October 2026 is projected to be the warmest October on record, with Climate Dashboard estimating a global temperature anomaly of about 1.76°C above the preindustrial level. The estimate uses 28 days of the month and remains a projection, rather than an observed, completed-month temperature value.
The dashboard gives October an estimated first-place ranking and displays a projected interval from 1.73°C to 1.78°C above the preindustrial level. It labels the result as a Monte Carlo projection, refreshed daily. The ranking therefore describes the expected outcome for October, not a record already established from a complete month of observations.
Researchers Andrew Dessler and Zeke Hausfather have projected October at around 1.76°C above the preindustrial level, based on Copernicus Climate Change Service data. That rounded figure matches the central estimate displayed on the dashboard. The comparison is with the preindustrial level and concerns the global temperature anomaly, rather than a local temperature reading.
Observed data and the projected monthly result
The dashboard’s monthly projection data, labelled as regenerated on 30 October 2026, separates the month-to-date anomaly from the estimate for the full month. It reports a month-to-date figure of 1.7372°C above the preindustrial level and a projected monthly anomaly of 1.7591°C. These are distinct values serving different purposes within the calculation.
The month-to-date figure reflects the observations included so far. The projected figure combines those observations with forecast information to estimate the monthly result. The data’s accompanying description says the calculation blends 28 days of month-to-date observations with an ECMWF EC46 46-day ensemble, initialised on 29 October 2026.
The forecast component is therefore part of the basis for the headline estimate. Although most of October is represented by observations in the calculation, the full-month figure still depends on a blend of observed and forecast information. The dashboard explicitly leaves the current monthly mean blank because the month is incomplete.
The same data gives a 5th-percentile projection of 1.7349°C and a 95th-percentile projection of 1.7832°C. Those values correspond to the rounded interval of 1.73°C to 1.78°C shown on the dashboard. They describe the spread of the projection, rather than a range of completed-month observations.
The more precise central value, 1.7591°C, rounds to the displayed estimate of 1.76°C. Its additional decimal places do not change the result’s status: it is still a projected anomaly. Likewise, the estimated first-place ranking is attached to that projection and should not be confused with the blank field for the incomplete monthly mean.
Where the dashboard’s figures come from
Copernicus says the dashboard automatically retrieves temperature data from C3S Climate Pulse. Climate Pulse provides near-real-time updates of key global climate variables from the Copernicus Climate Change Service, which is implemented by ECMWF for the European Commission. This identifies the underlying temperature data service used by the dashboard.
Copernicus also distinguishes that data supply from the dashboard’s forecasting work. It says the dashboard’s predictions and projections are based on Zeke Hausfather’s own model and El Niño forecasts. The use of C3S temperature data does not, by itself, make the dashboard’s projected monthly result a completed-month assessment from Copernicus.
That distinction is central to the October estimate. The dashboard draws on near-real-time climate information, but its record ranking remains an output of a projection. The published calculation sets out both the observations available for 28 days and the forecast ensemble used alongside them, making clear why the result is not yet an observed monthly value.
For now, the dashboard points to October 2026 taking first place among Octobers, at approximately 1.76°C above the preindustrial level. Its daily refresh and stated projection interval preserve the uncertainty around that expectation. The incomplete-month status remains explicit, with no observed full-month mean reported in the projection data.
