Try this notebook in 
.
[1]:
!test -f temp_small.bufr || wget https://get.ecmwf.int/repository/test-data/pdbufr/test-data/temp_small.bufr
Temp: overview
[2]:
import pdbufr
The input BUFR data contains Temp (radiosonde) observations. In this notebook we will read this file with the temp reader, which uses pre-defined parameters instead of the ecCodes BUFR keys. Parameter is a high-level concept in pdbufr enabling data extraction without knowing the actual BUFR encoding and allowing more tailor-made access to specific data types.
The default settings
By default the “stnid”, “lat”, “lon”, “elevation”, “time” (see station parameters for details) and the upper level parameters are extracted. We only extract the first station from the BUFR file by specifying a filter. The resulting dataframe contains one row per pressure level.
[3]:
df = pdbufr.read_bufr("temp_small.bufr", reader="temp", filters={"count": 1})
df[:10]
[3]:
| stnid | lat | lon | elevation | time | pressure | z | t | td | wind_speed | wind_dir | |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 71907 | 58.47 | -78.08 | 26 | 2008-12-08 12:00:00 | 100300.0 | 250.0 | 258.3 | 255.7 | NaN | NaN |
| 1 | 71907 | 58.47 | -78.08 | 26 | 2008-12-08 12:00:00 | 100000.0 | 430.0 | 259.7 | 258.3 | 0.0 | 0.0 |
| 2 | 71907 | 58.47 | -78.08 | 26 | 2008-12-08 12:00:00 | 99800.0 | 630.0 | 261.1 | 259.6 | NaN | NaN |
| 3 | 71907 | 58.47 | -78.08 | 26 | 2008-12-08 12:00:00 | 99100.0 | 1160.0 | 261.7 | 258.3 | NaN | NaN |
| 4 | 71907 | 58.47 | -78.08 | 26 | 2008-12-08 12:00:00 | 92500.0 | 6240.0 | 258.1 | 256.2 | 5.0 | 275.0 |
| 5 | 71907 | 58.47 | -78.08 | 26 | 2008-12-08 12:00:00 | 85000.0 | 12470.0 | 253.1 | 251.0 | 9.0 | 250.0 |
| 6 | 71907 | 58.47 | -78.08 | 26 | 2008-12-08 12:00:00 | 75100.0 | 21340.0 | 245.3 | 244.2 | NaN | NaN |
| 7 | 71907 | 58.47 | -78.08 | 26 | 2008-12-08 12:00:00 | 72400.0 | 23900.0 | 242.9 | 239.7 | NaN | NaN |
| 8 | 71907 | 58.47 | -78.08 | 26 | 2008-12-08 12:00:00 | 71700.0 | 24580.0 | 242.3 | 239.3 | NaN | NaN |
| 9 | 71907 | 58.47 | -78.08 | 26 | 2008-12-08 12:00:00 | 70000.0 | 26240.0 | 241.7 | 238.3 | 12.0 | 235.0 |
Using parameter groups
By using the columns option we can control the parameters to extract. Using the pre-defined values of “station”, “location” or “geometry” we will only read station related information yielding one row per station. See details here.
