MOTIF/PMIP2 Materials

This page contains information about the following MOTIF/PMIP2-related subjects:


The pp-motif Post-Processing Utility

Introduction

The pp-motif utility is designed to produce CF-compliant netCDF files for incorporation into the PMIP2 database. It has two parts, a model-independent wrapper and infrastructure that deals with all of the issues surrounding writing CF-compliant files (using the CMOR library), and a model-dependent part that deals with reading model output files, describing the axes used by a model and so on.

To use pp-motif for a new model, it is necessary to produce appropriate model-dependent code. The current release of the software contains model-dependent code for the FOAM and HadCM3 models as used in Bristol, and it should be relatively clear how to implement the appropriate functions for a new model — the interface between the model-independent and model-dependent code is relatively thin, requiring the implementation of only five or six functions.

Prerequisites

pp-motif is written in Fortran 90. It has so far only been tested with the Intel Fortran compiler (versions 8.0 and 8.1). There are known problems with compiling pp-motif with the Portland Group compilers.

It relies on the CMOR library, which is included in the pp-motif distribution. It also requires an installation of the netCDF libraries with Fortran 90 module files. The version of the netCDF libraries currently in use in Bristol is 3.5.1. A patch is required to build this version of the libraries with version 8.0 of the Intel Fortran compiler. The version of the netCDF 3.5.1 source linked below incorporates this patch, and will work with the Intel compiler.

Other prerequisites are:

These are all available from the download section below.

Download


Benchmarking Plots

Benchmarking plots are available for the following models:

For each model, all possible benchmark plots are generated for all available runs. Note that not all variables are available for all models. The runs available at the time the plots were generated were:

Model 0K OA 0K OAV 6K OA 6K OAV 21K OA 21K OAV
CCSM X X X
ECBILTCLIO X X X X X
FOAM X X X X
GISSmodelE X X
HadCM3M2 X X X
IPSL-CM4-V1-MR X X X
MIROC3.2 X X X
MRI-CGCM2.3.4fa X X
MRI-CGCM2.3.4nfa X X
UBRIS-HadCM3M2 X X X X
UToronto X X
UVic X X X

The variables available for each model were (an X means the variable was available for all model experiments, while an asterisk means the variable was only available for some experiments):

Model tas pr evspsbl mrso sic tos
CCSM X X * X X
ECBILTCLIO X X * * X
FOAM X X X X X
GISSmodelE X X X X
HadCM3M2 X X X X X
IPSL-CM4-V1-MR X X * *
MIROC3.2 X X X X X
MRI-CGCM2.3.4fa X X X X X X
MRI-CGCM2.3.4nfa X X X X X X
UBRIS-HadCM3M2 X X X X X X
UToronto X X X X X X
UVic X X X X X X

The code used to produce these plots is available here. It relies on the NCAR Graphics Command Language (NCL), which is available from here.