Converting the output from the Met Office CSSP Brazil convecting-permitting model (CPRCM) output into a single-variable netCDF files consistent with CMOR 3.7, CF-1.7 CMIP-6.2 and CF standards.
The Brazil CPRCM experiment includes three simulations performed with the Met Office regional environment model (MOHC-REM3), using the regional atmosphere-land tropical configuration at 4.5 km grid spacing (MOHC-HadREM3-RAL1T-4.5km).
The Hindcast simulation (hindcast) is ran for 1998-2008. The boundary conditions are provided by the ERA-interim dataset, dynamically downscaled to the CPRCM 4.5km resolution using two-step nesting. ERA-Interim data is first downscaled to 25 km with MOHC-HadREM3-GA705-25km and then into 4.5km with HadREM3-RAL1T-4.5km.
The present-day control simulation (pdControl) covers the same period and the boundary conditions are provided by the HadGEM3 atmosphere-land model at 25km (MOHC-HadGEM3-GC31-N512). Sea surface temperatures (SSTs) and sea ice are from the present-day daily 0.25 deg. dataset (Reynolds et al., 2007).
The Future simulation (future2100) simulates 10 years under 2100 RCP8.5 forcing driven by the MOHC-HadGEM3-GC31-N512 model. The future forcing includes GHG concentrations for the year 2100 from the RCP8.5 scenario (Moss et al., 2010) and sea surface temperatures (SST) anomalies between the 1975-2005 and 2085-2115 from the MOHC-HadGEM2-ES CMIP5 RCP8.5 simulation, added to the present-days (Reynolds) SSTs.
We have used a Python-based tool for converting the UM output into CMOR standards. This tool has been developed by the Climate Data Dissemination System (CDDS) team at the Met Office (https://github.com/MetOffice/CDDS) and was extensively used to convert the Met Office UM data for CMIP6.
Each of the three simulations was split into two segments of 5 years (+1 year spinup), which were run in parallel to save computing time. The spinup yr for segment one is 2007 and the simulation period is 1998-2002. The spinup yr for segment two is 2002 and the simulation period is 2003-2008.
| Experiment | Calendar | UM model | suite ID (Segment I) | suite ID: Segment II | Driving model | Driving suite |
|---|---|---|---|---|---|---|
| hindcast | Gregorian | MOHC-HadREM3-RAL1T-4.5km | mi-ba751 | mi-bb055 | MOHC-HadREM3-GA705-25km | mi-ba898 |
| pcControl | 360-day | MOHC-HadREM3-RAL1T-4.5km | mi-ba488 | mi-bb069 | MOHC-HadGEM3-GC31-N512 | u-ab261 |
| future2100 | 360-day | MOHC-HadREM3-RAL1T-4.5km | mi-ba489 | mi-bb070 | MOHC-HadGEM3-GC31-N512 | u-ab268 |
Hindcast Simulations:
- MO rose suites mi-ba751 was initialised at 1998-01-01 from a dump file from its spinup suite (mi-ba451). This
was initialised from the RCM (MOHC-HadREM3-GA705-25km) dump at 1997-01-01. The RCM was initialised at 1992-01-01 from reconfiguration, which uses offline JULES run to create estimate of the soil moisture state. - MO rose suites mi-bb055 was initialised at 2003-01-01 from a spinup dump of suite mi-ba451.
pcControl Simulations:
- MO rose suite mi-ba488 was initialised at 1997-01-01 from a PD GCM dump (MOHC-HadGEM3-GC31-N512, MO rose suite u-ab261). The first year of mi-ba488 is regarded as spinup and the simulation begins at 1998-01-01.
- MO rose suite mi-bb069 was initialised at 2002-01-01 from a u-ab261 dump. The first year of mi-bb069 is its spinup and the simulation begins at 2003-01-01.
Future2100 Simulations:
- MO rose suite mi-ba489 was initialised at 1998-01-01 from a Future GCM dump (MOHC-HadGEM3-GC31-N512, MO rose suite u-ab268). The first year of mi-ba489 is regarded as spinup and the simulation begins at 1998-01-01.
- MO rose suite mi-bb070 was initialised at 2002-01-01 from a u-ab268 dump. The first year of mi-bb070 is its spinup and the simulation begins at 2003-01-01.
We have generated single-variable netCDF files consistent with CMOR 3.7, CF-1.7 CMIP-6.2 and CF standards. The filename is given in the CMIP6 format, as following:
{variable}{frequency}{model_id}{experiment_id}{variant_id}{grid_type}{start_date}-{end_date}.nc
For example: ‘pr_day_HadREM3-RAL1T_hindcast_r1i1p1f1_gn_19980101-19981231.nc’
Files from the future2100 also have dates from 1998-01-01 to 2008-01-01, even though their forcing are for 2100. This is to reflect the dates of the base SSTs that are used. For this reason, time series plots will have 1998-2008 dates for both the pdControl and the future2100 simulations.
The CMIP6 netCDF file names contain information on variant-id, with the format e.g.: r1i1p1f1. The letters express the following: r - realisation (i.e. ensemble member), I - initialisation method, p- physics, f- forcing. We have decided not to use the variant ID to distinguish between the different CSSP Brazil simulations (because of subtly different meaning of r, i, p and f in the CMIP experiments) and therefore filenames from all simulations are defined as r1i1p1f1.
The netCDF files metadata includes information about the timestep (60s or 1/1440 of a day), diagnostic name and units, the experiment name and suite id and a link to this url.
Citations: Halladay, K., Kahana, R., Johnson, B. et al. Convection-permitting climate simulations for South America with the Met Office Unified Model. Clim Dyn 61, 5247–5269 (2023). https://doi.org/10.1007/s00382-023-06853-0
Kahana, R., Halladay, K. et al. Future precipitation projections for Brazil and tropical South America from a Convection-permitting climate simulation. Front. Clim. Sec. Predictions and Projections Volume 6 (2024) doi: 10.3389/fclim.2024.1419704 https://www.frontiersin.org/journals/climate/articles/10.3389/fclim.2024.1419704/abstract