1.6k+

# Ancient Observers Riding Decadal Hydrologic Cycles

Authored by: Vikram M. Mehta

Print publication date:  April  2017
Online publication date:  March  2017

Print ISBN: 9781466554528
eBook ISBN: 9781315374482

10.1201/9781315374482-4

#### Abstract

The Great Plains region of North America is mostly semi-arid (Rosenberg 1987). East of 100°W longitude, the climate of land in the “Prairie Provinces” is sub-humid to humid. It is interesting to note that Zebulon Pike and Stephen Long, explorers of the early nineteenth century CE, passed through the Great Plains region in times of drought and referred to the region in their reports as a “desert” (Wishart 2004; Rosenberg 2007). Periodic droughts—as a part of decadal hydrologic cycles (DHCs)—after European settlement of the Plains region led to repeated out-migrations from the region, the last major one being in the 1930s CE, which resulted in 3.5 million people migrating from the Great Plains states between 1930 CE and 1940 CE. In the late nineteenth century CE, Powell recommended that irrigation be developed in this region to buffer the effects of recurrent droughts (Powell 1879), and, in the midst of the “Dirty-thirties” (the 1930s CE dry epoch), a report to President Franklin D. Roosevelt (Cooke et al. 1936) proposed a wide variety of agronomic adaptations and policy adjustments to stabilize the region. All of these explorers and observers were certain that droughts were a permanent feature of the region’s climate. None, of course, knew why they occurred. There are numerous other instances of such dry (and also wet) epochs in other countries and consequent deaths and out-migrations of people and animals from affected regions, as mentioned in Chapter 1 and as we will see in detail in this chapter.

#### 3.1  Introduction

The Great Plains region of North America is mostly semi-arid (Rosenberg 1987). East of 100°W longitude, the climate of land in the “Prairie Provinces” is sub-humid to humid. It is interesting to note that Zebulon Pike and Stephen Long, explorers of the early nineteenth century CE, passed through the Great Plains region in times of drought and referred to the region in their reports as a “desert” (Wishart 2004; Rosenberg 2007). Periodic droughts—as a part of decadal hydrologic cycles (DHCs)—after European settlement of the Plains region led to repeated out-migrations from the region, the last major one being in the 1930s CE, which resulted in 3.5 million people migrating from the Great Plains states between 1930 CE and 1940 CE. In the late nineteenth century CE, Powell recommended that irrigation be developed in this region to buffer the effects of recurrent droughts (Powell 1879), and, in the midst of the “Dirty-thirties” (the 1930s CE dry epoch), a report to President Franklin D. Roosevelt (Cooke et al. 1936) proposed a wide variety of agronomic adaptations and policy adjustments to stabilize the region. All of these explorers and observers were certain that droughts were a permanent feature of the region’s climate. None, of course, knew why they occurred. There are numerous other instances of such dry (and also wet) epochs in other countries and consequent deaths and out-migrations of people and animals from affected regions, as mentioned in Chapter 1 and as we will see in detail in this chapter.

Dry and wet epochs influence and can even have substantial and significant impacts on water resources, agriculture, public health, energy production and consumption, waterways and reservoir navigation, recreation, and other societal sectors. Therefore, it is very important to describe and quantify, if possible, characteristics such as onset time, duration, spatial extent, and severity of the epochs, in order to relate them to causes and effects. These characteristics also provide quantitative goals for prediction of such dry and wet epochs and their societal consequences. In this chapter, we will see characteristics of dry and wet epochs, each lasting 5 to 15 years, and the two epochs constituting a DHC, as found in indirect observations in various parts of the world going back 700 to 1000 years.

The plan of this chapter is as follows: The introduction is followed by descriptions of categories of dry and wet epochs in Section 3.2. Then, the use of trees as climate observers is described in Section 3.3. Sections 3.4, 3.5, and 3.6 describe DHCs in tree-ring-based data, supported by other sources of information, in North America, Europe and North Africa, and monsoon Asia, respectively. Global-scale DHCs are described in Section 3.7. Finally, important results and conclusions are summarized in Section 3.8.

#### 3.2  Categories of Dryness and Wetness

Typical water-related variables used to characterize DHCs are precipitation, river flow, soil moisture, groundwater, and various types of dryness/wetness indices. Since unusual dryness or a drought is considered a more debilitating condition than unusual wetness or a pluvial, droughts have attracted much more attention of researchers, stakeholders, and the general public than pluvials. Several types of droughts have been defined. They are as follows:

• Meteorological drought: Substantial decrease in precipitation from a long-term average.
• Hydrological drought: Substantial decrease in available surface and sub-surface water.
• Agricultural drought: Substantial decrease in soil moisture.
• Socioeconomic drought: Societal impacts of the above three types of droughts.

Figure 3.1 shows a conceptual diagram of these four types of droughts, their origin, and their impacts. As shown, climate variability, especially below-average precipitation and/or above-average temperature, leads to meteorological drought, which, in turn, leads to agricultural drought in the form of reduced soil water, vegetation stress, and reduced crop yields. These reductions, combined with hydrological drought in the form of reduced river flow and reduced water levels in reservoirs and wetlands, lead ultimately to socioeconomic drought via economic, social, and ecological impacts. The other part of a climate variability “cycle,” namely above-average precipitation and/or below-average temperature, can lead to meteorological, hydrological, agricultural, and socioeconomic pluvials. Pluvials, depending on intensity, areal extent, and timing, can cause positive or negative societal impacts.

Figure 3.1   Climate variability and various types of droughts.

Several types of drought indices are used to characterize intensity of dryness (or, wetness), combining precipitation, temperature, and other effects. There are several ways to define a drought index, depending on the purpose for which it is defined. A drought index typically encapsulates data on water-related variables such as precipitation, evaporation, river flow, and others into one number per unit of time. Some examples of drought indices * are percentage of normal, deciles, Standardized Precipitation Index, Palmer Drought Severity Index (PDSI), Crop Moisture Index, Surface Water Supply Index, and Reclamation Drought Index. No drought index is the best for all uses. An American meteorologist, Wayne Palmer, defined the PDSI in 1965 CE; it is widely used around the world for expressing intensity of dryness and wetness. Its calculation by the technique devised by Wells et al. (2004), called the self-calibrating PDSI (SC-PDSI), makes this index more representative of local conditions and makes worldwide comparisons more realistic. The original PDSI is based on a basic water balance model (Palmer 1965), whose basis is the difference between the amount of precipitation required to retain a normal water balance level and the amount of actual precipitation. The other parts of the PDSI calculation account for differences in climate between locations and seasons of the year. These PDSI values are scaled such that they fit Palmer’s 11 dryness/wetness categories, shown in Table 3.1, and allow for spatial and temporal comparisons. Empirical constants for climate characteristics and duration factors used in the calculation of Palmer’s PDSI, which directly affect the spatial comparability of the index, are based on averaging the constants from only a few locations in western Kansas and central Iowa in the United States, representing a small number of climates. The SC-PDSI replaces the empirically derived climate characteristics and duration factors based on Kansas and Iowa data, with values based on historical, local climate data at each location in the world, thereby making the SC-PDSI more representative of a location’s climate (Wells et al. 2004; Dai 2011) while still fitting Palmer’s original dryness/wetness categories. The same dryness/wetness categories in Table 3.1 also apply to SC-PDSI values.

