土地荒漠化是当今世界最为棘手的社会-经济-环境问题之一，荒漠化土地分布广泛，全球三分之二的国家和地区都受其影响，中国作为世界上荒漠化面积最大的国家之一，人们的日常生活已经受到荒漠化问题的影响。因此，对荒漠化进行动态监测并及时掌握其发展变化情况，对于荒漠化的治理具有重要的意义。本文以内蒙古自治区为研究区，以MODIS数据为数据源，针对已有模型对于轻度和中度荒漠化提取效果不佳的问题，基于空间距离指数法构造了综合荒漠化监测指数（Comprehensive Desertification Monitoring Index，CDMI），并与基于线性和点对点特征空间模型构造的荒漠化监测指数（Desertification Monitoring Index，DMI）进行对比。使用混淆矩阵来评估各模型的分类精度，然后基于最优模型提取2001-2020年的内蒙古荒漠化信息，在此基础上利用年际变化率、重心迁移和转移矩阵等方法来分析内蒙古荒漠化时空演变趋势。最后从气象、社会经济和地形三方面选取9种驱动因子，采用相关性分析和地理探测器法对荒漠化土地进行驱动力分析。主要内容和结论如下：

（1）基于空间距离指数法构建的综合荒漠化监测指数CDMI具有较好的适用性，其中Albedo-NDVI-LST监测精度最高，为85.4%；在基于一般线性模型构建的荒漠化监测指数中，LST-NDVI特征空间具有最高精度，达到了82.8%；在基于点对点模型构建的荒漠化监测指数中，LST-NDVI特征空间精度最高，为82.6%。综合分析三类荒漠化监测模型可以看出，空间距离指数法构建的模型精度整体高于线性和点对点模型。

（2）基于融合Albedo-NDVI-LST构造的综合荒漠化监测指数对内蒙古自治区进行土地荒漠化信息提取。制作了2001-2020年内蒙古自治区土地荒漠化分布图，在此基础上分析研究了不同等级荒漠化土地的动态变化情况。研究结果表明：在空间分布上，不同等级荒漠化土地分布界线较为明显，荒漠化严重程度总体呈现从东北向西南逐渐加重的趋势；在面积变化上，荒漠化土地面积整体呈减少趋势，具体面积变化情况为：未荒漠化和轻度荒漠化土地面积增加，中度、重度和极重度荒漠化土地面积减少；从迁移轨迹分析，整体荒漠化重心向西南方向迁移，东北部的未荒漠化土地增加程度要高于西南部极重度荒漠化增加程度，荒漠化呈恢复态势。内蒙古荒漠化土地变化总体呈“恢复-扩张-快速逆转-稳定”的发展格局。

（3）对驱动因素进行相关性分析的结果表明：年均气温和CDMI的相关系数为0.43，总体呈正相关；年降水量和CDMI的相关系数为-0.18，总体呈负相关；年均风速和CDMI的相关系数为0.04，总体呈正相关。对驱动因素进行地理探测器分析的结果表明：荒漠化对气候因素的响应尤为敏感，其中降雨量对土地荒漠化分异的解释力最大；在社会经济因素对荒漠化的影响中，农林牧渔业总产值起主要影响，说明经济水平的变化在一定程度上将影响荒漠化土地的分布；在地形因素中，驱动因子的解释力整体偏小，对荒漠化影响程度不大。各驱动因子交互作用后对荒漠化呈现增强作用，表明荒漠化的分异是各种因素共同作用的结果。

Land desertification is one of the most challenging socio-economic-environmental issues in the world today. Desertification is widely distributed and affects two-thirds of countries and regions worldwide. As one of the countries with the largest desertification area in the world, China's daily lives have been affected by desertification issues. Therefore, dynamic monitoring of desertification and timely understanding of its development and changes are of great significance for desertification control. With Inner Mongolia as the research area and MODIS data as the data source, this paper constructs the Comprehensive Desertification Monitoring Index (CDMI) based on the spatial distance index method to solve the problem of poor extraction effect of existing models for mild and moderate desertification, And compare it with the Desertification Monitoring Index (DMI) constructed based on linear and point-to-point feature space models. The Confusion matrix is used to evaluate the classification accuracy of each model, and then the desertification information of Inner Mongolia from 2001 to 2020 is extracted based on the optimal model. On this basis, the spatial-temporal evolution trend of desertification in Inner Mongolia is analyzed using the methods of inter annual change rate, gravity center migration and Stochastic matrix. Finally, 9 driving factors were selected from three aspects: meteorology, socio-economic, and terrain, and correlation analysis and geographic detector methods were used to analyze the driving forces of desertification land. The main content and conclusions are as follows:

