ICON in Climate Limited-area Mode (ICON Release Version 2.6.1): a new regional climate model

For the first time the limited-area mode of the new weather and climate model ICON has been used for a continuous long-term regional climate simulation over Europe. Building upon ICON-LAM, ICON-CLM (ICON in Climate Limitedarea Mode, hereafter ICON-CLM, available in ICON Release Version 2.6.1) is an adaptation for climate applications. A first version of ICON-CLM is now available and has already been integrated into a starter package (ICON-CLM_SP Version Beta1). The starter package provides users with a technical infrastructure that facilitates long-term simulations as well as 5 model evaluation and test routines. ICON-CLM and ICON-CLM_SP were successfully installed and tested on two different computing systems. Test with different domain decompositions showed bit-identical results, and no systematic outstanding differences were found in the results with different model time steps. Comparison was done between ICON-CLM and COSMOCLM (the recommended model configuration by the CLM-Community) performance. For that, an evaluation run with ERAInterim boundary conditions was carried out with the setups similar to the COSMO-CLM recommended optimal setups. ICON10 CLM results showed biases in the same range as those of COSMO-CLM for all evaluated surface variables. This is remarkable since the COSMO-CLM simulation was carried out with the latest model version which has been developed for two decades and was carefully tuned for climate simulations on the European domain. Furthermore, ICON-CLM already showed a better performance for air temperature, its daily extremes, and total cloud cover. Results for precipitation and mean sea level pressure did not show clear advantage from any model. However, as ICON-CLM is still in the early stage of development, there is still 15 much room for improvement. 1 Background information In 1999, the limited-area weather forecast model LM (Lokalmodell, Doms and Schättler (1999), later COSMO, Baldauf et al. (2011)), which was developed by the Deutscher Wetterdienst (DWD, the German Meteorological Service), went operational together with the global model GME (Majewski and Ritter, 2002). A few years later, it was renamed into "COSMO model" in 20 order to reflect that further development has become a joint task of the COnsortium for Small scale MOdelling (COSMO). In 1 https://doi.org/10.5194/gmd-2020-20 Preprint. Discussion started: 7 July 2020 c © Author(s) 2020. CC BY 4.0 License.

2. Section 3 gives details of the ICON-CLM model configuration and setup for the evaluation run as well as the evaluation methods we used. Results of this evaluation run in comparison to observational data and to the results of the latest COSMO-CLM version are shown in Section 4. Conclusions are provided in Section 5.

Model development
As the Limited-Area Mode of ICON, which ICON-CLM builds upon, has originally been developed for NWP applications, 5 several adaptations and technical extensions were necessary to prepare the model for climate applications. Apart from the adjustments in the code, long-term climate simulations require a technical infrastructure for data and job management. Such an infrastructure has also been developed based on the existing infrastructure of COSMO-CLM.

The regional climate model ICON-CLM
Weather forecasting, which predicts the state of the atmosphere only up to about 2 weeks in advance, often does not involve 10 the development of the ocean state. The ocean surface condition, hence, is often kept constant during the forecasts in weather prediction models or just slightly adjusted with a climatological trend for the forecast period. Thus, the sea surface temperature (SST) and sea-ice cover in ICON-LAM can only be updated monthly. For ICON-CLM, it is necessary to have a more frequent (up to hourly) update of SST and sea ice from external data. For this purpose, an option for higher update frequencies of these boundary conditions was implemented in ICON-CLM. Time-dependent SST and sea-ice data can now be read from external 15 data files and are fed to ICON-CLM with an user-defined interval (e.g. 1 hourly or 6 hourly). The user can select this option of frequent update of SST and sea ice via namelist settings. The external SST and sea ice data must be prepared and remapped to the ICON grid.
Similarly, the green house gas (GHG) values are usually kept constant in weather forecast models, because the changes during the forecast period are negligible. In climate projections, however, it is necessary to use the time-dependent GHGs from 20 climate change scenarios. Such an option was already available in ICON, but only in combination with the ECHAM physics package. The corresponding read routine was therefore extended so that it works for the NWP physics as well. Some additions to the NWP radiation scheme were made with respect to the GHG vertical profile with a new option to get the profile from external gas data. A file that contains yearly values of CO 2 , CH 4 , N 2 O and Chlorofluorocarbons (CFC) for all years of the experiment needs to be prepared in advance. These features of the time-dependent SST and GHG were largely based on the 25 corresponding implementations in the ICON-A (Giorgetta et al., 2018).
For the upper boundary, ICON-LAM offers an option of prescribing the upper boundary conditions by using the same driving data source as for the lateral boundary conditions (nudging option). Users can define the height of the nudging zone as well as the nudging coefficients for the horizontal wind and for the thermodynamic variables via namelist settings. If this vertical nudging option is turned off, a Rayleigh damping is applied to the vertical wind speed within the damping layer in order to 30 prevent unphysical reflection of vertically propagating gravity waves.
In the NWP configuration of ICON-LAM, the number of soil layers is always constant with eight layers. The depths of half soil layers are also fixed at values between 5 mm and 14.5 m. However, for climate simulations in domains other than Europe (e.g. Africa, Asia) or to achieve better simulation of the soil variables for European domain, it is usually reasonable to adjust these soil parameters. Therefore, an option for a flexible number and depth of the soil layers has been implemented in the ICON-CLM code.

