Notes from Dr. Thor Karlstrom U. Mass Lecture Lecture Notes

Dr. Thor N. V. Karlstrom: “Natural Climatic Cycles,” June 5, 2007, University of Massachusetts Lecture

Summary by Dr. Eric Karlstrom
To view a streaming video of this lecture, click here:
Thor Karlstrom Umass

“Natural Climatic Cycles from Seasonal to Ice Age Scales and Their Affects on Linear Predictions of Future Climate Change“

Introduction and Short Biography of Dr. Thor Karlstrom by Professor Mike Williams

Thor Karlstrom did his undergraduate degree in geology at Augustana College in Rock Island, Illinois. He received his Ph.D. in geology from the University of Chicago in 1952, with his dissertation in structural geology. Karlstrom was head of the geology department for three years at Upsala University, in East Orange, New Jersey (1946-1949). He was a research geologist with the U.S. Geological Survey between 1949 and 1983; For the first fifteen of these years (1949-1964), he investigated the glacial history of Alaska. This research led to the first surficial geological map of Alaska as well as many published papers. In 1965, he moved to the Astro-geology branch of the U.S. Geological Survey in Flagstaff, Arizona. There, he mapped the moon, planned geologic missions for Apollo astronauts, and was project head of mission planning for the first three Apollo moon missions. Between 1972 and 1983, Karlstrom’s research focused on higher-frequency climate cycles and Quaternary issues on the Colorado Plateau. Since his retirement in 1983, he has continued to analyze paleoclimatic data in order to reconstruct past climates.

This presentation is based on a poster session Dr. T. Karlstrom prepared for the Paclim (Pacific Climate) Workshop in Asilomar, California.

T. Karlstrom Lecture: “Natural Climatic Cycles”

This presentation features a series of figures I have put together by analyzing the international literature on paleoclimates since my retirement from the U.S. Geological Survey in 1983.

If we reconstruct climatic records on the basis of proxy data where there is a high degree of accuracy in dating, we see that a whole series of different climate frequencies have operated during the last past 2 to 3 million years. We can examine a whole series of time series that go back into that go back into the Ice Ages (last 2 to 3 million years). Time series are graphs with time plotted on the X axis and the amplitude of climatic inferences that can be derived from the geologic data plotted on the Y axis.

1) Figure 1: “Southwest Droughts and the 139-Year Event Cycle” (times scale: last 2,000+ years, 8 time series)

This is a very important graph that shows one of the very high frequency cycles that is replicated in nine paleoclimatic records from California to New Mexico and including Arizona and Nevada. The time series’ are based on the most precise of dating procedures, tree rings. With proper analyses we can get resolution down to the year. Here we look at tree-ring records in the Southwest and see that all these records indicate a cycle of about 139 years. (This is a very short time span from a geologic time period). This 139-year cycle is a robust, rigorous cycle because it is repeated over a large geographic region, i.e., most of the Southwest. These paleoclimatic records extend back to before the beginning of the Christian Era (in New Mexico) and all show the 139-year event cycle.

These time series diagrams show amplitude variations that reflects cyclic changes in both temperature and moisture. (The records taken near upper tree line reflect the influence of temperature changes whereas those from lower tree line reflect the influence of moisture variations.) There are a few places where there are gaps in the cycles either due to noise in the records or resonance interference. (The half cycle is alternately in phase and out of phase with the host cycle. If it is a three to one resonance or third resonance it comes into phase every third time).

Note: For the purposes of correlation, I have converted the original amplitude data of each paleoclimatic record to Z scores. This procedure normalizes the record to pluses or minuses and makes it possible to compare different kinds of records. It will be possible to refer to the original data to specify the actual temperature and precipitation variations.

