Chapter 2: Observations

Although Earth’s climate is currently changing rapidly relative to past changes, in most regions where we live changes are slow enough that we do not notice them directly during our daily lives. However, older people may have noticed changes during their lifetimes and in some regions changes are larger and more obvious than elsewhere. In this chapter we will discuss some observations from the past 100 years and data from regions that are particularly sensitive to climate change, where the most dramatic effects have occurred. This will not be a comprehensive documentation of existing observations. Additional observations will be discussed throughout the remainder of this book.

a) Atmosphere

Observations show that climate is changing on a global scale. Surface air temperature data, averaged over the whole Earth, indicate warming of about 1°C over the last ~100 years (Fig. 1). But that warming was neither steady nor smooth. From 1880 to about 1910 there was cooling, followed by warming until about 1940. After this slight cooling or approximately constant temperatures were observed until the 1970s, followed by a rapid warming until the present. Each year’s temperature is somewhat different from the next. Not all of these year-to-year changes are currently understood, but natural variations, which are not caused by humans, do play a role. For instance, strong El Niño years such as 1997-98 or 2015 show up as particularly warm years, whereas La Niña years such as 1999-2000 or 2011 show up as relatively cool. El Niño and La Niña, also known as ENSO (El Niño/Southern Oscillation), is climate variability in the tropical Pacific that impacts many other regions of the Earth.

Figure 1: Global average surface temperature change from 1880 to 2018. Anomalies from the 20th century average are shown. From the National Oceanographic and Atmospheric Administration’s (NOAA) web site noaa.gov. There are a few other groups around the world who calculate global average surface temperatures, e.g. the National Aeronautics and Space Administration’s (NASA) Goddard Institute for Space Science (GISS), the Climate Research Unit (CRU) of the University of East Anglia, or the Berkeley Earth Surface Temperatures (BEST), and all come to very similar results. Errors due to unequal sampling have been estimated to be about 0.2°C during the early part of the record and decrease to 0.1°C during the more recent part. This is a key figure. The linear trend over the full period is 0.08°C/decade, that over the last 100 years is 0.10°C/decade and that over the last 50 years is 0.18°C/decade.

Explore Temperature Data

Go to NOAA’s Climate at a Glance website to explore their global temperature data. Select “annual” to get the annual averages. Now list the 20 warmest years by clicking on the ANOMALY column in the table below until you have the warmest year listed on top.

Which is the record warmest year?
How many of the 20 warmest years have occurred since the year 2000?

Click on the “Display Trend” button and enter different start and end dates.

What is the trend over the last 50 (last 100) years?

Interested in how temperature measurements from long ago have been recorded? Have a look at this fascinating interactive article.

The increase in atmospheric temperatures over the last 100 years has not been uniform everywhere. Fig. 2 shows that temperatures over land changed more than over the oceans and the Arctic warmed more than the tropics. These patterns are called land-sea contrast and polar amplification and are quite well understood and simulated in climate models as we will see later. The warming is indeed almost global. The only exception is the northern North Atlantic, which has been slightly cooling for reasons we will discuss later.

Surface Temperature Change Map Last 50 Years

Figure 2: Global Map of Recent Surface Temperature Change. From Wikipedia’s Climate Change based on data from NASA’s Godard Institute for Space Studies (GISS).

Analysis of thousands of temperature records into an estimate of global mean temperature change such as Fig. 1 is not trivial. Issues such as changes in station locations, instrumentation and data coverage have to be taken into account. The fact that five different groups analyzing the data using different methods come to the same conclusions suggests that the results are robust. The reliability of the data has been demonstrated and it has been shown that station locations (e.g. urban versus rural) don’t matter. See this blog for more discussion.

