Our climate has been constantly changing since Earth began, with periods of global warming and global cooling long before human beings and their activities began. Modern weather measurements go back only 100 to 150 years. So how do we know what Earth and its climate were like in prehistoric times?

Some clues come from Earth itself. Natural recorders of climate variations are called proxies; they include ocean and lake sediments, ice cores, fossils, tree rings, and corals. Through proxies, scientists can deduce long-term regional temperature trends as well as changes in the atmosphere's chemical makeup. For example:

• Tree rings show what climate factors shaped each year of a tree’s life.
• Bubbles trapped inside ice fields hold air from many thousands of years ago.
• Tiny ocean creatures called foraminifera reflect the state of temperatures and nutrients.
• Shells and other debris deposited in layers of sediment on the bottom of lakes and oceans can be analyzed to provide insights into past climate, year by year.

We also know that the orbits of the Sun and Earth undergo a variety of cyclic shifts over thousands and millions of years. Tiny changes in the tilt of Earth and the asymmetry of its path around the Sun can make a big difference to regional climate—sometimes enough to trigger an ice age.

Research groups at NCAR and other centers use data from climate proxies like tree rings and air bubbles trapped in ice cores to reconstruct the ups and downs in global average temperature.

With these clues in hand, NCAR scientists peer into our climate history using numerical models run on large supercomputers. These complex, sophisticated software packages can trace the flow of climate over hundreds and even thousands of years. They rely on our basic understanding of atmospheric physics and chemistry, as well as the clues gathered from Earth itself.

Paleoclimate models can look even further back, reproducing slices of atmospheric time from millions of years ago. One NCAR study of the climate 251 million years ago lends support to the notion that a sudden increase in carbon dioxide helped trigger the greatest mass extinction in Earth’s history.

Scientific and public debate has swirled around overall portraits of temperature increase like the one shown here. These are often dubbed "hockey stick" graphs because of the sharp rise on the right-hand side. Recent research at NCAR and elsewhere has bolstered the case for this representation of a dramatic temperature increase over the last century.

Critics have questioned the statistical methods used to convert temperature proxies into global averages for the period prior to the late 1800s (when modern observations began). A number of recent studies have re-examined the methods used and suggested adjustments. Yet the basic message of the hockey stick remains valid, as a panel of the National Academy of Sciences found in a 2006 report. That panel concluded that the warming trend since 1900, and especially since the 1970s, is highly unusual and perhaps unprecedented in the last millennium.

The black line starting in the late 1800s shows actual temperature measurements, known as the instrumental record. All the other colors show reconstructions based on individual proxies (such as tree rings or ice cores from different ice caps), or on combinations of proxies. The multiproxy studies take advantage of the strengths of the separate proxies while minimizing the influence of limitations associated with the individual records.

The gray background gives a qualitative idea of how uncertainty increases further back in time. The darker the background, the greater the uncertainty regarding the temperature signal emerging from a reduced number of samples.

Temperatures by the mid-20th century were at or above the levels seen during the "Medieval Warm Period," near the start of the graph in 1000 C.E. Since the 1970s, the global average has risen more than 0.4°C (0.7°F). (Image courtesy National Research Council, National Academy of Sciences.)

Research groups at NCAR and other centers use data from climate proxies like tree rings and air bubbles trapped in ice cores to reconstruct the ups and downs in global average temperature. Research groups at NCAR and other centers use data from climate proxies like tree rings and air bubbles trapped in ice cores to reconstruct the ups and downs in global average temperature. This graphic from a 2006 report by a panel of the National Academy of Sciences shows seven different reconstructions of the past 1100 years of climate, from 900 to 2000. Each of these studies used different methods and a different set of proxies collected from a different part of the Northern Hemisphere.

View a recently modeled animation of Earth's warming. Movie file is large, but worth the wait. For more than 100 years, Earth's surface temperature has been monitored by a global network of land-based weather stations. These reports are supplemented by sea-surface and air temperature readings taken at points across the oceans that cover 70% of the globe. The ups and downs of air temperature are modulated by the sea, so the uppermost ocean serves as a good index of the average air temperature just above it.

