Home

GENIE User Manual: Global biogeochemical cycling

Contents

1. either just this single directory AND all its contents or down load and unpack the tar ball of various forcing subdirectories The forcing subdirectories live in genie_forcings Finally to run something old_rungenie sh genie_eb_go_gs_ac_bg genie_userconfigs worbe2_preindustrial 11 all one line 21 B Available GENIE modules Component Abrv LME Reference IGCM 3 D AGCM de Forster et al 2000 Atmosphere EMBM 2 D Fanning and Weaver 1996 ixe GOLDSTEIN Edwards and Marsh 2005 A Slab de Forster et al 2000 fo Fixed 2 F gs GOLDSTEIN multi option Edwards and Marsh 2005 Sea ice ss Slab de Forster et al 2000 fs Fixed GENIE land Meissner et al 2003 mil IGCM land de Forster et al 2000 Land surface el ENTS Williamson et al 2006 Fixed a E o ae ee j TRIFFID Cox 1998 Vegetation ER ENTS Williamson et al 2006 ixe Ocean biogeochemistry bs siocem Ridawetlet al 20070 Deep sea sediments sg seooem _Ridgwell and Hargreaves 20071 Atmospheric chemistry ac AtcHem Ridgwel etal 20078 Terrestrial weathering re rorem Figure 2 Table summarizing the different GENIE modules 22 C GENIE directory structure This section describes the directories and files pertinent to configuring and running biogeochemistry experiments in GENIE 1 genie atchem genie biogem genie embm genie goldstein genie i
2. 25 initialise_gem f90 Tracer property and relationship definition genie genie matlab MUTLAB scripts for plotting fields from the 2 D and 3 D BIOGEM and SEDGEM netCDF results files plus related plotting configuration files e genie genie rokgem data input genie genie rokgem src fortran e genie genie sedgem data input Input data files for the deep sea sediment biogeochemistry module SEDGEM This is the location searched for input data files if the default input directory namelist setting is used whihc is set by the namelist parameter sg_par_indir_name In addition to any files added locally e g for specific model runs the SVN server holds the following files Masks which specify the sediment grid locations at which sediment cores are generated x sedgem_save_mask 36x36 x sedgem_save_mask 72x72 Bioturbational mixing profile x sedgem_sed mix_k dat Sediment grid topography bathymetry x sedgem_topo D 36x36 x sedgem_topoD 72x72 e genie genie sedgem src fortran Source code for the ocean sediments biogeochemistry module SEDGEM cpl_comp sedocn f90 cpl1 flux_sedocn f90 end_sedgem f90 Final auditing and reporting initialise_sedgem f 90 Initialize SEDGEM read in name list settings sedgem f90 Primary ocean biogeochemical cycling subroutine loop saving of time series and time slice data re start file saving
3. 32 F FAQ aka has this dumb question been asked before F 1 Running experiments General F 1 1 Help My experiment has died If when using the old_rungenie sh shell script to run a GENIE 1 experiment at the command line interactively it all goes horribly pear shaped 1 Dies absolutely immediately Check that the old_rungenie sh shell script has executable permis sions Also check that the directory you are trying to run the model from is your home directory C Does it die almost immediately before the line gt gt Here we go appears If so there should be an error message telling you that a particular file or directory cannot be found Check e All the files and directories you have specified exist e You have not omitted spaces where you should not have nor added spaces where a _ separator was required e You have not mis spelt anything a common cause of problems is in reading the number one C1 for the letter el 1 or vice versa in the computer font Courier Another potential source of files not being found either as the run script starts or as GENIE initializes is that may not necessarily get expanded into the path of your home directory e g home mushroom In this situation can simply be replaced with HOME Note that as well as making this substitution at the command line the user config file may also contain instances of J
4. i and latitude j 15 5 Saving data This section covers how to generate results files from the model runs and what file is which and which contains what Confused you soon will be The following sections cover 1 The two basic ways of saving data as spatially explicit time slices and as a continuous time series of a single global or otherwise representative variable 2 File naming conventions of results files 3 Output produced in netCDF format 4 Plotting using the MATLAB scripts in genie matlab 5 Restart files Results files as well as re start files are all saved in module specific results directories So for BIOGEM results files you will need to go to genie_output RUNID biogem where RUNID is the model experi ment name and for SEDGEM results genie_output RUNID sedgem Note that ATCHEM currently does not save its own data files which are instead saved by BIGOEM by accessing the atmospheric com position information placed on the surface atmosphere tracer array Also note that SEDGEM saves data only at the end of an experiment and does not have access to the time series and time slice information available to BIOGEM or even the current time during an experiment 5 1 Time slices One of the most informative data sets that can be saved is that of the spatial distribution of properties However saving full spatial distributions a 36 x 36 x 8 array for any or all of the t
5. sedgem_box f90 Ocean biogeochemical cycling calculations time dependent forcing updat ing tracer auditing calculations sedgem_box_archer1991_sedflx f90 sedgem_box_ridgwe112001_sedf1lx f90 sedgem_data f90 Load re start initialize ocean grid and tracers ASCII time series data saving ASCII time slice data saving read in name list sedgem_data_netCDF f90 netCDF format time slice data saving sedgem_lib 90 Parameter and array definitions having SEDGEM module scope set namelist definitions sedgem_nnutils f90 Neural network utilities makefile Errr the makefile 0 e genie_archive This directory IS NOT created automatically for you and is where the old_rungenie sh run script will attempt to put archived results e genie_forcings This directory IS NOT created automatically for you e genie_log The location where standard output and error streams are directed into a file This directory IS NOT created automatically for you 26 e genie_output The default directory where the results will be sent This directory IS created automatically for you e genie_userconfigs 27 D Namelist parameter definitions and defaults Parameters in the model are controlled via namelists lists of parameter names whose values are passed into GENIE when it initializes The default namelist parameter values are listed in a set of Tables which can be downl
6. 33 E2 aE a iaeoa ce ae wt a a a AA eed rae hh he a ee SS ee SE SS 34 FE 3 Biogeochemistry 2 s s es e waa Bk ke ow eR ads we a ws 34 F 3 3 How do I do an ocean CO i injection experiment 208 36 F 3 4 I have been told ERROR MESSAGE Particulate tracer CaCO3 does does not have the corresponding r ggy ac 37 F 4 1 Do I have to submit experiments to the queue rather than running interactively 37 F 4 2 Can I leave all my experiment results on the cluster for ever 37 38 eee te es by e abastos 38 cease ius an bh ooh oe eee 38 AAA 38 a Diane Re he eee a oh ee 38 H Contact Informatio 40 1 README blurb This document covers the directory structure of the source code and ancillary data input files instructions for compiling the GENTE executable and how to configure and execute an experiment i e how to get some actual work done Also covered is how to specify the results you want saved and at what points in time in what format they are saved and some ways of getting the results visualized A Quick Start Guide is provided in Appendix A page 20 if you prefer to just jump in semi blind However it is advisable to at least read the FAQ section lest any of the FAQs become EMFAQs Even More Frequently Asked Questions For additional information refer to the companion Tutorial genie tutorial and HOW TO genie howto documents The source for this document is in TEX and resides
7. etc Altering the value of ea_20 thus lets you control climate and circulation without having to adjust CO2 and the carbon cycle Next SSTs in BIOGEM can be specified independently of the climate model You achieve this by setting up a restoring forcing of ocean temperatures at the surface Note that by default prescribing SST s or SSSs in BIOGEM does not propagate through to the climate model which does its own independent climate thing based on the value of ea_20 This allows you to retain the surface temperatures and thus solubility associated with a x1 CO2 World but have a warmer more stratified ocean appropriate for a much warmer World 38Tt is OK to change the flavor of GENIE as linking is done by the C compiler 39Namelist ea_36 y 34 What actually happens is that BIOGEM receives both the altered circulation field and the altered SSTs due to x4C02 but sets its own SSTs internally rather than use those calculated by the climate model Setting up the SST restoring is in principal just like for PO4 The values for the SST field you can simply copy and paste out of a prior x1C02 experiment The converse experiment is to have circulation and biological productivity not change but explore the effect of changes in SST driven solubility i e to separate the solubility pump from circulation change effects on glacial CO2 F 3 2 What is tracer auditing should I have it switched on When developing a new model paramet
8. Many of the results files are created during initialization and are periodically updated appended to during the exper iment Even before the experiment has finished they may contain data that is useful to view 22 The mid points at which time slices are saved are specified as described above 23 The mid points at which time slices are saved are specified as described above 24 This data is saved only at the termination of an experiment i e the netCDF file contains only a single time slice 17 any of sediment cores to be saved is specified in a plain text ASCIT file pointed to by the string value of the namelist parameter sg par sedcore_save_mask_name In the mask file a 1 indicates a location to save a sediment core at and a 0 indicates that no sdiment core should be saved at this location This file must be present so to save no sediment cores simply populate the file with all zeros in an xx by yy grid 3 sedcore_ These files contain pseudo time series of surface sediment environmental properties at each of the requested sediment core locations if any 4 seddiag_misc_DATA_GLOBAL res A summary of mean global sedimentation dissolution and preservation fluxes and surface sediment composition 5 seddiag misc_DATA FULL res Surface sediment and bottom water properties at each and every sediment grid point 5 4 netCDF format output The primary format for saving spatial 2 and 3 D data by BIOGEM and
9. a patch file The altered configuration details are added patched to create a fine tuned configuration config file with a name unique to the model experiment For exampld 4 taking a basic science module flavor and configuration genie_eb_go_gs_ac_bg config and modifying it according to the biogeochem experiment defined by genie_userconfigs worbe2_preindustrial_1 results in the creation of a new configuration file genie_eb_go_gs_ac_bg worbe2_preindustrial_1 config The new configuration file will contain additional settings such as e The number of time steps in the model in order to exactly achieve your requested run length e Whether a re start file is to be used or not and if so where the re start files are located e The location of parameter files required by the various science modules 3 Finally the working directory is changed to genie genie main and the model is automatically invoked in the usual way i e genie_example job f configs genie_eb_go_gs_ac_bg worbe2_preindustrial_1 config BUT NOTE that this is to illustrate what the rungenie sh script does automatically for you and DOES NOT mean that you should necessarily enter in the genie_example job f command by hand unless you have completely configured an experiment manually E 2 technical details In glorious and wonderful and completely tedious detail old_rungenie sh actually does the following 1 ACCEPT PASSED PARAMETERS Check that