The global warming trend has made the probability of warm and extreme warm temperatures more likely and the probability of cold and extreme cold temperatures less likely. E.g. extremely warm (more than 3°C warmer than the 1951 to 1980 average) summer temperatures over Northern Hemisphere land areas had only a 0.1% chance of happening from 1951 to 1980, whereas during the decade from 2005 to 2015 the chance of those extreme warm temperatures to occur was 14.5%. Conversely, relatively cold summer temperatures (less than 0.5°C cooler than the 1951-1980 average) that happened about every third year from 1951 to 1980 only occurred 5% of the time during 2005 to 2015. See this article and this blog for additional information and graphs.

b) Cryosphere

One of the most sensitive regions to climate change is the Arctic. Sea ice cover there has decreased dramatically over the past 40 years particularly in late summer (Fig. 3). Arctic sea ice experiences a strong seasonal cycle. In late winter it covers about 15.4×10⁶ km² and decreases to about 6.4×10⁶ km²in late summer (Fig. 4). Therefore, the relative changes are larger in late summer, with a reduction of about 46 % or 2.9×10⁶ km² from 1980 to 2015. In winter the reduction was only about 9 % or 1.5×10⁶ km². The lost area of summer Arctic sea ice is more than 10 times the size of Oregon (255,000 km²) or four times the size of France (~650,000 km²).

Figure 3: Arctic sea ice extent in September 1979 (left) and 2020 (right) from satellite observations. The purple line denotes the median ice extent from 1981-2010. In 1979 the ice extent was 7.1 million sq km, in 2020 it was 3.9 million sq km. Images from the National Snow and Ice Data Center (NSIDC).

In the Antarctic, sea ice experiences an even larger seasonal cycle but it has changed less over the last 40 years compared to the Arctic. Southern hemisphere sea ice has slightly increased. Note that year-to-year fluctuations are larger Antarctic than in the Arctic, whereas the long term trends in the Antarctic are much smaller than in the Arctic. This combination of larger short-term fluctuations and a smaller long-term trend makes Antarctic sea ice trends less statistically significant. A trend is statistically significant if it is larger than the uncertainties from short term fluctuations.

sea ice

Figure 4: Changes in late summer sea ice extent in the Arctic (top) and Antarctic (bottom), and global monthly sea ice extent (center). From NOAA.

Globally, Earth has lost about 2 million square kilometers of sea ice from 1980 to 2015, which is about 10% of the total. Compare that area to that of your favorite state or country.

Explore Sea Ice Changes

Go to the National Snow and Ice Data Center’s (NSIDC) website and click on a few years to see how Arctic sea ice cover has changed over the year. To get a time series for the current month go to their sea ice index site and click on the Monthly Sea Ice Extent Anomaly Graph in the lower right corner.

By how much has the sea ice cover decreased since the 1980s? Estimate the decrease both in relative terms (percentage) and in absolute terms (million square kilometers).

An animation is available here.

Mountain glaciers are also sensitive to climate change. Fig. 5 shows an example from Muir Glacier, which has retreated dramatically since 1941. This is typical for most glaciers around the world. In fact, only a small number of glaciers show advances, whereas the vast majority of glaciers melt and retreat from the valleys up into higher elevations.

Figure 5: Muir Glacier in Glacier Bay National Park and Preserve, Alaska. The picture on top is from 1941, that on the bottom from 2004. From the National Snow and Ice Data Center (NSIDC).

The World Glacier Monitoring Service (WGMS) has compiled information on hundreds of glaciers world-wide. Fig. 6 shows that since 1980 glaciers in all regions have been losing mass with an acceleration of loss in recent years. Watch this video of glacier changes in Iceland.

Figure 6: Cumulative mass balance in meters water equivalent (w.e.) of mountain glaciers. From the World Glacier Monitoring Service.

Explore Glacier and Ice Sheet Change

Go to the Glacier Browser and select a glacier of your choice.

What do you observe?

Explore Greenland ice sheet change with this interactive chart, which shows the surface area experiencing melting. Click on a few years in the early part of the record (e.g. 1980s) and in the more recent part (e.g. 2010s).

How has the melt area changed?

Ice sheets are also melting. Observations from the Gravity Recovery and Climate Experiment (GRACE) satellites, which measure very precisely Earth’s gravity field and can detect changes in mass, show that since 2002 the Greenland ice sheet has lost about 4,000 Gt of mass, and the Antarctic ice sheet has lost about 2,500 Gt (Fig. 7). Here is a presentation about Greenland melting. Melting of the Greenland ice sheet contributes currently about 0.8 mm/yr to global sea level rise. Antarctica’s contribution is 0.4 mm/yr and mountain glaciers add about 0.6 mm/yr, for a total of 1.8 mm/yr sea level rise from melting of ice (IPCC 2019).

