United States Climate Change Almanac - The Weather of the Future: Heat Waves, Extreme Storms, and Other Scenes from a Climate-Changed Planet - Heidi Cullen

The Weather of the Future: Heat Waves, Extreme Storms, and Other Scenes from a Climate-Changed Planet - Heidi Cullen (2010)

Appendix 1. United States Climate Change Almanac

Clearly, climate change will leave no part of the world untouched; but just as all the places I wrote about have their own indicators showing that climate change is taking place, so does your town or city. What follows here is an almanac that includes current and future temperature trends for a variety of cities around the United States. I’ll start with a look at the big picture in the United States.

1. Total Monthly Record High versus Low Temperatures for the United States

One way to get an idea of how climate change is making itself felt in terms of day-to-day weather in the United States is to look at the total number of daily record high and low temperatures that have been set around the country. A record daily high means that the high temperature recorded at a specific weather station was higher on a specific day than on that same day in previous years. A record daily low means that the lowest temperature on a specific day at a specific weather station was lower than on that same day in previous years.

In a recent study, scientists gathered and tallied records from across the United States to get a better picture of the long-term changes in record highs and record lows.1 They analyzed millions of daily high and low temperature readings taken over a period of six decades at about 1,800 weather stations across the country. The temperature measurements were collected by the National Climatic Data Center of the National Oceanic and Atmospheric Administration (NOAA) and underwent a strict quality control process that can pick out potential problems, such as missing data and inconsistent readings caused by changes in thermometers and station locations. What makes the tables below quite fascinating is that there is such a large discrepancy between the total number of record high temperatures and the total number of record low temperatures. Technically, if the Earth’s temperature was not increasing, you would expect that the number of record daily highs and lows being set each year should be about even. But that is far from the case. For the period from January 1, 2000, to December 20, 2009, the continental United States set 294,276 record highs and only 145,498 record lows. And if you look back over the past sixty years, that picture is reinforced. The ratio of record daily high to record daily low temperatures was almost one to one in the 1950s but has been rising steadily since the 1980s.

As you’ll notice in the tables below, there is still a lot of variability from year to year. In fact, October 2009 serves as an excellent example of a month that was quite cold: the number of record lows was 1,908 while the number of record highs was only 1,017. On average, October 2009 was the third-coolest October since record keeping began in 1895. The average October temperature of 50.8°F was 4.0°F below the twentieth-century average. But when you step back and look at the big picture, global land and ocean surface temperature for October 2009 was the sixth-warmest on record, with an anomaly of 1.03°F above the twentieth-century average of 57.1°F. The lesson here is that there will always be natural variability in space and time!

What’s also interesting is that the current two-to-one ratio of highs versus lows across the country comes from the fact that there is a relatively small number of record lows. In other words, much of the nation’s warming is taking place at night, and, as a result, temperatures dip down less often to set new record lows. This finding is completely consistent with climate models showing that higher overnight lows are to be expected as the planet warms.

In addition to looking at historical temperatures in recent decades, scientists also used a computer model to simulate how record high and low temperatures are likely to change in the future. The model indicates that if greenhouse gas emissions continue to increase in a “business-as-usual” scenario, the U.S. ratio of daily record high to record low temperatures would increase to about twenty to one by 2050 and fifty to one by 2100.

Number of Record High and Low Temperatures for the United States for 2000


Number of Record High and Low Temperatures for the United States for 2001


Number of Record High and Low Temperatures for the United States for 2002


Number of Record High and Low Temperatures for the United States for 2003


Number of Record High and Low Temperatures for the United States for 2004


Number of Record High and Low Temperatures for the United States for 2005


Number of Record High and Low Temperatures for the United States for 2006


Number of Record High and Low Temperatures for the United States for 2007


Number of Record High and Low Temperatures for the United States for 2008


Number of Record High and Low Temperatures for the United States for 2009


Total Number of Record High and Low Temperatures for the United States from 2000 to 2009


2. Increasing Number of Hot Days in U.S. Cities

The bar plots that follow illustrate how extremely hot days are likely to become more common and more intense as the overall climate warms during the next century. They also help demonstrate that a general warming of the climate has significant implications for the number and severity of extreme weather and climate events: in this case, heat waves.

For selected large cities or regions in the United States, the plots compare the average number of extremely hot days observed during the summer months in the twentieth century with projections for extremely hot days in the middle and end of the twenty-first century. Climate research indicates that heat waves are likely to be more stifling, and potentially more deadly, in coming decades as climate change progresses. For many locations, heat waves are projected to be more frequent, more intense, and longer lasting, and extreme heat events that are currently considered rare will become more common in coming years.2

Extreme weather and climate events can cause significant damage, and heat waves are considered public health emergencies because of their effects on human health. Hot temperatures contribute to increased emergency room visits and hospital admissions for cardiovascular disease and can cause heatstroke and other life-threatening conditions. The elderly are particularly vulnerable to extreme heat.