10. configuration specified by genie_eb_go_gs_ac_bg and using as a restart the experiment saved in this location genie_output genie_eb_go_gs_ac_bg worbe2_preindustrial In this methodology for each model experiment you will require a plain text ASCII file that details any namelist parameter changes compared to the GENIE defaults and the basic flavor and configuration defined in the config file If you require any tracer forcings then you will also need to specify the directory containing the necessary forcing specification files The location of the directory where GENIE will go looking for forcing specification files is set by a namelist parameter bg par fordir name Full details of the specification and application of tracer forcings is given later Full details regarding what old_rungenie sh actually does can be found in Appendix E page 29 START l SELECT available base configs are listed in BASE CONFIG gt genie gemine main configs gt select one e g based on one used l in a similar pre existing experiment CONSTRUCT simplest to A ee USER CONFIG P copy and rename existing user config from genie_userconfigs y CHECK i biogeochem parameters are consistent with base config ii time slice and time series specifications are consistent with the desired run length y CHECK Design new SAVE 4 iii desired tracer or re use pre existing USER CONFIG forcing s are lt gt tracer configuratio
11. e mean atmospheric tracer concentrations plus total inventory e mean ocean tracer concentrations plus total inventory e mean plus total global productivity e mean plus total global sedimentation In the SEDGEM results directory some or all of the following files will be written 1 fields_sedgem_2d nc Contains 2 D fields of sediment surface and ocean bottom properties 24 2 sedcore_ These files contain the stacked records of accumulated deep sea sediment composition Files with a name of the form sedcore_xxyy res are also only saved at the very end of the model run Each is a plain text ASCII file format containing the down core properties of a specific sediment grid point The xxyy identifier corresponds to the i j grid location of the artificial sediment core as a zero padded 4 digit integer The data is saved in a space delineated column format with column headers In some of the column headers there is no space to include the units for the solid components e g POC CaCO3 det FeO opal ash the units are either wt the default or vol depending on whether the option report sediment data as a mass fraction in sedgem_config par has been selected or not while isotopic components are reported as per mil The locations if 20For historical reasons the maximum number of time series and time slice data points was set to 4096 This is set by the parameter n_data_max in biogem_1ib f90 and can be altered if required 21
12. in time or not according to the logical value of the namelist parameter bg_ctrl_misc_t_BP which is false by default 5 2 Time series The second data format for model output is that of a time series of of change in a single integrated property of the Earth system Model characteristics must be reducible to a single meaningful variable for 17 With the exception of sediment core location environmental properties which is saved whenever the sediment column grows or contracts by a 1 cm unit 18 An empty list is valid time slices will then be populated for you at an interval set by the time slice integration interval But if you really don t want any time slices just set the first or only time point to occur beyond the end year of the run 19 The location of this file is specified by the BIOGEM data input directory namelist parameter bg_par_indir_name which by default is genie biogem data input 16 this to be practical i e saving the time varying nature of 3 D ocean tracer distributions is not Suitable reduced indicators include the total inventories in the ocean and or atmosphere of various tracers or equivalently the mean global concentrations partial pressures respectively global sea ice coverage A file containing a series of model times t at which data points need be saved is defined as before in a file whose name is specified by the namelist parameter bg_ par_infile_sig name The data values saved in the t
13. involve a change in compiled array dimension The compute nodes of the cluster do not have access to the FORTRAN compiler Hence submit ted jobs cannot recompile modules and all science modules must be already compiled when a job is 33 submitted 38 To recompile you must first force a clean of the compiled modules From genie genie main issue make cleanall Now you need to recompile and re link the science modules To do this first start an interactive run of the experiment you want to conduct This will ensure that it is correctly compiled This also serves as a visual check that you have requested a user config restart etc that actually exists Start the run for the length of time you intend to use when submitting the experiment as a job to the queue but kill it keyboard command Ctrl C once it is compiled and you are happy that it is running OK say after 10 years At this point you can be reasonably confident that the experiment is safe to submit the job to the cluster and all files and inputs are as they should be If you have multiple experiments all with the same resolution and number of tracers you DO NOT need to re run interactively or attempt to recompile Also you can add modules and not recompile i e you can interactively run an ocean only carbon cycle And then submit it And then submit an experiment using SEDGEM as well Because when the model is compiled ALL sciences modules are compiled
14. meaning that all there is to do is just link them which does not require the ifort FORTRAN compiler F 2 Climate F 3 Biogeochemistry F 3 1 Can I separate out solubility SST related changes from stratification and circula tion changes With BIOGEM coupled to the climate model core of GENIE 19 a change in atmospheric CO2 will induce a change in SSTs which in turn affect the carbon cycle and feedback on CO2 via changes in solubility and via changes in circulation stratification and thus biological productivity There are times when is it helpful to separate out solubility related changes from circulation related changes This equally applies to dissolved 02 and CO2 The problem is that you need a change in climate and surface temperatures in the climate model in order to induce a change in circulation There is a way of having an altered climate and circulation which then affects the marine carbon cycle yet specify the SSTs seen by BIOGEM and thus used in solubility calculations First of all control the radiative forcing of climate internally in the EMBM rather than externally by the atmospheric CO2 concentration calculated by ATCHEM Turn off explicit CO2 forcing of climate by setting ea_36 n The namelist parameter ea_20 will then dictate the EMBM radiative forcing a value of 1 0 default gives no change in radiative forcing CO2 278 ppm a value of 2 0 corresponds to the effect of doubling CO2 4 0 x4 CO2
15. on branches all handled by the Subversion SVN code repository We will assume that the model directory structure genie results directory and optional model configuration directory and run scripts all reside in a directory given by which may look something like home username 1 stands for the home directory 2 At the prompt where username is your username Model output the directory where the results of the model experi ments will be saved will then appear here home username genie_output or genie_output Rather than having multiple directories and ancillary files all cluttering up your home directory level and making it look untidy the everything can be placed at an alternative location e g my_model_stuff genie To install and run GENIE with a non standard directory structure the following files in genie genie main will have to be edited user mak user sh The science modules currently available are listed in Appendix B page B2 1 2 Model background blurb The core model component is the climate model C GOLDSTEIN as described in Edwards and Marsh 2005 calibrated in Hargreaves et al 2005 with minor bug fixes described in Ridgwell et al 2007a It comprises three modules e GOLDSTEIN the 3 D ocean circulation model of Edwards and Shepherd 2001 e dynamic thermodynamic sea ice and e 2 D EMBM Energy Moisture Balance Model following Weaver et al 2001 In GENIE 1 terminology
16. such as in specifying particular forcings Further trouble can occasionally arise when using Windoz and editing files e g the user config file and it is possible to corrupted the format of the file For what file s you have edited use the command dos2unix to strip off Windoz formatting characters which are invariably invisible in most editors The syntax for this or see the linux Man pages or even Google it is dos2unix FILENAME If the model starts running but dies with a reported failure to solve the aqueous carbonate system it may be that you need to force a re compile Running GENIE with array dimensions which do not match the number of tracers selected is a common cause of failures to solve the aqueous carbonate system as often calcium ion or other tracer concentrations become corrupted and get assigned nutty and all but impossible values F 1 2 When do I have to recompile GENIE You will need to recompile GENIE in the following situations 1 2 You have just carried out one of the GENIE tests e g MAKE TEST or MAKE TESTBIOGEM You have changed the dimension of the climate model grid which also means an automatic change in the biogeochemistry modules either horizontally e g going from 36x36 to 64x64 or vertically e g going from 8 levels in the ocean to 16 You have changed the number of selected biogeochemical tracers either gaseous dissolved and or particulate The latter two
17. the CO2 injection pyyyyz FC02_UNIFORM This is provided on mygenie seao2 org Download and unpack tar xfz pyyyyz FCO2_UNIFORM tar gz from the directory genie forcings As it stands this is configured to stuff 1 PgC yr 1 of CO2 into the ocean over the course of one year The location of the CO2 injection is some random default place that probably does not exist which is not very good So you need to specify your ocean location For this add the following lines to a user config file bg_par_force_point_i 22 bg_par_force_point_j 33 bg_par_force_point_k 5 which corresponds to a cell in the N Atlantic i j 22 33 at an intermediate depth k 5 The i j k coordinates are counted from left to right with longitude i from bottom to top with latitude j and form top to bottom with depth for ocean level k The land sea mask and maximum depth lowest k integer you are allowed can be got from the BIOGEM 2D netCDF variable grid_level This is a map of the k values 90 means land for the 8 level ocean the ocean depths will be between 1 and 8 8 being the surface So the map is of the depth of the ocean and thus lowest k value you are allowed to use By default using the CO2 injection forcing template you will get 1 PgC emitted to the ocean in the location you specify You can scale the amount of carbon up via the namelist parameter bg_par_ocn_force_scale_val_3 xxx where xxx is the multiple of 1 PgC you want to inject NOT your f
