Interview: Dr. Edward R. Cook, Director, Tree-Ring Laboratory, LDEO

Dr. Edward R. Cook is the Ewing Lamont Research Professor at the Lamont-Doherty Earth Observatory (LDEO) of Columbia University and serves as the Director of the observatory's Tree-Ring Laboratory. 

Dr. Cook co-founded the Tree-Ring Laboratory in 1975 alongside Dr. Gordon C. Jacoby. Dr. Cook is a world-renowned paleoclimatologist specializing in dendrochronology (the study of tree rings) to reconstruct past climate variability and change. 

He led the development of several continental-scale "drought atlases," including the North American Drought Atlas (NADA), the Monsoon Asia Drought Atlas (MADA), and the Old World Drought Atlas (OWDA). In so doing, he showed that “megadroughts” (periods of dryness lasting multiple decades in some cases) were far more common in the past compared to modern times and were a global phenomenon. 

He is also credited with developing new statistical methods for processing tree-ring data and for creating networks of chronologies for climate reconstruction and the development of drought atlases. 

1. Tell us about your journey to becoming the Director of LDEO's Tree Ring Laboratory.

I was extremely fortunate to meet Gordon Jacoby in 1975 when I was finishing up my Master of Science (MS) degree at the University of Arizona Laboratory of Tree-Ring Research (LTRR). Gordon had been invited by famed Lamont geochemist Wallace S. Broecker to set up a tree-ring lab at Lamont and needed help in doing that. Gordon hired me, and we worked in tandem in establishing the Lamont-Doherty Tree-Ring Lab (TRL) to study climatic change and other earth science processes from tree rings. Over many years of effort, Gordon and I succeeded in developing a globally significant TRL. Then, after Gordon retired as the first Director of the LDEO TRL, I assumed that leadership position in the lab.

2. Did you have a role model or an event that inspired you to become a scientist?

I have had many role models over the years, but the biggest event was when my biological sciences professor at Trenton Junior College suggested that I consider taking a course in dendrochronology offered at the University of Arizona, where I was transferring to in 1968 for my Batchelor of Science (BS) degree. I took this course in my senior year in 1970 and realized that there was nothing else I wanted to study. After a hiatus in the army, I returned to the University of Arizona in 1973 with a commitment to study dendrochronology for my Master of Science (MS) degree.

3. Your work is foundational to the field of paleoclimatology, particularly in establishing the long-term context of drought. What got you interested in this field of research?

My field research for my MS at the University of Arizona was conducted in the Hudson Valley of New York in the Shawangunk Mountains. I sampled four long-lived tree species growing on dry (xeric) sites that were perfect to study past drought from tree rings. After I was hired by Lamont in 1975, I used these tree-ring records to publish the first quantitative reconstruction of past drought in the eastern United States extending back to 1730. Since that time, I have focused on reconstructing drought from tree rings around the world.

4. Tell us about your study on Low-frequency signals in long tree-ring chronologies (Science, 2002) that improved how scientists reconstruct past temperature variability over a thousand years.

I met Jan Esper, a German student in dendrochronology, in 2000. He expressed an interest in coming to Lamont on a postdoctoral research fellowship to study tree-ring analysis methods with me. He brought with him a large collection of tree-ring wood density measurements made by Fritz Schweingruber in Switzerland from a range of tree-ring sites around the Northern Hemisphere (NH) with very strong summer temperature signals in them. From these data, Jan decided to produce a reconstruction of NH summer temperatures extending back over 1,000 years for comparison to the ‘Hockey Stick’ estimate of past temperatures published by Mike Mann and others in 1999.

Jan used a different method of processing the tree-ring density data to preserve as much multi-centennial variability as possible in the tree-ring density chronologies used for reconstruction. In so doing, he showed the occurrence of Medieval Period warmth, followed by pronounced Little Ice Age cooling, and then 20^th century warming, which was not well expressed in the Hockey Stick. This result created additional controversy about the Hockey Stick when it came out, but the story described in the 2002 Science paper has been largely confirmed by other reconstructions of past NH temperatures since that time.

5. Tell us about your study the smoothing spline: a new approach to standardizing forest interior tree-ring width series (Tree-Ring Bulletin, 1981), which became a cornerstone for processing tree-ring data.

I have always been a ‘methods’ person in dendrochronology, and one of the fundamental problems was how to extract climate information from trees growing in closed-canopy forests subjected to large amounts of non-climatic competition and disturbance. The methods of detrending tree-ring series to remove non-climatic biological growth trends in open-canopy tree-ring series in the western United States did not apply to the eastern forests where I was working. I therefore investigated alternative methods of detrending that were more naturally adaptive to the changing growth trends found in closed-canopy forest tree-ring series.

In so doing, I came upon a paper describing the smoothing spline, a curve fitting method that was very adaptive, but hard to fit in an easily controlled way. At the time, the Lamont TRL had working for it an extremely capable applied mathematician named Ken Peters, who had previously worked on the analysis of lunar heat flow measurements from probes installed by Apollo astronauts into the lunar surface. Ken looked at the mathematics of the smoothing spline and realized that it could be re-expressed in the form of a digital low-pass filter with theoretically defined filtering characteristics. 

