Track entry and quality
Entry is tested anywhere along the trajectory, including between track points.
Timing and regions
Genesis region
Lysis longitude sector
Route archetype
Dominant 24 h precipitation region
A system is assigned to the analysis box with its largest peak regional precipitation.
Intensity, precipitation and trajectory
Climate and circulation at genesis
These are contemporaneous regime descriptors, not causal attribution. Monthly indices use the genesis month; RMM uses the genesis date.
System explorer
Tracks and density
Every map layer is selectable: clicking near a line or density cell chooses the geometrically nearest filtered trajectory. Open a disturbance dossier to step through its track-centred diagnostics and contemporaneous ERA5 weather.
Selected-system evolution
Choose up to three variables. Drag slider to move map and weather time.
Select a disturbance to inspect its evolution.
Accessible selected-system values
Subset evolution
Solid line: median · filled band: IQR · dashed line: all-WD median.
Accessible subset summary
Selected-system vertical structure
Track-centred time–pressure section: vorticity at seven levels from 850 to 300 hPa; other fields at 850, 700 and 500 hPa. Drag the dashed marker to move map and weather time.
Selected-system precipitation footprint
ERA5 precipitation accumulated across the published WD lifetime. The 1° display grid does not partition precipitation into liquid and solid phases.
Climatology
Counts describe genesis events in the filtered subset. They are catalogue diagnostics, not a trend attribution.
Annual genesis count
Systems beginning in each calendar year.
Accessible annual data
Seasonal cycle
Monthly genesis count and mean peak upper-tropospheric vorticity.
Accessible monthly data
Precipitation-impact regions
Dominant analysis region among filtered systems.
Accessible impact-region data
Genesis density
Unique genesis systems per one-degree cell; colour intensity uses a square-root count scale.
Winter activity and multi-WD spells
Daily activity from November to April by winter year; outlined cells belong to objective same-impact-region spells.
Accessible winter activity summary
Catalogue diagnostic extremes
Extremes are internal to this ERA5-derived catalogue and are not authoritative meteorological records.
Data, definitions and reproducibility
The catalogue grain, upper-level diagnostic and regional-impact definitions are kept explicit.
Catalogue and time
WD v7 contains 49,200 regional candidate trajectories and 1,493,423 three-hourly track points from January 1950 through December 2025.
track_id identifies one Lagrangian trajectory. Genesis is its first published track point and lysis its last. Display names use genesis year and within-year genesis order, for example WD 2017 042; exports retain the catalogue ID.
Year filters use calendar year; select exact dates to isolate a particular winter. A meridian crossing is the first linearly interpolated crossing of the selected longitude; trajectories and charts remain on their original clocks.
The base catalogue retains trajectories lasting at least 24 h with peak T42 vorticity ≥2 × 10⁻⁵ s⁻¹ and a track point in 55–90°E, 20–50°N. Entry filters test the complete trajectory: an eastward meridian crossing, entry into a box, or entry into a latitude band east of a longitude. Short gaps up to 9 h are interpolated; boundaries are inclusive.
The initial selection uses DJFM genesis, lifetime ≥48 h, eastward entry across 70°E at 20–50°N, and the quality screen. The screen requires ≥80% temporal coverage, ≤20% ambiguous links and reverse-check disagreements, and no detector-boundary or source-time censoring flags. All these selection controls can be relaxed.
Positions and diagnostics are derived from ERA5. This is a research catalogue, not an operational warning service or official best track.
WD-specific physics
- Upper-tropospheric vorticity: track-centred relative vorticity averaged through the 450–300 hPa layer and spectrally truncated to T42, reported in 10⁻⁵ s⁻¹.
- 24 h precipitation: track-centred and regional accumulated precipitation, reported in mm.
- Path length: great-circle distance summed between consecutive observed track points.
- Percentiles: fixed against all 49,200 catalogue trajectories; filters do not rescale them.
The evolution selectors expose the core series plus 70 lazily loaded weather, vortex and link diagnostics. Up to three line variables use separate axes while precipitation bars remain visible. The vertical vorticity panel includes 850, 700, 500, 450, 400, 350 and 300 hPa; other fields use 850, 700 and 500 hPa. The tracked T42 maximum and native ERA5 regional means are distinct diagnostics.
