Release of EC-EARTH ver 3.0 beta

A new version of the earth system model EC-EARTH has recently been released by the Rossby Centre to the EC-EARTH community. The new version 3.0beta includes upgraded components, namely the ECMWF atmospheric model (IFS) cycle 36r1 and version 3.2 of the Nemo/LIM ocean and sea-ice models.

The coupling between atmosphere and ocean models has been revisited and the number of coupled fields has been halved. IFS has shown major improvements since the version that is part of EC-EARTH version 2.1. Results from ECMWF with the changes in the convection and radiation schemes of cycle 36r1 raise expectations for a better representation of the atmosphere and its variability. The first tests with the new EC-EARTH version look promising.

Improvements in the annual precipitation

As an example we show the improvements in the annual precipitation. For this experiment, the ocean has been switched off and the atmosphere is forced with observed SST and sea-ice. Model results are compared against the precipitation estimate from the GPCP dataset that is a combination of satellite retrieval and rain-gauge observations.

RCN2010, Precip, EC_EARTH, v3
Figure 1. Precipitation in EC_EARTH version 3.
Enlarge Image
RCN2010, Precip, EC_EARTH, GPCP
Figure 2. Precipitation in the GPCP dataset.
Enlarge Image
RCN2010, Precip, EC_EARTH, v2
Figure 3. Precipitation in EC_EARTH version 2.1.
Enlarge Image

The precipitation pattern over the Pacific Ocean has clearly improved with Version 3. The narrow structure of the ITCZ over the equator is captured better than with the previous version of EC-EARTH although the area with highest precipitation remains too large. The amount of precipitation over the Amazon basin is more realistic with Version 3. Over the Kongo basin, the new release has improved the results but it is still not precipitating as much as the observations suggest.

The development continues

The development of EC-EARTH Version 3 continues. Some work remains with the technical setup for long simulations and the inclusion of time-varying forcing fields for the planned CMIP5 simulations (see article EC-EARTH and CMIP5in the newsletter). Further testing of the new model system is needed and the balance of mass and energy in long-term integrations needs to be established.