How to include a time-varying solar cycle in WACCM

[gchiodo@fis_ucm_es]

Dear all,

I want to use WACCM-v3.1.9 to make a future climate simulation (2000-2100) with a prescribed solar cycle. Basicly, my experiments would be similar to the CCMVal REF-B2 simulations, but including some "random" solar variability.

Unfortunately, I don't know exactly how to use a time-varying solar cycle as input for WACCM (v 3.1.9). I have been browsing for a bit inside the code and the Marsh et al. (2007) paper to figure out how to do this, but I am still not sure, so please correct me if I'm wrong.

For photolysis calculations, WACCM needs f10.7 and f10.7a, as on eq.(1) and (2) of Marsh et al. 2007

So if a time-varying cycle is to be used, these parameters must be a function of time (as in wasolar.nc, for the REF-B1 simulations?)
How can this external parameter be used as input?
Is this the "solar_parms_file" namelist variable?

In the default namelist version, a single value in "wa_smin.nc" (77) is used (I think you insert a single value if the model is forced with fixed solar min /max, right?), so do I just need to insert a time-varying f107, and WACCM will use it as input for eqs.(1)(2)(5)(6)(7) of Marsh et al.?

Additionally, Marsh et al. says that WACCM needs the TSI value to compute the scaling factors (5) (6) in the stratosphere and troposphere (below 65 km), and minimize the difference between TSI and the integral spectral irradiance (1)(2) and R350. So, TSI must also be used as a time-varying variable in the namelist variable "bndtvscon"?

For my simulation, I only have a table with TSI values from 2000 to 2100 in a yearly resolution.
How could I use it as input?
Should I interpolate these values to daily resolution (assuming the same observed amplitude for 27 day rotational period), and use it as "bndtvscon" (and set scon='RAMPED') in the namelist?
What should I do with the f107 parameter? Am I supposed to somehow deduce it from my TSI values, and insert it as en external file for "solar_parms_file"?

Many thanks for your help!

 

[marsh]

Hello,

To put in a time-varying solar input, all that is needed is to replace wa_smin.nc with a new file that has time-dependent parameters. If you dump wa_smin.nc you will see:

f107 = 77.15, 77.15 ;

f107a = 77.15, 77.15 ;

kp = 2.7, 2.7 ;

ap = 12, 12 ;

isn = 0, 0 ;

date = 18000101, 22000101 ;

WACCM interpolates between the dates in the file, so in this case it always returns the same solar parameters since the endpoints in this two date file are the same. So, to add time dependence create a new file with a timeseries into the future at the temporal frequency you require. The date format is YYYYMMDD. You have the option of fixed geomagnetic -- repeat the same value of Ap for each day, or generate a timeseries.

You do not need to supply a TSI.

I hope that helps,

Dan

 

[gchiodo@fis_ucm_es]

Thanks for the message.

I'm planning to use the new WACCM version 3.5.48, instead of the 3.1.9.

Does this procedure (by only using the f107 as proxy for solar output variations) also apply to that one?

 

[gchiodo@fis_ucm_es]

I'm doing a PhD on the solar-cycle in WACCM simulations. I am currently planning to carry out a few new WACCM-3.5
simulations which, in view of the forcings we want to introduce, will
mostly resemble the CCMVal-2 REF-B2 simulations (aside from a prescribed
time-varying solar forcing).

The simulation period will be 2000-2100, so we would need the initial
conditions for the year 2000. Is it possible to get them for the 1.9x2.5
resolution?

We need to drive the model integrations with the following
forcings:

1) The A1B scenario for future CO2 concentrations.
2) The A1 WMO-halogen future scenario.
3) The SSTs and SICs from coupled ocean-atmosphere models.
4) The ozone and aerosol precursors, according to the B2 baseline IIASA
scenario through 2100.
6) A prescribed future solar forcing scenario

Would it be possible to get the nc-files for the first 4 forcings?
Unfortunately, I could not find any of these input data on the public
WACCM directory.

As regards the QBO and solar cycle forcing, do we need to change the
dates for the observed QBO and f107-series in the nc-file, and
WACCM-v3.5 will work with them for a future climate simulation? Is the
f107 index (along with Kp) the only needed input for solar variability
in WACCM-3.5, as in the older 3.1 version?

If I want to run the model under an hypothetical new Mounder-Minimum
(thus with significantly lower TSI values than now), do I just need to
convert the TSI values into f107, or is the TSI value also needed for
the scaling in the new CAM 3.5 SW-radiation scheme?

Moreover, I also planned to carry out an
additional sensitivity study on the observed solar input in the
UV-spectrum. Is it possible to change the spectral properties of the
solar irradiance in WACCM3.5, even though the integrated TSI remains constant?

