基于遥感的东北地区水稻延迟型冷害动态监测

Dynamic monitoring rice delayed-type chilling damage based on remote sensing in northeast China

水稻延迟型冷害大范围准确的同步跟踪监测,可为提高稻米品质和产量提供帮助。在明确东北地区水稻种植区域和水稻低温冷害年的基础上,基于MODIS MOD13A2数据产品划分了东北地区水稻各发育普遍期植被指数判识指标的参考范围,构建了基于地理信息的水稻发育期模型,在此基础上制作了水稻静态发育期空间分布图,并以2009年5月24日和8月7日为例开展了东北地区水稻延迟型冷害动态监测的应用研究。结果表明:东北三省发生较大范围水稻延迟型冷害的年份有2002年、2003年、2008年和2009年,其中黑龙江省是水稻延迟型冷害的重灾区;EVI在水稻发育期判识中较NDVI更具优势;利用经度、纬度和海拔高度与日序的相关关系构建水稻发育期模型,推算无观测地区的发育期具有一定的可行性;在同一时间东北三省各地的水稻发育期分布推算结果的基础上,实现了2009年5月24日和8月7日东北地区水稻延迟型冷害动态监测,所得监测结果与2009年农业气象观测记录在时间和空间上有很好的一致性,监测结果正确。

Rice chilling damage remains one of the major agricultural disasters in northeast China. Remote sensing technology can easily monitor this disaster on a relatively large scale. Satellites and ground multi-source data were used in this study to monitor rice delayed-type chilling damage by combining with extraction results of rice planting area, growth stage calculation and spatial distribution of static development stage. Based on these results, we can provide technical supports for the agricultural production sectors to monitor rice chilling damage in the wide area and to improve rice quality and yield. Using remote sensing data, daily mean temperature of 196 meteorological stations located in northeast China from 2000 to 2009, rice development stages and related geographic information data, the degree of rice chilling damage and main year of chilling damage were analyzed according to technical specifications for evaluation of rice cold damage. The information of rice planting area was extracted by the products of MODIS MCD 12Q 1 and Landsat data. Normal differential vegetation Index （NDVI） and Enhanced Vegetation Index （EVI） were acquired from the products of MODIS MCD 12Q1. The reference range of NDVI and EV1 for development stages of rice was divided by the method of statistical analysis. According to the correlation between Julian days office development stages and geographical factors, the regression equations office development stages were constructed. On the basis of the equations, daily distribution maps of development stages were obtained by a piece of sliding filter board through ENVI＋IDL. Finally, monitoring of rice delayed-type chilling damage was done based on remote sensing in northeast China on May 24, 2009 and August 7, 2009. The results showed that the large-range distribution of damage happened in 2002, 2003, 2008 and 2009 in northeast China. Heilongjiang province was particularly affected by the disaster. Also, both of NDVI and EVI values of rice in each development stage obeyed the normal distribution. But the accuracy rate of EVI values in determining the rice development stages was over 76%. Obviously, EVI values were more effective than NDVI values in detecting rice phenology. In addition, rice development stage models were established according to the relationship between longitude, latitude, altitude and day series. The dates of simulated rice development stages were earlier than the actual observation dates as a whole with an error of 1 to 3 days. And the difference days in simulated results and observed values were much smaller than that between two rice development stages. The estimate accuracy can meet the need of the research. Moreover, on the basis of models, we made the spatial distribution maps of every static development stage and answered how to confirm the ownership of the rice development stages by infinite function. Finally, rice delayed- type chilling damage was dynamically monitored in two days by using this method. The monitoring results had a better consistence with the observational records in time and space.