Proxy Streamflow Records for Water Planning In Western Canada

Water News, Fall, 2006 Proxy Streamflow Records for Water Planning In Western Canada Dave Sauchyn1, Jodi Axelson1 and Antoine Beriault2 1

2

Prairie Adaptation Research Collaborative, University of Regina, Regina, SK, S4S 7J7 Water Stewardship, Kootenay Division, BC Ministry of Environment, 1902 Theatre Rd, Cranbrook, BC, V1X 7G1

Introduction Climate and water monitoring networks provide the hourly and daily data used to derive and calibrate operational water supply forecasts that support decision making about reservoir storage, flood and drought mitigation, and hydropower production (Pagano et al. 2005). Instrumental data have more limited application, however, to long-term water planning, policy and infrastructure, because most records are too short to capture the decadal and longer-term variation in the regional climate and hydrology. Sustained departures from mean hydroclimate are observable only with proxy (surrogate) climatic and hydrometric data. Most drivers of climate variation have a periodicity that approaches or exceeds the length of instrumental records. Proxy hydroclimate data can provide water resource planners and engineers with • a context for reference hydrology to evaluate baseline conditions and water allocations • worst-case scenarios - what is possible in terms of the severity and duration of drought • long-term probability of hydroclimate conditions exceeding specific thresholds • a context for scenarios of water supply under climate change • a much broader perspective on the variability of water levels to assess the reliability of water supply systems under a wider range of flows than recorded by a gauge • the geographic extent of multi-year periods of low-and-high flows, including the synchroneity of droughts Tree-rings are the preferred proxy for records of climate variability at annual to multi-decadal scales spanning centuries to millennia (Briffa 2000). They are the source of both hydroclimate information and a chronology with absolute annual resolution. Annual variations in tree-ring width reflect daily and seasonal growth limiting processes. Where available soil moisture limits tree growth, standardized tree-ring widths correlate with hydrometric variables. Stream flow records correlate with moisture-sensitive tree-ring chronologies, because stream flow and tree growth have a similar muted response to episodic inputs of rainfall and snow melt water. Hydrological peaks are usually underestimated by tree rings, given a biological limit to the growth response to excess soil moisture. Therefore proxy records do not provide precise volumes of streamflow, yet they capture the timing and duration of periods of high and low flow. Tree rings are an especially good indicator of drought; dry years produce narrow rings. Until recently networks of moisture sensitive tree-ring chronologies have been lacking for western Canada, and streamflow has been reconstructed using just a few chronologies (e.g. Case and Macdonald 2003). Researchers at the University of Regina Tree Ring Lab (www.parc.ca/urtreelab) have collected tree rings at 85 sites (Figure 1) to enable the inference of long-term moisture and streamflow variability from a pool of predictor chronologies that capture more of the regional climatic variability. Nearly all of these collections are from open-canopy forests on ridge crests, southand west-facing slopes, and/or rapidly drained soils. At these dry sites, tree growth is limited by available soil moisture and therefore our tree-ring chronologies are proxies of summer and annual precipitation, soil moisture and runoff.

Water News, Fall, 2006

Figure 1. The network of tree-ring chronologies established by the University of Regina Tree Ring Lab (www.parc.ca/urtreelab). In this paper streamflow is reconstructed for the two gauges marked with blue triangles: Oldman River at Waldron’s Corner (AB) and Churchill River at Otter Rapids (SK). Examples of Streamflow Reconstructions Any technical paper in dendrohydrology (e.g. Beriault and Sauchyn 2006) will describe the methods for constructing tree-ring chronologies and developing models for inferring streamflow. Here we present two examples of streamflow reconstructions: the Churchill River at Otter Rapids (Figure 2) and the Oldman River at Waldron’s Corner (Figure 3), from master’s thesis research (Beriault 2005, Axelson 2006, respectively) funded by Manitoba Hydro. The streamflow data are plotted as departures from the mean; above normal flows are blue and below normal flows are red. Comparisons to the gauge records indicate that the tree rings capture the sequences of multi-year high and low flow.


400 200

-200

100

0

1840 1860 1880 1900 1920 1940 1960 1980 2000 Figure 2. Positive (blue) and negative (red) departures from mean annual streamflow (m3 sec-1) for the Churchill River at Otter Rapids. Streamflow was inferred from tree-rings for the period 1840 to 2003 (Beriault 2005). The gauge record (insert) is from 1960-2001. These proxy streamflow records reveal periodic shifts in hydroclimatic regime from intervals of dominantly low frequency variation to intervals of dominantly interannual variation. Large interannual variability represents a different challenge for water management than extended wet and dry intervals. The Churchill River reconstruction extends the 40-year gauge by 120 years but extensive and frequent fire limits the age of trees in the boreal forest and so the length of tree-ring records. The 500-year reconstruction for the Oldman River captures much more hydroclimatic variability than the gauge record, in terms of extremes and duration of high and low flows.