(1) The comprehensive desertification monitoring index CDMI constructed based on the spatial distance index method has good applicability, with Albedo-NDVI-LST having the highest monitoring accuracy of 85.4%; In the desertification monitoring index constructed based on a general linear model, the LST-NDVI feature space has the highest accuracy, reaching 82.8%; In the desertification monitoring index constructed based on point-to-point models, LST-NDVI has the highest spatial accuracy of 82.6%. A comprehensive analysis of the three types of desertification monitoring models shows that the overall accuracy of the models constructed by the spatial distance index method is higher than that of linear and point-to-point models.

(2) Extracting land desertification information from Inner Mongolia Autonomous Region using a comprehensive desertification monitoring index constructed based on the fusion of Albedo-NDVI-LST. A distribution map of land desertification in Inner Mongolia Autonomous Region from 2001 to 2020 was created, and based on this, the dynamic changes of different levels of desertification land were analyzed and studied. The research results indicate that in terms of spatial distribution, the distribution boundaries of desertification land at different levels are more obvious, and the overall severity of desertification shows a trend of gradually increasing from northeast to southwest; In terms of area change, the overall area of desertification land shows a decreasing trend, with specific area changes as follows: the area of non desertification and mild desertification land increases, while the area of moderate, severe, and extremely severe desertification land decreases; From the analysis of migration trajectory, the overall focus of desertification is shifting towards the southwest direction, and the increase in non desertification land in the northeast is higher than the increase in extremely severe desertification in the southwest, indicating a recovery trend of desertification. The overall development pattern of desertification land change in Inner Mongolia is "restoration expansion rapid reversal stability".

(3) The correlation analysis of the driving factors shows that the correlation coefficient between annual average temperature and CDMI is 0.43, indicating a positive overall correlation; The correlation coefficient between annual precipitation and CDMI is -0.18, showing a negative overall correlation; The correlation coefficient between annual average wind speed and CDMI is 0.04, showing a positive overall correlation. The results of geographic detector analysis on driving factors indicate that desertification is particularly sensitive to climate factors, with rainfall having the greatest explanatory power on land desertification differentiation; In the impact of socio-economic factors on desertification, the total output value of agriculture, forestry, animal husbandry, and fishery plays a major role, indicating that changes in economic level will to some extent affect the distribution of desertification land; In terms of terrain factors, the explanatory power of driving factors is generally small, and the impact on desertification is not significant. The interaction of various driving factors enhances desertification, indicating that the differentiation of desertification is the result of the joint action of various factors.

[1] 朱震达. 中国土地荒漠化的概念、成因与防治[J]. 第四纪研究, 1998, (02): 145-155.

[2] Carla F, Samaneh S, Georgia D, et al. Soil degradation in the European Mediterranean region: Processes, status and consequences[J]. Science of the Total Environment, 2022, 805: 150106.

[3] Sandeep K, Devesh S, Aditya S. Spatio-temporal variation in desert vulnerability using desertification index over the Banas River Basin in Rajasthan, India[J]. Arabian Journal of Geosciences, 2021, 14(1): 1-13.

[4] Shen X, Liu Y, Liu B, et al. Effect of shrub encroachment on land surface temperature in semi-arid areas of temperate regions of the Northern Hemisphere[J]. Agricultural and Forest Meteorology, 2022, 320.

[5] 张亚男, 王旭, 甄莹. 利用“修改型”植被盖度反演模型提取辽西北沙化土地信息[J]. 测绘通报, 2018, (05): 93-96.

[6] 孙技星, 钟成, 何宏伟, 等. 2000—2015年中国土地荒漠化连续遥感监测及其变化[J]. 东北林业大学学报, 2021, 49(03): 87-92.

[7] Liu S, Zhang J, Zhang J, et al. Assessing Controversial Desertification Prevention Policies in Ecologically Fragile and Deeply Impoverished Areas: A Case Study of Marginal Parts of the Taklimakan Desert, China[J]. Land, 2021, 10(6): 641.

