形坡−0.21
    植被归一化指数<0.151.07400~700−0.46内向直坡、直线直坡−1.79
    0.15~0.25−0.30700~14000.72外向直坡、内向凹坡−0.82
    >0.25−0.70>1400−1.84直线凹坡、外向凹坡−0.99
    坡向/(°)0~180−1.33地形湿度指数
    0~6−0.35堆积层厚度/m0~0.8−9.98
    180~2340.016~120.830.8~1.6−3.35
    234~252−0.1312~180.231.6~2.4−1.04
    252~3420.58>18−1.24>2.42.54
    342~360−0.60
    下载: 导出CSV

    表  2   标准差和变异系数

    Table  2   Standard deviation and coefficient of variation

    模型平均数标准差变异系数95%置信区间下限95%置信区间上限
    DT90.3040.7340.81390.16090.448
    GBDT95.6120.0620.06595.60095.624
    XGBoost97.2810.0100.01097.27997.283
    下载: 导出CSV
  • [1] 武雪玲,沈少青,牛瑞卿. GIS支持下应用PSO-SVM模型预测滑坡易发性[J]. 武汉大学学报(信息科学版),2016,41(5):665 − 671. [WU Xueling,SHEN Shaoqing,NIU Ruiqing. Landslide susceptibility prediction using GIS and PSO-SVM[J]. Geomatics and Information Science of Wuhan University,2016,41(5):665 − 671. (in Chinese with English abstract)

    WU Xueling, SHEN Shaoqing, NIU Ruiqing. Landslide susceptibility prediction using GIS and PSO-SVM[J]. Geomatics and Information Science of Wuhan University, 2016, 41(5)665-671(in Chinese with English abstract)

    [2] 殷跃平,张晨阳,闫慧,等. 三峡水库蓄水运行滑坡渗流稳定和防治设计研究[J]. 岩石力学与工程学报,2022,41(4):649 − 659. [YIN Yueping, ZHANG Chenyang, YAN Hui, et al. Research on seepage stability and prevention design of landslides during impoundment operation of the Three Gorges Reservoir, China[J]. Chinese Journal of Rock Mechanics and Engineering,2022,41(4):649 − 659. (in Chinese with English abstract)

    Yin Yueping, Zhang Chenyang, Yan Hui, et al. Research on seepage stability and prevention design of landslides during impoundment operation of the Three Gorges Reservoir, China[J]. Chinese Journal of Rock Mechanics and Engineering, 2022, 41(4): 649-659.(in Chinese with English abstract)

    [3] 杨何,汤明高,许强,等. 长江三峡库区滑坡变形统计特征研究[J]. 灾害学,2021,36(2):37 − 42. [YANG He,TANG Minggao,XU Qiang,et al. Research of statistical characteristics of deformation of landslides in the Three Gorges Reservoir area of the Yangtze River[J]. Journal of Catastrophology,2021,36(2):37 − 42. (in Chinese with English abstract)

    YANG He, TANG Minggao, XU Qiang, et al. Research of statistical characteristics of deformation of landslides in the Three Gorges Reservoir area of the Yangtze River[J]. Journal of Catastrophology, 2021, 36(2)37-42(in Chinese with English abstract)

    [4] 朱宇航,黄海峰,殷坤龙,等. 基于滑坡破坏模式分析的易发性评价—以三峡库区首段泄滩河左岸为例[J]. 中国地质灾害与防治学报,2023,34(2):156 − 166. [ZHU Yuhang, HUANG Haifeng, YIN Kunlong, et al. Evaluation of landslide susceptibility based on landslide failure mode analysis: A case study of the left bank of Xietan River in the first section of Three Gorges Reservoir[J]. The Chinese Journal of Geological Hazard and Control,2023,34(2):156 − 166. (in Chinese with English abstract)

    [ZHU Yuhang, HUANG Haifeng, YIN Kunlong, et al. Evaluation of landslide susceptibility based on landslide failure mode analysis: a case study of the left bank of Xietan River in the first section of Three Gorges Reservoir[J]. The Chinese Journal of Geological Hazard and Control, 2023, 34(2)156-166(in Chinese with English abstract)