18. the external reservoir boundaries is calculated 3 At a periodic pre defined time the inventories are re calculated The difference between old and new inventories should be equal to the integrated net flux If the relative difference between re calculated inventory and estimated on the basis of net flux differs by more than a predefined threshold then an error message is raised and the model halted if requested 4 The integrated net flux variable is re set to zero and steps 2 4 repeated In short if you are not modifying the code then you can take it on trust that the model distribution is free of major bugs and that spurious gain or loss of tracers does not occur It you don t trust me then switch the auditing feature on Auditing is inactivated by default To activate it bg_ctrl_audit true To adjust the threshold relative tolorancd4 bg_par_misc_audit_relerr value To halt the model 1 if it fails the tracer drift tolorance bg_ctrl_audit_fatal true A secondary benefit of tracer auditing when running the model interactively is that it reports back to you the maximum and minimum value of all the tracers and locations of where this occurs as follows gt gt gt SAVING BIOGEM TIME SERIES year 0 50 278 069 6 501 16 522 3 843 18 543 temp min 0 2713E 03 18 36 8 max 0 3030E 03 4 18 8 40By default this is set to 1 0E 08 41 By default the model will continue running even if there is a
19. this EMBM atmosphere plus GOLDSTEIN ocean plus sea ice linked to GOLD STEIN has the flavor i e specific combination of science modules genie_eb_go_gs The addition of ocean and atmosphere biogeochemistry gives you a coupled climate carbon cycle model which was called CB GOLDSTEIN in its non modular pre GENIE incarnation In addition to the GOLDSTEIN sea ice EMBM flavor genie_eb_go_gs CB GOLDSTEIN has an ocean biogeochem istry module called BIOGEM for BIOGEochemistry Model a module to update and keep track of the atmospheric chemistry greenhouse gas concentration and isotopic composition called ATCHEM AT mospheric CHEmistry Model and a coupler to exchange fluxes between ocean and atmosphere across the air sea interface The coupler also provides BIOGEM with information about the overlying atmo spheric composition and if required can drive the EMBM with trace gas concentrations for calculating radiative forcing of climate As well as indirect coupling via the atmosphere and EMBM BIOGEM and GOLDSTEIN are directly coupled through the exchange of 3 D array information regarding the geochem ical composition of the ocean BIOGEM also makes copies of various key configuration parameters at start up as well as paying attention to sea ice extent etc and provides these all as output independently of GOLDSTEIN The GENIE 1 flavor equivalent to CB GOLDSTEIN is genie_eb_go_gs_ac_bg The configuration cali
20. view of obtaining the source code wiki username and password required In summary If you are using linux then you will most likely have the command line client already installed Try typing which svn to see if the client is in your path Command line and GUI clients for SVN are available for just about any operating system tortoisesvn is popular for windows SVNx is a good choice for the Mac Once you have Subversion installed you will be prompted for your password when required SVN caches your password and so you often do not need to type it You can get check out a copy of GENIE from SVN by typing svn co http source ggy bris ac uk subversion genie trunk username your_username genie To update an existing checkout simply cd to the highest level directory and issue the update command svn update note that you could update just a sub tree of your checkout by changing directory to say genie genie embm and then calling svn update which would result in an updating of just the EMBM module Or you could just update a single file with svn update filename from within the appropriate directory or by additionally supplying the path 2 1 Where am I what is in the file tree The pertinent directories associated with configuring and running experiments with the carbon cycle and EMBM based climate using GENIE 1 are listed in full in Appendix C pagel23 2 2 Configuring GENIE for you local softwar
21. CH4 partial pressure atm F 1 0 F F F 2 01 O1 d13C CH4 0 00 Fordi 0 gt Bs CES E 2 01 O1 d14C CH4 0 00 F 1 0 F F F 2 01 01 sulphur hexafloride SF6 partial pressure atm F 1 0 F F F 2 01 01 nitrous oxide N20 partial pressure atm 14 F 1 0 F F F 2 01 01 d15N N20 0 00 F 1 0 F F F 2 01 01 hydrogen sulphide H2S partial pressure atm F 1 0 F F F 2 01 01 d32S of H2S 0 00 f 0 1 F F F 2 01 01 CFC 11 partial pressure atm f 0 1 F F F 2 01 01 CFC 12 partial pressure atm END OF DATA AAA IN IN IN IN IN ON 010203 0405 06 07 08 09 The description of each column is detailed at the bottom of the file COLUMN 01 include restoring forcing of tracer DATA FORMAT T F t f COLUMN 02 time constant of restoring forcing years DATA FORMAT real COLUMN 03 include flux forcing of tracer DATA FORMAT T F t f COLUMN 04 scale flux forcing of tracer DATA FORMAT T F t f COLUMN 05 force ocean surface in equilibrium with atmosphere needs ocean restoring forcing set DATA FORMAT T F t f COLUMN 06 make forcing uniform over this dimension 2 2D O point ELSE spatially explicit forcing DATA FORMAT integer COLUMN 07 i grid location of point forcing DATA FORMAT integer COLUMN 08 j grid location of point forcing DATA FORMAT i
22. GENIE User Manual Global biogeochemical cycling Andy Ridgwell October 3 2008 Contents 1 README blurb 3 1 1 Naming conventions blurbh aoa aoea a e a 1 2 Model background blah 2 Getting your grubby little mitts on the mode 6 2 1 Where am I what is in the file tree 2 2 Configuring GENIE for you local software environment 3_ Running GENI Sil x anial a Al a a a e a a a adod Ge Rw ahs AS ML AL AL es ah a a wae aS aD 32 Via a shell script 4 il 0 Sa ee ee eek eee eee SEE RRS 3 3 Using old rimgenie sh ura ba eee PO ee Ee Se Se A a 3 4 Job submission to a cluster 2 ee 10 co 0 N 12 13 14 16 Dsl Pimeslicesik ard tet Wine we a Gh we ee ae ee ae ee Oe WA He ee oe EKO 16 BZ Limesseriesi basar FS a AO a oS Be OE REE RAE ER RS PO eH ee 16 Geek vasa as a ae el 17 5 4 netCDF format output o a ao ee 18 5 4 1 Plotting with Panoply o o ooa a a 18 5 4 2 MUTLAB plotting oaa a a 18 5 5 Re start filled a wa eaaa a d uaaa aa e a a a a a a a aa a 19 20 B Available GENIE modules 22 C_GENIE directory structure 23 D Namelist parameter definitions and default 28 E The old _rungenie sh script 29 a ia ee dd 29 E2 technical detailse s vico amp tp RR ewe PP a ras os 29 F_FAO aka has this dumb question been asked before 33 F 1 Running experiments General 33 F 1 2 When do I have to recompile GENIE 0 0 0 0 002 000
23. SEDGEM binary in netCDF network Common Data Form More information on the netCDF format and the libraries necessary to compile the model can be found The writing of netCDF follows roughly the NetCDF Climate and Forecast CF Metadata Convention The netCDF output is written for BIOGEM and SEDGEM separately and both modules have a flag that suppresses data file saving in ASCII format with netCDF format being the default Under unix linux the contents of a netCDF file can always be seen with ncdump h filename nc This will provide the header information of the file The command is included in the netCDF library which has to be present to run the model anyway It s useful to get the NCO software package helping to concatenate files or extract variables as shell command A full list of available software to manipulate or graphically illustrate netCDF files can be found herel If you really really must insist on using Windozzzz a recommended viewer for netCDF is In fact the can be run under linux and on the Mac OS X as well There is also ncBrowsel Again this will also run under LINUX and on the Mac OS X You can also view netCDF files from MUTLAB However first you will need to ensure that you have the netCDF toolbox for MUTLAB Once the toolbox is included you can simply load netCDF files into MUTLAB with the command gt gt ncload filename parametername Or write it in the form gt gt vari ncload filen
24. a job from the current working directory qsub cwd e Request an email is sent when the job starts and or when it finishes see the man pages for qsub for the required syntax e A complete example would be qsub j y o genie_log S bin bash old_rungenie sh lt options gt which merges standard output and error streams and redirects the resulting file to the directory genie_log Note that to redirect output as per in this particular example the directory genie_log MUST be present I have no idea what happens if it is not but it can t be good You can check the status of the SGE job queue with the command qstat f and you can kill a job with the qdel command the job numbers being given by the qstat command On almond ggy bris ac uk you can graphically check what is going on via the www In an X window enter ssh almond ggy bris ac uk firefox no remote amp and follow the links Cluster Status Ganglia and then Job Queue to get to the actual queue listing and cluster status NOTE It is common on clusters for the FORTRAN compiler to not be accessible by the computer nodes almond ggy bris ac ukis no exception in this The implication of this is that the GENIE 1 modules must be already compiled BEFORE a job is submitted to the queue In other words if you have just issued a make cleanall you MUST briefly run your desired experiment or equivalent interactively i e in the shell window to ens
25. ame parametername where vari is going to be a cell array Whatever For more information on using the MUTLAB NetCDF Toolbox also see here 5 4 1 Plotting with Panoply 5 4 2 MUTLAB plotting There are some plotting scripts available for BIOGEM and SEDGEM netCDF output which can be found in genie matlab The plotting scripts are all written as functions and have on line help i e type gt gt help functionname where functionname m is one of 25 The location of this file is specified by the SEDGEM data input directory namelist parameter sg_par_indir_name which by default is genie sedgem data input 18 e plot fields _biogem_2d m Plot a 2 D field from fields _biogem 2d nc e plot_fields _biogem_3d_i m Plot a vertical meridional 2 D slice through the ocean i e all cells have the same i longitudinal coordinate value from fields_biogem_2d nc e plot_fields_biogem_3d_i_AVE m This is based on plot_fields_biogem_3d_i m but plots a slice constructed by averaging longitudinally over a supplied mask For example if the mask is of the entire ocean mask_worbe2_ALL dat the result is a global merid ional cross sectional mean If the mask is just of a single basin such as the Atlantic mask_worbe2_Atlantic dat the result is the Atlantic meridional cross sectional mean Masks can also be constructed that are only a single cell wide longtudinally but which take a meandering path following an ocean tr
