| Forecasting the weather is a lot harder than it | | | | Surprised? Well, don't worry, you're not alone. I'm |
| appears when you see a weather man on | | | | sure you've seen on your local television during a |
| television. There are several key features that a | | | | storm situation about model A, B, and C. Well |
| meteorologist must look at in order to make a | | | | what is model A, B, and C? Actually there are lot |
| great forecast, and those that don't follow these | | | | more than three models, but in most case the |
| rules are easy to pick out. So let's break down | | | | TV weatherman is talking about the GFS, ETA |
| the important aspects of a forecast. | | | | WRF, and possibly the old NGM model. However, |
| The most important part of a forecast should be | | | | there are so many more models out there that |
| the first thing a meteorologist should look at, | | | | you can look at right on the internet. I'll name a |
| current observations. Current observations cover | | | | few. In the short term, there is the MM5 and RUC |
| everything from current and previous conditions, | | | | which go out to roughly 12 to 48 hours. These |
| radar data, and satellite images. You can't know | | | | models are great for severe thunderstorms, |
| what the future is going to be without an | | | | lake-effect snow, and other small scale weather |
| understanding of the past, and forecasting the | | | | impacts. In the medium and long term, there are |
| weather is no different. Having a strong | | | | several more models. Of course there is the WRF |
| understanding of current weather conditions is | | | | NAM/ETA model group, the GFS, the SREF, the |
| extremely important. The direction of the current | | | | ECMWF, the Canadian model groups, the UKMET, |
| wind may be a clue as to whether a location | | | | the Korean models, and then ensemble models. |
| receives rain or snow. The pressure trends can | | | | Ensemble models are basically several runs of a |
| give a meteorologist clues on where a deepening | | | | model with various small changes in the input of |
| low pressure is going. Satellite trends shows us at | | | | the model. With all these choices, forecasting the |
| what stage in development a storm is in or | | | | weather should be easy, right? |
| where dry air is impacting a tropical low. The | | | | Well, not exactly. The problem with models is that |
| water vapor satellite image in particular can show | | | | they were created by man, and as such have a |
| a meteorologist a small, yet potent disturbance | | | | lot of error. Some have errors in handling latent |
| that could produce a severe weather outbreak or | | | | heat, which is released during thunderstorm |
| a burst of heavy snow. Meanwhile, radar data | | | | events. Other models have errors in handling the |
| gives us warnings of tornado development, | | | | position of upper lows for certain seasons or tend |
| intensifying precipitation, and the trends of that | | | | to phase jet streams and produce unrealistic |
| precipitation which can produce flooding. | | | | forecasts, but how would you know this? Well, |
| Understanding the current environment is the | | | | first you have to have a strong handle on the |
| difference between a bad forecast and a good | | | | physics of the atmosphere and what should |
| one. | | | | happen based on those laws of physics. Also, |
| The next important part of a forecast is | | | | simply learning from your past mistakes is also |
| climatology. While current observations are | | | | extremely important in using these models. As |
| extremely important, understanding the normal | | | | you continue to forecast, you pick up on model |
| climatologically conditions is extremely important | | | | trends and errors. You learn what to look for and |
| as well. I put indices like the North Atlantic | | | | the clues they show. However, you must |
| Oscillation (NAO), Arctic Oscillation (AO), the ENSO | | | | remember that the model is not predicting the |
| states (El Nino and La Nina), and also the use of | | | | weather, it is only showing a possible outcome. It |
| comparing previous patterns to the current | | | | is you, the forecast that must determine if that |
| pattern. All this falls under climatology. Once again, | | | | outcome is realistic or not. |
| understanding the past and learning lessons from | | | | Now you are ready to forecast. We've collected |
| the past is extremely important to a forecast. | | | | and analysis the current weather data so we |
| What does it mean when the NAO is negative? | | | | know what is going on now and what has been |
| How does this impact a storm forecast 3 to 5 | | | | going on recently. We've developed a strong |
| days away? What type of pattern is expected in | | | | understanding of the current climatological |
| an El Nino or La Nina winter, and which part of the | | | | conditions so we understand what is normal and |
| country will be impacted the most from these | | | | what is abnormal. And finally, we've looked at all |
| ENSO states? Answering these questions every | | | | the model guidance and look over all the possible |
| day is crucial to a forecast and must be | | | | outcomes. Now, take the knowledge you gained |
| understood to avoid error. | | | | and make your forecast! Good luck! |
| Finally, the last part of a forecast is the models. | | | | |