1/8/2011

Global Warming
This page shows my early global warming investigations using Excel charts.
Data workbooks and charts made with Excel by D Kelly O'Day; available for download

"Warming of the climate is unequivocal, as is now evident from observations of increases in global average air and ocean temperatures, widespread melting of snow and ice, and rising global mean sea level." Intergovernmental Panel on Climate change (IPCC),  Summary for Policymakers, Feb. 2007

Introduction

Global warming, increasing atmospheric CO₂, ice age, fossil fuels, rising sea levels, El Nino, hurricanes, climate models, Gulf Stream, “Inconvient Truth”, Kyoto … ! Global warming is a very serious, complex and controversial subject.

I’m really curious about these topics. I really want to understand what’s going on with our climate. I’d like to see the numbers, chart the trends, understand the science. I want to separate the facts from the opinions and hype.

 This page will document my journey from a climate science ignoramus to hopefully a better informed citizen who can help solve the climate challenges. I'll point to videos and web sites that I have found useful in my journey.

 This page will use Excel charts and other desktop tools to help assess the climate trends. Each post will include a link to the data sources and where appropriate, sites that provide the scientific/theoretical underpinnings to the demonstrated analysis.

(Top)

Green House Effect and Global Warming

EPA's site provides a good explanation of climate and the greenhouse effect. A summary of EPA's discussion is reproduced below:

"Energy from the sun drives the earth’s weather and climate, and heats the earth’s surface; in turn, the earth radiates energy back into space. Atmospheric greenhouse gases (water vapor, carbon dioxide, and other gases) trap some of the outgoing energy, retaining heat somewhat like the glass panels of a greenhouse.

 Without this natural “greenhouse effect,” temperatures would be much lower than they are now, and life as known today would not be possible. Instead, thanks to greenhouse gases, the earth’s average temperature is a more hospitable 60°F. However, problems may arise when the atmospheric concentration of greenhouse gases increases."

Since the natural greenhouse effect is so important to the climate on earth, we need to be sure that we do not adversely effect it by increasing greenhouse gases and altering our climate in unanticipated ways. EPA states:

"Since the beginning of the industrial revolution, atmospheric concentrations of carbon dioxide have increased nearly 30%, methane concentrations have more than doubled, and nitrous oxide concentrations have risen by about 15%. These increases have enhanced the heat-trapping capability of the earth’s atmosphere. Sulfate aerosols, a common air pollutant, cool the atmosphere by reflecting light back into space; however, sulfates are short-lived in the atmosphere and vary regionally."

(Top)

I have additional temperature trend analysis at my Chart & Graphs blog.

NASA's Goddard Institute of Space Studies (GISS) tracks atmospheric global temperature climate trends as part of its long term climate assessment efforts. Temperature changes by location, day of year and time of day. To provide meaningful year to year comparisons, climatologists calculate the global mean annual land and ocean temperature.

To facilitate assessments of long term trends, climatologists compare the mean for a base period with the annual mean. Differences between the annual mean and baseline mean are called anomalies.  GISS uses the 1951 - 1980 period for their baseline period. They use the difference between the annual mean and the baseline mean to determine the global temperature anomaly for the year.

This Excel based chart, using GISS data available here, shows the annual global temperature anomalies for the period 1880 - 2006. GISS uses the 1951-1980 period mean to establish the baseline.

In the 1880 - 1935 period, the temperature anomaly was consistently negative.  In contrast, the 1980 - 2005 period has had a consistently positive temperature anomaly. The 1917 temperature anomaly (-0.47^oC) was the lowest year on record. Since 1917, global temperature has warmed, with the most recent years showing the highest anomalies of +0.4/ 0.6 ^oC  in the past 120 years.

Download the Excel file here.

 (Top)

Northern Hemisphere Temperature Trend Lines This Excel based chart shows the Northern Hemisphere temperature anomaly - F for the period 1880 - 2006. Trend lines have been added to compare trends over the 126 year period.   Trend Line Change Rate & r2 Period Change  Rate oF/Year r2 1880 - 1921 0.007 0.11 1922 - 1938 0.036 0.44 1939 - 1972 - 0.017 0.38 1972 - 2006 0.054 0.79

There have been at least 4 distinct N Hemisphere temperature anomaly periods in the 1880 - 2006 period. The rate of change has varied from a low to -0.017 ^oF per year in the 1939 - 1972 period to a high of 0.054 ^oF in the 1973 - 2006 period.

