Initial quality controlled subset release

Table of Contents

1. Introduction
2. Available data fields
3. Perturbations
4. References

Introduction

This document describes data from the climateprediction.net project which was the basis for the analysis published in the paper [Stainforth et al.]. The data comprises a grand ensemble of 2578 simulations of the HadAM3 GCM exploring 6 independent physical perturbations and perturbed initial conditions.

A brief description of the format of the data, and associated metadata, is given below. See [Stainforth et al.] for a detailed discussion of the experiment.

Further enquiries can be directed to Stephen Pascoe at the Rutherford Appleton Laboratory, UK.

Available data fields

The GCM experiment is composed of phases of 15 years:

1. 1811-1825. Spin up.
2. 1826-1840. Pre-industrial CO₂
3. 2051-2065. Double CO₂

The data fields available for each model simulation are described in Table 1. These fields are distributed amongst 17 netCDF files depending on the experiment phase, UM ouput stream and field domain. Each filename is of the form:

<job_id>.<output_stream>.<domain>.x<phase>.nc

The filenaming convention is demonstrated in Table 2. Note that when constructing filenames the UM job id is composed of the 4 character experiment id followed by either aa, ba or ca according to the experiment phase. Also note from Table 2 that some climatology fields are seasonal means and some are monthly means. The heat convergence field (UM output stream pc) is also available in UM ancillary format with the file extension .anc.

Perturbations

The grand ensemble includes perturbations of 7 physical parameters and one initial condition parameter. These parameters are listed in Table 3 and are also included as attributes within each netcdf file. The physical parameters CW_LAND and CW_SEA are always varied together and therefore there are 6 independent physical perturbations.

The actual parameter values for each simulation are stored as attributes in the netCDF file and are summarised in Table 4.

The perturbations applied to each simulation are tabulated in the file perturbations.dat which accompanies the dataset. This is a text file which can be easily imported into a spreadsheet or parsed by scripts. The format is two header lines, prefixed with the '#' character, followed by tab delimeted column data:

Column 1

lists the simulation directory.

Column 2

Perturbation id. This is a base 3 integer uniquely identifying the physical perturbations applied.

Columns 3-9

describe the 7 physical parameters as being unperturbed (stnd), perturbed low (low) or perturbed high (high).

Column 10

lists the initial condition ensemble parameter DTHETA.

The perturbation id repeats the information in columns 3-9 in a form that makes identifying duplicates easier. Each digit in the perturbation id corresponds to the value of a physical parameter using the encoding 0=low, 1=stnd and 2=high. The file has been sorted by perturbation id such that initial condition ensemble members occur together.

References

Stainforth 2004

Stainforth D.A. et al. Uncertainty in predictions of the climate response to rising levels of greenhouse gases, Nature, January 27.