26. an sect The trivial usage would be to construct a mask consisting of a vertical slice with equal i the result is then in effect exactly the same as using plot_fields_biogem_3d_i m The mask is a 2 D array of model grid points on the BIOGEM grid A value of 1 represents a vertical column of ocean cells to include whereas a value 0 will exclude all cells in the water column at that particular grid point e plot_fields_biogem_3d_k m Plot a horizontal slice through the ocean i e all cells having the same k vertical coordinate from fields_ biogem_2d nc Note that for the 8 layer version of the ocean circulation model legal values for k range from 1 through 8 with k 8 representing the uppermost surface layer For the 16 level version of GOLDSTEIN legal values for k are 1 16 e plot_fields_sedgem_2d m plot a 2 D fields from fields_sedgem_2d nc 5 5 Re start files Re start files are saved in the results directories by default For ATCHEM BIOGEM and SEDGEM these are in a binary format form unformatted and the filenames take no extension and are simply named after the module For the climate modules of GENIE 1 GOLDSTEIN SEAICE EMBM re start files can be saved in either plain text ASCIT or netCDF format 26e g as in file mask_worbe2_WOCE_P16 dat 19 A Quick start guide The quick start guide assumes that you already have access to the SVN code repository i e have a username and pass
27. an I leave all my experiment results on the cluster for ever NO Nothing is backed up on the cluster and space is not infinite So periodically transfer archived tar gz results off of the cluster and delete both the archive file and the results directory 37 G Known issues G 1 Radiocarbon tolerance The comparison tolerance for make testbiogem has been raised 01 05 08 from 5 to 19 Ulps units of last place think of notches on a ruler marked with numbers resolved by a variable of so many bytes so that the test will pass when using the g95 compiler version 0 91 The variables in question is ocn DIC_14C G 2 Ocean tracer number related compilation problems If the number of dissolved ocean tracers or the grid resolution is changed then the model will need to be recompiled from clean This is because the ocean circulation model GOLDSTEIN is compiled with array dimensions sufficient only for the actual number of selected tracers in the ocean including temperature and salinity Theoretically it might be possible to clean only the GOLDSTEIN and BIOGEM modules as well as the interfacing glue However it is much safer to request that all modules libraries objects and dependency information is cleaned up You can do this from the command line within genie genie main by issuing the command make cleanall One symptom of an incorrectly compiled number of tracers is page after page after page after
28. atchem src fortran Source code for the atmospheric chemistry module ATCHEM atchem f90 Updating of atmospheric composition re start file saving 23 atchem_box f90 currently just an empty template module will hold atmospheric chem istry subroutines atchem_data f90 Load re start initialize atmospheric grid and tracers read in namelist settings atchem_1ib f90 Parameter and array definitions having ATCHEM module scope set namelist definitions cpl_comp_atmocn 90 Set surface atmosphere trace gas interface composition cpl_flux_ocnatm 90 Integrate net trace gas flux to atmosphere from surface end_atchem f90 currently does not do anything except print to screen initialise_atchem f90 Initialize ATCHEM read in name list settings makefile Errr the makefile 0 e genie genie biogem data input Input data files for the ocean biogeochemistry module BIOGEM This is the location searched for input data files if the default input directory namelist setting is used which is set by the namelist parameter bg_par_indir_name In addition to any files added locally e g for specific model runs the SVN server holds the following files Defaul configuration of tracer forcing of BIOGEM configure forcings atm dat Selection and configuration of forcing of atmospheric gas tracers x configure_forcings_ocn dat Selection and configur
29. ation of forcing of ocean dissolved tracers x configure_forcings_sed dat Selection and configuration of forcing of sediment par ticulate matter tracers Configuration files for use in the BIOGEM test make testbiogem genie_eb_go_gs_ac_bg_test_save_sig dat Specification of the times of time series data saving genie_eb_go_gs_ac_bg_test_save_timeslice dat Specification of the times of time slice data saving genie_eb_go_gs_ac_bg_test_windspeed dat Prescribed wind field for calculating air sea gas exchange Various illustrative template time series data saving specification files save_sig historical dat Time series saves specified to cover 1765 5 to 2000 5 at 1 0 year intervals with 0 5 to 1000 5 following save_sig_log10 dat below save_sig_log10 dat Time series saves specified at intervals of 1 0 years up until 10 5 years then 10 0 years until 100 5 years etc etc save_sig_log10_25kend dat As above but terminating at 25000 5 i e the mid point of the year 25000 save_sig_log10_50kend dat As above but terminating at 50000 5 i e the mid point of the year 25000 Various illustrative template time slice data saving specification files x save_timeslice historical dat Time slice saves specified with mid points at 1765 5 1994 5 2000 5 save_timeslice_log10 dat Time slice saves specified with mid points at 0 5 10 5 100 5 1000 5 etc ete 28 T
30. avorite movie viewer rating e g 100 PgC bg_par_ocn_force_scale_val_3 100 0 Note that 100 0 PgC is quite a lot of carbon to be injecting into a single location cell in the ocean model By default the time scale of injection is set as 1 year To increase the time over which the CO2 injection takes place use the namelist parameter bg_par_ocn_force_scale_time_3 which simply scales the time interval i e bg_par_ocn_force_scale_time_3 10 0 36 causes the CO2 injection to take place over 10 years But since the flux is in units of PgC per year you will get 1000 0 PgC carbon total 10 years x 100 PgC yr 1 So a combination of both namelist scaling parameters both flux scaling and interval scaling will be needed for the required total CO2 injection Note that the integer at the end of the namelist parameter name corresponds to the index of the ocean tracer 3 is DIC 12 would allow you to inject alkalinity into the ocean but the you would need to create additional forcing specification files The slightly harder way involves entering in the i j k location explicitly in the forcing configuration file configure_forcings_ocn dat Altering the magnitude and or duration of the flux release requires editing biogem_force_flux_ocn_DIC_sig dat The hardest way requires that two 3D fields explicitly specifying the spatial nature of the forcing flux are created and modified For these alternative options see earlier section on tracer forc
31. bration and evaluation of GENIE 1 is described in Ridgwell et al 2007al Earlier incarnations of the ocean biogeochemistry are described in Cameron et al 2005 and Lenton et al 2006 The global carbon cycle can be further extended with the addition of a representation of the mass balance of solid material deposited to deep ocean sediments a module called SEDGEM SEDiment GEochemistry Model This is coupled via an interfacing module to BIOGEM The overall coupled climate ocean biogeochemistry sediment geochemistry model was previously termed CBS GOLDSTEIN but in GENIE 1 terminology is 3When editing these files particularly in Windoz remember that these files must be executable genie_eb_go_gs_ac_bg_sg The configuration calibration evaluation and application of GENIE 1 with deep sea sediments is described in Ridgwell and Hargreaves 2007 and Ridgwell et al 2007b Further developments to the sediment model are described in Ridgwell 2007 When allowing burial of biogenic material in deep sea sediments using the SEDGEM module a further module is available to provide the long term balance the loss of matter from the ocean This is called ROKGEM for ROcK GEochemistry Model and calculates the solute supply to the coastal ocean resulting from the weathering on land of exposed rock surfaces and soil minerals The flavor of GENIE run with this additional weathering module is then termed genie_eb_go_gs_ac_bg_sg_rg T
32. d tracers are initialized with default values applied to all wet cells while all particulate concentrations in the ocean are zero e In ATCHEM all tracer gas concentrations and isotopic properties are initialized with uniform default values everywhere in the atmospheric grid e In SEDGEM the sediments are set to a homogeneous detrital or ash tracer composition through out A continuing run will carry on from where ever a previous run left off using a set of files called restart files that between them completely define a snapshot state of the system More on this plus examples later and in the tutorial document For GENIE 1 a model experiment can be initiated in two different ways 3 1 manually The basic flavor and configuration of GENIE is run from genie genie main by typing genie_example job genie_examp1le job is a shell script which sets up a basic base configuration of the model A different flavor and configuration of the model is enabled by specifying an alternative base configuration file genie_example job f another config where another config is a specified model configuration flavor config file GENIE is supplied with a number of pre prepared base config files These live in genie genie main configs The names of the base config files are typically structured as genie_aa_oo_ss_xxxx config where e aa is the atmosphere name ig igcm eb embm e oo is the ocean name
33. e environment Some hints for instally and adapting GENIE for your local software environment and machine archi tecture can be found in the documents README and readme main in genie main docs but note that they are specific to GENIE not genie and refer to CVS rather than SVN for code repository access as well as some other out of date details no doubt these will be updated in due course A general GENIE user manual can also be created in a variety of formats including ps and pdf from within genie genie main doc by issuing the command make user manual However this currently suffers from some of the same documentation issues as the ASCII files in genie main is as much as it is specific to GENIE trunk and assumes use of a CVS rather than SVN code repository Again an update will follow ha ha ha ha 4syn co is synonimous with svn checkout 3 Running GENIE Pl discuss this in the reverse order just for the hell of it setting a model experiment going first and only then describing what you should have done in the first place to have set it up correctly It is important to recognize from the outset that there are two fundamental modes of operating GENIE as a new or continuing run A new run is one which starts off from a default state of the system e In the case of the ocean model GOLDSTEIN this is with the ocean set to uniform temperature and salinity throughout and no momentum e In BIOGEM the dissolve
34. erization it is of fundamental importance that careful track is kept of the total tracer inventory of the system in the face of internal mass transfer and any inputs e g prescribed restoring or flux boundary conditions or outputs e g sedimentation No spurious gain or loss of tracer mass must occur as a result of bugs introduced to the code The tracer inventories of the ocean can be periodically calculated and compared to that predicted to have occurred on the basis of any net input or output occurring in the intervening time to help catch bugs The simplest implementation would be an audit carried out at system start up before any transformation of tracer mass has taken place and at the very end after the last transformation of the tracer fields However integrating over over an extended time period can lead to the excessive accumulation of numerical truncation errors Instead the audits are carried out periodically during the model run The periodicity of tracer auditing follows the times specified for time series data saving i e at time listed in the file specified by bg par_infile_sig_name The entire audit procedure is as follows 1 First an initial inventory is calculated achieved by summing the product of the concentration of each selected tracer with the mass of each each cell across all wet cells 2 During the model run the net spatially and time integrated transfer of tracer mass arising from all transfers across
