[RegCNET] IMPORTANT - "BUG" FOUND IN REGCM3 - PLEASE READ

[FILIPPO GIORGI]

Dear RegCNETers

With this email I wanted to alert you that, thanks to Jonathan, we have 
found a bug in the current RegCM code. If you have it in your code, you 
need to correct it.

The bug is in the calculation of the eccentricity in subroutine getdat.f.

More specifically, the line of code

       calday = dble(julday)+mod(NNNNNN-NSTRT0,4)/4.+(xtime/60.+gmt)/24.

should be replaced with

       calday = dble(julday)+(NNNNNN-NSTRT0)/4. + (xtime/60.+gmt)/24.

The bug sets the eccentricity at the value of the initial day (idate0 in 
regcm.in) and then does not change it any more.

So, for example, if you started your run on Jan 1, the eccentricity will 
remain that of Jan 1 for the whole run. The eccentricity affects the value 
of solin, i.e. the incoming solar radiation at the top of the atmosphere. 
This values varies by about 6-7% from a maximum in January to a minimum in 
July. So, depending on when you start your simulation and how long it is, 
you might have an error of up to 6-7% in the incoming solar radiation. 
This may be more or less important for your simulation, but you need to 
change it and check older runs. If you completed only short runs (say 
3-months or so), the bug should not affect your simulations much. Also, in 
the winter we do not expect a big effect. The largest effect is for the 
summer climate in a long run started in January, when summer solar 
radiation is too high by up to 6-7%.

Unfortunately, we are not sure of when this bug was put into the model, 
but we suspect it has been there for some time, probably since a late 2004 
version. It also seems that it is not in all versions of the model that 
have been circulating since then, both serial and parallel versions, so 
you need to check yours. In fact, if you have any idea of why, when and 
how this code change was put in, please let us know, so that we can try to 
figure out what happened.

Note that the code in the public website has been already updated and 
corrected, so you can also check there how to remove the bug.

In the process of checking this, we also found that, for reasons too long 
to explain, you should always start your run on 00:00 GMT, otherwise the 
model has some problems in calculating the declination angle with the 
proper timing. Fortunately the model routinely starts at 00:00 GMT, but 
anyways you should be aware of this, especially because we have not been 
able to correct this problem (we are working on it).

It is unfortunate to find such bugs, but it happens. It is part of 
the modeling life, although certainly not a nice part.

Should you have any questions, or any comments that might help us 
understand what happened, do not hesitate to contact me, Bi, or Sara.

best regards,   Filippo Giorgi

################################################################
# Filippo Giorgi, Head                                         # 
# Earth System Physics Section                                 #
# The Abdus Salam International Centre for Theoretical Physics #
# P.O. BOX 586, (Strada Costiera 11 for courier mail)          #
# 34100 Trieste, ITALY                                         #
# Phone: + 39 040 2240 425                                     # 
# Fax: + 39 040  2240 449                                      #
# email: giorgi at ictp.it                                        #
################################################################