Australia's alpine ski season typically runs from the June long weekend through to early October, with resorts across the Snowy Mountains (Perisher, Thredbo, Charlotte Pass), Victoria's High Country (Falls Creek, Mt Hotham, Mt Buller), and smaller fields like Selwyn and Mt Baw Baw all dependent on natural snowfall supplemented by snowmaking. The season's quality hinges on a narrow window of cold, wet southerly weather — a combination that is strongly influenced by large-scale climate drivers including ENSO, the Indian Ocean Dipole, and the Southern Annular Mode.
This forecast is modelled exclusively on Spencer's Creek snow depth records at 1830 m elevation in the Snowy Mountains — the longest continuous alpine snow record in Australia. It is most applicable to high-elevation terrain in the Kosciuszko region. Lower-elevation resorts such as Mt Buller (1,707 m summit) or Mt Baw Baw (1,564 m) typically receive less snow and may diverge significantly from this outlook, particularly in marginal seasons. Snowmaking capacity at individual resorts also means on-piste conditions can differ substantially from natural depth indicators.
⚠ Mid-season update (14 July 2026): The season is now underway and the climate picture has shifted materially since March. Scroll past the original forecast tool for a full mid-season review, including how conditions are tracking against our pre-season model. For day-to-day alpine weather, check the ECMWF and GFS forecasts for the Snowy Mountains region on AerisCast.
Mid-Season Update: 14 July 2026
Six weeks into the season, the picture is now much clearer than it was in March — and unfortunately, the risk scenarios flagged in our original outlook have largely materialised. The 2026 Australian snow season has had one of its worst starts in the 72-year Spencer's Creek record, and the climate drivers behind it are now locked in for the remainder of winter.
What has happened so far
The season opened with a promising 20 cm snowfall just before the King's Birthday long weekend in early June, but warm, moist air masses from the tropics quickly washed it away. June was characterised by above-average temperatures — Perisher's average minimum and maximum for June ran approximately 2°C above their long-term means — and frequent rain rather than snow. Spencer's Creek was effectively snowless on 1 July, only the second time that has occurred in the entire record dating back to 1954.
A meaningful cold front arrived in the first week of July, depositing 20–40 cm across the resorts and finally giving the season a working base. A second, smaller system on the weekend of 11–12 July added another 10–15 cm at mid-elevations, with slightly more up high. As of the latest Snowy Hydro reading on 7 July, Spencer's Creek sat at approximately 35 cm — well below the long-term average of around 100 cm for that date. The snowpack has since consolidated in cold, clear weather, forming a firm base — but it remains thin.
For a sense of how the coming days look at resort level, the extended ECMWF and GFS forecasts for the Snowy Mountains on AerisCast show the detail hour by hour, up to 15 days out. Comparing both models side by side is especially useful in marginal seasons like this one, where the difference between rain and snow often comes down to 1–2°C — exactly the kind of disagreement the two models can flag. For a deeper look at how ECMWF and GFS differ and when to trust each one, see our ECMWF vs GFS explainer.
The climate drivers — what changed
When we published the original outlook in March, ENSO was neutral and the question was whether a weak El Niño would emerge. It did — and then some. As of mid-July, El Niño is firmly established, with the Niño 3.4 index at +1.24°C (well above the +0.80°C threshold) and the SOI strongly negative at −25.2. NOAA now assigns a 97% probability that El Niño will persist through early 2027, and multiple forecast agencies are pointing toward a strong to very strong event — potentially among the highest since 1950.
The Southern Annular Mode (SAM) has been positive for most of the season so far, recently hitting a three-year high. Positive SAM pushes the band of westerlies — and the cold fronts they carry — south toward Antarctica, keeping them away from the Australian Alps. The SAM index does now appear to be trending back toward neutral, which would be a welcome shift, but the damage from the persistently positive phase in June is already done.
The Indian Ocean Dipole (IOD) remains neutral for now, with the index at −0.02°C in late June. However, models increasingly favour a positive IOD developing through winter and spring, with probability exceeding 60% by July and rising above 90% by September. A positive IOD combined with El Niño would be the worst-case combination for snow — historically associated with the leanest seasons in the record.
In short: The climate state has shifted from the “neutral ENSO, wait and see” picture of March to a confirmed El Niño with positive SAM and a likely positive IOD — all three drivers pointing the wrong way for snow. This is essentially the risk scenario our model assigned a 15% probability back in March.
How the season is tracking against the forecast
Our March model produced a median peak depth forecast of 174 cm (P10–P90 range: 111–264 cm), weighted across four climate scenarios. The scenario most consistent with what has actually eventuated — El Niño + positive IOD — carried a pre-season mean of 142 cm and a P10 of just 94 cm.
