Met Office vs BBC Weather – which weather app is best?

Turn to the Met Office for the best temperature forecasts, and to BBC Weather for predictions of rain, meteorologists at the University of Reading suggest. 

These are the preliminary conclusions of scientists at the University of Reading, who have taken the first steps in measuring the accuracy of two of the UK’s most popular weather apps.  

The most comprehensive open research of its kind initially shows the Met Office temperature forecasts are more reliable than BBC Weather, but the latter was slightly more useful at forecasting rain.  

Dr Rob Thompson, who led the research, said: “Whether you use BBC Weather or Met Office for checking the weather, you’re in good hands. The forecasts are reliable for several days, helping you to know what clothes to pack for a weekend trip or when to expect a frost in the garden. BBC Weather appears to be slightly better at predicting rain, although its variability means forecasting rain is much harder than temperature and needs more data to draw firm conclusions. As we would expect, we also find that the further ahead the forecast, the more uncertainty there is. 

“Modern weather forecasting has advanced significantly thanks to more atmospheric observations (particularly from satellites), improved computer models, and a deeper understanding of weather science. These tools allow meteorologists to predict large-scale weather patterns with remarkable accuracy, even a week or more ahead. However, pinpointing the exact timing and location of rain remains a challenge, explaining the remaining uncertainty in rainfall forecasts.” 

Looking ahead

Dr Rob Thompson and meteorology PhD student Rosie Mammatt compared live weather forecast data from BBC Weather and the Met Office for the location of Reading, against actual weather data recorded at the University of Reading Atmospheric Observatory. They have initially compared temperature and rainfall only, having begun gathering data in October 2024.  

Forecast data is collected hourly and compared to observations every three hours, around the clock. The study has so far analysed 85,000 forecasts from both apps across 704 observation times during autumn and winter 2024.  

As of Monday, 13 January, the Met Office had predicted temperatures more accurately on 5.4% more occasions than BBC Weather. BBC Weather   gave the more useful forecast on 6.1% of occasions. 

The data and full methodology is available from Reading’s Department of Meteorology website – with rolling assessments of the success of the forecasts provided, along with the cumulative results, updated every three hours. 

Comparisons were limited to a six-day forecast as this is the approximate limit for useful forecasts. Key early findings include:   

• Temperature forecasts are highly accurate: Both apps predict temperatures correctly to within 2°C up to 3 days in advance and 3°C up to 5 days ahead, with biases smaller than 1°C. 

• Consistently similar predictions: Because both apps report temperatures to the nearest degree, forecasts from the two apps often match—more than 71% of the time for predictions for the next hour, and still 57% of the time even at 5 days out. 

• Met Office edges ahead in temperature accuracy: When forecasts differ, the Met Office predictions are more often closer to observed temperatures than BBC Weather, with a noticeable difference for longer-range forecasts (3-6 days ahead). 

• Rainfall predictions are slightly pessimistic: Both apps overestimate the chance of rain, predicting that rain will occur 17% of hours, when it has actually rained   just under 10% of the observation hours. BBC Weather is more pessimistic when predicting further in the future (as it becomes less willing to predict a low chance of rain), whereas the Met Office has a less dramatic change with time. This extra chance of rain, by both forecasts, could in part account for by very light rain that a person would notice – but a tipping bucket rain gauge might not detect.  

• BBC excels in confidence: BBC Weather forecasts for rain tend to display higher confidence of no rain, providing greater certainty for users that it will stay dry – and edging the Met Office on our usefulness measure in the process – more data is needed to clarify if the extra confidence is justified.  

The team intends to continue these comparisons with a view to submit an academic article for publication in a peer-reviewed journal. It is expected that as more data is collected and comparisons made, the results will become more meaningful over time.  

Results are recorded on the University of Reading Department of Meteorology website.