The atlas intentionally omits LPS-specific pressure-deficit classes, IBTrACS matching, BSISO filters and Indian-state precipitation fills.
Regions, routes and sequences
Genesis regions reproduce the four-cluster method behind Figure 5 of the 2025 WD review. Classification uses its transformed longitude–latitude cluster centres; points outside the original 99.5% distance envelopes are Other.
Route archetypes are an exploratory eight-cluster k-means classification of longitude and latitude interpolated at nine common elapsed-life fractions. The labels describe cluster medians and are not meteorological event classes.
Multi-WD spells link systems within the same winter and dominant precipitation-impact region when the next genesis is no more than 72 h after the latest lysis. The five impact boxes overlap and are analytical conventions, not administrative boundaries.
Analogues compare full-trajectory shape, intensity, lifetime, path length and regional precipitation impacts across different catalogue years.
Maps and impact context
Density counts each track at most once per one-degree cell, but remains selectable through the underlying trajectories. The slider and selected-system chart both move the track point and contemporaneous weather frame.
Independent ERA5 weather backgrounds provide positive 350-hPa vorticity, trailing 24 h precipitation, 500-hPa wind speed, temperature and specific humidity, and mean-sea-level pressure. A staged field is enabled only after its complete 1950–2025 archive passes validation.
The selected-system footprint accumulates hourly ERA5 precipitation from published genesis through lysis on a 1° grid. It provides spatial context but does not partition liquid and solid precipitation or estimate impacts.
Jet and circulation context
The jet relationship uses the daily latitude and speed of the maximum 200-hPa wind between 20 and 55°N at the WD longitude from 50 to 80°E. Times absent from the dedicated jet source, including 29–31 October 2023, remain unavailable.
Genesis-month NOAA PSL ONI, NAO, AO and PNA values use ±0.5 category thresholds. Daily MJO phase comes from the Bureau of Meteorology RMM index; amplitude below 1 is weak. These filters describe contemporaneous regimes and do not imply causation.
Comparisons and composites
The month definition can follow genesis, any active month, peak vorticity or peak precipitation. Peak ties use the first occurrence. Annual rates divide event counts by the actual selected calendar-day exposure; WD-days integrate elapsed lifetimes within those dates. Systems overlapping a date range need not have their genesis or peak inside it.
Pinned references retain their filter selection while the current subset changes. Climate-state comparisons use all WDs in the same time/month selection, including unavailable indices in the denominator.
Matched IMDAA T42 and ERA-Interim T63 tracks use legacy detectors. Matches require temporal overlap and geographical agreement; ambiguous matches are withheld. Storm-centred sections average nine actual track points nearest equally spaced elapsed times. Precipitation composites average complete UTC-day totals, centred near noon, separately from the lifetime footprint. Downloads retain cell-level sample counts.
Operational forecasts and climate projections
Experimental WD forecasts derive positive 450–300-hPa relative vorticity from global winds, then apply T42 truncation, a forecast-specific maxima detector and spherical predictive linking at six-hour intervals. Tracks must last at least 36 h, progress eastwards and intersect 60–80°E, 20–50°N. Track efficiency is endpoint great-circle displacement divided by total path length; values below 10% are rejected.
GFS supplies all four layer levels. GEFS, IFS and AIFS require interpolation of 350 and 450 hPa from 300, 400 and 500 hPa. Precipitation is the exact trailing 24 h accumulation and is unavailable before lead 24 h. Ensemble groups average contemporaneous matched member tracks; membership can vary with lead time. Group support is not a calibrated probability.
Forecast sources: NOAA/NCEP GFS, GEFS and ECMWF Open Data (IFS/AIFS; CC-BY-4.0). The displayed guidance is derived and modified, not an official forecast product.
The climate-change tab uses the existing CMIP5 WD archive, with source-time coverage checked in each model’s native calendar. It compares equal-weight paired models, not a rerun of ERA5 v7 or a CMIP6 projection.
Download and provenance
Exports carry the current filter state, units and WD track identifiers.
Full v7 catalogue
Track points: Parquet · CSV (gzip) · Coordinates only (CSV, gzip)
Track summaries (CSV) · Weather variable definitions · Catalogue provenance
Source and archived releases
Archived dataset releases · asset checksums · WD climatology and jet-context paper
Build
This static deployment loads compact gzip assets in the browser; large diagnostics and selected-system footprints are fetched only when needed.