Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

The expected climate change within the next decades is predicted to be caused by an increase of the emissions of anthropogenic CO2 and other greenhouse gases. These processes will have a huge influence on the future climate and consequently on the society. Although climate change will be noticeable on the global scale, the effects will be much more diverse and extreme regarding the regional scale.

General circulation models (GCMs) currently are the most advanced tools for simulating the response of the global climate system to increasing greenhouse gas concentrations. They are able to numerically solve the equations of physics hence, GCMs have the ability of representing physical processes of the atmosphere, ocean, cryosphere and the land surface using a 3-D grid over the globe with a typical resolution. Nowadays, typically GCMs use a horizontal resolution of 100–200 km. However, to better understand also regional climate phenomena such as local extremes and in order to assess the effect of the expected climate change, scientists and end users like federal agencies and climate impact and adaptation researchers require projections on the regional scale with a higher horizontal resolution.

The Coupled Model Intercomparison Project Phase 5 (CMIP5) [1] provides a framework for coordinated climate change experiments, where 20 different modelling groups carried out global climate projections with their GCMs, contributing to the latest assessment report of the Intergovernmental Panel on climate change (IPCC). Those projections are based on Representative Concentration Pathways (RCPs) [2], representing four different possible greenhouse gas (GHG) concentration scenarios of the future climate. Namely, these scenarios are the RCP8.5, RCP6, RCP4.5 and RCP2.6 scenario, indicating the possible range in the change of radiative forcing (in W/m2) by the year 2100 relative to pre-industrial values.

In 2009 the COordinated Regional climate Downscaling EXperiment (CORDEX, http://wcrp-cordex.ipsl.jussieu.fr) was established by the World Climate Research Programme (WCRP), in order to provide ensembles of regional climate simulations on a higher spatial resolution. The task within CORDEX is to apply the GCMs which contributed to the CMIP5 database for the boundary forcing of different regional climate models (RCMs). CORDEX is separated into different sub-groups, covering all continental regions of the globe, serving as input for climate change and adaptation studies. EURO-CORDEX, the european branch of the CORDEX initiative, is focusing on grid sizes of 0.11\(^{\circ }\) (\(\sim \)12 km), but simulations using a resolution of 0.44\(^{\circ }\) (\(\sim \)50 km) are carried out simultaneously.

The ReKliEs-De project funded by the Federal Ministry of Education and Research (BMBF), contributes to EURO-CORDEX by carrying out a certain number of regional climate projections. Several institutes in Germany are part of ReKliEs-De applying different downscaling techniques (dynamical and statistical downscaling), by forcing their RCMs with CMIP5 data. The goal of the ReKliEs-De project (http://www.reklies.hlnug.de/) is to provide robust informations about the future evolution of the climate in Germany based on regional climate projections using a high spatial resolution of 0.11\(^{\circ }\). Beside scientifically analysing the results of the simulations, the focus is also laid on producing user friendly data by providing special climate indices which can be applied easily by the end users. Today, the dynamical downscaling of GCM data is a well respected standard technique to produce high resolution climate projections.

An ensemble of RCMs participating in EURO-CORDEX, forced with ERA-Interim reanalyses data [3] were evaluated e.g. by [4]. Although model biases, predominantly cold and wet biases, emerged in the majority of the experiments, the authors documented the general ability of the RCMs to represent the basic spatiotemporal patterns for the european climate.

The scope of this part of the WRFCLIM project at HLRS (see also [5]), is to produce high resolution regional climate projections in the framework of ReKliEs-De on 0.11\(^{\circ }\). We will carry out this long term simulations by dynamically downscaling different GCM model output from 1958 to 2100 with the Weather Research and Forecasting (WRF) [6] Model.

In the following chapter a technical description to the simulations within WRFCLIM at HLRS is presented. A summary of the current analysis and some preliminary results of the simulations which were carried out from April 2015 to April 2016 are reported as well.