Together, these data show that Earth's surface air temperature has risen more than 1.1°F (0.7°C) since the late 1800s. This warming of the average temperature around the globe has been especially sharp since the 1970s. Global models at NCAR have simulated 20th century climate and found three main factors at work:

1. Solar activity contributed to a warming trend in global average temperature from the 1910s through 1930s.
2. As industrial activity increased following World War II, sun-blocking sulfates and other aerosol emissions helped lead to a slight global cooling from the 1940s to 1970s.
3. Since 1980, the rise in greenhouse gas emissions from human activity has overwhelmed the aerosol effect to produce overall global warming.

Some urban areas have also warmed due to the heat-island effect, where buildings and pavement retain more heat than undeveloped areas and cause more runoff and thus drier conditions as well. NCAR scientists and their colleagues have worked carefully to remove urban heat-island effects and other potential biases from the global record. Even after these adjustments, the rise in global temperature remains clear.

There are other signs of a warming planet. Glaciers are retreating, especially atop lower-latitude mountains. In the Arctic the thickness and extent of summer sea ice have decreased dramatically over the last 50 years, and recent modeling by NCAR scientists shows that the Arctic’s summer ice may virtually disappear by 2040. Meanwhile, snowfall over much of Antarctica is increasing, a paradoxical sign of warming temperatures in this frozen, arid land. The annual cycle of plants and migrating animals shows a lengthening of the warm season over much of the Northern Hemisphere.

Since the late 1970s, satellites have measured the temperature in a broad layer of the troposphere several miles above Earth. For years, they showed a smaller temperature rise at these heights than at ground level. It now appears that most of the disagreement was due to errors in the satellite data and how it was interpreted. While there are still differences between tropospheric and surface warming in some regions, the discrepancy is no longer apparent on a global scale, according to a 2006 U.S. Climate Change Program report.

Studies of past, present, and future climate benefit continually from these and other improvements in data gathering, computer modeling, and analysis. For example, recent research at NCAR re-examines the role of decades-long cycles of solar variation in explaining the observed warming in the first half of the 20th century.

Researchers have identified an additional impact of humanity's increasing consumption of fossil fuels: changes in seawater chemistry as the oceans absorb more carbon dioxide. These changes pose a real threat to marine organisms, including those that build coral reefs around the world.

Global average temperature since 1890 as reproduced by the NCAR/DOE Parallel Climate Model.

Why should Earth be warming? The amount of energy emitted by the Sun has risen a fraction of a percent since 1900. However, climate simulations at NCAR have shown that solar changes explain less than a third of the warm-up during the last century. The most straightforward explanation for a warming Earth is the greenhouse gases emitted when fossil fuels are burned in homes, gas and coal-fired power plants, vehicles, and factories.

The average number of frost-free days per year is projected to rise across much of the globe by the 2080s, with the largest increases (red and orange) across the western fringes of North America and Europe. (Illustration courtesy Gerald Meehl and Claudia Tebaldi, NCAR.) Water vapor, carbon dioxide (CO2), and other greenhouse gases act to warm Earth’s atmosphere by absorbing, and trapping, some of the outgoing radiation from the Earth and reradiating some of this energy back to the surface. The amount, or concentration, of CO2 gas in the atmosphere has risen more than 30% since widespread fossil-fuel use began with the start of the industrial revolution in the late 1700s. CO2concentration in the atmosphere is now at its highest point in more than 600,000 years.

Each year, the CO2 concentration increases by about 0.5%. Because CO2 has a lifetime in the atmosphere of over 100 years, its atmospheric concentration will continue to increase as long as emissions from human activities continue.

Many NCAR scientists are part of a global team studying this problem and its meaning for our planet's future. The Intergovernmental Panel on Climate Change (IPCC) includes more than 1,000 experts from a variety of climate specialties. The next IPCC report is expected to be issued around 2013.

In their previous report (2001), the IPCC predicted that increasing levels of greenhouse gases will warm the globe by a significant amount. The most probable range, according to the IPCC, is between 2.5 and 10.5°F (1.4–5.8°C) over 1990 levels by the year 2100. Also in 2001, an NCAR scientist and his colleague estimated a 90% likelihood that the range will fall between 3 and 9°F (1.7–4.9°C).