    [5] 许嘉慧,张虹,文海家,等. 基于逻辑回归的巫山县滑坡易发性区划研究[J]. 重庆师范大学学报(自然科学版),2021,38(2):48 − 56. [XU Jiahui,ZHANG Hong,WEN Haijia,et al. Landslide susceptibility mapping based on logistic regression in Wushan County[J]. Journal of Chongqing Normal University (Natural Science),2021,38(2):48 − 56. (in Chinese with English abstract)

    Xu Jiahui, Zhang Hong, Wen Haijia, et al. Landslide susceptibility mapping based on logistic regression in Wushan County[J]. Journal of Chongqing Normal University (Natural Science), 2021, 38(2): 48-56. (in Chinese with English abstract)

    [6] 陈丽霞, 徐勇, 李德营. 武陵山区城镇地质灾害风险评估技术指南及案例分析[M]. 武汉: 中国地质大学出版社, 2019

    CHEN Lixia, XU Yong, LI Deying. Guidelines and applications for geohazard risk assessment of urban areas in Wuling Mountain region[M]. Wuhan: China University of Geosciences Press, 2019. (in Chinese)

    [7]

    BAI Shibiao,WANG Jian,LÜ Guonian,et al. GIS-based logistic regression for landslide susceptibility mapping of the Zhongxian segment in the Three Gorges area,China[J]. Geomorphology,2010,115(1/2):23 − 31.

    [8]

    LONG Jingjing,LIU Yong,LI Changdong,et al. A novel model for regional susceptibility mapping of rainfall-reservoir induced landslides in Jurassic slide-prone strata of western Hubei Province,Three Gorges Reservoir area[J]. Stochastic Environmental Research and Risk Assessment,2021,35(7):1403 − 1426. DOI: 10.1007/s00477-020-01892-z

    [9] 周超,殷坤龙,曹颖,等. 基于集成学习与径向基神经网络耦合模型的三峡库区滑坡易发性评价[J]. 地球科学,2020,45(6):1865 − 1876. [ZHOU Chao,YIN Kunlong,CAO Ying,et al. Landslide susceptibility assessment by applying the coupling method of radial basis neural network and adaboost:a case study from the Three Gorges Reservoir area[J]. Earth Science,2020,45(6):1865 − 1876. (in Chinese with English abstract)

    ZHOU Chao, YIN Kunlong, CAO Ying, et al. Landslide susceptibility assessment by applying the coupling method of radial basis neural network and adaboost: a case study from the Three Gorges Reservoir area[J]. Earth Science, 2020, 45(6): 1865-1876. (in Chinese with English abstract)

    [10] 田乃满,兰恒星,伍宇明,等. 人工神经网络和决策树模型在滑坡易发性分析中的性能对比[J]. 地球信息科学学报,2020,22(12):2304 − 2316. [TIAN Naiman,LAN Hengxing,WU Yuming,et al. Performance comparison of BP artificial neural network and CART decision tree model in landslide susceptibility prediction[J]. Journal of Geo-Information Science,2020,22(12):2304 − 2316. (in Chinese with English abstract)

    TIAN Naiman, LAN Hengxing, WU Yuming, et al. Performance comparison of BP artificial neural network and CART decision tree model in landslide susceptibility prediction[J]. Journal of Geo-Information Science, 2020, 22(12): 2304-2316. (in Chinese with English abstract)

    [11]

    REGMI N R,GIARDINO J R,VITEK J D. Assessing susceptibility to landslides:using models to understand observed changes in slopes[J]. Geomorphology,2010,122(1/2):25 − 38.

    [12] 王佳佳,殷坤龙,肖莉丽. 基于GIS和信息量的滑坡灾害易发性评价—以三峡库区万州区为例[J]. 岩石力学与工程学报,2014,33(4):797 − 808. [WANG Jiajia,YIN Kunlong,XIAO Lili. Landslide susceptibility assessment based on GIS and weighted information value:A case study of Wanzhou district,Three Gorges Reservoir[J]. Chinese Journal of Rock Mechanics and Engineering,2014,33(4):797 − 808. (in Chinese with English abstract)

    WANG Jiajia, YIN Kunlong, XIAO Lili. Landslide susceptibility assessment based on GIS and weighted information value: a case study of Wanzhou district, Three Gorges Reservoir[J]. Chinese Journal of Rock Mechanics and Engineering, 2014, 33(4): 797-808. (in Chinese with English abstract)

