chapter 2

Canadian Prairies Drought from a Paleoclimate Perspective

David Sauchyn and Samantha Kerr

Introduction

Recurring drought is characteristic of the climate of the Canadian Prairies (Bonsal et al. 2011). It has serious consequences given the sub-humid climate (potential evapotranspiration exceeds precipitation in an average year) and predominance of agricultural land use in this region, which accounts for more than 80% of Canada’s agricultural land. The impacts of snowpack deficits, soil moisture depletion, and decreased streamflow and lake levels on the agricultural sector, and on water supplies in general, are well documented (e.g., Bonsal et al. 2011; Wheaton et al. 2008). Prolonged drought is especially damaging because its impacts are cumulative and can lower the resistance of ecosystems and soil landscapes to disturbance from hydroclimatic events to a point that thresholds of landscape change are exceeded and recovery of natural systems can take decades or centuries (Wolfe et al. 2001).

Drought is understood as a deficit of water: “Drought originates from a deficiency of precipitation over an extended period of time—usually a season or more—resulting in a water shortage for some activity, group, or environmental sector” (National Drought Mitigation Center 2006). According to this typical definition, a drought exists when the water deficit crosses a threshold in terms of duration and degree. These thresholds are a function of regional social and historical circumstances: sensitivity to water shortages and the adaptive capacity to deal with their adverse impacts. While precipitation and water level data are used to measure meteorological and hydrological drought, respectively, whether a drought is occurring depends on whether it is having an impact. The impacts of socio-economic drought, and specifically agricultural drought, range from a lack of soil moisture for dryland farming to the eventual depletion of water stored for irrigation.

In the Canadian Prairies, aridity and drought define the landscape and have punctuated the human history with periodic impacts and adaptation. Since ecosystems and rural communities in the driest areas are adapted to drought, and weeks without rain are characteristic of the summer climate, a season is probably an appropriate minimum duration for defining drought in this region. Summer water deficits are the norm for a semi-arid climate, and thus a season without much rain is tolerable, provided there is either access to irrigation or adequate soil moisture early in the season to enable germination and emergence of the crop. The impacts of a water deficit will therefore depend almost entirely on how much it lasts beyond one season—the “more” in “a season or more.” In recent years, droughts have only rarely persisted for more than several seasons or at most two to three years. These droughts were recorded by water and weather gauges and thus both perceived and defined meteorologically as the “normal” maximum duration. As a result, droughts of longer duration would conceivably exceed the adaptive capacity of Prairie agricultural communities. This chapter explores the question of whether droughts recorded and experienced by agrarian communities in western Canadian are as bad as they can get or whether we can expect droughts of greater intensity and longer duration based on our knowledge of past droughts. Thus, this chapter puts our recent experience with drought in a paleoclimatic context. Prairie drought is a recurring theme in paleo-environmental research (Sauchyn and Bonsal 2013; Bonsal et al. 2012; Lapp et al. 2012; St. George et al. 2009; Sauchyn et al. 2002, 2003). This chapter provides an overview of this research and presents a case study of paleodrought based on the reconstruction of the annual flow of Swift Current Creek (Saskatchewan) over the past four centuries.

Future climate will be a combination of natural climate cycles and the effects of anthropogenic climate change. Because the period of weather observations, since the 1880s, is short relative to some natural climate cycles and the return period of rare severe events, knowledge of pre-instrumental climate is required to determine the full range of variability and extremes in the regional climate and hydrology. Longer proxy hydroclimate records provide water resource managers and engineers with a historical context to evaluate 1) baseline conditions and water allocations, 2) worst-case scenarios in terms of severity and duration of drought, 3) long-term probability of hydroclimate conditions exceeding specific thresholds, 4) scenarios of water supply under climate change, 5) variability of water levels to assess reliability of water supply systems under a wider range of flows than recorded by a gauge, and 6) geographic extent of multi-year periods of low-and-high flows, including the synchronicity of droughts in adjacent watersheds (St. George and Sauchyn 2006).

Drought Proxies

The climate of the past, or paleoclimate, is preserved in biological and geological archives. Ecosystems and soil landscapes evolve under a certain range and seasonality of heat and moisture, and thus they correspond to regional climate regimes. But climate is not static, and as it varies, ecosystems and sediments preserve the climate changes and variability; they act as recorders of environmental change enabling the reconstruction of past climate variability on seasonal, yearly, and century-long time scales. This reconstruction of environmental history, irrespective of the proxy, is based on an understanding of the natural systems and their relationship to the current climate: “the present is the key to the past.” Thus, the interpretation of proxy data is only as good as the contemporary ecological, hydrological, meteorological, and geological data used to calibrate and interpret the proxy. There must be an appropriate measure of drought if environmental history is to be reconstructed.

