You get the perfect weather forecast at exactly your position. Or do you?
As we all know, weather models are not perfect. Every day, people complain about the forecast (and the meteorologists). Perhaps they get wet although their weather app was showing no rain, or the promised sun is covered by clouds.
It is well known to most people, that the uncertainty of a weather forecast increases with time. What perhaps fewer people realize, is that a point forecast (as presented in a weather app for example) is not representing a larger area but only that specific forecast point and this makes the forecast more uncertain than an area forecast. It is also a widespread assumption that a meteorologist is involved in the raw forecast data presented in a point forecast, which often is not the case.
Two main issues with a point forecast
1. Due to the limitation in computer power, weather models can only present weather data in a certain amount of grid points. This means that some local features, can be overseen by the model. An example is a narrow peninsula, where the model grid points are located on each side of the peninsula. As the grid points are located in open water, the model will oversee the peninsula and therefore probably overestimate the wave height on the leeside of the landmass.
2. Even though a certain weather situation overall is well forecast, the displayed weather conditions in a single forecast point can change dramatically from one model run to another or differ a lot from the observed weather conditions. This is due to the small area that the single point forecast is representing. Below are some examples of this.
A. Isolated showers:
Especially in situations with showers, the point forecast should be treated with caution.
The weather models can only predict areas where the showers are more likely but will rarely predict the exact position. The model is perhaps predicting heavy showers in the area near your location but is placing the showers just 5km away from the forecast point you are looking at. The forecast will then show no or just light precipitation, but the showers may as well pass directly over your location in real life.
B. A minor, but intense, low pressure that is approaching the forecast position:
The general track of the low may be well forecast, but there can still be a great uncertainty in the individual forecast point. If the forecast run has the low pressure to pass just south of the forecast point, the wind will probably be weak and with a low wave height. But if the low takes a slightly more northerly track, there will be strong westerly winds and rising waves.
C. Stationary fronts close to the forecast position, with a significant difference in weather conditions on either side of the front:
A front is the border between two types of air masses. In some cases, there could be a big difference between the weather on the cold side and the warm side. There could for example be much stronger winds on one side or a big difference in wind direction (influencing the wave height) or perhaps heavy thunderstorms. In either case, if the forecast point is very close to the front there will be a great uncertainty in the point forecast.
The first issue is a model problem, that will be less significant if the model has a high resolution. Local limited area models also usually have a more correct modeling of the terrain and bathymetry.
The second issue is more of an educational issue. Although the models are getting better, they are not perfect and will always have some degree of uncertainty.
It is therefore recommended to try to understand the synoptic weather situation by looking at weather charts, model charts and read the latest weather forecast for your area. You will then see what type of weather that is near your position, and by looking at ensemble forecasts (see our previous article about ensemble forecast) you can determine how reliable the latest forecast for your location is.
Next time you take a look at a point forecast – for and offshore wind farm, oil platform or just for your home, then remember to think about above statements.
Be sure that you know the overall weather pattern before you evaluate the point forecast – this will most likely benefit your use of a point forecast.