This page summarises the important features of the EFAS seasonal outlook products. Further information on the details of the different steps generating these products is available at CEMS-flood sub-seasonal and seasonal forecast product generation methodology.

'Seasonal Outlook' and 'Seasonal Outlook - Basins' layers

The EFAS seasonal outlook layers show maps of the 'expected to happen' forecast anomaly and related uncertainty for the next 7 months for all river pixels over 250 km2 (river network layer called 'SeasonalOutlook' on the mapviewer) and for 204 major river basins over the EFAS domain (basin layer called 'SeasonalOutlook - Basins' on the mapviewer), using colour-coded categories (see Figure 1). The products are updated once a month and published around the 8th of the month, showing:

  • forecast anomaly expressed in one of 5+2 categories from extreme low to extreme high, which are derived from specific model climatology percentiles
  • forecast uncertainty is categorised into low, medium and high levels, based on the spread of the ensemble member extremity levels

The forecast anomaly and uncertainty are based on the extremity levels of the monthly averaged river discharge of the ensemble members in the context of the model climatology. The model climatology is range-dependent (change with the lead time) and generated from 20 years of reforecast hydrological simulations (2004-2023).

The forecast evolution of the anomaly is also given in popup window products for fixed reporting points (grey squares along the river network) and basin-representative points (black circles, one point for each of the 204 predefined basins).

Figure 1: EFAS seasonal outlook on the EFAS-IS map viewer. A lead-time selection panel on the mapviewer allows the users to select the forecast month.


Reporting point pop-up window product

From the river network map, the reporting point pop-up products can be accessed by clicking on the point markers on the river network layer (grey squares showing the EFAS fixed reporting points - see more in CEMS-Flood reporting points and dynamic point generation algorithm, and the black circles showing the basin-representative points - see more in CEMS-flood sub-seasonal and seasonal basins and representative stations). The pop-up products contain metadata information about the stations, a hydrograph with the evolution of the climatological, antecedent and forecast conditions (see Figure 2), and the probability values for the extended 7-value anomaly categories (the 5 main categories shown on the maps and hydrographs, extended by two additional categories within the 'Near normal' category) and the evolution of these probabilities from previous forecast runs.

The hydrographs are retrospectively updated each month once, when the proxi-observed river discharge values become available for one further month. This happens usually on the 2nd of each month, but sometimes it can be delayed. This means, one extra black dot will appear on the hydrograph showing what actually happened during that month. One by one, each month, a further black dot will appear in the forecast section of the hydrographs. The users are encouraged to go back and check previous forecasts to see how well earlier forecasts predicted the anomalies. 

Figure 2: Discharge hydrograph showing weekly-averaged antecedent and climate conditions and forecast evolution out to 7 months.


Figure 3: Forecast probability evolution table with (a) probabilities values (as numbers) for all anomaly categories with the extended 7-category version ('Extreme low' (EL), 'Low' (L), 'Bit low' (BL), 'Normal' (N), 'Bit high' (BH), 'High' (H) and 'Extreme high' (EH), as left to right); (b) ‘expected to happen’ anomaly category as the one coloured cell at each monthly forecast lead time, the colour dependent on the anomaly and on the related forecast uncertainty category (how uncertain the forecast is; either low, medium or high). Each main column (7-column groups) represent a monthly forecast lead time and the 7 rows show the 7 most recent forecast start dates (the most recent being on top), which are always the 1st of the month (indicated only by month and year).

 

Different legends to help interpreting the forecast signal

Figure 4: Combination of anomaly and uncertainty categories, defined by the mean and standard deviation of the ensemble member ranks, used in the EFAS Seasonal Outlook products. Note, only the 5 main categories are used in the map layers and the hydrograph, with the near-normal categories merged together into 'Near normal'.

Table 1: Anomaly categories defined by the mean of the ensemble member ranks, used for classifying the forecast signal (colours) in the EFAS Seasonal Outlook products.

Table 2: Uncertainty categories defined by the standard deviation of the ensemble member ranks, used for classifying the forecast uncertainty (colour intensity) in the EFAS Seasonal Outlook products.


EFAS seasonal products up to EFAS v5.3 (before 12 March 2025)

The EFAS Seasonal outlook layer shows the maximum area-averaged probability of unusually high (> 90th percentile, blue) or low (<10th percentile, red) weekly river flow occurring during the 8-week forecast horizon for 74 major river basins over the EFAS domain, using colour-coded categories:

  • Dark blue: high (> 90%) probability of high flows
  • Medium blue: medium (75-90%) probability of high flows
  • Light blue: low probability (50-75%) of high flows
  • White: Near normal flow (below 50% for both low and high flow)
  • Light red: low probability (50-75%) of low flows
  • Medium red: 75-90% probability of low flows
  • Dark red: > 90% probability of low flows


Pop-up window

Note that the area-averaged discharge is scale dependant and values will change with resolution change, between EFAS v4 and EFAS v5.