Its a question often asked, why is the weatherman always wrong ? Or people will say, “I don't listen to the weather their always wrong anyway.” As we will see this is not exactly correct. Such an attitude can also deprive a person of much needed information at a critical time, and as a result put them in much more danger than need be.
Much of the answer comes down to understanding how weather forecasting works. Modern forecasting makes use of various mathematical models. These models contain as many as hundreds of variables. They are in fact large multi-variable equations. The problem as any first year algebra student can tell you is you can't have a multi-variable equation. So forecasting in a way is based on a fallacy. have to do in this situation is assume a value for every variable except one. Of course modern meteorologists are very good at estimating all those missing values. These values can include temperature, wind speed, wind direction, elevation and many other things. It is for just this reason that nearly every forecaster, will use a half dozen or more models In forming their forecast. By combining input from several forecast models which use different kinds of data, and weigh that data differently you reduce the rate of error when your assumed values inevitably change. The tricky nature of working with forecasting models was made clear in a course a took called, “Challenges of Forecasting in the West.” It was based on a workshop the National Weather Service held for its own forecasters. Using several examples from the pacific northwest, the class showed how even if all the data in a model is correct the model may be wrong in its result. This is because of unique local conditions that differ from the base assumptions built into that specific model. This is why the knowledge and experience of the forecaster is so important, and “adds value” to the forecast. This is particularly true in terms of the forecasters knowledge of his or her local area.
So why use mathematical forecast models at all ? After all, the atmosphere is a fluid. The variables are never going to remain the same. They in fact will change constantly. This of course will reduce the accuracy of our model, and our forecast. Because despite their faults in many cases and in many ways they are the best tools we have. A weather forecasting model is the same as any other tool in this respect. Any tool, no matter what type has its strengths and weaknesses. Mathematical forecast models are no different, and it helps to understand them if you understand this basic reality.
There is also the concept of confidence level. Many times especially if you are dealing with the forecast discussion produced by the National Weather Service's local Weather Forecasting Office it will say something like “Confidence is low in this forecast.” Such a comment can refer to the forecast as a whole, or a single element of it. What this means is even though the best available information says the weather will play out in a certain way, the forecaster for a variety of reasons may think otherwise. Information on the forecaster's “confidence” in a forecast is most often not relayed to the general public.
A third factor relating to how accurately we can predict the weather relates to public safety. Weather information is rarely utilized in real time. In case of bad weather people need time to prepare, make alterations to their routines, or even evacuate. The more lead time a forecaster wants to give people, the more opportunity there is for the variables on which the forecast is based to change. For this reason many of not most forecasters build a safety margin of sorts into their predictions. In terms of public safety it is better to risk the public being over prepared, than under warned. Of course the danger here is if you over predict enough, often enough you will be seen as “crying wolf.” If this happens your creditibility, and usefulness to the public will be greatly damaged. So as you can imagine this part of the process can be quite a balancing act !
So we can see when we evaluate how wrong we think the weatherman is, it helps to understand the strengths and weaknesses of the process involved. No human process of course is perfect. Weather forecasting is no exception. Also as we have seen, when dealing with public safety being “wrong” tp a point may be the best thing to do.
Robert A. Crutchfield is a veteran weather spotter working with NWS/Skywarn. The Severe Weather Spotter Team of the Fort Bend County (Texas) Office of Emergency Management, and Storm Pins a cell phone app affiliated with Houston's NBC TV affiliate. Other emergency management/disaster recovery functions he has performed in the past include shelter volunteer, mass care logistics, crowd control, search and rescue, and fire department chaplain. Follow him on Twitter at www.twitter.com/revcrutchfield.
Comments