2 Simulation Setup with WRF and Forcing Data

The climate simulations are carried out on the CRAY XC40 System at the HLRS. The downscaling process of the coarse GCM output data to the required 0.44\(^{\circ }\) and 0.11\(^{\circ }\) grid resolution is performed by the WRF model version 3.6.1 with its coupled land surface model NOAH [7]. WRF uses the dynamical technique by numerically solving the governing equations of the atmosphere on a finer grid. The simulations are forced 6 hourly at the lateral boundaries with the sea surface temperature and the 3-D fields of temperature, horizontal winds and moisture from the forcing GCMs. We apply a one-way nesting approach via 0.44\(^{\circ }\) to 0.11\(^{\circ }\). WRF was compiled at HLRS with PGI 14.7 and applied in a hybrid configuration using MPI and OpenMP to optimize the speed of the simulation. The model configuration was set to the same parameterizations like it was applied in [8]. Table 1 shows some technical details of the planned and the so far partly finalized simulations performed on hazelhen.

The WRF projections presented in this chapter are carried out within the frame of EURO-CORDEX to create climate projections. The simulation domains which were specified within CORDEX are displayed in Fig. 1 (left). The analysis focuses on Germany and its contributing river catchments area as shown in Fig. 1 (right). For the future climate projections, four different GCMs and two different RCP scenarios of the CMIP5 project, are applied as boundary forcing with the WRF model. The historical runs of the GCMs cover the period from 1850 to 2005. This period is forced by observed atmospheric composition changes of anthropogenic and natural sources. The RCP scenarios of the GCMs cover the period from 2006 to 2100. They represent mitigation scenarios that assume policy actions will be taken into account to achieve certain emission targets [1]. The numbers of the RCPs give a rough estimate of the range in the change of the radiative forcing by the year 2100 relative to the preindustrial values. The forcing data we applied, the resolution of the GCMs, its scenarios and the chosen simulation period is presented in Table 2.

Table 1 Technical details of WRF simulations. Note: the raw model output size is minimized after simulation in the postprocessing process as the amount of the results cannot be stored in its raw output
Full size table
Fig. 1
figure 1

Simulation domain as specified by CORDEX (a), and focal domain within ReKliEs-De with Germany (red) and the areas of its main river catchments (coloured) (b)

Full size image
Table 2 Applied GCM data for the WRF simulations
Full size table

3 Results

Based on the WRF downscaling simulations to the 50  km grid, first results of the different forcing GCMs are presented (see GCM description in Table 2). Simulations were carried out for the EURO CORDEX domain with lateral boundaries from approximately 25\(^{\circ }\)N to 75\(^{\circ }\)N and 30\(^{\circ }\)W to 50\(^{\circ }\)E. Since our model simulations are still ongoing by this time (May 2016) exemplary results of the 2 m temperature will be presented. In the first part the focus is based on a time period of the historical simulations from 1971–2000 represented by the four forcig GCMs which were applied in our study. The second part shows a time period of the projected simulations for the two future scenarios RCP2.6 and RCP8.5 represented by the MPI-ESM-LR model for the years 2071–2099. The average near surface temperature is presented in the following based on the annual and on the seasonal scale for the Northern Hemisphere (NH) summer and winter. It is emphasized that the results of this study are based on just one regional model and will need further analyses within the EURO-CORDEX ensemble and the ReKliEs-De project.

Fig. 2
figure 2

Average near surface temperature on the 50 km resolution simulations from WRF. Boundary forcing with the different GCMs: EC-EARTH (a), HadGEM2-ESM (b), MIROC5 (c) and MPI-ESM-LR (d)

Full size image
Fig. 3
figure 3

Average near surface temperature on the 50 km resolution simulations from WRF during winter. Boundary forcing with the different GCMs: EC-EARTH (a), HadGEM2-ESM (b), MIROC5 (c) and MPI-ESM-LR (d)

Full size image

3.1 Historical Temperatures

The simulated annual average near surface temperature from 1971–2000 is given in Fig. 2. The four simulations show a common agreement of the average temperature pattern for Europe, which indicates a generally good performance of the models. The temperature ranges from around 264 K (\(-9\,^{\circ }\)C) in the North to 294 K (21 \(^{\circ }\)C) in the Southern part of the simulation domain. However, some differences in the temperature range is apparent among the GCMs. Compared to HadGEM2-ESM Fig. 2b and MPI-ESM-LR Fig. 2d, lower temperatures are simulated by MIROC5 Fig. 2c and EC-EARTH Fig. 2a in the Northern and in the Eastern part of the CORDEX domain. Most of the GCMs show a cold bias. MIROC5 indicated to have the strongest bias compared to the other three which are presented here.