Figure 7: Changes in the Antarctic (top) and Greenland (bottom) ice sheet mass. Pale lines are from satellite gravity measurements. From the US Environmental Protection Agency (EPA; epa.gov). Satellite data are from NASA.

Box 1: Rates of Change

The rate of change of a variable X between two points in time t₁ and t₂ can be calculated as the difference (denoted by the greek letter delta Δ) of the value of the variable at time t₂ minus the value of the variable at time t₁ ΔX = X₂ – X₁ divided by the difference in time Δt = t₂ – t₁

(B1.1) $\begin{equation*} \frac{\Delta X}{\Delta t} = \frac{X_2 - X_1}{t_2-t_1}\ . \end{equation*}$

Thus the units of the rate of change are the units of the variable divided by time. You can determine the rate of change from a timeseries graph such as Fig. 1 or Fig. 6 by selecting two points in time on the horizontal axis and reading the corresponding values of X₁ and X₂ from the vertical axis. Using Fig. 1 as an example our variable will be the temperature anomaly T. Choosing t₁ = 1940 and t₂ = 2010 we can read off T₁ = 0ºC and T₂ = 0.7ºC. Thus, ΔT = 0.7ºC, Δt = 70 years and the rate of change ΔT /Δt = 0.01ºC/yr = 0.1ºC/decade.

Obviously, calculating the rate of change in this way the resulting value will depend on the two times picked. The rate of change of a whole set of data points can be calculated by assuming a linear relationship and minimizing the distance of all points from a straight line X = S×t + I, where S is the slope and I the intercept with the vertical axis. This is called linear regression. Simple formulae to calculate S and I can be found here. Linear regressions are commonly used to estimate rates of change. E.g. the straight lines in Fig. 4 in this chapter and Figs. 7-12 in chapter 1 have been calculated using the formulae for linear regression. Regression lines are also often referred to as trend lines.

c) Ocean

Subsurface temperature measurements in the oceans document warming over the last 60 years (Fig. 8). The ocean’s heat content has increased by about 30×10²² Joules during that time. The heat content of the ocean is its temperature Ttimes the heat capacity of water c_p= 4.2 J/(gK). Prior to 2005 subsurface temperature measurements were more limited in space and time because they were taken from ships by lowering CTD(conductivity, temperature, depth) instruments on a cable into the ocean. Since 2005 autonomous, free-drifting Argofloats measure temperature, salinity, pressure, and velocity of the upper 2 km of the water column. Currently there are about 4,000 floats out there, which provide much better spatial and temporal coverage than the previous ship-based measurements.

Figure 8: Observational estimates of global ocean heat content (0-2 km) from NOAA. From 1960 to 2005 ocean temperatures were measured mainly using research ships (blue line). Since 2005 autonomous ARGO floats have increased the data density both in space and time (red and black lines).

The melting of mountain glaciers and ice sheets leads to increased runoff into the ocean, which contributes to sea level rise (Fig. 9). Sea level rise is also caused by warming sea water, which causes expansion and by increased runoff from pumping of groundwater out of aquifers. Estimates based on tide gauge records indicate that sea level has risen by about 20 cm from the 1870s to the year 2000 and another 6 cm since. The melting of mountain glaciers and ice sheets contributes about similarly to the current sea level rise, but if current trends continue it is likely that many mountain glaciers will completely disappear and the large ice sheets will contribute more and more to global sea level rise. Note that sea level rise is not spatially uniform.

Figure 9: Observed global mean sea level estimated from satellites (red) and tide gauges (blue). Data from nasa.gov and csiro.au. The pie chart shows the relative contributions from melting of land ice (1.8±0.1 mm/yr), thermal expansion of sea water (1.4±0.3 mm/yr) and groundwater pumping (0.4 mm/yr; estimated as the residual) during 2005-2015 (IPCC, 2019). The total rate of sea level rise during this period (3.6±0.5 mm/yr) is larger than during earlier periods. This acceleration of sea level rise is caused by recently increased melting of the Greenland and Antarctic ice sheets.

Explore Sea Level Changes

Goto noaa.gov and explore local sea level changes from tide gauges.

Where is sea level increasing?
Where is it decreasing?
Select a tide gauge. What is the time period covered?

