Surface layer set at ~15m

[mvieno]

Hi,
how small can the surface layer in the EMEP model. I've done a quick test with WRF with an additional ~15m layer at the surface (see the EMEP log below).

1 10000.00 Pa 16179.71 m
2 13600.00 Pa 14205.09 m
3 17200.00 Pa 12701.53 m
4 20800.00 Pa 11487.02 m
5 24850.00 Pa 10345.22 m
6 30250.00 Pa 9042.60 m
7 37180.00 Pa 7623.86 m
8 45280.00 Pa 6216.14 m
9 53290.00 Pa 5012.83 m
10 60580.00 Pa 4039.37 m
11 67240.00 Pa 3229.93 m
12 73270.00 Pa 2551.40 m
13 78670.00 Pa 1981.20 m
14 83440.00 Pa 1503.36 m
15 87580.00 Pa 1106.27 m
16 91090.00 Pa 781.36 m
17 93970.00 Pa 522.27 m
18 96220.00 Pa 324.30 m
19 97840.00 Pa 184.07 m
20 98920.00 Pa 91.63 m
21 99370.00 Pa 53.35 m
22 99820.00 Pa 15.22 m
23 100000.00 Pa 0.00 m

In the EMEP guide it advise against it:
"NB: it is important not to define a lowest layer thinner than about 45 meters; the deposition scheme will fail if the middle of the lowest layer is smaller than the highest defined vegetation"

I've only did a quick test for two months and EMEP didn't complain. Is ok to use this setup?

many thanks
Massimo

[mifads]

There are some issues. One is that the model uses a tile approach, with an assumption that the centre of the grid-cell represents the blending height in some loose way. In this case you are assuming that the O3 over a forest at 7m is the same as the O3 over a lake, which is not realistic. Now, we have some complex "2-layer" calculation in DryDep which attempts to account for that, but I am not sure it applies to such thin layers.

Also, is it reasonable to have grid cells of perhaps 1-5km wide and 15m deep? A horizontal transport of ca. 500m would usually disperse (rule-of-thumb 10%) through ca. 50m, whereas you are trapping it under a very thin ceiling. Lots of O3 titration I would think...?!

Then there are numerical issues and that sentence you quote. I can't answer without digging into the code, and don't have time for that right now. I'll have to leave you to check n test ;-)

[mvieno]

Hi,
I am running a full test now (well two months). In principle I fully agree with what you say, but I see lots of ATCM applications in literature where the surface layer is <=20m and I was wondering how the results may change in EMEP if I use a thinner layer. I'll let you know what I find.