Publication Date:
2024-02-28
Description:
〈title xmlns:mml="http://www.w3.org/1998/Math/MathML"〉Abstract〈/title〉〈p xmlns:mml="http://www.w3.org/1998/Math/MathML" xml:lang="en"〉High spatio‐temporal resolution near‐surface projected data is vital for climate change impact studies and adaptation. We derived the highest statistically downscaled resolution multivariate ensemble currently available: daily 1 km until the end of the century. Deep learning models were employed to develop transfer functions for precipitation, water vapor pressure, radiation, wind speed, and, maximum, mean and minimum temperature. Perfect prognosis is the particular statistical downscaling methodology applied, using a subset of the ReKIS data set for Saxony as predictands, the ERA5 reanalysis as during‐training predictors and the CORDEX‐EUR11 ensemble as projected predictors. The performance of the transfer functions was validated with the VALUE framework, yielding highly satisfactory results. Particular attention was given to the three major perfect prognosis assumptions, for which several tests were carried out and thoroughly discussed. From the latter, we corroborated their fulfillment to a high degree, thus, the derived projections are considered adequate and relevant for impact modelers. In total, 18 runs for RCP85, 1 for RCP45, and 4 for RCP26 were downscaled under both stochastic and deterministic approaches. This multivariate ensemble could drive more accurate and diverse impact studies in the region. Generally, the projected climatologies are in agreement with coarser resolution projections. Nevertheless, statistical particularities were observed for some projections, thus, a list of caveats for potential users is given. Due to the scalability of the presented methodology, further possible applications with additional datasets are proposed. Lastly, several potential improvement prospects are discussed toward the ideal subsequent iteration of the perfect prognosis statistical downscaling methodology.〈/p〉
Description:
Plain Language Summary: There is a great worldwide demand for high spatio‐temporal resolution projections to develop climate change adaptation and mitigation schemes. Despite recent improvements, the resolution of both global and regional climate models is still too coarse to properly represent local variability, particularly in complex terrains. Depending on the application, impact modelers and decision makers require kilometer‐scale projections, with a minimum daily temporal resolution, of near‐surface variables. To fill this information gap, we employed artificial intelligence algorithms to downscale, to a novel daily 1 km resolution, a projection ensemble until the end of the century consisting of precipitation, water vapor pressure, radiation, wind speed, and, maximum, mean and minimum temperature. The ensemble comprises 18 runs of the business‐as‐usual worst‐case scenario (RCP85), 1 run of the stabilization scenario (RCP45), and 4 of the optimistic low‐emissions scenario (RCP26). The main assumptions of the methodology were thoroughly tested and discussed. The validation carried out yielded highly satisfactory results. Thus, we consider the projections to be adequate and relevant for impact studies. The region studied is located in Saxony (Germany), still, the methodology shown is potentially applicable anywhere in the world.〈/p〉
Description:
Key Points:
〈list list-type="bullet"〉
〈list-item〉
〈p xml:lang="en"〉Highest statistically downscaled spatio‐temporal resolution multivariate ensemble currently available, consisting of 23 projection runs〈/p〉〈/list-item〉
〈list-item〉
〈p xml:lang="en"〉We downscaled precipitation, water vapor pressure, radiation, wind speed, and, maximum, mean and minimum temperature〈/p〉〈/list-item〉
〈list-item〉
〈p xml:lang="en"〉The methodology complied to a high degree with the three perfect prognosis assumptions and is scalable to other spatio‐temporal resolutions〈/p〉〈/list-item〉
〈/list〉
〈/p〉
Description:
European Social Fund, Freistaat Sachsen
http://dx.doi.org/10.13039/501100004895
Description:
https://rekis.hydro.tu-dresden.de/
Description:
https://doi.org/10.5281/zenodo.7570247
Description:
https://doi.org/10.5281/zenodo.7559173
Description:
https://doi.org/10.5281/zenodo.7558945
Description:
https://doi.org/10.5281/zenodo.8059248
Description:
https://doi.org/10.5281/zenodo.8198925
Keywords:
ddc:551.6
;
climate change
;
statistical downscaling
;
perfect prognosis
;
ERA5
;
CORDEX
;
deep learning
;
multivariate ensemble
Language:
English
Type:
doc-type:article
Permalink