### Table 3.1   Categories of Wet and Dry Epochs by PDSI Value

PDSI Value

PDSI Category

Greater than 4

Extremely wet

3.99 to 3.00

Severely wet

2.99 to 2.00

Moderately wet

1.99 to 1.00

Mildly wet

0.99 to 0.50

Incipient wet

0.49 to –0.49

Normal

–0.50 to –0.99

Incipient dry

–1.00 to –1.99

Mildly dry

–2.00 to –2.99

Moderately dry

–3.00 to –3.99

Severely dry

Less than –4.00

Extremely dry

We will now see what tales do long PDSI time series tell of past DHCs in various parts of the world.

#### 3.3  Trees as Climate Observers

Let us digress for a few moments to see how trees are used as climate observers. Annual tree growth appears as rings in the cross-section of a tree. Changes in the thickness of tree rings over time indicate changes in length of, or water availability during, the growing season. That makes certain types of trees climate observers.

A tree ring consists of a layer of lighter color that grows in the spring and a layer of darker color that grows in late summer, as shown in Figure 3.2. * At locations where tree growth is limited by water availability, trees will produce wider rings during wet and cool years than during hot and dry years. Drought or a severe winter can cause narrower rings. If the rings have the same width over the entire tree’s cross-section, it implies that climate was the same year after year. By counting the rings and measuring their widths from the center of the tree trunk outward, the age and health of the tree and the growing season of each year can be determined with high accuracy. In Figure 3.2, light-colored rings show growth during spring/early summer and dark-colored rings show growth during late summer. During the first 69 years of this 110-year-old tree, the tree grew in very crowded conditions and lacked adequate water and light to grow well, resulting in narrow growth rings. At 69 years, tree density (number of trees per unit area) was reduced by cutting smaller, adjacent trees so as to give this particular tree more light and water, and its growth started to improve, until surrounding trees grew tall and wide enough to limit light and water again after 23 years. At 92 years of age, tree density was again reduced, resulting in much wider spacing among residual trees. With adequate water and light, this tree then again started to grow closer to its optimum ability.

Figure 3.2   Annual growth rings of a 110-year-old tree. See text for details.

(Courtesy of Peter Kolb, Montana State University.)

There are several factors that affect tree ring width, such as climatic effects via precipitation, temperature, cloudiness, and amount of CO2; age of a tree; soil conditions; fire; pests; and diseases. Although it is difficult to remove confounding effects and isolate climate-related effects alone, the ring widths can be translated into quantitative information about durations and lengths of each type of hydrologic epoch. The quantitative information can be calibrated in terms of precipitation, temperature, or even PDSI, thereby enabling a time history of such epochs to be formed over the lifetime of a tree. Quantitative tree-ring information has also been translated in terms of climate variability indices such as the Pacific Decadal Oscillation (PDO) described in Chapter 2. Certain species of trees are more sensitive to their environment than others, and only the more sensitive species can be used to assemble the so-called proxy climate records. Also, trees are susceptible to diseases, and a disease affected tree can lose its climate record, just as hand-written records on parchment or paper can be made illegible or can even be destroyed by chemical processes or insects. Moreover, such tree species do not exist everywhere in the world, or social–political situations may not be conducive to assembling long tree-ring records. After all these problems are overcome, techniques must be developed to accurately relate tree rings to time (“dating”), so that a time series of ring widths and dry/wet epochs—the so-called dendrochronology—can be assembled. Braving these and other difficulties, however, many paleoclimatologists have charted a pioneering trail to assemble multicentury or even millennial records of tree-ring-based climate reconstructions and dared to imagine what tales these ancient observers tell of past climate. It is instructive to see evidence of DHCs through the eyes of ancient observers – namely, proxies for precipitation and temperature such as tree rings that are affected by dry and wet epochs - who have been silent witnesses to ancient DHCs in many parts of the world and to compare their tales with societal impacts of these DHCs, as recorded by human observers.

#### 3.4  North America

There are many published studies of tree-ring and derived climate data analyses, mainly about past climate in North America and also about past climates in other continents (see reviews in Cook et al. 2007, 2010, 2015 and references therein). Although these studies have shed considerable light on decadal to multicentury timescale climate variability, they did not use proxy climate data calibrated at a high spatial resolution from tree-ring data by using the same techniques for different regions/continents for the same multicentury to millennial period. Learning from and building on past studies, Cook et al. (2007, 2010, 2015) recently produced reconstructed PDSI in June–July–August (JJA) (the growing season in the Northern Hemisphere) from tree rings in North America from 0 CE to 2005 CE on a 0.5° longitude–0.5° latitude grid, in Monsoon Asia from 1300 CE to 2005 CE on a 2.5° longitude–2.5° latitude grid, and in Europe and northern North Africa from 1013 CE to 2012 CE on a 0.5° longitude–0.5° latitude grid. They showed that instrumental records of JJA precipitation and temperature in the overlap period match the tree-ring-based PDSI reconstructions reasonably well. Therefore, this unique data set consisting of reconstructed PDSI data in four continents was specially analyzed for this book to focus on DHCs and its societal impacts, and results are described in this and the following three sections. Data for North America, Europe and North Africa, and Monsoon Asia were analyzed separately as well as jointly. In each case, the empirical orthogonal function (EOF)–principal component (PC) analysis technique was used to isolate major patterns of PDSI variability. The Fourier spectrum analysis technique was used on the PC time series to see if there were dominant timescales of PDSI variability.

The most reliable and extensive network of tree-ring-based climate reconstructions is in North America, going back well over 1000 years. Calibrations of various lengths of annual tree-ring data in various locations in terms of the PDSI have allowed researchers to quantify periods and severities of specific, long-term droughts with reasonable accuracies (see, e.g., Douglass 1929, 1935; Fritts et al. 1971; Fritts and Shatz 1975; Stockton and Meko 1975; Mitchell et al. 1979; Briffa et al. 1986; Fritts 1991; Meko et al. 1993; Cook et al. 1994, 1999; Woodhouse and Overpeck 1998; Stahle et al. 2000; and Zhang et al. 2004). There are several comprehensive reviews of this research, including the techniques used to calibrate tree-ring data in terms of the PDSI (see, e.g., Cook et al. 2007 and references therein).

For computational and visualization convenience and for verification of major DHCs in the reconstructed PDSI with other records, the new North America data analyses for this book were limited to the 1006–2005 CE period. Results of the EOF–PC analysis of the 1000 years’ PDSI data are shown and discussed here; the prefix N is used to denote EOFs and PCs of the North America data. Figure 3.3a shows the NEOF 1 spatial pattern “explaining” 13.3% of the total PDSI variance in North America. This pattern is centered on the “Four Corners” region in the western United States, where Colorado, New Mexico, Arizona, and Utah states have common borders, and spans almost the entire western United States, with PDSI anomalies of the same sign. The NEOF 2–NPC 2 pair “explains” approximately 8% variance; NEOF 2 is a bipolar pattern between the Missouri–Mississippi River Basins and the New Mexico–Arizona–northern Mexico region. NEOF 3–NPC 3 pair “explains” approximately 6% variance. This NEOF is a biploar spatial pattern between the northwestern United States and the Ohio–Mississippi River Basins. Only the NEOF 1–PC 1 pair is shown here.

