بررسی خصوصیات الگوی توزیع مکانی فرسایندگی باران در استان هرمزگان

Introduction
Water and wind erosion are the two main causes of land degradation. The extent of water erosion is greater, and its outcomes are much more complicated, than the other types of erosion in Iran (Sadeghi et al. 2011). Rainfall erosivity is one of the key parameters for soil erosion risk assessment under future landuse and climate change (Meusburger et al. 2012). Rainfall erosivity is the potential ability for rainfall to cause soil loss. Erosivity can be quantified by means of the R factor calculation of the universal soil loss equation (USLE). The universal soil loss equation (USLE) is a mathematical model to predict soil loss. Therefore, the awareness of its level and variation is critical to improve the management of the land. Thus, one can note that isoerodent maps are essential in evaluating the potential of soil erosion over specific region. The objective of this study was to investigate the spatial pattern of R factor using the modified Fournier index (MFI) over Hormozgan province.
 
Methodology
Hormozgan province is located in south of Iran and north of Persian Gulf (25°25΄ N, 52°39΄ E to 27°18΄ N, 59°14 ΄E) and covers an area of 70712 km2. The mean annual precipitation varies from 100 mm to 410 mm.
The original method to calculate the rainfall erosivity (R factor) for a storm event requires pluviographic records. This information is difficult to obtain in many parts of the world, and its computing process is difficult and time-consuming. Therefore, simplified methods for estimating rainfall erosivity by using readily available data have been presented and are used in many countries. In other words, these simplified methods such as the modified Fournier index can estimate monthly or annual values of rainfall erosivity by using mean annual and mean monthly rainfall records. The modified Fournier index (MFI) is defined as the annual sum of the square of the monthly maximum amount of rainfall divided by annual precipitation. The MFI was defined as: (MFI= ). Where P < /em> is annual precipitation (mm), and Pi is the monthly rainfall amount (i = 1, . . ., 12) from January to December. The R factor can be obtained by the regression equation, that its predictor variable is MFI (Renard and Freimund, 1994).
It is indispensable for the rainfall erosivity data to follow a normal distribution. In this study, the normality of data is assessed by the Kolmogorov-Smirnov test. Furthermore, a semivariance analysis is performed by GS+ software for evaluating the spatial correlation of the data and opting the best-fit variogaram model and the superior interpolation technique for mapping rainfall erosivity at a regional scale.
 
Results and discussion
Testing different variogram models such as the linear, spherical, exponential, and gaussian models revealed that the general spatial pattern of rainfall erosivity was better captured by the linear model in Hormozgan province. Additionally, testing different algorithms for spatial interpolation including the inverse distance weighting (IDW), kriging and co-kriging methods indicated that according to the results of performance criteria (the RMSE and R2), the IDW method -with the value of 3 for the distance power- was the superior algorithm for generating rainfall erosivity map in the study area. Cross-correlation matrix showed a significant correlation between the annual erosivity and the annual rainfall in Hormozgan province. Meanwhile, any significant correlation between the annual erosivity and other considered variables (i.e., latitude, longtitude, elevation) was not recorded. The Results of this study indicated that the amounts of rainfall erosivity varied from 32 MJ mm ha-1h-1yr-1 at Jask station to 414 MJ mm ha-1h-1yr-1 at Esteghlal Dam station. The highest frequency of rainfall erosivity was belong to a range between 100 MJ mm ha-1h-1yr-1 and 140 MJ mm ha-1h-1yr-1 that covers about 40% of the study area. Additionally, 7.6% of the total area of Hormozgan province affected by rain drops with erosivity force more than 200 MJ mm ha-1h-1yr-1.
 
Conclusion
The objective of this study was to investigate the characteristics of spatial distribution of the rainfall erosivity over Hormozgan province. The modified Fournier index (MFI) was computed at 67 weather stations during a 21-year that was chosen as a common period. Then, the R factor was obtained according to the MFI. The Authors revealed that the variations of rainfall erosivity not followed a regular spatial pattern in Hormozgan province. Meanwhile, Mapping the interpolated values of R factor is very useful to illustrate how the rainfall erosivity affects the soil erosion and to deliver an important source of information for conducting erosion risk assessment at a regional scale such as Hormozgan province.