What does this mean for society? A vast majority of climate scientists agree with the IPCC consensus that Earth will warm along with increasing greenhouse gases. However, the effects will be far more varied than a simple and uniform warming over the entire planet, because heating also alters the water cycle, among other changes. As a result, some regions will become considerably hotter or cooler, or wetter or drier, than others.

Several national studies have addressed these regional consequences, including those for the United States, Canada, and the United Kingdom. Some aspects of regional climate change are already well established. For instance, high-latitude areas such as Canada, Russia, and the Arctic are warming more rapidly than the tropics, as predicted by computer models. This trend is expected to continue. In many nations, rainfall and snowfall are becoming more concentrated in heavier bursts, and regions poleward of latitude 40 degrees north are expected to see more days with heavy precipitation. NCAR scientists and colleagues are working to improve understanding of other potential regional changes in climate, such as where U.S. rainfall and snowfall patterns might shift.

Researchers are also working to improve techniques for assessing the impacts of a changing climate on environment and society. One method is to translate temperature changes from a model into trends that affect people's everyday lives. A 2004 NCAR study found that, by the period 2080-99, American and European heat waves will be more severe, frequent, and long-lasting.

A related study found that frost days (those in which temperatures dip to 0°C or 32°F) will decline in many parts of the globe by 2080-99. The largest decreases are projected across the northwest parts of Europe and North America, as mild marine air becomes more prevalent in winter. Such a change would affect agriculture and tourism as well as natural ecosystems.

New research in 2006 by NCAR scientists and colleagues looked more specifically at the potential for an increase in weather extremes in a warmed climate. The researchers used simulations from nine different climate models to demonstrate the risk of dangerous heat waves, intense rains, and other kinds of extreme weather in the next century.

These three studies, along with many others, have been used in preparing the 2007 IPCC assessment. One theme for the new assessment will be the extent to which our planet is committed to some amount of human-induced climate change, regardless of what actions we take in the future.

Forecasting regional climate poses a challenge similar to that faced by weather forecasters every day. Forecasters may call for a 70% chance of rain with confidence, even if they cannot say where each thunderstorm will form. In much the same way as that forecast for a chance of rain, scientists are finding probabilities a useful way to describe how regional climate will change in a warming world. Long-range climate outlooks may give the chance of a week of 100-degree weather, for example, and how that chance would increase as the decades unfold. NCAR scientists are developing methods to determine the probabilities of future changes to the climate on regional scales.

The average number of frost-free days per year is projected to rise across much of the globe by the 2080s, with the largest increases (red and orange) across the western fringes of North America and Europe. (Illustration courtesy Gerald Meehl and Claudia Tebaldi, NCAR.) Water vapor, carbon dioxide (CO2), and other greenhouse gases act to warm Earth’s atmosphere by absorbing, and trapping, some of the outgoing radiation from the Earth and reradiating some of this energy back to the surface. The amount, or concentration, of CO2 gas in the atmosphere has risen more than 30% since widespread fossil-fuel use began with the start of the industrial revolution in the late 1700s. CO2concentration in the atmosphere is now at its highest point in more than 600,000 years.

Most climate change simulations are created with models that simulate the global scale and produce global averages as their results. But to understand how global warming will affect drinking water storage or the ability to grow wheat, corn, and other staples, regional simulations and impact studies are needed.

Some Challenges of Regional Climate modeling

Regional models depict the climate of a smaller area in more detail, which is challenging because:

• scientists are still studying the influence of clouds and local storms and how best to portray them in regional models; global models do not try to represent these fine-scale processes
• many of the other equations created to represent physical processes in global models must have more spatial detail
• the models must reflect changes in the atmosphere, land, oceans, and other parts of the envirronment over shorter time scales
• all of this higher resolution usually requires greater computing power and longer time to run a simulation

NCAR scientists and their colleagues are addressing these issues through development of a nested regional climate model (NRCM). The model's developers plan to:

• Seamlessly integrate the weather and climate models based at NCAR to capture all important spatial and temporal scales
• Collaborate with impacts researchers to address both scientific and societal issues
• Provide a new community model freely available to researchers around the world, such as the NCAR-initiated Community Climate System Model (CCSM) and Weather Research and Forecasting Model (WRF).

