It can be diabolical to analyze stuff like that. Remember that they don't just feed fuel into the airflow, but a premixed fuel/air emulsion. I think the temperature of the fuel (and the vapor pressure of the particular fuel you're using) can have quite an effect under some conditions.

I doubt the A/F ratio is actually changing very much -- it's more that the quality of the emulsion changes enough that it doesn't combust well, and you end up with unburnt fuel in the exhaust. If the problem is at 2000 rpm, it's probably running strictly on the low-speed circuit ("probably" because I don't know what chokes or aux venturis you've got), so I'd look for idle jets with a different air component (the F# thing).

I've got a similar problem on hot days at small throttle openings, varying in a small range between 2500 and 3200 rpm. On cool days it's fine. Logic would indicate that it's going rich when hot (hot air has less oxygen, so tends to run richer), but noooo... going richer clears it up. Easy to drive around, but aggravating.

Diabolical, I say.