Once images.aavso.org was restored about two weeks ago, the image inspecting volunteers quickly discovered that the dessicant on BSM_NM had failed and needed to be replaced. Unfortunately, this meant that nights from 160528 through 160622 were compromised. Normally these faults are picked up quickly, as the volunteers keep up with the daily volume of images, and we don't have such a long stretch of compromised data!
The main effect of moisture on the sensor is a "picture frame" rim around the outside of the field, with what seems to be decent imaging in the middle. This moisture condenses when the detector is initially cooled at the beginning of each night, and then slowly evaporates through the night. This means that images taken, say, after local midnight, may be completely free of moisture, while those close to dusk twilight may have quite a bit of the field impacted.
One of the real big problems with this early moisture - later clear effect is that we take flats at dusk, and that is when most of the moisture is present. When you then apply those flats to science images taken later in the night, a clean science image can then be contaminated by a poor flat.
While the corners were contaminated around 160528, as the dessicant became less effective, later nights will show more moisture, and even remaining through the night.
We are in the process of redoing the 160528-160622 nights, using earlier flats that seem to be in good shape. You will see "you've got data" emails if you have targets taken with BSM_NM. You should probably redo a few of these targets, and see if they are affected. If the target is centered, and you have comps nearby, you may see zero difference between the old processing and the new. If your data only uses the Sloan filters, don't bother reprocessing, as those flats are only taken in the dawn when the sensor was clear of moisture.
Arne