Study On Forest Fire Early Warning And Monitoring Methodology Using Remote Sensing And Geography Information System Techniques

Forest fire is a kind of worldwide natural calamity. It is extensively distributed with high occurrence frequency and destroys forest resources thus disturbing normal living order of people and leading to environmental deterioration. It has been paid more attention to by many governments. It's a study hotspot in worldwide how to using Modern high-tech information technologies, such as Remote Sensing (RS) and Geographic Information Systems (GIS), to forest fire danger rating predicting and forest fire monitoring. Based on the review of internationally and domestically published papers, Moderate spatial resolution RS data (MODERATE RESOLUTION IMAGING SPECTRO RADIOMETE, MODIS) and GIS Techniques have been used as main techniques in this application study on forest fire early warning and monitoring.Based on the review of internationally and domestically published papers on MODIS, research on how to process MODIS data and their application in forest fire danger prediction have been carried out. Main work includes comparison study of geometric correction method on MODIS data, Chinese national forest fuel distribution and classification using MODIS data, Chinese national vegetation status, Chinese vegetation burning information retrieval and forest fire danger rating. Application methods based on remote sensing data that use GIS as the platform and MODIS as the main information source have been developed in forest fire prediction. Summarization of this dissertation is as follows:(1) Approaches on geometric correction and fuse of MODIS LIB data have been obtainedThe advantages and disadvantages of using geometric correction to process MODIS LIB data by Georeference from Input Geometry (GLT) method and Geometry Control Point (GCP) method were discussed through the comparison study. A followed suggestion about Geometric Correction of MODIS LIB data has been given. Fuse method of different spatial resolution of MODIS LIB data has also been obtained and program of fuse implement has been developed.(2) Classification of Large Scale Forest Fuels Map Based on MODIS Data has been generated.Based on previous researches and achievements on the classification of vegetation, landcover, and land utilization, large scale forest fuels classification system based on MODIS data has been bring forward in China for the first time. Under this classification system, China national forest fuels class has been studied by using an integrative unsupervised and supervised classification method. A national forest fuels distribution map has been produced. At the same times, the classification's precision and its verification by middle spatial resolution satellite data has been discussed. The precision of this classification was verified by overlaying the classifying results with background thematic data in GIS.(3) The processing method of different forest fire danger rating factors has been obtainedData used by forest fire danger rating model was classified into static data and dynamic data based on whether these data were stable against time. After lab test, the processing method of these two types of data was obtained as the followings:a) For the static data, after it was corrected by the background database, it was digitalized before being inputted into the rating model program.b) For the dynamic data, three kinds of dynamic data resource have been used in the forest fire danger rating system. The first one is the daily weather observation data. Based on the precision and efficiency of three kinds of spatial interpretation method, the best way is the use of Invert Distance Weight (IDW) spatial interpretation method. IDW can predict the weather condition of the region where there is no weather observation station by using weather observation station data from other regions. The second one is the growth of vegetation. Relative greenness parameter has been used to estimate the forest fuels' growth. The third one is the moisture content of forest fuel, which can be estimated by an integrative method using near infrared reflectance (NIR) channels and short wave infrared reflectance (SWIR) channels. The result shows that the best way to estimate moisture content of forest fuels is using the rate of channel 2 to channel 7 of MODIS according to comparing the results of channel 2 to channel 5, channel 6 and channel 7.(4) National Forest Fire Danger Rating Prediction method has been obtainedIn this paper, the dynamic and static data related with forest fire danger rating prediction have been quantified and used to calculate fire danger index. The fire danger index is the quantified index and grading criterion for Chinese national forest fire danger rating prediction. Consequently, it helps the realization of describing forest fire danger rating from quality to quantification. The method has been verified by the tests in North East of China. The results demonstrated the RS and GIS based method is useful in forest fire prediction.(5) Method to retrieve the vegetation burning information using MODIS data has been obtainedBase on previous research achievement of fire monitoring using MODIS data, the information of related MODIS bands' data in daytime was analyzed and the accuracy of the result by using Surface Bright Temperature method and Bright Temperature-NDVI method was compared. A model was developed to identify forest fire by using Bright Temperature-NDVI method and integrating GIS technology. In experiment, the totally precision of identification was above 80%, which can satisfy the need of forest fire check-up. At the same time, it can also identify the vegetation type of fire.