ALBERT

Description: Much of our knowledge about future changes in precipitation relies on global (GCMs) and/or regional climate models (RCMs) that have resolutions which are much coarser than typical spatial scales of precipitation, particularly extremes. 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. Here we present an extension of this approach that separates the downscaling from the bias correction and in principle is applicable to free-running GCMs/RCMs. In a first step, we bias correct RCM-simulated precipitation against gridded observations at the same scale using a parametric quantile mapping (QMgrid) approach. In a second step, we bridge the scale gap: we predict local variance employing a regression-based model with coarse-scale precipitation as a predictor. The regression model is calibrated between gridded and point-scale (station) observations. For this concept we present one specific implementation, although the optimal model may differ for each studied location. To correct the whole distribution including extreme tails we apply a mixture distribution of a gamma distribution for the precipitation mass and a generalized Pareto distribution for the extreme tail in the first step. For the second step a vector generalized linear gamma model is employed. For evaluation we adopt the perfect predictor experimental setup of VALUE. We also compare our method to the classical QM as it is usually applied, i.e., between RCM and point scale (QMpoint). Precipitation is in most cases improved by (parts of) our method across different European climates. The method generally performs better in summer than in winter and in winter best in the Mediterranean region, with a mild winter climate, and worst for continental winter climate in Mid- and eastern Europe or Scandinavia. While QMpoint performs similarly (better for continental winter) to our combined method in reducing the bias and representing heavy precipitation, it is not capable of correctly modeling point-scale spatial dependence of summer precipitation. A strength of this two-step method is that the best combination of bias correction and downscaling methods can be selected. This implies that the concept can be extended to a wide range of method combinations.

Description: We demonstrate both analytically and with a modelling example that cross-validation of free running bias-corrected climate change simulations against observations is misleading. The underlying reasoning is as follows: a cross-validation can have in principle two outcomes. A negative (in the sense of not rejecting a Null hypothesis), if the residual bias in the validation period after bias correction vanishes; and a positive, if the residual bias in the validation period after bias correction is large. It can be shown analytically that the residual bias depends solely on the difference between the simulated and observed change between calibration and validation period. These changes, however, depend mainly on the realisations of internal variability in the observations and climate model. As a consequence, also the outcome of a cross-validation is dominated by internal variability, and does not allow for any conclusion about the sensibility of a bias correction. In particular, a sensible bias correction may be rejected (false positive) and a non-sensible bias correction may be accepted (false negative). We therefore propose to avoid cross-validation when evaluating bias correction of free running bias-corrected climate change simulations against observations. Instead, one should evaluate temporal, spatial and process-based aspects.

Description: Compound events are multivariate extreme events in which the individual contributing variables may not be extreme themselves, but their joint – dependent – occurrence causes an extreme impact. The conventional univariate statistical analysis cannot give accurate information regarding the multivariate nature of these events. We develop a conceptual model, implemented via pair-copula constructions, which allows for the quantification of the risk associated with compound events in present day and future climate, as well as the uncertainty estimates around such risk. The model includes meteorological predictors which provide insight into both the involved physical processes, and the temporal variability of CEs. Moreover, this model provides multivariate statistical downscaling of compound events. Downscaling of compound events is required to extend their risk assessment to the past or future climate, where climate models either do not simulate realistic values of the local variables driving the events, or do not simulate them at all. Based on the developed model, we study compound floods, i.e. joint storm surge and high river runoff, in Ravenna (Italy). To explicitly quantify the risk, we define the impact of compound floods as a function of sea and river levels. We use meteorological predictors to extend the analysis to the past, and get a more robust risk analysis. We quantify the uncertainties of the risk analysis observing that they are very large due to the shortness of the available data, though this may also be the case in other studies where they have not been estimated. Ignoring the dependence between sea and river levels would result in an underestimation of risk, in particular the expected return period of the highest compound flood observed increases from about 20 to 32 years when switching from the dependent to the independent case.

Description: Compound events (CEs) are multivariate extreme events in which the individual contributing variables may not be extreme themselves, but their joint – dependent – occurrence causes an extreme impact. Conventional univariate statistical analysis cannot give accurate information regarding the multivariate nature of these events. We develop a conceptual model, implemented via pair-copula constructions, which allows for the quantification of the risk associated with compound events in present-day and future climate, as well as the uncertainty estimates around such risk. The model includes predictors, which could represent for instance meteorological processes that provide insight into both the involved physical mechanisms and the temporal variability of compound events. Moreover, this model enables multivariate statistical downscaling of compound events. Downscaling is required to extend the compound events' risk assessment to the past or future climate, where climate models either do not simulate realistic values of the local variables driving the events or do not simulate them at all. Based on the developed model, we study compound floods, i.e. joint storm surge and high river runoff, in Ravenna (Italy). To explicitly quantify the risk, we define the impact of compound floods as a function of sea and river levels. We use meteorological predictors to extend the analysis to the past, and get a more robust risk analysis. We quantify the uncertainties of the risk analysis, observing that they are very large due to the shortness of the available data, though this may also be the case in other studies where they have not been estimated. Ignoring the dependence between sea and river levels would result in an underestimation of risk; in particular, the expected return period of the highest compound flood observed increases from about 20 to 32 years when switching from the dependent to the independent case.