    [13] 周超,殷坤龙,向章波,等. 基于GIS的淳安县滑坡易发性定量评价[J]. 安全与环境工程,2015,22(1):45 − 50. [ZHOU Chao,YIN Kunlong,XIANG Zhangbo,et al. Quantitative evaluation of the landslide susceptibility in Chun'an County based on GIS[J]. Safety and Environmental Engineering,2015,22(1):45 − 50. (in Chinese with English abstract)

    ZHOU Chao, YIN Kunlong, XIANG Zhangbo, et al. Quantitative evaluation of the landslide susceptibility in Chun'an County based on GIS[J]. Safety and Environmental Engineering, 2015, 22(1): 45-50. (in Chinese with English abstract)

    [14] 张林梵,王佳运,张茂省,等. 基于BP神经网络的区域滑坡易发性评价[J]. 西北地质,2022,55(2):260 − 270. [ZHANG Linfan, WANG Jiayun, ZHANG Maosheng, et al. Evaluation of regional landslide susceptibility assessment based on BP neural network[J]. Northwestern Geology,2022,55(2):260 − 270. (in Chinese with English abstract)

    [ZHANG Linfan, WANG Jiayun, ZHANG Maosheng, et al. Evaluation of regional landslide susceptibility assessment based on BP neural network[J]. Northwestern Geology, 2022, 55(2): 260-270.(in Chinese with English abstract)

    [15] 许冲,徐锡伟. 基于GIS与ANN模型的地震滑坡易发性区划[J]. 地质科技情报,2012,31(3):116 − 121. [XU Chong,XU Xiwei. GIS and ANN model for earthquake triggered landslides susceptibility zonation[J]. Geological Science and Technology Information,2012,31(3):116 − 121. (in Chinese with English abstract)

    XU Chong, XU Xiwei. GIS and ANN model for earthquake triggered landslides susceptibility zonation[J]. Geological Science and Technology Information, 2012, 31(3): 116-121. (in Chinese with English abstract)

    [16]

    BRAGAGNOLO L,DA SILVA R V,GRZYBOWSKI J M V. Artificial neural network ensembles applied to the mapping of landslide susceptibility[J]. CATENA,2020,184:104240. DOI: 10.1016/j.catena.2019.104240

    [17] 陈飞,蔡超,李小双,等. 基于信息量与神经网络模型的滑坡易发性评价[J]. 岩石力学与工程学报,2020,39(S01):2859 − 2870. [CHEN Fei,CAI Chao,LI Xiaoshuang,et al. Evaluation of landslide susceptibility based on information volume and neural network model[J]. Chinese Journal of Rock Mechanics and Engineering,2020,39(S01):2859 − 2870. (in Chinese with English abstract)

    CHEN Fei, CAI Chao, LI Xiaoshuang, et al. Evaluation of landslide susceptibility based on information volume and neural network model[J]. Chinese Journal of Rock Mechanics and Engineering, 2020, 39(S01): 2859-2870. (in Chinese with English abstract)

    [18]

    YU Chenglong,CHEN Jianping. Landslide susceptibility mapping using the slope unit for southeastern Helong city,Jilin Province,China:a comparison of ANN and SVM[J]. Symmetry,2020,12(6):1047. DOI: 10.3390/sym12061047

    [19]

    ZHAO Shuai,ZHAO Zhou. A comparative study of landslide susceptibility mapping using SVM and PSO-SVM models based on grid and slope units[J]. Mathematical Problems in Engineering,2021,2021:1 − 15.

    [20]

    LEE S,HONG S M,JUNG H S. A support vector machine for landslide susceptibility mapping in gangwon Province,Korea[J]. Sustainability,2017,9(1):48. DOI: 10.3390/su9010048

    [21]

    PRADHAN B. A comparative study on the predictive ability of the decision tree,support vector machine and neuro-fuzzy models in landslide susceptibility mapping using GIS[J]. Computers & Geosciences,2013,51:350 − 365.

    [22]

    PARK S,HAMM S Y,KIM J. Performance evaluation of the GIS-based data-mining techniques decision tree,random forest,and rotation forest for landslide susceptibility modeling[J]. Sustainability,2019,11(20):5659. DOI: 10.3390/su11205659

    [23]

    SONG Yingxu,NIU Ruiqing,XU Shiluo,et al. Landslide susceptibility mapping based on weighted gradient boosting decision tree in Wanzhou section of the Three Gorges Reservoir area (China)[J]. ISPRS International Journal of Geo-Information,2018,8(1):4. DOI: 10.3390/ijgi8010004

    [24]

    YANG Yanguo, YU Jiaqi, FU Yubin, et al. Research on geological hazard risk assessment based on the cloud fuzzy clustering algorithm[J]. Journal of Intelligent & Fuzzy Systems,2019,37(4):4763 − 4770.

