A Dissertation Submitted to the Council of the College of Agricultural Engineering Sciences at Salahaddin University-Erbil in Partial Fulfillment of the Requirements for the Degree of PhD in Soil Science (Soil and Water Conservation)
Estimate Rainfall Erosivity Factor from Rainfall Depth
Soil erosion by water is a major threat to sustainable agriculture and food production across the globe. Rainfall erosivity (R-Factor) is one of the important parameters for water erosion risk assessment. One obstacle to estimate this parameter in the study region is the lack of detailed rainfall intensity data. Furthermore, many attempts were made worldwide to derive the functional relationship between rainfall erosivity and more readily available rainfall data, but most of the derived models have limited application out of the area where they were developed without careful validation. To overcome the problem of data scarcity, univariate and multivariate models were derived for estimating rainfall erosivity along with the validation of different models derived inside and outside of the Mediterranean region. The database for model development were rainfall data at different time scales obtained from 25 stations distributed across the study region recorded during 2000-2018. About 50% of the data set obtained from pluviographic stations. The computed rainfall erosivity was in the context of USLE and RUSLE models. The explanatory variables encompassed annual rainfall (P), Fournier index (FI), modified Fournier index (MFI) and precipitation concentration index (PCI) and geographical coordinates of the stations. A host of statistical indices was selected to evaluate adequately the model’s performance. Further, the models were cross-validated using K-fold procedure and unseen data.