This opened up the smoothing spline for direct use in adaptively filtering and removing non-climate variability in closed-canopy forest tree-ring series. It is still extensively used for this purpose and also as a general time series low-pass filtering method.

So while I recognized the potential for using the smoothing spline for detrending closed-canopy forest tree-ring series, it was Ken Peters who made it a practical method to apply.

6. Tell us about your study Tree-ring standardization and growth-trend estimation (The Holocene, 1995), which is a primary reference for "standardizing" tree rings to remove non-climatic biological noise.

This paper investigated the so-called ‘segment length curse’ in tree-ring standardization and its effect on how much low-frequency variability could be preserved in a tree-ring series after it had been detrended. Based on a paper by the famous economist C.W.J. Granger, my 1995 paper showed how it is impossible to preserve timescales of variability longer than the lengths of tree-ring series being detrended.

This problem had been recognized earlier by my colleague Keith Briffa in his seminal 1992 paper on preserving climate variability at short and long timescales, but was described in more general terms in my 1995 paper. 

The ‘segment length curse’ has profound implications when one is trying to reconstruct climate variability at all time scales from detrended tree rings. It was the leading motivation for how the tree-ring data were processed by Jan Esper in his 2002 Science paper, and was based on Keith Briffa’s method to recover multi-centennial summer temperature variability.

7.  What are the causes, and consequences of the southwestern North American megadroughts and its impact on intense wildfires?

The causes of megadroughts in the Southwest are still not completely understood. We know that the cold phase of the El Niño-Southern Oscillation (ENSO) called La Niña, commonly leads to drought in the Southwest. Sea surface temperature conditions in the Atlantic Ocean have also been implicated. But what has not been adequately explained in my opinion is how the past megadroughts in the Southwest persisted for decades back in medieval times without any need for anthropogenic climate forcing. This suggests that the climate system has a natural ‘megadrought mode’ of variability built in to it that we do not understand very well.

Regarding the impact of megadroughts on wildfires, prolonged dryness and increasing evapotranspiration demand due to warming will lead to increased tree mortality and, therefore, an increase in fuel wood for intensifying wildfires. Add to this the long-term effects of forest fire suppression leading to increases of forest in-growth and the fuel wood loading is even greater. This condition has led to the occurrence of intense stand-replacing fires in the Southwest that may not have happened in the past.

8.  Is climate change increasing drought risk and severity globally?

I am not an expert here, but I think there is evidence of increasing drought globally that is probably related to the effects of global warming. Increasing temperature leads to increasing evapotranspiration demand, which leads to increasing water stress in plants during periods when precipitation is inadequate to offset increasing water loss. The occurrence of “hot droughts” is likely to increase in the future as temperatures increase and place increasing stress on agriculture and water supply systems in the affected regions. But this is not happening everywhere, which complicates the evaluation of climate change increasing drought risk.

9. Tell us about your  "Drought Atlases"—large-scale, gridded reconstructions of past climate. 

As I mentioned previously, in 1977, I published a paper describing past drought from tree rings in the Hudson Valley in New York extending back to 1730. A couple of years later, I extended this reconstruction back to 1694. I thought these results were pretty cool. A couple of years after that, I presented my results at a meeting in Grenoble, France. After my talk, I asked a respected climate dynamicist about how my reconstruction might help him understand the causes of drought variability. He politely replied, “Ed, I am sorry to say it, but I can’t think of a single thing. Climate scientists need spatial patterns to interpret the causes of climate variability.”

This “quote” is not verbatim after so many years, but is reasonably accurate. It seared a hole in my brain, but it was the best advice I ever received over my long career. I returned home to Lamont, dedicating myself to developing a tree-ring chronology network (a spatial array of tree-ring sites) that would enable the development of spatial patterns of drought (a “drought atlas”) across the United States. This was done in collaboration with colleagues at the Arizona LTRR and at a newly established TRL at the University of Arkansas.

At the same time, this push towards the spatial reconstruction of drought from tree rings also led me to develop a new method of spatial climate reconstruction called “Point-by-Point Regression” (PPR) that was specifically designed for reconstructing spatially complex hydroclimate fields. This eventually led to the development of the first drought atlas in 1999, spanning the coterminous United States. It was followed not long after by the NADA in 2004. As global tree-ring networks developed, subsequent drought atlases followed: the MADA in 2010, the OWDA in 2015, the Australia-New Zealand Drought Atlas (ANZDA) in 2015, and the South American Drought Atlas (SADA) in 2020.

The global expansion of drought atlases continues to this day, all based on the PPR reconstruction method. They have propelled our understanding of the causes of hydroclimatic variability in important new ways and have also revealed the impacts of extreme wetness and dryness on past societies and cultures in previously unknown ways.

10. How can people reach you?

https://lamont.columbia.edu/directory/edward-r-cook