At 35 cm in early July, the season is tracking below even the pessimistic end of that scenario. For context, the long-term average depth at Spencer's Creek in early July is around 100 cm, and the 2025 season (a good La Niña year) was above 150 cm at the same point. It is not impossible for the season to recover — the famous 1991 season started poorly before sustained falls from mid-July took the peak to nearly 300 cm — but it would require a significant and sustained shift in the synoptic pattern that the current climate drivers do not favour.
Realistically, the best remaining hope is for short negative-SAM windows through late July and August to deliver a handful of meaningful cold fronts. These are difficult to predict more than a couple of weeks out, which is exactly why watching both the ECMWF and GFS models on AerisCast day to day matters more than ever — when the models agree on an approaching front, confidence is high; when they diverge, plan cautiously. Check forecasts for Sydney, Melbourne, or Canberra and the Snowies for trip-planning detail.
Updated outlook for the remainder of the season
Given the confirmed El Niño, the likely positive IOD, and the poor base established to date, we now consider a peak season depth in the range of 100–160 cm to be the most likely outcome, with significant risk of finishing below 120 cm if no substantial fronts arrive through August. An above-average season (200+ cm) remains theoretically possible but would require an exceptional and sustained pattern break that the broad-scale drivers do not support.
The resorts will lean heavily on snowmaking for the remainder of the season. Cold, dry nights — which the current high-pressure-dominated pattern does at least deliver between fronts — are ideal for snowmaking, and on-piste conditions at the major resorts will often be substantially better than the natural depth readings suggest.
Original Pre-Season Forecast (March 2026)
How the forecast works
The simulation draws on Spencer's Creek snow depth records dating back to 1954 — 71 years of continuous alpine observations managed by Snowy Hydro. This dataset captures the full range of Australian alpine variability, from the record-breaking depths of 1974–75 and 2010–11, to the meagre seasons of 1982 and 2019.
Rather than a single-number forecast, we use Monte Carlo simulation — running 100,000 probabilistic seasonal trajectories — to express the genuine uncertainty in seasonal forecasting. Each iteration draws from a log-normal distribution calibrated to the historical record, weighted by the current climate state.
Climate scenario weighting (as at March 2026)
The original forecast was composed of four climate scenarios, each assigned a probability weight based on the Bureau of Meteorology's outlook at the time:
- Neutral ENSO + neutral IOD (50%): The baseline scenario. Historically produces average seasons around 190 cm with wide variance.
- Weak El Niño + neutral IOD (25%): A moderate headwind. Historical mean around 172 cm.
- El Niño + positive IOD (15%): The risk scenario — historically associated with the leanest seasons, mean around 142 cm. ← This is essentially what has eventuated.
- Neutral ENSO + negative IOD (10%): The upside scenario. Mean around 225 cm historically.
Trend correction
Australian alpine snowfall has been declining at approximately 5.5 cm per decade since the 1950s. This structural trend is applied as a correction to the forward-looking forecast, which is why the simulated median (174 cm) sits below the raw historical mean of 182 cm.
The original P10–P90 forecast range for 2026 spanned 111 to 264 cm. Given the confirmed El Niño and poor start, the realistic range has narrowed significantly. The season is currently tracking at or below the P10 line from the March model.
What to Watch From Here
The key variables for the rest of the season are whether the SAM returns to neutral or negative (which would allow cold fronts to reach the Alps), whether the IOD formally enters positive territory (which would further suppress moisture), and whether any individual storm systems manage to break through the El Niño pattern. Historical analogues like 1991 show that individual storms can still deliver big totals even in otherwise poor seasons.
For day-to-day monitoring, AerisCast remains the tool we use ourselves — comparing the ECMWF and GFS forecasts for the Snowy Mountains region gives you a clear read on whether the models agree on incoming fronts, and the hourly detail helps distinguish rain from snow on those critical borderline days. You can also track conditions for the Victorian High Country via Melbourne or search any location worldwide.
We will continue to update this outlook as the season progresses.
About This Analysis
This outlook is produced by Aeris Spatial as part of our ongoing work in climate data visualisation and spatial analysis — capabilities we apply to professional client projects in planning, development, agriculture, and infrastructure. Our core business is GIS consulting and geospatial mapping across NSW, and this analysis is built on the same data infrastructure that powers AerisCast ⛅, our free weather forecast platform showing ECMWF and GFS side by side for any location on earth. For bespoke climate analysis or custom data tools, see our Climate & Data Solutions service.
If you have a project that could benefit from data visualisation, GIS processing, or custom mapping — get in touch.