Figure 3.3   First empirical orthogonal functions (EOF 1s) of tree-ring-based PDSI data in (a) North America, (b) Europe and North Africa, (c) monsoon Asia, and (d) combined North America and monsoon Asia. Color scale shows the range of EOF coefficients. See text for details.

The corresponding NPC 1 time series from 1006 CE to 2005 CE is shown in Figure 3.4. The sign conventions in Figures 3.3a and 3.4 (and in figures in Sections 3.5, 3.6, and 3.7) are such that a positive multiplicative product of signs of the EOF pattern and the PC time series denotes wet conditions and a negative product denotes dry conditions. These sign conventions are: (1) the product of positive sign (blue shades) of the EOF pattern and positive PC value (blue bar) is positive and denotes wet conditions; (2) the product of positive sign (blue shades) of the EOF pattern and negative PC value (red bar) is negative and denotes dry conditions; (3) the product of negative sign (red shades) of the EOF pattern and positive PC value (blue bar) is negative and denotes dry conditions; and (4) the product of negative sign (red shades) of the EOF pattern and negative PC value (red bar) is positive and denotes wet conditions. Major dry and wet epochs for which descriptive records from independent sources are available are marked with light red and light blue boxes, respectively. These DHC epochs in North America and their societal consequences based on existing records are briefly described in Table 3.2. A few multidecadal wet epochs, revealed by the EOF–PC analysis, are also marked in Figure 3.4 and mentioned in Table 3.2. It is obvious from Figure 3.4 that many DHC and multidecadal dry/wet epochs can be visually identified. Fourier spectrum analysis of the NPC 1 time series shows that there are several prominent and significant hydrologic cycle periods at 5–10 years, 11–12 years, 18–22 years, 25–30 years, 35 years, 55–75 years, and longer than 88 years, indicating the presence of decadal and multidecadal megadroughts and pluvials since 1006 CE. It is also obvious from Figure 3.4 that the period from the late 1600s CE to the end of the eighteenth century CE was a relatively wet period in North America.

Figure 3.4   First principal component (PC 1) time series of tree-ring-based North American PDSI data. PC 1 “explains” 13.3% of the PDSI variance. Light blue and light red boxes mark wet and dry epochs, respectively. See text for details.

### Table 3.2   Epochs of North American Decadal Hydrologic Cycles and Their Societal Consequences

Years (CE)

Epoch

Name

Societal Consequences

1273–1298

Dry

The Puebloan Drought

Rapid declines and migrations of populations of the Anasazi, the Fremont, and the Lovelock cultures in the Western United States and the Cahokian culture in the Mississippi River valley (Axtell et al. 2002; Benson et al. 2007)

1299–1336

Wet

1346–1355

Dry

Mississippian Drought

Decline of complex societies in the Mississippi River valley, partly due to widespread, poor crop harvests and limited food storage facilities (Anderson et al. 1995; Milner 1998)

1378–1388

Wet/dry

Mississippian Drought

Wet in the West and dry in the Mississippi River valley; further decline of societies in the Mississippi River valley

1449–1458

Dry

Mississippian Drought

Severe drought in the Mississippi River valley and Southern Plains; abandonment of Mississippian settlements in eastern Oklahoma and the area around the confluence of the Ohio River and the Mississippi River (Thomas 2000; Cobb and Butler 2002)

1563–1573

Wet/dry

Mississippian Drought

Wet in the West and dry in the Mississippi River valley and the Northeast

1666–1674

Dry

The Puebloan Drought

Disease, deaths, and village abandonment due to famine (Sauer 1980)

1806–1808

Dry

The Great American Desert

Reports of the Great Plains as a desert by the Zebulon Pike expedition (Wishart 2004; Rosenberg 2007)

1810–1820

Wet

Westward migration of people from the eastern United States tripled Missouri’s population in this epoch and demands for statehood for Missouri

1820–1826

Dry

The Great American Desert

Reports of the areas along the Missouri River and the Platte River as a desert by the Stephen Long expedition (Kane et al. 1978)

1856–1867

Dry

The Civil War Drought

Large-scale crop failure, major dust storms (Malin 1946), severe depletion of bison population due to competition for water with native Americans and European immigrants

1867–1871

Wet

The Garden Myth

Cyrus Thomas and other scientists and non-scientists expound that increased cultivation and settlement bring increased rain—“Rain follows the plow” (Reisner 1986)

1872–1884

Dry; wet during 1877–1879

Rocky Mountain locust swarms covered area of the western United States, comparable to the area of mid-Atlantic states and New England (Lockwood 2004; Seager and Herweijer 2011); soon after the wet years, Wilber (1881) extols the western United States’ agricultural potential

1890–1898

Dry

Major dust storms; beginning of large-scale irrigated agriculture with the Reclamation Act of 1902 (Seager and Herweijer 2011)

1905–1930

Wet

The Big Muggy

Most intense pluvial period in 1200 years; longest period of predominantly high flow in the Colorado River in 450 years; decadal doubling of human population in the western United States (Stockton and Jacoby 1976; Woodhouse et al. 2005); large-scale agricultural development

1934–1940

Dry

The Dust Bowl or Dirty Thirties Drought

Blowing topsoil due to drought and poor land management practices; crop production sharply reduced; 3.5 million people migrated out of the drought-affected area; Federal Government financial assistance $1 billion (1930s dollars; Warrick, 1980); large-scale farm bankruptcies and sell-offs; drought and its consequences became subject of books (Of Mice and Men and The Grapes of Wrath by John Steinbeck; Whose Names Are Unknown by Sanora Babb), music (Woody Guthrie), and photographs (Dorothy Lange); this drought became inspiration for the 2014 movie Intersteller 1941–1952 Wet Rapid expansion of agriculture due to this wet epoch and rapid industrialization due to the Second World War, resulting in a rapid economic boom; return to inappropriate farming and grazing practices (http://drought.unl.edu/DroughtBasics/DustBowl.aspx) 1952–1958 Dry Crop production sharply reduced; extreme low river flows, major groundwater depletion (Nace and Pluhowski 1965); dry pastures affected cattle production; thousands of new irrigation wells and hastening of development of center-pivot irrigation systems; drought and consequences became subject of a book (The Rainmaker by N. Richard Nash), later made into a movie 1980–1987 Wet Floods damaged dams, levees, and riverside built environment; at least$1 billion (1980s dollars) loss

1988–1991

Dry, mainly in the Missouri and Mississippi River Basins

Substantially reduced river flows and run-offs into reservoirs, as much as 50% of average in some reservoirs; substantial reduction in crop production; water use restrictions in many cities; shortening of river navigation season and restrictions on draughts and tow lengths of barges; reduction in hydropower and thermal power production; associated heat waves killed 5000–17,000 people in the United States; cost $80 billion–$120 billion (2008 dollars) in damages, perhaps the costliest natural disaster in the history of the United States; enactment of state and federal legislation on water use efficiency and groundwater management (Mehta et al., 2013a)