Any systematic analysis of the intensity, duration, timing, frequency, and spatial extent of drought, including the inference of these characteristics from biological and geological archives, requires an operational definition based on one or more drought-related variables (Zargar et al. 2011). Among the quantitative expressions of drought, the most popular has been the Palmer Drought Severity Index (PDSI), which is based on precipitation, temperature (evapotranspiration), and soil water recharge rates. One complaint about the PDSI is that scaling of the index is sensitive to the soil moisture balance component. As McKee et al. (1993) pointed out, all measures of drought frequency, duration, and intensity are a function of implicitly or explicitly established time scales. They introduced the Standardized Precipitation Index (SPI) and demonstrated its applicability over intervals of 3 to 48 months. McKee et al. (1993: section 3.0) found a maximum correlation between the PDSI and SPI at 12 months, “suggesting that the PDSI does indeed have an inherent time scale even though it is not explicitly defined.” Vicente-Serrano et al. (2010) added a temperature term to the SPI to create the Standardized Precipitation-Evapotranspiration Index (SPEI). They concluded that “the PDSI is not a reliable index for identifying either the shortest or the longest time-scale droughts, which can have greater effects on ecological and hydrological systems than droughts at the intermediate time scales . . . only hydrological and economic systems that respond to water deficits at time scales of 9–18 months can be monitored using the PDSI” (2010: 9).

No common index or definition is available for paleodrought. The PDSI is the basis for most previous studies of recent and past prairie drought (Bonsal et al. 2012; Lapp et al. 2012; Sauchyn and Skinner 2001) and for the North American Drought Atlas, a continent-wide reconstruction of past drought from thousands of tree-ring chronologies (Cook et al. 2007). Use of the SPI or SPEI is likely to yield similar results as the PDSI, but at least these indices have the advantage of explicit time scales over which the applicability of the index is consistent (Vicente-Serrano et al. 2010; McKee et al. 1993). When the SPI is averaged over periods of up to 48 months, drought occurs with decreasing frequency, although these infrequent droughts of long duration represent the integration of series of water deficit events, with intervening periods of precipitation that are insufficient to overcome the accumulating water deficit. This statistical averaging is analogous to the integration of weather events, and the smoothing of short-term hydroclimatic variability, by drought proxies, whether annual tree growth or the gradual accumulation of plant and animal remains at the bottom of a lake. The higher the resolution of a proxy, the closer it comes to capturing a discrete drought event.

Drought indices were developed for analyzing instrumental meteorological and hydrometric time series and expressing the frequency and severity (intensity and duration) of water deficits. Their use for calibrating drought proxies is a unique application. Normally the use of a drought index, and particularly the choice of an averaging period, depends on the sensitivity of a system to water deficits of varying duration and intensity (Maliva and Missimer 2012). With drought proxies, however, that “averaging period” is a function of the sampling of the ecological or geological archive. The use of numerical drought indices is best suited for proxies, such as laminated sediments and tree rings, where the temporal resolution is high (years) and consistent, and the proxy is a measured physical or chemical property of the natural archive. Where the temporal resolution is low (decadal), and the indicator is simply the relative abundance of a climate proxy (e.g., plant pollen), inferred drought is typically described as an interval of dry climate or low water levels. Long paleo-environmental records encompass changes in climate, including shifts in aridity, which is a permanent water deficit as opposed to the temporary weather condition of drought (Maliva and Missimer 2012).

Each climate proxy represents a unique response of natural systems and processes to environmental change. Therefore, there is no universal definition of paleodrought in terms of duration and intensity, and human impacts are not considered unless there is an archaeological component to a paleoclimate study. Each proxy is a signal of a particular scale and aspect of climate, from the response of terrestrial (upland) vegetation to regional temperature and precipitation over multiple years, to the sensitivity of aquatic organisms to lake salinity, and carbonate mineralogy to lake water chemistry and temperature. The use and interpretation of climate proxies are subject to the following universal limitations and factors.

Location and timing: The sensitivity of natural systems to climate change and variability fluctuates over time and space. On the margins of regional ecosystems, species are at the limits of their ranges and thus are climatically sensitive. Island forests and permanent wetlands in the Prairie Ecozone provide a valuable source of information on environmental change, because they exist in an otherwise semiarid region and thus the terrestrial and aquatic species in these forest and wetlands are living on the margins climatically. The availability of indications of environmental change also varies over time. For example, lake sediments can yield detailed information about climate during dry periods (i.e., when lakes are low and sensitive), but during wet periods, lake sediments tend to yield less information about climate, because high water levels buffer the effects of fluctuations in temperature and precipitation. For 21 lakes in central Saskatchewan, Pham et al. (2008) found a coherent response to climatic variability in dry years but a lack of synchrony in wet years. Similarly, where heat, light, moisture, and nutrients are sufficient, tree growth is complacent—the rings have consistent width and no signal of inter-annual climate variability.

