Satellite-based prediction of surface dust mass concentration in southeastern Iran using an intelligent approach
Por favor, use este identificador para citar o enlazar este ítem:
http://hdl.handle.net/10045/134920
Título: | Satellite-based prediction of surface dust mass concentration in southeastern Iran using an intelligent approach |
---|---|
Autor/es: | Asadollah, S. Babak H.S. | Sharafati, Ahmad | Motta, Davide | Jódar-Abellán, Antonio | Pardo Picazo, Miguel Ángel |
Grupo/s de investigación o GITE: | Ingeniería Hidráulica y Ambiental (IngHA) |
Centro, Departamento o Servicio: | Universidad de Alicante. Departamento de Ingeniería Civil |
Palabras clave: | Surface dust mass concentration | Gradient boosting regression | Prediction | Khuzestan |
Fecha de publicación: | 5-jun-2023 |
Editor: | Springer Nature |
Cita bibliográfica: | Stochastic Environmental Research and Risk Assessment. 2023, 37: 3731-3745. https://doi.org/10.1007/s00477-023-02473-6 |
Resumen: | The southeastern section of Iran, especially the province of Khuzestan, experience severe air pollution levels, such as high values of surface dust mass concentration (SDMC). The province lacks accurate and well-placed ground observational stations, therefore the only viable approach for evaluating SDMC is via remote sensing. In this study, meteorological, hydrological and geological data on 11 input variables from Modern-Era Retrospective analysis for Research and Applications Version 2 (MERRA-2), global precipitation measurement (GPM) and Global Land Data Assimilation System (GLDAS) for the year 2018 are used for prediction of the SDMC values, also obtained from another MERRA-2 mission. For real-time prediction, Pearson’s correlation coefficient (PCC) analysis shows that wind-related variables—surface wind speed, surface aerodynamic conductance and surface pressure—are those with the highest correlation with SDMC. Using the gradient boosting regression (GBR) algorithm, these three variables can simulate SDMC with good accuracy (R2=0.76,NSE=0.76,N-RMSE=0.48andN-MAE=0.34). Furthermore, near-future SDMC forecasting down to 8 days prior of SMDC occurrence is also carried out. A sequential forward feature selection of the input variables, based on PCC, is used for four lead times and results show that surface pressure and heat flux govern near-future predictions. With R2=0.46 and N-RMSE=0.74, GBR shows good potential for forecasting SDMC 8 days in advance. Real-time and near-future simulation results generally show that robust SDMC prediction can be obtained using exclusively remote sensing data, without ground-based observations. |
URI: | http://hdl.handle.net/10045/134920 |
ISSN: | 1436-3240 (Print) | 1436-3259 (Online) |
DOI: | 10.1007/s00477-023-02473-6 |
Idioma: | eng |
Tipo: | info:eu-repo/semantics/article |
Derechos: | © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2023 |
Revisión científica: | si |
Versión del editor: | https://doi.org/10.1007/s00477-023-02473-6 |
Aparece en las colecciones: | INV - IngHA - Artículos de Revistas |
Archivos en este ítem:
Archivo | Descripción | Tamaño | Formato | |
---|---|---|---|---|
Asadollach_etal_2023_StochEnvironResRiskAssess_preprint.pdf | Preprint (acceso abierto) | 2,31 MB | Adobe PDF | Abrir Vista previa |
Asadollach_etal_2023_StochEnvironResRiskAssess_final.pdf | Versión final (acceso restringido) | 2,14 MB | Adobe PDF | Abrir Solicitar una copia |
Todos los documentos en RUA están protegidos por derechos de autor. Algunos derechos reservados.