Here is a map of sea level measured from satellites.

What time period is covered by the satellite data?
Where is sea level increasing?
Where is it decreasing?

d) Biosphere and Carbon Cycle

Plant and animal species have been observed to move poleward and upward in the mountains (Parmesan and Yohe (2003). This response is consistent with global warming and the tendency of organisms to stay in a temperature range to which they have adapted. Flowering dates of many plants have also shifted earlier in the spring such as for cherry blossoms.

Carbon dioxide (CO₂) has been measured in the atmosphere since 1958 at Mauna Loa Observatory in Hawaii (Fig. 10). At that time concentrations were just below 320 parts per million (ppm). Subsequently they increased to values of just over 400 ppm today. That is a 25 % increase. Overlaid on the long term trend is a seasonal cycle. Growth of the terrestrial biosphere in northern hemisphere spring leads to CO₂ drawdown and decay of organic matter such as fallen leaves increases CO₂ in the fall. As we will see later, CO₂ is an important greenhouse gas, and its increase over the past decades is the main cause of the recent global warming.

Figure 10: Atmospheric CO₂ measured at Hawaii’s Mauna Loa Observatory. The measurements were pioneered by Charles Keeling from Scripps Institution of Oceanography in 1958. From noaa.gov.

Chapter 1: Weather

a) Weather and Climate

Weather and climate are related but they differ in the time scales of changes and their predictability. They can be defined as follows.

Weather is the instantaneous state of the atmosphere around us. It consists of short-term variations over minutes to days of variables such as temperature, precipitation, humidity, air pressure, cloudiness, radiation, wind, and visibility. Due to the non-linear, chaotic nature of its governing equations, weather predictability is limited to days.

Climate is the statistics of weather over a longer period. It can be thought of as the average weather that varies slowly over periods of months, or longer. It does, however, also include other statistics such as probabilities or frequencies of extreme events. Climate is potentially predictable if the forcing is known because Earth’s average temperature is controlled by energy conservation. For climate, not only the state of the atmosphere is important but also that of the ocean, ice, land surface, and biosphere.

In short: ‘Climate is what you expect. Weather is what you get.’

b) The Climate System

Earth’s climate system consists of interacting components (Fig. 1). The atmosphere, which is the air and clouds above the surface, is about 10 km thick (more than two thirds of its mass is contained below that height). The ocean covers more than two thirds of Earth’s surface and has an average depth of roughly 4 km. Contrast those numbers with Earth’s radius which is approximately 6,400 km and you’ll find that Earth’s atmosphere and ocean are very thin layers compared to the size of the planet itself. In fact, they are about 1,000 times thinner. They are comparable perhaps to the outer layer of an onion or the water on a wet soccer ball. Yet all life is constrained to these thin layers. The major ocean basins are the Pacific, the Atlantic, the Indian, and the Southern Ocean. Ice and snow comprises the cryosphere, which includes sea ice, mountain glaciers and ice sheets on land. Sea ice is frozen sea water, up to several meters thick, floating on the ocean. Ice sheets on land, made out of compressed snow, can be several kilometers thick. The biosphere includes all living things on land and in the sea from the smallest microbes to trees and whales. The lithosphere, which is the solid Earth (upper crust and mantle), could also be considered an active part of Earth’s climate sytem because it responds to ice load^[1] and impacts atmospheric carbon dioxide (CO₂) concentrations and climate on long timescales through the movements of the continents.

The blue marble. A composite image of Earth from space. It shows all four components of Earth’s climate system. The atmosphere with its complex cloud patterns. The ocean, which covers about 70% of Earth’s surface. The cryosphere is visible as the white areas on the top: sea ice covering the Arctic Ocean and the Greenland ice sheet. Green colors on land and turquoise shades along the ocean’s margin indicate the biosphere as forests and phytoplankton blooms. Notice in the lower left the thin layer of the atmosphere surrounding Earth.

Figure 1: The blue marble. A composite image of Earth from space. It shows all four components of Earth’s climate system. The atmosphere with its complex cloud patterns. The ocean, which covers about 70% of Earth’s surface. The cryosphere is visible as the white areas on the top: sea ice covering the Arctic Ocean and the Greenland ice sheet. Green colors on land and turquoise shades along the ocean’s margin indicate the biosphere as forests and phytoplankton blooms. Notice in the lower left the thin layer of the atmosphere surrounding Earth. From nasa.gov.