Resolution: Temporal resolution varies among proxy records according to the time span represented by individual samples and measurements. For example, in shallow and dry prairie lake basins, a single sample of lake sediment can represent material accumulated over decades, because sediments are re-suspended in the water on windy days and also lakes can periodically disappear. Some unusually deep freshwater lakes, on the other hand, contain continuous, undisturbed, and in some cases, annually laminated sediments (St. Jacques et al. 2015).

Non-climatic controls: Significant variations in proxy data can reflect the response of natural systems to internal thresholds or to events that are indirectly or not related to climate. Forests expand and become denser, and lakes fill with sediment and evolve chemically; proxy data from these systems can contain a signature of these processes. Land-use change also has a strong influence on the pollen and chemical record from prairie lakes (Pham et al. 2008).

Chronological control: Establishing the timing of climatic changes and resolving climatic variability depend entirely on chronological control, typically based on radiometric dating of organic and mineral carbon. Only tree rings and varves (annually laminated sediments) can be assigned to individual calendar years. Even these often represent floating chronologies, which must be dated by other means or correlated (cross-dated) with modern samples or strata to obtain absolute dates.

In the northern Great Plains of North America, changes in climate are recorded in the shifting of vegetation, fluctuations in the level and salinity of lakes, patterns of tree rings, and the age and mobility of sand dunes. In this dry environment, where lake water levels and chemistry, prairie vegetation, and rates of runoff and erosion reflect the soil and surface water balance, most proxies record fluctuations in hydroclimate, including periods of water deficits. Prairie paleodrought is identified mainly from studies of sediments, archival documents, and tree rings.

Lake and Terrestrial Sediments

Soils and sediments are ubiquitous climate proxies. They provide paleoclimate records that span the geologic history of a surface or sediment sink, although the age and origin of sediments usually can be resolved only to within decades (with the rare exception of annually laminated sediments). The sediments in permanent lakes represent a continuous accumulation of mineral and biological proxies. Although lakes are less common in semi-arid environments, they are important climate archives where drought is frequent and has ecological consequences. Pham et al. (2009) determined that the long-term chemical characteristics of prairie lakes were regulated mainly by changes in winter precipitation or groundwater flux. This finding has important implications for the hydro-climatic interpretation of the abundance of type of organisms found in lake sediments.

The postglacial climate history of the Canadian Prairies is known mostly from the analysis and interpretation of the type and abundance of fossil plants and aquatic organisms found in lake sediments. The analysis of bulk samples representing multiple decades limits the inference of hydroclimate to indications of relative aridity rather than drought. More recently, the continuous sampling and precise dating of lake sediments at fine intervals has yielded time series of higher resolution. The fine sampling of diatom assemblages from prairie lakes has revealed droughts embedded in multi-centennial shifts in moisture regimes (Michels et al. 2007; Laird et al. 2003). Using paleo-environmental information from the Peace–Athabasca Delta (PAD), Wolfe et al. (2008) determined that the levels of Lake Athabasca have fluctuated systematically over the past millennium. The lowest levels were during the eleventh century, whereas the highest lake levels coincided with maximum glacier extent during the Little Ice Age (sixteenth to nineteenth centuries). This important work has demonstrated that recent water level fluctuations on the PAD are within the range of long-term natural variability and therefore are very unlikely to be caused by the impoundment of water upstream (Wolfe, Hall, et al. 2012).

The frequency and duration of droughts also has been inferred from the age and origin of sand dune deposits (Wolfe, Hugenholtz, et al. 2012). Dry periods lasting years to decades will trigger the reactivation of a dune field; but the most severe droughts may not be detectable if continuous and extensive sand dune activity prevents the preservation of biological or geological evidence. From the precise optical dating of quartz grains, Wolfe et al. (2001) identified widespread reactivation of sand dunes in southwestern Saskatchewan about 200 years ago and correlated this geomorphic activity with tree-ring records of prolonged drought during the mid-to-late eighteenth century. A lag occurred between peak dryness around 1800 and the onset of dune activity at about 1810. Dune stabilization has occurred since 1890. The droughts of the 1930s and 1980s were insufficient to reactivate dunes.

