Building Better Forecasts

This time next year, when the local weather forecast is for a 30 percent chance of rain, the odds for a correct forecast will be greater than they are now. Meteorologists around the world will soon have access to more and better data about Earth's atmosphere -- information they need to improve their forecasts -- thanks to a new NASA instrument planned for launch this spring.

The Atmospheric Infrared Sounder, flying onboard the Aqua satellite, will measure radiation as it leaves Earth's atmosphere. Scientists will be able to translate these measurements into precise information about atmospheric temperature, water vapor, clouds, ozone and other gases.

JPL meteorologist Dr. Eric Fetzer, who is working with the science team says the instrument was designed for "basic research -- monitoring the climate and understanding the processes that occur within the climate system -- which means weather. At the same time, it provides detailed information that can be used now by weather forecasters."

He explains some of the challenges involved in weather forecasting and how the instrument will fit into this complicated process.

Q: Why aren't weather forecasts better than they are?

A: Forecasting is necessarily a combination of observations and computer modeling of the atmosphere. One major challenge is getting sufficiently good observations over a lot of places.

Right now, the observations we have primarily come from a combination of weather balloons and some satellite information. Data is very sparse in some places, particularly over the oceans.

Also, the most up-to-date weather service global computer models have a grid point every .7 degrees - every 80 kilometers or about 50 miles. That means that they try to numerically describe the physics of the atmosphere at that particular spot. For example, there could be a grid point here in Pasadena, another grid point at San Bernardino and another 30 miles off shore. That's not a lot of detail considering how much different the weather is 30 miles off shore or the climate is in San Bernardino.

Q: How is the Atmospheric Infrared Sounder going to help?

A: One key to improving forecasts is to get detailed information at higher spatial resolutions. The instrument is going to do that.

It will make a suite of atmospheric observations, but the big three as far as weather forecasting is concerned are temperature, water vapor (or humidity) and cloud state -- that's how cloudy it is, how much water is in the clouds, and how high their tops are.

The instrument is going to provide these detailed observations over most of Earth at intervals of about 50 kilometers or 31 miles -- that's about three times better spatial resolution than current global models.

These observations clearly become especially important over the oceans where we don't have weather balloons or people making weather measurements.

Airs will provide several observations at each point -- it will be actually over-observing what the models use. It will reveal a lot of subtleties that will help improve forecasts.

Q: Will better observation lead to perfect computer models of the atmosphere?

A: A computer model is always an imperfect picture of the way the atmosphere works. You don't have the computing resources to do it in great detail, and you simply don't understand the physical processes involved. Something like clouds, for example, are very difficult to put into a computer model. The detailed processes that go on, for example, how cloud droplets turn into rain, are not fully understood and yet are crucial to all kinds of weather modeling. Many processes in the real atmosphere simply can't be correctly modeled, so you have to observe them.

Q: How soon will the Atmospheric Infrared Sounder make a difference in my local weather forecast?

A: We will be providing data to weather forecasters around the world within a year. We're working with several institutions, including the NASA Data Assimilation Office, the National Oceanic and Atmospheric Administration, the European Center for Medium-Range Weather Forecasts, the Australia Bureau of Meteorology, the British Meteorological Office and the Meteorological Service of Canada. They'll get our data within three hours. It's going right through the pipeline out to the forecasting community.

Q: What kind of impact will the Atmospheric Infrared Sounder have on weather forecasting?

A: It will be useful in short-term forecasting for a day or two and should also have a significant impact on three- to five-day forecasts.

People plan their lives around the weather -- it's not just a question of whether I'm going to bring my umbrella to work, but it affects things like airline scheduling, transportation, agriculture and construction. For a lot of businesses, a more reliable two-day forecast means a lot.

The instrument may well reduce the number of disaster forecasts -- those are the ones that really go off the mark. One famous example was the October 1987 storm in Britain. It was the biggest storm to hit the country in more than 200 years, yet forecasters didn't see it coming. Those are the ones that matter. If the forecast is for a 30 percent chance of rain today and our instrument can refine that down to 32 percent, well, so what? But if it is likely to rain a week from now and you can say whether that is "highly likely" or "highly unlikely," that's a pretty significant forecast.

Also, in the long run, the scientific insights the instrument will provide into basic atmospheric processes such as convection--a major component in such events as hurricanes, tornados and El Nino --will eventually lead to better weather models.

Q: Will we ever get to the point where we have perfect weather forecasts?

A: No. It was recognized as long as 40 years ago when the first chaos computer models were analyzed by Edward Lorenz at the Massachusetts Institute of Technology.

The general rule is that you can't make a good weather forecast beyond 7 to 10 days, but you can make climate forecasts.