Scientists hope to improve understanding and simulation of complex, two-way scale interactions, with emphases on:

• downscaling from global climate simulations (especially in investigations of societal impacts)
• upscaling from regional processes, including the effects of land and ocean processes
• the impact of regional weather, particularly the organization of moist convection on larger scales.

Our Current Understanding

Some aspects of regional climate change are already well established. For instance, high-latitude areas such as Canada, Russia, and the Arctic are warming more rapidly than the tropics, as predicted by computer models. This trend is expected to continue. In many nations, rainfall and snowfall are becoming more concentrated in heavier bursts, and regions poleward of latitude 40 degrees north are expected to see more days with heavy precipitation. NCAR scientists and colleagues are working to improve understanding of other potential regional changes in climate, such as where U.S. rainfall and snowfall patterns might shift.

Researchers are also working to improve techniques for assessing the impacts of a changing climate on environment and society. One method is to translate temperature changes from a model into trends that affect people's everyday lives. A 2004 NCAR study found that, by the period 2080-99, American and European heat waves will be more severe, frequent, and long-lasting.

A related study found that frost days (those in which temperatures dip to 0°C or 32°F) will decline in many parts of the globe by 2080-99. The largest decreases are projected across the northwest parts of Europe and North America, as mild marine air becomes more prevalent in winter. Such a change would affect agriculture and tourism as well as natural ecosystems.

Both of these studies, along with many others, have been used in preparing the 2007 IPCC assessment. One theme for the new assessment will be the extent to which our planet is committed to some amount of human-induced climate change, regardless of what actions we take in the future.

Uncovering North America's climate future

What will climatic conditions look like for the United States and Canada by the middle of this century? An international team of scientists is focusing in on North American climate from 2040 to 2070, laying the groundwork to create regional simulations with unusually fine detail.

Most projections of future conditions rely on global climate models run on supercomputers that, despite their sophistication, lack the detail to simulate behavior within a state or region. Led by NCAR, the North American Regional Climate Change Assessment Program (NARCCAP) will use an ensemble of global climate models and high-resolution regional climate models to produce simulations with about triple the resolution of most projections of future climate.

The combination of tools will allow scientists to incorporate relatively small topographical features, such as mountain ranges, lakes, and complex coastlines, that can have a significant influence on local and regional climate. An important research benefit of the effort, which is designed to complement similar projects in Europe and South America, will be the ability to compare the results of fine-scale with coarser-scale modeling to more clearly determine the added value of high-resolution projections of future climate.

NCAR scientists are collaborating with colleagues at U.S. universities and laboratories, the University of Quebec and the Ouranos Consortium in Montreal, and Britain’s Hadley Centre for Climate Prediction and Research. The project is expected to spawn additional international collaborations as researchers continue to fine-tune projections of the impacts of climate change on North America.

A thunderstorm cloud passes over the plains east of Denver. Researchers believe that some regions will see warmer and wetter weather, while others will have droughts.

Researchers are developing new approaches to fine-scale computer modeling to allow greater focus on regional impacts of a changing climate. (Image courtesy North American Regional Climate Change Assessment Program, NCAR.)

For almost five centuries, coastal residents of Peru noted and recorded a strange feature of the eastern Pacific Ocean waters that border their home. In the first months of each year, a warm southward current usually modified the normally cool waters. But every few years, this warming started earlier—in December. It was far stronger than usual, and it lasted as long as a year or two. Torrential rains fell on the arid land; as one early observer put it, "The desert becomes a garden." However, warm waters flowing south also shut off the deeper, cooler waters that are crucial to sustaining the region's marine life.

El Niño Visualizations

This is El Niño, "The Baby Boy," so named by the Peruvians because of its typical appearance around Christmas time. Once thought to affect only a narrow strip of water off Peru, El Niño is now recognized as a large-scale oceanic warming that affects most of the tropical Pacific. The weather impacts related to El Niño and its counterpart, La Niña (a cooling of the eastern tropical Pacific), extend throughout the Pacific Rim to eastern Africa and beyond.