Historical (Archival) Records

The Euro-Canadian (non-Aboriginal) history of the northern plains is several centuries longer than the instrumental observation of weather that began with agrarian settlement. Explorers and fur traders reported extreme weather and related events (e.g., fires, floods, ice cover). These documents are archived in libraries, museums, government repositories, and notably in the Hudson’s Bay Company Archives in Winnipeg. This archival information is valuable for verifying paleoclimate data from other sources (Rannie 2006; Blair and Rannie 1994). Severe, and at times prolonged, drought in the late eighteenth and mid-nineteenth centuries, evident in tree-ring and sand-dune chronologies, are described by explorers and fur traders. The archives of the Hudson’s Bay Company contain this report:

At Edmonton House, a large fire burned “all around us” on April 27th (1796) and burned on both sides of the river. On May 7th, light canoes arrived at from Buckingham House damaged from the shallow water. Timber intended to be used at Edmonton House could not be sent to the post “for want of water” in the North Saskatchewan River. On May 2nd, William Tomison wrote to James Swain that furs could not be moved as “there being no water in the river.” (Johnson 1967: 33–39, 58)

At the end of this dry decade, reports from Fort Edmonton House describe poor trade with both the Slave and Southern Indians due to “the amazing warmness of the winter” (Johnson 1967: 33-39, 58) diminishing both the bison hunt and creating a “want of beaver.” There were reports of smoke that almost obscured the sun and remarks like “the country all round is on fire.” The “amazing shallowness of the water” prevented the shipment of considerable goods from York Factory (the headquarters of the Hudson’s Bay Company on Hudson Bay).

In the 1850s, Captain John Palliser was dispatched from London by the Royal Geographical Society to evaluate the potential for British settlement of western Canada. He concluded that

this large belt of country embraces districts, some of which are valuable for the purposes of the agriculturalist, while others will forever be comparatively useless. . . . The least valuable portion of the prairie country has an extent of about 80,000 square miles, and is that lying along the southern branch of the Saskatchewan, and southward from thence to the boundary line [the US border]. (Palliser 1862: n.p.)

Palliser filed these remarks in 1860 in the midst of a 25-year drought. Despite his warning, settlers were drawn from Europe, eastern Canada, and the United States. The railroad and communities like Medicine Hat, Alberta, were built. In the very first edition of the Medicine Hat Times, dated 5 February 1891, an editorial entitled “Our True Immigration Policy” stated, “It would be almost criminal to bring settlers here to try to make a living out of straight farming” (Jones 2002: 18). As it turned out, the next several decades were relatively wet, settlers flooded in, and the populations of Saskatchewan and Alberta increased by nearly 500% in one decade. Certainly they were not aware of decadal-scale climatic variability and the fact the climate would later flip again and bring the devastating droughts of the 1920s and “Dirty 30s.”

Tree Rings

Tree rings provide a source of hydroclimatic data, such as data on available water and heat, and a chronology with absolute annual resolution spanning centuries to millennia. During the summer growing season in Canada’s western interior, there is usually more than enough light and heat. In this dry continental climate, soil moisture is the most limited determinant of tree growth. Therefore, the increment of annual growth is a proxy of hydrological or agricultural drought; dry years consistently produce narrow rings. Tree rings from living and dead trees that were growing at the same time for at least a few decades can be cross-dated, and calendar years can be transferred from the living to the dead trees. This process has produced tree-ring chronologies spanning the past millennium in western Canada. The mathematical relationship between standardized tree-ring widths and hydroclimatic data from nearby gauges is applied to the tree-ring data to reconstruct the relative moisture levels each year for the entire tree-ring record. Because the soil moisture that supports tree growth is derived mainly from melting snow and early-season rain, and because winter precipitation is strongly linked to large-scale climate oscillations, tree rings capture these teleconnections and the associated inter-annual to multi-decadal climatic variability, including periodic severe and prolonged drought.

Over the past 25 years, researchers in the Tree-Ring Lab at the University of Regina have collected more than 8,000 samples from old trees at more than 170 sites in the boreal, montane, and island forests of the Rocky Mountains and northern Great Plains. This network of tree-ring sites encompasses the semi-arid Prairie Ecozone. Because the tree rings were collected at dry sites (south- and west-facing slopes, sandy soils, ridge crests), where tree growth is moisture-sensitive, a strong correlation exists between the ring-width chronologies and drought and moisture indices. These tree-ring data have been the basis for a series of studies of Prairie drought (e.g., Bonsal et al. 2012; Lapp et al. 2012; St. George et al. 2009; Sauchyn et al. 2002, 2003; Sauchyn and Skinner 2001), including recent research by Kerr (2013), which is the source of the following case study.