5
The input/output of ICON-CLM has also been adjusted to have more flexibility. In NWP mode, the precipitation data are accumulated from the start of the forecast till the end without any reset. This is suitable for short weather forecasts, but for long climate simulations, this procedure is inconvenient and could, in the worst case, cause problems due to data imprecision. Furthermore, the maximum and minimum 2-m temperature values are calculated for 6 hourly intervals in NWP applications, while for climate simulations the standard for these output variables is usually 24 hours. To control this flexibility extensions, 10 new namelist parameters were introduced in ICON-CLM.
At the lateral boundaries, ICON-LAM requires, by default, information on cloud liquid water content and cloud ice water content from the global forcing data. These input fields are usually available if the ICON-CLM lateral boundary conditions are taken from reanalysis data like ERA-Interim. But if global climate projections are used as lateral boundary conditions, these fields are usually not provided. Thus, the model code has been adjusted so that if cloud liquid water content and cloud ice 15 content are not available in the lateral boundary data, these variables are initialized with zero. Table 1 provides an overview of some differences between COSMO-CLM and ICON-CLM.

The starter package ICON-CLM_SP
In order to facilitate long-term climate simulations, we developed a run time infrastructure called starter package and a separate evaluation tool. Both are provided along with the ICON-CLM model source code. The starter package ICON-CLM_SP contains 20 a run routine, a climatological testsuite, all necessary utilities and configure scripts for different super computing environments.
At the moment, two system settings for Cray (DWD) and Atos/Bull (DKRZ) are supported and tested. Settings for other machines could be easily added if necessary.
The run routine in ICON-CLM_SP, called "subchain", was adapted from the routine of the existing COSMO-CLM package.
The "subchain" contains five sub-routines for input preparation (prep), converting input data (conv2icon), ICON-CLM job 25 management (icon), archiving (arch) and postprocessing (post) of the model output. Sub-routine "prep" copies and checks all the global forcing data as input for "conv2icon". Then "conv2icon" preprocesses and interpolates the initial data and the lateral, lower and upper boundary data onto the ICON-CLM model grid for the current model simulation. Sub-routine "icon" does the job management for ICON-CLM model. After that, all model output data are compressed by "arch" and some post-processing steps like the provision of time series of selected output variables are done in "post". 30 A climatological testsuite (CTS) was also created based on the CTS from COSMO-CLM. In the CTS, 5-year test simulations can be done automatically with "subchain". The users can choose one simulation as a reference. The test simulations then will be compared with the reference simulation with respect to observational data (E-OBS and CRU, see Section 3.2 for more details) by an extra sub-routine called "eval". At the end, the results are visualized with standardized plots. This CTS was built https://doi.org/10.5194/gmd-2020-20 Preprint. Discussion started: 7 July 2020 c Author(s) 2020. CC BY 4.0 License. for the purpose of testing different versions of model source codes, or different setups of the same model version. Hence, it is a very helpful tool for model development and tuning.
Besides the sub-routine "eval" in CTS, a separate evaluation tool called "ETOOLS" was also adapted from the COSMO-CLM evaluation tool. This tool provides comparisons of the simulation results with observation data sets and creates standardized plots to visualize the results. In order to facilitate the transition from COSMO-CLM to ICON-CLM for the users, both ICON-5 CLM_SP and ETOOLS were created such that the "look and feel" as well as the usage of the software packages is as similar as possible to the corresponding packages that exist for the COSMO-CLM model. The output structure of ICON-CLM or postprocessed time series from "subchain/post" are also similar to those of COSMO-CLM for the same reason. Furthermore, users should be able to use all existing scripts and programs that were developed for COSMO-CLM output also for ICON-CLM data. The model atmosphere was initialized with ERA-Interim data. The soil temperature and soil moisture were taken from a previous test simulation which is long enough to ensure that the spin-up of the soil has been completed. For the upper boundary, as described in Section 2, there are two options: (1) Using the driving data and nudging gradually in the relaxation zone; (2) Damping the vertical wind beneath the upper boundary. To assess the impact of these two options, two 10-year simulations 25 (1979)(1980)(1981)(1982)(1983)(1984)(1985)(1986)(1987)(1988) were done with the same setups, with and without global data nudging. Analysis from these 10-year runs resulted in very minor differences on surface variables and none of the options showed any advantage over the other. For the evaluation run, we chose the option with nudging data at the upper boundary with ERA-Interim data, as later we wanted to compare the results with those from a COSMO-CLM run using a similar nudging option. The nudging zone started from the height of 12 km to the model top of atmosphere (30 km). 30 In order to find a suitable model configuration for ICON-CLM at the resolution R2B8, an optimized namelist setup was used, namely the setup from ICON-LAM for R3B7 with nested domain on R3B8 grid (approximately 13 km and 6.5 km respectively). The tuning parameters were taken over from the global settings. In this setting, the Tiedke/Bechtold (Bechtold et al., 2008) convection parameterization scheme and the Rapid Radiation Transfer Model (RRTM) radiation scheme (Mlawer et al., 1997) were used. These setups were checked to make sure that they are appropriate for climate applications and were used in all simulations. Former simulations (not published) with COSMO-CLM showed that in some cases the model results depend on the chosen model time steps. There is one particular time step that leads to larger biases in precipitation and surface pressure, especially over the Alps and the south western area of the model domain. This issue has been analyzed by the CLM-Community 5 but is still not fully understood yet. To ensure that such a dependency of the results on the model time step is not present in ICON-CLM, different fast physics/advection time step (hereafter: time step) choices were tested. At R2B8 (approximately 10 km) resolution, the time step should not exceed 120 seconds for stability reasons. With the common model and experiment setups described above, we carried out multiple one-year simulations for the year 1979 with time steps of 60, 80, 90, 100 and 120 seconds. Figure 3 shows the biases compared to reference data of 2-m temperature, mean sea level pressure (MSLP), total 10 precipitation and total cloud cover. The biases were averaged for each month and for the Alpine region (sub-region denoted AL in Figure

