As is, need to manually query and download data for many ghcnd stations, unzip it to data/raw folder. There are functions to prep raw data once they are in that folder, and data for a good number of stations only need to be downloaded once because their record stopped in the 80s-00s (no longer useful for infilling current data). However I didn't know the rnoaa package exists and had I, I definitely would have written code to dynamically fetch precip and temp data. It can be a little time consuming to manually download + unzip data, and which stations to download can be confusing (I'm looking at you, Nederland [many stations over the years and some short lived]).
If code is developed to dynamically fetch NOAA data, I would change the climate data prep workflow to entirely using R packages to fetch data (i.e., do away with manual data request and download). Needing to manually download GHCNd was always my biggest reason for not making everything dynamic.
milesalanmoore
commented
4 months ago
As is, need to manually query and download data for many ghcnd stations, unzip it to data/raw folder. There are functions to prep raw data once they are in that folder, and data for a good number of stations only need to be downloaded once because their record stopped in the 80s-00s (no longer useful for infilling current data). However I didn't know the rnoaa package exists and had I, I definitely would have written code to dynamically fetch precip and temp data. It can be a little time consuming to manually download + unzip data, and which stations to download can be confusing (I'm looking at you, Nederland [many stations over the years and some short lived]).
If code is developed to dynamically fetch NOAA data, I would change the climate data prep workflow to entirely using R packages to fetch data (i.e., do away with manual data request and download). Needing to manually download GHCNd was always my biggest reason for not making everything dynamic.