A Tree-Ring Reconstruction of Hydrological Drought

Throughout the world, agriculture is the dominant use of water, and the major impacts of drought are directly or indirectly related to food production. Therefore, most indices of agricultural drought are expressions of the soil moisture balance, which unlike precipitation is not routinely measured. The best index of hydrological drought is streamflow; it is extensively monitored and integrates the net precipitation (in excess of evapotranspiration) over time (days to seasons) and watersheds. There is a relatively dense network of water-level gauges in the southern Prairies, since there is a strong demand for a limited surface water supply. This hydrometric network was originally established in the early twentieth century—not for the study of hydrology or climate, but rather to identify supplies of water initially for steam locomotives and irrigation (Greg McCullough, Water Survey of Canada, personal communication, June 2011). Therefore, just a few gauges have operated continuously for more than 50 years, recording only a few periods of sustained low water levels.

Tree rings are an effective streamflow proxy; they record the timing and duration of high and low water levels, and they have a similar muted response to episodic inputs of precipitation. When watersheds are wet (dry), streams rise (fall) and tree growth is enhanced (suppressed). Tree rings usually underestimate hydrological peaks, because there is a maximum positive biological response to available moisture; other factors constrain growth when soil moisture is not lacking. Thus, tree-ring data from moisture-sensitive ring-width chronologies are a better proxy of drought than of excess moisture.

Recently, Kerr (2013) completed a study of paleohydrology in the dry core of the northern Great Plains. Much of this region, at the junction of Alberta, Saskatchewan, and Montana (Figure 1), receives less than 330 mm of annual precipitation. Wetter conditions prevail in the uplands, so they contain island forests and the headwaters of all local rivers and streams. Kerr (2013) augmented and updated a network of tree-ring chronologies derived from lodgepole pine (Pinus contorta) and white spruce (Picea glauca) in the Cypress Hills (Alberta and Saskatchewan) and from Douglas fir (Pseudotsuga menziesii) and ponderosa pine (Pinus ponderosa) in the Sweet Grass Hills and Bears Paw Mountains of north-central Montana. Statistical tree-ring models explained 40%–55% of the recorded summer and annual flow of the Frenchman River, Battle Creek, and Swift Current Creek in southwestern Saskatchewan. The water-year (October–September) data from a gauge on Swift Current Creek are plotted in Figure 2 along with the flow predicted by a statistical tree-ring model for the same period (1979–2009). The two curves match in terms of the timing of high and low flows, although the tree rings underestimate the magnitude of the highest flows. Thus, they are a better proxy of drought than excess water.

Figure 1. Tree-ring sites (triangles) and streamflow gauges (squares) for a study of paleohydrology in the dry core of the northern Great Plains
(Source: Kerr 2013)

Figure 2. A plot of water-year (October–September) streamflow (m3/sec), from 1979 to 2009, as recorded at the gauge on Swift Current Creek below Rock Creek and reconstructed using tree rings.

Figure 3. A tree-ring reconstruction of the flow of Swift Current Creek since 1672. Top plot: Water-year (October–September) streamflow (m3/sec) showing the mean flow and 10th and 25th percentiles. Bottom: Water-year (October–September) streamflow (m3/sec) plotted as departures from the mean reconstructed value. (Source: Kerr 2013)

By applying the statistical tree-ring model of streamflow to the entire length of the tree-ring chronologies, water-year flow from 1672 to 2009 was reconstructed. Two versions of this paleo-flow time series are plotted in Figure 3. The top plot shows the inferred annual flow, mean flow, and two thresholds of low flow—the 10th and 25th percentiles. In the bottom plot, departures from the reconstructed mean flow highlight the inter-annual variability and inter-decadal pattern, with extended periods of low flow evident in the 1790s to early 1800s and the late 1840s through 1870s.

Because much of the unexplained variance in the calibration period (1979–2009; Figure 2) can be attributed to the underestimation of high flows, more confidence can be applied to the interpretation of low flows, which consistently correspond to narrow tree-rings, capturing the timing and duration of drought. The late eighteenth through mid-nineteenth centuries have the most sustained low flows. Severe hydrological droughts occurred from 1794 to 1798, 1816 to 1821, and 1854 to 1860 (Table 1). The repetitive nature of moisture surpluses and deficits in the streamflow reconstruction suggests some quasi-cyclical behaviour in the hydroclimatic regime. Spectral analyses provided evidence of this periodic hydroclimatic variability at the inter-annual (~2–6 years) and multi-decadal (~20–30 years) scales corresponding to the dominant frequencies of the El Niño–Southern Oscillation (ENSO) and the Pacific Decadal Oscillation (PDO).

Hydrologic droughts in the Swift Current Creek paleohydrology coincide with low flows and below-average precipitation in other paleoclimate records from western North America. The early to mid-1700s was a period of prolonged drought documented by paleoclimatic investigations across western North America (Cook et al. 1999; Woodhouse and Overpeck 1998; Laird et al. 1996; Stockton and Meko 1983). Various paleoclimatic studies emphasize the sustained nature of drought during the mid-nineteenth century, with very little relief in a few scattered wet years. This intense, long-lasting drought is well documented as occurring from the 1840s through mid-1860s throughout the western United States, Canada, and Mexico (Stahle and Dean 2011; Stahle and Cleaveland 1988; Fritts 1983; Stockton and Meko 1983; Hardman and Reil 1936).