2) Figure 2: “German Tree Ring Record Indicating European Warm Periods Contemporaneous with Major Southwest Droughts”

This German tree ring record is, in effect, a thermograph, where the broad tree rings coincide with warm periods and narrow rings coincide with cooler periods. This graph shows that the warm periods of northern Germany are in phase with major drought cycles in the Southwest. This is consistent with my model that climate oscillates between relatively warm and dry vs. relatively cool and moist. And the correlation of these sensitive records in widely separated regions suggests that the controlling factor is atmospheric circulation which in turn is dependent on pressure changes (from low to high pressure). Both the American Southwest and Germany are affected by the same mid-latitude storm tracks. (So in effect, it appears that the Aleutian Low and the Icelandic or Northern Atlantic Low are in phase, that is they change simultaneously).

3) Figure 3: “Southwest Droughts, Failing Water Supplies, and Prehistoric Cultural Responses” (last 2,235 years, 2 time series)

This is a rather complicated graph. The upper time series (approximately 1,100 years) is the tree-ring record from the Colorado Plateaus, showing smoothing by 50-year intervals and by my half cycle (half of 139-year cycle or 69.5 years). The 139-year cycle shows up well in this time series. Point boundaries are the time intervals where all the dates cluster. The lower time series is based on my Southwest alluvial chronology extending back at least 2,000 years based on the tree-ring, radio-carbon, and archaeological dating of many hundreds of point boundaries (buried archaeological sites, trees, soils, unconformities, etc.). Here, we see that the dating of the point boundaries and buried soils, etc., in the Southwest tie in beautifully with the tree-ring record.

Interestingly, these two upper time series also coincide extremely closely with dated cultural changes, as represented by site elevations of Anasazi dry farmer dwellings as well. Euler et al (1979) and Berry et al (1982) have plotted the number of dated archaeological sites at various elevations on the Colorado Plateau at 10-year intervals. These data indicate that during relatively dry periods, Puebloan farmers moved to higher, wetter and cooler sites where there was more available water. And during intermediate, relatively wet periods, they farmed in intermediate elevations where water was increasingly available in the arroyos and climate conditions permitted farming.

So here we have a whole series of periods of natural warming and natural cooling that affected prehistoric peoples in a specific way.

4) Figure 4: “Iceland Temperature Record of Warmer Periods Contemporaneous with Major Southwest Droughts” (AD 877 to 1989)

This is a temperature graph inferred from sea-ice conditions, that is, it is based on ice conditions around the coast of Iceland. Here we see that the dates clustering at point boundaries (breaks in alluvial stratigraphy) in the American Southwest coincide with warmer periods in Iceland.

5) Figure 5: “Correlation of Midwestern Drought Record with Average U.S. Runoff and Carbon Dioxide on Timescale of the 25/1 Resonance Subphase Cycle”

1/25 of the 278 year subphase cycle is 11.12 years. This data is from a meteorologist and tree-ring man (Mitchell and Jacobi, I believe). We come out with a 22 year cycle. There is a change of phase in one of the wetter periods.

In 1911, O. Petterson showed that the movements of the earth, moon and sun cause gravitational changes between these bodies that in turn, change the tidal forces on the earth. He implied that these tidal changes affect the ocean currents which, in turn, affect climate. He calculated from celestial mechanics that the last major tidal force interval was 1433 AD.

So all the timing in all these graphs are referenced to 1433 AD. There is always a 1433 AD in the graphs. All records converge on the same period of time.

6) Figure 6: U.S. Stream Runoff as Decadal and Single-Year Cycles

This is a record of U.S. Geological Survey hydrologic information from California to New England, going back to 1911. Langbein and Slack selected records that showed the effects of natural rather than human-altered stream flows. This record shows that what’s happening on the east coast of the U.S. is also going on west coast. The record shows a surprising in-phase relationship all the way from the west coast to the eastern seaboard. That is, the resonance curve is in phase all the way across the country in a transect which is a storm track moving northeastward from California to New England.

