We’re currently working on a new local forecast that will be introduced to Ferny Grove Weather.
This will be a new addition to the website, built using multiple sources of forecast data to better reflect the local suburban conditions experienced at Ferny Grove.
Weather forecasting
Most weather forecasts are based on weather models — large computer simulations run by weather agencies around the world that calculate how the atmosphere is likely to evolve over time.
Each model has its own strengths and limitations. Some handle certain weather patterns better than others, and none of them are perfect, especially at a local, suburban site where weather can be experienced in quite unique ways.
A widely used and proven way to improve forecasts is to take a consensus approach. This means looking at several different models rather than relying on just one, and combining their information to reduce individual weaknesses. This approach is commonly used in operational forecasting because it tends to produce more reliable and more consistent results.
In addition to combining models, forecasts can also be improved by applying bias corrections. Over time, models often show repeatable tendencies — for example, running slightly too warm or too cool in specific situations as a simple example. By comparing recent forecasts against actual observations, these tendencies can be identified and adjusted for.
The new forecast we’re developing brings these concepts together. It combines information from many different models, applies various adjustments based on how they’ve been performing locally, and blends the results into a single forecast.
The aim is to provide a forecast that reflects the most likely outcome for Ferny Grove, using methods that are widely used in forecasting to improve accuracy and reliability.
How is this different from official forecasts?
Official forecasts produced by national weather agencies are designed to be representative over large areas and are based on observations from standardised weather stations.
This new local forecast isn’t intended to replace those forecasts.
Instead, it focuses on one specific location — Ferny Grove — and looks closely at how different forecast sources tend to behave here over time. By comparing forecasts against local observations and adjusting for consistent local differences, it aims to provide an additional, site-specific perspective.
Official forecasts and this local forecast are best seen as complementary, rather than competing.
What will be included first?
The first version will focus on temperature, as this is a core part of weather forecasts and a good place to introduce the new approach.
As development continues, this temperature forecast may also be blended with locally generated WXSIM output, which has been in use here for over a year. WXSIM is a locally run weather simulation that can help capture terrain and other local influences, though whether it appears in the initial release will depend on further testing and results.
Other forecast elements, such as rainfall guidance and probabilities, are planned for later phases once they can be added in a way that is clear, useful, and reliable.
What’s happening now?
Work is continuing behind the scenes to develop and refine the new forecast, with results being compared against actual observations as part of that process.
Early verification suggests excellent results, with mean absolute error for daily temperature typically around 1 °C or better at this site, though performance naturally varies with weather conditions.
Accuracy can be reduced on days with changing cloud, thunderstorms, or other localised weather effects. Improving how these situations are handled is an ongoing part of development, and the forecast will continue to be refined over time.
Looking ahead
Updates will be shared as this forecast becomes available and as new elements are added over time.
It’s hoped this helps explain what is quite a complex topic and marks the start of a new addition to Ferny Grove Weather. Please feel free to comment or share this post. You’re also welcome to get in touch if you have any questions related to these forecasts or other information on the site.