35. escribing flux fields or boundary conditions to the model These are termed forcings There is a serious difficulty in devising a means for implementing a boundary condition to the model because the most interesting and relevant forcings tend to be time dependent i e the magnitudes vary with time Worse the boundary conditions are often spatially heterogeneous i e the magnitude of the forcing is not the same at each and every grid point or cell Dust fluxes are a good example of spatial heterogeneity in a flux input to the system To fully define a time varying 2D flux forcing for say 1000 years of model integration at a 0 01 yr time step equates to 36 x 36 x 1000 x 100 data points clearly beyond simple and efficient data manipulation and processing One end member solution would be to define the complete spatial distribution but leave the fluxes invariant with respect to time The opposite end member would by to fully define the time varying nature of the forcing but distribute the flux uniformly in space Both approaches have serious drawbacks for many applications A compromise scheme has therefore been devised to encapsulate data input and output as efficiently and as flexibly as possible Firstly the forcing of atmosphere ocean and sediment tracers are treated separately For the at mosphere atm_ tracers atmospheric composition is manipulated from the 2 D surface ocean grid of BIOGEM This means that no forcing is applied to the a
36. fi fixed go goldstein ocean s1 slab e ss is the sea ice name fi fixed gs goldstein seaice sl slab and e xxxx is for any more info e g ac ATCHEM bg BIOGEM sg SEDGEM rg ROKGEM A relatively simple configuration file is genie_ig_fi_fi config a dynamical atmosphere IGCM plus prescribed ocean surface and sea ice boundary conditions In addition to specifying the flavor and grid resolution this configuration file also changes the run length from one month to 10 years and turns on daily IGCM atmosphere output It also changes the name of the directory in which the experimental output is placed 5See HOWTO for details of how to initialize with non uniform initial T and S distributions 3 2 via a shell script For configuring and running experiments interactively or submitting jobs to a cluster queue an executable shell script called old_rungenie sh is available old_rungenie sh is invoked from i e the same level as the genie code tree and genie_output results directory If you are going to run the model using old_rungenie sh several additional directories are recom mended which you must create e genie_userconfigs Directory containing user configuration filed These file contain namelist parameter values specific to a particular model experiment These details are appended onto a basic configuration config file specifying general flavor and model configuration see Appendix E page P9 for full detai
37. g for all tracers is mol yr 1 for bulk fluxes and per mil for isotopic composition For restoring forcings the units are atmospheres atm for atmospheric gaseous tracers and mol kg 1 for ocean dissolved tracers For isotopic tracers the units are per mil There are then two alternative ways of providing the final information needed to fully describe a tracer forcing the spatial component by explicitly defining 2 or 3 D distributions the sole original methodology or by assuming a much simpler point source or uniform distribution The latter is rather easier to grasp and implement but lack the flexibility of the former 4 1 Tracer forcing as an explicit 2 3 D distributions The tracer forcing in the explicit 2 3 D methodology requires two further files with filenames e biogem_force_ _I dat e biogem_force_ _II dat These hold information about spatial distributions The data format of these is either 2 D for the atmosphere or surface ocean only rows j and columns i or 3 D for the whole ocean with the successive depth layers k repeated as i j blocks down through the file The way in which these three files are employed to define a spatially explicit time varying input field is as follows The first field file _I dat defines the baseline distribution and the second _11 dat an alternative distribution The difference between the two fields defines the spatial component of a time varying forcing The trace
38. goldsteinseaice genie i configs genie main ee genie matlab 4 docs genie rockgem genie sedgem data input data input files mak file ae fortran fle framework glue makefile E genie_archive results archives a i forcing i tracer forcing E i name 1 configuration files O genie forcings i forcing i name 1 H forcing atchem E biogem g x i experimen o genie_output name 41 embm B results directories tee experiment le i name 2 goldstein a E have deje a vil rokgem 5 i seaice genie_userconfigs sedgem user config files inputifiles model exe Figure 3 Schematic of the GENIE file tree Note that only the sub directory structure for the genie biogem module is shown expanded near identical sub structures are present in all the science modules The dark gray shaded parts of the directory tree are not part of the official GENIE release but instead required when using the old_rungenie sh script The light gray branch is required only when submitting cluster jobs exactly as described in this document The directory in which the genie code tree is installed is assumed to be your home directory and is designated by The details of the directory contents are as follows e genie genie atchem data input Input data files for the atmospheric chemistry module ATCHEM Currently there are no data files that are required to be present here e genie genie
39. he SPACES rungenie sh genie_eb_go_gs_ac_bg genie_userconfigs worbe2_preindustrial 10001 5 The full path and name of any restart required 4 If the 5th optional parameter is not passed then GENIE will run from cold i e no re start If the 5th parameter is present then GENIE will attempt to run from a previously generated re start state 6The old_ bit designates use of the non xml configuration framework TThis can be obtained from mygenie seao2 org 8This directory can be anywhere as it is defined in a passed parameter to rungenie sh see main body of text Can be elsewhere see main body of text 10Note that the script file must be executable in order to run it if not set the correct permissions by entering the command at the linux command line chmod u x old_rungenie sh 11Please note that this is an example and does not necessarily mean that this is the same config file as you want to use 12This must be entered on the command line as an integer even though GENIE will be treating it as a real 13There must exist a subdirectory in genie_output with this name A typical command line to run the model using a specified re start might thus look something like all one line rungenie sh genie_eb_go_gs_ac_bg genie_userconfigs worbe2_historical 2001 genie_output genie_eb_go_gs_ac_bg worbe2_preindustrial in which the experiment worbe2_historical will be run for 2001 years using the flavor and basic
40. he Table Pin Appendix B summarizes the different science modules in GENIE Suggested background readings to help understanding the model its typical applications its successes and perhaps more importantly its limitations are e Annan et al 2005 e Edwards and Marsh 2005 for the fast climate model core component C GOLDSTEIN e Edwards and Shepherd 2001 e Hargreaves et al 2004 For atmosphere ocean sediment carbon cycling and biogeochemistry try e Ridgwell et al 2007a b e Ridgwell and Hargreaves 2007 e Ridgwell 2007 e Cameron et al 2005 for an earlier representation of biogeochemical cycling in the ocean and atmosphere e Lenton et al 2006 e Ridgwell 2001 e Ridgwell et al 2002 Versions of most of the relevant model references can either be downloaded from e the official GENIE project site or e my own publications server Also refer to the companion documentation to this manual e genie tutorial tex Documentation describing various tutorials and example model configura tions and experimental designs e genie howto tex HOW TO documentation potted explanations of how to get stuff done 2 Getting your grubby little mitts on the model First you will a copy of Subversion abbreviated SVN installed on the machine you want to run the model on or at least on the machine you want the source code to be maintained on See e the Wikipedia entry or e the GENIE wiki for a basic over
41. hese are configured for no selected forcings 24 save_timeslice_10g10_25kend dat As above but terminating at 25000 5 save_timeslice_10g10_50kend dat As above but terminating at 50000 5 Modern observed annual mean wind speed for use as an air sea gas exchange boundary con dition in BIOGEM2 windspeed dat e genie genie biogem src fortran Source code for the ocean biogeochemistry module BIOGEM biogem f90 Primary ocean biogeochemical cycling subroutine loop saving of time series and time slice data re start file saving biogem_box f90 Ocean biogeochemical cycling calculations time dependent forcing updat ing tracer auditing calculations biogem_data f90 Load re start initialize ocean grid_and tracers ASCII time series data saving ASCII time slice data saving read in name list 4 biogem_data_netCDF f90 netCDF format time slice data saving biogem_lib 90 Parameter and array definitions having biogem module scope set namelist definitions end _biogem f90 Final auditing and reporting initialise_ biogem f90 Initialize BIOGEM read in name list settings makefile Errr the makefile o e genie genie main see GENIE wiki pages e genie genie main configs see e genie genie main data input Input data files required by the cmngem module defining the tracers in the atmosphere ocean and sediments and the relationships bet
42. ime series frequency itstp ea_5 go_5 and gs_5 set the frequeny of integral diagnostics reporting These are set equal to ma_genie_timestep 1 effectively disabling this feature e f Set climate model components average frequency ianav ea_6 go_6 and gs 6 set the output averaging interval These are set equal to ma_genie_timestep 1 effectively disabling this feature e g Set ROKGEM terrestrial weathering model component reporting frequency The value of rg_par_screen_output is set equal to ma_genie_timestep 1 effectively disabling this feature 5 SET CLIMATE MODEL RE START FILE DETAILS e a Set netCDF restart saving flag The y n value of ea_31 go_19 and gs_14 determine whether a netCDF format restart file is to be saved These are set to n e b Set ASCII restart output flag The y n value of ea_32 go_20 and gs_15 determine whether an ASCII format restart file is to be saved 35For GENIE 1 flavors a different time step is used for GENIE 2 flavors 30 e c Set ASCII restart number ea_29 go_17 and gs_12 contain strings used to form the ASCII restart file names The string is appented by n where n is an integer The value of n is incremented at each new re start save However because the frequency of re start saving has been configured to create only a single restart save at the end of the run the extension of the restart file will always be 1 when using rungenie sh un