The components interact with each other by exchanging energy, water, momentum, and carbon thus creating a deliciously complex coupled system. Imagine water evaporating from the tropical ocean heated by the sun (Fig. 1). The air containing that water rises and cools. The water condenses into a cloud. The cloud is carried by winds over land where it rains. The rain sustains a forest. Trees are dark, having a low albedo. This influences the amount of sunlight absorbed by the Earth. Dark surfaces absorb more sunlight and get warmer compared to bright surfaces such as desert sand or snow. Air warmed by the surface rises and affects the wind.

c) Processes

Fig. 2 illustrates some of the important processes that contribute to the complex interactions within the climate system. Earth’s energy source is the sun. Both solar and terrestrial radiation are affected by gases, aerosols, and clouds in the atmosphere. Thus, the atmospheric composition affects the heating and cooling of the earth. Heating and cooling affect the temperature and circulation of the atmosphere and oceans. Circulations of the air and sea affect temperatures and precipitation over both ocean and land, which impact the biosphere and cryosphere. Atmosphere and oceans exchange heat, water (evaporation and precipitation), and momentum. Wind blowing over the ocean pushes the surface water ahead. Air temperatures and snow fall affect the growth and melting of glaciers and ice sheets. Water from melting ice flows through rivers into the ocean affecting its salinity, its density and movement. Variations in solar irradiance can cause global climate to change. Volcanoes can eject large amounts of aerosol into the atmosphere with climatic implications. Humans are influencing the climate system through emissions of greenhouse gases, aerosols, and land use changes.

Figure 2: Schematic view of components of the climate system and processes involved in their interactions. From Le https://www.ncdc.noaa.gov/cagTreut et al. (2007).

The complexity makes studying the climate system challenging. Scientists from many disciplines contribute such as physicists, chemists, biologists, geologists, oceanographers, atmospheric scientists, paleoclimatologists, mathematicians, statisticians, and computer scientists. To me, the challenges and interdisciplinary nature of climate science are fascinating and fun. I learn something new about it every day.

Explore

Go to one of the following websites and explore temporal variations of surface temperature observations in a location of your interest. Plot today’s data, yesterday’s, last week’s, last month’s, last year’s, the last decade’s, as far back as you can and make some notes on what you observe. Note regular, predictable cycles such as the diurnal and annual cycle in contrast with irregular, unpredictable variations such as day-to-day and year-to-year fluctuations. What causes the regular cycles? Try to order the variations with respect to their amplitude (strength). Discuss with fellow students and instructor.

Hourly and daily data are available at: https://www.wunderground.com/history/. Enter a location of your interest. Click on the “Weekly”, “Monthly”, or “Custom” tabs to select different time periods. Unfortunately, when I tried it Jan. 30th, 2017 the “Custom” option didn’t allow one to plot more than one year of data. However, you may be able to identify diurnal (daily) cycles in temperature. Fig. 3 shows an example. Although some nights can be warmer than some days and sometimes days and nights have the same temperature, on average days are warmer than nights. Thus the average diurnal cycle is part of climate. It is forced by the diurnal cycle of solar irradiance. Does pressure have a diurnal cycle? Answer: no. What is the typical timescale of pressure variations? Often it is about a week or so. This timescale is associated with the transition of weather systems (high and low pressure systems) passing by.

Weather data from Corvallis, Oregon from January 2017. In the upper panel we can see diurnal cycles in temperature. Pressure varies on longer timescales. Low pressure is associated with storms (high winds and wind gusts).

Figure 3: Weather data from Corvallis, Oregon from January 2017. In the upper panel we can see diurnal cycles in temperature. Pressure varies on longer timescales. Low pressure is associated with storms (high winds and wind gusts).

Help on finding past weather data is available at https://www.weather.gov/help-past-weather. Following the instructions and clicking on Oregon brings up http://www.wrh.noaa.gov/pqr/.

