|
|
 Home Research Seasonal
forecasting User
guide GloSea
and EURO-SIP |
 |
| Weather and climate | Aviation | Leisure | Research | Education | Products and services | |
 |
|
Each forecast starts from the observed state of the ocean, land and atmosphere on the first day of the month of issue - the 'initial conditions'. Unfortunately, the initial conditions are not known precisely (one reason is that observing stations/instruments are sparsely distributed in some regions), and individual forecasts can be very sensitive to small uncertainties in the initial conditions. The sensitivity to initial conditions is taken into account by making a number of predictions each with slight variations to the starting conditions which reflect our uncertainty in the initial state. This technique generates an 'ensemble' of forecasts with each individual forecast referred to as an ensemble 'member'. The forecast products are created by analysing the output from all the ensemble members. In addition to forecast sensitivity to starting conditions, forecasts can also be sensitive to the way we represent the real climate system in the computer models - so called sensitivities to model formulation. We can take account of sensitivities to model formulation by combining ensembles from models that have been developed at different institutes using different formulation strategies. It has been shown in the recent DEMETER project that combining ensembles in this way to construct a multi-model ensemble improves the skill of seasonal forecasts (relative to the skill obtained using any one single model from the multi-model set).
With default settings of the menu-driven map selection, the forecasts displayed are from the Met Office global seasonal prediction system, referred to as 'GloSea'. An extensive set of performance validation information is provided for the GloSea system from analysis of retrospective forecasts over the period 1987-2002. The products are based on the output from forecasts made using a coupled ocean-atmosphere General Circulation Model (GCM). The GCM, known as the GloSea (Global Seasonal) model, is similar to the HadCM3 climate version of the Met Office Unified Model (UM), with a number of enhancements for seasonal forecasting purposes. Details of the model physics and discussion of the performance of HadCM3 can be found in Gordon et al, Climate Dynamics (2000). The atmospheric component is version HadAM3 (see Pope et al, Climate Dynamics (2000) for a description), with a horizontal resolution of 3.75° east-west and 2.5° north-south, and 19 vertical levels. The oceanic component has 40 vertical levels (compared to 20 in HadCM3), zonal grid spacing at 1.25°, and meridional grid spacing of 0.3° near the equator increasing to 1.25° poleward of the mid-latitudes (compared to 1.25° resolution east-west and north-south in HadCM3). A coastal tiling scheme has been included, to enable specifications of the land-sea mask at the ocean resolution. Like HadCM3, the GloSea coupled GCM contains no flux corrections or relaxations to climatology. Each forecast requires initial ocean, land and atmosphere conditions. The land and atmosphere conditions are specified from atmospheric analyses that are produced seperately for weather prediction purposes. The ocean initial conditions are taken from ocean analyses generated specifically for seasonal forecasting, using the ocean GCM component of GloSea. The ocean GCM is run using surface fluxes of momentum, heat and water prescribed from atmospheric analyses, while assimilating sub-surface ocean obervational data, with temperatures in the top layer(s) constrained to be close to surface observations. Each month forecasts are run with starting conditions at the beginning of the month, to create a 41-member ensemble. The different ensemble members are created by perturbing the ocean state, through windstress perturbations throughout the ocean analysis stage and initial sea surface temperature perturbations.
In addition to seasonal forecasts of temperature and precipitation, the Met Office’s dynamical seasonal prediction model (GloSea) is used to forecast the number of tropical storms forming over each of the tropical ocean basins out to 6 months ahead. Latest forecast for the North Atlantic basin The predictions are based on GloSea representation of dynamical and physical processes characteristic of tropical storms, and are achieved by counting the frequency of tropical storms in the model forecasts. As the dynamical model grid does not fully resolve tropical storms, numbers are calibrated using behaviour in retrospective forecasts for past years. Recent studies have shown that GloSea and other European models have substantial skill in predicting the number of tropical storms during the North Atlantic season, nominally June to November (see eg. Vitart et al 2007). Importantly, the models distinguished between the exceptionally active season of 2005 and the below-normal activity of the 2006 season. This marked interannual change was missed by a number of statistical prediction methods, which have to date formed the basis of most issued predictions. Dynamically-Based Seasonal forecasts of Atlantic Tropical -Storm
Activity issued in June by EUROSIP Vitart, F., Seasonal forecasting of tropical storm frequency using a multi-model ensemble, Quart. J. Roy. Meteor. Soc., 132, 647-666 (2006). Vitart, F. & Stockdale, T.N., Seasonal forecasting of tropical storms using coupled GCM integrations, Mon. Wea. Rev., 129, 2521-2537 (2001). |
|||||||||
|
 |