1992–2001

Wet

Wide-spread flooding and restrictions on river navigation in some years; substantial damage to built environment and infrastructure in some metropolitan areas; increased demands on urban water systems for water purification; substantial crop losses (Mehta et al., 2013a); several billion dollars’ (1990s dollars) damages

2001–2004

Dry

The Aughts Drought

Mandatory water restriction in some major cities; low crop and forage yields; increased sales of cattle; increased investments in center-pivot irrigation systems; increase in no-till farming; decreased populations of fish, waterfowl, and deer; more intense storms overwhelmed storm sewers and water treatment facilities; increased water pollution and restrictions on power plant operations due to low flows in rivers (Mehta et al., 2013a)

The archeological and circumstantial evidence cited in Table 3.2 strongly suggests that several dry epochs dealt “death blows” to native civilizations between the thirteenth and seventeenth centuries. Notable among these are the Puebloan droughts from 1273 CE to 1298 CE and from 1666 CE to 1674 CE and the Mississippian drought (mentioned in Chapter 1) during 1449–1458 CE. Major dry epochs in the nineteenth century, apparent in Figure 3.4 and supported by documentary and other evidence cited in Table 3.2, are the Great American Desert during 1806–1808 CE, mentioned by the Zebulon Pike Expedition, and the 1856–1867 CE Civil War drought.

NPC 3 shows the four dry epochs mentioned as Mississippian droughts in Table 3.2 in the fourteenth, fifteenth, and sixteenth centuries and also other dry/wet epochs. Some of the wet and dry, especially the latter, epochs in the Mississippi River Basin and in the areas around it, mentioned in Table 3.2, are captured by the NEOF 3–NPC 3 combination. Cook et al. (2007) plotted the reconstructed PDSI patterns for these four dry epochs; these patterns are shown in Figure 3.5. In all four dry epochs, the PDSI was –2 to –3 in the Mississippi and Ohio River Basins, indicating moderate droughts. In Figure 3.5, the red-shaded areas show the presence of moderate to severe droughts lasting approximately a decade. Figure 3.5 also shows a mild to moderate wet epoch in western and eastern Canada and a moderate to intense wet epoch in northern and central Mexico in the 1449–1458 CE period. Similarly, during the dry epoch in the Mississippi River Basin and northeast United States in the 1564–1573 CE period, there was a wet epoch in the upper Missouri River Basin. NPC1 (Figure 3.4) also shows the 1344–1353 CE and 1449–1458 CE dry epochs in the Mississippi River Basin.

Figure 3.5   Tree-ring-based reconstructions of the June–July–August Palmer Drought Severity Index (PDSI) in North America during four intense decadal drought epochs in and around the Mississippi River valley. The PDSI scale is shown below. See text for details.

(From Cook et al. Earth-Science Reviews, 81, 93–134, 2007.)

In the twentieth century, several dry epochs/major droughts occurred, for which there are more instrumental data evidence and more detailed documentary evidence, as cited in Table 3.2. For example, the 1934–1940 CE “Dust Bowl” drought in the United States (mentioned in Chapter 1) caused severe agricultural and socioeconomic problems, including out-migrations of millions of people from affected areas, as mentioned in Table 3.2, and motivated fundamental changes in agricultural and soil conservation practices. The 1988–1991 CE dry epoch, which occurred mainly in the Missouri and Mississippi River Basins, currently two major “bread baskets” not only of the United States but also of the world, is very well captured by the NEOF 2–NPC 2 pair. Finally, in the first decade of the twenty-first century CE, the Aughts Drought impacted a wide variety of societal sectors, as mentioned in Table 3.2. Since all these dry epochs/droughts and the ones in the western United States, as shown in Figure 3.4 and mentioned in Table 3.2, lasted for several years to a decade, their effects devastated water resources and agriculture, ultimately resulting in mass migrations and even the end of civilizations settled in these areas.

As Figure 3.4 and Table 3.2 also show, there have been many wet epochs in the United States since the thirteenth century CE. Since early nineteenth century CE, such wet epochs—and consequent water availability and blooming of vegetation—created the “Garden Myth,” cited by Cook et al. (2007), and spurred westward migration, unsustainable farming practices, and unsustainable water rights allocations, instigated in no small measures by land speculators and “boosters” (Wilber 1881; Stegner 1992). Then, alternating wet and dry epochs in the late nineteenth and twentieth centuries CE witnessed human migration surges and development of agriculture and cities, followed by crop failures, water shortages, large-scale economic losses and suffering, and water rights disputes within and among states and between the United States and Mexico.

Thus, tree-ring-based climate reconstructions clearly show the presence of DHCs in North America for at least the last 1000 years. Herweijer and Seager (2008) showed that the spatial patterns of DHCs in the tree-ring-based PDSI data were indistinct from the patterns in the PDSI data calculated from precipitation and temperatures, measured by instruments since the mid-nineteenth century CE, and that the magnitudes of PDSI extremes—implying occurrences of extreme DHC epochs—were essentially comparable to those based on instrument-measured data. Thus, tree-ring-based climate reconstructions not only extend the historical climate record from the instrument era back for over at least 1000 years but also illuminate periods when civilizations rose, flourished, and died. In addition, tree-ring-based climate reconstructions provide a continuity in DHC characteristics and a context for simulations of past DHCs and predictability and prediction studies for future applications with Earth System Models.

These and previous tree-ring-based climate reconstructions also show that the western United States was also prone to multidecadal megadroughts during the last 1000 years (Cook et al. 2007). Among others, Sauer (1980), Anderson et al. (1995), Milner (1998), Gumerman and Dean (2000), Thomas (2000), Acuna-Soto et al. (2002), Axtell et al. (2002), Cobb and Butler (2002), and Pauketat (2004) have documented actual and hypothesized impacts of these multidecadal megadroughts on native civilizations in North America.

#### 3.5  Europe and North Africa

There have been numerous studies of past climate in Europe that use tree-ring data going back many centuries and multicentury precipitation inferences/measurements (see, e.g., Briffa 1999; Büntgen et al. 2011; and references therein). These studies have shed considerable light on past hydrologic events, especially on multidecadal to century timescales, in various regions of Europe.

Using the new JJA-only reconstructed PDSI data from 1013 CE to 2012 CE (Cook et al. 2015), EOFs and PCs were calculated; the prefix E is used to denote EOFs and PCs of the Europe-North Africa data. The EEOF 1–EPC 1 pair “explains” 14.1% of the variance. The EEOF 1 spatial pattern (Figure 3.3b) has same-sign PDSI anomalies from Ireland eastward to isolated regions in western Russia and in southern Scandinavia and opposite-sign anomalies in central Iberian Peninsula, the North Cape region of Scandinavia, and northwest Algeria in North Africa. EPC 1 is shown in Figure 3.6. The sign conventions in Figures 3.3b and 3.6 are explained in Section 3.4. Major dry and wet epochs for which descriptive records exist (only in western and northern Europe) are marked with light red and light blue boxes, respectively. These DHC epochs in Europe and their societal consequences based on existing records are briefly described in Table 3.3. Dry and wet epochs in the Iberian Peninsula and northwest Algeria are described separately.

