Ardashir Yousefzadeh; Battol Zeynali; Khalil Valizadeh Kamran; Saayad Asghari Sar Eskanrood
Volume 6, Issue 18 , June 2019, , Pages 181-205
Abstract
Introduction According to Cornelsen (2015), soil moisture is one of the most important variables in the hydrological cycle. In Manson's studies (2010), soil moisture was identified as one of the major climatic variables by the World Meteorological Organization, the Global Climate Observing System, and ...
Read More
Introduction According to Cornelsen (2015), soil moisture is one of the most important variables in the hydrological cycle. In Manson's studies (2010), soil moisture was identified as one of the major climatic variables by the World Meteorological Organization, the Global Climate Observing System, and the Observational Satellite Observatory. Remote sensing provides a powerful tool for detecting and monitoring soil moisture near the Earth's surface (0 to 5 cm). Also, according to Babaeian research (2015), optical reflection of the soil and thermal emission to Eliit (1979) and Microwave backed by Das researches (2008) is related on soil moisture. Remote sensing techniques based on microwave waves are effective techniques for estimating soil moisture. Surface water levels can be extracted using the NDVI index in Landsat images (Maryam Khosravian et al., 2012, p. 115), and user variations in time series can also be identified. (Malian et al., 1395, p. 49). Due to the limitation of access to radar information, the focus of the study is on the near-visible infrared range and the amount of heat from the surface of the earth is measured from 3.5 to 14 micrometers (Curran, 1985). Soil moisture content with this method requires the estimation of soil surface temperature and vegetation index (Wang & Co, 2009). Vegetation and surface temperature have a complex dependence on soil moisture (Carlson, 1994). According to Gillies et al. (1997), the combination of these two indicators can be used to estimate soil moisture with an acceptable accuracy.In 2017, a model for estimating soil moisture using a visual distance assay was proposed based on the linear physical relationship between soil moisture and the short-range infrared reflection (STR), which is based on the distribution of pixels inside the surface temperature space and the normalized vegetation index (STR-VI) (Sadegi et al., 2017). A trapezoid or triangle model is one of the models used in remote sensing to estimate soil moisture. The study area is the Simineh River basin which is one of the sub basins of Lake Urmia Basin, with an of 3279 km2. Methodology The main data in this study are Landsat 8 satellite imagery. After applying atmospheric and radiometric corrections, the processing of images, between 2016-2017, was done according to the process of view of Figure 1. Figure (1) Research process (Source: Writers) -Thermal-Optical Trapezoid Model (TOTRAM) This model is based on the distribution of pixels in the surface temperature and vegetation cover space that is fitted to estimate soil moisture using a linear equation in space (LST-VI) (Sadegi et al., 2017). Equation (1) -Optical Trapezoid Model (OPTRAM) The base of this model is the insertion of surface temperature to estimate the soil moisture in the visible wavelength range. In this physical model, the linear relationship between soil moisture and infrared reflection is expressed. Equation (2) Result According to the results of this study, the lowest average temperatures of satellite images were respectively -3.23 and 2.12 C in 2015 and 2016, indicating an increase in temperature. In 2017, the highest amount of vegetation density was 0.66. The correlation between the OPTRAM model in 2015 and the STR and NDVI variables, were positive and the correlation indices were respectively 0.709 and 1. These figures for STR and NDVI in 2016 were respectively -0.648 and 1, which indicated a negative correlation between STR and soil moisture; soil moisture decreased with increasing STR and increased with increasing NDVI. And the positive correlation between OPTRAM model and NDVI confirmed it. In 2017, the positive correlation between STR and NDVI with soil moisture were respectively 0.672 and 1. The TOTRAM model in 2015 had a negative correlation with the LST and NDVI indices and they were respectively -0.574 and -1. It indicated low accuracy of this model compared to the OPTRAM model in estimating soil moisture. In 2016, the correlation between LST and NDVI with soil moisture were respectively -0.974 and 0.409. They respectively reached -0.940 -0.787 in 2017. Discussion and Conclusion In this research, due to the limitations of the field information, soil moisture was extracted without the use of ground control points. The comparison of the accuracy of the two models in the region was investigated. The results indicated that soil moisture can be extracted from the STR index with high accuracy, compared to LST index, based on NDVI Triangular space. Due to the low cost and the availability of visible images, radar images were accurately obtained and the correlation between OPTRAM model and soil moisture estimation was confirmed. According to the extraction results, the OPTRAM model can estimate the soil moisture better than the TOTRAM model, due to the fact that it is not influenced by environmental factors and global parameters. According to research results, TOTRAM has two main constraints. First, it cannot be used for a satellite without thermal bonding. Secondly, in addition to soil moisture, the LST depends on environmental factors to be calibrated for each image. To overcome the limitations of the TOTRAM model as well as the empirical visibility of indicators, a new physical trapezoidal model, called OPTRAM, is proposed. It is based on the physical relationship developed between soil moisture and the "reflected infrared reflection" (Sadegi et al., 2015).
