Planting Area Monitoring And Yield Estimation Of Rice In Jiangsu Province Based On Remote Sensing And Meteorological Data

As the proverb goes, "The country takes people as origin, food is god to the people". Under the circumstance of drastic changes in climatic conditions and economic situation, timely and effective implementation of monitoring crop growth status and yield estimation in China and the world’s major grain-producing areas closely relates to the issue of food security. China, as the world’s most populous country, the first rice producer and consumer country brings forward high request for monitoring research on rice growth status. Against this background, the present study take the rice monitoring by remote sensing in Jiangsu province as the research target, By using the advantages of remote sensing in objective timeliness, macro-scale, economic benefits et al., rice planting areas were extracted and rice yields were estimated, which may provide evidences for a more rational development of agricultural policy, macro-control of food prices, and food crisis alleviation.In this paper we used MOD09A1 data to extract the rice area of Jiangsu province in 2004-2010. Secondly, we used Landsat 8 OLI, a moderate spatial resolution data, to extract the rice area of Jiangsu province. The Landsat data extraction results were adopted as the reference to evaluate the matching degree of spatial location extracted by MOD09A1 data. The EVI values at different growth stages of rice during 2004-2006 and statistical rice yield were combined to establish yield estimation model. The model was used to estimate rice yield of Jiangsu province in 2007. Eventually, the yield map at administrative unit was converted to the counterpart at pixel-level. Thirdly, we combined MOD15A2 LAI data with four meteorological data, low temperature, high temperature, relative humidity and downward shortwave radiation to simulate daily GPP of rice area using a two-leaf rectangular hyperbolic model. The GPP results were used to estimate rice yield of Jiangsu province.Main achievements and innovations of this paper were as follow:1) The results illustrated that the rice yield extracted by MODIS data had an acceptable accuracy, while it didn’t match well on the spatial distribution of the overall area. When we modified the parameters of rice extraction algorithm, making the rice area of MODIS data extraction consist with statistical data to analyze the matching degree of spatial location, we found there was a large error on spatial matching, especially in Nanjing and Qidong. But the main rice planting areas have been extracted.2) The estimation accuracy of rice yield at the municipal level was around 95%. The spatialization of rice yield from the municipal level will refine the regional rice yield estimation results for more intuitive understanding of the paddy rice production information.3) Add maximum carboxylation rate at 25℃ parameters (Vcmax,25) was introduced to build a new two-leaf rectangular hyperbolic model to estimate daily GPP and subsequent rice yield.