[8] Wang T. Aeolian desertification and its control in Northern China[J]. International Soil and Water Conservation Research, 2014, 2(4): 34-41.

[9] 边振. 基于遥感技术的荒漠化监测方法研究[D]. 北京林业大学, 2011.

[10] 赵方. 榆林市荒漠化动态变化与驱动因素分析[D]. 西北农林科技大学, 2022.

[11] 郭瑞霞, 管晓丹, 张艳婷. 我国荒漠化主要研究进展[J]. 干旱气象, 2015, 33(03): 505-513.

[12] 杨雪栋. 内蒙古自治区荒漠化和沙化土地监测概述[J]. 内蒙古林业调查设计, 2020, 43(02): 86-88.

[13] 丁雪. 内蒙古自治区土地荒漠化动态变化研究[D]. 东北农业大学, 2018.

[14] 赵文博. 河北省荒漠化时空演变及驱动因素分析[D]. 河北工程大学, 2022.

[15] 丁雪, 雷国平, 许端阳, 等. 1981-2010年内蒙古沙漠化演变对区域生态系统服务价值的影响[J]. 水土保持研究, 2018, 25(01): 298-303.

[16] 边振, 张克斌. 我国荒漠化评价研究综述[J]. 中国水土保持科学, 2010, 8(01): 105-112.

[17] 赵媛媛, 高广磊, 秦树高, 等. 荒漠化监测与评价指标研究进展[J]. 干旱区资源与环境, 2019, 33(05): 81-87.

[18] 张云芝, 胡云锋, 韩月琪, 等. 全球主要生态退化区和研究热点区的空间分布与演变[J]. 生态学报, 2021, 41(19): 7599-7613.

[19] 郭强. 中国北方荒漠化遥感动态监测与定量评估研究[D]. 中国科学院大学(中国科学院遥感与数字地球研究所), 2018.

[20] Wang T, Yan C, Song X, et al. Monitoring recent trends in the area of aeolian desertified land using Landsat images in China’s Xinjiang region[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2012, 68: 184-190.

[21] 陈长委, 伍永秋, 谭利华, 等. 青藏铁路错那湖段沙漠化土地变化及成因分析[J]. 干旱区地理, 2019, 42(04): 885-892.

[22] 于钧. 基于特征空间的科尔沁沙地荒漠化信息提取研究[D]. 辽宁工程技术大学, 2022.

[23] 邱立. 毛乌素沙漠景观的遥感分析[D]. 西安科技大学, 2008.

[24] 王鹏新, 陈晓玲, 李飞鹏. 典型干草原退化草地的时空分布特征及其动态监测[J]. 干旱地区农业研究, 2002, (01): 92-94+106.

[25] 闫峰, 吴波. 近40a毛乌素沙地荒漠化过程研究[J]. 干旱区地理, 2013, 36(06): 987-996.

[26] 宋伟东, 张亚男, 高琳, 等. 面向对象与决策树模型的辽西北地区沙化信息提取[J]. 辽宁工程技术大学学报(自然科学版), 2018, 37(03): 595-601.

[27] Chen F, Wang S, Bai X, et al. Assessing spatial-temporal evolution processes and driving forces of karst rocky desertification[J]. Geocarto International, 2021, 36(3): 262-280.

[28] 黄晓君, 颉耀文, 卫娇娇, 等. 基于变化检测-CART决策树模式自动识别沙漠化信息[J]. 灾害学, 2017, 32(01): 36-42.

[29] Prăvălie R, Patriche C, Bandoc G. Quantification of land degradation sensitivity areas in Southern and Central Southeastern Europe. New results based on improving DISMED methodology with new climate data[J]. Catena, 2017, 158: 309-320.

[30] 敏玉芳, 张耀南, 康建芳, 等. 基于MODIS影像的中巴经济走廊荒漠化程度时空动态监测研究[J]. 遥感技术与应用, 2021, 36(04): 827-837.

[31] Piao S, Fang J, Liu H, et al. NDVI‐indicated decline in desertification in China in the past two decades[J]. Geophysical Research Letters, 2005, 32(6): 402.