Figure 3.6   First principal component (PC 1) time series of tree-ring-based European and North African PDSI data. PC 1 “explains” 14.1% of the PDSI variance. Light blue and light red boxes mark wet and dry epochs, respectively. See text for details.

### Table 3.3   Epochs of Decadal Hydrologic Cycles in Europe and Their Societal Consequences

Years (CE)

Epoch

Name

Societal Consequences

1315–1322

Wet

The Great Famine

Very wet in Britain, northern France, Scandinavia, Belgium, Netherlands, Germany, and western Poland; crop failures and food price increases; extreme levels of criminal activity, diseases, cannibalism, and mass migrations from rural areas to cities; 10%–25% of people in many cities died; in one instance of food scarcity, King Edward II of England unable to find bread for himself and entourage (Warner 2009); anecdote about King Louis X of France forced to retreat from invasion of Flanders due to soggy ground; apparent failure of prayers to alleviate the famine undermined the Catholic Church’s authority; confidence in governments undermined (Jordan 1996)

1590s

Dry

Worst famine in centuries due to crop failures; high food prices coupled with high human populations

1618–1648

Dry

Famines in Europe exacerbated by the Thirty Years’ War

1693–1710

Dry

Seven Ill Years (in Scotland)

Scotland: droughts, multiyear crop failures from the 1680s, famines (Mitchison 2002; Cullen 2010); possible influences of eruptions of Hekla (1693) in Iceland and Serua (1693) and Aboina (1694) in Indonesia (Morrison 2011); 5%–15% of Scotland’s population dead (Wormald 2005); large-scale emigration from Scotland to North America, the West Indies, and Ireland (Smout et al. 1994; Cullen 2010); consequent agrarian, trade, and banking reforms, eventually leading to formation of the United Kingdom with England (Mitchison 2002); in France, starvation and diseases due to crop failures, exacerbated by ongoing wars, resulted in 2 million deaths (6% of France’s population) (Ó Gráda and Chevet 2002)

1740s

Dry

Very cold winters, summer droughts, famines; high mortality

1783–1795

Dry

The French Revolution Drought

France: long-running dry conditions since the early 1780s, worsened to droughts in the mid-decade; primitive agricultural technology, 20%–25% of harvest used as seeds (Taine 1958; Sée 1958); following a decade (1770s) of recession and unemployment after France’s entry in the American War of Independence (Neumann 1977; Labrousse 1958); 90% of France’s population poor; Rye or oat bread staple diet, with 55%–88% of income spent on bread (Lefebvre 1954; Sée 1958); high prices of all staple foods, including wine, due to the droughts; droughts and consequent famine a catalyst of “The Great Fear of 1789” and the French Revolution (Lefebvre 1973; Neumann 1977)

1815–1827

Wet/dry

The Tambora Famines

Following very explosive eruption of Mount Tambora volcano in Indonesia in April 1815, “Year Without a Summer” in 1816 and subsequent years of cool and wet or warm and dry climate caused major food crises around the world; severe famines in Britain, Ireland, Germany; riots and looting in many European cities; worst famine in the nineteenth century in Europe (Post 1977; Gore 2000)

1832–1848

Dry

The 1848 Revolutions Drought

Long-running dry conditions; crop failures, especially in 1846, led to hardships for rural and urban workers; food prices soared; demand for manufactured goods decreased; bankruptcies and urban unrest increased; a potato blight caused widespread famines in Ireland and continental Europe; in conjunction with political, economic, and social factors, droughts and consequent famines also responsible for revolutions in over 50 countries (Breuilly 2000)

1975–1977

Dry

The Grovel Drought

Multiyear drought in the United Kingdom, the worst drought and heat in summer 1976, the hottest summer in at least 350 years (Cox 1978); crop failures, devastating heath and forest fires in southern England; significant increase in food prices; extremely low water levels in reservoirs and some rivers; widespread water rationing; government appointed a minister for drought; on drought-parched English cricket grounds in summer 1976, the West Indies team—fired up by a seemingly racist and patronizing public comment by the captain of the England team—utterly dominated all matches between the two teams, beginning the domination of the West Indies in world cricket for the next two decades (Tossell 2007)

1999–2002

Wet

Replenishment of groundwater and reservoirs depleted by the 1995–1998 drought

2008–2012

Wet

2012, the second wettest year on record in the United Kingdom (UK Met. Office data)

It is obvious from Figure 3.6 that many DHC and multidecadal dry/wet epochs can be visually identified. It is also obvious that western and northern European climate since the eleventh century CE, as represented by the reconstructed PDSI, was largely dry until the 1950s. There were, however, a few, brief wet epochs in each century; the Iberian Peninsula, the North Cape, and northwest Algeria were largely wet from the eleventh century CE to the 1950s CE and have been drying since then. During the Medieval Climate Anomaly (Diaz et al. 2011) from approximately 1000 CE to 1200 CE, the climate of western and northern Europe was very dry as seen in Figure 3.6. There was also a multidecadal megadrought in early fourteenth century CE. Europe experienced the Little Ice Age (Matthews and Briffa 2005) from approximately 1550 CE to 1750 CE. This was also a very dry 200-year epoch, except for an embedded, relatively wet epoch from approximately 1720 CE to 1740 CE. Since approximately 1800 CE, European climate has been gradually becoming less dry, with fewer and milder decadal–multidecadal dry epochs, and it has been positively wetter since the 1950s CE as mentioned earlier.

The brief descriptions of dry epochs in Table 3.3 indicate that these epochs were catalysts and contributing causes of social and political events of historic proportions. For example, approximately three DHCs from 1618 CE to 1648 CE—with dry epochs much stronger and more persistent than wet epochs—caused famines in Europe; these famines were a contributing factor as well as a consequence of the Thirty Years’ War in Central Europe, mainly Germany (see, e.g., Wilson 2011). Then, in 1693–1710 CE, the “Seven Ill Years” drought and famine in Scotland caused large-scale deaths and human migrations as well as sweeping reforms, ultimately leading to the formation of the United Kingdom with England. The same droughts and consequent crop failures in France apparently resulted in up to 2 million deaths. In the eighteenth century CE, France appears to have been particularly vulnerable to crop failures caused by droughts and consequent increased food prices, sociopolitical instabilities, and revolutions. There were famines or food scarcities in 1709 CE, 1725 CE, 1749 CE, 1775 CE, 1785 CE, and 1788–1789 CE (Rose 1956, 1959). The droughts partially responsible for these famines/food scarcities are represented by the EOF 1–PC 1 pair (Figure 3.6). As described in Table 3.3, the last famine in 1789 CE, partially caused by a long-persistent drought in the 1780s CE, was one of the contributing factors of the French Revolution beginning in 1789 CE as mentioned in Chapter 1. The Mount Tambora eruption in Indonesia in April 1815, the subsequent “Year Without a Summer,” and alternating wet–cool and dry–warm epochs till 1827 CE devastated agriculture in many parts of the world. There were severe famines in Britain, Ireland, and Germany; looting and rioting in many European cities; and possibly the worst famine in the nineteenth-century Europe. It is tempting to speculate that the heavy rain and resulting sodden ground on June 17–18, 1815, which were a contributing factor in Napoleon Bonaparte’s defeat in the Battle of Waterloo on June 18, 1815 (Jomini 1864), were a consequence of the Mount Tambora eruption in April 1815! The last dry epoch in the nineteenth century, but certainly not the least intense in Table 3.3, was from 1832 CE to 1848 CE, culminating in the 1848 CE revolution in France. This drought and consequent famine may have contributed not only to a potato famine in Ireland and continental Europe, as mentioned in Table 3.3, but also to revolutions in some or all of the over 50 countries where revolutions broke out in 1848 CE. Finally, the “Grovel Drought” in the United Kingdom in 1975–1977 CE resulted in crop failures and wildfires, widespread water rationing, and even the appointment of a government minister for drought! This drought is labeled “The Grovel Drought” here, because the captain of the English cricket team, a white cricketer born in apartheid-era South Africa, publicly derided the visiting West Indies cricket team in the summer of 1976 CE and said, “I intend, … , to make them grovel” (Tossell 2007). The word “grovel” had connotations of slavery ancestry to many members of the black West Indies cricket team and acted as a spur that resulted in superlative performances on drought-parched English cricket grounds, starting the West Indies team’s domination of world cricket for the next two decades.

