Comprehensive evaluation of an improved large‐scale multi‐site weather generator for Germany

Abstract

In this work, we present a comprehensive evaluation of a stochastic multi-site, multi-variate weather generator at the scale of entire Germany and parts of the neighbouring countries covering the major German river basins Elbe, Upper Danube, Rhine, Weser and Ems with a total area of approximately 580,000 km 2 . The regional weather generator, which is based on a first-order multi-variate auto-regressive model, is setup using 53-year long daily observational data at 528 locations. The performance is evaluated by investigating the ability of the weather generator to replicate various important statistical properties of the observed variables including precipitation occurrence and dry/wet transition probabilities, mean daily and extreme precipitation, multi-day precipitation sums, spatial correlation structure, areal precipitation, mean daily and extreme temperature and solar radiation. We explore two marginal distributions for daily precipitation amount: mixed Gamma-Generalized Pareto and extended Generalized Pareto. Furthermore, we introduce a new procedure to estimate the spatial correlation matrix and model mean daily temperature and solar radiation. The extensive evaluation reveals that the weather generator is greatly capable of capturing most of the crucial properties of the weather variables, particularly of extreme precipitation at individual locations. Some deficiencies are detected in capturing spatial precipitation correlation structure that leads to an overestimation of areal precipitation extremes. Further improvement of the spatial correlation structure is envisaged for future research. The mixed marginal model found to outperform the extended Generalized Pareto in our case. The use of power transformation in combination with normal distribution significantly improves the performance for nonprecipitation variables. The weather generator can be used to generate synthetic event footprints for large-scale trans-basin flood risk assessment.

Process‐Based Flood Risk Assessment for Germany

“…The RFM is a chain of four coupled models: (a) a stochastic multi‐site, multivariate Regional Weather Generator (RWG) simulates synthetic time series of precipitation, temperature, humidity and solar radiation at climate station locations for entire Germany and upstream parts of riparian countries (Hundecha et al., 2009; Nguyen et al., 2021). (b) These time series are fed into the hydrological model ‐ Soil and Water Integrated Model (SWIM) (Krysanova et al., 1998) to obtain runoff input into the river network.…”

“…In the second stage, if necessary, it generates daily non-precipitation variables such as temperature (maximum, minimum, average), relative humidity and solar radiation conditioned on the state (dry/wet) of the generated precipitation. RWG was originally introduced by Hundecha et al (2009) and has recently been improved and evaluated for all major German river basins by Nguyen et al (2021). For non-zero precipitation values at individual stations, a theoretical distribution is fitted to simulate daily precipitation sums.…”

“…For non-zero precipitation values at individual stations, a theoretical distribution is fitted to simulate daily precipitation sums. The RWG provides two options: (a) a six-parameter mixed distribution (Frigessi et al, 2002;Hundecha et al, 2009) combining a Gamma distribution for bulk precipitation and a Generalized Pareto distribution for extreme precipitation, (b) a parsimonious three-parameter extended Generalized Pareto distribution allowing a smooth transition between bulk distribution and extreme values (Naveau et al, 2016;Nguyen et al, 2021). In the present study, the mixed distribution was used because it generally shows a better fit to observational data although the fitting procedure is much more computationally expensive.…”

Process-based flood risk assessment for Germany

“…The RFM is a chain of four coupled models: (a) a stochastic multi-site, multivariate Regional Weather Generator (RWG) simulates synthetic time series of precipitation, temperature, humidity and solar radiation at climate station locations for entire Germany and upstream parts of riparian countries (Hundecha et al, 2009;Nguyen et al, 2021). (b) These time series are fed into the hydrological model -Soil and Water Integrated Model (SWIM) (Krysanova et al, 1998) to obtain runoff input into the river network.…”

“…The first component of the RFM is the multi-site, multi-variate stochastic Regional Weather Generator (RWG) based on a first-order multivariate autoregressive model considering spatial correlation structure (Hundecha et al, 2009;Nguyen et al, 2021). The model has two simulation stages.…”