Add specific humidity to the ADPUPA BUFR Python Converter

[PraveenKumar-NOAA]

The main goal of this issue is to:

• Extract the specific humidity calculation informations from the prepBUFR codes
• Document them
• Add this to the the ADPUPA BUFR Python Converter
• Run end-to-end tests using global-workflow

[PraveenKumar-NOAA]

The specific humidity informations were extracted from the prepBUFR codes, added to the ADPUPA BUFR Converter. Run the converter and created the IODA/ netCDF file.

The following plots were created for the BUFR and prepBUFR specificHumidity using the IODA/netCDF files created using the BUFR and prepBUFR converters, respectively.

BUFR: specificHumidity

prepBUFR: specificHumidity

[PraveenKumar-NOAA]

Analysis for BUFR data:
Total Count: 45594 Max: 0.273 Min: 0.000 Mean: 0.005 Std: 0.008

Analysis for prepBUFR data:
Total Count: 32005 Max: 0.027 Min: 0.000 Mean: 0.002 Std: 0.004

It seems that prepBUFR QC reduced the number of obs from 45594 to 32005.

@emilyhcliu please look at this.

[PraveenKumar-NOAA]

BUFR (corrected) vs prepBUFR analysis for specificHumidity:

BUFR - Total Count: 37556 Max: 0.027 Min: 0.000 Mean: 0.003 Std: 0.004
prepBUFR - Total Count: 32005 Max: 0.027 Min: 0.000 Mean: 0.002 Std: 0.004

The updated plot of specificHumidity for BUFR:

Plot for prepBUFR:

@emilyhcliu please check these plots and data analysis for specificHumidity.

[PraveenKumar-NOAA]

@nicholasesposito please look at the sH plots for BUFR vs prepBUFR.

[nicholasesposito]

Most of the data points look the same, which I think is good. The max, min and std are the same, which is also good, though the mean is slightly different. It's hard to say if that's ok since the values are so low to begin with.

Two things I think you should look into.

1. Do a histogram of sH values to see how they compare.
2. Since they're sondes, I'm imagining the data lower to the ground is greater than sH higher up. So I would also plot a Pressure (y-axis) vs sH (x-axis) so you can see where in the atmosphere the biggest differences are.

[PraveenKumar-NOAA]

specificHumidity vs pressure scatter plots:

prepBUFR:

BUFR: