Mohseni, F; Mokhtarzade, M (2020). A new soil moisture index driven from an adapted long-term temperature-vegetation scatter plot using MODIS data. JOURNAL OF HYDROLOGY, 581, 124420.

The estimation of soil moisture content on the global scale is a key research issue in the field of remote sensing and, to date, a range of methods have been developed to achieve this. The temperature-vegetation (T-V) technique, which is perhaps one of the most common and successful, suffers from being inaccurate in some cases, especially when the daily T-V scatter plots are not able to define the wet and dry edges in a reliable way. A further challenge is the scale inconsistency between the results obtained on different days, which is a serious concern when monitoring soil moisture changes over periods of time. To address these restrictions, this paper introduces a new soil moisture index, called SMiSEE, driven by a new long-term T-V scatter plot that is established from data covering a 1-year period. The normalized NDVI is set as the horizontal axis and three different temperature factors are suggested as the vertical axis in the proposed 1-year scatter plot. From these three temperature factors, the results proved that "LSTday-Ta-day (10:30 am)" is the most appropriate choice. In contrast to most linear co-distance indexes reported in the literature, the proposed SMiSEE is a novel non-linear soil moisture index that defines the locus of co-moisture points in the scatter plot as unequal distance curves. To investigate the proposed method, two different soil moisture observation networks with different climate and vegetation conditions, namely, SMAPVEX12 in Canada and REMEDHUS in Spain, were applied. The results demonstrated that the efficiency of the proposed 1-year scatter plot, which was even applicable to, and improved upon, other indexes in the literature. In addition, SMiSEE outperformed SEE, iTVDI and phi as the most recent similar indexes, achieving correlation coefficients of 0.65 and 0.74 for the SMAPVEX12 and REMEDHUS networks, respectively. These results appear to be promising, especially for the vegetated area in the SMAPVEX12 network.