Simulation of multisite precipitation using an extended chain‐dependent process

Abstract

[1] The chain-dependent process is a popular stochastic model for precipitation sequence data. In this paper, the effect of daily regional precipitation occurrence is incorporated into the stochastic model. This model is applied to analyze the daily precipitation at a small number of sites in the upper Waitaki catchment, New Zealand. In this case study, the probability distributions of daily precipitation occurrence and intensity, spatial dependences, and the relation between precipitation and atmospheric forcings are simulated quite well. Specifically, some behaviors which are not well modeled by existing models, such as the extremal behavior of daily precipitation intensity, the lag 1 cross correlation of daily precipitation occurrence, spatial intermittency, and spatial correlation of seasonal precipitation totals, are significantly improved. Moreover, a new and simpler approach is proposed which successfully eliminates overdispersion, i.e., underestimation of the variance of seasonal precipitation totals.Citation: Zheng, X., J. Renwick, and A. Clark (2010), Simulation of multisite precipitation using an extended chain-dependent process, Water Resour. Res., 46, W01504,

ApproximateBayesian computation methods for daily spatiotemporal precipitation occurrence simulation

“…Hidden Markov models have been used to model occurrence [ Hughes et al ., ] as well as intensity [ Ailliot et al ., ; Charles et al ., ]. Other methods include nearest‐neighbors resampling [ Apipattanavis et al ., ; Buishand and Brandsma , ; Rajagopalan and Lall , ], generalized chain‐dependent processes [ Zheng and Katz , ; Zheng et al ., ], power transformation to normality [ Sansó and Guenni , ; Yang et al ., ], artificial neural network methods [ Cannon , ], copula‐based approaches [ Bárdossy and Pegram , ], and multifractal rainfall models [ Gupta and Waymire , ; Lovejoy and Schertzer , ; Menabde et al ., ; Deidda , ; Veneziano and Langousis , ; Langousis and Veneziano , ; Veneziano et al ., ; Veneziano and Langousis , ]. Current approaches to this problem typically seek the assistance of latent multivariate normals, sometimes including a transformation, to generate the occurrence/intensity values over a spatial domain.…”

Daily spatiotemporal precipitation simulation using latent and transformed Gaussian processes

“…It is critical to capture the domain aggregate behavior of precipitation intensity and dry or wet spells, which play important roles in hydrologic planning and water resource management. There are a number of approaches to spatial‐temporal modeling of precipitation, including hidden Markov models for occurrence [ Hughes and Guttorp , 1999] and intensity [ Ailliot et al , 2009; Charles et al , 1999], resampling based on nearest neighbors [ Apipattanavis et al , 2007; Buishand and Brandsma , 2001; Rajagopalan and Lall , 1999], generalized chain‐dependent processes [ Zheng and Katz , 2008; Zheng et al , 2010], power transformation of precipitation to normality [ Sansó and Guenni , 2000; Yang et al , 2005], artificial neural networks [ Cannon , 2008], or copula‐based approaches [ Bárdossy and Pegram , 2009]. One of the main advantages of stochastic precipitation generators is uncertainty quantification over a spatial domain, which is used, for example, in assessment of impacts of climate change [ Kilsby et al , 2007; Mehrotra and Sharma , 2010]; see also Burton et al [2008] for recent spatiotemporal precipitation simulation software.…”

A stochastic space-time model for intermittent precipitation occurrences

“…The spatio-temporal dependence in rainfall zeros is a critical aspect of any spacetime stochastic model for precipitation. On the daily time scale, Katz (1977) used a Markov chain model to describe the temporal dependence of precipitation occurrence at individual locations, Zheng and Katz (2008) and Zheng, Renwick and Clark (2010) extended the Markov chain model for simulations of the multisite precipitation, Hughes and Guttorp (1999) introduced a spatio-temporal model of precipitation occurrence using hidden Markov models, and Ailliot, Thompson and Thomson (2009) developed a hidden Markov model using censored Gaussian processes.…”