Evaluation of the Effect of Soil Moisture, Wind Speed and Land Cover Properties on Dust Emission Using Remote Sensing and Land Surface Data

  • Sagar Prasad Parajuli

Student thesis: Master's Thesis


Dust storms are common natural phenomena in arid and semi-arid regions. They can affect human health, economy, environment and earth radiation budget at both regional and global scale. Dust emission has a large temporal and spatial variation and its modeling is extremely challenging since it is associated with several factors such as topography, wind speed, soil moisture, surface roughness land cover features and soil properties. The UAE is a unique place for aerosol research since the atmosphere is laden sometimes with mineral dust, sometimes with pollution and mostly with a mixture of the two. Aerosol optical thickness (AOT) is a widely used quantitative indicator of dust since it gives indication of the attenuation of radiation due to the presence of the dust layer. However, its use is limited either because the availability of AOT data measured at ground is very limited or satellite-derived AOT suffers from the coarser temporal and spatial resolution. SEVIRI data is widely used for qualitative detection of dust because of the availability of thermal channels and their high temporal resolution. However, no SEVIRI-derived AOT products are available yet. In this work, effect of wind speed and soil moisture on dust emission is examined using historical datasets of ground measured wind speed, satellite measured soil moisture (AMSR-E), SEVIRI brightness temperatures (BTs), MODIS AOT and AERONET AOT. Furthermore, Artificial Neural Network (ANN) based method is developed to estimate AOT over the UAE using SEVIRI, wind speed and soil moisture data. The results show that the developed combinations of SEVIRI channels can be effectively used to retrieve AOT within RMSE of 0.1. Also, SEVIRI channels and MODIS AOT show good agreement. However, contrary to the findings of previous studies, no quasi-linear relationship was observed between AERONET AOT and SEVIRI BT at 10.8 μm. Similarly, dust emission, as measured by SEVIRI dust indicators, MODIS AOT and AERONET AOT, didn't show any significant correlation with wind speed which is against the widely accepted hypotheses that dust emission can be expressed in terms of power of wind speed.
Date of Award2012
Original languageAmerican English
SupervisorMohamed Hosni Ghedira (Supervisor)


  • Dust --Environmental Aspects
  • Remote Sensing
  • Soil Moisture.

Cite this