The Excel workbook is available here.

(Top)

Open Mind Blog's 4/3/07 post presents an analysis of 5 Swiss temperature datasets for the period 1901 - 2004. The source data was obtained from the European Climate Assessment And Dataset project (ECA).

Open Mind smoothed the daily data sets with a 1-year timescale wavelet transform.  Open Mind's chart is shown on the right.

The interpretation is difficult because of the static nature and size of the display. The charted points are the wavelet transformation  values, not the actual measurements.

Several Open Mind readers had questions about the data set and proper interpretation of the data. ProcessTrends.Com has developed an Excel workbook which provides the source data as well as an interactive chart tool so that users can toggle series on/off and move a cursor to get exact values for all series for a particular year. The ProcessTrends.Com analysis used mean annual temperature trends without smoothing.

This workbook reproduces the Open Mind analysis by:
1. Documenting Swiss Station identification codes, lat/long locations and elevations
2. Calculating mean annual temperature values for the 5 stations
3. Charting 1901 - 2004 mean temperature trends for the 5 stations
4. Calculating 1975 - 2004 temperature increase rates based on mean temperatures
5. Calculating projected 100 year mean temperature increase rates for the 5 stations
6. Comparing projected 100 year mean temperature increase rates with Open Mind's wavelet approach

Users can download the Excel workbook for these 5 Swiss stations at this link.

Change Point Analysis

Several Open Mind blog readers raised questions about the temperature trends, noting potential trend shifts in several series. To address these questions, I performed a Change Point Analysis (CPA) based on techniques described by Dr. Wayne Taylor ( Variation.Com). CPA combines the use of CuSum charts and a bootstrapping technique to compute 1,000 or more iterations of the CuSum chart. 

The 5 Swiss station CuSum charts are shown to the right. Notice that all 5 series had change points in 1988. Geneva had change points 4 years, 1920, 1943, 1962 in addition to 1988.  Lugano had change points in 1919,  1940, and 1988,  Saentis had change points in 1920 and 1988, and Zurich had change points in 1943, 1951 as well as 1988.

While the 5 stations are physically quit distant, and the temperature values also differ, the patterns of the 5 CuSum charts are comparable. For example, the 1920 change points in Saentis and Geneva are close to the 1919 change point for Lugano. Zurich's 1919 - 1920 CuSum chart shows a change in direction, it is simply to small a change to be classified as a change point.

This table shows change point years for each station.

CuSum charts and change point analysis provide comparative information that can be useful in analysis of multiple temperature series.

The Excel workbook, with the CuSum analysis for these 5 Swiss stations, is available at this link.

(Top)

"Carbon dioxide is the most important anthropogenic greenhouse gas. The global atmospheric concentration  of carbon dioxide has increased from a pre-industrial value of about 280 ppm to 379 ppm in 2005. The atmospheric concentration of carbon dioxide in 2005 exceeds by far the natural range over the last 650,0000 years (180 to 300 ppm) as determined from ice cores." (IPCC),  Summary for Policymakers, Feb. 2007

Monthly CO₂ air samples have been collected at Mauna Loa Observatory, Hawaii since 1958 (data link). This CO₂ data and a series of Excel charts are included in this workbook. The CO₂ data can be plotted as a trend chart, as shown to the right. The CO₂ data has increased from 315 parts per million - volume (ppmv) in 1958 to  380 ppmv in 2006. The data also exhibits a significant seasonal effect.

Respiration by land vegetation causes the seasonal cycle. Most of the land vegetation is in the northern hemisphere. During northern-hemisphere summer, plants use atmospheric CO₂ to grow, extracting CO₂ from the air and lowering its atmospheric concentration. As plant matter decays in the winter months, the CO₂ is returned to the atmosphere.

Monthly cycle charts can be used to analyze seasonal cycles. The CO₂ monthly cycle chart for the Mauna Loa observatory is presented in the chart to the right. It shows the 1958 to 2004 data trends for each month (Jan - Dec) as well as the monthly average CO₂ levels. Two key points are clear from this monthly cycle chart:

1. CO₂ levels have increased for every month over the 1958 - 2004 period.
2. CO₂ varies by month, with the maximum CO₂ level in May-June and minimum in September - October.

