Remote sensing approach
Remotely sensed vegetation indices are among the main tools used in vegetation surveys (Mirzaei et al., 2015; Abdolalizadeh et al., 2020; Zarei et al., 2020). In arid and semi-arid regions, the soil brightness may accelerate the surface reflectance and causes exaggeration in vegetation cover condition; therefore, some Vegetation Indices (VIs) are developed to be used in such regions by considering the soil line parameters (slope and intercept) that are produced by regression between red and near-infrared bands in bare soil patches. In the current study, we used the Modified Soil-Adjusted Vegetation Index (MSAVI) suggested by Qi et al. (1994). The MSAVI is based on a modification of the L factor (soil adjustment factor) of the SAVI index (Soil-Adjusted Vegetation Index) that tended to make a better correction of the soil background brightness in different vegetation cover conditions. Previously, in the SAVI index, the L factor defined according to the density of the vegetation and the climatic conditions of an area (Qi et al., 1994), which replaced by its calculation based on the slope of the background soil line and other vegetation indices (NDVI and WDVI). The MSAVI vegetation index expressed as the following formula (Qi et al., 1994):
\begin{equation} \frac{NIR-Red}{(NIR+Red+L)}\times\left(1+L\right);\ \ L=1-2\gamma\ (NDVI\times WDVI)\ \nonumber \\ \end{equation}
Where;
NIR= reflectance in the near-infrared band (expressed as reflectance)
RED= reflectance in the visible red band (expressed as reflectance)
NDVI = Normalized Difference Vegetation Index ((NIR-RED)/(NIR+RED))
WDVI = Weighted Difference Vegetation Index (NIR- γRED)
γ = Slope of the background soil line
In this formula, the soil adjust factor is selected as an empirical equation to decrease with decreasing vegetation cover, as in the case in semi-arid regions (Qi et al., 1994). In addition, the L factor ranges from 0 to 1 and multiplied by two (2L) to increase the L dynamic range (Eastman, 2016).
The differencing method was used to detect spatio-temporal trend changes (increasing, decreasing and no changes) between pre- and post-mulch treatment (Singh, 1989). In this method, the digital value of the second map (i.e., MSAVI map of 2017) is subtracted from the first map (i.e., MSAVI map of 2019), pixel by pixel (Mirzaei et al., 2015). This method’s results include positive, negative and zero pixels, indicating increasing, decreasing, and no changes in vegetation cover, respectively. However, this method needs to determine the change threshold, to distinguish the changing area (increasing and decreasing) from the no-change area (Fung and Ledrew, 1988). In the current study, we employed a statistical method (Mirzaei et al., 2015) based on the following equation:
\begin{equation} Z=\ \frac{X_{i}-\overset{\overline{}}{X}}{S}\nonumber \\ \end{equation}
Where Xi is the numerical value of each pixel,\(\overset{\overline{}}{X}\) is the mean score of the pixels, and S is standard deviations.
Finally, to statistically compare the control and mulch treated area, we design a network of random points including 100 points in both the control and the mulch treated area (Fig. 1). The MSAVI values are extracted and analyzed (see statistical analysis section).