7) Figure 7: “Sunspots, Climatic Records and Decadal Cycles”

OK. We’ve been going through a whole series of oscillations between cold and warm, wet and dry. Now I’m showing these changes in relation to the length of the sunspot cycle as shown by Lassen and Christenson (1991). They found a better than 90% correlation between sunspot cycle length and the northern hemisphere average temperatures. This is a very high correlation. Showed a correlation between sunspot cycle length and of over 90%.

What I’ve done is to introduce 1) the Iceland curve, 2) the Santa Barbara marine record, and 3) a tree-ring record that shows the stable isotope temperatures. These time series show the 90-year Gleissburg sunspot cycle that relate to the tidal force resonance.

1711 is one of my important boundaries. 1/3 of that is 1757, another third resonance is 1803, the last is 1942. Look at the number of reversals of the temperature record since the beginning of the industrial revolution! Your projections of future climates from each part of the record would have to vary significantly. So the amount of smoothing you have in these time series is very important in terms of projection.

8) Figure 8: Two “Standard Glacial” Melt-Water Chronologies of the Ice Ages and Climatic, Millenia-Long, Oscillations.

Many, many geologists at a consortium of universities worked for many years on the deep-sea core record of the earth. And they came to the conclusion that the record of climate change recorded in the deep-sea cores correlates with the climate model developed by Milankovitch. Milankovitch was a Serbian mathematician who calculated the amount of caloric changes from solar insolation impacting on the outermost part of the atmosphere.

This climatic model indicated that during summertime, you would get higher temperatures, more melt-water from glaciers. If you put these two records together you can calculate the amount of meltwater over a period of thousands of years based on insolation changes as predicted by Milankovitch.

This record goes back over 300,000 years. And it is considered to be the standard for the Pleistocene. It features the Alaska glacial chronology that I developed back in the 1950’s and 1960’s. the marine record,

The conventional wisdom is that climate changes in the northern and southern hemisphere are in phase. The belief is there was so much more ice in the northern hemisphere that climate changes here would have triggered similar climate changes in the southern hemisphere. To test this, we need to look at terrestrial records on both side of the equator.

9) Figure 9: Ice-Age Climate Oscillations Modulated by Solar Insolation Cycles that are Tens of Thousands of Years Long and of Opposite Signal Across the Equator” (last 150,000 years, 12 time series).

Milankovitch demonstrated, and others have confirmed, that climate in the polar regions is dominated by the approximately 40,000-year obliquity cycle. Another input in the insolation is the precession cycle, which is about half as long or about 20,000 years.

When you go to the equator, the dominant influence is the 20,000-year precession cycle. But when you go across the equator, you have reciprocal relationship. Because you are dealing only with the summer half year.

Here we have a whole series of records in the southern hemisphere, including pollen, marsh and so on, that show no synchronaeity between the northern and hemisphere climate changes. There is about a 10,000-year difference between the northern and southern hemisphere records.

10) Figure 10: “Greenland Ice Core (GIS P2), Solar Insolation, Isotope Temperature, and Greenhouse Gas” (last 128,000 years, three time series)

This is done by the paleoclimatic laboratory people in England. They have accepted the Milankovitch curve. This is GIS P2 ice core from Iceland. After counting the layers, they fine-tuned it according to the Milankovitch curve. The second graph is their O18 curve (a proxy for temperature curve), and methane curve. Methane is a greenhouse gas. It is actually more powerful than carbon dioxide.

This graph is long before man entered scene. So the question here is whether the temperature drove the methane or the methane determined the temperature.

11) Figure 11: “Southern and Northern Hemisphere Glaciations, Glacio-Eustatic Levels, and Modulation By Solar Insolation Cycles” (300,000 years ago to 50,000 years in the future)

Glacio-eustacy is sea-level changes that result from melting and increasing continental ice. So when the continental glaciers grow, water is removed from the ocean and sea levels drop. When these huge glaciers melt, more water is returned to the ocean and sea level rises.

