Continuous-time Modelling of Hydrologic Time Series: Shot Noise Models

To simulate the runoff process at short-term aggregation scale (typically one day) the shot noise model has gained considerable reputation. In this paper, the model developed by Murrone et al. (1997) is taken as a notable example of physically-consistent framework to discuss the different aspects related to the building and estimation of a stochastic tool for time-series generation. In particular, applications of the model in various geographic, morphological and climatic conditions give the opportunity to present its features, also in relation to an alternative, bivariate, model configuration. After a general presentation of the model typology and of the literature background, the conceptually-based shot noise model framework is briefly resumed, with specific attention to the inverse estimation of the effective rainfall and to the assessment of the quality of runoff generation. Various issues related to model application in temperate, semi arid and alpine watersheds are then presented and discussed. Even though additional efforts are required to adapt the conceptualization to alpine environments, the shot noise framework demonstrates to be a valid tool, also to accompany physically-based modelling or frequency analysis of extreme events, as it reproduces in a treatable way the basic mechanisms of runoff formation. 1 SHOT NOISE PROCESSES