search

HydrologicEngineeringCenter / Vortex

data processing utilities
MIT License
25 stars 7 forks source link

Current version importer writes both MRMS and AORC precipitation fields with same D and E part timesteps #113

Closed openSourcerer9000 closed 3 weeks ago

openSourcerer9000 commented 3 weeks ago

I'm including datasets from both of these sources for you to test. It seems it's been doing this for the last ~ dozen versions. I've found versions that write the time steps correctly but have a plethora of other bugs, such as writing blank DSS grids or simply freezing up indefinitely. Unfortunately I don't have exact version numbers for each bug documented.

I've tried playing with the metadata etc but with no luck. When running time shifter on these DSS's, E part drops out completely. Using 'dataType':'PER-CUM','units':'MM', Cpart:'PRECIPITATION' cellsize 1000

image March_2018 MRMS.zip AORC.zip

tombrauer commented 3 weeks ago

It looks like you are trying to import a non-CF compliant NetCDF file. If the NetCDF file doesn't include a time_bnds variable, which defines the interval over which the data applies, vortex doesn't make any guesses about what that might be. For this case data processing happens in two steps 1) Use the importer to import the data to HEC-DSS and 2) use the time-shifter to shift the start time or end time of the HEC-DSS grids by some amount, e.g. shift the start time by -1 hours. I would contact the originator of the data and ask them to include a time_bnds variable to be more explicit about the data interval.

openSourcerer9000 commented 3 weeks ago

Thanks, I've never heard of that one. I'm assuming that should be an auxiliary coordinate variable, not a second data variable next to precipitation? Do you have an example dataset with time_bnds so I can see what dims, coordinates, and attributes it would have? My data is cumulative hourly precipitation depth.

openSourcerer9000 commented 3 weeks ago

NVMD, I didn't see your link. Works like a charm