Datasets:

deepguess
/

severe-weather-env-hrrr-v1

Error code:   DatasetGenerationError
Exception:    CastError
Message:      Couldn't cast
dataset_name: string
source: string
source_file: string
rows: int64
columns: int64
source_summary: struct<rows: int64, features_full: int64, features_env_only: int64, labels: list<item: struct<source (... 224 chars omitted)
  child 0, rows: int64
  child 1, features_full: int64
  child 2, features_env_only: int64
  child 3, labels: list<item: struct<source_package: string, label_family: string, label_tornado: int64, n: int64>>
      child 0, item: struct<source_package: string, label_family: string, label_tornado: int64, n: int64>
          child 0, source_package: string
          child 1, label_family: string
          child 2, label_tornado: int64
          child 3, n: int64
  child 4, years: list<item: struct<target_year: int64, label_tornado: int64, n: int64>>
      child 0, item: struct<target_year: int64, label_tornado: int64, n: int64>
          child 0, target_year: int64
          child 1, label_tornado: int64
          child 2, n: int64
  child 5, skipped_optional_tables: list<item: string>
      child 0, item: string
  child 6, feature_matrix_mb: double
created_utc: string
checksums_sha256: struct<data/hrrr_event_control.parquet: string, data/preview_1000.csv: string, metadata/columns.json (... 60 chars omitted)
  child 0, data/hrrr_event_control.parquet: string
  child 1, data/preview_1000.csv: string
  child 2, metadata/columns.json: string
  child 3, metadata/manifest.json: string
  child 4, README.md: string
to
{'text': Value('string')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1821, in _prepare_split_single
                  num_examples, num_bytes = writer.finalize()
                                            ^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/arrow_writer.py", line 781, in finalize
                  self.write_rows_on_file()
                File "/usr/local/lib/python3.12/site-packages/datasets/arrow_writer.py", line 663, in write_rows_on_file
                  self._write_table(table)
                File "/usr/local/lib/python3.12/site-packages/datasets/arrow_writer.py", line 773, in _write_table
                  pa_table = table_cast(pa_table, self._schema)
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2321, in table_cast
                  return cast_table_to_schema(table, schema)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2249, in cast_table_to_schema
                  raise CastError(
              datasets.table.CastError: Couldn't cast
              dataset_name: string
              source: string
              source_file: string
              rows: int64
              columns: int64
              source_summary: struct<rows: int64, features_full: int64, features_env_only: int64, labels: list<item: struct<source (... 224 chars omitted)
                child 0, rows: int64
                child 1, features_full: int64
                child 2, features_env_only: int64
                child 3, labels: list<item: struct<source_package: string, label_family: string, label_tornado: int64, n: int64>>
                    child 0, item: struct<source_package: string, label_family: string, label_tornado: int64, n: int64>
                        child 0, source_package: string
                        child 1, label_family: string
                        child 2, label_tornado: int64
                        child 3, n: int64
                child 4, years: list<item: struct<target_year: int64, label_tornado: int64, n: int64>>
                    child 0, item: struct<target_year: int64, label_tornado: int64, n: int64>
                        child 0, target_year: int64
                        child 1, label_tornado: int64
                        child 2, n: int64
                child 5, skipped_optional_tables: list<item: string>
                    child 0, item: string
                child 6, feature_matrix_mb: double
              created_utc: string
              checksums_sha256: struct<data/hrrr_event_control.parquet: string, data/preview_1000.csv: string, metadata/columns.json (... 60 chars omitted)
                child 0, data/hrrr_event_control.parquet: string
                child 1, data/preview_1000.csv: string
                child 2, metadata/columns.json: string
                child 3, metadata/manifest.json: string
                child 4, README.md: string
              to
              {'text': Value('string')}
              because column names don't match
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1347, in compute_config_parquet_and_info_response
                  parquet_operations = convert_to_parquet(builder)
                                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 980, in convert_to_parquet
                  builder.download_and_prepare(
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 882, in download_and_prepare
                  self._download_and_prepare(
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 943, in _download_and_prepare
                  self._prepare_split(split_generator, **prepare_split_kwargs)
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1646, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1832, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

