Map of irrigated cropland in northern China at 30m resolution from 1990 to 2020

High-resolution irrigated cropland maps are essential for optimizing agricultural productivity and managing freshwater resources. China has the world’s largest irrigated cropland area, but rapid irrigation expansion in recent decades has exacerbated regional water stress. We develops an automated machine learning framework to map irrigated croplands over the past four decades at 30 m resolution across northern China, a region that accounts for more than 70% of China’s irrigated cropland and is highly susceptible to water shortages. The main contributions of our framework include: (1) delineating irrigated areas by analyzing the time-series characteristics of irrigated crops by constructing a composite irrigation performance index based on greenness, humidity, and temperature, and refining the sample pool with two specific irrigated crop types; (2) incorporating synthetic phenological features and a 30 m surface temperature dataset to capture the cooling effect of irrigation; (3) leveraging 98,697 Landsat scenes to train 673 localized random forest models per time period to enhance classification accuracy across diverse regions. Validation results indicate that the overall accuracy of the generated irrigated cropland maps ranges from 0.78 to 0.85. Produced maps show strong agreement with statistical data and outperform three existing irrigation products. The 30m spatial resolution map of irrigated cropland also captures the loss of irrigated cropland around the urban, highlighting the increased competition for regional water resources. By minimizing reliance on statistical data and integrating irrigation-induced cooling effects and phenological variations, our approach improves mapping accuracy and provides a valuable resource for regional irrigation planning and water resource management.Coordinate system EPSG:4326Value 0: non-cropland; Value 1: non-irrigated; Value 2: irrigated

高分辨率灌溉农田地图对于优化农业生产效率与淡水资源管理而言至关重要。中国拥有全球规模最大的灌溉农田面积，但近数十年间灌溉面积的快速扩张加剧了区域水资源压力。本研究构建了一套自动化机器学习框架，以30米分辨率绘制中国北方地区过去四十年的灌溉农田分布图——该区域占中国全国灌溉农田总面积的70%以上，且极易遭受水资源短缺问题的影响。本框架的核心贡献如下：(1) 基于植被绿度、湿度与温度构建综合灌溉表现指数，通过分析灌溉作物的时间序列特征划定灌溉区域，并结合两类特定灌溉作物类型优化样本池；(2) 融合合成物候特征与30米分辨率地表温度数据集，以捕捉灌溉带来的降温效应；(3) 利用98697景Landsat影像，针对每个时段训练673个本地化随机森林模型，以提升不同区域的分类精度。验证结果显示，所生成的灌溉农田地图总体精度介于0.78至0.85之间。生成的地图与统计数据吻合度极高，且优于三款现有灌溉制图产品。这款30米空间分辨率的灌溉农田地图还能够识别城市周边灌溉农田的流失情况，凸显出区域水资源竞争的加剧。本方法通过降低对统计数据的依赖，融合灌溉诱导的降温效应与物候变化特征，提升了制图精度，可为区域灌溉规划与水资源管理提供极具价值的参考数据。坐标系：EPSG:4326；像素值说明：0代表非农田；1代表非灌溉农田；2代表灌溉农田

搜集汇总

背景与挑战

背景概述

该数据集提供了中国北方1990年至2020年期间30米空间分辨率的灌溉农田地图，通过自动化机器学习框架（利用随机森林模型和Landsat卫星数据）生成，总体分类精度为0.78-0.85，旨在支持农业水资源管理和区域规划。数据集包含四个时间点的TIF文件，覆盖中国70%以上的灌溉农田区域，并整合了物候和温度特征以提升映射准确性。