Looks like the sensor at BTM is conked off, and this data may not be correct.

The readings from this sensor has swung wildly between 40-800 in a matter of 2 days.

So it might be safe to say that the reading of 438 might also be wrong, comparing the historical data of this station and nearby points.