more values are in specific humidity calulation

[met-sree]

Hi everyone,
I tried to calculate the specific humidity from ERA5 data by using metpy.calc. But, it provides higher values (more than 1/10 th) as compared with ERA5 specific humidity (like the attached image). Please let me know why these huge variations appear. Is it due to the disparity in calculating approximations? Please suggest me.

The code:

from netCDF4 import Dataset
import metpy.calc
from metpy.units import units
import numpy as np
import matplotlib.pyplot as plt
import pint
ureg = pint.UnitRegistry()

data=Dataset('D:/Data/WV_TEMP/ncmwrf/RH/ecmwf.nc')
lat1=data['latitude'][:]
lon1=data['longitude'][:]
lev1=data['level'][:]
tim1=data['time'][:]
temp=data['t'][:]
rh=data['r'][:]
sh=np.array(data['q'][:])

siz=np.shape(temp);

press1=np.empty([siz[2],siz[3]])
press2=np.empty([siz[1],siz[2],siz[3]])
press3=np.empty([siz[0],siz[1],siz[2],siz[3]])


for pp in range(0,siz[1]):
        press1[:,:]=lev1[pp]
        press2[pp,:,:]=press1
        
for qq in range(0,siz[0]):
    press3[qq,:,:,:]=press2
    
aa=metpy.calc.dewpoint_from_relative_humidity((np.array(temp)*units.K),(np.array(rh)/100)).to(units('K'))
bb=np.array(metpy.calc.specific_humidity_from_dewpoint(np.array(aa)*units.K,press3*units.mbar))

fig = plt.gcf()
fig.set_size_inches(30, 10)
plt.subplot(1,3,1)
plt.contourf(lon1,lat1,bb[11,2,:,:])
plt.title('metpy')
plt.colorbar(ticks=np.linspace(0, 0.000200, 7))
plt.clim(0,0.000200)

plt.subplot(1,3,2)
plt.contourf(lon1,lat1,sh[11,2,:,:])
plt.title('era5')
plt.colorbar(ticks=np.linspace(0, 0.000200, 7))
plt.clim(0,0.000200)

plt.subplot(1,3,3)
plt.contourf(lon1,lat1,bb[11,2,:,:]-sh[11,2,:,:])
plt.title('metpy-era5')
plt.colorbar()

Thank you.

[dopplershift]

Interesting. Any chance you could share that data file somewhere so I could take a look directly? How big is that file?

Also, did something motivate you to compare the ERA5 specific humidity with the output from MetPy's calculation?

[met-sree]

Thank you for your reply. The file size is 38 MB. I tried to reduce the size of the file in so many ways (like zip, gzip, bzip etc...). But, its size remains the same. So, below is attached the python script of that ERA5 data. Please look at this.
Thank you.

script.py.gz

[dopplershift]

Thanks, that script worked great (once I changed to grab specific_humidity rather than rainwater). Based on what @DanielAdriaansen posted below, I think looking at a comparison of specific humidity is the wrong way to go since it seems that's more of a direct variable, while relative humidity is a derived quantity.

Comparing ERA5 relative humidity to MetPy's calculated one, there's definitely some discrepencies. Here's a plot of the relative error between the two as a function of temperature:

There's something going on even above 0C, but I'm not sure what yet.

[DanielAdriaansen]

I did not look too much at the details here, but did notice the use of dewpoint_from_relative_humidity() and specific_humidity_from_dewpoint().

From the ERA5 documentation for relative humidity here https://apps.ecmwf.int/codes/grib/param-db?id=157:

For temperatures over 0°C (273.15 K) it is calculated for saturation over water. At temperatures below -23°C it is calculated for saturation over ice. Between -23°C and 0°C this parameter is calculated by interpolating between the ice and water values using a quadratic function.

If the MetPy functions assume RH is with respect to water, then a temperature mask has to be applied to the ERA5 RH field so that you only get RH values where it is RH saturated over water.