Description: The spatial dependence of meteorological variables is crucial for many impacts, for example, droughts, floods, river flows, energy demand, and crop yield. There is thus a need to understand how well it is represented in downscaling (DS) products. Within the COST Action VALUE, we have conducted a comprehensive analysis of spatial variability in the output of over 40 different DS methods in a perfect predictor setup. The DS output is evaluated against daily precipitation and temperature observations for the period 1979–2008 at 86 sites across Europe and 53 sites across Germany. We have analysed the dependency of correlations of daily temperature and precipitation series at station pairs on the distance between the stations. For the European data set, we have also investigated the complexity of the downscaled data by calculating the number of independent spatial degrees of freedom. For daily precipitation at the German network, we have additionally evaluated the dependency of the joint exceedance of the wet day threshold and of the local 90th percentile on the distance between the stations. Finally, we have investigated regional patterns of European monthly precipitation obtained from rotated principal component analysis. We analysed Perfect Prog (PP) methods, which are based on statistical relationships derived from observations, as well as Model Output Statistics (MOS) approaches, which attempt to correct simulated variables. In summary, we found that most PP DS methods, with the exception of multisite analog methods and a method that explicitly models spatial dependence yield unrealistic spatial characteristics. Regional climate model‐based MOS methods showed good performance with respect to correlation lengths and the joint occurrence of wet days, but a substantial overestimation of the joint occurrence of heavy precipitation events. These findings apply to the spatial scales that are resolved by our observation network, and similar studies with higher resolutions, which are relevant for small hydrological catchment, are desirable.

Description: Water availability is fundamental to societies and ecosystems, but our understanding of variations in hydroclimate (including extreme events, flooding, and decadal periods of drought) is limited because of a paucity of modern instrumental observations that are distributed unevenly across the globe and only span parts of the 20th and 21st centuries. Such data coverage is insufficient for characterizing hydroclimate and its associated dynamics because of its multidecadal to centennial variability and highly regionalized spatial signature. High-resolution (seasonal to decadal) hydroclimatic proxies that span all or parts of the Common Era (CE) and paleoclimate simulations from climate models are therefore important tools for augmenting our understanding of hydroclimate variability. In particular, the comparison of the two sources of information is critical for addressing the uncertainties and limitations of both while enriching each of their interpretations. We review the principal proxy data available for hydroclimatic reconstructions over the CE and highlight the contemporary understanding of how these proxies are interpreted as hydroclimate indicators. We also review the available last-millennium simulations from fully coupled climate models and discuss several outstanding challenges associated with simulating hydroclimate variability and change over the CE. A specific review of simulated hydroclimatic changes forced by volcanic events is provided, as is a discussion of expected improvements in estimated radiative forcings, models, and their implementation in the future. Our review of hydroclimatic proxies and last-millennium model simulations is used as the basis for articulating a variety of considerations and best practices for how to perform proxy-model comparisons of CE hydroclimate. This discussion provides a framework for how best to evaluate hydroclimate variability and its associated dynamics using these comparisons and how they can better inform interpretations of both proxy data and model simulations.We subsequently explore means of using proxy-model comparisons to better constrain and characterize future hydroclimate risks. This is explored specifically in the context of several examples that demonstrate how proxy-model comparisons can be used to quantitatively constrain future hydroclimatic risks as estimated from climate model projections.

Description: VALUE is an open European network to validate and compare downscaling methods for climate change research. VALUE aims to foster collaboration and knowledge exchange between climatologists, impact modellers, statisticians, and stakeholders to establish an interdisciplinary downscaling community. A key deliverable of VALUE is the development of a systematic validation framework to enable the assessment and comparison of both dynamical and statistical downscaling methods. In this paper, we present the key ingredients of this framework. VALUE's main approach to validation is user- focused: starting from a specific user problem, a validation tree guides the selection of relevant validation indices and performance measures. Several experiments have been designed to isolate specific points in the downscaling procedure where problems may occur: what is the isolated downscaling skill? How do statistical and dynamical methods compare? How do methods perform at different spatial scales? Do methods fail in representing regional climate change? How is the overall representation of regional climate, including errors inherited from global climate models? The framework will be the basis for a comprehensive community-open downscaling intercomparison study, but is intended also to provide general guidance for other validation studies.