Hi @Gkikas-LGPZ - Thanks for responding. In most forecasts I check I do mostly see consistency as you do, but even the numbers in your time series add up to 14.0 mm, not 14.5 mm. I think you're suggesting that the model's reference time may be earlier than what is displayed in the time series, and therefore is greater as it's counting precip that is not displayed?
The model time is 2020-03-27 12:00 Zulu time/UTC, which converted to the local region is 7:00 AM CDT (-5). I'm assuming that the time series converts to the local time zone, so it shouldn't be referenced from a time prior to the converted time series. If the time series is also in UTC it still wouldn't be referencing values off the table so long as its only summing from the reference time.
The NAM model (North American only) does report a 48 hour duration.