This case study of the paleohydrology of southwestern Saskatchewan demonstrates tree rings are an effective proxy of annual streamflow. The proxy record for Swift Current Creek reveals periods of sustained low flow, including pre-settlement droughts that exceed in intensity and duration the worst conditions that have affected modern agriculture on the northern plains. Water deficits of this severity will reoccur in the future in a climate of rising temperatures.

Conclusion

Seasonal, and sometimes prolonged, moisture deficits are so characteristic of the climate of the Canadian Prairies that they define the region ecologically and ultimately limit forest and farmland productivity. Severe drought of high intensity and/or long duration, with serious consequences, is relatively infrequent. Thus over the past 13 decades, since Euro-Canadians first came to ranch and farm, people have experienced and recorded relatively few severe droughts. Although the meteorology and socio-economic impacts of these relatively few severe droughts have been studied extensively (e.g., Bonsal et al. 2011; Wheaton et al. 2008), the sample size is too small to analyze the frequency of these events in relation to regional climate variability and change. Thus, in this chapter, we examined the paleoclimate record of drought extending back over the past millennium.

A robust conclusion of the paleoclimate research on the Prairies is that the climate of the instrumental period is representative of the longer-term frequency of one- to two-year droughts but does not capture the full range of intensity and duration. The dry periods of greatest severity and duration occurred before the Prairies were settled. These include the intense drought years of the late eighteenth century (and the sand dune activity described above) and the sustained drought of the 1840–60s. Thus, the proxies suggest that the climate of the twentieth century (especially since the 1930s) was relatively favourable for the settlement of the Prairies, because the region has lacked the sustained droughts of preceding centuries. While the twentieth-century droughts may have been characterized by relatively modest precipitation deficits compared to earlier events, they have been hotter droughts than the cooler moisture-deficient periods of preceding centuries. This finding has important implications for studying and projecting future drought in a period of rapid global warming. The most serious impact of a warming climate in this region would be realized if the droughts of the 1790s or 1850s, and associated natural forcing, were to reoccur in the much warmer greenhouse gas climate of the twenty-first century.

The paleoclimatic records capture the tempo of natural climate variability, including the near-regularity of wet and dry cycles at certain frequencies. They show that the hydroclimatic regime periodically shifts from predominantly interannual variation to intervals with extended wet and dry spells and that there is a significant difference in the likelihood of drought according to phase of ocean-atmosphere oscillations (ENSO and PDO). This knowledge of long-term climate variability contributes to our understanding of the climate system at scales that exceed the length of instrumental records. The longest and most intense droughts, and the factors that cause them, reoccur so infrequently that a pre-instrumental paleoclimate perspective is required to validate the modelling and prediction of these events.

Our capacity to withstand and prepare for water scarcity has developed in response to the droughts that have occurred since the Prairies were first settled for agriculture, which have been shown to be much less intense than those that occurred before the Prairies were settled (i.e., those in the paleoclimate record presented here). Greater adaptive capacity will be required if future drought conditions are more intense or prolonged than those previously experienced. Significant adaptations may be required, particularly to water management practices and policies, starting with a scientific knowledge base that extends beyond instrumental records and the scale at which water supplies seem relatively secure and stationary, and then encompassing the longer view provided by paleoclimate records and model projections of future climate. Communities and governments are investing effort and resources in adaptation planning, in large part to mitigate the potential impacts of a warmer and more extreme climate. To inform this process, and be perceived as a credible source of information on exposure to drought, reconstructions of long-term climatic variability must be based on definitions of drought that are applicable to agriculture and water resource management.

The characteristics of drought detected using natural and historical archives must be related to droughts of recent experience and to the nearest modern analogues. To communicate the severity of paleodroughts, an example might be given of a situation in which one intense recent drought is followed immediately by another similarly intense drought. In this way, the characteristics of the megadroughts in the paleoclimate record can be translated into terms that are used and understood by other natural and social scientists, and by engineers and resource managers responsible for monitoring and managing drought.

Acknowledgments

Funding for the case study was provided by the Rural Community Adaptation to Drought project through the Social Sciences and Humanities Research Council and by the Prairie Adaptation Research Collaborative. Jessica Vanstone, Cesar Perez-Valdivia, Ben Brodie, and Tiffany Vass assisted Samantha Kerr with field and laboratory work.