Fig. 4
figure 4

Average near surface temperature on the 50 km resolution simulations from WRF during summer. Boundary forcing with the different GCMs: EC-EARTH (a), HadGEM2-ESM (b), MIROC5 (c) and MPI-ESM-LR (d)

Full size image

The average near surface temperature for the time period 1971–2000 as simulated by WRF with the four different forcing GCMs is given for the winter season (DJF) in Fig. 3 and for the summer season (JJA) in Fig. 4. The average temperatures during NH winter vary from around 258 K (\(-15\,^{\circ }\)C) in North-East to around 290 K (17 \(^{\circ }\)C) over the atlantic ocean in the south-west of the domain. During NH summer the lowest temperatures occur over Iceland in the north-east and over the arctic ocean with 276 K (3 \(^{\circ }\)C). Highest temperatures in Europe aside from the Sahara desert are around 296 K (23 \(^{\circ }\)C) over the mediteranean and the greek islands. Similar as for the annual average, EC-EARTH and MIROC5 indicate lower temperatures (Figs. 3a–c and 4a–c) in both presented seasons compared to the two GCMs HadGEM2-ESM and MPI-ESM-LR (Figs. 3b–d and 4b–d).

Fig. 5
figure 5

Differences of the average near surface temperature on the 50 km resolution simulations from WRF. Boundary forcing with MPI-ESM-LR for 1971–2100 RCP 26 (a) and RCP8.5 (b)

Full size image

3.2 Projections for Temperature Changes

The future changes, based on the differences of the simulated average temperatures from 1971–2000 and 2071–2099, is shown in Fig. 5. Following the RCP2.6 scenario, the highest temperature increase reaches 2–2.4 \(^{\circ }\)C above the Norwegian Sea and the Russian peninsula Kola. Following the RCP8.5 scenario, highest temperature changes are simulated above the Russian peninsula with an increase of around 5 \(^{\circ }\)C.

The seasonal changes are displayed in Fig. 6 for the winter and for the summer averages. During winter the RCP2.6 scenario indicates a maximum increase of the temperatures over the peninsula of Russia with up to 4.5 \(^{\circ }\)C (Fig. 6a). For the RCP8.5 scenario the temperature difference is even higher, revealing an increase of up to 6 \(^{\circ }\)C (Fig. 6b). Beside Russia, also the eastern part of Turkey will experience as simulated for this high emission scenario a high temperature increase of around 5 \(^{\circ }\)C compared to the historical period.

Fig. 6
figure 6

Differences of the average near surface temperature on the 50 km resolution simulations from WRF. Boundary forcing with MPI-ESM-LR for 1971–2100 for the winter season: RCP 26 (a) and RCP8.5 (b) and for the summer season RCP 26 (c) and RCP8.5 (d)

Full size image

Compared to the temperature changes in winter, the increase of the average summer temperature is less intense for both scenarios. The increase in the temperature will reach around 1.5 \(^{\circ }\)C in central Europe and in the southern part of the norwegian sea for the RCP2.6 scenario (Fig. 6a). For the RCP8.5 scenario the highest temperature increases are simulated over south and south eastern Europe and northern africa indicating a temperature increase of around 5 \(^{\circ }\)C in Spain and up to 6 \(^{\circ }\)C in Turkey.

4 Conclusion

The results shown in this chapter highlighted preliminary results based on the near surface temperature of the downscaling of four GCMs (EC-EARTH, HadGEM2-ESM, MIROC5 and MPI-ESM-LR) from the original model grid of around 150 km with WRF to a refined grid on 50 km resolution for the EURO-CORDEX domain. Further downscaling simulations to the 12 km grid for the historical and future time periods are still ongoing. It is expected to finalize the simulations as described in Table 1 by the end of this year (2016) accompanied with further improvements of the projection skills due to a better representation of the orographic effects.

The short insight of the seasonal differences which was shown in this chapter, pointed out the need of further analyses. In particular different variables and climate indices need to be investigated in more detail as the information about future changes of the climate is a major need for cultivators world wide. Precipitation is a parameter which needs to be pointed out and analysed carefully as it showed to be highly variable throughout the year in germany and also Europe.

We would like to point out the importance of analysing a model ensemble rather than choosing a single RCM when predictions of the evolution of climate parameters are being done. To achieve robust predictions, ReKliEs-De is an appropriate platform. Within this project a huge ensemble member of climate simulations can be realized with the contribution of several RCMs applying different techniques and GCMs.