Integrated quantitative and qualitative GIS based drought analysis and salt/dust storm impacts on food security

Yariyan, Peyman; Feizizadeh, Bakhtiar; Blaschke, Thomas; Karimzadeh, Sadra; Yakar, Murat

doi:10.22059/eoge.2026.409715.1196

Integrated quantitative and qualitative GIS based drought analysis and salt/dust storm impacts on food security

Document Type : Original Article

Authors

Peyman Yariyan ¹

Bakhtiar Feizizadeh ²

Thomas Blaschke ³

Sadra Karimzadeh ²

Murat Yakar ⁴

¹ Department of Remote sensing and GIS, University of Tabriz, Tabriz, Iran

² Department of Remote Sensing and GIS, University of Tabriz, Tabriz, Iran

³ Department of Applied Geoinformatics Z_GIS, University of Salzburg, Austria

⁴ Department of Geoinformatics, University of Mersin, Mersin, Turkey

10.22059/eoge.2026.409715.1196

Abstract

Dust storms are a growing environmental hazard in the drought-affected Urmia Lake Basin, Iran, threatening agricultural productivity and public health. This study employs four advanced machine learning algorithms—Support Vector Machine (SVM), Gradient Boosting Decision Trees (GBDT), Random Forest (RF), and Logistic Regression (LR)—to model dust storm susceptibility using a comprehensive set of environmental, climatic, and soil variables. The models were trained and validated on remote sensing and ground data, achieving strong performance metrics, with SVM attaining the highest Area under the Curve (AUC) of 0.94, indicating excellent discriminative ability. Susceptibility maps produced by these models reveal significant spatial variability in vulnerability, highlighting extensive agricultural lands and populated areas at risk. The increasing frequency of dust storms due to Lake Urmia’s desiccation directly threatens food security by reducing crop yields and exacerbates respiratory and cardiovascular health issues among local populations. These findings emphasize the urgent need for targeted environmental management and public health strategies to mitigate dust storm impacts and enhance resilience in this ecologically sensitive region.

Keywords

Dust Storm

Machine Learning

Food Security

GIS

Susceptibility Mapping

Subjects