A hybrid stochastic rainfall model that reproduces some important rainfall characteristics at hourly to yearly timescales

Abstract

Abstract. A novel approach to stochastic rainfall generation that can reproduce various statistical characteristics of observed rainfall at hourly to yearly timescales is presented. The model uses a seasonal autoregressive integrated moving average (SARIMA) model to generate monthly rainfall. Then, it downscales the generated monthly rainfall to the hourly aggregation level using the Modified Bartlett–Lewis Rectangular Pulse (MBLRP) model, a type of Poisson cluster rainfall model. Here, the MBLRP model is carefully calibrated such that it can reproduce the sub-daily statistical properties of observed rainfall. This was achieved by first generating a set of fine-scale rainfall statistics reflecting the complex correlation structure between rainfall mean, variance, auto-covariance, and proportion of dry periods, and then coupling it to the generated monthly rainfall, which were used as the basis of the MBLRP parameterization. The approach was tested on 34 gauges located in the Midwest to the east coast of the continental United States with a variety of rainfall characteristics. The results of the test suggest that our hybrid model accurately reproduces the first- to the third-order statistics as well as the intermittency properties from the hourly to the annual timescales, and the statistical behaviour of monthly maxima and extreme values of the observed rainfall were reproduced well.

Modelling rainfall with a Bartlett–Lewis process: New developments

“…As a result, we find that these Bartlett-Lewis models still tend rather to underestimate the variability 450 at scales coarser than a week, which provides a confirmation of the wisdom of developing combinations of Bartlett-Lewis models with simple coarse-scale models to capture long-term variability (e.g. see Park et al (2019) and forthcoming work).…”

“…If we want the model to be able to capture longer term variability (as would certainly be required to reproduce climate variability for instance), then this 55 issue must be addressed. The most promising ways forward in this respect come from combining the Poisson-cluster model with a coarse-scale model that captures much of the longer-term variability (Park et al, 2019), or from letting climatological information guide the weighting to be assigned to different months in the data in calibrating the model (Kaczmarska et al (2015); Cross et al (2019)). Both approaches represent important developments.…”

Assessing Spatial Flood Risk from Multiple Flood Sources in a Small River Basin: A Method Based on Multivariate Design Rainfall

“…However, the precision of generated rainfall time series data is sometimes not satisfactory. Nonetheless, these approaches are being continuously developed [51,52] and could be better in the future. In recent studies, long-term climate ensemble forecasting data simulated by global circulation model (GCM) or regional climate model (RCM) are used for flood risk assessments [15,46].…”

Modelling rainfall with a Bartlett–Lewis process: new developments

“…This can be by defining the number cells in a storm as a random variable, with another random variable modelling the delays from the storm to the cell arrival time. This defines a Neyman-Scott process (see Cowpertwait, 1998;Evin and Favre, 2008;Paschalis et al, 2014). Alternatively, a second homogeneous Poisson process defines the cell arrival times over a duration of storm activity that defines a random variable (see Onof and Wheater, 1993;Khaliq and Cunnane, 1996;Verhoest et al, 1997;Kossieris et al, 2018).…”

“…If we want the model to be able to capture longer-term variability (as would certainly be required to reproduce climate variability for instance), then this issue must be addressed. The most promising ways forward in this respect come from combining the Poisson cluster model with a coarse-scale model that captures much of the longerterm variability (Park et al, 2019) or from letting climatological information guide the weighting to be assigned to different months in the data in calibrating the model (Kaczmarska et al, 2015;Cross et al, 2020). Both approaches represent important developments.…”