It was mentioned earlier that climate of northern and western Europe was mainly dry from the eleventh century to the 1950s CE. Figure 3.6 and Table 3.3 show an unusual occurrence of famines and consequent societal turmoil caused by a wet epoch lasting 7 years from 1315 CE to 1322 CE. This wet epoch was partially responsible for many millions of deaths; migrations of many more millions; and social, economic, and political upheavals. In more recent times, there have been wet epochs in late 1990s to early 2000s CE and from 2008 CE to 2012 CE.

In the Iberian Peninsula and northwest Algeria, the reconstructed PDSI EEOF1–EPC1 combination (see explanation of sign convention earlier in this section) indicates generally wet conditions till the second decade of the twentieth century CE, when long-term drying began. As seen in Figure 3.6, during the generally wet period from the eleventh century CE, there were several dry epochs in the 1090s CE, from 1140 CE to 1160 CE, from the 1240s CE to the 1270s CE, from 1505 CE to 1530 CE, in the early to mid 1560s CE, and from 1712 CE to 1740 CE. During the twentieth century CE, wet epochs were superimposed on the long-term drying in the 1910s to early 1920s CE, late 1920s CE, mid 1930s CE, early 1940s and late 1940s CE, early 1960s CE, and mid 1970s CE. Vicente-Serrano (2006) has calculated the Standardized Precipitation Index (SPI) from 1910 CE to 2000 CE at 51 precipitation-measuring stations in the Iberian Peninsula. The SPI averaged over the entire Peninsula shows the dry and wet epochs seen in the reconstructed PDSI data.

Using these PDSI data and documentary records, Cook et al. (2015) reconstructed several major hydrologic events that occurred in Europe since the fourteenth century. As Cook et al. (2015) noted, these maps represent the drought of 1921 CE from Ireland to central Europe; the drought of 1893 CE from Britain to southern Scandinavia and southern and eastern Europe, including modern-day Turkey; the drought of 1740s CE from Ireland to southern Scandinavia and central Europe, which could have contributed to the well-known famine in Ireland; the central and eastern Europe droughts of 1616 CE and 1540 CE; and the excessive wetness in 1315 CE, which is described in Table 3.3. A close examination of these seemingly 1-year events shows that while each of these hydrologic events may be at an extreme in the year identified with it, that year was a part of a multiyear event or a DHC.

Thus, as the foregoing shows, tree-ring-based PDSI data over 1000 years show the distinct presence of multiyear to decadal and longer dry and wet epochs in Europe and North Africa. Independent evidence strongly suggests that some of these epochs made civilization-scale impacts on European societies and were contributing factors for the sociopolitical revolutions that changed the course of not only European history but also world history.

#### 3.6  Monsoon Asia

Cook et al. (2010) also calibrated June–July–August tree-ring data from over 300 sites in the Northern Hemisphere summer monsoon regions comprising the Indian, East Asian, and Southeast Asian–Australian monsoons in terms of the PDSI. They also included Asian land areas to the north of the greater Asian monsoon regions in their reconstruction.

Using the JJA-only reconstructed PDSI data from 1300 CE to 2005 CE at 2.5° longitude–2.5° latitude resolution (Cook et al. 2010), EOFs and PCs were calculated; the prefix A is used to denote EOFs and PCs of the Asia data. The AEOF 1–APC 1 pair “explains” 14.8% of the total variance. The AEOF 1 spatial pattern is shown in Figure 3.3c. This pattern has anomalies of one sign in India and parts of Afghanistan, the Tibetan Plateau, Myanmar, and southeast Asia. The regions from central Afghanistan in the south to northern Kazakhstan in the north and from western China and western Mongolia in the east to Uzbekistan and Turkmenistan in the west have anomalies of the opposite sign. APC 1 is shown in Figure 3.7. The sign conventions in Figures 3.3c and 3.7 are explained in Section 3.4. Major dry and wet epochs for which descriptive records exist (only in India) are marked with light blue and light red boxes.

Figure 3.7   First principal component (PC 1) time series of tree-ring-based Monsoon Asia PDSI data. PC 1 “explains” 14.8% of the PDSI variance. Light red boxes mark dry epochs. See text for details.

These DHC epochs in India and their societal consequences based on existing records are briefly described in Table 3.4. It is obvious from Figure 3.7 that many DHC and multidecadal dry/wet epochs can be visually identified, without any low-pass filtering or spectrum analysis. The brief descriptions of dry and wet epochs in Table 3.4 indicate that these epochs were catalysts and partial or entire causes of major famines that killed millions of people and left an indelible impression on history. The Damaji Pant Famine during the dry epoch in the 1460s CE is remembered largely for the good deeds of Damaji Pant in distributing grain to famine-stricken people. At least 2 million people are believed to have died in the 1630s CE dry epoch and famine in eastern India. As DHCs continued and human and animal populations increased, subsequent dry epochs and famines took larger and larger tolls in terms of human deaths and out-migrations. These natural DHCs, coupled with indifferent or self-serving rulers and a lack of preparedness, resulted in a series of multimillion human deaths famines throughout the eighteenth, nineteenth, and early twentieth centuries, as seen in Figure 3.7 and Table 3.4. While the droughts were the major causes of these famines, an almost-deliberate disregard of the British East India Company and the British and colonial governments for the native people’s well-being and mercantilist policies—such as forced cultivation of opium and indigo for export, rather than of grains for the local population, and grain exports while the local population starved—were also substantially responsible for these famines, as briefly described in Table 3.4. Native merchants’ profiteering and grain hoarding also played a substantial role in causing these famines. During this almost 250-year period, 50–100 million people may have died due to starvation and malnutrition, and many millions and their cattle migrated out of affected regions. As mentioned in Table 3.4, however, the famines were practically everywhere in India, so the out-migrations from one region to other regions often exacerbated the situation where the migrants went. Beginning in the 1860s CE, the British and colonial governments ruling India began to plan and prepare for the next—inevitable—famine, including the appointment of Famine Commissions. These recurring famines and consequent human sufferings and deaths became an important argument that was used by Indian leaders and their British sympathizers alike against rule by a non-representative colonial government and resulted in the founding of the Indian National Congress in the 1880s CE to kindle awareness among Indians for their rights. Thus, DHCs and consequent famines became an important factor in India’s demand for independence from Britain. The last major famine during colonial rule was the early 1940s CE drought and famine, whose causes are still debated (Table 3.4), but there is evidence that the British government deliberately exported food to Britain during the Second World War while tens of millions of Indians were starving (Mukerjee 2010). Since the 1940s CE, buffer stocks of food, modern transportation systems, and preparedness have combined to very substantially reduce human deaths due to droughts in India—and the Indian subcontinent, in general.

