Investigation of statistical analysis of desertification integrated indicators by using the ANOVA Non-Parametric tests

Desertification defines land degredation in the arid, semi arid and subhumid dry regions in consequence of some parameters such as climate change, and human activities. Assessment of sensitivity of a region to desertification by combination of indeces and utilization of methods and standard analysis, leading us to understanding more about desertefication. Many studies have done to assess desertification by combination of desertification indicators. Kasmas et al (2014), analysised correlation between some variables and DESIRE project to considere desertification of a region and found significant correlation between some of such variables. Salvati et al. (2014), studied risk assessment of soil degradation and desertification in the Mediterranean by using nonparametric Kruskal-Wallis test and found that 20 variables from 47 variables have greatest impacts on desertification. Kairis et al. (2013) found that indicators with correlation coefficient of greater than 0.4 (P value = 0.05) are most effective indicators to evaluate risk assessment of desertification and land degredation.
This study aim to evaluate statistical correlation of indicators, effecting desertification by using nonparametric analysis in Khorasan e Razavi province at eastern Iran.
In this study, six indecators include Vegetation cover, Precipitation, Surface temperature, Soil moisture, soil temperature at 0-100 cm underground and Evaporation (evapotranspiration, soil evaporation) for four periods of 2001, 2005, 2009 and 2013 were used. These indeces were products of MODIS, MERRA and TRMM sensors onboard on TERRA satellite. In order to do statistical analysis of the variables, the Spearman correlation tests were used between vegetation index and the other indicators. Nonparametric analysis of Kruskal-Wallis tests were also used to determine variance in different indicators at the four periods,
High correlation between the vegetation index and the evapotranspiration in 2009 shows that increase in vegetation cover causes increasing in evapotranspiration due to a specific climate condition over this area. The same but converse correlation showed less vegetation consequence increase humidity in the subsurface depths to 2013. Kruskal-Wallis test results also showed that there are significant differences in the study time periods for rainfall, evapotranspiration and vegetation due to different climatic conditions prevailing in the region. Highest significant difference is for evapotranspiration in 2001 -2013 due to transition of drought period to a normal climate.
Since a large part of Khorasan e Razavi province is at desertification risk, combination indicators is an appropriate method to estimate potential relationships among of factors contributing to desertification. Methods used in this study may be an effective method to 1) choice appropriate variables associated to desertification 2) Identify spatiotemporal changes of variables by using remote sensing for different periods 3) analysis between variables Based on statistical test. Statistical analysis of variables can be identified the most influence and sensitive variables to desertification