摘要
农作物种植类型的地理分布差异,气候条件差异、土壤环境不同等因素的影响,需要开展农作物生长模型参数区域化、本地化的研究工作;通过改善区域气象数据空间化方法以提升插值精度的研究,也需要得到应用的重视。针对以上问题,该文以SWAP(soil-water-atmosphere-plant model,土壤-水-大气-作物模型)模型为基础,以中国黑龙江省南部地区作为研究区域,以其主要农作物春玉米为目标作物,确定研究春玉米的作物生长模型参数,并综合考虑纬度及海拔对气温的影响情况,研究将协同克里金(coKriging)方法引入作物生长模型气象数据插值获取中,从而提高模型输入参数中气象数据精度,并以叶面积指数(leaf area index,LAI)及蒸散发(evapotranspire,ET)数据作为同化遥感数据源,通过优化玉米灌溉量和出苗日期,获取了研究区2013年的玉米产量空间分布成果,与统计资料结果对比,玉米总产量监测结果的R2达到了0.939 4,均方根误差(root mean squared error,RMSE)达到了148 065 t,平均绝对误差(mean absolute error,MAE)为114 335 t。研究区15个县市区的预测单产和统计单产之间的决定系数达到了0.724 5,RMSE为598.5 kg/hm^2,MAE为531.5 kg/hm^2。研究结果表明,利用SWAP模型,以协同克里金方法获取气象数据空间插值成果作为输入数据,通过同化LAI和ET遥感数据,可以有效进行黑龙江南部区域的玉米产量遥感监测,为区域作物生长及生产力的遥感监测预测提供参考。
Crop yield monitoring and forecast has great impact on food security, ecological environment, and farmers’ incomes. Crop growth and yield monitoring and forecast by using crop growth model has the advantages of clear mechanism, high precision, and high monitoring frequency, but its monitoring scale is usually limited to land block level. Along with the fast development of remote sensing satellite technology, using remote sensing data combined with crop growth model to accurately monitor regional crop growth and yield in a large regional scale has gradually become an important means of regional level and even national level crop growth monitoring. However, due to differences in crop types, climatic conditions, soil conditions and monitoring areas, the regionalization and localization of crop growth model is the major bottleneck of crop growth monitoring by using crop growth model combined with remote sensing data, and it is urgent to conduct targeted studies on the identification of assimilation parameters of crop growth model, pre-processing of meteorological data, and the setting of crop parameters. Based on soil-water-atmosphere-plant model(SWAP), and by taking the major commodity grain production base of China, Northeast China Region as a study region, in this the paper, we conducted a study by taking the major crop of spring maize of the region as its target crop. Firstly, we used Landsat to obtain maize(Zea mays) planting area in the study area, and used it as the basic data for estimating the total maize yield in the study area. The overall accuracy of maize area classification was 93.2%, with R2 of 0.951 2. By considering the influence of latitude and altitude on temperature, in the study, we used the co Kriging method in crop growth model meteorological data interpolation acquisition, so as to improve the precision of input parameters of the model. The result showed that the average standard error of minimum temperature of coKriging method was 0.31 ℃, while that of the Kriging method was 1.51 ℃. The
作者
王利民
姚保民
刘佳
杨玲波
杨福刚
Wang Limin;Yao Baomin;Liu Jia;Yang Lingbo;Yang Fugang(Institute of Agricultural Resources and Regional Planning,Chinese Academy of Agricultural Sciences,Beijing,100081;Colleage of Environmental and Resource Sciences of Zhejiang University,Hangzhou 310012)
出处
《农业工程学报》
EI
CAS
CSCD
北大核心
2019年第22期285-295,共11页
Transactions of the Chinese Society of Agricultural Engineering
基金
国家重点研发计划“粮食作物生长监测诊断与精确栽培技术”课题“作物生长与生产力卫星遥感监测预测”(2016YFD0300603)
关键词
作物
遥感
模型
玉米
叶面积指数
蒸散发
协同克里金
同化
SWAP模型
crops
remote sensing
models
corn
leaf area index (LAI)
evapotranspiration
co Kriging
assimilation
SWAP model
