基于GEE平台和生育期差异的巴基斯坦主要农作物分类数据集（2000-2020）

巴基斯坦是传统农业国家，农业经济占国民总经济较高，水资源短缺及水分利用效率提升是影响地区农业发展的关键问题。获取长时序、准确的作物空间分布信息，不仅为作物生产、估产等提供基础信息，还对区域内水资源的调配管理起重要意义，从而提高农业用水效率，缓解农业用水压力。本数据集以巴基斯坦为研究区，运用GEE平台，结合MODIS遥感数据、植被特征指数和作物物候期构建决策树模型，提取2000-2020年巴基斯坦主要农作物类型，通过GEE平台和地面验证数据对巴基斯坦农作物类型进行精度验证，作物种植提取的准确率均在73%以上，具有较高的精度，Kappa系数为0.70，小麦、玉米、水稻、棉花和甘蔗的相关系数分别为0.73、0.89、0.81、0.70、0.75、总面积相关性为0.70，结果具有可信度。Pakistan is a traditional agricultural country, where the agricultural economy accounts for a large share of its gross national economy. Water resource shortage and the improvement of water use efficiency are key issues restricting regional agricultural development. Obtaining long-time-series, accurate spatial distribution information of crops not only provides basic data for crop production and yield estimation but also plays a critical role in the allocation and management of regional water resources, thereby improving agricultural water use efficiency and alleviating agricultural water stress. This dataset takes Pakistan as the study area, uses the Google Earth Engine (GEE) platform, and combines MODIS remote sensing data, vegetation characteristic indices and crop phenology to construct a decision tree model for extracting major crop types in Pakistan from 2000 to 2020. The crop types in Pakistan were verified for accuracy via the GEE platform and ground validation data. The overall extraction accuracy of crop planting areas exceeded 73%, indicating high precision, with a Kappa coefficient of 0.70. The correlation coefficients for wheat, maize, rice, cotton and sugarcane are 0.73, 0.89, 0.81, 0.70 and 0.75 respectively, and the correlation coefficient for total crop area is 0.70, demonstrating that the results are reliable.

搜集汇总

背景与挑战

背景概述

该数据集是基于Google Earth Engine平台，结合MODIS遥感数据、植被特征指数和作物物候期差异，通过决策树模型生成的巴基斯坦主要农作物分类数据，时间覆盖2000至2020年（包含五个年份：2000、2005、2010、2015、2020），空间分辨率为100米至1公里，作物类型包括小麦、水稻、玉米、棉花和甘蔗，并区分了冬季（Rabi）和夏季（Kharif）季节。数据集经过精度验证，总体准确率在73%以上，Kappa系数为0.70，具有较高的可靠性，旨在支持农业水资源管理和作物估产等应用。