3. Results
The results showed that gathered data satisfy assumptions of normal distribution (all P > 0.05) and homogeneity of variance (all P > 0.05) (Table 3). Based on the results, vegetation components significantly varied between control and mulch treated sites (all P < 0.05), in which total cover and litter increased both about 76% while bare soil decreased about 26% under mulch treated (Fig. 2). Species richness was not varied between control and mulch treated sites (t -value= -0.067 andP =0.947). In contrast, species evenness, Shannon, and Simpson diversity indices were negatively affected by mulch treatment (Table 3 and Fig. 2). For example, applying mulch treatment resulted in a decline in species evenness, Shannon, and Simpson diversity indices about 88%, 63%, and 71%, respectively (Fig. 2).
Table 3. Results of normal distribution (Shapiro-Wilk Test), homoscedasticity (Levene’s Test), and independent t-test between vegetation and diversity component in control and mulch treated sites. significant values are shown by bold numbers. The non-significant values in Shapiro-Wilk and Levene’s tests indicate normal distribution and homogeneity in variance, respectively.
Figure 2. Mean ± SE comparison of vegetation and diversity components in control and mulch treated sites. significant differences are indicated by different letters.
All RCS parameters (BS, LF and BP) and total RCS significantly differed between control and mulch treated sites (all P < 0.001) while WPC was not significantly different between control and mulch treated sites (t -value= 1.226 and P = 0.236) (Table 4). In addition, RCS class in control and mulch treated areas were very poor and poor, respectively.
Table 4. The comparison results of RCS and their components. BS: bare soil, LF: litter frequency, WPC: weighted palatability classes, BP: biomass production, RCS: rangeland condition score. Significant values are shown by bold numbers.
Before calculating MSAVI maps of 2017 and 2019, the slope and intercept of soil lines regression were calculated by fitting a linear regression between the red and infrared bands (Table 5).
Table 5. The soil line parameters (slope and intercept) are based on a linear regression between the red (as the dependent variable) and infrared bands (as an independent variable) and good of fitness indices.
Then, the MSAVI maps of 2017 and 2019 were produced and reclassified (Fig. 3). As shown in Fig. 4, the vegetation cover significantly increased in 2019 compared to those found in 2017.
Figure 3. The vegetation cover maps based on remotely sensed MSAVI index
The results of map differencing showed that vegetation cover of extent area of studied regions was not varied (52.4%) or increased (44.8%) between 2017 and 2019; however, vegetation cover in some small areas (northwest) was decreased (2.8%) (Fig. 4).
Figure 4. Vegetation cover change monitoring using the differencing method
The results of map differencing showed that vegetation cover of extent area of our studied regions was not varied (52.4%) or increased (44.8%) between 2017 and 2019; however, vegetation cover in some small areas (northwest) was decreased (2.8%) (Fig. 4).
Finally, two-way ANOVA results showed that years (2017 vs. 2019), treats (control vs. mulch), and their interaction had significant effects on vegetation cover value extracted by the MSAVI index (Fig. 5). Based on the results, vegetation cover was greater in 2019 than that of 2017 and also greater in the mulched site in 2019 than the controls site, while there was no difference between vegetation cover in controls and mulch sites in 2017 (Fig. 5).
Figure 5. Results of two-way ANOVA of vegetation cover extracted by MSAVI index for pre and post applying mulching. The significant differences are shown by different letters. ***, indicate significant effects at P < 0.001.