References

Blair, D., and W.F. Rannie. 1994. “Wading to Pembina: 1849 Spring and Summer Weather in the Valley of the Red River of the North and Some Climatic Implications.” Great Plains Research 4, no. 1: 3–26.

Bonsal, B.R., R. Aider, P. Gachon, and S. Lapp. 2012. “An Assessment of Canadian Prairie Drought: Past, Present, and Future.” Climate Dynamics. doi: 10.1007/s00382-012-1422-0.

Bonsal, B., E. Wheaton, A. Chipanshi, C. Lin, D. Sauchyn, and L. Wen. 2011. “Drought Research in Canada: A Review.” Atmosphere-Ocean 4: 303–19. http://dx.doi.org/10.1080/07055900.2011.555103.

Cook, E.R., D.M. Meko, and D.W. Stahle. 1999. “Drought Reconstruction for the Continental United States.” Journal of Climate 12: 1145–62.

Cook, E., R. Seager, M. Cane, and D. Stahl. 2007. “North American Drought: Reconstructions, Causes, and Consequences.” Earth-Science Reviews 81: 93–134.

Fritts, H. 1983. “Tree-ring Dating and Reconstructed Variations in Central Plains Climate.” Transactions of the Nebraska Academy of Science and Affiliated Societies 11: 37–41.

Hardman, G., and O.E. Reil. 1936. “The Relationship between Tree-growth and Stream Runoff in the Truckee River Basin, California–Nevada.” Agricultural Experiment Station 41: 1–38.

Johnson, A.M. (ed.). 1967. Saskatchewan Journals and Correspondence: Edmonton House and Chesterfield House, 1795–1800. London: Hudson’s Bay Record Society.

Jones, D.C. 2002. Empire of Dust: Settling and Abandoning the Prairie Dry Belt. Calgary: University of Calgary Press.

Kerr, S.A. 2013. “A Dendroclimatic Investigation of Southwestern Saskatchewan.” M.Sc. thesis, University of Regina.

Laird, K.R., B.F. Cumming, S. Wunsam, O. Olson, J.A. Rusak, R.J. Oglesby, S.C. Fritz, and P.R. Leavitt. 2003. “Lake Sediments Record Large-Scale Shifts in Moisture Regimes across the Northern Prairies of North America during the Past Two Millennia.” Proceedings of the National Academy of Sciences 100: 2483–88.

Laird, K.R., S.C. Fritz, K.A. Maash, and B.F. Cumming. 1996. “Greater Drought Intensity and Frequency before A.D. 1200 in the Northern Great Plains.” Nature 384: 551–54.

Lapp, S., J.-M. St. Jacques, D. Sauchyn, and J. Vanstone. 2012. “Forcing of Hydroclimatic Variability in the Northwestern Great Plains since AD 1406.” Quaternary International 310: 47–61.

Maliva, R., and T. Missimer. 2012. Arid Lands Water Evaluation and Management, Environmental Science and Engineering. Berlin Heidelberg: Springer-Verlag.

Michels, A., K.R. Laird, S.E. Wilson, D. Thomson, P.R. Leavitt, R.J. Oglesby, and B.F. Cumming. 2007. “Multi-Decadal to Millennial-Scale Shifts in Drought Conditions on the Canadian Prairies over the Past Six Millennia: Implications for Future Drought Assessment.” Global Change Biology 13: 1295–1307.

McKee, T.B., N.J. Doesken, and J. Klesit. 1993. “The Relationship of Drought Frequency and Duration to Time Scales.” Pp. 179–84 in Proceedings of the Eighth Conference on Applied Climatology. Anaheim, CA, 17–22 January 1993.

National Drought Mitigation Center. 2006. “What is Drought? Understanding and Defining Drought.” National Drought Mitigation Center website. http://drought.unl.edu/DroughtBasics/WhatisDrought.aspx.

Palliser, J. 1862. “Journals, Detailed Reports and Observations Relative to Captain Palliser’s Exploration of a Portion of British North America.” In I.M. Spry (ed.) (1968), The Papers of the Palliser Expedition. Toronto: The Champlain Society.

Pham, S.V., P.R. Leavitt, S. McGowan, and P. Peres-Neto. 2008. “Spatial Variability of Climate and Land-use Effects on Lakes of the Northern Great Plains.” Limnology and Oceanography 53, no. 2: 728–42.

Pham, S.V., P.R. Leavitt, S. McGowan, B. Wissel, and L.I. Wassenaar. 2009. “Spatial and Temporal Variability of Prairie Lake Hydrology as Revealed Using Stable Isotopes of Hydrogen and Oxygen.” Limnology and Oceanography 54, no. 1: 101–18.

Rannie, W.F. 2006. “A Comparison of 1858–59 and 2000–01 Drought Patterns on the Canadian Prairies.” Canadian Water Resources Journal 31, no. 4: 263–74.

