Spatial analysis of environmental factors influencing dust sources in the east of Iran using a new active-learning approach

Abstract

The frequency and intensity of dust storms in Iran has increased significantly in recent years. This study identifies the sources of dust using hybrid algorithms – probability density-index of entropy (PD-IOE), probability density-radial basic function neural network (PD-RBFNN), probability density-self-organizing map (PD-SOM), and probability density-fuzzy ARTMAP (PD-FAM). Hybrid models employed several effective environmental factors: land cover, slope, soil, land use, wind speed, geology, temperature, and precipitation. A random selection of 70% of the data points were used for training the spatial models and the remainder (30%) were used to test the effectiveness of the models to determine the best algorithm. The results reveal that the PD-FAM algorithm produced the most accurate predictions of dust sources. Geology, slope, and soil were the factors that were most effective predictors of dust generation in eastern Iran. Comprehensive management is needed to manage dust production in Iran and these findings may ease identification of locations most likely to produce dust.

Keywords:

• Dust
• effective variables
• GIS
• hybrid algorithms
• Eastern Iran

Disclosure statement

No potential conflict of interest was reported by the authors.