This glossary explains technical terms frequently used within the GloFAS documentation. It is complemented by the FAQ also available in this GloFAS User Guide.
| Term | Definition |
|---|---|
| Analysis | An analysis is our best estimate of reality, used as a proxy for observations which do not exist everywhere. It is the archived lead time 0 of a forecast. For example, Initial 2m temperature is an analysis. |
Archived forecast | Archived forecasts are past forecast runs; these differ in that they are not re-run when the model changes, so the model version is not consistent throughout the time period. |
Climatology | Set of reference simulations, including historical simulations forced with observations or reanalysis. Typically longer time periods of 20 to 30 years. |
Control forecast | Ensemble forecast sub-set only including simulation obtained from the control member of the corresponding Numerical Weather Prediction (NWP). The control member is the unperturbed forecast run at the resolution of the ensemble forecast. Recommended to be used in combination with the perturbed forecasts. |
ECMWF-ENS | Probabilistic meteorological forecast (ensemble forecast) consisting of 51 members run with the global ECMWF model. The spatial resolution is ~18 km up to 15 days, decreasing to 36 km from day 16 to 30. One forecast, called the control forecast, is run from the operational ocean and atmosphere ECMWF analyses. 50 additional integrations, the perturbed members, are made from slightly different initial atmospheric and oceanic conditions, which are designed to represent the uncertainties inherent in the operational analyses. |
Ensemble forecast (ENS) | A number of simulations (members) using the same NWP model, but varying the initial conditions and/or the tendencies. The ensemble members are perturbed (pushed away from the observed state) with regards to the best estimate of the initial conditions, and the ensemble spread should ideally reflect the model error. Ensemble forecasts are often used to express the uncertainty of a forecast by associating a probability with different forecast values. |
Ensemble streamflow prediction (ESP) | A forecasting method consisting of starting a hydrological model with the latest initial hydrological conditions (e.g. soil moisture, river level, snow cover, etc.) and forcing it with reanalysis (e.g. precipitation, temperature, evaporation, etc.). Each ensemble member will thus correspond to a year of historical meteorological observations. For example, to generate hydrological forecasts on the 1st of February 2022 for the next 7 months of lead time, the hydrological model would be started with the hydrological conditions observed on this day. Meteorological forcing would then be taken from each available year of meteorological observations for the 1st of February - 31st August (+7 months). |
Exceedance probability | The probability of exceeding a defined threshold. Within GloFAS, return period values for different discharge values are calculated and the number of ensemble members exceeding a return period threshold (e.g. 31 out of 51 members are above the 20-yr return period threshold) is expressed as exceedance probability (in this example 60% exceedance probability). |
Forced simulation | Hydrological time series simulation forced with observations, reanalysis or other proxy for observations. Typically used to generate initial conditions for a hydrological forecast and reference long-term run. The latter can be used to calculate return periods and for validation of forecasts. |
GloFAS-IS | GloFAS Information System. It is the Map Viewer that can be accessed from GloFAS homepage and that contains all layers necessary for the analysis of the forecasts. |
GloFAS Map Viewer | Another term for the GloFAS-IS, an interactive map for analyzing GloFAS forecasts. |
High-resolution forecast (HRES) | Single simulation with one meteorological forecast input. The high-resolution NWP input is usually better than ensemble forecasts in resolving local features, such as orographic precipitation, but small deviations in the model simulation, for example, geographical displacement, can cause large deviations in the hydrological output in comparison with observations. |
Hindcast | Hindcast is another commonly used term for reforecast; while these mean the same thing, hindcast is used less and less to be replaced with reforecast. |
Historical simulation forced with observations | Long (> 20 year) hydrological time series simulation forced with proxy-observed meteorological fields (e.g. reanalysis such as the ERA5 product). Typically used to generate reference statistics, for example, return period discharge magnitude against which to compare forecast simulations to generate e.g.reporting points, and to validate hindcasts. |
Lead time | The length of time between the issuance of a forecast and the future time step one is trying to predict (e.g. the time step the event peaks). |
LISFLOOD-OS | The hydrological model used within GloFAS. LISFLOOD-OS is an open source, GIS-based, distributed hydrological rainfall-runoff-routing model specifically designed to be used in large scale catchments. The source code and the documentation are available from https://github.com/ec-jrc/lisflood-code |
Medium-range forecast | Forecasts produced for a lead time of several days (typically 3-15 days). The HRES and the ENS constitute medium-range forecasts. |
Numerical weather prediction model (NWP) | Mathematical models of the atmosphere, oceans, and land processes that predict the weather based on current weather conditions. |
Perturbed forecast | Ensemble forecast subset excluding its control member. Generally used in combination with the control forecast. The order of the numbering of the perturbed forecasts is random. |
Reanalysis | A reanalysis is the same as analysis but produced using a frozen model version giving the benefit that the model has not changed over the time period of the reanalysis (note that the observations assimilated will still change over time depending on what was available). |
Reforecast | Forecast simulations performed for a set of past dates, based on a configuration as close as possible to the operational setting. They are used to evaluate the forecast skills against a benchmark or reference forecast. |
Return period threshold | Also known as a recurrence interval, is an estimated average time between events such as river discharge flows to occur. |
Seasonal-range forecast | Forecasts produced for a lead time of several months (typically 30 days to 7 months) |
Time step | Time increments for each forecast. A daily time step means that the forecast has a value every 24 hours until the maximum forecast horizon. |
Upstream area | The total area that contributes with water to a specific point on the river network. Also known as a catchment area. |