mozhdeh Jamei; mohammad mousavi; Amin Alizadeh; parviz irannejad
Abstract
Introduction: Surface soil moisture is one of the most important variables in the hydrological cycle, and plays a key role in scientific and practical applications such as hydrological modelling, weather forecasting, climate change studies and water resources managements. Microwave radiometry at low ...
Read More
Introduction: Surface soil moisture is one of the most important variables in the hydrological cycle, and plays a key role in scientific and practical applications such as hydrological modelling, weather forecasting, climate change studies and water resources managements. Microwave radiometry at low frequencies (1.4GHz) is an established technique for estimating global surface soil moisture with a suitable accuracy. In recent years, soil moisture measurements have become increasingly available from satellite-based microwave sensors. The ESA’s Soil Moisture and Ocean Salinity (SMOS) satellite was launched in November 2009. It carries the first L-band 2-D synthetic aperture microwave radiometer to provide global estimates of soil moisture with an averaged ground resolution of 43 km over the field of view. The main objective of this research was to validateSMOS soil moisture retrievals over the west and south west of Iran.
Materials and Methods:The study area is located in the west and southwest of Iran which contains five areas belongingto the Ministry of Power. For the validation of SMOS dataover the study area, the SMOS soil moisture retrievals from MIR_SMUDP2 productswere compared with ground-based insitu measurements. The validation process was carried out using Collocation techniquefor the period 2012-2013. Collocation technique is a method used in the field of remote sensing to verify compliance measurements from two or more different instruments. In this study, the collocation codes were developed in Matlab Linux programming language. The ground-based in situ measurements included direct soil moisture measurements at the 5cm depth which were collected from five meteorological stations in the study area. We prepared a file for each station which contained daily soil moisture, date and time, geographical coordinates of metrological stations as input for validation model. The SMOS Level 2 Soil Moisture User Data Product (MIR_SMUDP2 files) version 551, which were provided through the ESA, contains the retrieved soil moisture and simulated TB, dielectric constants, etc. In this work, the ESA’s SMOS Matlab tool on RedHat Linux was used to read and derivesoil moisture data from MIR_SMUDP2 files.Four statistical metrics and Taylor diagram were used for the evaluation error of validation; the Root Mean Squared Difference (RMSD), the centered Root Mean Square Difference (cRMSD), the Mean Bias Error or bias and the correlation coefficient (R).
The Taylor diagrams wereused to represent three different statistical metrics (R, centered Root Mean Square Difference (cRMSD) and standard deviation) on two dimensional plots to graphically describe how closely SMOS dataset matched ground-based observations .
Results and Discussion: Based on the research algorithm and using MATLAB, the Validation model for SMOS soil moisture data was obtained. This model was appliedfor five metrological stations and the collocated soil moisture data from SMOS data and in situ data was saved as output of model to error evaluation. The results of validation errorshoweda good correlation between the SMOS soil moisture andin situ measurements. The highestand lowest correlation coefficientswere shown at Ahvaz (R=0.88) and Sarableh(R=0.75)stations, respectively.According to the bias values, the SMOS soil moisture retrievals had underestimation atAhvaz(MBE=0.04 m3m−3),Sararod(MBE=0.011 m3m−3), Sarableh(MBE=0.048 m3m−3) stations, whereas a slight overestimation of the SMOS product was detectedatthe Darab (MBE=-0.01 m3m−3) andEkbatan (MBE=-0.031 m3m−3) stations. In addition, the Root Mean Squared Difference (RMSD) values between the SMOS data and in situ data varied from 0.02 to 0.062 m3m−3 and at Ahvaz station withRMSD=0.048 m3m−3is close to the targeted SMOS accuracy of 0.04 m3m−3.Based on the Taylor diagrams, SMOS data had the highest correlation (R=0.88) with in situ measurements at Ahwaz stationand the lowest difference (cRMSD=0.008 m3m−3) between two data setswas found at Darab station.
Conclusions:The objective of this paper was to validateESA’s SMOS (Soil Moisture and Ocean Salinity) satellite products in the west and southwest of Iran for the period of 2012-2013. The validation of SMOS soil moisture retrievals from MIR_SMUDP2 products was done by using soil moisture measurements from five meteorological stations. The SMOS soil moisture retrievals showed underestimations at Ahvaz, Sararod andSarableh stations, whereas a slight overestimation werefound at Darab, Ekbatan stations. The validation results and Taylor diagrams showed thatthe SMOS soil moisture retrievals with R=0.88, RMSD=0.048 m3m−3, cRMSD=0.021 m3m−3at Ahvaz stationwasvery close to the targeted SMOS accuracy objectiveof 0.04 m3m−3 and then at Darab station SMOS data with R=0.82, RMSD=0.028 m3m−3,cRMSD=0.008 m3m−3indicateda good agreement with ground soil moisture measurements. Overall, the SMOS soil moisture data hadan acceptableaccuracy and agreement with in situ data at all stations. Therefore, we can use these data sets as a tool to derive soil moisture maps at study areas.
M. Jamei; M. Mousavi Baygi; M. Bannayan Awal
Abstract
Available accurate and reliable precipitation data are so important in water resources management and planning. In this study,to determine the best method of regional precipitation estimate in Khuzestan province, estimated daily precipitation data from the best interpolation method and APHRODIT Daily ...
Read More
Available accurate and reliable precipitation data are so important in water resources management and planning. In this study,to determine the best method of regional precipitation estimate in Khuzestan province, estimated daily precipitation data from the best interpolation method and APHRODIT Daily Grid Precipitation data during the 2000-2007 years were compared with 44 meteorological stations. Four interpolation methods i.e. Inverse Distance Weighted, Ordinary Kriging, Cokriging, and Regression Kriging were assessed to determine the most appropriate interpolation method for daily precipitation.For the variography analysis in Kriging models, five variogram models including spherical, exponential, linear, linear to sill and Gaussian fitted on the precipitation data. Near neighbor method was used to compare APHRODIT Daily Precipitation data with station recorded data. Cross validation technique was employed to evaluate the interpolation methods and the most appropriate method was determined based on Root Mean Square Error,Mean Bias Error, Mean Absolute Error indices and regression analysis. The result of error evaluation of interpolation methods showed that regression Kriging method has the highest accurate to interpolation of daily precipitation data in Khuzestan province. Therefore, regression-based interpolation methods which using covariates would be improved precipitation evaluate accurate in the area. Comparison of error indices and regression analysis of regression Kriging interpolation method and estimate of APHRODITE show that on most days the accurately estimate of regression Kriging is higher than the APHRODITE. Therefore to understanding of spatial distribution and estimate of daily precipitation data in Khuzestan Province, Regression Kriging interpolation method is more accurate than available APHRODITE data