IRI Home
Tutorial 1

IRI Climate Prediction Tutorial 1

What Is the Problem?

  • Year to year differences in the timing, intensity and duration of the rainy/dry and warm/cold seasons can have HUGE impacts on people: on our prosperity, on our health, and on our environment.

  • The winter of 1982/1983 was exceptionally hard on the planet. The dry seasons in the Southwest United States, in Peru and Chile were very rainy and the rainy season in Indonesia was extremely dry. You can see some of the effects of abnormal or "anomalous" temperatures and rainfall in 1982/83

  • One of IRI's jobs is to predict  when a rainy season might fail or when flooding or temperature extremes might be likely.

Laying the Foundation for a Solution

  • If we can't predict the weather next week, you might ask yourself, why do we think we can make predictions for next season?
    Well, we can't predict the weather for next season, but under some conditions we can to say something useful about the climate for next season.

  • So then what is climate? And what makes it different from weather?
    Check out our Climate Dictionary for definitions. Climate is what you expect, weather is what you get. The below plots help to illustrate. The panel on the left shows the precipitation rate for a single day (January 15, 1999) and the panel on the right shows the precipitation for a season (December-January-February) averaged over  many years. If we look at Southeast Asia for example, we would expect (based on our long term seasonal average) it to be relatively dry this time of year but on January 15 what we got was rain.


NOAA NCEP-NCAR CDAS-1 DAILY Diagnostic surface
[ prate * 2.67800E+06 ] 15 Jan 1999


NOAA NCEP CPC CAMS_OPI climatology precipitation

  • If you want to predict the weather tomorrow (without watching the weather channel) what do you need to know?
    The most important thing to know for tomorrow's forecast is today's weather (for example, temperature, pressure, winds), in as much detail as possible.

  • If you want to predict the climate next season (without looking at IRI forecasts) what do you think you'd need to know?
    The things we need to know for next season's forecast are things that affect the atmosphere and change  slowly, like the temperature of the  ocean. Ocean temperatures change relatively slowly and can be very different from one year to the next, especially when there is an El Niño. You can see ocean or "sea surface" temperature changes in our Maproom.

  • How is El Niño important?
    El Niño involves changes in sea surface temperatures over large areas of the tropical Pacific. It's associated with year-to-year variations in the character of the seasons.  You can get a simple explanation of the infamous El Niño in our Climate Dictionary. There are other things that are related to year-to-year variations in the seasons, too. The North Atlantic Oscillation is a pattern in surface pressure that can shift rainfall and temperature patterns over eastern North America and Europe. You can also read more about how El Niño and its sister La Nina affect seasonal patterns all over the globe and how the North Atlantic Oscillation affects surface temperatures.

What Is the IRI Doing?

  • With a knowledge of current sea surface temperatures in the tropical ocean basins, and model predictions of how the oceans will likely evolve during the next several months, we use numerical models of the atmosphere to predict mean seasonal patterns of temperature and rainfall. But because the atmosphere is a chaotic system, small changes in the current conditions can result in big differences in the seasonal patterns the model predicts. So, a single prediction does not tell the whole story.  Each prediction is just one possible outcome. We deal with this problem by doing a number of predictions. The combination of several of these predictions is our best estimate of the next season's mean climate.


S-O-N Precipitation Anomalies


Masked Anomalies

  • This average prediction for rainfall anomalies is shown in the pictures above for September-October-November 1999. The left panel shows what the models forecast. The right panel shows the same forecast but we've masked out the regions where the model has not historically done a good job (note: the model may do better in a particular region in one season than another--this is an example of one model in one season).

  • By combining predictions from several different computer models, the IRI Net Assessment Forecast  provides more information than the average of a single model's  predictions.  These Net Assessment maps tell us the probability of above normal/normal/below normal precipitation or temperature - this gives us an idea of the differences between the individual predictions.

  • The IRI Net Assessment forecasts are only the beginning of what IRI is doing. To maximize the utility of seasonal forecasts, we are also developing a variety of monitoring and applications tools  for the global community so that we may be better prepared for the variations that climate dishes out.

The IRI Climate Information System

  • Our Climate Information System is being developed and has a wealth of information and insights on topics ranging from climate monitoring to impacts of climate on health. The Climate Information Digest offers a summary of recent events and forecasts. Check it out on our site monthly.  Explore our web site and please give us some feed back. Thanks for visiting.

To top   Back