    [25]

    SRIKANTH B,LAKSHMI V PAPINENI S,SRIDEVI G,et al. Adaptive XGBOOST hyper tuned meta classifier for prediction of churn customers[J]. Intelligent Automation & Soft Computing,2022,33(1):21 − 34.

    [26]

    TEHRANY M S,PRADHAN B,MANSOR S,et al. Flood susceptibility assessment using GIS-based support vector machine model with different kernel types[J]. CATENA,2015,125:91 − 101. DOI: 10.1016/j.catena.2014.10.017

    [27]

    WU Yuanyuan,WU Li,ZHU Huacheng,et al. Design of high temperature complex dielectric properties measuring system based on XGBoost algorithm[J]. Materials,2020,13(6):1419. DOI: 10.3390/ma13061419

    [28] 赵建华,陈汉林,杨树锋,等. 基于决策树算法的滑坡危险性区划评价[J]. 浙江大学学报(理学版),2004,31(4):465 − 470. [ZHAO Jianhua,CHEN Hanlin,YANG Shufeng,et al. Landslide risk assessment based on decision tree arithmetic[J]. Journal of Zhejiang University (Science Edition),2004,31(4):465 − 470. (in Chinese with English abstract)

    Zhao Jianhua, Chen Hanlin, Yang Shufeng, et al. Landslide risk assessment based on decision tree arithmetic[J]. Journal of Zhejiang University (Science Edition), 2004, 31(4): 465-470. (in Chinese with English abstract)

    [29] 蔡文学,罗永豪,张冠湘,等. 基于GBDT与Logistic回归融合的个人信贷风险评估模型及实证分析[J]. 管理现代化,2017,37(2):1 − 4. [CAI Wenxue, LUO Yonghao, ZHANG Guanxiang, et al. Personal credit risk assessment model and empirical analysis based on the fusion of GBDT and logistic regression[J]. Modernization of Management,2017,37(2):1 − 4. (in Chinese with English abstract)

    [Cai Wenxue, Luo Yonghao, Zhang Guanxiang, et al. Personal credit risk assessment model and empirical analysis based on the fusion of GBDT and logistic regression[J]. Modernization of Management, 2017, 37(2(in Chinese with English abstract)

    [30]

    CHEN Tianqi, GUESTRIN C. XGBoost: a scalable tree boosting system[C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. August 13 – 17, 2016, San Francisco, California, USA. New York: ACM, 2016: 785 – 794.

    [31]

    ZHONG Jiancheng,SUN Yusui,PENG Wei,et al. XGBFEMF:an XGBoost-based framework for essential protein prediction[J]. IEEE Transactions on NanoBioscience,2018,17(3):243 − 250. DOI: 10.1109/TNB.2018.2842219

    [32] 黄发明,殷坤龙,蒋水华,等. 基于聚类分析和支持向量机的滑坡易发性评价[J]. 岩石力学与工程学报,2018,37(1):156 − 167. [HUANG Faming,YIN Kunlong,JIANG Shuihua,et al. Landslide susceptibility assessment based on clustering analysis and support vector machine[J]. Chinese Journal of Rock Mechanics and Engineering,2018,37(1):156 − 167. (in Chinese with English abstract)

    Huang Faming, Yin Kunlong, Jiang Shuihua, et al. Landslide susceptibility assessment based on clustering analysis and support vector machine[J]. Chinese Journal of Rock Mechanics and Engineering, 2018, 37(1): 156-167. (in Chinese with English abstract)

    [33] 张俊,殷坤龙,王佳佳,等. 三峡库区万州区滑坡灾害易发性评价研究[J]. 岩石力学与工程学报,2016,35(2):284 − 296. [ZHANG Jun,YIN Kunlong,WANG Jiajia,et al. Evaluation of landslide susceptibility for Wanzhou District of Three Gorges Reservoir[J]. Chinese Journal of Rock Mechanics and Engineering,2016,35(2):284 − 296. (in Chinese with English abstract)