Clicking on “local”, choosing the “Local Data/Records” tab, and clicking on “Daily Records for cities in Oregon” brings me to the website of the Western Regional Climate Center http://www.wrcc.dri.edu/summary/climsmor.html where I can choose a city. Now I choose “Corvallis Oregon State University” and scroll down on the left side menu to “Graph and Lister” under “Daily Data”. Now select the time period (I choose 20120101 to 20160101) and click on the “Daily Precipitation” button and then “Create Graph”.

Figure 4: Local weather data (daily minimum and maximum temperature and daily precipitation) from Corvallis, Oregon, from 2012 to 2016. Large variations in temperatures from ~40°F in winter to ~65°F in summer indicate the seasonal (or annual) cycle. Precipitation, which is very episodic, also shows a large annual cycle with lots of rain in winter and less rain in summer. The mean seasonal cycle is shown in Fig. 5. From wrcc.dri.edu

Averaging many years of these weather data yields a climatology. At the WRCC website you can see plots of climatologies over 30 years by scrolling back up the left side menu and clicking on one of the “Daily Temp. & Precip.” links. For Corvallis it shows average temperatures around 40°F (4°C) in winter and 65°F (18°C) in summer. It also shows a larger diurnal cycle in summer than in winter. What could be the reason for that? Answer: less cloud cover in summer. Clouds lead to cooler days and warmer nights, thus decreasing the diurnal cycle. As expected, precipitation is low in summer and high in winter. Curves are smooth in the climatology. In contrast to Fig. 4 the weather variations have been averaged out.

Climatology of temperature and precipitation in Corvallis, Oregon. The difference between the red and blue curve denotes the average diurnal cycle.

Figure 5: Climatology of temperature and precipitation in Corvallis, Oregon. The difference between the red and blue curve denotes the average diurnal cycle.

Now try a forecast of temperature for next August and next January. What is your guess? Answer: a good guess would be the climatology, i.e. 65°F for August and 40°F for January. Thus the average seasonal cycle is climate and it is predictable because it is forced by the seasonal cycle of solar irradiance. However, deviations from the climatology for a certain day next year are unpredictable weather.

Statistical properties of climate can be summarized in a frequency histogram. Let’s explore those with an example. Download from the WRCC website daily average temperature data from a certain day of year, e.g. May 1st. You can do this by clicking on the “Lister” link in the left menu. Next choose start and end days (e.g. 19851001 and 20160101) to create a list of 30 years of data. Enter “wrcc35” as the password. Let’s choose “Average Temperature” as the element to list. Now select a day of the year (e.g. May 1st) for both “Starting Date” and “Ending Date”, and click the “Get Listing” button. This gives the following list (or vector mathematically speaking) of N=30 temperatures T = ( T 1, T 2, … , T N) = (52, 52, 44, 52, 52, 54, 48, 52, 50, 56, 59, 46, 64, 58, 55, 48, 54, 54, 63, 54, 51, 54, 42, 52, 52, 48, 48, 48, 66, 56). Here N = 1 corresponds to 1985, N = 2 to 1986 and so on. In order to construct the frequency histogram we first need to choose bin sizes. I propose to choose 6 bins: 40-44, 45-49, 50-54, 55-59, 60-64, and 65-70. Now simply count how many years fall into each bin. I get 2, 6, 14, 6, 2, 1. Using python I produced the following graph (Fig. 6).

Climate change can be expressed as a change in the mean, which would correspond simply to a shift of the whole histogram to warmer or colder temperatures. However, it can also change the shape of the histogram. For example by making the distribution wider or narrower. This would increase or decrease the occurrences of extreme events. And, of course, it can both change the mean and the width. Most of the time, including in this text, discussions of climate change consider only changes in the mean. But we should keep in mind that changes to the tails of the distribution may be equally important because it is those tails that can have large impacts (heat waves, cold spells, droughts, floods, etc).

Histogram of 30 year (1985-2016) temperatures in Corvallis on May 1st. Most years (14 out of the 30) the temperature is between 50 and 55°F. The mean of the distribution is 52.8°F, its standard deviation of σ =5.4°F represents its width (about 2/3 of all years have temperatures within ±1 σ of the mean). The histogram, although it approximates well a Gaussian (normal) distribution, is slightly asymmetric such that very warm years (65-70) occur slightly more often than very cold extremes (35-40). The tails of the distribution are the upper and lower bins. They represent rare or extreme events. Only one year was warmer than 65°F. This year (2014) can be regarded as an extreme event. It was a record warm year in Oregon.

