Corporate Targets

The following Key Performance Targets (KPTs) have been set for the Chief Executive of the Met Office and announced in Parliament for the financial year (FY) 2008/9. The targets are designed to drive further improvements to the Agency's performance and are as follows.

Forecast Accuracy

To achieve at least three out of four of the following forecast accuracy measures. However, any failed element will be required to meet the same level as the FY 2007/8 outturn for the overall KPT to be met.

1. Improve the forecasting skill using the combined Numerical Weather Prediction Index to at least 125.8
2. More skilfully predict whether precipitation will occur at selected locations to achieve a skill score of at least 0.438
3. Predict maximum temperature at selected locations to within 2 degrees accuracy 86.2% of the time
4. Predict minimum temperature at selected locations to within 2 degrees accuracy 84.4% of the time.

Business Profitability

To achieve a business profitability target of £7.0 million.

Return on Capital Employed

In line with Treasury requirements to achieve a Return on Capital Employed of at least 3.5%.

Support to wider government controls

To deliver the outputs of the Customer Supplier Agreements (CSA) for Public Weather Services, Defence and Defra within the tolerances agreed with the customers and defined in the CSAs.

Weather forecast verification – how accurate are Met Office forecasts?

The Met Office has an open and transparent policy in the verification of its forecasts.

Verification is needed to measure how we are doing against the Key Performance Targets (KPT) for forecast accuracy, as set by government. These KPTs are part of our ongoing long-term commitment to improving the accuracy of our forecasts to the public. We will do this by reaching accuracy targets set for:

* next-day rainfall forecasts for 11 places within the UK (the nearest major towns to these places are shown on the map opposite);
* next-day temperature forecasts (max and min) for 11 places within the UK (the nearest major towns to these places are shown on the map opposite);
* the forecasting skill of our computer models.

Computer model forecast accuracy

All forecasts, of whatever type, are ultimately based on the predictions from our suite of sophisticated atmospheric and oceanic models, run on our powerful supercomputer.

This form of forecasting is known as Numerical Weather Prediction (NWP). The accuracy of the Met Office computer model forecasts is measured by the NWP Index.

The NWP Index takes into account computer forecasts for several weather elements, including mean sea level pressure, for regions covering the entire globe from one to five days ahead. The higher the NWP Index, the more accurate the computer forecasts are.

Rainfall

Accurate forecasts of rainfall are essential for partner organisations such as the Environment Agency for flood warnings and the Highways Agency for road safety

Unlike temperature, rainfall may or may not happen. Rain can be highly variable, especially when falling as showers; one side of town may get rain while the other remains dry. So for the purpose of verification, forecasts of rain are expressed in terms of the “chance” (probability) of a place being wet or dry.

The Met Office uses a verification method that:

• gives a score (called the Brier Skill Score) that takes account of the probability applied to the forecast;
• takes account of how difficult or easy the forecast was by measuring it against a climatological reference.

A positive score indicates that the forecasts are, on average, better than climatology. A completely accurate forecast would have a score of 1.

High scores are achieved when:

• a high probability of rain is forecast and rain subsequently occurs
• a low probability of rain is forecast and it subsequently remains dry

Conversely low scores are achieved when:

• a high probability of rain is forecast and it subsequently remains dry
• a low probability of rain is forecast and rain subsequently occurs

Minimum temperature

Accurate forecasts of temperature are vital for utility companies, for road authorities and for health forecasting

This graph shows how accurate our forecasts of minimum temperatures are for the next day ahead. It shows the percentage of forecasts that are within 2 °C of the actual temperature recorded. The accuracy of minimum temperature forecasts is lower than that for maximum temperature. This is because minimum temperature is more sensitive to variations in cloud cover, wind speed and wind direction and so is harder to predict.

Maximum temperature

Accurate forecasts of temperature are vital for utility companies, for road authorities and for health forecasting

This graph shows how accurate our forecasts of maximum temperatures are for the next day ahead. It shows the percentage of forecasts that are within 2 °C of the actual temperature recorded.

How our forecasts have improved

Through continual investment in research, supercomputing and observations, Met Office scientists have steadily improved the accuracy of our forecasts. All of the forecasts we produce are stored and their accuracy assessed, so that we can learn from what went wrong with inaccurate forecasts and make sure that they keep getting better. Our targets for forecast accuracy are set each year by the government.

There are a number of ways of measuring the accuracy of a forecast. One method used at the Met Office is called the NWP index. This combines the accuracy of a number of different elements into a single measure of overall accuracy. An example of how our computer-generated sea-level pressure forecasts have improved over the years is shown below.

Forecasting further ahead

Our three-day forecasts are now as accurate as our one-day forecasts were 20 years ago. This improvement in forecast accuracy stems from investment in research, faster supercomputers and greater coverage by observations.

In addition to improving the accuracy of the one- to five-day forecast, research has enabled us to make forecasts that were previously impossible. We can now forecast further into the future allowing regular seasonal forecasts to be produced, and predictions of climate change are continually improving. New research is expected to provide further significant improvements in our ability to forecast heavy thunderstorms a few hours in advance.