SATELLITE BASED DROUGHT ASSESSMENT OVER LATUR , INDIA USING SOIL MOISTURE DERIVED FROM SMOS

Climatological variables such as rainfall, temperature have been extensively used by researchers for drought monitoring at a larger spatial region. These variables have a direct influence on the soil moisture which in turn extends the application of soil moisture in drought assessment. With the advancement of technology, various satellites provide soil moisture data at different spatio-temporal resolutions. In this article, soil moisture obtained from Soil Moisture Ocean Salinity (SMOS) is used to analyze the drought condition over Latur district in Maharashtra, India. The monthly soil moisture derived by averaging the daily data for the years 2010 to 2015 is compared with two drought indices, i.e. Standardized Precipitation Index (SPI) calculated for years 2010 to 2015 and Standardized Precipitation-Evapotranspiration Index (SPEI) calculated for years 2010 to 2013. Even though the overall correlation among the indices with the soil moisture is not significant, the seasonal (summer) correlation is significant. From the results, it is identified that SMOS derived soil moisture can be used as a potential parameter in drought assessment. * Corresponding author


INTRODUCTION
Drought is a major disaster affecting the society with increasing severity, duration and spatial extent (Mallya et al., 2016).Broadly, drought is classified into three classes namely meteorological, agricultural and hydrological droughts.If the mean annual rainfall in a region is less than 75% of its normal rainfall, it is known as meteorological drought (Kendale, 2011).Knowledge of type, severity, and spatial location of drought helps the government agencies to take effective measures and policy formulation.
Standardized Precipitation Index (SPI) (McKee et al., 1993) and Standardized Precipitation-Evapotranspiration Index (SPEI) are the well-established indices for drought monitoring over larger spatial regions.SPEI is an extension of SPI that also captures the effect of potential evapotranspiration (PET) on drought (Vicente-Serrano et al., 2010).These indices can be calculated for various time scales.The indices computed with one or two months time scale is used for meteorological drought assessment.Based on the range of these indices value the drought intensity of a region is assessed as shown in  et al., 1993).
The variables used for computing these indices have a direct influence on the soil moisture which in turn extends the application of soil moisture in drought assessment.Many studies have shown the application of soil moisture derived from in-situ measurements, and remote sensing techniques in drought assessment.Due to the limitations (spatial coverage, temporal resolution and high resource consumption) in the insitu measurements, soil moisture derived from remote sensing techniques are often used by the researchers.Soil moisture can be obtained from both the optical and microwave remote sensing techniques.However, the optical data has limitations such as cloud cover, less spatial coverage and low temporal resolution.Therefore, passive microwave remote sensing is more preferred due to its all weather capacity, day-night coverage, high temporal resolution and sensitivity to dielectric constant of soil.
A study done by (Thiruvengadam and Rao, 2016) shows a positive spatio-temporal correlation of AMSR-E soil moisture with precipitation.Many studies were carried out to find an association of soil moisture with drought condition using various drought indices.Palmer Drought Severity Index (PDSI), a widely used drought index, shows a significant correlation with surface soil moisture (Dai et al., 2004).A similar study done by (Scaini et al., 2015) found that relations between soil moisture and drought in-dices such as SPI and SPEI are promising; but higher correlation is seen with in-situ measurements than SMOS soil moisture.Soil moisture is an effective tool for monitoring agricultural drought as well.CMI (Crop Moisture Index) and the AWD (Atmospheric Water Deficit index) are interrelated with soil moisture regarding time series variation, as well as in correlation (Mart´ınez-Fernandez´ et al., 2015).Soil Water Deficit Index (SWDI) computed based on satellite soil moisture data can pose an effective tool for agricultural drought monitoring.(Mart´ınez-Fernandez´ et al., 2017).Soil Moisture Agricultural Drought Index (SMADI), has been used by (Sanchez´ et al., 2017) for drought monitor-ing which shows a good correlation with different agricultural drought indices.Effect of drought on crop yield is studied with the variation of soil moisture (Chakrabarti et al., 2014).
The main aim of the study is to identify the potential of SMOS derived soil moisture in drought assessment for Latur district in Maharashtra state, India.The next section describes the selected study area and the data sets used.Section III describes the methodology and the results are shown in section IV and conclusion in section V.

STUDY AREA AND DATASET
Latur district in Maharashtra state, India spans between 17 52'N to 18 50'N and 76 18'E to 79 12'E with total area of 7157 sq km as shown in Figure 1.The district lies on Deccan Plateau

METHODOLOGY
The methodology adopted for the study is shown in the Figure 2. The daily temperature provided by IMD is available in The SMOS soil moisture that is obtained in netCDF format is converted to geotiff format in ArcGIS 10.1 for ease of analysis.Similar to the indices, the soil moisture is also clipped to the study area.One raster data is converted into point data and all the temporal raster data is extracted to this point file by using "Extract Multi Values to Points" in ArcGIS.The daily ascending and descending pass soil moisture is averaged spatially and temporally to monthly data for the years 2010 to 2015.

RESULTS AND DISCUSSIONS
The   In order to do the quantitative assessment of the SMOS soil moisture for drought monitoring, a correlation analysis with the indices is carried out.The indices derived with one month time scale are used in correlation analysis as three-month time scale is not significant.The range of values for the variables such as

CONCLUSION
The results of the study revealed that SMOS derived soil moisture can be used as a potential parameter in drought assessment.The study also found that the better performance of SPEI with soil moisture is attributed to temperature variable.A more detailed study, considering the utility of other variables like wind velocity and surface humidity is required to identify the robustness of soil moisture in drought assessment.The study can be extended to agricultural and hydrological drought, as they involve several climatological variables which influence soil moisture.

Figure 1 .
Figure 1.Study area showing Latur district in Maharashtra state of India.

Figure 2 .
Figure 2. Methodology for drought assessment using SMOS and IMD rainfall and temperature data time series plot of all the variables (BAL, PET, TMIN, TMAX, PRCP) is shown in the Figure 3.As the influence of rainfall on soil moisture does not hold for a long duration, indices derived with one and three-month time scale are considered for analysis.The monthly temporal variation of SPI for the years 2010 to 2015 is shown in the Figure 4.The time series plots of SPI and SPEI are shown in Figures 5 and 6 respectively.It can be seen from the Figures 5 and 6 that Latur has experienced severe drought during 1972 and 2009.In the duration of SMOS soil moisture availability, Latur experienced drought conditions during 2011 and 2015.

Figure 3 .
Figure 3. Distribution of Precipitation, TMAX, TMIN, PET, and Climate water balance data over Latur district.

Figure 4 .
Figure 4. Monthly temporal variation of SPI for the years 2010 to 2015 over Latur district.

Figure 5 .
Figure 5. SPI calculated for one and three month time scales for Latur district.

Figure 6 .
Figure 6.SPEI calculated for one and three month time scales for Latur district.

Figure 7 .
Figure 7. Monthly temporal variation of soil moisture against SPI.

Figure 9 .
Figure 9. Monthly temporal variation of soil moisture against SPI during summer.

Figure 10 .
Figure 10.Monthly temporal variation of soil moisture against SPEI during Summer.

Figure 11 .
Figure 11.Monthly temporal variation of soil moisture against SPI during Monsoon.

Figure 12 .
Figure 12.Monthly temporal variation of soil moisture against SPEI during Monsoon.

Table 1
. SPI values and corresponding condition (McKee