This page explains how to modify the *.conf files in the HWRF parm directory, and explains the purpose of the various conf files. There are a number of subpages with more detailed information:

Todo:: Add input source and WRF configuration pages.

The *.conf files configure various aspects of the HWRF system. Users can also override these configuration settings manually on the command line, or programmatically in the hwrf_expt.prelaunch() function.

When the workflow is started, the selected list of conf files are read in sequence by the hwrf.launcher.launch() function during the scripts.exhwrf_launch job. That creates the storm1.conf file for the cycle. Each job thereafter reads the storm1.conf in the hwrf.launcher.load() function. That means configuration settings changed in the HWRF install directory do not apply to cycles that were already launched.

These are the standard conf files read in by all workflows, in the order they are read in:

File	Purpose
parm/hwrf_input.conf	Default input file paths and names on all clusters.
parm/hwrf.conf	Detailed configuration of most aspects of HWRF.
parm/hwrf_holdvars.conf	For creating the storm1.holdvars.txt
parm/hwrf_basic.conf	Basic configuration settings.
parm/system.conf	System-specific configuration.

The first four files, hwrf_input.conf, hwrf.conf, hwrf_holdvars.conf and hwrf_basic.conf, should not need to be modified unless one is making extensive changes to the HWRF, or adding new functionality to the underlying scripts. Instead, one can specify additional configuration files that override the default settings. See HWRF Experiment Configuration for a list of alternate HWRF configurations, and how to enable them.

The system.conf should be modified manually. In fact, it does not even exist in the repository. One must create it by copying one of the existing system.conf.* files and modifying to suit the user.