### Table 3.4   Epochs of Decadal Hydrologic Cycles in the Asian Monsoon Region and Their Societal Consequences

Years (CE)

Epoch

Name

Societal Consequences

1393–1407

Dry

Dvadasavarsha Panjam (12-year famine) and the Durga Devi Famine

South India, Deccan Plateau

1460–1465

Dry

Damaji Pant’s Famine

Deccan Plateau: drought and famine; Damaji Pant, a high revenue official in the Bahamani Kingdom in the Deccan plateau and a devotee of God Vithoba, is believed to have distributed grain to starving people from the royal granary (Kosambi 2000), but Damaji Pant is also mentioned as the savior in the 1393–1407 famine (The Journal of the Administrative Sciences 1979); a temple in Damaji Pant’s honor is in Mangalwedha, Solapur district, Maharashtra, India

1628–1635

Dry

Western India, Deccan Plateau: drought and consequent crop failures; 2 million deaths (Ó Gráda 2007)

1765–1773

Dry

Great Bengal Famine

Northern and Central Bengal, Bihar, Odisha, Bangladesh: drought and consequent large-scale crop failures exacerbated by forced cultivation of opium and indigo by the British East India Company, reducing food availability, large-scale outmigration of people from affected areas, 10 million deaths (Dutt 1908; Fiske 1942; Chaudhury 1999)

1780–1784

Dry

Chalisa Famine

Southern, Western, and Northern India: droughts followed by a volcanic eruption in Iceland (Laki fissure) may have caused decreased monsoon rainfall for several years; it was called Chalisa due to occurrence in the year 1840 of King Vikram (40 translates to “Chalis” in Hindi); total death toll may have been up to 11 million people (Grove 2007); serious, worldwide consequences of the Laki eruption (Wood 1992;Steingrímsson and Kunz 1998; Thordaldson and Self 2003; Oman et al. 2006)

1789–1801

Dry

Skull Famine

Western, Central, and Southern India: drought followed by severe famine for several years; prices of essential foods increased 300%–800% in 4–5 years in some regions (Gazetteer of the Bombay Presidency 1885); at least 11 million deaths (Grove 2007); known as “the Skull Famine” due to human skeletons lying on roads and fields; large-scale intra- and inter-regional migrations of people seeking food

1865–1874

Dry

Odisha–Rajasthan–Bihar Famines

Odisha, Bihar, Western India, Punjab: drought, price speculation and profiteering, insufficient food storage, improvement in British colonial administration for drought and famine relief as the multiyear event unfolded (Nisbet 1901; Imperial Gazetteer of India 1907; Yang 1998; Hall-Matthews 2008); 4 million to 5 million deaths due to starvation, malnutrition, and diseases; large-scale outmigration of people and cattle from Western India

1876–1878

Dry

Great Famine

Western and Southern India: grain exports by the British colonial government from India despite large-scale crop failures; 5.5 million people dead due to starvation, malnutrition, and diseases (Fieldhouse 1996); emigration of a large number of agricultural laborers to British tropical colonies as indentured laborers (Roy 2006); drought and famine subject of Tamil and other South Indian folk and literary traditions; famine led to the founding of the Indian National Congress in the next decade and became an important argument against British colonial rule (Hall-Matthews 2008)

1896–1900

Dry

Indian Famines

Western and Central India: large-scale crop failures; grain exports from some affected areas for profiteering; food riots in some areas; millions of cattle dead; 6 million to 8 million people dead due to starvation, malnutrition, and diseases (The Cambridge Economic History of India 1983; Seavoy 1986; Fieldhouse 1996; Maharatna 1996; Fagan 2009); appointment of the Famine Commission of 1898

1942–1944

Dry

Bengal Famine

West Bengal, Bihar, Odisha, Bangladesh: controversy about causes of famine; combination of effects of cyclonic storm, crop disease, and drought on food production; cutting off of rice imports from Myanmar (Burma) due to Japanese occupation, export of substantial quantities of rice to Britain, a deliberate policy of denial toward India’s food requirements by the British Government, price speculation and profiteering (Sen 1983; Mukerjee 2010); 3 million deaths due to starvation, malnutrition, and diseases; famine became subject of books and movies in Bengali and English

1964–1967

Dry

Western and Southern India: over 160 million affected

1983–1987

Dry

Western and Southern India: 300 deaths, 310 million affected

The Khmer Civilization, based in Angkor in present-day Cambodia, existed from approximately the eighth–ninth century CE to the seventeenth century CE. The Khmer people built an elaborate system of reservoirs and canals for irrigation to grow rice and other crops. It is plausible that the dry epoch from 1350 CE to 1369 CE (Figure 3.7) was a major cause of the decline of the Khmer Civilization, because the nearly two-decade-long drought may have severely strained the capacity of the Khmer agricultural system to produce necessary food to sustain its increasing human population. It is possible that some of the later dry (e.g., from 1681 CE to 1699 CE) and wet epochs were also partly responsible for the gradual withering away of the Khmer Civilization due to decreased agricultural production and damage to the water infrastructure (Buckley et al. 2010). Unfortunately, no written records about this multicentury period of the Khmer Civilization have survived and, so, further corroboration is not possible.

Cook et al. (2010) reconstructed four major drought events in the Asian summer monsoon region with their PDSI data. These four events are the Ming Dynasty Drought during 1638–1641 CE, the “Strange Parallels” drought during 1756–1768 CE, the East India Drought during 1790–1796 CE, and the “Great Drought” during 1876–1878 CE. A visual inspection of the APC1 time series (Figure 3.7) shows that the last two droughts, specific to India, reconstructed by Cook et al. (2010), are also present in the APC1 time series. The APC1 time series also shows that these droughts were embedded in DHCs (i.e., there were pluvials before or after these drought events) and lasted for several years to a decade or longer.

As in the cases of North America and Europe, tree-ring-based PDSI data over the last approximately 700 years in Monsoon Asia show that there were multiyear to decadal and longer dry and wet epochs. Supporting evidence in South Asia, and also to some extent in Southeast Asia, shows that there were major famines associated with many of the dry epochs that may have resulted in the deaths of tens of millions of people and may have disrupted water supply and irrigation systems in Southeast Asia; this may have caused a death blow to the Khmer civilization based in Angkor. However, the evidence does not unequivocally point to the dry epochs as the only cause of these famines and water system disruptions, but it does indicate that these epochs were a major contributing factor.

