Predicting of dust storm source by combining remote sensing, statistic-based predictive models and game theory in the Sistan watershed, southwestern Asia

作     者：Mahdi BOROUGHANI Sima POURHASHEMI Hamid GHOLAMI Dimitris G KASKAOUTIS Mahdi BOROUGHANI;Sima POURHASHEMI;Hamid GHOLAMI;Dimitris G KASKAOUTIS

作者机构：Research Center for Geoscience and Social StudiesHakim Sabzevari UniversitySazbevar ***Iran Faculty of Geography and Environmental SciencesHakim Sabzevari UniversitySabzevar ***Iran Department of Natural Resources EngineeringUniversity of HormozganHormozgan ***Iran Institute for Environmental Research and Sustainable DevelopmentNational Observatory of AthensAthens 15784Greece Environmental Chemical Processes LaboratoryDepartment of ChemistryUniversity of CreteCrete 70013Greece 

出 版 物：《Journal of Arid Land》 (干旱区科学（英文版）)

年 卷 期：2021年第13卷第11期

页      面：1103-1121页

核心收录：

学科分类：0830[工学-环境科学与工程（可授工学、理学、农学学位）] 07[理学] 070601[理学-气象学] 0706[理学-大气科学] 

基　　金：The study was financially supported by the Fund for Support of Researchers and Technologists of Iran(97022330) Panhellenic Infrastructure for Atmospheric Composition and Climate Change(PANACEA,MIS 5021516) Competitiveness,Entrepreneurship and Innovation(NSRF 2014-2020) co-financed by Greece and the European Union(European Regional Development Fund) 

主　　题：potential dust source remote sensing frequency ratio weight of evidence dust emission 

摘      要：Dust storms in arid and desert areas affect radiation budget,air quality,visibility,enzymatic activities,agricultural products and human *** to increased drought and land use changes in recent years,the frequency of dust storms occurrence in Iran has been *** study aims to identify dust source areas in the Sistan watershed(Iran-Afghanistan borders)-an important regional source for dust storms in southwestern Asia,using remote sensing(RS)and bivariate statistical ***,this study determines the relative importance of factors controlling dust emissions using frequency ratio(FR)and weights of evidence(WOE)models and interpretability of predictive models using game *** this purpose,we identified 211 dust sources in the study area and generated a dust source distribution map-inventory map-by dust source potential index based on RS *** addition,spatial maps of topographic factors affecting dust source areas including soil,lithology,slope,Normalized difference vegetation index(NDVI),geomorphology and land use were *** performance of two models(WOE and FR)was evaluated using the area under curve(AUC)of the receiver operating characteristic *** results showed that soil,geomorphology and slope exhibited the greatest influence in the dust source *** 55.3%(according to FR)and 62.6%(according to WOE)of the total area were classified as high and very high potential dust sources,while both models displayed acceptable accuracy with subsurface levels of 0.704 for FR and 0.751 for WOE,although they predict different fractions of dust potential *** on Shapley additive explanations(SHAP),three factors,i.e.,soil,slope and NDVI have the highest impact on the model s ***,combination of statistic-based predictive models(or data mining models),RS and game theory techniques can provide accurate maps of dust source areas in arid and semi-arid regions,which can be helpful for mitigation of negative effects of d