Evaluation methods
For model assessment and evaluation, output fields from six variables were analyzed, namely 2-m temperature, daily maximum and minimum values of 2-m temperature, MSLP, total precipitation and total cloud cover. Monthly average values of these variables were calculated and used for further analysis. For total precipitation, the monthly accumulated amounts were calculated. The evaluation within COPAT for 2-m temperature, MSLP and cloud cover was done using E-OBS version 10.0 and CRU version 3.22 as reference data. Therefore, in order to compare our evaluation to the one from COPAT, we also used the same versions of the data sets. The comparison period is 20 years from 1981 to 2000, the same as the evaluation period in COPAT. 15 The reference total precipitation data was taken from E-OBS version 12.0 because this dataset (among versions from 10.0 to 17.0) shows the fewest missing data for precipitation over the area of Poland. Table 2) were calculated from ICON-CLM, COSMO-CLM and E-OBS 2-m temperature and total precipitation data for the entire period 1981-2000. The number of days that fulfils the definition (in Table   2) was counted for each horizontal grid cell, then averaged over a sub-region.

20
Root-mean-square error (RMSE) was calculated from the model and observed monthly values:

Precipitation
The mean annual precipitation bias ranged from -50 mm/month to 50 mm/month in both models ( Figure 4). Overall, both models simulated more precipitation than E-OBS data. However, one should keep in mind that E-OBS precipitation data tend the spatial distribution of precipitation biases was similar in both models.

25
Looking at the spatial variability of the seasonal biases within the sub-regions in Figure 8, we see that although for some sub-regions in certain seasons, the bias medians were close to zero, the ranges of biases were large. This is expected because precipitation is a highly inhomogeneous variable. Summer and autumn tended to have small median bias in ICLM-REF and RMSEs and STDEVs of cloud cover did not show much difference between the two models. No concrete conclusion could be drawn from those numbers, therefore they were not shown here.

25
The new regional climate model ICON-CLM has been derived from the weather forecast model ICON-LAM along with the necessary technical infrastructure allowing users to carry out long-term regional climate simulations. An evaluation run from climate applications for more than 20 years and was well-tuned with a large number of tested namelist combinations in the COPAT project. Therefore, ICON-CLM has still great potential to improve the model setup and with growing experience we expect that ICON-CLM results will improve further in the upcoming years. 15 The next step in the ICON-CLM preparation will be a thorough model tuning by testing the sensitivity of the model to a variety of namelist parameters and their different combinations of namelist settings in order to find an optimal configuration.
Climate simulations on different domains, e.g. CORDEX Africa, will be done to evaluate the ability of ICON-CLM to simulate different climates. So far only re-analysis driven simulation have been performed with ICON-CLM, but historical simulations driven by the results of global climate simulations will also be done to test the model performance for this experiment type.

20
Based on a well-evaluated model configuration, regional climate projections will be performed in order to address the impact of climate change at regional scale. Thus, climate projections, e.g. in the framework of CORDEX, will also be provided with ICON-CLM in the future. We plan to further develop ICON with the aim of unifying the different physics packages currently existing for the numerical weather prediction and the climate mode in order to pursue a "Seamless prediction" system with one forecasting system which can produce produce forecasts and projections for all time-scales from weather prediction to seasonal 25 and decadal predictions and climate projections. 19 https://doi.org/10.5194/gmd-2020-20 Preprint. Discussion started: 7 July 2020 c Author(s) 2020. CC BY 4.0 License.