43. ime series file do not represent discrete values in time but an average calculated from time t At 2 to t At 2 as per the construction of time slices The averaging interval At is set by the value of the namelist parameter bg_par_data_save_sig_dt If there are less than two elements present in the list a default frequency of data saving will be invoked set equal the averaging interval except in the situation that this results in an unreasonably large amount of data when an order of magnitude or more than one order of magnitude where necessary fewer save points are assumed 24 5 3 Results files At the end of an experiment the BIOGEM results directory will contain some of all of the following files 1 fields biogem_2d nc 2 D fields of mostly ocean bottom ocean surface and sediment surface properties 4 Also water column integrals of certain geochemistry diagnostics air sea gas exchange fluxes atmospheric composition 2 fields_biogem_3d nc 3 D fields ocean dissolved and particulate tracer properties 24 3 biogem_series_ Time series results file 4 biogem_year_ _diag_GLOBAL res Miscellaneous global diagnostic information It is saved at each requested time slice with the file name string containing the mid point of the time slice as years The diagnostics includes e time mid point and integration interval e global ocean surface area and volume e mean global sir sea gas exchange coefficient for CO2
44. ings Section 4 F 3 4 Ihave been told ERROR MESSAGE Particulate tracer CaCO3 does does not have the corresponding ocean tracer Ca selected is this a problem No You are simply being reminded that you have calcium carbon CaCO3 selected as a particulate tracer in the model but although when it dissolves it releases Ca and removes Ca t when CaCO is precipitated you do not have Ca selected as an explicit dissolved tracer in the ocean This is not a problem as by far the most important effect on the carbon cycle of adding subtracting Ca2 is a change in alkalinity which is implicitly account for Only on very long time scales or in deep time situations when the Ca t Mg ratio was very different form today might you need to select Ca t and Mg as an ocean tracer F 4 Running experiments The almond ggy bris ac uk cluster F 4 1 DoT have to submit experiments to the queue rather than running interactively Yes Except for developing the model and debugging testing new experimental designs and forcing a re compile The number of instances of the model that can be run simultaneously interactively is limited by the number of processing cores 4 on the head node The more experiments that are run interactively the slower everything will go Additionally if you even temporarily lose your internet connection an interactively run experiment will die The queue is there for your convenience believe it or not F 4 2 C
45. just T and S giving climate only simulation capabilities but this value is adjusted in the config files Any further deviation from the ocean tracer selection requires a new value to be assigned in the user config For example for 16 selected ocean tracers including temperature and salinity add the line GOLDSTEINNTRACSOPTS DEFINE GOLDSTEINNTRACS 16 to the end of the user config file For setting initial values of tracers the syntax is gm_atm_init_nn 278 0E 6 in the case of an atmospheric gaseous tracer Again nn is the index of the tracer Ocean tracers are initialized similarl gl The units of the tracer initialization parameters are given in the summary Tables 31 There is no user configurable initialization of deep sea sediment composition in SEDGEM 28 E The old_rungenie sh script E 1 Overview The overall strategy for the old_rungenie sh methodology of running GENIE is as follows 1 Copy a config fed defining the required flavor of GENIE plus configuration and calibration of the climatology from genie genie main configs 2 Append a series of namelist parameter values changes specific to a particular model experiment particularly those involving the biogeochemistry and carbon cycling such as the specification of a CO2 emissions forcing 23 or selection of feedbacks e g CO2 and climate CO2 and calcification These namelist parameter alternations compared to the defaults are provided in the form of
46. lative to the total size of the temporaries the program will terminate with a segmentation fault The a solution is to increase the CPU stack space Try ulimit s unlimited G 4 Re starts There is a minor bug associated in how re starts are done The shortwave radiation incidence at the ocean surface is used to calculate biological productivity The spatial field of SW radiation is passed to BIOGEM after the biogeochemical ocean update has been carried out Although physical boundary conditions such as SW radiation are made available to BIOGEM during initialization the SW radiation field is only calculated during the time stepping loop Thus the very first time step that BIOGEM takes 38 has zero in the SW radiation field This affects experiments from a cold start but only trivially More noticeable is that from a re start the ocean carbon cycle experiences a brief hick up with no biological productivity associated with the very first time step This bug was not present in older version of the model and was introduced associated with a re ordering of the time stepping of BIOGEM relative to the climate model components 39 H Contact Information e Andy Ridgwell bandy seao2 org e Gethin Williams gethin williams bristol ac uk 40
47. ls regarding what exactly old_rungenie sh does e genie_forcings Directory where prescribed tracer forcings are defined The location of this directory is defined in a namelist parameter bg par fordir name The default is to use no tracer forcings which is specified by the default forcing files found in the directory genie genie biogem data input e genie_archive Directory containing tar ed and gzip ed results archives e genie_log Directory for storing experiment job logs 3 3 Using old_rungenie sh You use old_rungenie sq to run GENIE taking example flavor genie_eb_go_gs_ac_bg and experiment design configuration patch file worbe2_preindustrial as follows old_rungenie sh lt options gt Now pay attention there are 4 FOUR parameters that must be passed to the script as lt options gt 1 The name of the required flavor and basic configuration but omitting the config bit egi genie_eb_go_gs_ac_bg 2 The full path of the directory containing the user experiment configuration file i e the file containing the list of namelist changes for a particular experiment In the example this might be something like genie_userconfigs 3 The name of the experiment and specification of user config file is to be used i e worbe2_preindustrial 4 The run length of the experiment in years A typical command line to run the model might thus look something like all one line and be careful with t
48. modified 6 CONFIGURE USE OF RESTART Configure GENIE to use a restart the alternative being to start from cold Restart namelist items are set according to the presence or absence of the 5th passed parameter restart experiment name If the 5th parameter is present when rungenie sh is invoked the following action is taken e a Check that the restart experiment directory exists Generate an error message and exit if not e b Set the climate model components to use a restart namelist items ea_7 go_7 gs_7 The syntax is c for a continuing run i e using a restart and n for new from cold e c Set the biogeochemistry model components to use a restart namelist items ac_ctrl_continuing bg_ctrl_continuing sg_ctrl_continuing rk_ctrl_continuing The syntax is LOGICAL true or false or t or f e d Set whether a netCDF restart is used for the climate model components ea_30 go_18 gs_13 The syntax is n for no netCDF restart file format instead it will be ASCII e e Set the climate model ASCII restart file name namelist values ea_35 go_23 gs_18 The name in each case is rst 1 rst being the default saved restart filename and 1 indicating the first saved restart see above e f Set re start location given by optional parameter 5 The location of the re start file for each module is given by a namelist parameter xx_rstdir_name in the case of the climate model modules and where x
49. n apparent spurious drift in tracer inventories occurring 35 sal min 0 3337E 02 10 35 8 max 0 3891E 02 30 29 8 DIC min 0 1878E 02 35 24 8 max 0 2581E 02 33 21 1 DIC_13C min 4225E 00 3 16 3 max 0 2792E 01 25 13 8 DIC_14C min 1779E 03 33 21 1 max 0 2197E 02 30 29 8 P04 min 0 7071E 07 29 28 8 max 0 3806E 05 3 16 3 02 min 4521E 04 27 30 5 max 0 3363E 03 24 35 8 ALK min 0 2212E 02 10 35 8 max 0 2724E 02 33 21 1 DOM_C min 4159E 05 21 34 3 max 0 1517E 04 32 25 8 DOM_C_13C min 1000E 20 1 3 2 max 0 5817E 01 29 36 8 DOM_C_14C min 1000E 20 1 3 2 max 0 2236E 04 29 36 8 DOM_P min 3924E 07 21 34 3 max 0 1431E 06 32 25 8 Ca min 0 9769E 02 10 35 8 max 0 1136E 01 30 29 8 CFC11 min 0 0000E 00 1 3 2 max 0 0000E 00 1 3 2 CFC12 min 0 0000E 00 1 3 2 max 0 0000E 00 1 3 2 Mg min 0 5050E 01 10 35 8 max 0 5888E 01 30 29 8 F 3 3 How do I do an ocean CO2 injection experiment There is a hard way but maximum flexibility a less hard way and an easy way To cut the shit what follows is the easy way First you want to use the updated tracer forcing format bg_ctrl_force_oldformat false Put this line in the user config file if it is not already there perhaps under FORCINGS section You will need a forcing template for
50. n of each column and allowed values are summarized at the bottom of each file Options include whether a flux or restoring forcing is required the time constant of any restoring forcing etc A set of tracer forcing configuration files with default values for no forcings of any sort live in 16No provision is currently made for restoring forcing 12 genie biogem data input Files to specify alternative forcing configurations will typically be located in sub directories of genie_forcings For a tracer forcing top be fully defined one further set of files are required and also located in sub directories of genie_forcings taking filenames of the format e biogem_force_ _sig dat and contains two columns of information the first is a time marker year and is paired with a corre sponding tracer scalar value in the second column These files all live happily together in the directory defined by the namelist parameter i j The information in the signal file define the time varying information For a model year falling outside of the maximum or minimum time markers specified in the signal file no forcing will be applied this in effect allows a restoring forcing to be turned on and then off again later For model years inbetween specified time points the tracer scalar values are linearly interpolation It should be noted that biogem_force_ _sig dat file must contain at least one row entry The units for the flux forcin
51. n set from specified if any genie forcings NAME A CITETTITTTETTETTTTTE ie CHECK Is restart available if required RUN experiment RE COMPILE RUN make E NEEDED Veg cleanali P ieomple NO SUBMIT TO QUEUE test interactively to check every is OK Figure 1 Schematic of the sequence of events in configuring and running an experiment using the old_rungenie sh methodology 3 4 Job submission to a cluster For submitting jobs with Sun Grid Engine SGE on a cluster E4 qsub S bin bash subgenie sh lt options gt Here the S bin bash part is to ensure that SGE uses the BASH shell to submit the job because this is the language that subgenie sh has been written in Take care that the installed FORTRAN compiler can be seen by the cluster nodes If not the GENIE executable must have already been built prior to submitting a job The easiest way to do this is to run GENIE interactively briefly e g with a run length of just a couple of years or kill it and then submit the full run to the cluster Other useful submission options for SGE e To redirect the standard output stream qsub o genie_log e To redirect the standard error stream qsub e genie_log e To merge standard output and error streams into standard output qsub j y e To specify particular resources such as the nodes with 8 GB of RAM qsub 1 mem_total 8 0G e To decrease the priority of a jor qsub p 1 e To submit