[4]:
pdbufr.read_bufr("temp_small.bufr", reader="temp", columns="station")
[4]:
| stnid | lat | lon | elevation | time | |
|---|---|---|---|---|---|
| 0 | 71907 | 58.47 | -78.08 | 26 | 2008-12-08 12:00:00 |
| 1 | 71823 | 53.75 | -73.67 | 302 | 2008-12-08 12:00:00 |
| 2 | 89009 | -90.00 | 0.00 | 2835 | 2008-12-08 12:00:00 |
| 3 | 91348 | 6.97 | 158.22 | 38 | 2008-12-08 12:00:00 |
| 4 | 91408 | 7.33 | 134.48 | 30 | 2008-12-08 12:00:00 |
| 5 | 71836 | 51.27 | -80.65 | 11 | 2008-12-08 12:00:00 |
| 6 | 83612 | -20.47 | -54.67 | 567 | 2008-12-08 12:00:00 |
[5]:
pdbufr.read_bufr("temp_small.bufr", reader="temp", columns="location")
[5]:
| lat | lon | |
|---|---|---|
| 0 | 58.47 | -78.08 |
| 1 | 53.75 | -73.67 |
| 2 | -90.00 | 0.00 |
| 3 | 6.97 | 158.22 |
| 4 | 7.33 | 134.48 |
| 5 | 51.27 | -80.65 |
| 6 | -20.47 | -54.67 |
[6]:
pdbufr.read_bufr("temp_small.bufr", reader="temp", columns="geometry")
[6]:
| lat | lon | elevation | |
|---|---|---|---|
| 0 | 58.47 | -78.08 | 26 |
| 1 | 53.75 | -73.67 | 302 |
| 2 | -90.00 | 0.00 | 2835 |
| 3 | 6.97 | 158.22 | 38 |
| 4 | 7.33 | 134.48 | 30 |
| 5 | 51.27 | -80.65 | 11 |
| 6 | -20.47 | -54.67 | 567 |
We can can combine these groups with “upper”, which represents the pressure level parameters. Please note thath at the moment it is not possible to use individual parameters in columns.
[7]:
df = pdbufr.read_bufr("temp_small.bufr", reader="temp", columns=["location","upper"], filters={"count": 1})
df[:5]
[7]:
| lat | lon | pressure | z | t | td | wind_speed | wind_dir | |
|---|---|---|---|---|---|---|---|---|
| 0 | 58.47 | -78.08 | 100300.0 | 250.0 | 258.3 | 255.7 | NaN | NaN |
| 1 | 58.47 | -78.08 | 100000.0 | 430.0 | 259.7 | 258.3 | 0.0 | 0.0 |
| 2 | 58.47 | -78.08 | 99800.0 | 630.0 | 261.1 | 259.6 | NaN | NaN |
| 3 | 58.47 | -78.08 | 99100.0 | 1160.0 | 261.7 | 258.3 | NaN | NaN |
| 4 | 58.47 | -78.08 | 92500.0 | 6240.0 | 258.1 | 256.2 | 5.0 | 275.0 |
Adding units columns
With units_columns=True we can add units to the resulting DataFrame.
[8]:
df = pdbufr.read_bufr("temp_small.bufr", reader="temp", filters={"count": 1}, units_columns=True)
df[:5]
[8]:
| stnid | lat | lon | elevation | time | pressure | pressure_units | z | z_units | t | t_units | td | td_units | wind_speed | wind_speed_units | wind_dir | wind_dir_units | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 71907 | 58.47 | -78.08 | 26 | 2008-12-08 12:00:00 | 100300.0 | Pa | 250.0 | m2 s-2 | 258.3 | K | 255.7 | K | NaN | m/s | NaN | deg |
| 1 | 71907 | 58.47 | -78.08 | 26 | 2008-12-08 12:00:00 | 100000.0 | Pa | 430.0 | m2 s-2 | 259.7 | K | 258.3 | K | 0.0 | m/s | 0.0 | deg |
| 2 | 71907 | 58.47 | -78.08 | 26 | 2008-12-08 12:00:00 | 99800.0 | Pa | 630.0 | m2 s-2 | 261.1 | K | 259.6 | K | NaN | m/s | NaN | deg |
| 3 | 71907 | 58.47 | -78.08 | 26 | 2008-12-08 12:00:00 | 99100.0 | Pa | 1160.0 | m2 s-2 | 261.7 | K | 258.3 | K | NaN | m/s | NaN | deg |
| 4 | 71907 | 58.47 | -78.08 | 26 | 2008-12-08 12:00:00 | 92500.0 | Pa | 6240.0 | m2 s-2 | 258.1 | K | 256.2 | K | 5.0 | m/s | 275.0 | deg |
[ ]: