Ray:

I think I have finally solved the "Te problem". Recall
the "O2+ problem", where O2+ was wrong in the upper half
of the model. This was solved a few weeks ago in a loop 
near the end of qinite:

Original snoe mods in qinite that apparently caused bad O2+:

      DO 15 I=1,LEN3
        F(I,NQO2PK)=F(I,NQO2PK)+0.67*S6(I,1)*S5(I,1)+1.e+2
      ENDDO  

Original tgcm13mt code that when reinstated, "solved" the 
O2+ problem:

      DO 15 I=1,LEN3
        F(I,NQO2PK)=F(I,NQO2PK)+0.67*S6(I,1)*S5(I,1)
        F(I,NQOPK) =F(I,NQOPK) +(0.33*S6(I,1)+S7(I,1))*S5(I,1)
        F(I,NQN2PK)=F(I,NQN2PK)+0.86*S8(I,1)*S5(I,1)
        F(I,NQNPK) =F(I,NQNPK) +0.14*S8(I,1)*S5(I,1)
   15 CONTINUE

However, after you left I finally realized that using the 
above loop in qinite (i.e., without adding 100 to NQO2P)
was apparently creating the TE problem (the TE "typhoon"
around -60 deg lat).

Now I was faced with the following: adding 100 to NQO2P
resulted in good TE, but bad O2+ (and incidentally also 
created the O2 problem above zp +4.), but NOT adding 100 
to NQO2P caused bad TE and good O2+ (and O2). This was true 
in both tgcm14a and tgcm13mt+snoe.

At this point I went back to tgcm13mt+kibo. I found that the same 
problem existed with your current tgcm13mt+modsrc.kibo, but that 
it was NOT a problem with *my* modsrc.kibo, which I had obtained 
from you on 9/9/99 (when they were merged with tgcm13mt). I then 
did a careful comparison between tgcm13mt+snoe and tgcm13mt + *my* 
9/9/99 modsrc.kibo, and found the following:

Recall you pointed out the use of DIPMIN (minimum DIPMAG) in 
lamdas.f (setting it to .005 vs .17, etc). I discovered that if 
DIPMIN was not used at all in lamdas (i.e., use DIPMAG unconditionally), 
everything worked fine, even with all other snoe mods. (original 
qinite code resulted in good TE, and good O2+ and O2). 

DIPMAG is used only in lamdas and settei. I then found that if
DIPMIN is used in lamdas, that it must also be used in settei.
Even with DIPMIN=.005, if it is used in both lamdas and settei,
everything is fine. This was strange tho, because dipmin should
only effect latitudes near the equator, and the TE problem was
around -60. Then I realized that DIPMAG is negative in the southern
hemisphere, and the conditional was based on dipmag > dipmin, so
it was using dipmin throughout the southern hemisphere.

Currently I am using dipmin=.005, and (abs(dipmag) > dipmin) as the 
conditional in both lamdas and settei, all other snoe mods, and 
everything appears to be ok:

In lamdas (DO 1):
      if (abs(DIPMAG(I,J)) >= dipmin) then
        t4(i) = sin(dipmag(i,j))
      else
        t4(i) = sin(dipmin)
      endif

In settei (DO 11):
        if (abs(DIPMAG(I,J)) >= dipmin) then
          T2(I) = (SIN(DIPMAG(I,J)))**2
        else
          T2(I) = (SIN(DIPMIN))**2
        endif

So now I will get back to work on tgcm14a on blackforest..

Hope you're having a good trip -- see you on 6/5.

--Ben