Sauchyn, D.J., E.M. Barrow, R.F. Hopkinson, and P. Leavitt. 2002. “Aridity on the Canadian Plains.” Géographie physique et Quaternaire 66: 247–69.

Sauchyn, D.J., and B. Bonsal. 2013. “Climate Change and North American Great Plains’ Drought.” In A.-H. El-Shaarawi and W. Piegorsch (eds.), Encyclopedia of Environmetrics. Chichester, UK: John Wiley & Sons. doi: 10.1002/9780470057339.vnn123.

Sauchyn, D.J., and W.R. Skinner. 2001. “A Proxy PDSI Record for the Southwestern Canadian Plains.” Canadian Water Resource Journal 26, no. 2: 253–72.

Sauchyn, D.J., J. Stroich, and A. Beriault. 2003. “A Paleoclimatic Context for the Drought of 1999–2001 in the Northern Great Plains.” The Geographical Journal 169, no. 2: 158–67.

St. George, S., D.M. Meko, M.P. Girardin, G.M. MacDonald, E. Nielsen, G.T. Pederson, D.J. Sauchyn, J.C. Tardif, and E. Watson. 2009. “The Tree-Ring Record of Drought on the Canadian Prairies.” Journal of Climate 22: 689–710.

St. George, S., and D. Sauchyn. 2006. “Paleoenvironmental Perspectives on Drought in Western Canada.” Canadian Water Resources Journal 31, no. 4: 197–204.

St. Jacques, J.-M., B.F. Cumming, D.J. Sauchyn, and J.P. Smol. 2015. “The Bias and Signal Distortion Present in Conventional Pollen-based Climate Reconstructions as Assessed by Early Climate Data from Minnesota, USA.” PLoS One 10, no. 1: e0113806.

Stahle, D.W., and M.K. Cleaveland. 1988. “Texas Drought History Reconstructed and Analyzed from 1698 to 1980.” Journal of Climate 1: 59–74.

Stahle, D., and J. Dean. 2011. “North American Tree Rings, Climatic Extremes, and Social Disasters: Social Impacts of Climate Extremes during the Historic Era.” In M. Hughes, T. Swetnam, and H. Diaz (eds.), Dendroclimatology Progress and Prospects. Developments in Paleoenvironmental Research. Volume 11. London/New York: Springer Dordrecht Heidelberg.

Stockton, C., and D. Meko. 1983. “Drought Recurrence in the Great Plains as Reconstructed from Long-term Tree-ring Records.” Journal of Climate and Applied Meteorology 22: 17–29.

Vicente-Serrano, S., S. Begueria, and J. López-Moreno. 2010. “A Multiscalar Drought Index Sensitive to Global Warming: The Standardized Precipitation Evapotranspiration Index.” Journal of Climate 23: 1696–1718.

Wheaton, E., S. Kulshreshtha, V. Wittrock, and G. Koshida. 2008. “Dry Times: Hard Lessons from the Canadian Drought of 2001 and 2002.” Canadian Geographer 52: 241–62.

Wolfe, B.B., R.I. Hall, T.W.D. Edwards, S.R. Jarvis, R.N. Sinnatamby, Y. Yi, and J.W. Johnston. 2008. “Climate-Driven Shifts in Quantity and Seasonality of River Discharge over the Past 1000 Years from the Hydrographic Apex of North America.” Geophysical Research Letters 35: 1–5.

Wolfe, B.B., R.I. Hall, T.W.D. Edwards, and J.W. Johnston. 2012. “Developing Temporal Hydroecological Perspectives to Inform Stewardship of a Northern Floodplain Landscape Subject to Multiple Stressors: Paleolimnological Investigations of the Peace–Athabasca Delta.” Environmental Reviews 20: 1–20.

Wolfe, S.A., C.H. Hugenholtz, and O.B. Lian. 2012. “Palliser’s Triangle: Reconstructing the ‘Central Desert’ of the Southwestern Canadian Prairies during the late 1850s.” The Holocene 23: 669–707.

Wolfe, S.A., D.J. Huntley, P.P. David, J. Ollerhead, D.J. Sauchyn, and G.M. Macdonald. 2001. “Late 18th Century Drought-induced Sand Dune Activity, Great Sand Hills, Southwestern Saskatchewan.” Canadian Journal of Earth Sciences 38: 105–17.

Woodhouse, C., and J. Overpeck. 1998. “2000 Years of Drought Variability in the Central United States.” Bulletin of the American Meteorological Society 79, no. 12: 2693–2714.

Zargar, A., R. Sadiq, B. Naser, and F. Khan. 2011. “A Review of Drought Indices.” Environmental Reviews 19: 333–49.