This description is a short summary (extract) of the original case description by Andy Brown. Please read the original extended version when you would like to know more about the case and the measurements on which the case is based. Also read the specific instructions.

Setup Specifications

Note that when times are given in UTC they are in the form HH:MM e.g. 11.50 is eleven fifty (rather than half past eleven). However, for many purposes (e.g. specification of time-varying surface and large-scale forcings) it will be more convenient to use time given in seconds after 00.00 UTC on 21$^{st}$June.

The runs start at 11.30 UTC, or t = 41400 s.

The surface pressure should be set to 97000 Pa, and 100000 Pa should be used as the reference pressure for converting between potential temperature and temperature i.e.

$$T = \theta {\left( {P(z) \over 100000} \right)}^{R/C_p}$$

Initial Conditions

The initial potential temperature and total water content values to be used are given in Table 1. Linear interpolation should be used to obtain values at intermediate heights.

Table 1: Initial profiles of $\langle \theta \rangle$, $\langle r_T \rangle$, $\langle u \rangle$ and$\langle v \rangle$.

z (m)	$\langle \theta \rangle$ (K)	$\langle r_T \rangle$ (g/kg)	$\langle u \rangle$ (m/s)	$\langle v \rangle$ (m/s)
0.0	299.00	15.20	10.0	0.0
50.0	301.50	15.17	10.0	0.0
350.0	302.50	14.98	10.0	0.0
650.0	303.53	14.80	10.0	0.0
700.0	303.70	14.70	10.0	0.0
1300.0	307.13	13.50	10.0	0.0
2500.0	314.00	3.00	10.0	0.0
5500.0	343.20	3.00	10.0	0.0

As well as ensuring that all models start with the same potential temperature profiles, it is important to ensure that they diagnose similar temperature profiles (as the diagnosed saturation mixing ratio depends strongly on temperature, changing by approximately 1 g/kg per degree at 700 m). For a quick check, my model gives p = 89658 Pa at 700 m, and p = 72584 Pa at 2500 m. This gives initial temperature values of 294.4 K and 286.5 K at these two heights. Please check that your values are close to these.

In the absence of detailed estimates of the time and height variation of the geostrophic wind, and of any large-scale advective tendencies of the wind components (see later), there seems to be little point in attempting to closely match the observed wind profiles. Accordingly, for simplicity, set the initial horizontal wind ($\langle u \rangle ,\langle v \rangle$) to (10,0) ms$^{-1}$ at all levels, as indicated in Table 1.
 

Surface boundary condition

It has been decided that the surface scalar fluxes should be imposed in the intercomparison run. Table 2 gives values of the fluxes at a number of different times, and linear interpolation should be used to obtain the values to impose at intermediate times.

Table 2: Surface sensible ($H$) and latent ($LE$) heat fluxes to be imposed in model runs. Linear interpolation should be used to obtain the fluxes between the specified times. Please note that the specified times here are not the same as those used in specifying the large-scale forcing.

Time (s)	H (Wm$^{-2}$)	LE (Wm$^{-2}$)
41400	-30	5
55800	90	250
64800	140	450
68400	140	500
77400	100	420
86400	-10	180
93600	-10	0

The surface roughness length should be set to $0.035$ m (a characteristic value for the ARM site).

Large-scale forcing and radiation

Table 3 gives the values of $A_{\theta}$, $R_{\theta}$$A_{qt}$ which are to be used to scale the values of the large-scale advective tendency of $\theta$, of the large-scale advective tendency of $q_t$, and of the radiative tendency to be applied to $\theta$. These values are given at various times, and linear interpolation should be be used to obtain values at intermediate times.

Table 3: Magnitudes of large-scale advective forcing and radiative tendencies to be applied in lowest 1000 m (note units!). Reduced tendencies are to be applied above 1000 m, as described in the text. Linear interpolation should be used to calculate forcings at intermediate times. Please note that the specified times here are not the same as those used in specifying the surface fluxes.

Time (s)	$A_{\theta}$	$R_{\theta}$	$A_{qt}$
	(K/hour)	(K/hour)	((g/kg)/hour)
41400	0.000	-0.125	+0.080
52200	0.000	0.000	+0.020
63000	0.000	0.000	-0.040
73800	-0.080	0.000	-0.100
84600	-0.160	0.000	-0.160
93600	-0.160	-0.100	-0.300

Tendencies should then be applied to $\theta$ and $q_t$ as follows.

For $z<1000$ m,

$$\frac{\partial \theta}{\partial t}=ADV_{\theta}+RAD_{\theta}=A_{\theta}+R_{\theta}$$
$$\frac{\partial q_t}{\partial t}=ADV_{qt}=A_{qt}$$

For $1000$ m $\le z<3000$ m,

$$\frac{\partial \theta}{\partial t}=ADV_{\theta}+RAD_{\theta}......eft(A_{\theta}+R_{\theta}\right)\left(1-(z-1000)/2000 \right)$$
$$\frac{\partial q_t}{\partial t}=ADV_{qt}=A_{qt}\left(1-(z-1000)/2000 \right)$$

and for $z \ge 3000$ m,

$$\frac{\partial \theta}{\partial t}=ADV_{\theta}+RAD_{\theta}=0$$
$$\frac{\partial q_t}{\partial t}=ADV_{qt}=0$$

No estimates of the geostrophic wind are presently available. Noting that the observed initial wind speeds above the boundary layer are approximately 10 ms$^{-1}$, it is proposed to set the geostrophic wind to (10,0) ms$^{-1}$. The Coriolis parameter should be set to $8.5\times 10^{-5}$ s$^{-1}$, as appropriate for a latitude of$36 ^o$N.
 

Acknowledgements. We would like to thank Andy Brown for providing us with the text of the original case description.