search

ecmwf / pdbufr

High-level BUFR interface for ecCodes
Apache License 2.0
23 stars 8 forks source link

columns for TC tracks are out of sync #76

Open ncubukcu opened 1 month ago

ncubukcu commented 1 month ago

when you read a .bufr file from ecmwf public sites at https://data.ecmwf.int/forecasts/, you get dataframe like the one below. 

                             0         1      2         3      4    \

stormIdentifier 03E 03E 03E 03E 03E
latitude 20.5 20.7 21.3 20.8 21.4
longitude -127.4 -128.2 -128.4 -128.9 -129.1
pressureReducedToMeanSeaLevel NaN 100000.0 NaN 100400.0 NaN
windSpeedAt10M NaN NaN 18.0 NaN 17.0 

                                5      6         7      8         11   \

stormIdentifier 03E 03E 03E 03E 03E
latitude 20.9 22.0 20.8 21.0 20.7
longitude -129.5 -129.5 -130.0 -130.9 -131.4
pressureReducedToMeanSeaLevel 100400.0 NaN 100800.0 NaN 101000.0
windSpeedAt10M NaN 16.5 NaN 14.9 NaN 

                           ...    108    109       110    111       112  \

stormIdentifier ... 05E 06E 06E 06E 06E
latitude ... 26.3 15.2 15.9 16.7 16.7
longitude ... -139.3 -107.3 -108.8 -108.3 -110.7
pressureReducedToMeanSeaLevel ... NaN NaN 100200.0 NaN 100400.0
windSpeedAt10M ... 13.4 NaN NaN 19.6 NaN 

                             113       116    117      118    119

stormIdentifier 06E 06E 06E 06E 06E
latitude 12.7 18.5 13.6 19.4 14.9
longitude -114.5 -113.4 -114.7 -115.7 -114.8
pressureReducedToMeanSeaLevel NaN 100300.0 NaN 99900.0 NaN
windSpeedAt10M 22.6 NaN 27.3 NaN 23.7
``

As you see pressure and wind and timePeriod are not aligned correctly. Is thare something I am missing? Thanks

Code: df = pdbufr.read_bufr(bufr_file_path,columns=("stormIdentifier","latitude", "longitude","windSpeedAt10M","pressureReducedToMeanSeaLevel"),required_columns=False)

sandorkertesz commented 1 month ago

Thank you for reporting this issue. It seems to be the same as #69 and my interpretation is that this is a problem with the data itself.