(Top)

Continuous temperature records based on thermometers only go back about 100 years. Paleoclimatology, the study of past climate, uses several proxy data techniques to estimate climate conditions over geologic time scales. These techniques include:

• Historical records - farmer logs, travel logs, newspaper accounts can be used to construct climate estimates
• Corals calcium carbonate composition - oxygen isotopes and trace metals are used to reconstruct climate during that period of time that the coral lived
• Fossil Pollen -  pollen grains are well preserved in the sediment layers of water bodies, estimates of climate conditions can be based on the types of pollen found in sediment layers
• Tree Rings -  tree-ring widths, density, and isotopic composition can be used to estimate climate conditions
• Ice Cores - ice in glaciers and ice caps over many centuries. Ice cores contain dust, air bubbles, or isotopes of oxygen, that can be used to interpret the climate at the time the ice formed  
• Ocean & Lake Sediments-  ocean and lake sediments include tiny fossils and chemicals that can be used to interpret past climate

Vostok , Antarctica Ice Core

The collaborative ice-core project between Russia, the United States, and France at the Vostok station in Antarctica reached a depth of 3,623 meters, extending back approximately 420,0000 years in time. The chart below was made in Excel using panel chart techniques described here.

 The parallel nature of the methane (CH₄), CO₂ and temperature difference is striking.

From a chart making standpoint, there are several interesting aspects to this paleoclimate chart:

1. Time scale is thousands of years before present (Kyr BP)
2. The vertical panel chart uses a single x axis scale for the three time lines
3. The Y Axis labels alternate to avoid axis scale data label crowding

The Vostok data workbook is available here.

Law Dome Antarctica Ice Core

Etheridge et al reconstructed atmospheric CO₂ and CH₄ levels from thee ice cores taken from Law Dome, East Antarctica  between 1987 and 1993.  The Law Dome CO₂ data extends from 1006 AD until 1978 AD, a 972 year period.

The Law Dome atmospheric CO₂ trends show a relatively stable trend between 1006 and 1750. After 1750, the atmospheric CO₂ level increased, with an accelerating trend in the 20th century.

 (Top)

The Vostok ice core data shows that the atmospheric CO₂ levels ranged from 182 - 299 ppmv in the past 420,000 years before present. The Law Dome Ice Core data set (period 1006 - 1978) shows that atmospheric CO₂ levels ranged from a minimum of 274 in 1604 to a maximum of 335 in 1978.The annual Mauna Loa Observatory, Hawaii data set shows a continual rise in atmospheric CO₂; 1959 levels were 316 ppmv increasing to 381.9 ppmv in 2006.

The figure below consolidation of the Vostok and Law Dome Ice Core data and Mauna Loa Observatory data. 

Several important points are apparent:

• All Vostok CO₂ levels were less than or equal to 299 ppmv over the 420,0000 years before present
• All Law Dome pre industrial revolution (circa 1750) CO₂ data was below 300 ppmv.
• All Law Dome readings after 1912 exceeded 300 ppmv
• All Mauna Loa Observatory readings after 1958 exceed 300 ppmv
• Law Dome and Mauna Loa CO₂ trends show and clear increasing pattern after 1912

This workbook is available here.

What role has human activity played in this increase of atmospheric CO₂?  The following section discusses CO₂ emissions.

 (Top)

CO₂ Emission Trends

The use of fossil fuels, clearing and burning of forests and production of cement release  CO₂ into the atmosphere. The Carbon Dioxide Information Analysis Center (CDIAC) estimates of global CO₂ emissions from 1751 to 2003 (trends) are shown below. Global CO₂ emissions have increased from an estimated 3 million metric tons per year in  1751 to nearly 7,300 million metric tons per year in 2003, a 2,433 fold increase.

Rollover to see breakdown by source

By rolling your cursor over the Global CO₂ Emissions chart you can see the breakdown by source. The source of emissions has evolved over time, with solid (wood, coal) sources representing nearly 100% until the late 1800's when liquid and gas sources became more common. Notice the role of cement production, which converts calcium carbonate to lime, releasing  CO₂.

(Top)

The Excel based dot plot below compares the population, CO₂ emissions and CO₂ emission per capita for the USA, Europe, china, India and the rest of the world (ROW) for the year 2000 (CDIAC) .

