Ensemble analogue downscaling of the Twentieth Century Reanalysis over France for 140-year-long hydrological reconstructions
| Autor: | Vidal, Jean-Philippe, Caillouet, L., Sauquet, Eric, Graff, B. |
|---|---|
| Přispěvatelé: | Hydrologie-Hydraulique (UR HHLY), Institut national de recherche en sciences et technologies pour l'environnement et l'agriculture (IRSTEA), Compagnie Nationale du Rhône (CNR), Irstea Publications, Migration |
| Jazyk: | angličtina |
| Rok vydání: | 2017 |
| Předmět: | |
| Zdroj: | International Workshop on Stochastic Weather Generators for Hydrological Applications (SWGen-Hydro) International Workshop on Stochastic Weather Generators for Hydrological Applications (SWGen-Hydro), Sep 2017, Berlin, Germany. pp.54 |
| Popis: | International audience; The record length of streamflow observations is generally limited to the last 50 years even in data-rich countries like France. This is not enough to properly explore the long-term hydrometeorological variability, which is a key to better disentangle the effects of anthropogenic climate change from natural variability. In order to overcome this limit, the SCOPE (Spatially COherent Probabilistic Extension) ensemble analogue downscaling method is used to bridge the scale gap between the large-scale Twentieth Century Reanalysis and local meteorological variables relevant for catchment-scale hydrology. SCOPE extends the SANDHY method (Stepwise Analogue Downscaling Method for Hydrology) originally developed for daily quantitative precipitation forecast. SCOPE consists in refining analogues dates and associated predict and values found by applying SANDHY with locally optimized predictor domains over France. SCOPE first includes two additional analogy levels with large-scale 2m-temperature and sea surface temperature as predictors. The driest analogue dates are then removed to correct for the remaining wet bias. Finally, the Schaake Shuffle procedure ensures the spatial coherence of 8-km precipitation and temperature fields for each of the 25 ensemble members of this meteorological reconstruction called SCOPE Climate. SCOPE Climate presents very little bias in comparison to the reference Safran surface reanalysis on their common period (1958-2012). There is a slight overestimation of precipitation in spring (around 10%), and an underestimation/overestimation (within 1 deg C) in summer/winter mean temperature, respectively. SCOPE Climate is able to capture the seasonal variability of precipitation and temperature as well as the increasing trend in temperature since the 1990s. It also highlights the large multi decadal variability in precipitation that occurred over the last 140 years. A continuous hydrological modelling using SCOPE Climate as forcings then allowed reconstructing 25-member ensembles of daily streamflow time series for more than 600 near-natural catchments in France over the 1871-2012 period. The resulting SCOPE Hydro reconstruction dataset was recently used to identify and characterize spatiotemporal extreme low-flow events in a homogeneous way over France and over a 140-year period. This analysis built on innovative aspects for defining local extreme low-flow events and for matching events across the whole country. On top of recent events like 1976 or 1989-1990, it brought forward the outstanding 1921 and 1940s events but also older and less known ones that still are the most severe ones to date in specific regions, like 1893 in the North-East of France and 1878 around the Mediterranean coast. The SCOPE Climate and SCOPE Hydro thus pave the way for further analyses of long-term meteorological and hydrological variability and trends in a comprehensive way over the whole France. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|