[32] 何磊, 王超, 别强, 等. 利用MOD13Q1产品监测肯尼亚2001—2010年荒漠化动态[J]. 中国沙漠, 2013, 33(01): 46-52.

[33] Liang S. Narrowband to broadband conversions of land surface albedo I[J]. Remote Sensing of Environment, 2001, 76(2): 213-238.

[34] Wei H, Wang J, Han B. Desertification Information Extraction Along the China–Mongolia Railway Supported by Multisource Feature Space and Geographical Zoning Modeling[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2020, 13: 392-402.

[35] 曾永年, 向南平, 冯兆东, 等. Albedo-NDVI特征空间及沙漠化遥感监测指数研究[J]. 地理科学, 2006, 26(1): 75-81.

[36] Guo B, Zang W, Han B, et al. Dynamic monitoring of desertification in Naiman Banner based on feature space models with typical surface parameters derived from LANDSAT images[J]. Land Degradation & Development, 2020, 31(12): 1573-1592.

[37] Wu Z, Lei S, Bian Z, et al. Study of the desertification index based on the albedo-MSAVI feature space for semi-arid steppe region[J]. Environmental Earth Sciences, 2019, 78(6): 232.

[38] 李艳华, 丁建丽, 孙永猛, 等. 基于三维特征空间的土壤盐渍化遥感模型[J]. 水土保持研究, 2015, 22(04): 113-117+121.

[39] Nick M. Rangeland management and climate hazards in drylands: dust storms, desertification and the overgrazing debate[J]. Natural Hazards, 2018, 92(S1): 57-70.

[40] 赵文博, 冯莉莉, 赵安周, 等. 河北省2000-2017年荒漠化的时空演变及其气候驱动因子[J]. 水土保持通报, 2021, 41(04): 252-259.

[41] 姚正毅, 李晓英, 董治宝. 黄河源区玛多县沙漠化成因与发展过程[J]. 冰川冻土, 2015, 37(05): 1245-1256.

[42] Tucker C, Dregne H, Newcomb W. Expansion and Contraction of the Sahara Desert from 1980 to 1990[J]. Science, 1991, 253(5017): 299-300.

[43] 刘易. 论沙漠污染行为的法律规制[D]. 中央民族大学, 2016.

[44] Shoba P, Ramakrishnan S. Modeling the contributing factors of desertification and evaluating their relationships to the soil degradation process through geomatic techniques[J]. Solid Earth, 2016, 7(2): 341-354.

[45] 张博. 1999-2018年青海省土地荒漠化遥感监测及其驱动力分析[D]. 中国地质大学(北京), 2020.

[46] Xie L, Zhong J, Chen F, et al. Evaluation of soil fertility in the succession of karst rocky desertification using principal component analysis[J]. Solid Earth, 2015, 6(2): 515-524.

[47] Jiang M, Lin Y, Chan T, et al. Geologic factors leadingly drawing the macroecological pattern of rocky desertification in southwest China[J]. Scientific reports, 2020, 10(1): 1440.

[48] Kundu A, Patel N, Saha S, et al. Desertification in western Rajasthan (India): an assessment using remote sensing derived rain-use efficiency and residual trend methods[J]. Natural Hazards, 2017, 86(1): 297-313.

[49] Hua D, Hao X. Spatiotemporal change and drivers analysis of desertification in the arid region of northwest China based on Geographic Detector[J]. Environmental Challenges, 2021, (prepublish): 100082.

[50] 厉静文, 董锁成, 李宇, 等. 中蒙俄经济走廊土地利用变化格局及其驱动因素研究[J]. 地理研究, 2021, 40(11): 3073-3091.

[51] 李迎双. 离子吸附型稀土矿区土地荒漠化遥感动态监测及驱动因素研究[D]. 江西理工大学, 2022.

[52] 高尚武, 王葆芳, 朱灵益, 等. 中国沙质荒漠化土地监测评价指标体系[J]. 林业科学, 1998, (02): 3-12.

[53] 杨达, 易桂花, 张廷斌, 等. 青藏高原植被生长季NDVI时空变化与影响因素[J]. 应用生态学报, 2021, 32(04): 1361-1372.

[54] Xiao X, Boles S, Liu J, et al. Characterization of forest types in Northeastern China, using multi-temporal SPOT-4 VEGETATION sensor data[J]. Remote Sensing of Environment, 2002, 82(2): 335-348.

