Application Of Cold Disaster Monitoring And Predicting By Retrieval Of Land Surface Temperature

Guangdong province which locates in the area of tropical monsoon climate and subtropical monsoon climate, abounds with fruits, such as banana, litchi, longan, pineapple and mango. The cold weather is harmful to growing fruits whose production costs are expensive, and causes great loss. In fact cold weather has become the third natural disaster following flood and typhoon. Recently, many statistical methods have been applied into cold weather in winter in Guangdong. It is in first stage that regional cold disaster is monitored and predicted with high-technology approach, and there are few research and application of cold disaster with remote sensing technology.The paper starts with the basic theory of thermal infrared remote sensing. Then, the case study in the Shanwei city of Guangdong province uses ASTER data and MODIS data to monitor cold disaster by land surface temperature retrieval. We choose reference-channel method which is one classical method of Temperature/Emissivity Separation to retrieval transient LST in the early days of the cold disaster on 23rd, December 2004 by ASTER data, and choose split window method to retrieval almost simultaneous LST by MODIS data, too. Furthermore, the main economic crop is classified and extracted from study area ASTER data by remote sensing image distinguished, and the distribution map of banana growing is generated. In accordance with the main economic crop the quantitative research revealed the relationship of many factors, such as LST, NDVI, elevation and aspect. The results show ASTER based LST is more accuracy than MODIS based LST in coefficient between LST andobserved atmospheric temperature by 34%, observed Ocm ground temperature by 27%, and by 18% between LST and NDVI. With geostatistical methods and GIS technology, the forecast model of retrieval LST amended during cold disaster is established, which adds quantitative influence of terrain factors. And then, we can predict spatial distribution of corps suffered cold disaster and estimate the loss of cold disaster based on the model according to some assessment indexes of cold disaster.The research realized integration application of ASTER/MODIS data and land surface temperature retrieval methods, and achieves the innovation of cold disaster monitoring and predicting at the aspects of method and technology. The paper provides a timely and efficiently quantitative research approach on cold disaster monitoring and predicting in Guangdong winter, and provides scientific basis for loss monitoring and estimating of cold disaster. This achievement which can directly service cold disaster alleviation and precaution is of momentous current significance and is of obvious social and economic benefit.