I normally don't have trouble with craters with the icing recipe I use. However, a few weeks ago I was working on some cookies and was filling in the lettering on the logo. I was flooding very small sections, just as you described. I experienced the worst cratering EVER.... in all 12 of the cookies I made. I scraped out the flooded areas (as they hadn't fully dried), and re-flooded, being sure to flood them extra full and run a toothpick through the icing this time, but I STILL had cratering problems. I actually had to go back later with icing on a toothpick and carefully try to fill in the craters - it was a nightmare! Out of curiosity, I took the same icing and flooded several 3" circle cookies to see what would happen. Those cookies had NO cratering problems whatsoever. So, the only thing I figured out was that the icing was fine, but something about flooding in those small areas was causing the cratering issue. Next time I will have to try the dehumidifier idea that some of you mentioned above..... maybe that's the secret! Thanks for the idea!
This may be getting too geeky on you all, but my theory is that smaller spaces crater more due to changing surface tension effects on the icing as it dries. Meaning: the icing tends to dry faster in thinner areas near edges and corners, and if those edges are close together, the icing on the edges in effect weighs down the center icing as it dries, causing the center to collapse in on itself. I need to find a scientist to validate this for me! But, let's just say, I've never seen cratering in areas bigger than about 1/2 inch, and it's almost always worse in angular shapes with corners. But if you can quick-set the icing, then you minimize any differences in drying time across the surface of the icing and therefore minimize craters. That all said (hypothesized), I think the size and shape of the space is the biggest driver of cratering. If you use a quick-drying (often thicker) icing this can minimize craters too, along with fanning and dehydrating.