GENIE:GENIE Tuning rel-2-7 phase 3

=Strategy for tunings of release 2.7.n - Phase 3=

Summary
TODO: to be written

Model setup and compilation
Analogous to the setup and compilation of the base version of  described for the phase 1 of the tuning exercises, except that the source code is available from   instead of from.

Additions, modifications, corrections, and updates made in branch 'rel-2-7-7-tuning'

 * Addition of code for the computation and output of normalised root-mean-square (RMS) differences between model output and data based fields of phosphate and alkalinity (see section on Target data below for details) along with the normalised RMS differences used for the model-data evaluation of the model runs of phases 1 and 2.

Base configuration
The phase 3 tuning exercise of the model is applied to the 36x36x16 equal-area-grid resolution of the model as set up for experiment  of the second phase of the tuning exercise (see the page for the phase 2 of the tuning exercises for further details]] with the ATCHEM and BIOGEM model components additionally enabled. The configuration file for a 5000 year spin-up simulations for setup    is provided in the directory.

Base parameters
Except for the parameters listed in the two tables below, the base parameters are set to the same values as used for the base version of  used in phase 2 of the tuning exercises.

The values which have been varied in exercise  of phase 2 are now set to the values of one of the solutions (point 69) of the final ensemble of tuning exercise  :

TODO: add list of non-default parameter values for BIOGEM/ATCHEM

Tuning configuration
The same 6 parameters of BIOGEM as varied in the study by Annan and Hargreaves [2010] (parameters 7-12 in Table 2 of Annan and Hargreaves [2010]) have been selected for parameter estimation in tuning exercise :

Build macro definitions
Except the definition listed in the table below, the settings are the same as used for the base version of  used in phase 1 of the tuning exercises.

Wind-related boundary conditions
All settings for the wind-related boundary conditions are the same as used for the base version of  used in phase 1 of the tuning exercises.

Target data
For the evaluation of the individual simulations, annually-averaged model output from the final year of the simulations is compared with observational data for the two three-dimensional fields of phosphate and alkalinity in the ocean as output by BIOGEM. Two RMS error values are returned (in addition to the four RMS differences returned for the model-data comparison as used in phases 1 and 2) for each field, as well as a composite error function value (in addition to the composite error function value for C-GOLDSTEIN).

The phosphate field used in the model-data comparison is the objectively-analysed annual-mean climatology of phosphate in the ocean [Garcia et al., 2006] of the World Ocean Atlas 2005 (WOA05; http://www.nodc.noaa.gov/OC5/WOA05/pr_woa05.html) from the National Oceanographic Data Center (NODC; http://www.nodc.noaa.gov/). The alkalinity field used in the model-data comparison is the total alkalinity field from the Global Ocean Data Analysis Project (GLODAP) data set [Key et al., 2004], and is available from the GLODAP data server. See also the report by Sabine et al. [2005] for further details on this dataset.

TODO: Description of the interpolation of the target datasets onto the model grid and computation of the normalised RMS difference.

List of required input files
Names in bold denote input files which are required in phase 3 of the tuning exercises, but do not correspond to any of the input files required during the second phase.

Validation
TODO: To be described and performed.

Tuning 3b
A different set for the parameters which have been varied in exercise  of phase 2 is used to perform the tuning. These parameters are now set to the values of a different solution (point 22) of the resulting ensemble of tuning exercise :