Data Description¶
One of the corner stones of the NUKLEUS project is the provisioning of climate model output data. NUKLEUS provides a wide variety of data, not only regional model output data. Besides climate model output data of different flavours, we also offer different types of observational datasets. Altogether we offer roughly 7 million data files spread across 8 data projects. Most prominent are the cmip6, cordex and the nukleus data projects. Information how to access the data can be found in section Data Access.
Overview¶
The most important datasets provided by NUKLEUS are briefly outlined here:
cordex:
This data project holds all cordex data available at the German Climate Computing Centre (DKRZ).cordex-nukleus:
A recommended subset of cordex models that represent the German climate specifically well. The subset consists of 25 different model members. Those members consist of output from 5 global models (GCMs) that has been downscaled by 5 different regional models (RCMs). The resulting ensemble reflects the best estimate of uncertainty and accuracy at the given cordex simulation scale (roughly 12 km) and domain (Central Europe).nukleus:
This data project is the heart of the NUKLEUS project and hosts the model output of the high resolution (roughly 3 km) historical and climate change scenario simulations. Currently, some of these simulations are still in process. You can find further information and the simulation status in the section NUKLEUS Simulations. This data project has also a number of bias-corrected products, which can generally be seen as more realistic than their non-corrected counter parts (see Currently Available Bias-corrected Data).reanalysis:
This data project offers a variety of reanalysis products. Most prominent is the ERA5 Reanalysis Dataset at a global resolution of 30 km in space and 1 hour in time.observations:
Various observational datasets such as gridded radar based precipitation estimates RADOLAN, gridded station based precipitation estimates HYRAS, and others.
Some more useful information on climate modelling in general and the NUKLEUS simulations specifically can also be found in the climate register glossary.
NUKLEUS Simulations¶
Key product of NUKLEUS are climate simulations at two different horizontal resolutions. This novel ensemble consists of output of 3 global models (GCMs) that has been downscaled by 3 different regional models (RCMs). The following table gives an overview of the used GCMs and RCMs.
| GCMs | RCMs |
|---|---|
| EC-EARTH3-Veg | COSMO-CLM (CCLM) |
| MIROC6 | ICON-CLM |
| MPI-ESM1-2-HR | REMO |
Model Domains and Resolution¶
In a first step, the GCMs are downscaled by RCMs to a resolution of 12 km (EUR-11). These simulation domain covers entire Europe (see Figure 1, left). In a second step, the EUR-11 simulations are further downscaled by the same RCMs to a high resolution of 3 km (CEU-3) for a smaller domain covering Germany and surrounding areas (see Figure 1, right).
Simulation types and time periods¶
Within NUKLEUS, different types of simulations are performed.
At first, an evaluation run done with each of the used regional models (RCMs) driven by the ERA5 reanalysis data. The evaluation periods for the 12 km and the 3 km simulations are given in the table below. The aim of the evaluation runs is to best reproduce the historical atmospheric condition. To reach this, the ERA5 data is assimilated into the model at any available time step (every 3 hours).
The second simulation type is the so called historical run. For this, a historical reference period is simulated. The historical period for the 12 km and the 3 km simulations are also given in the table below. In contrast to the evaluation runs, the historical runs are initialized ones with observational/reanalysis data and then run free for the entire simulation period. The purpose is to check, whether a model is capable to reproduce the historical climate appropriate.
| Type | EUR-11 | CEU-3 |
|---|---|---|
| evaluation | 1979-2020 | 2005-2014 |
| historical | 1951-2014 | 1961-1990 |
The third type of simulations are the scenario runs. These simulations are initialized with current climate conditions and run into the future. For the future evolution, e.g., of the emission of greenhouse gases, various climate scenarios (SSP) are defined and commonly used in climate modelling. Within NUKLEUS, transient (continuous) scenario runs are performed at the 12 km resolution covering the entire period 2015-2100. Main focus is on the scenario ssp370. With 7 W/m² by the year 2100, this scenario is in the upper-middle part of the full range of scenarios. Within NUKLEUS, other scenarios are only used for sensitivity studies with the model combination of EC-EARTH and ICON.
At the 3 km resolution, two time slices of 30 years each are chosen representing a period in which climate reaches a 2 degree global warming (global warming level, GWL2K) or a 3 degree global warming (GWL3K) compared to preindustrial conditions (1881-1910).[1] The 30-year periods are centered around the year, when the global model (GCM) reaches the corresponding warming level. Due to their varying climate sensitivity, the periods for GWL2K and GWL3K varies among the used GCMs. The table below gives an overview of the different simulation periods for the 3 km scenario runs.
| Scenario | GWL | EC-EARTH3-Veg | MIROC6 | MPI-ESM1-2-HR |
|---|---|---|---|---|
| ssp245 | GWL2K GWL3K |
2026-2055 2062-2091 |
- | - |
| ssp370 | GWL2K GWL3K |
2024-2053 2047-2076 |
2038-2067 2070-2099 |
2037-2066 2066-2095 |
| ssp585 | GWL2K GWL3K |
2019-2048 2040-2069 |
- | - |
Please note that in NUKLEUS, there are only simulations with EC-EARTH using the scenarios ssp245 and ssp585. It should also be noted that the historical reference period (1961-1990) already shows a global warming signal of 0.42 degrees compared to the preindustrial period (1881-1910) according to the HadCRUT5 dataset.[2]
Available Data¶
Below you can find an overview of the NUKLEUS simulations and the used scenarios for the two different domains EUR-11 and CEU-3. Simulations/scenarios marked in green are already available via the Freva Data-Browser. The simulations/scenarios marked in orange are completed and the post-precessing to include them in the Freva Data-Browser is ongoing. In case the simulations/scenarios are marked in red, the simulations are still ongoing and, therefore, data is currently not available.
