last modified
Feb 14, 2018 01:12
Data analysis of CMIP6 simulations using the Earth System Model Evaluation Tool (ESMValTool)
With the aim to improve the complex evaluation of earth system models with observational data and monitor the running climate model simulations, the DLR is developing the ESMValTool together with other international partners. The tool is installed on the ESGF data note at the DKRZ which shall serve as example for integrations into environments of other institutions worldwide. For CMIP6 the model evaluation results will be provided in a web-based portal (Freva) developed at the FUB.
In order to improve the complex evaluation of earth system models with observational data in the long term and to build up an evaluation standard, the DLR is developing the ESMValTool together with other international partners (the Earth System Model Evaluation Tool, https://www.esmvaltool.org/). The ESMValTool is a diagnostic tool for routinely checking the quality of earth system models using observational data. The representation of climatological mean values, trends and variability of basic climatic variables in the models is tested on different spatial and temporal scales and additionally a processor-oriented evaluation is carried out. With the ESMValTool, CMIP models can be effectively and extensively evaluated for their representation of various important climatic variables and processes (see Figure 1, for example).
In subproject 2, the ESMValTool is installed on the ESGF data node at the DKRZ. Compared to other CMIP models, this allows MPI-ESM1 / 2 and EMAC2 to provide an extensive range of diagnostics for comprehensive characterization at an early stage. The tool is expanded with additional evaluation diagnostics and observational data. The integration of the ESMValTool at the DKRZ is intended as an example for integrations into the environments of other institutions worldwide and makes a substantial contribution to CMIP6 +. The results of the CMIP6 model evaluation are provided in a web-based portal (see subproject 4).
In addition, the ESMValTool is further developed in such a way that it can be used for scientific quality control during the running of model simulations at the DKRZ. For this purpose, quicklooks of important climate variables are created and visualized for monitoring on an internal website. These Quicklooks can be used by the relevant scientists in order to recognize possible quality problems in the data at an early stage.
In addition, the ESMValTool is further developed in such a way that it can be used for scientific quality control during the running of model simulations at the DKRZ. For this purpose, quicklooks of important climate variables are created and visualized for monitoring on an internal website. These Quicklooks can be used by the relevant scientists in order to recognize possible quality problems in the data at an early stage.
Figure 1: Relative mean square deviation of the CMIP5 model performance for selected climatic parameters, based on a global average climatology of the year (1980-2005) and calculated with the ESMValTool.
