o-b looks a lot better in skylab if the u and v wind components are reversed first. to test whether this is true before applying hofx, we need to adjust the converter, re-convert the data, and re-run JEDI for the 2023 sept period.
After talking to Windborne, their labeling was backwards and speed_x is now changed to speed_v and speed_y to speed_u. the converter is updated to handle both cases (speed_u or speed_y).
The variable names from their previous data were compared to their updated obs website, and they match, aside from the addition of speed_u and _v. (march 21st 2024). There is expected to be a specific humidity variable to come in the future.
# Right now, we simply hardcode these values to reasonable averages
# However, in the future we may calculate these ourselves
wind_x_error = 1.18
wind_y_error = 1.18
relative_humidity_error = 8.23
temperature_error = 0.83
Description
o-b looks a lot better in skylab if the u and v wind components are reversed first. to test whether this is true before applying hofx, we need to adjust the converter, re-convert the data, and re-run JEDI for the 2023 sept period.
After talking to Windborne, their labeling was backwards and
speed_x
is now changed tospeed_v
andspeed_y
tospeed_u
. the converter is updated to handle both cases (speed_u or speed_y).The variable names from their previous data were compared to their updated obs website, and they match, aside from the addition of speed_u and _v. (march 21st 2024). There is expected to be a specific humidity variable to come in the future.
Their converter script: https://github.com/windborne/to-prepbufr/blob/main/wb_to_prepbufr.py ^in which, they convert RH to Q and give error estimates:
Issue(s) addressed
Resolves #1489