【作者】 李诚志；

【导师】 刘志辉；

【作者基本信息】 新疆大学 ， 自然地理学， 2012， 博士

【摘要】 沙漠化是当前世界上最严重的土地退化问题之一，对沙漠化的监测与预警是沙漠化防治的基础，研究意义重大。由于沙漠化的成因复杂，相关观测数据的匮乏，迄今为止对沙漠化的监测还不充分，而对沙漠化的预警才刚刚开始，值得进一步深入研究。本文针对沙漠化监测与预警的研究主题，从沙漠化监测与预警的基本理论出发，运用先进的技术手段，构建沙漠监测与沙漠化预警模式。以新疆作为研究区，对新疆的沙漠化进行了长期的监测与预警研究。得出的主要结论如下：（1）通过对沙漠化内涵的整理，沙漠监测指标的归纳、总结与帅选，沙漠程度分级体系的对比，以及沙漠化预警的方法与模式的比较研究，初步形成沙漠化监测与预警理论体系。（2）研制了一套全自动风沙监测仪器，实现了沙通量、风速、风向、温度数据的自动采集与远距离的数据传输，实现了沙漠化风沙活动的远距离自动监测、自动存储、自动分析。（3）选取沙漠、固定沙丘、戈壁、草地、森林、耕地六种典型地类，对MODISNDVI数据监测沙漠化的可行性进行了研究，其结果表明：低植被盖度的NDVI数据各月份、各年份的稳定性都很高，而高植被盖度的NDVI因月份、年份的不同而存在差异，不能满足监测沙漠化变化的要求。进而对4~9月份的最大NDVI进行合成，并对其稳定性进行分析发现，该数据比较稳定，满足沙漠化监测要求。（4）以4~9月份的合成最大MODIS NDVI数据，以流动沙地、半流动沙地、固定沙地、平沙地、盐碱地、戈壁作为新疆沙漠化监测范围，利用像元二分模型反演植被覆盖度，构建新疆沙漠化的常规年度监测模式，实现了沙漠化年度监测。（5）利用灰色GM（1,1）预测模型，结合GIS栅格运算功能，创建了栅格灰色GM（1,1）预测模型，并对沙漠化变化的关键性因子NDVI进行了预测，其预测结果显示：该模型预测的总精度在5%以内，并利用此模型对新疆2012、2013、2014、2015、2020年的NDVI进行了预测。（6）创建了沙漠化栅格累加预警模型，并利用沙漠化监测与预测的结果对新疆2012、2013、2014、2015、2020年的沙漠化状态进行预警，结果发现南疆沙漠化的警情大于北疆，其中沙漠化警情最大是塔里木河中游，特别是英巴扎到恰拉段，另外在乌鲁木齐市的达坂城区，乌伦古湖的西北侧和西南侧，车儿臣河中游，策勒县、于田县北侧，且末县西部，麦盖提县西部，巴楚县西部均出现不同程度的沙漠化警情。（7）在沙漠化累加预警模型的基础上，以人类自身为视角，考虑沙漠化致灾因子、土地产值，创建了沙漠化损失风险预警模型。并对新疆的沙漠化损失风险进行预警，其结果显示：沙漠化损失风险预警确定的警情比沙漠化状态预警更加偏向于人类活动，更适合于沙漠化治理与防治。（8）运用GIS技术、WEBGIS技术和.NET技术构建了新疆沙漠化监测与预警系统，实现了新疆沙漠化监测与预警的自动化，并将其监测与预警的结果进行网络发布，实现了研究成果共享。更多还原

【Abstract】 Desertification is one of the most serious Land degradation problems in globaltoday. The premise to prevent and control desertification is monitoring andearly-warning desertification. However, due to the complex causes of desertification,the lack of observation data, and so far, the monitoring of desertification is not sufficient,and the early warning of desertification is just beginning, it needs to further research.This paper aimed at the key scientific problems of the monitoring and earlywarning desertification. Through using the basic theory of the monitoring and warning,utilizing the advanced technology of GIS, the paper constructed the monitoring andwarning model of desertification, and this model has be used in Xinjiang, as thelong-term monitoring and early-warning of desertification research. The majorconclusions obtained from the paper as follows:(1) Through sorting the connotation of desertification, summarizing and choosingthe monitoring indices of desertification, comparing the classification systems ofdesertification degree, and contrasting the early warning methods and models used indesertification, the paper produced the initial monitoring and early warning theorysystem of desertification.(2) The paper manufactured a set of automatic sand-moving monitoring instrument.The instrument can automatically gain the data of sand flux, wind speed, wind directionand temperature obtained from the instrument. It can remotely transmit the date fromfield to the monitoring center, and the data can be automatically saved to the data-baseat the monitoring center, and the data can be automatic analysis. Through the instrument,the sand activities of desertification realized the distance automatic monitoring.(3) By selecting six typical land types that is desert, fixed dune, gobi, grassland,forest, and cultivated land, this paper researched the MODIS NDVI data stability thatinfluences the monitoring desertification feasibility. The research results showed that the NDVI stabilities of low vegetation coverage were very high monthly and yearly, theNDVI stabilities of high vegetation coverage were shifty with the different year and thedifferent month, the primary data did not satisfy the requirement of monitoringdesertification changes. Then it composed the yearly biggest NDVI with the monthsNDVI from April to September, and testing the composed data stability, it founded thatthe data stability improved, it met the requirements of monitoring desertification.(4) It utilized the yearly biggest MODIS NDVI data composed by the monthsNDVI from April to September, and it established the flow sand land, the half flow sandland, the fixed sand land, the gobi land, and the saline-alkali land as the monitoringdesertification ranges. Using the binary inversion model to invert the vegetationcoverage, it constructed the annual conventional monitoring desertification mode, andrealized the annual monitoring desertification in Xinjiang, China.(5) By using the grey GM(1,1) prediction model and combining with rasteroperation function of GIS, the paper created a grid gray GM(1,1) prediction model.Used the grid gray GM(1,1) prediction model, it forecasted the key factor (NDVI) ofdesertification changes. The forecast results showed that the total model predictionaccuracy was within5%. Using the model, it forecasted the NDVI in2012,2013,2014,2015,2020, Xinjiang.(6) The study created the grid accumulated desertification early warning model.Used the model and the monitoring and predicting desertification results in2012,2013,2014,2015,2020, it published the desertification warning degrees in Xinjiang. Itfounded that the desertification warning degrees in southern Xinjiang was greater thanthe degree in northern Xinjiang. The largest alert of desertification was the area in themiddle reaches of Tarim River, especially in region from Yingbaza to Qala. In theDabancheng, Urumqi, in the northwest and southwest sides of the Ulungur Lake, in themiddle reaches of the Qarqan River, the north of CeLe county, the north of Yutiancounty, in the west of Qiemo county, in the west of Makit county, and in the west ofBachu county, all appeared different alerts of desertification. (7) Based on the perspective of human beings selves, considered the hazard factorof desertification and the output value of land, and integrated the grid accumulateddesertification early warning model, this study also founded the desertification loss riskearly warning model. The desertification loss risk early warning model was used toalarm the desertification in Xinjiang. The results showed that the warning determined bythe loss risks model of desertification more tended to human activities than the gridaccumulated desertification early warning model. The results were more suitable forprevention and control of desertification.(8) Used GIS technology, WEBGIS technology and internet technology, the studyconstructed the desertification monitoring and early warning system in Xinjiang. Itrealized the automatic monitoring and early warning of desertification in Xinjiang, theresults of the desertification monitoring and early warning was published in network. Itrealized the sharing research results.更多还原