High-resolution global soil salinity and sodicity mapping (1980–2024): Box-Cox-based sample optimization, multi-source remote sensing features, and uncertainty quantification

IF 11.4 1区 地球科学 Q1 ENVIRONMENTAL SCIENCES
Tiantian Wang , Jinwei Dong , Binyan Lyu , Xuan Gao , Nan Wang , Zhou Shi
{"title":"High-resolution global soil salinity and sodicity mapping (1980–2024): Box-Cox-based sample optimization, multi-source remote sensing features, and uncertainty quantification","authors":"Tiantian Wang ,&nbsp;Jinwei Dong ,&nbsp;Binyan Lyu ,&nbsp;Xuan Gao ,&nbsp;Nan Wang ,&nbsp;Zhou Shi","doi":"10.1016/j.rse.2025.114991","DOIUrl":null,"url":null,"abstract":"<div><div>Global soil salinization and sodification threaten food security by causing excessive salt accumulation in soils, degrading productivity. However, their spatiotemporal patterns have not been well documented due to the lack of high-accuracy estimates of soil salinity and sodicity with a long-term perspective. Here we propose a novel framework combining the Box-Cox transformation that addresses the skewed distribution of samples and more critical predictors as well as remote sensing indices, to predict global soil salinity (electrical conductivity of the saturated soil extract, ECe) and sodicity (exchangeable sodium percentage, ESP) from 1980 to 2024 with a 1 km × 1 km resolution using random forest. Model accuracy is higher than previous studies, with root mean square error (RMSE) of ECe and ESP as 2.24 and 6.04 respectively, and an <em>R</em><sup>2</sup> of 0.65, and 0.60 respectively. The implementation of the Box-Cox transformation significantly improved the model performance (<em>R</em><sup>2</sup>) by approximately 100 % (from 0.35 to 0.79 for ECe and 0.59 to 0.85 for ESP), while the additional predictors further enhanced the performance (<em>R</em><sup>2</sup> increased by 15 %), ranking in the top 30 % of the feature importance list. Results revealed that global multiple-year average salinization and sodification are primarily concentrated in arid regions characterized by low precipitation and high temperatures. We also found a significant increasing trend of soil salinization in 20 % of global land and of sodification in 48 % of global land from 1980 to 2024, both most pronounced near the equator, as well as in central and eastern North America, Europe, southeastern China, and Mongolia. This study provides updated long-term soil salinity and sodicity maps with improved accuracies, offering critical insights for sustainable land management under climate change, serving as an essential resource for addressing food security and land degradation challenges worldwide.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"330 ","pages":"Article 114991"},"PeriodicalIF":11.4000,"publicationDate":"2025-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Remote Sensing of Environment","FirstCategoryId":"5","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0034425725003955","RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENVIRONMENTAL SCIENCES","Score":null,"Total":0}
引用次数: 0

Abstract

Global soil salinization and sodification threaten food security by causing excessive salt accumulation in soils, degrading productivity. However, their spatiotemporal patterns have not been well documented due to the lack of high-accuracy estimates of soil salinity and sodicity with a long-term perspective. Here we propose a novel framework combining the Box-Cox transformation that addresses the skewed distribution of samples and more critical predictors as well as remote sensing indices, to predict global soil salinity (electrical conductivity of the saturated soil extract, ECe) and sodicity (exchangeable sodium percentage, ESP) from 1980 to 2024 with a 1 km × 1 km resolution using random forest. Model accuracy is higher than previous studies, with root mean square error (RMSE) of ECe and ESP as 2.24 and 6.04 respectively, and an R2 of 0.65, and 0.60 respectively. The implementation of the Box-Cox transformation significantly improved the model performance (R2) by approximately 100 % (from 0.35 to 0.79 for ECe and 0.59 to 0.85 for ESP), while the additional predictors further enhanced the performance (R2 increased by 15 %), ranking in the top 30 % of the feature importance list. Results revealed that global multiple-year average salinization and sodification are primarily concentrated in arid regions characterized by low precipitation and high temperatures. We also found a significant increasing trend of soil salinization in 20 % of global land and of sodification in 48 % of global land from 1980 to 2024, both most pronounced near the equator, as well as in central and eastern North America, Europe, southeastern China, and Mongolia. This study provides updated long-term soil salinity and sodicity maps with improved accuracies, offering critical insights for sustainable land management under climate change, serving as an essential resource for addressing food security and land degradation challenges worldwide.
高分辨率全球土壤盐度和碱度制图(1980-2024):基于box - cox的样本优化、多源遥感特征和不确定度量化
全球土壤盐碱化和固体化通过造成土壤中盐的过度积累,降低生产力,威胁粮食安全。然而,由于缺乏从长远角度对土壤盐分和碱度的高精度估计,它们的时空格局尚未得到很好的记录。本文提出了一个结合Box-Cox变换的新框架,该框架解决了样本的倾斜分布和更关键的预测因子以及遥感指数,利用随机森林以1公里× 1公里的分辨率预测1980年至2024年全球土壤盐度(饱和土壤提取物的电导率,ECe)和碱度(交换钠百分比,ESP)。模型精度高于以往研究,ECe和ESP的均方根误差(RMSE)分别为2.24和6.04,R2分别为0.65和0.60。Box-Cox转换的实施显著提高了模型性能(R2)约100% (ECe从0.35提高到0.79,ESP从0.59提高到0.85),而额外的预测因子进一步提高了性能(R2提高了15%),在特征重要性列表中排名前30%。结果表明,全球多年平均盐碱化和固体化主要集中在低降水和高温的干旱地区。我们还发现,从1980年到2024年,全球20%土地的土壤盐碱化和48%土地的土壤固化趋势显著增加,这两个趋势在赤道附近、北美中部和东部、欧洲、中国东南部和蒙古最为明显。该研究提供了更新的长期土壤盐度和碱度地图,提高了准确性,为气候变化下的可持续土地管理提供了重要见解,是应对全球粮食安全和土地退化挑战的重要资源。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
Remote Sensing of Environment
Remote Sensing of Environment 环境科学-成像科学与照相技术
CiteScore
25.10
自引率
8.90%
发文量
455
审稿时长
53 days
期刊介绍: Remote Sensing of Environment (RSE) serves the Earth observation community by disseminating results on the theory, science, applications, and technology that contribute to advancing the field of remote sensing. With a thoroughly interdisciplinary approach, RSE encompasses terrestrial, oceanic, and atmospheric sensing. The journal emphasizes biophysical and quantitative approaches to remote sensing at local to global scales, covering a diverse range of applications and techniques. RSE serves as a vital platform for the exchange of knowledge and advancements in the dynamic field of remote sensing.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:604180095
Book学术官方微信