text string
sample_key
source_package
sample_id
event_id
label_tornado
label_hail
label_wind
label_severe
label_family
source_report_type
sample_role
target_time_utc
target_date
target_year
hrrr_valid_time_utc
sample_lat
sample_lon
target_month
target_dayofyear
target_hour_utc
target_dayofyear_sin
target_dayofyear_cos
target_hour_sin
target_hour_cos
surface_native_era5_cape_jkg
surface_native_era5_cin_jkg
surface_sp_pa
surface_mslp_pa
surface_t2m_k
surface_tcc_fraction
surface_tcwv_kgm2
surface_td2m_k
surface_u10_ms
surface_v10_ms
surface_wind10_dir_deg
surface_wind10_speed_ms
hrrr_composite_reflectivity_dbz
hrrr_contamination_cloud_flag
hrrr_contamination_lightning_flag
hrrr_contamination_precip_flag
hrrr_contamination_reflectivity_flag
hrrr_high_cloud_fraction
hrrr_lightning
hrrr_low_cloud_fraction
hrrr_medium_cloud_fraction
hrrr_native_srh_0_1km_m2s2
hrrr_native_srh_0_3km_m2s2
hrrr_native_storm_motion_dir_deg
hrrr_native_storm_motion_speed_ms
hrrr_native_storm_motion_u_ms
hrrr_native_storm_motion_v_ms
hrrr_precip_rate_kgm2s
hrrr_reflectivity_1km_agl_dbz
hrrr_reflectivity_4km_agl_dbz
hrrr_total_precip_kgm2
hrrr_vil_kgm1
hrrr_wind_gust_ms
profile_levels_t_k_50hpa
profile_levels_t_k_75hpa
profile_levels_t_k_100hpa
profile_levels_t_k_125hpa
profile_levels_t_k_150hpa
profile_levels_t_k_175hpa
profile_levels_t_k_200hpa
profile_levels_t_k_225hpa
profile_levels_t_k_250hpa
profile_levels_t_k_275hpa
profile_levels_t_k_300hpa
profile_levels_t_k_325hpa
profile_levels_t_k_350hpa
profile_levels_t_k_375hpa
profile_levels_t_k_400hpa
profile_levels_t_k_425hpa
profile_levels_t_k_450hpa
profile_levels_t_k_475hpa
profile_levels_t_k_500hpa
profile_levels_t_k_525hpa
profile_levels_t_k_550hpa
profile_levels_t_k_575hpa
profile_levels_t_k_600hpa
profile_levels_t_k_625hpa
profile_levels_t_k_650hpa
profile_levels_t_k_675hpa
profile_levels_t_k_700hpa
profile_levels_t_k_725hpa
profile_levels_t_k_750hpa
profile_levels_t_k_775hpa
profile_levels_t_k_800hpa
profile_levels_t_k_825hpa
profile_levels_t_k_850hpa
profile_levels_t_k_875hpa
profile_levels_t_k_900hpa
profile_levels_t_k_925hpa
profile_levels_t_k_950hpa
profile_levels_t_k_975hpa
profile_levels_t_k_1000hpa
profile_levels_t_k_1013hpa
profile_levels_td_k_50hpa
profile_levels_td_k_75hpa
profile_levels_td_k_100hpa

End of preview.

Severe Weather Environment HRRR v1

This dataset contains HRRR event/control severe-environment samples for tornado, hail, wind, and spatial-control experiments. It is the high-resolution short-range component of the severe-environment data release.

This is a research dataset, not operational guidance. It is not an official NOAA, NWS, SPC, or HRRR product, and it should not be used as a stand-alone warning or forecast system.

Files

data/hrrr_event_control.parquet: primary HRRR feature table.
data/preview_1000.csv: small preview.
metadata/manifest.json: counts and source metadata.
metadata/columns.json: column list.
metadata/checksums.sha256: checksums.

Rows: 370,160
Columns: 816

Labels

Use the binary label columns directly:

label_tornado
label_hail
label_wind
label_severe

Label-family counts:

hail_nontor: 74,706 rows
wind_nontor: 124,390 rows
spatial_control: 57,014 rows
tornado_positive: 114,050 rows

Features

The table includes surface fields, pressure-level thermodynamic and wind fields, profile-extra fields, hydrometeor/profile-extra fields, and native HRRR diagnostics. Feature columns are intentionally kept wide for direct ML ingestion.

Suggested Uses

Tornado-vs-non-tornado classification with hail/wind severe cases as hard negatives.
Multihazard classification with separate tornado, hail, wind, and severe heads.
Feature ablation across surface-only, pressure-level-only, derived-diagnostic, hydrometeor/profile-extra, and full feature sets.
Testing clean-inflow and contamination-screening choices by filtering on available sample metadata and upstream manifests.

QC And Construction

Rows come from the local HRRR v2+ and hail/wind non-tornado pipelines.
Hail/wind hard negatives were built upstream with tornado-report exclusion logic, including a 25-mile and +/-2-hour tornado-report exclusion rule in the source hail/wind non-tornado package.
Failed model extractions were not backfilled with another model system.
The clean-inflow screen is automated and should not be treated as manual radar-mode classification.
Use time-based splits for honest testing; random row splits can overstate skill.

Recommended baseline split: train through 2021, validate on 2022-2023, and test on 2024+.

Year Coverage

2016, label_tornado=0: 9,096 rows
2016, label_tornado=1: 1,334 rows
2017, label_tornado=0: 27,614 rows
2017, label_tornado=1: 13,242 rows
2018, label_tornado=0: 26,007 rows
2018, label_tornado=1: 10,032 rows
2019, label_tornado=0: 28,281 rows
2019, label_tornado=1: 13,878 rows
2020, label_tornado=0: 26,256 rows
2020, label_tornado=1: 10,024 rows
2021, label_tornado=0: 26,950 rows
2021, label_tornado=1: 12,288 rows
2022, label_tornado=0: 26,600 rows
2022, label_tornado=1: 11,066 rows
2023, label_tornado=0: 27,745 rows
2023, label_tornado=1: 12,158 rows
2024, label_tornado=0: 29,745 rows
2024, label_tornado=1: 17,084 rows
2025, label_tornado=0: 27,708 rows
2025, label_tornado=1: 12,728 rows
2026, label_tornado=0: 108 rows
2026, label_tornado=1: 216 rows

Caveats

These are model-environment samples, not direct observations.
Severe-weather reports have spatial, temporal, and population biases.
A non-tornado label means no tornado report matched the configured upstream exclusion rules; it does not prove that no tornado occurred.
Storm mode is not a complete target in this table.
HRRR fields describe the sampled model environment. They do not resolve individual tornado-scale processes.

Attribution

This package is derived from public severe-weather reports and public numerical weather prediction data processed through a local rustwx severe-environment pipeline. Cite the original data providers and this dataset when using it in research or downstream model training.

Downloads last month: 42

Total file size:

526 MB