DYCOMS II SCM results: Martin Köhler, ECMWF

15 Oct 2003

Here are some first results using the ECMWF SCM.

Setup

I had to do several minor adjustments to the DYCOMS LES setup listed below. I am recommending them to other DYCOMS SCM modelers:

1. Initial state:

   Bjorn provided us with profiles of z, p, T, qt and ql from 8m to 30000m, which he used for the radiation calculation. I have trivially extrapolated them to 0m and 100000m. When running the ECMWF SCM, the level from 7 to 11km immediately condensed into an everlasting cirrus. This might be because of different qsat formulas used by Bjorn and ECMWF. To prevent that upper level condensation from happening, I have moved the gradual decay of qv from originally 7600m - 13000m to 5100m - 8100m. The updated inistial state file can be found here.

   Note, that you have to specify a surface pressure of 1016.31hPa in your SCM consistent with the initial state table and not the 1017.8hPa specified in the LES setup paper.

2. Nocturnal for ever:

   The case is designed to investigate nocturnal stratocumulus. The desire for SCM simulations will be to run the case for longer than the 4 hours postulated in the LES case (e.g. 2 or even 10 days). Yet, I suggest to run it with solar radiation off for the whole duration. For the stratocumulus diurnal cycle one should refer to the EUROCS FIRE case.

3. Location & Time:

   Longitude=31.3N, latitude=121.7W, initial time=6UTC, 10 July 2001 (see DYCOMS BAMS paper).

4. Horizontal velocity:

   As initial condition, I used the geostrophic wind (u=7m/s & v=-5.5m/s) throughout the atmosphere.

5. Vertical velocity:

   Divergence D is set at 3.75 x 10^-6 s^-1 appropriate for lower tropospheric levels (I suggest 0-5000m). Further aloft this has to be balanced with a convergence. I suggest an equivalent convergence for 5000m-10000m.

I am attaching the ECMWF SCM input file in netcdf and postscript format.

Preliminary results

This simulation uses the ECMWF SCM CY25R3 (operational cycle Jan-Oct 2003). The results can be browsed in its entirety as netcdf or postscript with a few being replicated below.

The main feature is the decaying cloud water and intermittend regeneration through convective updrafts after about 30 hours. Cloud fraction reduces quickly to 50%. The radiative cloud top cooling results in a reduced cloud top entrainment. The reduced radiative cooling allows the PBL to warm (see theta profile) to approach the SST. The reduced entrainment allows the PBL to moisten. Reduced surface gradients result in a decay in sensible and a 50% reduction in latent heat fluxes.

Liquid water mixing ratio (red = 0.2 g/kg)
Liquid water path [kg/m^2]
Water vapor mixing ratio [kg/kg]   Time: 0 to 48h from black to red. Thick line is mean.
Potential temperature [K]   Time: 0 to 48h from black to red. Thick line is mean.
Convective flux of dry static energy [W/m^2]
Surface sensible heat flux [W/m^2]