Saturation-excess overland flow in the European loess belt:An underestimated process?

A major challenge in runoff and soil erosion modelling is the adequate representation of the most relevant processes in models while avoiding over parameterization.In the European loess belt,pro-gressive soil crusting during rainfall events,resulting in infiltration-excess runoff,is usually considered the dominant process generating runoff on catchments covered with silty soils.Saturation-excess may also occur and affect their runoff and erosion behavior.However,saturation-excess runoff occurrence and quantification have rarely been performed and is usually not taken into account when modelling runoff and erosion in these environments.Accordingly,a continuous simulation of the Austreberthe catchment(214 km^(2)),located in the European loess belt(Normandy,France),was conducted with the new Water and Sediment(WaterSed)model over 12 years,corresponding to more than 780 individual rainfall events,at a 25 m spatial resolution.The inter-annual variability of runoff and erosion was closely linked to the number of intense events per year and their distribution through the year.The model was properly calibrated over a representative set of 35 rainfall events,considering either infiltration-excess and/or saturation-excess runoff.It was also able to reproduce the measured runoff volume for most of the monitoring period.However,the three years with most rainfall were adequately modelled only including saturation-excess runoff.An analysis performed at the seasonal scale revealed that saturation was modelled in the catchment during almost all of the modelling period,suggesting the importance of this often overlooked process in current modelling attempts.

Impact of Rainfall Temporal Heterogeneity on Relationship between Curve Number and Rainfall Depth in the Zagros Mountain Region,Iran

The impact of temporal variation of rainfall on the relationship between rainfall and catchment response is investigated in a catchment with high temporally variable rainfalls and a high percentage of permeable soils in the southwest of Iran.Twenty-nine storm events are classified into two classes, High Temporal heterogeneous(HT) and Low Temporal heterogeneous(LT) events using the variogram technique and the storm events of each class are analyzed to detect the relationship between Curve Number(CN) and rainfall depth. It is found that there is not a similar correlation between CN values and rainfall depths for both temporally variable classes, and hence, two different responses can be observed in the catchment according to rainfall temporal heterogeneities. For HT events, a complacent behavior is detected in which the CNs decline as rainfall depth increases while a different response, violent behavior, is observed for LT events in which the CNs rise and asymptotically approach a constant value with increasing storm size. This considerable difference between CN-P relationships derived from the two temporally variable classes of rainfall is attributed to the provocation of different runoff generation mechanisms, infiltration-excess and saturation-excess caused by rainfall temporal heterogeneities. Moreover, the results support the validity of variogram technique to classify storm events into two LT and HT classes.