It may have just been a temporary gauge malfunction. I noticed it last night, and when I looked again a half hour later, it was back to what it had been. There was no corresponding drop in the data from the Glencoe gauge.
The downward spike has been removed from the chart. There are now two gaps in the data, one between 10 and midnight and the other between about 4 and 8 this morning. The gaps lead me to believe that the very low number had more to do with the gauge than the actual flow.