Stochastic Model Output Statistics for Bias Correcting and Downscaling Precipitation Including Extremes

Abstract

Precipitation is highly variable in space and time; hence, rain gauge time series generally exhibit additional random small-scale variability compared to area averages. Therefore, differences between daily precipitation statistics simulated by climate models and gauge observations are generally not only caused by model biases, but also by the corresponding scale gap. Classical bias correction methods, in general, cannot bridge this gap; they do not account for small-scale random variability and may produce artifacts. Here, stochastic model output statistics is proposed as a bias correction framework to explicitly account for random small-scale variability. Daily precipitation simulated by a regional climate model (RCM) is employed to predict the probability distribution of local precipitation. The pairwise correspondence between predictor and predictand required for calibration is ensured by driving the RCM with perfect boundary conditions. Wet day probabilities are described by a logistic regression, and precipitation intensities are described by a mixture model consisting of a gamma distribution for moderate precipitation and a generalized Pareto distribution for extremes. The dependence of the model parameters on simulated precipitation is modeled by a vector generalized linear model. The proposed model effectively corrects systematic biases and correctly represents local-scale random variability for most gauges. Additionally, a simplified model is considered that disregards the separate tail model. This computationally efficient model proves to be a feasible alternative for precipitation up to moderately extreme intensities. The approach sets a new framework for bias correction that combines the advantages of weather generators and RCMs.

Multivariate bias adjustment of high-dimensional climate simulations: the Rank Resampling for Distributions and Dependences (R<sup>2</sup>D<sup>2</sup>) bias correction

“…Of course, there is no reason why the rank chronology should be the same. This also implies that this multivariate BC provides deterministic corrections, while some studies pointed out the need for stochastic corrections or at least the need for introducing some stochasticity and variability in the BC process (e.g., Wong et al, 2014;Mao et al, 2015;Volosciuk et al, 2017). Hence, the goals of this paper are -to propose a multivariate BC (MBC) method for both multi-site and multi-variable simulations;…”

A combined statistical bias correction and stochastic downscaling method for precipitation

“…We calibrate the probabilistic regression model Figure 1. Schematic of (black) our combined statistical bias correction and stochastic downscaling model, and (grey) the Wong et al (2014) model. between gridded and point-scale observations and then apply it to the corrected grid-scale time series in the validation period.…”

“…The major problems with these projections are both climate model biases and the gap between gridbox and point scale. Wong et al (2014) developed a model to jointly bias correct and downscale precipitation at daily scales. This approach, however, relied on pairwise correspondence between predictor and predictand for calibration, and, thus, on nudged simulations which are rarely available.…”

“…Standard downscaling approaches in turn have a limited ability to correct systematic biases. Wong et al (2014) developed a model that jointly bias corrects and downscales precipitation at daily scales. However, this approach relies on pairwise correspondence between predictor and predictand for calibration that is only provided by nudged GCM/RCM simulations, and is not able to post-process standard, free-running GCM simulations .…”

“…Here we present a modification of the Wong et al (2014) approach that is designed to also work in principle for free-running GCM/RCMs, such as those available from ENSEMBLES (van der Linden and Mitchell, 2009) or CORDEX (e.g., Jacob et al, 2013). With the aim of combining their respective advantages we combine a statistical bias correction and a stochastic downscaling method.…”