Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Assessing dust storm risks in water-scarce regions : a machine learning approach

Naghibi, Amir LU ; Hashemi, Hossein LU orcid ; Mousavi, Seyed Mohsen ; Zhao, Pengxiang LU and Mansourian, Ali LU orcid (2026) p.13-24
Abstract

Climatic factors and water and land use management mainly impact dust storm frequency and intensity. Water scarcity and a drier climate will lead to more severe and frequent dust storms in the near and far future. This makes dust storm susceptibility a dynamic phenomenon requiring spatiotemporal analyses. Regarding the complexity of this phenomenon, this chapter discusses AI-based methods for spatiotemporal modeling of dust storm susceptibility. Those include mixture discriminant analysis (MDA) and heterogeneous discriminant analysis (HDA) algorithms. The findings depicted that the MDA and HDA algorithms modeled dust storm susceptibility with area under the receiver operating characteristic curves ranging between 0.79 and 0.87 for four... (More)

Climatic factors and water and land use management mainly impact dust storm frequency and intensity. Water scarcity and a drier climate will lead to more severe and frequent dust storms in the near and far future. This makes dust storm susceptibility a dynamic phenomenon requiring spatiotemporal analyses. Regarding the complexity of this phenomenon, this chapter discusses AI-based methods for spatiotemporal modeling of dust storm susceptibility. Those include mixture discriminant analysis (MDA) and heterogeneous discriminant analysis (HDA) algorithms. The findings depicted that the MDA and HDA algorithms modeled dust storm susceptibility with area under the receiver operating characteristic curves ranging between 0.79 and 0.87 for four dry and wet hydrological conditions periods, demonstrating robust applicability across regions facing similar environmental challenges. The findings also revealed that elevation, precipitation, wind speed, and Normalized Difference Vegetation Index (NDVI) were most important in dust storm source susceptibility. NDVI reflects human impacts through land use, land cover, or cropping-type changes. Vegetation is also affected by sociopolitical aspects in the region, such as war, land use, and water management in a transboundary context. Therefore, it is essential to consider and investigate human and natural impacts. As a key transboundary water resource, the Tigris and Euphrates River Basin is critical in mitigating dust storm risks, necessitating strong and coordinated regional governance.

(Less)
Please use this url to cite or link to this publication:
author
; ; ; and
organization
publishing date
type
Chapter in Book/Report/Conference proceeding
publication status
published
subject
keywords
Dust storm, heterogeneous discriminant analysis, machine learning, mixture discriminant analysis, mixture learning, Tigris and Euphrates
host publication
Water Scarcity Management : Towards the Application of Artificial Intelligence and Earth Observation Data - Towards the Application of Artificial Intelligence and Earth Observation Data
pages
12 pages
publisher
Elsevier
external identifiers
  • scopus:105026854217
ISBN
9780443267239
9780443267222
DOI
10.1016/B978-0-443-26722-2.00012-X
language
English
LU publication?
yes
id
9209ee55-7214-4610-8d82-b6b6dfdf03dd
date added to LUP
2026-02-16 09:49:16
date last changed
2026-02-17 02:59:25
@inbook{9209ee55-7214-4610-8d82-b6b6dfdf03dd,
  abstract     = {{<p>Climatic factors and water and land use management mainly impact dust storm frequency and intensity. Water scarcity and a drier climate will lead to more severe and frequent dust storms in the near and far future. This makes dust storm susceptibility a dynamic phenomenon requiring spatiotemporal analyses. Regarding the complexity of this phenomenon, this chapter discusses AI-based methods for spatiotemporal modeling of dust storm susceptibility. Those include mixture discriminant analysis (MDA) and heterogeneous discriminant analysis (HDA) algorithms. The findings depicted that the MDA and HDA algorithms modeled dust storm susceptibility with area under the receiver operating characteristic curves ranging between 0.79 and 0.87 for four dry and wet hydrological conditions periods, demonstrating robust applicability across regions facing similar environmental challenges. The findings also revealed that elevation, precipitation, wind speed, and Normalized Difference Vegetation Index (NDVI) were most important in dust storm source susceptibility. NDVI reflects human impacts through land use, land cover, or cropping-type changes. Vegetation is also affected by sociopolitical aspects in the region, such as war, land use, and water management in a transboundary context. Therefore, it is essential to consider and investigate human and natural impacts. As a key transboundary water resource, the Tigris and Euphrates River Basin is critical in mitigating dust storm risks, necessitating strong and coordinated regional governance.</p>}},
  author       = {{Naghibi, Amir and Hashemi, Hossein and Mousavi, Seyed Mohsen and Zhao, Pengxiang and Mansourian, Ali}},
  booktitle    = {{Water Scarcity Management : Towards the Application of Artificial Intelligence and Earth Observation Data}},
  isbn         = {{9780443267239}},
  keywords     = {{Dust storm; heterogeneous discriminant analysis; machine learning; mixture discriminant analysis; mixture learning; Tigris and Euphrates}},
  language     = {{eng}},
  pages        = {{13--24}},
  publisher    = {{Elsevier}},
  title        = {{Assessing dust storm risks in water-scarce regions : a machine learning approach}},
  url          = {{http://dx.doi.org/10.1016/B978-0-443-26722-2.00012-X}},
  doi          = {{10.1016/B978-0-443-26722-2.00012-X}},
  year         = {{2026}},
}