Open lekoenig opened 2 years ago
Just to clarify, the expected temp column name for river-dl is specified in the config file, so we could change what we use without causing any errors or changes to the behavior for any other project. It's just a question of whether there would be confusion / perceived inconsistencies if what's really the same input variable has different names in different models.
@lekoenig I believe we can close this now?
The 2022 forecasting data release contains both aggregated (by NHM segment-date) and unaggregated temperature observations, both of which have columns for
min_temp_c
,mean_temp_c
, andmax_temp_c
. As a result, these are the columns we carry through when we createp2_drb_temp_obs_by_comid
, which is a data frame containing one row of temperature data per COMID-date.Janet pointed out that river-dl is expecting a column called
temp_c
and so to run river-dl she changes the column frommean_temp_c
totemp_c
. This issue is a reminder to check in with others who run river-dl to see if switching tomean_temp_c
in that workflow would cause any issues. If so, we should adjust the column names used in our pipeline to match what river-dl is expecting.