2016-2023年松嫩平原西部盐碱地水体数据集  

A dataset of water bodies on the saline-alkali land in western Songnen Plain from 2016 to 2023

作　　者：刘昊泽 隗晓琪 李含含 黄泽晖 薛坤[1,2] LIU Haoze;WEI Xiaoqi;LI Hanhan;HUANG Zehui;XUE Kun(Key Laboratory of Lake and Watershed Science for Water Security,Nanjing Institute of Geography and Limnology,Chinese Academy of Sciences,Nanjing 210008,P.R.China;Key Laboratory of Watershed Geographic Sciences,Nanjing Institute of Geography and Limnology,Chinese Academy of Sciences,Nanjing 210008,P.R.China;University of Chinese Academy of Science,Beijing 100049,P.R.China;School of Remote Sensing and Geomatics Engineering,Nanjing University of Information Science and Technology,Nanjing 210044,P.R.China)

机构地区：[1]中国科学院南京地理与湖泊研究所,湖泊与流域水安全重点实验室,南京210008 [2]中国科学院南京地理与湖泊研究所,中国科学院流域地理学重点实验室,南京210008 [3]中国科学院大学,北京100049 [4]南京信息工程大学,遥感与测绘工程学院,南京210044

出　　处：《中国科学数据(中英文网络版)》2025年第1期189-202,共14页China Scientific Data

基　　金：黑土地保护与利用科技创新工程专项(XDA28110503);国家自然科学基金(42371371);全球变化背景下中国区域湖泊响应数据库(CAS-WX2021SF-0306);中国科学院网络安全和信息化专项(CAS-WX2022SDC-SJ05)。

摘　　要：由于自然环境和人为等因素的影响,松嫩平原西部地区具有大面积的盐碱地,是世界三大苏打盐碱地的主要分布区之一。水资源在人类生产生活和自然环境中起着至关重要的作用,对于生态环境脆弱的松嫩平原西部盐碱地地区来说更是如此。本文以Google Earth Engine(GEE)平台上的Sentinel-2A/B MSI为数据源,选取松嫩西部水体提取的样本点,采用随机森林分类方法生产了2016–2023年松嫩平原西部月度水体数据集,并生成2016–2023年松嫩平原西部的水体频率数据。由于2019年以前经过大气校正的大气底层反射率数据产品Level-2A并未覆盖研究区,因此2019–2023年采用的是Level-2A数据产品,2016–2018年使用经过校正的大气表观反射率数据产品Level-1C进行补充。经过同步数据提取结果的对比,使用Level-1C数据产品对Level-2A数据产品缺失的时间范围进行补充对水体的分类结果影响不大。本研究随机森林模型分类的总体精度为91.96%,Kappa系数为0.9128,该结果表明生成的月度水体数据集具有良好的精度和可靠性,可为松嫩西部的水资源监测、盐碱地治理、盐碱地高效治理模式的评价和辐射推广提供数据支持和帮助。Due to natural and anthropogenic factors,western Songnen Plain has vast areas of saline-alkaline land,making it one of the world’s three major distribution areas of soda saline-alkaline land.Water resources play a crucial role in both human production and living activities and the natural environment,and this is especially true for the ecologically fragile saline-alkali land in western Songnen Plain.In this paper,we used Sentinel-2A/B MSI data on the Google Earth Engine(GEE)platform as the data source,selected the sample points extracted from the water bodies in western Songnen Plain.Then,we used Random Forest Classification to produce a dataset of monthly water bodies and generate the water body frequency data in western Songnen Plain from 2016 to 2023.Since the atmospherically corrected Level-2A did not cover the study area prior to 2019,the Level-2A data product was used for the period from 2019 to 2023.For the years 2016-2018,the corrected atmospheric apparent reflectance data product Level-1C was used as a supplement.After comparing the results of simultaneous data extraction,the use of Level-1C data product to supplement the missing timeframe of Level-2A data product had little effect on the classification results of water bodies.The overall accuracy of the random forest model classification in this study was 91.96%,and the kappa coefficient was 0.9128.This indicates that the dataset is highly accurate and reliable.The dataset can provide data support for water resource monitoring,saline-alkali land management,and the evaluation and promotion of efficient management models for saline-alkali in western Songnen Plain.

关 键 词：松嫩西部 盐碱地 水体 随机森林 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论] P332[自动化与计算机技术—计算机科学与技术]