    Zhang Jun, Yin Kunlong, Wang Jiajia, et al. Evaluation of landslide susceptibility for Wanzhou district of Three Gorges Reservoir[J]. Chinese Journal of Rock Mechanics and Engineering, 2016, 35(2): 284-296. (in Chinese with English abstract)

    [34] 王芳,殷坤龙,桂蕾,等. 不同日降雨工况下万州区滑坡灾害危险性分析[J]. 地质科技情报,2018,37(1):190 − 195. [WANG Fang,YIN Kunlong,GUI Lei,et al. Landslide hazard analysis under different daily rainfall conditions in Wanzhou district[J]. Geological Science and Technology Information,2018,37(1):190 − 195. (in Chinese with English abstract)

    WANG Fang, YIN Kunlong, GUI Lei, et al. Landslide hazard analysis under different daily rainfall conditions in Wanzhou district[J]. Geological Science and Technology Information, 2018, 37(1): 190-195. (in Chinese with English abstract)

    [35] 刘渊博,牛瑞卿,于宪煜,等. 旋转森林模型在滑坡易发性评价中的应用研究[J]. 武汉大学学报(信息科学版),2018,43(6):959 − 964. [LIU Yuanbo,NIU Ruiqing,YU Xianyu,et al. Application of the rotation forest model in landslide susceptibility assessment[J]. Geomatics and Information Science of Wuhan University,2018,43(6):959 − 964. (in Chinese with English abstract)

    LIU Yuanbo, NIU Ruiqing, YU Xianyu, et al. Application of the rotation forest model in landslide susceptibility assessment[J]. Geomatics and Information Science of Wuhan University, 2018, 43(6): 959-964. (in Chinese with English abstract)

    [36]

    BASHEER I A,HAJMEER M. Artificial neural networks:fundamentals,computing,design,and application[J]. Journal of Microbiological Methods,2000,43(1):3 − 31. DOI: 10.1016/S0167-7012(00)00201-3

    [37]

    TIAN Liwei,FENG Li,SUN Yu,et al. Forecast of LSTM-XGBoost in stock price based on Bayesian optimization[J]. Intelligent Automation & Soft Computing,2021,29(3):855 − 868.

    [38]

    CHEN Yu,WEI Yongming,WANG Qinjun,et al. Mapping post-earthquake landslide susceptibility:A U-net like approach[J]. Remote Sensing,2020,12(17):2767. DOI: 10.3390/rs12172767

    [39]

    ZHANG Kaixiang,WU Xueling,NIU Ruiqing,et al. The assessment of landslide susceptibility mapping using random forest and decision tree methods in the Three Gorges Reservoir area,China[J]. Environmental Earth Sciences,2017,76(11):405. DOI: 10.1007/s12665-017-6731-5

  • 期刊类型引用(6)

    1. 王勇,邢振涛,李锁,闫勇,司甜. 基于SBAS-InSAR和光学遥感的天津市北部山区潜在滑坡识别研究. 灾害学. 2025(01): 30-35 . 百度学术
    2. 谢哲,侯照方,谢杰. 基于钻探及室内试验的蒙城县地面沉降量预测. 能源技术与管理. 2025(01): 24-26 . 百度学术
    3. 赵志远,葛超英,徐雯佳. 时序InSAR技术对淮扬区域地面沉降监测. 四川地质学报. 2025(01): 162-169 . 百度学术
    4. 李幸丽,戴华阳,方军,张豪磊. 基于SBAS-InSAR技术的老采空区注浆充填地表变形时空分布特征分析. 中国矿业. 2024(11): 86-94 . 百度学术
    5. 马恩华,李益敏,俞文轩,吕圣彬. 基于InSAR的玉溪市红塔区地面沉降时空分布特征. 人民长江. 2024(S2): 113-120 . 百度学术
    6. 李松虎. 亳州市地下水超采现状及存在问题分析. 治淮. 2023(11): 7-9 . 百度学术

    其他类型引用(1)

图(12)  /  表(2)
计量
  • 文章访问数:  362
  • HTML全文浏览量:  785
  • PDF下载量:  141
  • 被引次数: 7
出版历程
  • 收稿日期:  2022-06-16
  • 修回日期:  2022-08-25
  • 网络出版日期:  2023-07-12
  • 刊出日期:  2023-10-30

目录

/

返回文章
返回