Visual inspections of the three PC 1 time series and other data indicated that there may be definite phase relationships in dry and wet epochs among the three regions, especially between North America and Monsoon Asia. Therefore, joint analyses of the three data sets were also carried out. Results are described in the next section.

#### 3.7  Global-Scale Decadal Hydrologic Cycles

In order to identify global-scale DHCs in the reconstructed PDSI data in North America, Europe and North Africa, and Monsoon Asia, data for the first two were averaged from a 0.5° longitude–0.5° latitude resolution to a 2.5° longitude–2.5° latitude resolution, since the Monsoon Asia data are at the coarser resolution. Then, EOF analysis was performed on the three combined data grids from 1306 CE to 2005 CE. It was found that the resulting EOF spatial patterns showed covariations only between North America and Monsoon Asia. Therefore, EOF analysis was also performed on combined data of these two regions; the prefix C is used to denote EOFs and PCs of the combined North America–Monsoon Asia data. CEOF 1–CPC 1 pair “explains” 8% of total variance. The CEOF 1 spatial pattern is shown in Figure 3.3d, and the corresponding CPC 1 time series is shown in Figure 3.8. The CEOF 1 spatial pattern clearly shows an opposite-phase oscillation in the PDSI between North America and India–Southeast Asia and between India–Southeast Asia and Central Asia. The latter is also seen in the Monsoon Asia AEOF 1 pattern (Figure 3.3c). The sign conventions in Figures 3.3d and 3.8 are explained in Section 3.4.

Figure 3.8   First combined principal component (CPC 1) time series of tree-ring-based North America and monsoon Asia PDSI data. CPC 1 “explains” 8% of the combined PDSI variance. Light yellow boxes mark wet and dry epochs. The bottom panel shows the CPC1 (bars) and the Pacific Decadal Oscillation index time series (line). See text for details.

Many of the dry and wet epochs common to North America and Monsoon Asia are marked on CPC 1 in Figure 3.8 with light yellow boxes. These epochs are also present in NPC1 and APC1, even if not marked. Thus, these three PCs clearly establish that some of the hydrologic cycles represented in the tree-ring-based PDSI are parts of global-scale DHCs. In order to see if these global-scale DHCs are associated with any DCV phenomenon, the Pacific Decadal Oscillation (PDO) index time series, calculated from the twentieth and early twenty-first centuries’ instrument-measured sea-surface temperatures (SSTs), is plotted on the CPC 1 time series in the lowest panel in Figure 3.8; this panel shows the CPC 1 and PDO index time series. As is obvious from these two time series, the two especially many major dry/wet epochs, are correlated with the correlation coefficient between time series without filtering at 0.4, which increases to 0.5 after low-pass filtering the two time series. Cook et al. (2010) also showed that a positive PDO phase—like SST pattern in the Pacific Ocean is associated with droughts in India and southeast Asia during the instrument data period from 1856 CE to 2004 CE. This correlation among the PDO and dry and wet epochs in North America and Monsoon Asia, especially the Indian subcontinent, during the instrument observations period suggests the possibility that some (or many) of the past DHC epochs revealed by the PDSI analyses going back many centuries may also be due to the PDO variability.

As mentioned in Section 3.3, several studies have calibrated tree-ring information in terms of the PDO index. Biondi et al. (2001) used tree-ring information from Southern California and Baja California to reconstruct annual PDO index from 1661 CE to 1991 CE. They found that frequency spectra of the reconstructed PDO index time series had the dominant peak at approximately 17 to 23 years periods; the strength of the peak varied over the 300 years of the record. In another reconstruction of the PDO index from 1565 CE to 1988 CE, D’Arrigo and Wilson (2006) used tree-ring information from northeast Asia during the March–April–May season. Although both sets of tree-ring information are influenced by local or regional weather and climate, there is a considerable degree of similarity between the two reconstructed PDO index time series in the overlap period. A comparison between these two reconstructed PDO index time series and the North American PDSI PC 1 time series in Figure 3.4 shows that six dry epochs (1665–1675, 1752–1758, 1818–1826, 1861–1866, mid-1890s, and 1948–1958 CE) and seven wet epochs (1675–1685, 1688–1697, 1745–1750, 1793–1796, 1827–1840, 1867–1872, and 1905–1930 CE) in North America from 1665 to 1958 CE occurred when both reconstructed PDO indices were generally in a negative or a positive phase, respectively. Because there is an opposite-phase relationship between wet and dry epochs in North America and Monsoon Asia, as shown in Figure 3.8 and described in the previous paragraph, the association between reconstructed PDO index time series and dry and wet epochs in North America applies to Monsoon Asia as well. These associations are consistent with the phases of the instrument data based PDO index’s association with dry and wet epochs as described in the previous paragraph. Thus, this extension of the PDO’s association with dry and wet epochs in North America and Monsoon Asia back to the seventeenth century strengthens the plausibility that the PDO SST pattern’s variability is indeed responsible for causing DHCs in these two regions.

These results show the presence of global-scale DHCs, at least in the 700-year period of the joint PDSI analysis. More reliable instrument-measured data analysis also strongly indicates the presence of such global-scale DHCs as we will see in the next chapter.

#### 3.8  Summary and Conclusions

In this journey of discovery of dry and wet epochs, we saw that tree rings can be very useful proxy climate indicators that can be calibrated in terms of climate variables such as precipitation, temperature, and PDSI. Such carefully calibrated, multicentury PDSI data in North America (1006–2005 CE), Europe and North Africa (1013–2012 CE), and Monsoon Asia (1300–2005 CE) show that there were DHCs on all four continents, some of which may be catalysts for, or even causes of, major famines and other strifes—even rise and fall of civilizations—resulting in deaths of many millions of people in some of these hydrologic epochs. We saw many instances of such catastrophes attributable to DHCs in North America, Europe, and India. As described in this chapter, it is intriguing that truly momentous events such as the French Revolution in 1789 CE may have resulted, along with other causes, from a series of dry epoch-driven food shortages in France. Tree-ring-based PDSI data also showed the presence of a global-scale oscillation of dry and wet epochs between North America and the India–southeast Asia regions in the era since at least 1300 CE. Major volcanic eruptions appear to have played substantial roles in initiating or sustaining multiyear to decadal dry and wet epochs.

Major conclusions of this chapter are as follows: (1) Tree-ring-based PDSI data appear to contain very useful information about climate variability in at least the last 1000 years; this information is supported reasonably well by societal impacts recorded by human observers; (2) DHCs have been very much a part of global climate variability for at least the last several hundred to a thousand years, as seen in the tree-ring-based PDSI data; (3) while natural DCV phenomena may be major drivers of worldwide DHC patterns, at least since the mid-twentieth century and possibly for a much longer period as the reconstructed PDO’s association with DHC in North America and Monsoon Asia since the 17th Century shows, major volcanic eruptions also appear to have important roles; (4) DCV impacts on the hydrologic cycle can be explicit as well as subtle; and (5) we have just started on this journey of discovery about DHCs.