52. nown good set of results files that are held on SVN and installed along with the code in genie genie knowngood 5 At this point the science modules are currently compiled in a grid and or number of tracers configuration that is unlikely to be what you want for running experiments Clean up all the compiled GENIE modules ready for re compiling from the source code by make cleanall 6 Change back to your home directory cd or cd 7 Downloaded the old_rungenie sh script from mygenie seao2 org under GENIE resources MODEL CONFIGURATIONS AND EXPERIMENTAL DESIGNS This file MUST be executable chmod u x old_rungenie sh 8 For running GENIE 1 as described in this manual you will need to create the following directories genie_archive genie_forcings genie_log genie_userconfigs 27 These directories are not an official part of the GENIE distribution nor way of running the model 20 10 11 You will notice that genie_output is created automatically when running GENIE for the first time Downloaded user config file worbe2_preindustrial from mygenie seao2 org under User config experimental design namelist parameter files either just this single file or download and unpack the tar ball of various user config files The user config files live in genie_userconfigs Downloaded the forcing subdirectory worbe2_preindustrial from mygenie seao2 org under Tracer forcing definitions
53. nteger COLUMN END TRACER DESCRIPTION The columns pertinent to the this alternative way of implementing tracer forcings are 01 do you want to restore atmospheric composition toward a specified concentration or isotopic ratio t for true yes for false no Note that it does not matter whether you use upper case capitals T or lower case t If selected column 2 determines the time constant of the restoring 03 if not 1 do you want to apply a flux of gas to the atmosphere 06 what degree of homogenization of the flux do you wish to apply 2 apply flux evenly throughout the atmosphere 0 apply the flux at a single point only 1 use the original more complex flux forcing methodology with the flux or fractional weighting of total flux specified explicitly at each individual grid point If 0 a forcing at a point the next 2 columns 7 and 8 then define the i j location on the model grid longitude i and latitude j of the point source The forcing configuration files for dissolved tracers configure_forcings_ocn dat and particulates configure_forcings_sed dat should be self explanatory again there is a description of each column at the bottom of the file However for dissolved tracer forcings there is an additional column for the k depth level in the ocean location of a flux applied at a single point in addition to longitude
54. oaded from mygenie seao2 org These are the values and settings that GENIE will use if not told otherwise To effect a change in parameter value the parameter name is simply assigned the new value either in the base config used i e one of the config files in genie genie main configs Or more typically when using the old_rungenie sh model job configuration and submission script by appending the new namelist assignment s to the user config file defining the experiment in genie_userconfigs Either way the re assignment generally takes the form namelist value Where the value is a string the syntax is namelist string The syntax for logical flag e g true assignments is namelist true For selecting or de selecting tracers the syntax is gm_atm_select_nn true for an atmospheric tracer gas nn is the integer index of the tracer as detailed in the Tables For ocean and sediment particulate tracers the namelist names take the same form except with ocn or sed in the namelist parameter name If the number of selected tracers in the ocean is changed so to must the value of GOLDSTEINNTRACS which sets the ocean array dimension and the number of tracers that must be advected convected and diffused The value of GOLDSTEINNTRACS must be equal to the number of selected ocean tracers i e the number of gm_ocn_select_nn that are true including temperature T and salinity S The default value is 2
55. on the SVN reposi tory under genie genie main doc All the documentation including this can be generated from the TEX source in both postscript ps and Adobe pdf formats To build documentation from the TEX source from witin genie main doc issue the commandA e make user manual Builds the oritinal GENIE user manual NOW REDUNDANT SEE GENIE WIKI e make genie user manual Builds the biogeochemistry and GENIE 1 focused user manual e make genie tutorial Builds documentation describing various tutorials and example model configurations and experi mental designs e make genie howto The HOW TO documentation potted explanations of how to get stuff done To clean up previously built documentation e make clean Cleans the built user manual documentation files e make clean genie Cleans all the built files for the genie documentation PLEASE edit and update the documentation and populate it with wonderful useful things Spread the love o 1 1 Naming conventions blurb The modular Earth system model is called which stands for Grid ENabled Integrated Earth system model There are 2 primary variants one with a simple 2 D EMBM see below atmosphere referred to as GENIE 1 and one which employs a 3 D dynamical GCM atmosphere called GENIE 2 In addition there are temporary developmental variants that branch off of the primary model development pathway the trunk and which reside
56. page of output looking something like xxx WARNING gt Originating location in code module subroutine gem_carbchem f90 sub_calc_carb gt ERROR MESSAGE Numerical instability at step 0001 Data dum_DIC dum_ALK dum_Ca dum_PO4tot dum_Si02tot dum_Btot dum_SO4tot dum_Ftot dum_H2Stot dum_NH4to t loc_H4B04 1loc_0OH 1loc_HP04 2 0 loc_P04 1loc_H3Si04 1loc_HN3 1loc_HS 1loc_H 1loc_HSO 4 loc_HF loc_H3P04 pH SWS pHfree pHtotal pH OLD pH guess 1 pH guess 2 C02 CO32 HCO3 gt ERROR DATA 1 801730109436933E 007 gt ERROR DATA 1 575704792562387E 004 ds On the other hand appropriate output from BIOGEM looks like depending on the exact options set something rather more like gt gt gt SAVING BIOGEM TIME SERIES year 0 50 493 088 0 000 14 659 4 479 18 815 34 815 gt gt gt SAVING BIOGEM TIME SLICE year 0 500000000000000 gt gt gt SAVING BIOGEM TIME SERIES year 1 50 707 058 0 000 14 659 4 553 18 831 34 815 G 3 Stack space You may encounter issues with regards to the ifort Intel FORTRAN compiler an maybe others particularly when using SEDGEM because of the size of the arrays holding sediment information The Intel Fortran Compilers 8 0 or higher allocate more temporaries on the stack than previous Intel Fortran compilers Temporaries include automatic arrays and array sub sections corresponding to actual arguments If the program is not afforded adequate stack space at runtime re
57. r of internal time steps in the model is calculated based on the run length and length of each internal time step The length of each time step is determined by the value of ma_genie_timestep in seconds and is defined in the config file by ma _genie_timestep 365 25 24 0 500 3600 0 i e ma_genie_timestep 63115 2 giving 500 time steps per yeap Note that also in config file ma ksic loop 5 sets the updating of sea ice to occur every 5 GENIE time steps and ma_kocn_loop 5 set the updating of ocean circulation to occur every 5 GENIE time steps The runtime_defaults sh value of ma_katm_loop 1 is unchanged meaning that the atmo sphere the EMBM in this case is updated each and every GENIE time step Finally ma dt_write sets a default interval of output in multiples of the GENIE time step and is set equal to the total number of time steps ma_genie_timestep to give output writing at the end of the experiment only e c Set the climate model components restart saving frequency iwstp ea_4 go 4 and gs 4 set the frequency of restart saving multiples of the GENIE time step These are set equal to ma_genie_timestep to give re start saving only at the end of the run e d Set climate model components health check frequency npstp ea_3 go_3 and gs_3 set the frequency of health check diagnostics reporting These are set equal to ma_genie_timestep 1 effectively disabling this feature e e Set climate model components t
58. r scalar value in the signal file modifies the difference between the two spatial fields Putting it another way the magnitude of the forcing that is applied at any location i j and at point in time t F is equal to the baseline field defined in biogem_force_ Iuj plus the time dependent modifier S z times the difference between the two fields i e Faja biogem_force_ Taj S t x biogem_force x IIu 5 biogem_force_ Iaj Remember that the value of the time dependent tracer scalar S t is interpolated from the contents of the forcing signal file biogem_force _sig dat Because what I have just written above is probably vanishingly indistinguishable from complete gob bledygook a couple of examples might help or not 1 One way of using the forcing functionality of BIOGEM would be to assign a minimum forcing field to I a maximum forcing field to TI and specify a normalized i e taking values between 0 0 and 1 0 series of modifier values in the the signal file For example one could continually vary the surface temperature field seen by the marine carbon cycle over the deglacial transition based on available end member reconstructions for glacial maximum and modern CLIMAP 1980 and by assuming a semi representative normalized signal with which to interpolate between these two time slice reconstructions see Ridgwell 2001 2 A second way of using the functionality of tracer forcing in BIOGEM would be
59. racers at each and every time step is clearly not practical not only in terms of data storage but also because of the detrimental effect that repeated disk access has on model run time Instead BIOGEM saves the full spatial distribution of tracer properties only at one or more predefined time points years These are called time slices At the specified time points tn a set of spatially explicit data fields are saved for all the key tracer flux and physical characteristics of the system Rather than taking an instantaneous snapshot the time slice is averaged over a specified integration interval At in years defined by the namelist parameter bg_par_data_save_slice_dat 4 The model state is thus integrated from time tn At 2 to tn At 2 For instance setting a value of At 1 0 year results in all seasonal variability being removed from the saved time slices and successive time slices then only reflect long term gt 1 year trends in system state The mid point years tn for which time slices should be saved are specified in a single column pain text ASCII file whose name is specified by the namelist parameter bg_par_inf ile slice_namd The order in which the time sequence is ordered i e ascending or descending time values does not actually matter in practice as long as the list of times is ordered sequentially the list will be internally re ordered if necessary according to the selection of BP the model running backwards