The USA, with a population of 280 million, emitted 5.8 billion metric tons in 2000, equivalent to 20.6 metric tons per capita. India,  with a population of 1.0 billion, emitted 1.0 billion metric tons, equivalent to 1.0 metric tons/capita. An average US resident emitted 20.6 times as much CO2 as an Indian resident. European residents, with a standard of living comparable to the US level, emitted 7.7 metric tons per capita, only 37% of the US rate.

The per capita US rate of CO₂ emissions is the highest, 2.6 times the rate for Europe and nearly 8.6 times the rate for China.   Current CO₂ emissions have increased atmospheric CO₂ concentrations and global temperatures. As the lesser developed countries like China and India, with their large populations, expand economically, their CO₂ emissions per capita will increase, adding even more CO₂ to the atmosphere.

(Top)

The United Nations Environment  Programme reports that

"Over the last 100 years, the global sea level has risen by about 10 to 25 cm."

Sea level change is difficult to measure. Relative sea level changes have been derived mainly from tide-gauge data. In the conventional tide-gauge system, the sea level is measured relative to a land-based tide-gauge benchmark....

It is likely that much of the rise in sea level has been related to the concurrent rise in global temperature over the last 100 years."

 ".. the warming and the consequent thermal expansion of the oceans may account for about 2-7 cm of the observed sea level rise"

".. the observed retreat of glaciers and ice caps may account for about 2-5 cm. "

"Other factors are more difficult to quantify. The rate of observed sea level rise suggests that there has been a net positive contribution from the huge ice sheets of Greenland and Antarctica, but observations of the ice sheets do not yet allow meaningful quantitative estimates of their separate contributions. The ice sheets remain a major source of uncertainty in accounting for past changes in sea level because of insufficient data about these ice sheets over the last 100 years. "

The New England Integrated Sciences and Assessment (NEISA) has an Excel workbook of North Atlantic sea level trends from 1856 to 2004. The source data, maintained by the Permanent Service for Mean Sea Level (PSMSL), uses a Revised Local Reference (RLR) datum. The RLR datum for each station is defined to be approximately 7000mm below mean sea level to eliminate negative numbers in the resulting RLR monthly and annual mean values.

The linked workbook includes the source seal level data and a charting tool that allows the user to select the location and generate a trend chart of the seal level data for that location.

(Top)

Background Information

Open Mind's 11/1/07 post presents an analysis of the beginning day of snowmelt for the Red River of the North, using river flow data from Fargo, ND. Open Mind's post is based on data collected by Patrick Neuman, a retired snow hydrologist.

This post and downloadable workbook review Neuman's original data and the Open Mind analysis to confirm the results and provide the data set to others who might wish to analyze trends in snowmelt.

To evaluate trends in the beginning day of snowmelt each year, Neuman analyzed river flow data for three basins in the Midwest, as shown on this Google Earth map.