Many thanks in advance!

 

[marsh]

As regards the QBO and solar cycle forcing, do we need to change the dates for the observed QBO and f107-series in the nc-file, and WACCM-v3.5 will work with them for a future climate simulation?

The model will crash if you run it past the dates in the file. If you need a repeating QBO series it can be provided. For f10.7 / kp - you can use a fixed file with just two data points with dates very far apart (e.g. years 100 and 2500), or write some code to create a file with repeating f10.7 / kp into the future from a reasonable period (e.g. the last 4 solar cycles). The file for f10.7 / kp from the REFB1 run could be used for this -- just add on to the end of it.

Is the f107 index (along with Kp) the only needed input for solar variability in WACCM-3.5, as in the older 3.1 version?


No, solar spectral irradiance is also needed. We use the Lean spectral irradiance. In REFB1 it varied, in B2 I believe it was constant. Again, you can create a future varying input by copying the last solar cycles forward in time. The spectral irradiance files can be made available.

If I want to run the model under an hypothetical new Mounder-Minimum (thus with significantly lower TSI values than now), do I just need to convert the TSI values into f107, or is the TSI value also needed for the scaling in the new CAM 3.5 SW-radiation scheme?


You would need to create f10.7/kp and spectral irradiance files for these conditions. I suggest taking a look at the ssi data on the SOLARIS website, which goes back to 1610. For f10.7 you could pick a number representative of solar minimum (~70).

Moreover, I planned to carry out an additional sensivitity study on the observed solar input in the UV-spectrum. Is it possible to change the spectral properties of the solar irradiance (even though the integrated TSI remains constant)?

I'm not sure what you are trying to do here - WACCM does not use TSI, just f10.7 and spectral irrandiance. You could vary either to test the effects of changing UV (spectral irradiance) or EUV (f10.7).

I hope that helps.

 

[gchiodo@fis_ucm_es]

Thanks for your help.

I'm currently about to start these future runs, though this time with the new CESM / WACCM-4 version. These runs will be started as a continuation-run from a present-day transient simulation.

I am still wondering how to get the right spectral solar forcing for these future runs; for this reason, I have been checking out the implementation of solar variability in the code, and found out that, as it was the case in the previous version 3.5 of WACCM, the SSI is scaled to the reference_SSI in order to calculate a spectral variability factor. In this way, the solar radiative fluxes in the radSW module are scaled with this factor, and the spectral variability (which is high in the UV) is taken into account.

However, I do not understand why such reference_SSI is taken from the mean of the SSI between the cycles 8 and 10 (around year 1850) in the external file "spectral_irradiance_Lean_1950-2140_daily_c100804.nc" (see the notes with ncdump).
This nc-file is used for present-climate simulations. I assume that this reference_SSI is consistent with the reference_TSI (1360.8), which is also provided in this nc-file. Nevertheless, such reference_SSI of the year 1850, which is lower than the SSI of 1950-2000 at most of the time-steps, would just scale down all radiative fluxed in radSW at any time-step in a present-day simulation?

Could you please explain why the reference_SSI is taken from the cycles 8-10, and not from cycles 19-23?

Thanks a lot,
Regards

 

[gchiodo@fis_ucm_es]

I am also going to carry out some equilibrium simulations with WACCM-3.5, which means perpetual solar MAXIMUM conditions and perpetual solar MINIMUM conditions.

In such a set-up, what should the "ref_flx" --> ref_photon_fluxes (and reference_TSI) be like? Should It be the same as the actual photon_fluxes (so, ref_flx=sMAX_photon_fluxes and ref_flx=sMIN_photon_fluxes), or should it be taken as a "mean reference state", which could be the mean between cycles 19 and 23 (so, ref_flx=mean-photon_fluxes_sc19-sc23)?

Many thanks for your help!

 

[mmills@ucar_edu]

Gabriel,

The tsi_ref and ssi_ref fields record the values that were used in the pre-industrial (1850) control run. These values should remain the same in all solar input files for all WACCM4 runs. They record the baseline to which the model was tuned, and from which differences are compared.

If you look at the corresponding file for our 1850 WACCM compset (spectral_irradiance_Lean_1850_cntl_c100407), you will see that there are 2 dates (101, 25000101) that bracket all dates in the run. Both dates have the same tsi (1360.88677126701) and ssi wavelength variability. Dates in between are interpolated, so will have these same values.

In the 1850 case, the tsi and ssi are the same as tsi_ref and ssi_ref. For a solar maximum or minimum case, you should leave tsi_ref and ssi_ref the same, and change the tsi and ssi for the begin and end dates in your new file, so that both represent values appropriate for solar max or min.

Cheers,
Mike