Error with examples in MetPy Mondays #22 (CAPE/CIN): zero-size array

[rolypolyman]

I have had a couple of issues come up working through the "Metpy Mondays 22 - CAPE/CIN" video on YouTube, in which when we get to the line:

mpcalc.surface_based_cape_cin(p, T, Td)

It errors out with "ValueError: zero-size array to reduction operation minimum which has no identity". This is using all the same code and case examples given in the video.

I did an output to a separate array rather than overwriting the old values and ran some basic checks on the data and get this, so I'm not sure what the problem is.

After layer extraction:
Pressure layer array size: 18
Temperature layer array size: 18
Dewpoint layer array size: 18
Contents:
[969.0 964.0 939.6 925.0 906.8 904.0 874.9 850.0 844.0 819.0 814.0 784.8 757.0 729.1 722.0 700.0 690.0 677.2865329512894] hectopascal
[21.0 20.8 19.1 18.0 16.6 16.4 15.0 13.8 13.6 12.6 12.3 10.2 8.2 4.6 3.6 4.8 5.2 4.420385116433] degree_Celsius
[18.2 18.3 17.8 17.5 16.6 16.4 15.0 13.8 13.6 12.6 12.3 10.2 8.200000000000003 -24.0 -32.4 -29.2 -21.8 -22.579614883567] degree_Celsius

I also had a separate warning come up with the very first Skew-T examples:

UserWarning: Duplicate pressure(s) [17.1 27.1] hPa provided. Output profile includes duplicate temperatures as a result.

This warning didn't come up in the video example. I wasn't sure what to do about this so I wrote some code to remove duplicates. Is there anything in metpy that will clean up these arrays automatically?

Here is the complete code I have for MetPy Mondays 22 with the changes I added (but which still produces errors).

from datetime import datetime
from siphon.simplewebservice.wyoming import WyomingUpperAir
import numpy as np  # Added by AI for handling duplicates
df = WyomingUpperAir.request_data(datetime(2012, 4, 14, 12), 'OUN')

from metpy.units import units
p = df['pressure'].values * units(df.units['pressure'])
T = df['temperature'].values * units(df.units['temperature'])
Td = df['dewpoint'].values * units(df.units['dewpoint'])
heights = df['height'].values * units(df.units['height'])

import metpy.calc as mpcalc

# QUICK FIX for removing duplicate temperatures while preserving order
seen = set()
unique_indices = []
for idx, value in enumerate(T.magnitude):
    if value not in seen:
        seen.add(value)
        unique_indices.append(idx)

p = p[unique_indices]
T = T[unique_indices]
Td = Td[unique_indices]
heights = heights[unique_indices]

mpcalc.surface_based_cape_cin(p, T, Td)

mpcalc.most_unstable_cape_cin(p, T, Td)

from metpy.units import units

p_layer, T_layer, Td_layer = mpcalc.get_layer(p, T, Td, height=heights, depth=3 * units.km)

mpcalc.surface_based_cape_cin(p_layer, T_layer, Td_layer)

[kgoebber]

Hi @rolypolyman, it appears that with the layer you have chosen the equilibrium layer (EL) is undefined since the layer only goes to ~600 hPa and no EL level is found. So I don't think we currently support the calculation of CAPE/CIN without a valid EL.

The duplicate temperatures that you experienced were high up in the atmosphere. Another way to get around that would be to subset the data for pressure >= 100 hPa.

valid_levels = p >= 100*units.hPa

p_subset = p[valid_levels]
T_subset = T[valid_levels]
Td_subset = Td[valid_levels]
heights_subset = heights[valid_levels]

[rolypolyman]

Thanks, that gives me some perspective on the issue. I will work with the suggestions.