[55] 邓小进, 井长青, 郭文章, 等. 准噶尔盆地地表反照率时空变化特征及其影响因素分析[J]. 干旱区研究, 2021, 38(02): 314-326.

[56] Cézar B, Da S, Albuquerque M, et al. Dynamics of land cover and land use in Pernambuco (Brazil): Spatio-temporal variability and temporal trends of biophysical parameters[J]. Remote Sensing Applications: Society and Environment, 2022, 25.

[57] Verstraete M, Pinty B. Designing optimal spectral indexes for remote sensing applications[J]. IEEE Trans Geoscience and Remote Sensing, 1996, 34(5): 1254-1265.

[58] Li Y, Li H, Xu F. Spatiotemporal changes in desertified land in rare earth mining areas under different disturbance conditions[J]. Environmental Science and Pollution Research, 2021, 28(23): 1-12.

[59] Iosif V. Applying the change vector analysis technique to assess the desertification risk in the south-west of Romania in the period 1984-2011[J]. Environmental monitoring and assessment, 2017, 189(10): 524.

[60] Ridder D, Tax D, Lei B, et al. Classification, Parameter Estimation and State Estimation: An Engineering Approach Using MATLAB [M]. Wiley Online Library, 2017.

[61] Wei W, Li Z, Xie B, et al. Spatial distance-based integrated evaluation of environmentally sensitivity for ecological management in northwest China[J]. Ecological Indicators, 2020, 118.

[62] 郭泽呈, 魏伟, 石培基, 等. 中国西北干旱区土地沙漠化敏感性时空格局[J]. 地理学报, 2020, 75(09): 1948-1965.

[63] 段少洁, 杨军耀. 加权欧氏距离法和模糊综合评价法在阳泉市主要河流水质评价中的比较分析[J]. 水电能源科学, 2020, 38(07): 53-56.

[64] 张娟. 绿洲开发对干旱区生态环境的影响评价[D]. 兰州大学, 2016.

[65] Wei W, Guo Z, Xie B, et al. Spatiotemporal evolution of environment based on integrated remote sensing indexes in arid inland river basin in Northwest China[J]. Environmental science and pollution research international, 2019, 26(13): 13062-13084.

[66] 姜阳, 房龙. 混淆矩阵算法在质检工作中的应用[J]. 经纬天地, 2019, (01): 5-7.

[67] Olfa C. Improved remote-sensing land cover identification using parameterized feedforward neural networks[J]. Euro-Mediterranean Journal for Environmental Integration, 2021, 6(1).

[68] 王劲峰, 徐成东. 地理探测器:原理与展望[J]. 地理学报, 2017, 72(01): 116-134.

[69] Wang J, Li X, Christakos G, et al. Geographical Detectors-Based Health Risk Assessment and its Application in the Neural Tube Defects Study of the Heshun Region, China[J]. International Journal of Geographical Information Science, 2010, 24(1): 107-127.

[70] 陈发虎, 吴绍洪, 崔鹏, 等. 1949—2019年中国自然地理学与生存环境应用研究进展[J]. 地理学报, 2020, 75(09): 1799-1830.

[71] 武增海, 李涛. 高新技术开发区综合绩效空间分布研究—基于自然断点法的分析[J]. 统计与信息论坛, 2013, 28(03): 82-88.

[72] Ma Z, Xie Y, Jiao J, et al. The Construction and Application of an Aledo-NDVI Based Desertification Monitoring Model[J]. Procedia Environmental Sciences, 2011, 10: 2029-2035.

[73] 白蓉. 我国新疆地区荒漠化现状、成因及对策的研究[J]. 中国林业经济, 2017, (02): 81-82.

[74] 魏江龙. 川西高原荒漠化时空变化特征及驱动力分析[D]. 成都理工大学, 2017.

[75] Eze J, Onokala P. Pattern of household vulnerability to desertification in Yobe state, Nigeria[J]. GeoJournal, 2021, (prepublish): 1-19.

[76] Rosanna S, Andrea C, Leonardo B, et al. ‘Old’ Territorial Disparities and ‘New’ Spatial Patterns: Unraveling the Latent Nexus between Sustainable Development and Desertification Risk in Italy[J]. Economies, 2022, 10(2): 50.