Warning: converting a masked element to nan when plotting RTMA Data from THREDDS

[edrewitz]

Hey all!

I have noticed when plotting data from THREDDS about half the time I get an error saying "Warning: converting a masked element to nan." and then the RTMA data will not plot. I've tried looking online for ideas to get rid of these NaN values but couldn't find anything super helpful so I am reaching out here if anyone knows any tricks on how to resolve this issue. Thank you!

Regards,

Eric

[edrewitz]

One thing to add is it seems like there is data in the RTMA Temperature and Dewpoint arrays but then when calculating RH it sometimes converts them to NaN.

[edrewitz]

Good day Ryan,

I figured I'd follow up with you in regards to the nan values that show up when calculating relative humidity from temperature and dewpoint. I've had this issue show up again when ingesting the 7/13 RH data for 15z, 16z and 17z. The 13z and 14z came in fine with no errors. I'm not sure why sometimes it works while other times I get this error.

Thanks!

Regards,

Eric

[edrewitz]

Good day Ryan,

Sorry for all the posts here. I did more troubleshooting this morning with structuring my code differently and I think there were issues with the data arrays for temperature and dewpoint in the files on the THREDDS server during those times I get these errors rather than it being an issue with mpcalc.get_relative_humidity_from_dewpoint() in metpy. I'm not sure as to why there are these issues in the data arrays for the T and Td variables. If you have any insight on how to correct the data in my code I'd be happy to learn. If it is just an issue of no data existing, then I can make that work in my code to output a no data available statement on the images during the times of no data. Thank you for all that you do.

Regards,

Eric

[edrewitz]

Good day Ryan,

I did more troubleshooting this morning and I think the issue arises when taking the difference in RH values between two different times (comparing the current RH to the RH 24 hours ago). I got the error message when running the script to plot the difference at 7/20/2023 12z but when running the script to just plot the current RH with using 7/20/2023 12z as the current time, it plots with no issues. I'm not sure why this is but it is something I found this morning. Thanks.

Regards,

Eric

[dopplershift]

Thanks for the additional information. I was able to see the missing data for 12Z today. I think it's an issue with the TDS indexing for the best collection, because I don't see the missing data if I use the 12Z GRIB file directly on the server instead of Best. We'll keep digging...

[edrewitz]

Good day Ryan,

Thank you!! I just tried doing the same thing you did yesterday and got the same result. I changed my code to go directly into the latest file rather than indexing through the RTMA Best Collection and I do not see those nan values and get that error anymore. Before I did this, I did try indexing through the RTMA Best Collection for the same time and had those issues so I think this solves the problem. Thank you for all of your hard work and for helping us meteorologists with our learning of Python! 👍

Regards,

Eric