HWRF  trunk@4391
Running HWRF

This page and its subpages explain how to run the HWRF model, track what it is doing during the execution, and figure out what is going wrong.

Note that the multi-storm HWRF has its own special page with details on this matter:

Running HWRF

There are several ways to run the HWRF model:

The direct methods are more laborious than automation systems. However, direct methods are more useful for debugging if one is developing new capabilities in HWRF, or porting it to new supercomputers. For details on each method, visit one of the three above links.

For information about HWRF wrappers, see the HWRF User's Guide on the public HWRF website: http://www.dtcenter.org/HurrWRF/users/


Monitoring HWRF

After you start the Rocoto-based or ecFlow-based workflow, a sequence of jobs will run, and some may fail, requiring user intervention. In order to know this, you must monitor the progress of the HWRF forecast cycles. Both ecFlow and Rocoto have means by which to check which jobs are queued, submitted, completed or failed. See the HWRF Rocoto Workflow page for details on doing this in Rocoto.

HWRF Directory Structure

Knowing where data will show up can tell you a lot about what is going on in HWRF and why. The HWRF system has several key directories:

For great detail on the HWRF directory structure:

Todo:
Add HWRF directory structure page.

Detailed Logging

Logging in HWRF is quite extensive. For details on HWRF logging, we refer you to an entire page on the matter: