Study On The Yield Of Maize And Chilling Damage In Northeast China Based On Remote Sensing Crop Model

Northeast China is the biggest export base of commodity grain in China, of which maizeproduction accounts for about40%of the national maize production. Due to the lack of heatresource, maize production is vulnerable to chilling damage. Thus, real-time, accurate andquantitative impact assessment of agro-meteorological disasters on the crops yields hasimportant practical significance. Because of spatial heterogeneity, crop growth modelsbased-on-situ data faced up-scaling problems when applied to the monitoring of large-scalecrop production. In recent years, the complementarity of remote sensing technology and cropgrowth models and their development offer a possibility for their combination, and simple,easy to operate coupled methods are driving methods. The model named RS-P-YEC(Remote-Sensing-Photosynthesis-Yield Estimation for Crops) applied by the paper is aprocess-based remote sensing mechanism model. The C3metabolic pathway in thephotosynthetic sub-module of the RS-P-YEC model was modified to the C4metabolicpathway, and the yield of C4crop is calculated by multiplying the net primary productivity(NPP) by the harvest index (HI) derived from the ratio of grain to stalk yield. Firstly, thepaper extracted the acreage of maize in Northeast China and the spatial and temporalvariation of maize yield was simulated as well. The impact of chilling damage of the site,local and regional scales to the yield of maize was simulated according to the temperaturetests of phases, revision of harvest index and ground-state yield. The main research work andconclusions are as follows：(1) MODIS data product MCD12Q1, land utilization sets of1:1,000,000from ChineseAcademy of Sciences and digital elevation data were used to extract the acreage of maizein Northeast China. The statistical area and the extracted area were compared andcorrelation coefficients of Heilongjiang, Jilin and Liaoning provinces were0.67,0.74and 0.91, respectively, all of which were significant at0.01.(2) The leaf area index generated from S-G filtered surface albedo reflected the fieldexperiments much better than MODIS LAI, more suitable for the simulation of crop yieldby the remote sensing process model. The meteorological data of72meteorologicalstations and remote sensing were employed to drive the improved RS-P-YEC model tosimulate the yield of maize in Northeast China from2002to2011year. The111statisticaldata and simulated yields from study area were verified at county-leval, with thecorrelation coefficient (R) up to0.827, root mean square error (RMSE) of712kg/hm2andrelative error (RE) of9.3%, indicating that the improved model was suitable for yieldsimulation of maize in Northeast China. According to the sensitivity analysis of modelinputs, the effect of temperature on yield among meteorological elements was the highest,suggesting that the improved model could simulate the impact of temperature on the yieldof maize in Northeast China. The mean yield of mazie in Northeast China mainly rangedfrom4000to8000kg/hm2in early2000s. The spatial pattern of the yield of maize wasconsistent with leaf area index. The spatial distribution of yield was related to the meantemperature and total precipitation of the major growing season of maize (May toSeptember). In the10years, the yield was mainly increasing (p<0.10), and the averagecoefficient of variation was0.49, with the highest year of2010, and the lowest year of2009.(3) The simulated results derived by in-situ stations were consistent with the experimentalresults. The impact of local chilling damage to the maize yield simulated through revisingharvest index was concordant with literatures. When simulating chilling damage ofregional scale, the simulated yields of normal years derived by the mean temperatures ofyears were compared with the simulated yields derived by the actual temperatures. Theyield-reduced rate derived the yield gap was used as chilling damage index tocharacterize the spatial distribution of chilling damage. The cold damage leveldiscriminated by the yield-reduced rate of4typical disaster years were consistent withhistorical records, that was,1986and1989were the general chilling damage years, and1992and1995were the severe chilling damage years.