...
The generation of the sub-seasonal and seasonal forecast signal relies on few major steps. The process is illustrated by a flowchart in Figure 1. The revised sub-seasonal products cover calendar week periods (i.e. always Monday-Sunday), while the seasonal products are valid for whole calendar month periods. The forecast forecast anomaly and uncertainty signal is derived from the by comparing the real time forecast (top left section in Figure 1) to the 99-value model climate percentiles. The climatology is generated using reforecast over a 20-year period, which provides range-dependent climate percentiles that change with the lead time. The climate generation is described in Figure 1 in the bottom right corner. Further details of the real time forecast, reforecast and the generation of the climatologies are available on this page:
The revised sub-seasonal products cover calendar week periods (i.e. always Monday-Sunday), while the seasonal products are valid for whole calendar month periods. The forecast signal is derived from the relationship between the calendar weekly or monthly averaged river discharge and the climatological distribution of the same weekly- or monthly-averaged values. While this naturally works for the calendar months, the fixed calendar week lead times in the sub-seasonal allow the users to directly compare forecasts from different forecast runs, as the verification period is fixed onto the calendar weeks. This way, the evolution of the subsequent daily sub-seasonal forecast runs (always at 00 UTC) can be monitored by looking at the exact same verification period relationship between the calendar weekly or monthly averaged river discharge and the climatological distribution of the same weekly- or monthly-averaged values. While this naturally works for the calendar months, the fixed calendar week lead times in the sub-seasonal allow the users to directly compare forecasts from different forecast runs, as the verification period is fixed onto the calendar weeks. This way, the evolution of the subsequent daily sub-seasonal forecast runs (always at 00 UTC) can be monitored by looking at the exact same verification period.The forecast anomaly and uncertainty signal is derived by comparing the real time forecast (top left section in Figure 1) to the 99-value model climate percentiles. The climatology is generated using reforecast over a 20-year period, which provides range-dependent climate percentiles that change with the lead time. The climate generation is described in Figure 1 in the bottom right corner.
The following step in generating anomaly and uncertainty signals is to determine, how extreme the ensemble members of the forecast are in the context of the climatological behaviour, which is represented by the 99 percentiles and the corresponding 100 equally likely bins of the climate range. This is done by ranking each ensemble member and sorting them into one of the 100 climatological bins. This gives 51 ranks, that range from 1 to 100.
...