Try this notebook in .

[1]:

!test -f temp.bufr || wget https://get.ecmwf.int/repository/test-data/pdbufr/test-data/sample-data/temp.bufr

Generic: radiosonde data

[2]:

import pdbufr

The input BUFR data contains radiosonde observations. Each message represents a given station and contains vertical profiles on a set of pressure levels.

In this notebook we read this data with the generic reader, which is the default reader.

Example 1

Getting the WMO station ID and location for each message. Notice that since “WMO_station_id” is a computed key it appears as the last column in the resulting dataframe.

[3]:

df = pdbufr.read_bufr("temp.bufr",
    columns=("WMO_station_id", "latitude", "longitude"))
df.head()

[3]:

	latitude	longitude	WMO_station_id
0	58.47	-78.08	71907
1	53.75	-73.67	71823
2	-90.00	0.00	89009
3	18.43	-69.88	78486
4	21.98	-159.33	91165

Example 2

Extracting the same data but only for a given station.

[4]:

df = pdbufr.read_bufr("temp.bufr",
    columns=("WMO_station_id", "latitude", "longitude"),
    filters={"WMO_station_id": 71907},)
df

[4]:

	latitude	longitude	WMO_station_id
0	58.47	-78.08	71907

Example 3

Getting temperature profiles for two stations.

[5]:

df = pdbufr.read_bufr("temp.bufr",
    columns=("WMO_station_id", "data_datetime", "pressure", "airTemperature"),
    filters={"WMO_station_id": [71823, 71907]})
df.head()

[5]:

	pressure	airTemperature	data_datetime	WMO_station_id
0	100300.0	258.3	2008-12-08 12:00:00	71907
1	100000.0	259.7	2008-12-08 12:00:00	71907
2	99800.0	261.1	2008-12-08 12:00:00	71907
3	99100.0	261.7	2008-12-08 12:00:00	71907
4	92500.0	258.1	2008-12-08 12:00:00	71907

[6]:

df.tail()

[6]:

	pressure	airTemperature	data_datetime	WMO_station_id
44	30000.0	218.9	2008-12-08 12:00:00	71823
45	29700.0	218.9	2008-12-08 12:00:00	71823
46	25000.0	221.1	2008-12-08 12:00:00	71823
47	23200.0	223.1	2008-12-08 12:00:00	71823
48	20500.0	221.5	2008-12-08 12:00:00	71823

Example 4

Extracting temperature profiles for a given station only on standard pressure levels when temperature >= -30 C.

[7]:

df = pdbufr.read_bufr("temp.bufr",
    columns=("WMO_station_id", "data_datetime", "pressure", "airTemperature"),
    filters={"WMO_station_id": 71907,
            "verticalSoundingSignificance": 32,
            "airTemperature": slice(243.16, None)})
df

[7]:

	pressure	airTemperature	data_datetime	WMO_station_id
0	100000.0	259.7	2008-12-08 12:00:00	71907
1	92500.0	258.1	2008-12-08 12:00:00	71907
2	85000.0	253.1	2008-12-08 12:00:00	71907

Example 5

Extracting temperature, dewpoint and wind values on 500 hPa in the tropics:

[8]:

df = pdbufr.read_bufr("temp.bufr",
    columns=("latitude", "longitude", "pressure", "airTemperature", "dewpointTemperature", "windDirection", "windSpeed"),
    filters={"latitude": slice(-20, 20), "pressure": 50000})
df

[8]:

	latitude	longitude	pressure	airTemperature	dewpointTemperature	windDirection	windSpeed
0	18.43	-69.88	50000.0	267.5	247.5	275	4.0
1	18.03	-63.12	50000.0	267.5	244.5	340	5.0
2	13.47	144.78	50000.0	270.3	249.3	110	12.0
3	7.08	171.38	50000.0	268.9	226.9	220	6.0
4	7.45	151.83	50000.0	269.9	249.9	50	18.0
...	...	...	...	...	...	...	...
69	-10.17	123.67	50000.0	269.5	267.0	70	11.0
70	-1.18	136.12	50000.0	264.1	262.6	80	8.0
71	-2.17	106.13	50000.0	268.9	263.9	135	6.0
72	2.83	-60.70	50000.0	267.5	258.5	90	8.0
73	-3.67	-69.67	50000.0	266.7	266.6	350	4.0

74 rows × 7 columns