基于GF-1卫星数据和P-M方法的灌溉需水量空间格局研究  

Spatial Pattern of Irrigation Water Demand Based on GF-1 Satellite Data and P-M Method

作　　者：王志东 程锐 宋妮[4] 王景雷[4] 韩其晟 赵西宁[2,5] WANG Zhidong;CHENG Rui;SONG Ni;WANG Jinglei;HAN Qishen;ZHAO Xining(College of Water Resources and Architectural Engineering,Northwest A&F University,Yangling 712100,China;Ministry of Education Key Laboratory of Agricultural Soil and Water Engineering in Arid and Semiarid,Yangling 712100,China;Chinese Hydraulic Engineering Society,Beijing 100053,China;Farmland Irrigation Research Institute,Chinese Academy of Agricultural Sciences/Key Open Laboratory of Crop Water Demand and Regulation,Xinxiang 453002,China;Institute of Soil and Water Conservation,Chinese Academy of Science&Ministry of Water Resources,Yangling 712100,China)

机构地区：[1]西北农林科技大学水利与建筑工程学院,陕西杨凌712100 [2]西北农林科技大学旱区农业水土工程教育部重点实验室,陕西杨凌712100 [3]中国水利学会,北京100053 [4]中国农业科学院农田灌溉研究所/农业农村部作物需水与调控重点开放实验室,河南新乡453002 [5]中国科学院水利部水土保持研究所,陕西杨凌712100

出　　处：《灌溉排水学报》2021年第6期105-112,共8页Journal of Irrigation and Drainage

基　　金：南疆绿洲区主要作物高效节水灌溉技术研究与应用(2016YFC0400208);中国农业科学院所统筹项目(FIRI2019-03-01)。

摘　　要：【目的】分析作物需水量与灌溉需水量时空分布规律,为新疆灌溉用水管理和最严格水资源管理制度的实施提供基础数据和技术支撑。【方法】通过构建多时相归一化植被指数(NDVI)曲线,提取南疆地区作物种植结构;基于GF-1数据构建NDVI多时相序列,采用监督分类方法提取阿克苏河绿洲区作物种植结构,并结合Penman-Monteith公式与GIS空间分析功能,分析绿洲区作物需水量与毛灌溉需水量的时空分布特征。【结果】多时相NDVI序列下,监督分类中最大似然法分类结果最优,总体精度达93.08%,Kappa系数为0.913。监督分类结果显示粮食作物(水稻,玉米,小麦)主要分布在流域上游的乌什县和温宿县;经济作物棉花主要分布在阿克苏市、阿拉尔市及农一师的部分农场;果树主要分布在阿克苏市和阿瓦提县。南疆阿克苏流域作物需水量和灌溉需水量时空分布存在显著性差异,时间上,春夏需水量占全年需水量70%左右;空间上,温宿县附近区域需水量和灌溉需水量明显高于其他地区。同种作物需水量在不同区域存在差别,其中棉花年需水量范围在704.4~808.8 mm之间,不同地区需水量差达100 mm左右。【结论】基于GF-1数据的多时相NDVI序列与监督分类结合的方法可以实现大区域作物种植结构的提取。阿克苏流域绿洲区作物需水量和灌溉需水量在空间和时间上均存在明显差异,需水量和灌水量的空间分布图可以为灌溉制度的制定和农业水资源利用的格局优化提供参考。【Objective】Analyzing the temporal and spatial distribution of crop water demand and irrigation water demand.This study provides basic data and technical support for irrigation water management and the implementation of the strictest water resources management system in Xinjiang.【Method】We constructe the NDVI multi temporal sequence was based on the GF-1 satellite data.Used the supervised classification method to extract the crop planting structure in Aksu River Oasis.Combined with penman Monteith formula and GIS spatial analysis function,the temporal and spatial distribution characteristics of crop water demand and gross irrigation water demand in Oasis were analyzed.【Result】Under the multi temporal NDVI sequence,the maximum likelihood method was the best,with an overall accuracy of 93.08%and Kappa coefficient of 0.913,which could meet the requirements of large area classification.The results showed that the grain crops(rice,corn,wheat)were mainly distributed in Wushi and Wensu in the upper reaches of the basin,economic crops(cotton)were mainly distributed in Aksu,Alar and some farms of the first agricultural division,and fruit trees were mainly distributed in Aksu and Awati.The temporal and spatial distribution of crop water requirement and irrigation amount in the watershed is quite different.In terms of time,the water demand in spring and summer accounted for about 70%of the annual water demand;in terms of space,the water demand and irrigation amount in the area near Wensu county were significantly higher than those in other areas.At the same time,the water demand of the same crop is also different in different regions.The range of water demand of cotton is 704.4~808.8 mm,and the difference of water demand in different regions is about 100 mm in crops.【Conclusion】The method of combining multi temporal NDVI and supervised classification based on GF-1 data can realize the extraction of crop planting structure in large area.There are obvious differences in crop irrigation requirement in space and

关 键 词：阿克苏河绿洲区 GF-1 NDVI P-M 作物需水量 时空分布 

分 类 号：S274[农业科学—农业水土工程]