Some of the higher sea level stands during the Pleistocene are up hundreds of feet above present sea level. But part of the elevation is due to tectonic factors affecting the crust of the earth.

During the last glaciation (about 18,000 years ago), they’ve calculated based on the volume of ice involved that sea level dropped about 300 feet worldwide.

The top time series of this graph is a record by Canadians on the Lake Erie lobe of the continental ice cap (Laurentide glacier) that covered North America. You can see that it correlates with the insolation curve (second time series) that is the northern hemisphere component of the insolation at about 45 degrees N. latitude. The third time series from the top is the insolation at the equator (caloric equator). The bottom or fourth time series is insolation in southern hemisphere (about 45 degrees S.) which is reverse of the northern hemisphere.

Clomferton, an English geologist, has done a great amount of work in South America glacial geology. He came to the conclusion after dating organic material associated with marine shells, organic silts, wood, and so on, that there was very little he could do between the northern and southern hemisphere glacial records. I took a look at his data and found with the dates he gave, there was virtually a one-to-one correspondence between his glacial moraine complexes and the southern hemisphere insolation curve.

12. Figure 12. “Egyptian Nubian Pluvial, Solar Insolation, and the Substage Cycle in the Nile River Valley” (time scale: last 62,000 years)

OK. This is interesting. The data is in the Nile River Valley, north of the equator, these are the envelop of insolation, and these are the dates of the pluvials (period in which large lakes develop during or near the end of glacial phases). The Mausterian Pluvial dates about 52,000 to 42,000 years ago and the “Nubian Pluvial” dates about 29,000 to 22,000 years ago. And here you have the pluvials occurring during periods of low insolation and warmer, arid climate during intervals of high insolation.

The interesting thing about it; aggradation (deposition) periods in the Nile River Valley coincide with a 3300+ year cycle which I call the substage cycle. So here we have another example of the temperature of the higher frequency events following the local insolation curve and the substages following the tidal force curve.

13. Figure 13. “Time Frequency of Dated Tunisian Buried Soils Around the ca. 3300-year Substage Cycle”

This is a slide that indicates a way to deal with dated stratigraphy in sand dunes in Tunisia. Here, I believe, there are 81 dates in sand dunes. So the dunes represent drier phases and the soils represent relatively most periods when lakes appeared and/or the dunes were vegetated. Here, the radiocarbon dates cluster at the same time intervals.

(Nearly every time series curve that you work with is like this one in the sense that there are more radiocarbon dates as you get closer to the present time).

14. Figure 14. “Canadian Vegetation Changes and the about 2200-year Substage Cycle”.

This data is a record of pollen changes over time from Antifreeze Pond in the Yukon Territory by Rampton, of New York University. You notice from the vertical lines I’ve drawn here that there is a clear correlation between the pollen record and my substage cycles. The whole record indicates a change from tundra conditions at the beginning of the record to trees in the latter part of the record. The time series indicates a change from tundra conditions to tree conditions. And where the sampling interval (for pollen) is closer together, I get higher frequency cycles (the 1100-year cycle) showing up.

15. Figure 15. “European Post-Glacial and It’s Warmest Atlantic Period Concurrent with the North American Postglacial “Altithermal” (last 15,000 years, 4 time series)

OK, I think you recognize that the vertical lines here are the same as the previous diagram- with the scale changes depending how long the record is.

This top time series is a record from my Alaska work (Karlstrom, 1961), where I first detected the 550-year cycle. It is from the Bolling and Alerod from Denmark. The third time series is from a bog in Switzerland. Al these are radiocarbon dated.

This one is a record of timberline changes. During glacial times, tree line drops because of the colder temperatures. (Trees migrate down slope to ecological zones which have the appropriate temperature).

The correlation and internal consistency between these curves is surprising. You might ask me how I get dates from a bog when it is only dated by three or four horizons. This uses, as a first approximation, assuming uniform deposition rates throughout that interval.