Heat waves such as the Chicago heat wave of 1995 and the European heat wave of 2003, which killed an estimated 50,000 people, have proved especially deadly to vulnerable populations, including the elderly and persons with respiratory illnesses.3

Scientists at Climate Central conducted their own special analysis to generate the values in these graphics, using techniques and general climate projections that are well established in the peer-reviewed scientific literature.4

The projections for extreme heat in the years 2050 and 2090 are based on an average of twelve computer models that simulate climate. For these projections, Climate Central used a scenario of moderate-high greenhouse gas emissions, which currently appears optimistic, since global emissions have exceeded this scenario in recent years. Climate Central used a common technique to translate large-scale climate information from the computer models to provide useful information about local and regional conditions. This method involves calculating differences between time series data from current and future global climate model simulations and then adding these changes to time series of observed climate data.

Scientists at Climate Central first identified weather observation stations closest to each city, as well as the closest point in the output of computer models, which is known as a grid point. For the station data, Climate Central examined temperature information for the summer months (June, July, and August) during two twenty-year periods to determine how extreme heat events have evolved during the twentieth century. Those periods were 1951-1970 and 1981-2000.

For the computer model data, Climate Central looked at two future twenty-year periods of projected maximum temperature data, 2046-2064 and 2081-2100, as well as the simulated current climate period 1981-2000. These data shed light on how extreme heat events may evolve as the climate changes. The analyses are based on recent data from weather stations, regional-scale outputs from climate projection models, and a common technique for deducing best-guess local climate projections from regional projections. This method involves calculating differences between current and future global climate model simulations, and applying them to observed climate data from the same vicinity.

For a given month, Climate Central calculated changes in the twenty-year average monthly maximum temperature between the two periods in the station data and between future time periods in the computer simulations and the current climate simulation. This provided a comparison of the twenty-year climatology at the end of the twentieth century and the earlier period 1951-1970. It also permitted a comparison of the model-simulated average monthly maximum temperature in 2046-2064 with that of 1981-2000, and the end-of-the-century period 2081-2100 with 1981-2000.

Because data from twelve different computer models were used, Climate Central took a model average of the differences in simulated temperature changes between the time periods. Next, Climate Central turned to the station observations and used the daily temperature data for the period 1981-2000 to determine a climatology of daily temperature values. From this climatology came the set of numbers to go into the second bar of each group of four bars. Each group pertains to a given temperature threshold of 90°F, 95°F, and 100°F. Climate Central counted how many days exceeded the temperature threshold during each of the twenty months in the climatology and then averaged those numbers.

Climate Central generated the other bars similarly, simply shifting the climatology of daily maximum temperatures for the period 1981-2000 by the temperature changes computed from the model simulations (or the 1951-1970 observed data) and repeating the count on the new sets of daily maximum temperatures that were generated.

This last step created new, simulated data for each city for twenty Augusts in the middle of the twenty-first century. The same method that was used with actual 1981-2000 temperatures to estimate the average number of days over each temperature threshold in this future scenario was then applied.

The resulting projections give long-term averages, not predictions for any individual year; actual outcomes will vary significantly from year to year, owing to the natural variability of climate. Furthermore, because the modeling and methods used involve uncertainty, the projections should be taken as best guesses within a range of uncertainty. True long-term averages will be likely to prove somewhat higher or lower than the projections here. However, all twelve models are unanimous in projecting increased hot days (relative to the present) by the middle of the twenty-first century.

All model outputs used were based on a scenario of medium-high greenhouse gas emissions, called A1B by the Intergovernmental Panel on Climate Change. Carbon dioxide and other greenhouse gas emissions during the decade spanning 2000-2010 have already exceeded the A1B scenario, so the projections here are conservative and represent a future in which greenhouse gas emissions are reduced compared with the current trend.

The number of hot July and August days (above 90°F, 95°F, and 100°F) in twenty U.S. cities over four time periods (1951-1970, 1981-2000, 2046-2064, 2081-2100) are presented below. Cities are listed in order of population size. For a given month (in this case, July or August), changes in the twenty-year average monthly number of high temperature are represented by four columns. The first two columns represent two distinct twenty-year periods (1951-1970, labeled 1960; and 1981-2000, labeled 1990) in the station data. The second two columns represent best guesses about the future, with column three representing the twenty-year climatology in 2046-2064 (labeled 2050) compared with the twenty-year climatology in 1981-2000, both model derived; and column four representing the twenty-year climatology in 2081-2100 (labeled 2090) compared with the twenty-year climatology in 1981-2000, both model derived.

New York


Los Angeles








San Francisco






Washington, D.C.
























3. Temperature Trends by State

It’s always useful to look at the long-term temperature trend at a more regional level. For example, the table below shows how much each state in the continental United States has warmed, on average, over the period 1976-2005. The first column shows the long-term trend averaged over the entire year (January to December). The second column shows the long-term trend for winter (December to February). You’ll notice that winters have warmed significantly.


*Asterisk denotes statistical significance at the 90 percent confidence level. (SOURCE: C. TEBALDI, CLIMATE CENTRAL)