Climate-mode-conditioned exposure of a sporting mega-event: an event-based assessment of Rugby World Cup 2027 in Australia

Dimitri Defrance

The environmental risk of sporting mega-events is usually assessed from
host-city climatologies or, more recently, from tournament design (venue,
date, kick-off time). Both overlook a distinct axis: the interannual state
of the large-scale climate modes into which a given edition falls. We ask
to what extent the phase of the El Ni\~no--Southern Oscillation (ENSO) and
the Indian Ocean Dipole (IOD) shapes the mix of playing-condition regimes a
mega-event is exposed to, using the Rugby World Cup 2027 (Australia; the
first 24-team edition, 52 matches, 8 stadiums, 7 cities, October--November)
as a case study. Hourly ERA5-Land and ERA5 reanalysis is sampled around
each scheduled match ($\pm2$ days, kick-off window) over October--November
1996--2025; each realisation is assigned to one of five mutually exclusive
weather types (Fair, Hot, Wet, Windy, Compound) built from three hazard
families (thermal, wet, wind). The type distribution is estimated
conditionally on climate-mode phase. Moisture-driven matches
(Wet+Compound) roughly double under La Ni\~na and under negative IOD
relative to their opposite phases (year-level bootstrap contrasts +10.9
percentage points, 95% CI [+4.1,+17.4] for ENSO; +12.0 pp [+1.2,
+21.7] for IOD), whereas Windy matches are slightly more frequent in the
dry phases and heat is negligible at this latitude and season. The phase
sensitivity is spatially structured: north-eastern venues are ENSO-led,
south-eastern coastal venues IOD-led, and the west and tropical north are
weakly sensitive. Because mode phase is partially forecastable months ahead, the
phase-conditioned hazard profile becomes a planning variable with seasonal
lead time rather than an unforecastable background.