Our ESDV3 Seasonal Forecasting service now includes bespoke long-range hydro-meteorological forecast data for 51 ECMWF ensemble members. This allows users to test-case planning strategies against a selection of possible forthcoming weather scenarios and better manage risk using cutting-edge NWP data.
We’ve blogged previously about how forecasts of climate and weather are sensitive to small differences in the starting conditions of forecasts: the equations that predict future weather conditions need numbers to start with – i.e. observations of the current weather – and it is impossible to *perfectly* observe the current weather across the globe to provide these numbers, with no error. Due to the characteristics of the maths equations in leading weather forecasting systems (e.g. UMKO GLO-SEA-6, Meteo-France Sys 7, ECMWF SEAS-V, DWD GCFS2.0, NCEP-NOAA CFSystemv2), small differences in starting numbers can grow into different forecasts — the so-called chaos effect.
To combat this, forecasting centres can conduct multiple runs of forecasts each with slightly different starting numbers. Users then have the choice of how to use these forecasts:  take the average of all the individual runs and use that information – or  analyse all of the individual runs, and establish the worst and the best possible outcomes. (or option : do both  and !).
ESD Seasonal Forecaster enhancements: ESDV3
When we launched our Seasonal Forecasting service, we provided clients with information derived from five numerical weather prediction models in the form of , the average of all the individual runs for each centre. This information is a useful guide, but users have no information of probabilities of certain weather conditions being met. Is the average forecast the way it is because all the multiple runs came out with the same answer? Or because a small number had a result strongly in one direction and and equal number of runs had a result in the opposing direction? (e.g. cold temperatures and warm temperatures)
We started work several weeks ago to enhance the ESD Seasonal Forecaster to enable users to answer these types of questions and are pleased to say the ESDV3 Seasonal Forecaster is now available for our client’s use! In the same way as the initial Forecaster, we configure the code base with our user’s specific geographic and weather-dependent metrics, and run the code each month, but now provide a full ‘ensemble’ set of output (51 runs) for one of our primary component modelling systems ‘ECMWF’.
Identifying best & worst cases for planning purposes
We confer with our users to establish their unique weather criteria (in terms of exposure for their business / resources / operations) and then isolate the ‘best’ and ‘worst’ forecasts from the full set each month – allowing easy test-cases for their management strategies for the coming month, under best and worst potential cases.
Like the original ESD Seasonal Forecaster, ESDV3 has full global coverage, and a maximum lead time of seven months. They will be configured to output the most relevant weather-related intelligence needed by your organisation to make the most resilient planning decisions to protect your community, resources and customers.
The ESD Seasonal Forecaster and ESDV3 are monthly services utilising raw data from five of the world’s leading national seasonal forecasting models and transforming these data into user-relevant packages for onward, inhouse planning. Apply to join now as a user, or learn more via the buttons below.