10 0 -10 8 4 0 -4 -8 1500

1600

1700

1800

1900

2000

Figure 3. Positive (blue) and negative (red) departures from mean water year (Oct – Sept) streamflow (m3 sec-1) for the Oldman River at Waldron Corner. Streamflow was inferred from tree-rings for the period 1502 to 2003 (Axelson 2006). The gauge record (insert) is from 1950-2000. Prospects for Applying Proxy Records to Water Supply Planning In Canada, there are indications that water management agencies recognize that proxy hydroclimate records potential have some value for water supply planning. A number of hydropower utilities in Canada have supported tree-ring reconstructions of hydrometric conditions (e.g., Beriault and Sauchyn 2006, Begin 2003, Meko 2006). US scientists have not only produced more streamflow reconstructions (e.g., Meko et al. 2001, Woodhouse 2001) but they also have collaborated with water managers for the application of tree-ring records to water supply planning. Much of this collaboration is focused on the Colorado and Salt River (AZ) watersheds. Initiatives include • TreeFlow: Streamflow reconstructions for Colorado http://wwa.colorado.edu/themes/streamflow_reconstructions.html • Regional Integrated Sciences and Assessments http://www.ogp.noaa.gov/mpe/csi/risa/index.htm o Western Water Assessment Program - http://wwa.colorado.edu/index.html/ o Climate Assessment Project for the Southwest http://www.ispe.arizona.edu/climas/research/paleoclimate/background.html Tree-ring scientists and water managers involved in these projects met on May 5, 2005 in Tucson, Arizona at a “Planning Workshop to Develop Hydroclimatic Reconstructions for Decision Support in the Colorado River Basin” (http://www.ispe.arizona.edu/climas/conferences/CRBpaleo/index.html). Presentations and outcomes of this workshop included (Garrick and Jacobs 2005): • Water managers who have incorporated paleoclimatic reconstructions into their planning processes; For example, Denver Water is working to use tree-ring records in tandem with their water supply simulation model to estimate the occurrence, frequency, and intensity of drought in the Colorado and Platte River • Water managers at the Salt River Project were surprised that preliminary tree-ring records show that synchronous drought in their two water supply sources are more common than was previously thought …it has motivated


• •

development of new strategies to prepare for potential water supply shortfalls caused by broad regional drought events The most up-to-date science still shows that the Colorado is over-appropriated There is a great need to publish findings in publications commonly read by water managers, instead of in scientific journals — the usual practice by paleoscientists

Publications read by water managers in Canada include Water News and the Canadian Water Resources Journal, which currently has a special issue on paleoenvironmental evidence of drought in western Canada (St. George and Sauchyn, 2006). Because the rivers of the Canadian interior are exposed to a large range of climate variability, and potentially major climate change, they are a logical focus for attempts to apply proxy streamflow records to water supply planning following the example set by our colleagues in the western US. References Axelson, Jodi. 2006. Tree-Ring of Streamflow in the South Saskatchewan River Basin from a Network of Tree-ring Chronologies, M.Sc. thesis, University of Regina. Beriault, Antoine, 2005. Using Tree Rings to Infer Annual Stream Discharge in the Churchill River Basin. M.Sc. thesis, University of Regina. Beriault, A. and Sauchyn, D.J. 2006. Tree-ring reconstructions of streamflow in the Churchill River basin, Saskatchewan. Canadian Water Resources Journal. Bégin, Y., 2003. Analyse dendrochronologique des effets climatiques d'un vaste réservoir hydroélectrique au Québec nordique (Canada). Revue de géographie alpine, numéro spécial : Hydrosystèmes lacustres et changements environnementaux, 91(1): 67-80. Briffa, K.R., 2000 Annual climate variability in the Holocene: interpreting the message of ancient trees. Quaternary Science Reviews, 19, 87-106. Case, R.A.and MacDonald G.M. 2003. Tree ring reconstructions of streamflow for three Canadian rivers. J. Amer. Water Resources Assoc. June: 703-716. Garrick, Dustin and Jacobs, Kathy. 2005. Tree-ring records inform water management decisions: Workshop strives for better integration of past with present. Southwest Climate Outlook, May 2005, p. 2-4 Meko, D.M. 2006. Tree-ring Inferences on Water-level Fluctuations of Lake Athabasca. Paleoenvironmental Evidence of Drought in Western Canada, Special issue of the Canadian Water Resources Journal. Meko, D.M, M. D. Therrell, C. H. Baisan, and M. K. Hughes, 2001. Sacramento River Flow Reconstructed to A.D. 869 From Tree Rings. J. American Water Resources Assoc. 37(4): 1029-1039. Pagano, T. C., Erxleben, J. and T. Perkins 2005: Operational simulation modeling at the NRCS National Water and Climate Center. Proceedings of the Western Snow Conference. Sauchyn, D. and Beaudoin, A. 1998. Recent environmental change in the southwestern Canadian plains. Canadian Geographer, 42: 337-353. St. George, S. and Sauchyn, D. 2006. Paleoenvironmental Evidence of Drought in Western Canada, Special issue of the Canadian Water Resources Journal. Woodhouse, C.A., 2001: A tree-ring reconstruction of streamflow for the Colorado Front Range. Journal of the American Water Resources Association 31/3: 561-569.