Enhancement of Middle East Dust Plumes Based on Spectral Data of MODIS Sensor

Introduction
Dust storms are atmospheric phenomena in arid and semi-arid areas such as Middle East. Iran in west of Asia regularly experiences dust storms (Modarres and Silva, 2007) due to its location in arid and semi-arid belts of the world (Jalali et al., 2008). For example, intense dust storm occurred over Middle East in 18 Mars 2012. This dust storm affected many countries such as Iran, Iraq, Kuwait, Bahrain, Qatar, and Saudi Arabia.
Aerosols perturb the Earth’s energy budget by scattering and absorbing radiation and by altering cloud properties and lifetimes. They also exert large influences on weather, air quality, hydrological cycles, and ecosystems (IPCC, 2007). Industrialization and human activities in the past several decades have caused changes to the air quality and Earth’s climate by releasing excessive amounts of trace gases and aerosol particles. It is important to regularly monitor the global aerosol distributions and study how they are changing, especially for those aerosols with large spatial and temporal variability, such as dust storms. Detection of these highly variable aerosol events is challenging because of: episodic features, short lifetimes, multiple-scales, and strong impact on local surface and meteorological conditions (Zhao et al., 2010)
Based on the identification techniques, more recent studies have been carried out to quantitatively determine the physical parameters of dust storms, such as aerosol loading. The aerosol loading is a key parameter for dust storm assessment, modeling, and forecasting. The VIR technique can be used to retrieve this parameter over ocean (Tanré et al., 1997); but to do so over land there is a major challenge. Because dust storms mostly occur over desert or arid regions with bright surfaces, such as the Sahara Desert and the Gobi Desert, the surface contribution to the satellite signal is quite large and often unknown. As a result, estimates of the properties of dust aerosol are highly uncertain (Zhang et al., 2006).
Although many studies were conducted on dust detection, but there are still problems for the enhancement of dust. So far, considerable researches and appropriate methods have not been done for detection of dust using the satellite images over Middle East. Hence, in this paper, two methods used for detection of dust storm over Middle East and the results of each method are compared.
Material and Methods
The Moderate Resolution Imaging Spectroradiometer (MODIS) aboard NASA’s Terra and Aqua satellites is making near-global daily observations of the earth in a wide spectral range (0.41–15 µm) (Remer et al, 2005). MODIS has been acquiring daily global data in 36 spectral bands from visible to thermal infrared (29 spectral bands with 1km resolution, five spectral bands with 500m resolution, and two with 250-m resolution, nadir pixel dimensions) (Levy et al., 2007). Daily MODIS Level 1B (L1B) 1km data (MOD021KM=Terra) were used in this paper. Data were obtained from the Level 1 and Atmosphere Archive and Distribution System (LAADS; http://ladsweb.nascom.nasa.gov/)
Dust Detection Algorithm 1: (Zhao et al., 2010)
The following methods used for dust detection over Middle East:
Let’s first define some variables and their symbols that will be used throughout the paper
BT—brightness temperature (wavelength is given in subscript, e.g., BT11µm)
R—reflectance (wavelength is given in subscript, e.g., R0.64µm)
BTD—brightness temperature difference
MeanR—mean of reflectance for 3 x 3 pixels (wavelength is given in subscript, e.g., MeanR0.86µm)
StdR—standard deviation of reflectance for 3 x 3 pixels (wavelength is given in subscript, e.g.,
StdR0.86µm)
Rat1 = (R_0.64µm − R_0.47µm)/ (R_0.64µm + R_0.47µm)
Rat2 = (Rat1 × Rat1)/ (R_0.47µm × R_0.47µm)
R1 = R_0.47µm/_R0.64µm
R2 = R_0.86µm/R_0.64µm
NDVI = (R_0.86µm − R_0.64µm)/ (R_0.86µm + R_0.64µm)
MNDVI = NDVI²/ (R_0.64µm × R_0.64µm)
The specific visible reflectance and IR brightness temperature tests currently implemented are:
1.  Good data test for BT and R:
•  R_0.47µm, R_0.64µm, R_0.86µm, R_1.38µm > 0 
•  BT_3.9µm, BT_11µm, BT_12µm > 0K
2.  BTD and R tests:
•  BT_11µm − BT_12µm ≤ −0.5K & BT_3.9µm − BT_11µm ≥ 20K & R_1.38µm < 0.055
(Screen for pixels that are water cloud free.  If these conditions are not met, then the pixels are cloudy and terminate testing)
3.  Dust test:
•  If BT_3.9µm − BT_11µm ≥ 25K then dust
•  If MNDVI < 0.08 & Rat2 > 0.005 then dust
4.  Thick dust test:
•  BT_11µm − BT_12µm ≤ −0.5K & BT_3.9µm − BT_11µm ≥ 25K & R_1.38µm < 0.035
• MNDVI < 0.2
Dust Detection Algorithm 2
The flowchart of dust detection algorithm is shown in the following figure.
 
 
 
Discussion of Results
The natural color image (RGB₁₄₃) of MODIS sensor and atmospheric level 2 products, including Aerosol Optical Depth (AOD) and Fine Aerosol Mode Fraction (FMF) were used to compare and evaluate different methods of dust detection. The results showed that, although the algorithm 2 is able to detect dust, it couldn’t detect the exact boundaries of dust plumes very well. Instead, dust detection algorithm 1 extracted the dust plumes in a good level. In addition, by using the dust detection algorithm 1, it is possible to detect thick dust, as well.
 
Conclusion
The results showed that, although the algorithm 2 is able to detect dust, it couldn’t detect the exact boundaries of dust plumes very well. Instead, dust detection algorithm 1 extracted the dust plumes in a good level. Therefore, it is suggested that dust detection algorithm1 will be used for detecting dust plumes over Middle East countries.