60. t biogem_force_ _1 dat provide maximum flex ibility they are not always so easy to use nor perhaps understand An alternative simply way of specifying forcings has therefore been implemented You can access this by adding the namelist bg_ctrl_force_oldformat false which simply tells BIOGEM to expect a new file format for the tracer forcing configuration files As before the tracer forcing configuration files live in the directory specified by the namelist parameter bg_par_fordir_name e g bg_par_fordir_name genie_forcings worbe2_historical There are 3 files one for each type of tracer i e gas dissolved solid particulate e configure_forcings_atm dat gases e configure_forcings_ocn dat dissolved substances e configure_forcings_sed dat solid substances particulates For instance the format of the tracer forcing configuration file configure_forcings_atm dat looks like 01 02 03 04 05 06 07 08 VV VVWWNW V V START OF DATA F 0 0 F F F 2 01 01 surface air temperature K F 0 0 F F F 2 01 01 specific humidity n a f 0 1 f t F 2 01 01 carbon dioxide C02 partial pressure atm f 0 1 t E 2 01 01 d13C C02 0 00 F 1 0 F F f 2 01 01 oxygen 02 partial pressure atm F 1 0 F F F 2 01 01 d180 02 0 00 F 1 0 F F F 2 01 01 nitrogen N2 partial pressure atm F 1 0 F F F 2 01 O1 d15N N2 0 00 F 1 0 F F F 2 01 01 methane
61. the required parameters 4 have been passed Set local variables derived from passed parameters 2 SET LOCAL FILE AND DIRECTORY NAMES 3 CREATE EXPERIMENT CONFIGURATION FILE Make a copy of the specified template configuration file passed parameter 1 Give the configu 32As outlined earlier the specification of particular flavors unique combinations of science modules are contained in config files with names of the form genie_aa_oo_ss_xxxx config For instance a configuration parameter cal ibration of the 16 level version of the GOLDSTEIN ocean model together with ocean biogeochemistry is defined in genie_eb_go_gs_ac_bg_itfclsd_161_JH config 33The details of any biogeochemical forcings are stored in a directory specified by the value of the namelist parameter bg par_fordir_name 34Refer to the tutorial for a detailed description of this particular model experiment 29 ration file a unique filename by appending the model run ID passed parameter 3 4 SET MODEL TIME STEPPING Set up the time stepping control of model output data saving and experiment run length e a Set run length in BIOGEM bg_par_misc_t_start is the start year which is assumed to be zero by default bg parmisc_t_runtime is the run length years its value is taken from passed parameter 4 e b Based on the run length a consistent overall GENIE run length in terms of the number of internal time steps is set ma_koverall_total the overall numbe
62. tmosphere overlying dry land cells For the ocean ocn_ tracers all the wet cells in of the full 3 D ocean grid of BIOGEM can be forced For the sediment sed_ tracers all the cells comprising the 3 D ocean grid of biogem rather than the 2D SEDGEM grid are manipulated 4 Secondly two different types of forcings of the system are recognized a flux forcing and a restoring forcing The first type the flux forcing is pretty self evident it represents a flux of tracer that is applied to each cell The restoring forcing is less intuitive a flux is applied to the grid cells with a value calculated to bring the tracer values closer to the prescribed restoring value A time scale in years determines the rate at which the tracer values are brought toward the boundary condition This time scale value is termed the restoring constant The smaller the value of the restoring constant the harder the restoring and the more rapidly the model will be constrained to approach the boundary condition Whether a flux or restoring forcing is required together with the details of the forcing and what tracer s is are to be forced are specified in a set of 3 tracer forcing configuration files one for each type of tracer e configure_forcings_atm dat e configure_forcings_ocn dat e configure_forcings_sed dat Each file consists of rows of settings one row for each different tracer The settings are arranged into columns the descriptio
63. to set every wet cell in 13 biogem_force_ _I dat to a value of 0 0 and every corresponding location in biogem_force_ _II dat to a value of 1 0 The forcing applied to the model will then be the same as applying the value of the interpolated tracer scalar in the signal file equally to each and every cell i e Faso 0 S t x biogem_force x aj 0 An example of this usage would be in applying a spatially uniform time varying change in at mospheric composition such as restoring atmospheric CO2 to an observed historical or predicted future atmospheric concentrations trajectory It should be noted that although this scheme is generic and can equally be applied to any tracer atm ocn or sed and including isotopic properties as well as offering reasonable flexibility in representing the time varying characteristics of a boundary condition it does have limitations and cannot cover all possible eventualities For instance historical changes in atmospheric CFC concentrations and CO2 radiocarbon activity vary not only with time but the spatial heterogeneity changes in a complex way that cannot be represented as a interpolation between two alternative end member distributions In this way the built in framework for tracer forcings in GENIE does not conform to the OCMIP protocol REF for ocean carbon cycle model evaluation 4 2 Tracer forcing as a point source or uniform distribution While the spatial forcing fields biogem_force_ _I da
64. ure that everything is correctly compiled 14 With a different queue management environment it may be necessary to place the call to rungenie sh together with its list of parameters into an executable shell and submit that instead 15 The default priority is 0 A lower priority has a higher value 10 For instance either run the experiment for 1 or 11 years say or better start the experiment for the desired full duration but kill it Ctrl C once the experiment is running successfully Remember that if you have just changed model resolution or number of tracers you must force a complete recompile by issuing a make cleanall and then start the model running interactively in a shell 11 4 Specifying time dependent changes to tracer concentrations forcings Even the most highly coupled instance of GENTE is far from being a complete description of the Earth system For instance the lack of a dynamic atmosphere able to perform tracer advection means that dust deposition to the ocean surface cannot be internally calculated With respect to longer term biogeochem ical cycling the model currently lacks any explicit representation of shallow water carbonate deposition in banks and reefs In terms of global change and anthropogenic impacts on carbon cycling and climate GENIE is also unable to predict fossil CO2 emissions from human activities just how much do you want to burn in the future Provision is therefore made for pr
65. ween the different tracers tracer_define atm tracer_define ocn tracer_define sed The contents of these files must not be altered nor the files moved unless there is a very good reason to do so Since there never will be a very good reason HANDS OFF e genie genie main src fortran cmngem Source code for the geochemistry library modules common routines constants and parameters that need to be accessed by all biogeochemistry modules i e ATCHEM BIOGEM SEDGEM and ROKGEM The source code files are gem_carbchem f90 Definition and solution of aqueous carbonate chemistry routines for solving isotopic fractionation routines for estimating DIC and ALK given other parameters gem_cmn f90 Parameter and constant definitions gem_data f90 Read in namelist parameters gem_netcdf f90 Low level netCDF file I O routines gem_util f90 Miscellaneous routines such as for reading and writing ASCII file formats converting between isotopic notations numerical solution and multi dimensional linear inter polation initialization of tracers and inter tracer relationships 29The alternative is to estimate wind speed and hence calculate air sea gas exchange from the wind stress field ap plied to the GOLDSTEIN ocean circulation model The wind stress based reconstruction can be chosen by setting bg_ctr1_force_windspeed false 30Would more logically live in initialise_atchem f90
66. word assigned If not get one You will not get very far without any code 1 Log into your computing account 2 Check out the trunk source code for GENIE From your home directory 7 svn co http source ggy bris ac uk subversion genie trunk username YOURNAME genie all one line where YOURNAME will be your username for the SVN code repository You will be asked for a password 3 Change directory to genie genie main and make This compiles the default configuration of GENIE It serves to check that you have the software environment correctly configured If you are unsuccessful here too bad Try editing user mak user sh which are located in genie genie main and which set the environment 4 Next make assumedgood This creates a gold standard set of experimental results for GENIE 2 against which your version of GENIE 2 can subsequently be check to ensure that you have not broken it in any way These results cannot be stored centrally on SVN because GENIE 2 incorporating a dynamical atmosphere is chaotic Now make test These are short tests that check against your newly created assumed good experimental results You should get a pass here Followed by make testbiogem This is a short experiment using GENIE 1 and incorporating BIOGEM Alternatively make testebgogs will run a EMBM based climate only test Because the results are not chaotic the test compares the result against a k
67. x is the module abbreviation ea go gs and yy par rstdir_name in the case of the biogeochemical model modules with yy being the module abbreviation ac bg sg rg If restarts are not to be used then ea_7 go_7 gs_7 are all set to n and ac_ctrl_continuing bg ctrl_continuing sg_ctrl_continuing rk_ctrl_continuing are all set to false 7 APPEND EXPERIMENT SPECIFIC NAMELIST CHANGES Append the contents of the user configuration file specified by the run ID passed parameter 3 and its directory location passed parameter 2 to the basic flavor and configuration file parameter 1 But first in case one has a Windoz user infestation the namelist file is conditioned to avoid possibility of carriage return line feed screw upd 4 8 GO Run the model Change directory and from genie genie main invoke genie_example job f configs genie_eb_go_gs_ac_bg worbe2_preindustrial_1 config 9 CLEAN UP Move the configuration file created to the results directory in genie_output and archive entire 36 A string describing the full path filename see page 37This step can be omitted commented out deleted in the event that no Windoz users are involved 31 results directory of the run by tar cfz genie_archive MODELID RUNID tar gz 0UTPUTPATH This line will need to be edited if the instaled directory structure differs from the default assumed in this manual Or this line can simply be commented out if nto required

Download Pdf Manuals

Related Search

Related Contents

Manuale d`uso - eBook Reader Pyrus 2 LED  VacIon 10 l/s VacIon Pump User Manual  Bétonnière 190 L  Method and apparatus of synchronizing two computer systems  Kollmann - Test Equipment Depot    Overview Features System Requirements Application - A  Checklist – tudo o que precisa fazer antes da  

Copyright © All rights reserved. Failed to retrieve file