Figure 6: Histogram of 30 year (1985-2016) temperatures in Corvallis on May 1st. Most years (14 out of the 30) the temperature is between 50 and 55°F. The mean of the distribution is 52.8°F, its standard deviation of σ =5.4°F represents its width (about 2/3 of all years have temperatures within ±1 σ of the mean). The histogram, although it approximates well a Gaussian (normal) distribution, is slightly asymmetric such that very warm years (65-70) occur slightly more often than very cold extremes (35-40). The tails of the distribution are the upper and lower bins. They represent rare or extreme events. Only one year was warmer than 65°F. This year (2014) can be regarded as an extreme event. It was a record warm year in Oregon.

Now let’s explore effects of spatial scales on climate variations. Go to https://www.ncdc.noaa.gov/cag and start with the U.S. tab selection. We want to compare the annual average temperature change in a city of your choice (e.g. Salem in Oregon), averaged over the larger region of a state, for the U.S. as a whole, and the globe (use “Globe” tab and plot “Land and Ocean”, “Land”, and “Ocean” separately). For “Timescale” select “Annual”. Check the “Display Trend” box. This will include a trend line on the graph. Compare the trends and year-to-year variations. Are the trends larger for the global ocean or for the global land?

Compare the year-to-year variations and trends in regionally and globally averaged temperatures (Figs. 7-12) to the local diurnal cycle, day-to-day weather fluctuations (Fig. 3), and to the seasonal cycle (Figs. 4-5). Even though the global temperature trends (~0.1°C/decade) are much smaller than diurnal, day-to-day or seasonal fluctuations, which can be 10°C or more, we will see in the remainder of this course that global temperature changes of a few degrees C will have important impacts on natural systems and human societies.

Figure 7: Annual average temperature changes in Salem, Oregon. Note the large year-to-year variations of 2-3°C and the small overall trend (blue line).

Averaged over the state of Oregon temperature fluctuations from year-to-year are reduced (1-2°C) compared with the figure from Salem and there is a strong (0.1°C/decade) warming trend.

Figure 8: Averaged over the state of Oregon temperature fluctuations from year-to-year are reduced (1-2°C) compared with the figure from Salem and there is a strong (0.1°C/decade) warming trend.

Averaged over the contiguous U.S. the year-to-year variability has slightly decreased compared to Oregon. The warming trend is similar but slightly lower.

Figure 9: Averaged over the contiguous U.S. the year-to-year variability has slightly decreased compared to Oregon. The warming trend is similar but slightly lower.

This figure shows anomalies (differences) with respect to the 1901-2000 average of global averaged land temperatures. Averaged over all of Earth's land areas the year-to-year variations are further reduced (~1°C) and the trend is with 0.1°C/decade slightly larger than that in the U.S.

Figure 10: This figure shows anomalies (differences) with respect to the 1901-2000 average of global averaged land temperatures. Averaged over all of Earth’s land areas the year-to-year variations are further reduced (~1°C) and the trend is with 0.1°C/decade slightly larger than that in the U.S.

Averaged over the global ocean the year-to-year variations are still smaller (~0.5°C) than over land and the trend is also smaller (0.06°C/decade).

Figure 11: Averaged over the global ocean the year-to-year variations are still smaller (~0.5°C) than over land and the trend is also smaller (0.06°C/decade).

Averaged over land and ocean the increase in Earth’s temperature over the past 100 years has been approximately 0.7°C (0.07°C/decade). This result is very robust. Several groups around the world have analyzed the available data in different ways and came to the same conclusion. This is a key figure.

Figure 12: Averaged over land and ocean the increase in Earth’s temperature over the past 100 years has been approximately 0.7°C (0.07°C/decade). This result is very robust. Several groups around the world have analyzed the available data in different ways and came to the same conclusion. This is a key figure.

The weight of the ice depresses the lithosphere, which provides a feedback to the ice by lowering its surface elevation. Traditionally it has been thought that this feedback acts on long (millennial and longer) timescales, but recent research indicates that in certain regions such as West Antarctica, where the lithosphere is thin and the mantle viscosity is low the lithosphere can respond on much shorter (decadal to centennial) timescales. ↵