Supplying your own atmospheric sounding

Copyright (c) 2021 Thomas Schanzer.
Distributed under the terms of the BSD 3-Clause License.

In this notebook, we demonstrate dparcel’s ability to handle real atmospheric sounding data.

[1]:

import numpy as np
import matplotlib.pyplot as plt
import pandas as pd

from metpy.units import units
from metpy.plots import SkewT

from dparcel.environment import Environment
from dparcel.parcel import Parcel

As an example, we take sounding data over Sydney, Australia from http://slash.dotat.org/cgi-bin/atmos, reading it into a pandas dataframe:

[2]:

file = 'SYDNEY AIRPORT (94767) 16 Jul 2021 00Z.txt'
sounding = pd.read_csv(
    file, names=['pressure', 'height', 'temperature', 'dewpoint'],
    usecols=[0, 1, 2, 3], header=0)
sounding.head()

[2]:
pressure height temperature dewpoint
0 1021 35 11.2 9.6
1 1013 99 14.2 8.2
2 1005 165 15.0 9.0
3 1000 206 15.0 9.0
4 999 214 15.0 8.9

Then extract the columns as numpy arrays and assign them the correct units. We use the pint unit registry supplied in metpy.

[3]:

sounding = sounding.to_numpy()
pressure = sounding[:,0]*units.mbar
height = sounding[:,1]*units.meter
temperature = sounding[:,2]*units.celsius
dewpoint = sounding[:,3]*units.celsius

This is what the sounding looks like on a skew-T log-p diagram:

[4]:

fig = plt.figure(figsize=(8,6))
skew = SkewT(fig, rotation=45)
skew.plot(pressure, temperature, label='Temperature')
skew.plot(pressure, dewpoint, label='Dew point')
skew.ax.legend()
skew.ax.set(
    xlabel=r'Temperature ($^\circ$C)',
    ylabel='Pressure (mbar)',
    title='SYDNEY AIRPORT (94767) 16 Jul 2021 00Z')
fig.tight_layout()
plt.show()
../_images/examples_real_sounding_7_0.png

We then instantiate an Environment, supplying the sounding arrays and some information to identify the data later:

[5]:

sydney = Environment(
    pressure, height, temperature, dewpoint,
    info='SYDNEY AIRPORT (94767) 16 Jul 2021 00Z', name='Sydney')

dparcel interpolates the sounding we provide, allowing us to query the data at any height or pressure:

[6]:

sydney.pressure(5.32*units.km)

[6]:
526.3793103448276 millibar
[7]:

sydney.height(309.1*units.mbar)

[7]:
9256.609999999999 meter
[8]:

sydney.temperature(3.22*units.km)

[8]:
-0.6514506769825921 degree_Celsius
[9]:

sydney.dewpoint_from_pressure(299.4*units.mbar)

[9]:
-38.56 degree_Celsius

We may also calculate derived variables at any height:

[10]:

sydney.wetbulb_temperature(2.19*units.km)

[10]:
-2.133978225990675 degree_Celsius
[11]:

sydney.specific_humidity(6.77*units.km)

[11]:
0.0005668152743663845 dimensionless
[12]:

sydney.density(4.89*units.km)

[12]:
0.734725618621596 kilogram/meter3
[13]:

sydney.dcape_dcin()

[13]:

(123.01795081346282 <Unit('meter ** 2 / second ** 2')>,
 -450.43506401330427 <Unit('meter ** 2 / second ** 2')>)

If we wished to simulate the motion of downdraft parcels in this sounding, we would simply instantiate a Parcel object, supplying the same arguments used to create the Environment object, like so:

[14]:

parcel = Parcel(
    pressure, height, temperature, dewpoint,
    info='SYDNEY AIRPORT (94767) 16 Jul 2021 00Z', name='Sydney')

See the “Simulating the motion of a downdraft parcel” tutorial for information on performing parcel calculations.