Simulation Domain EUR-11¶
| RCM | GCM |
|||
|---|---|---|---|---|
| ERA5 | EC-EARTH3-Veg | MIROC6 | MPI-ESM1-2-HR | |
| ICON-CLM | Evaluation |
Historical |
Historical |
Historical |
| CCLM | Evaluation |
Historical |
Historical |
Historical |
| REMO | Evaluation |
Historical |
Historical |
Historical |
Simulation Domain CEU-3¶
| RCM | GCM |
|||
|---|---|---|---|---|
| ERA5 | EC-EARTH3-Veg | MIROC6 | MPI-ESM1-2-HR | |
| ICON-CLM | Evaluation |
Historical |
Historical |
Historical |
| CCLM | Evaluation |
Historical |
Historical |
Historical |
| REMO | Evaluation |
Historical |
Historical |
Historical |
Available Variables¶
Below is a selection of commonly used near-surface variables and the temporal resolution(s) at which they are available. In total, the output of the NUKLEUS simulations consists of over 100 different variables also at different heights or pressure levels. For a complete list, you can use the Freva Data-Browser. Please note that all below mentioned variables and temporal resolutions are available on the rotated grid.
| Description | Name | 1 hr | 3 hr | 6 hr | day | mon |
|---|---|---|---|---|---|---|
| Potential evapotranspiration | evspsblpot | X | X | |||
| Surface upward latent heat flux | hfls | X | X | X | ||
| Surface upward sensible heat flux | hfss | X | X | X | ||
| Relative humidity at 2m | hurs | X | X | X | ||
| Specific humidity at 2m | huss | X | X | X | ||
| Precipitation | pr | X | X | X | ||
| Hourly maximum precipitation rate | prhmax | X | X | |||
| Surface air pressure | ps | X | X | X | ||
| Sea level pressure | psl | X | X | X | ||
| Surface down-welling long-wave radiation | rlds | X | X | X | ||
| Surface up-welling long-wave radiation | rlus | X | X | X | ||
| Surface down-welling short-wave radiation | rsds | X | X | X | ||
| Surface up-welling short-wave radiation | rsus | X | X | X | ||
| Wind speed at 10m | sfcwind | X | X | X | ||
| Wind direction at 10m | sfcwinddir | X | X | X | ||
| Maximum wind speed at 10m | sfcwindmax | X | X | |||
| Surface snow amount | snw | X | X | X | ||
| Temperature at 2m | tas | X | X | X | ||
| Maximum temperature at 2m | tasmax | X | X | |||
| Minimum temperature at 2m | tasmin | X | X | |||
| Surface temperature | ts | X | X | X | ||
| Eastward wind component at 10m | uas | X | X | X | ||
| Northward wind component at 10m | vas | X | X | X | ||
| Maximum wind gust at 10m | wsgsmax | X | X | X |
1 hr = hourly data; 3 hr = 3-hourly data; 6 hr = 6-hourly data; day = daily data; mon = monthly data
Bias Correction¶
Small excerpts of the NUKLEUS simulations will undergo a bias-correction. For this limited sub-sample of the NUKLEUS ensemble, the data is converted to a regular latitude-longitude grid. These bias-corrected data can be distinguished by the “i” suffix within the product name. For example, CEU-3i would be bias-corrected data of the cordex data across Germany at 3 km. The bias-corrected variables get the suffix “Adjust”. For example, the bias-corrected version of the precipitation pr would be named prAdjust.
Available Data¶
It is planned to make the bias-corrected data available in the first half of 2025. The following table gives an overview of the current status. Again, simulations/scenarios marked in green are already available via the Freva Data-Browser, while simulations/scenarios marked in orange are completed but the post-precessing to include them in the Freva Data-Browser is still ongoing. In case the simulations/scenarios are marked in red, the data is currently not (yet) available.
Simulation Domain CEU-3i¶
| RCM | GCM |
||
|---|---|---|---|
| EC-EARTH3-Veg | MIROC6 | MPI-ESM1-2-HR | |
| ICON-CLM | Historical |
Historical |
Historical |
| CCLM | Historical |
Historical |
Historical |
| REMO | Historical |
Historical |
Historical |
Available Bias-Corrected Variables¶
The following table gives an overview on the variables that underwent a bias correction. As described above, daily observational data was used for the bias correction. Therefore, all variables are available at daily resolution only. Please note again, that this data is on the regular latitude-longitude grid.
| Description | Name | day |
|---|---|---|
| Bias-corrected precipitation | prAdjust | X |
| Bias-corrected daily mean 2m temperature | tasAdjust | X |
| Bias-corrected daily minimum 2m temperature | tasminAdjust | X |
| Bias-corrected daily maximum 2m temperature | tasmaxAdjust | X |
References
[1] R. Vautard et al. (2014): The European climate under a 2 °C global warming.
Link to Publication
[2] C.P. Maurice et al. (2021): An updated assessment of near-surface temperature
change from 1850: the HadCRUT5 dataset. Link to Publication
[3] Canon et al. (2025): Bias Correction of GCM Precipitation by Quantile Mapping: How
Well Do Methods Preserve Changes in Quantiles and Extremes? Link to Publication
Last modified: