Version 19 (modified by smasson, 10 years ago) (diff)

Executive Summary

This document describes the development and current status of the web NEMO configurations data base definition, development and implementation. The goal of this work is to document the diversity of the simulations that can be achieved with the ocean model NEMO and illustrate the impacts of different configurations of the same code.

To do this, we established and delivered a standard experiment protocol and a complete set of diagnostics to be applied to NEMO outputs in order to qualify and quantify the biases of simulations and the differences between existing experiments. All information and source code are available on http://forge.ipsl.jussieu.fr/nemo/wiki/core2_cmip and all simulations can be found on the following DODS server: http://dodsp.idris.fr/reee605/

1. INTRODUCTION

The main goal of this task is to make model developers and users aware of the important difference existing between a model version and a model configuration. This is a key point to ensure transparency and comprehension of the work done around climate models. The same model version can indeed be used in numerous configurations (physical options, choice of parameterizations, horizontal and vertical grid description and resolution, definition of the initial state…) that can lead to very different results. The model inter-comparisons on which is based a large part of IPCC report from Working group 1, are exploring differences between models. We are here illustrating differences between different configurations of the same model.

In this part of WP4, we are therefore targeting:

1) to document the diversity of the simulations that can be achieved with the ocean model NEMO in order to demonstrate the importance of a complete description of the model configuration.
2) to constitute a data base illustrating the impacts of different configurations of the same code. This catalogue of simulations based on the same model but with different configurations will later on be used to understand the processes at stake in ocean modelling.
3) to establish and deliver a standard experiment protocol that could be used by researchers to validate their own implementation of NEMO. Next, this protocol will be used to explore and quantify the specificities of a configuration of a particular user.

2. experimental Protocol

We decided to follow the experimental protocol defined by the CLIVAR Working Group on Ocean Model Development (Griffies et al. 2009). We therefore adopt the standard Common Ocean-ice Reference Experiments dataset (CORE2) to force the model (Large and Yearger 2009) that is available for download from the GFDL server. This dataset contains the air-sea fluxes of momentum, heat, fresh-water computed globally at frequencies ranging from 6-hourly to monthly. This dataset provides either one normal year of forcing (to force long climatological simulations) or 60 years interannual forcing (1948-2007, to force simulations over the recent period).

We decided to conduct 2 kinds of experiments:

  • a 2000-year long simulation forced with the perpetual normal year
  • a 180-year long simulation made of 3 time 60 years of the interannual forcing

The preparation of the input files has been done through cdo and nco. This work is make up of 3 parts:

1) mask the continents
2) fill masked values with neighbouring oceanic values
3) interpolation on ORCA grid with bilinear (scalar) or bicubic interpolation (vector)

Note that the 3rd point is not absolutely needed as NEMO can perform "on the fly interpolation". The rotation of the vector on ORCA grid is done within the code.

The complete description of this work is given on NEMO web page: http://www.nemo-ocean.eu/Using-NEMO/Pre-and-post-processing-packages/Pre-Processing/cdo-interpolation. Interpolated data can be found at http://dodsp.idris.fr/reee605/CORE2_clima/ and http://dodsp.idris.fr/reee605/CORE2_interan/

A first set of these simulations was performed with NEMO v3.2 and libIGCM 1.4 for the running scrip environment.

The use of libIGCM with NEMOv3.2 is detailed in the following web page: http://forge.ipsl.jussieu.fr/nemo/wiki/libIGCM

An updated manual for NEMO v3.3 is available here: http://forge.ipsl.jussieu.fr/nemo/wiki/libIGCM_nemo_v_3_3

The following web page details all the procedure to perform this simulation: http://forge.ipsl.jussieu.fr/nemo/wiki/core2_cmip

This page gives the "step by step" instructions to perform these simulations and lists the complete set of parameters (namelists, cpp keys…) that has been used to define the model configuration of these reference experiments.

All simulations outputs are freely available on the following DODS server:

3. output diagnostics

A set of diagnostics mostly based on Griffies et al. (2009) has been specifically written to fit the specificities of the global grid used in NEMO:

  • the tripolar curvilinear grid
  • the C-grid discretisation
  • the bottom partial steps
  • the linear free surface
  • the east-west boundaries condition and the north pole folding

These diagnostics are based on the SAXO package developed at LOCEAN, which uses the IDL software. IDL is a not a free software, however we also provide a compiled version of these diagnostics that allow to use it with a free version of IDL called the "IDL virtual Machine". There is therefore no need to have an IDL license to perform all those diagnostics. All the source code of these diagnostics (more than 2100 lines) are available at the following address:

https://forge.ipsl.jussieu.fr/nemo/browser/trunk/NEMOGCM/CONFIG/ORCA2_LIM/IDL_scripts.

More than 50 diagnostics have been developed and implemented. The diagnostics allow either a comparison with observed climatologies, or a comparison between 2 experiments. A special care has been taken to make the use of this tool as easy as possible and as flexible as possible. The user controls the description of the files on which the diagnostics will be applied trough a simple ascii file. The execution of the diagnostics is done through the call to a shell script: ./std_main.sh

Two sets of diagnostics are available:

  • maps (in surface or at depth, poles projections, …), vertical sections (global or basin mean, equatorial sections, …), stream functions, integrated heat transport… :
    ./std_main.sh -plots
    
  • global mean times series and time-depth diagrams:
    ./std_main.sh -ts
    

By default, this shell script produces postscript files. However, it can also produce pdf or html files with the use of -pdf or -html options of std_main.sh. A complete description of the use of this tool is detailed in the README file provided with the source code.

Several examples of the plots and time series produced by this tool can be found in the following web page: http://forge.ipsl.jussieu.fr/nemo/wiki/core2_cmip

As an illustration, there are 2 specific cases of plots and time series.

BIBLIOGRAPHY

  • Griffies, S M, A Biastoch, C Boning, F Bryan, G Danabasoglu, E P Chassignet, M H England, R Gerdes, H Haak, and R W Hallberg. 2009. Coordinated oceanice reference experiments (cores). Ocean Modelling 26 (1-2): 1-46. http://dx.doi.org/10.1016/j.ocemod.2008.08.007
  • Large, W and S Yeager, 2009.

Attachments (15)