El Niño is normally accompanied by a change in atmospheric circulation called the Southern Oscillation. The ENSO phenomenon (standing for El Niño–Southern Oscillation) is one of the main sources of interannual, or year-to-year, variation in weather and climate around the world.

Since recognizing some 30 years ago that the oceanic and atmospheric parts of ENSO are strongly linked, scientists have moved steadily toward a deeper understanding of El Niño. Climate forecasters have taken the first steps toward predicting the onset of El Niño and La Niña events months in advance.

NCAR scientists have been at the forefront of research placing El Niño in a global context, analyzing ENSO and its manifestations around the world.

In the 1970s and 1980s, social scientists at NCAR were among the first to examine how the phenomenon affected the people of South America and other regions. They have since collaborated on international projects involving dozens of nations in examining how forecasts of El Niño and La Niña could improve their citizens' lives.

Groundbreaking climate simulations have shown the occurrence of El Niño as far back as the last ice age. And NCAR scientists have demonstrated how El Niño, La Niña, and other regional cycles of the ocean and atmosphere can influence weather and climate thousands of miles away.

Computer models disagree on how El Niño and La Niña will be affected by long-term climate change. One possibility is a steady-state warming in the central Pacific that would have some characteristics of El Niño.

Water has been labeled the environmental issue of the 21st century. When reliable rainfall disappears from a region for years or even decades, the impacts on plants, animals, and people can be profound.

Many of the world's droughts are steered and shaped by large-scale climate cycles, such as El Niño and La Niña. Through exhaustive analyses, NCAR scientists have helped pinpoint how these cycles can produce drought at far-flung locations. Case studies show the value of prompt drought predictions when societies can relay the information to farmers, policy makers, and other key people.

As the global climate warms, how will that affect the location and duration of droughts? Globally, the average amount of rainfall is expected to increase, as the warmer temperatures lead to more evaporation of water from the sea. However, the same warmth will help dry out land more quickly, so the droughts that do occur may increase in intensity. The percentage of Earth's land area stricken by serious drought more than doubled from the 1970s to the early 2000s.

Computer-based simulations by NCAR scientists and their colleagues have been used to track areas of persistent dryness in the climates of the past and to project where these areas might develop in the future. These models do not yet agree on which specific areas will be most at risk for drought. However, the models tend to produce drying in the hearts of midlatitude continents, such as the central United States, eastern Europe, and western Asia. Southern Australia—already in a severe multiyear drought—is at enhanced risk for further drying, as the storm track encircling Antarctica shifts poleward.

Where drought does strike, the risk of wildland fire soars. NCAR takes a multidisciplinary approach to address this concern. Fine-scale models have simulated the erratic behavior of a fast-moving wildland fire. Mobile radars have sampled the hot, gusty winds of wildfires at close range. And atmospheric chemists have measured mercury and other environmental hazards in the witch's brew of smoke and gas spewed from a raging wildland blaze. A blend of satellite data and modeling revealed that wildfires in Alaska and Canada in 2004 emitted about as much carbon monoxide as did human-related activities in the continental United States during the same period. The fires also increased ground-level ozone across much of the Northern Hemisphere.

Wildland fires are part of a feedback loop that relates to global climate change. When trees and grasses burn, they release carbon dioxide, thus adding to the greenhouse effect and raising the risk of future heat-stoked wildfires.

Research at NCAR and elsewhere shows that the policy of suppressing fires over the past century has locked up about 25% of Earth's carbon budget in forest vegetation. If more forests burn, whether from wildfires or increased prescribed burning, more carbon dioxide will be released back into the atmosphere, where it will join the increased emissions from vehicles and industry.

This depiction of linear trends in the Palmer Drought Severity Index from 1948 to 2002 shows drying (reds and pinks) across much of Canada, Europe, Asia, and Africa and moistening (green) across parts of the United States, Argentina, Scandinavia, and western Australia. (Illustration courtesy Aiguo Dai, NCAR, and the American Meteorological Society.)