Force a read of 'forcing' streams on init

[matthewhoffman]

Previously, input files used as time-varying forcing failed to be read
at the initial model time. This commit forces input streams with an
input_interval set to a time duration to be forced to be read at the
model initial time. They are read in the same we these "forcing" files
are read during timestepping by using the most recent previous time
level from the input file.

[trhille]

I tested this on Cori with the Humboldt Glacier r04 mesh and forcings. The code hangs before all input streams are read.
This is from the end of log.landice.0000.out:

WARNING: Variable basalWaterThickness not in input file.
WARNING: Variable waterFrac not in input file.
WARNING: Variable eigencalvingParameter not in input file.
WARNING: Variable stiffnessFactor not in input file.
WARNING: Variable calvingMask not in input file.
WARNING: Variable upliftRate not in input file.
WARNING: Variable effectivePressure not in input file.
WARNING: Variable basalFrictionFlux not in input file.
WARNING: Variable damage not in input file.
WARNING: Variable calvingVelocityData not in input file.
WARNING: Variable ismip6shelfMelt_basin not in input file.
WARNING: Variable ismip6shelfMelt_gamma0 not in input file.
WARNING: Variable ismip6shelfMelt_3dThermalForcing not in input file.
WARNING: Variable ismip6shelfMelt_deltaT not in input file.
WARNING: Variable ismip6shelfMelt_zOcean not in input file.
WARNING: Variable ismip6shelfMelt_offset not in input file.
WARNING: Variable ismip6Runoff not in input file.
WARNING: Variable ismip6_2dThermalForcing not in input file.
WARNING: Variable xtime not in input file.
WARNING: File /global/cfs/cdirs/piscees/MALI_input_files/Humboldt_1to10km_r04/Humboldt_1to10km_r04_20210617.nc does not contain a seekable xtime variable. Forcing a read of the first time record.
WARNING: Variable xtime not in input file.
WARNING: File /global/cscratch1/sd/trhille/Humboldt_1to10km_r04_20210617/Humboldt_1to10km_r04_20210617_muFriction_m7.nc does not contain a seekable xtime variable. Forcing a read of the first time record.

[matthewhoffman]

@trhille , thanks for testing this. Can you send me the path to the full log and the streams file used?

[trhille]

@trhille , thanks for testing this. Can you send me the path to the full log and the streams file used?

/global/cscratch1/sd/trhille/Humboldt_1to10km_r04_20210617/m7/MIROC5/testPR10

[trhille]

Testing after 6301c56 produced desired behavior and code did not hang. From log.landice.000.out:

` Finished reading 'initial_only' input streams.
Looking for recurring input streams (forcing) that should be forced to be read at the initial time.

• Forced an initial read of input stream 'ismip6sfcMassBal' from time: 2007-07-01_00:00:00
• Forced an initial read of input stream 'ismip6surfaceAirTemperature' from time: 2007-07-01_00:00:00
• Forced an initial read of input stream 'ismip6_runoff' from time: 2007-07-01_00:00:00
• Forced an initial read of input stream 'ismip6_2dThermalForcing' from time: 2007-07-01_00:00:00
• Forced an initial read of input stream 'ismip6shelfMelt_3dThermalForcing' from time: 2007-07-01_00:00:00
  Finished processing recurring input streams at initial time.
  Using FO velocity solver.
  Initializing external velocity solver.
  Initializing external velocity solver grid data.
  Calculating N assuming perfect ocean connection.
  Initial timestep 2007-07-01_00:00:00
  Generating new external velocity solver FEM grid.
  Beginning velocity solve using external velocity solver.
  Completed velocity solve using external velocity solver.
  Starting timestep number 1
  Setting time step (days) to: 100.000000000000
  Starting face melt routine
  calling li_apply_front_ablation_velocity from grounded_face_melt_ismip6
  == Ablation complete. Total calved = 661178359.496096
  == Ablated from non-dynamic cells = 92101053.4246684
  == Ablated from dynamic cells = 263517186.514981
  == Stranded floating cells deleted = 361.393344835829
  == Unablated volumes: Non-dynamic cells=0.00000000000000; Dynamic cells=0.00000000000000
  Finished with li_apply_front_ablation_velocity
  basin=1, mean_TF=1.32037195650461
  iceshelf_melt_ismip6: gamma0=74628.5200000000
  Calculating N assuming perfect ocean connection.`

[trhille]

Testing produces desired results. Ready to merge.