Multi-model Forecast Maps
Multi-model median (ECMWF SEAS5 + Météo-France) forecast SST at leads L1-L6 for the latest init. The field is the variance-adjusted SST anomaly - detrended for global warming and standardized by the local observed SST standard deviation (σ units), so it shows how anomalous each region is relative to its own variability. Precipitation indicators (total, extremes, drought) will be added once the daily forecast fields are processed.
Drag to rotate · scroll to zoom. Values shown are the equal-weight multi-model median of the selected variable per cell.
-
Each centre shown as a coloured line (ensemble median P50) with shaded envelope (P10 - P90). Pool: all ensemble members × all country grid cells, per centre, per lead month.
About
Each contributing centre's ensemble mean is computed on its native grid and regridded to a common 1°×1° mesh. Centres are combined with equal weight (one centre = one vote), following our multi-model aggregation convention. The per-cell value shown is the cross-centre median of the selected variable at the chosen lead.
Maps for the current forecast only; past inits are not stored as maps (they're available as time-series on the Hindcasts tab). XRO is excluded from maps - it's an index-only model.