Neuman estimated the beginning day of snowmelt for each year in the period 1910 - 2003.  The beginning day was calculated in Julian days. Open Mind used this data for the Red River to produce a series of charts that show the trend in beginning day of snowmelt, 10 year moving average for beginning day and a 2 period regression model for the beginning day. Open Mind's charts are shown below. Open Mind Charts: Red River of the North Beginning Day of Snowmelt, 1910 - 2003 Click chart to see full size Trend Chart 1910 - 2003 Trend Chart with 10 Year Moving Average 2 Period Regression: (1919-1964 & 1965-2003) These charts show a significant change in the beginning day of snowmelt in the Red River basin. Open Mind states: " "It appears that up to 1964, snowmelt runoff timing for the Red River at Fargo, ND, had its ups and downs but remained constant over the long haul. Since 1965, snowmelt runoff into the Red River has been happening sooner. In fact, the slope of the line for the recent data indicates that snowmelt is coming earlier by about 0.62 days per year. From 1965 to 2003, its timing is about three and a half weeks earlier. That’s not a trivial difference, it’s quite significant." In looking at Neuman's original material, I was struck by his Figure 1 chart: Neuman's Original Figure 1: 10 Year Moving Average Chart for 3 Basins While Neuman's Figure 1 shows the same dramatic change in the Red River (Fargo) snowmelt beginning as Open Mind's chart,  Neuman's moving average trends for Scanlon and St Cr F were less dramatic, prompting me to compare the trends in the 3 basins rather than just rely on the Red River basin. ProcessTrends Analysis of 3 Basins Noting the differences in Neuman's moving average trends for the 3 basins, I redid Neuman's 10-year moving average chart to make it more readable and did CuSum/ Change Point analysis to look at the differences between the 3 basins in more detail. My 10 year moving average chart, comparable to Newman's,  is shown below: 10 Year Moving Average Beginning of Snowmelt - 3 Basins (ProcessTrends.Com) This 10 year moving average chart, similar to Neuman's chart, clearly shows that the Red River at Fargo exhibited a more rapid downward  trend than St Louis at Scanlon and St Croix at St Croix Falls. While Red River beginning day continued to decline in the 1990s, the Scanlon and St Croix beginning days stabilized. While all three basins showed a decline, the magnitude of the decline was more pronounced in the Red River than in the St Louis and St Croix basins. Open Mind found 1965 to be a change point in the Red River regression analysis, with the trend from 1910 to 1964 nearly random and the trend from 1965 to 2003 significant. I used the 1965 change point year to calculate regression line slopes for the 3 basins. The results are summarized in the table below. Regression Results: 1965 - 2003 Period    Red River @ Fargo St Louis R @ Scanlon St Croix R @ Falls Slope- days/yr -0.62 -0.14 -0.29 r2 0.39 0.01 0.05 The Red River slope for the 1965 - 2003 was -0.62 days per year, with an r2 of 0.39. The slope results for the St Louis and St Croix rivers were considerably less (-0.14 and  -0.29 per year) and the r2 values were very low (0.01 and 0.05). To further evaluate the differences among the basins, I prepared CuSum charts and conducted a change point analysis, as summarized in the graphic below: Comparison of CuSum Charts - 3 Basins (ProcessTrends.Com)

The St Louis and St Croix CuSum charts are similar to each other and quite different than the Red River CuSum chart . Change Point Analysis (CPA) of these stations showed that Red River had two change points (1971 and 1994) while the St Louis and St Croix stations had one change point each in 1983.

The CuSum and change point analysis confirm that the Red River snowmelt beginning date trend is different than the St Louis and St Croix river trends.

Conclusions

My analysis confirms the Neuman and Open Mind analysis results, however, like many analyses, it raises additional questions:

1. Why is the Red River snowmelt beginning day trend decline so much more dramatic than the St Louis and St Croix river trends?
2. Why do St Louis and St Croix river trends in 1995 - 2003 appear to have leveled off while Red River continues to decline?
3. Is there another factor at play that affects the St Louis and St Croix basins differently than the Red River basin?

(Top)

Dr. Wallace Broecker Lecture

Dr. Wallace S Broecker is a leading figure in climatology and the global warming discussion. I found this lecture at Columbia University well worth my time to listen and learn.

http://columbiauniversity.magnify.net/item/Q3WKX568D5H67542

Broecker’s lecture opened my eyes on several fronts:

• Automobiles add about 1 pound of CO₂ per mile traveled.
• Population adds about 3 tons CO₂ per person per year
• 50% of CO₂ absorbed by ocean and biosphere, leaving 50% to build up in atmosphere
• CO₂ rising at 2 parts per million (ppm) per year
• Pre Industrial Revolution CO₂ was 280 ppm , now 380 ppm
• Continued Business As Usual - CO₂ will increase to 850 ppm
• World produces 25 cubic kilometers of CO₂ per year
• Ice core analysis shows both temperature and CO₂ profiles over past 650,000 years.
• There have been periods of abrupt change in both CO₂ and temperature.
• Not sure why there were such abrupt shift
• Sea level affected by glaciers. 4-5 meter rise if Greenland glaciers melt
• Climate and history are linked. During the Medieval Warming period Eric the Red and Vikings settled in Greenland.  The population grew to about 5000 people. By 1320, the settlement was wiped out.
• The Little Ice Age started in the 1300’s.
• Great Ocean Conveyor - Dr. Broecker suspects that historical abrupt changes in temperature and CO₂ caused by shutdown of great ocean conveyor belt.

I found this lecture video a great place to start.

(Top)

(Top)

(Top)