Tutorial 2: Run a forecast

There are two options to initialize a forecast.

Initialization from sounding profiles, i.e. defined profiles of temperature, humidity and wind, from wrf_profile files (see WRF guide). Only usable for WRF’s idealized mode.
Initialization from an previous forecast, i.e. from WRF restart files - optionally with updates from data assimilation.

a) Initialization from sounding profiles

It is necessary to set the path to the prepared WRF input soundings as argument to dartwrf.utils.Config.

input_profile = '/jetfs/scratch/wrf_profiles/raso.fc.<iens>.wrfprof'

where <iens> is a placeholder, since there is one file for each member, from raso.fc.001.wrfprof to raso.fc.040.wrfprof.

The input profiles are then copied into the WRF ensemble directories by w.prepare_WRFrundir(cfg). Finally, with w.run_ideal(cfg), WRF’s ideal.exe program is called for all ensemble members to create initial condition files, called wrfinput_d01, for each member.

Now we can go to step 3 to run the forecast.

b) Initialization from a previous forecast

Let’s say you want to run a forecast starting at 9 UTC until 12 UTC. We use initial conditions of a previous experiment /user/test/data/sim_archive/exp_abc which was initialized at 6 UTC and there are WRF restart files for 9 UTC. Documentation can be found at dartwrf.workflows.WorkFlows.prepare_IC_from_prior().

import datetime as dt
cfg.update(
    time = dt.datetime(2008, 7, 30, 6),
    prior_init_time = dt.datetime(2008, 7, 30, 6),
    prior_valid_time = dt.datetime(2008, 7, 30, 9),
    prior_path_exp = '/user/test/data/sim_archive/exp_abc/',)
w.prepare_IC_from_prior(cfg)

Now, we are ready to run the forecast, or update the initial conditions with increments from assimilation.

Optional: Update posterior with increments from assimilation

To continue a forecast after assimilation you need the posterior = prior (1) + increments (2)

w.prepare_IC_from_prior(cfg)
w.update_IC_from_DA(cfg)

The second step updates the initial conditions, i.e. the WRFrst files (that were copied to the run_WRF directories), with assimilation increments at time. Now we can run the forecast ensemble.

Run the forecast

Set important parameters,

WRF_start: start time of the WRF simulation
WRF_end: end time of the WRF simulation
hist_interval_s (int): history output frequency in seconds
restart: whether it uses a restart file
restart_interval: interval in minutes to write restart files

and run

cfg.update( WRF_start=t,
            WRF_end=t+timedelta_integrate,
            restart=True,
            restart_interval=restart_interval,
            hist_interval_s=300,
)
id = w.run_WRF(cfg)

More documentation is in the docstring of dartwrf.workflows.WorkFlows.run_WRF().