The Excessive Heat product is designed to give emergency managers, forecasters and planners from 3 days to a week advance notice of regions where the combined effects of temperature and humidity are likely to create conditions ranging from uncomfortable to unhealthy and potentially dangerous.

The contours are the probability of an excessive heat event, which, at any location, is defined as:

At least 3 days in 5, or 7 with values of daily mean heat index (HI) of at least 85⁰ F .

Notice that the forecast probabilities on the maps are categorized and color coded into LOW, MEDIUM, and HIGH liklihood classes.

Also, outlooks are given for the 3-7, 6-10 and 8-14 day time ranges.

Those who want more information should consult the boxes to the left of the maps.

Note: Strictly speaking, the 8-14 day probabilities are not directly comparable to those for the other two forecast periods. This is because the criterion of 850F for 3 days in 7 is more easily reached than that for 3 days in 5. Therefore, probabilities on the 8-14 day maps are generally larger than those for 3-7 and 6-10 days.

EXCESSIVE HEAT PRODUCTION METHOD

RATIONALE. We wanted to predict the heat index for the 3-7, 6-10 and 8-14 day time ranges. To do this directly requires forecasts of temperatures and humidity near the ground at least 4 times per day out to 14 days. Model forecasts of near-surface temperature and humidity are either not yet available, or are not yet ready for use in operational forecasting. On the other hand, sets of 20 or more dynamical model forecasts of upper air height out to at least 14 days are available and are known to have useable skill for averages over 5 or more days at leads of 5 to 7 days.

Given that background, we decided to use an indirect method making use of the available forecasts of upper air height. This technique results in what is called a perfect prog model in which a forecast model is developed using observed data for the predictors and predictand. We used multiple regression with observed daily average 500 hPa height and 850 hPa temperature as predictors against observed heat index as the predictand.

Two sets of equations were developed. One set uses the daily maximum heat index, HImax, in the 5 or 7 day outlook period as the predictand. We chose this variable because it is easy to understand and is often quoted by the media.

The daily mean heat index, HImean, is more closely related to heat-induced illness than HImax, according to research by Dr. Larry Kalkstein, of the University of Delaware. In fact, for cities in the Midwest and the Northeast, heat-related illness and death increase sharply when values of HImean exceed 85⁰F for several consecutive days at a given location. As HImean increases, through 90⁰F, and 95⁰F, fewer consecutive days are required to produce ill-effects. This is the reason we chose three the thresholds listed above in the section entitled: A Brief explanation of the Excessive heat product.

PRODUCTION METHOD. Each day, a total of 23 daily dynamical model forecasts of 500 hPa height and 850 hPa temperature out to 14 days, all verifying at the same time, are available by about 8:00 AM ET. Each of the 23 dynamical model forecasts of the predictors is used, along with the regression equations, to produce forecasts of HImax and Himean. All 23 forecasts are then averaged to produce the final outlook for each variable.

In general, the atmosphere has more variability than the dynamical model. This causes the forecasts to have lower amplitude than they should. In order to correct this, the forecasts are inflated slightly, so that the variability of the predictors is accurately reflected in the heat index forecast.

DATA The data sets used to develop this product consist of:

• -hourly temperature and relative humidity at 202 stations from 1961-90
• -500 hPa height and 850 hPa temperature analyses, 4 times daily, 1961-90

DEFINITIONS OF TERMS

The heat index (HI) is a number that expresses, in degrees F, how it feels as temperature (T) and relative humidity (RH) vary. When T is high, but RH is low (HI lower than T), the body is generally able to cool itself efficiently through the evaporation of perspiration on the skin. At high T and high RH (HI greater than T), the efficiency of this natural cooling process declines, making us feel uncomfortable. If the body is exposed to hot humid conditions for long periods of time, our body finds it increasingly difficult to maintain a healthy core temperature. This is especially difficult for the very young, the very old, and others. These groups are at greater risk of heat-related discomfort and illness than much of the population.

Climatology: for example, for any station on the map and for the June 1-5 period, the climatology of daily mean heat index is found by 1) counting the number of times during the climatology period (1961-90) the observed daily mean heat index during each June 1-5 equaled or exceeded 85⁰ F at least 3 days out of the 5, 2) dividing that number by the total number of years (30) and, 3) multiplying that result by 100. If this event happened 10 times during 1961-90, the climatological probability would be 33 1/3 percent. If it only happened once, the climatological probability would be 3 1/3 percent. The same method is used to find the climatology for the other two thresholds.

probability of and event is the number of times an event actually occurs divided by the total possible number of time the event could have occurred, over the long term. This number is a fraction, less than one. Probabilities are often multiplied by 100 and followed by the symbol "%", the heat index maps are given in this form.