Deriving Bowen Ratio as a Forecast Variable in point_stat Using SHTFL and LHTFL

[briannen]

Is it possible to derive the Bowen Ratio (Sensible Heat Flux / Latent Heat Flux) as a forecast variable in point_stat, similar to how WIND is handled? I’ve reviewed the pcp_combine and stat_analysis documentation, but I couldn’t find a method for dividing two variables.

The forecast file (in GRIB2 format) contains Sensible Heat Flux (SHTFL) and Latent Heat Flux (LHTFL), but does not include the Bowen Ratio. However, the observations do contain a separate Bowen Ratio variable.

[j-opatz]

Hi @briannen,

And thank you for your patience on getting an answer to this question. While I will consult next week with the lead engineers when they return to office, I believe only UGRD and VGRD, and only when pulled from a GRIB or GRIB2-formatted file, can be used to compute a separate variable (WIND). This is because the relationship between the U and V components of wind and the final wind speed is well-defined. I will also note that this behavior (automatic calculation of wind speed) is desired from netCDF files, and has been captured in issue dtcenter/MET#1518 (although this issue is currently in the backlog of development and not in active development).

As you probably found during your search of the documentation, the Bowen Ratio is not a supported variable in MET (same for SHTFL and LHTFL). And as you correctly found, MET also does not support the division of one field by another. This is an intentional behavior; if division were allowed, there could conceivably be a situation where values of 0, -9999, NAs etc end up in the denominator. While the bad values (-9999, NAs, etc.) could be dealt with by assigning a bad data value and not calculating the result, a value of 0 could very well be good data, but cannot be calculated. To avoid this assigning bad data output to good data input locations, division is not allowed in PCP-Combine, nor any other user-facing tool option in MET.

If you want to calculate the Bowen Ratio in your data, I would suggest utilizing MET's Python Embedding to creating a Python Script that reads in your data, calculates the bowen ratio, and then passes the resulting field into MET for verification against the observed Bowen Ratio field. Information on what format that python field will need to look like is available here.

[briannen]

Thanks! I thought that would be the case, but I wanted to double-check before I spent the time working on a Python embedding script for the data.