Aggregating CONUS404 hourly data to daily values.

[amsnyder]

When aggregating the CONUS404 hourly data to daily values, it is important to keep track of whether the variable you are using is an instantaneous value or an aggregated value. Aggregated variables will have an attribute attached which indicates the period over which the data was aggregated, called integration_length.

If you were processing the PREC_ACC_NC variable, for example, the intregration_length attribute indicates that the data value is accumulated over prior 60 minutes. Therefore, to get a daily summary, you will want to sum the values for 01:00:00 of that day to 00:00:00 of the next day. You can do this with xarray by running a line of code like ds_daily_sum = ds.PREC_ACC_NC.resample(time="24H", base=1, loffset='-1H').sum(). The base=1 parameter shifts the daily aggregation ahead one hour (by default, it would use 00:00:00 to 23:00:00 of that day), and the loffset='-1H' adjusts the time stamp on the resulting aggregated dataset back to 00:00:00, rather that using the first time stamp it used of 01:00:00.

If you are aggregating an instantaneous data variable, you do not need to do this offset, and you can calculate your aggregation with something like: ds_daily_avg = ds.T2.resample(time="24H").mean().

[sfoks]

Very useful info!! Thank you!!