16. Figure 16. “Equatorial Pacific Ocean Core, Isotope Records and the ca. 1100-year Stadial Cycle.”(last 19,000 years, 5 time series)

This is a cycle which is 1/3 the substage cycle, which is 3300+ years. This is done on the basis of the isotopes, indicating the temperature of the surface water. The other is the highly smoothed melt-water cycle. When the curve is down there is more meltwater than when the curve is up. So you are going from glacial at the beginning of the graph to post-glacial time. And what I’ve done here is to amplify the difference between points by multiplying them. This is a mathematical derivative of the smoothed curve which, in effect, re-constitutes the original variation in the curve.

17. Figure 17. “Egyptian Dynastic History and the 139-year Event Cycle” (3571 BC to AD 43)

(chuckles). This is pretty speculative. This is the dating of the Egyptian dynastic periods by historians. And these boundaries are the boundaries between the dynasties. Here’s my 139-year cycle in the lower part of the graph. And I found that by plotting the two curves that the majority of the dynastic boundaries fall in the dry epicycles, suggesting that low Nile River levels was a contributing factor in the political stress of the time. Therefore, you change from one Pharoah to the next.

18. Figure 18. “Chinese Historic Floods and Frosts with Century- and Decadal-Long Cycles.”

The red line is a record of the number of floods from China. This blue line is the number of frosts. The frosts have a much smaller sampling intervals, it represents the number of frosts in a decade, whereas the flood record indicates the number of floods per century.

I think you can see again the similarities of the curves with the reconstruction.


My work has shown that past climate was strongly oscillatory in character long before humans entered the scene. Past climates were characterized by repeated periods of warming and cooling at many superposed frequencies ranging from seasonal to 10’s of thousands of years (ice ages). This reality must be considered when predicting future climates by extrapolating from very short segments of the climatic record. Current predictions of dire climatic consequences resulting solely from man-induced global warming since the industrial revolution (AD 1850 to present) are therefore insufficient and more rigorous predictions must attempt to quantify and separate the natural from anthropogenetic inputs.

Question and Answer Period

The warm climax in the insolation was about 10,000 years ago. In terms of the glacial record, you have a lag between the insolation and the glacial response. So the “Altithermal,” a term introduced by Ernst Antevs in the 1950’s, is the warmest period in the post-glacial time around 5,000 to 6,000 years ago. So on that basis you have a general trend toward cooling. But superimposed on that are a whole series of these other cycles, which sometimes interfere with each other and sometimes amplify each other. So you have a great variety up and down, even during the broad warm periods, based on the Milankovitch theory.

So we are somewhere along the line where we are coming out of the warmest period and we might be in a secondary warm period, in a trough, which is warm period. But eventually, we will be moving into a cooling period. If you accept the Milankovitch theory.

Q: What do you measure the (inaudible) cycle?

A: It’s a subharmonic of the tidal force cycle by Peterson. So right now, I’m thinking of it as a resonance, part of a whole series of cycles that affect directly the atmosphere as well as the ocean.

Q: It has nothing to with the 11-year solar cycle? Or that’s another story?

A: Well, if I take the tidal force and make it the 11.12 cycle, the real terrestrial cycle, it shows up beautifully in the sunspots. The question here is whether the tidal force by the planetary movements is the dominating factor or whether the sunspots themselves are a dominating factor.

Q: Not many people are saying we are going into a cooling period. On the relative influence of natural and human-induced climate change, how do you figure how much is which?

A: I really don’t know. It has to be quantified. And I have yet to come out with any figure that demonstrates the relative influence of natural and human-induced variation, the certain percent natural and a certain percent man-caused. To make any calculations about future climate, this is what needs to be done. If you notice, the carbon dioxide curve in my series, parallels one of the down-ward limbs of one of the natural cycles. The question is whether when the natural cycle starts going up, whether the carbon dioxide starts going up too.

Summary by Dr. Eric T. Karlstrom (May 9, 2009)