Revised 27 September 2026. Re-run on the current HURDAT2 file, hurdat2-nepac-1949-2025-091426.txt (the June run used an earlier download, 1,262 storms; the current file gives 1,260). The satellite-era northernmost-latitude lean moves from +0.24°/decade at p=0.11 to +0.26°/decade at p=0.080: still short of the four-test Bonferroni threshold (p<0.0125), still not a finding. Other trends move in the second decimal; no verdict changes.
The companion finding — an earlier, busier storm season in the 1949–2025 record, which from 1971 or 1978 shows no detectable change in first arrival, storm count or storm-days — raised an obvious follow-up: are the tracks themselves shifting geometry? If storms were recurving further north over time, more of them would reach the 1,500 km swing zone. Four metrics, 77 years, 1,260 storms.
Short answer: two metrics show significant long-record trends, but both are driven by the 1950s-to-1970s period when geostationary satellites were added and the catalog began capturing short-lived equatorial storms that were previously missed. Restrict the analysis to the satellite era (1978–2025) and all four trends become statistically inconclusive. The honest reading is: no confirmed geometric northward shift — though a weak northward lean in the satellite era (northernmost latitude: +0.26°/decade) sits at p=0.08, short of confidence but worth watching.
The tracks have not demonstrably shifted north in the satellite era. The long-record signals reflect surveillance improvement, not changed geometry.
Genesis latitude trended strongly southward across the full 77 years (p<0.001) — but that reflects the 1950s-to-1970s expansion of storm detection to shorter-lived equatorial systems. Post-1978, the same metric shows a weak, insignificant northward lean. The same detection change drives the phenology trends: on the satellite era no change in first arrival or storm count is detected, and the track geometry change is not confirmed either.
Two apparent trends — both driven by the pre-satellite period.
OLS regression of annual means; all four tested simultaneously (Bonferroni threshold: p<0.0125).
Restricting to the geostationary satellite era — when catalog quality is consistent — removes all significant trends. Genesis latitude: +0.14°/decade, p=0.16. Recurvature latitude: −0.04°/decade, p=0.87. Northernmost latitude: +0.26°/decade, p=0.080 — the weakest northward signal in the data, not yet significant but directionally consistent with a poleward expansion hypothesis.
The satellite-era northernmost-latitude trend is the one metric that would most directly explain more storms reaching the Saladita swing zone. At p=0.08 with n=48 years, it is suggestive but inconclusive. Another decade of data would likely settle it.
A large shift in the 1960s–70s, stable since.
The 1950s mean genesis was ~16°N. By the 1970s it was ~12.9°N — a 3° drop that coincides exactly with the expansion of geostationary satellite coverage, which detected shorter-lived, more equatorial storms that aircraft and ship reports missed. Since 1980, the mean has been nearly flat at 13–14°N.
Where storms turn poleward — no clear recent trend.
About 26% of EPAC storms show detectable recurvature by the algorithm (westward→eastward zonal motion with concurrent poleward motion). Among those, mean recurvature latitude was ~21°N in the 1950s and has drifted to ~17–18°N — again driven by the pre-satellite period. Since 1980 there is no clear direction.
Surveillance bias dominates the long record. The HURDAT2 catalog before 1966 is significantly incomplete — no geostationary satellites, sparse ship and aircraft reports. The 1950s mean genesis latitude of ~16°N is not a reflection of where storms formed; it is a reflection of which storms were detected. The big apparent southward shift in the 1960s–70s is catalog improvement, not climate change.
Recurvature is ambiguous. About 74% of EPAC storms track steadily westward and dissipate over cooler water or land — they never recurve. The recurvature subsample (26% of storms) is not a random draw from the full population; it skews toward storms that survive long enough to encounter mid-latitude steering flows. Trends in that subsample may reflect changing storm duration, not track curvature per se.
Annual averages mask within-year variability. A single long-tracked recurving storm can shift the annual mean recurvature latitude by several degrees. High interannual variance means confidence intervals are wide.
Four simultaneous tests. Bonferroni threshold is p<0.0125. The genesis-latitude result (p<0.001) clears it but is a surveillance artifact. The recurvature result (p=0.016) sits just above the threshold and is pre-satellite driven. Nothing clears threshold in the satellite era.
This matches the phenology finding. Over 1949–2025 the first storm within 1,500 km arrives earlier (−2.3 days a decade) and storms are more frequent, but both trends come from the pre-satellite years. From 1971 or 1978, when the catalogue is consistent, no change in first arrival, storm count or storm-days is detected (p 0.13–0.97). Peak wind within 500 km keeps a similar slope (+5 to +6 kt a decade) but is not significant on the satellite era (p≈0.07–0.09). Track geometry (where storms form and how they curve) is a separate question with the same answer so far.
On the satellite era neither the phenology trend nor the geometry trend is confirmed. A decade more of satellite data at the current +0.26°/decade northernmost-latitude rate would likely decide it.Synthesis · June 2026
Sources and method
NHC HURDAT2 NEPAC 1949–2025, file hurdat2-nepac-1949-2025-091426.txt (1,260 EP+CP storms; re-run 27 September 2026). Per-storm metrics derived from 6-hourly best-track fixes: genesis = first recorded fix; max latitude = northernmost fix over full track; mean heading = circular mean of bearings between consecutive fixes; recurvature = first inflection where 2-step net zonal displacement flips westward→eastward with concurrent poleward motion (2-fix smoothing window). Annual values = mean across all storms forming that year. Linear trends by OLS; p-values via normal approximation (~77 df). Script: scripts/analyze_hurricane_tracks.py → functions/api/_findings_hurricane_tracks.js.
Satellite era defined as 1978–present (GOES-1 operational late 1975; full geostationary EPAC coverage effectively established 1977–1978). Pre-1966 caveat per NHC HURDAT2 documentation.