Application of Streamflow Reconstruction to Water Resources Management - Dendroclimatology - page 232

In this chapter we place emphasis on the interactive aspect of streamflow recon-
struction with the water user or stakeholder. The statistical methodology of
reconstruction is addressed only peripherally, as that has been thoroughly reviewed
by Loaiciga et al. ( 1993 ) . Examples of recent modeling approaches adopted can be
found in recent papers (e.g., Hidalgo et al. 2000 ; Gedalof et al. 2004 ; Woodhouse
et al. 2006 ) .
This chapter begins with a brief historical background of streamflow reconstruc-
tion, followed by a section describing contributions of streamflow reconstruction
to the study of water resources. Case studies on a probabilistic approach to inter-
pretation of reconstructions (Colorado River, western United States) and on the
application of reconstructions to water resource management (Denver Water Board)
are then described. We close with a discussion of current challenges to streamflow
reconstruction and to the adoption of reconstructions by water managers.
8.2 Historical Background of Streamflow Reconstructions
In North America, dendrohydrological studies in the 1930s began to explore the
relationships between tree growth and streamflow, and the possible uses of tree-
ring records for extending gauge records (Hardman and Reil 1936 ; Hawley 1937 ;
Keen 1937 ) . Hardman and Reil's (1936) work concerning the flow of the Truckee
River, California-Nevada, was the first to examine these relationships in view of
possible applications to water resource management, particularly in the agricultural
regions of the Truckee River basin. In the 1940s, Schulman's ( 1945a , b, 1947 , 1951 ,
1956 ) interest in dendrohydrology led him to compare variations in ring widths and
annual runoff along the Pacific coast, in the Colorado River basin, the Missouri
River, and for several southern California rivers. Schulman's work in the Colorado
River basin was in part driven by the need to assess the long-term reliability of
power generation at Hoover Dam, and the record provided by a regional tree-ring
index allowed such an assessment (Schulman 1945a ; Stockton and Jacoby 1976 ) .
In work for the Denver Water Board, Potts ( 1962 ) collected and analyzed tree-ring
data to examine relationships with South Platte River, Colorado, annual flow and
to document recurrence of droughts. The Denver Water Board was interested in
estimating future storage requirements for the City of Denver's water supply and
hoped to use the record of past hydroclimatic variability from tree-ring data to sup-
port these estimates. Tree-ring-based analyses to this point consisted of comparisons
of ring widths and annual runoff, quantified by using correlation coefficients, and
frequently smoothed to account for persistence in annual runoff and tree growth
(Stockton and Jacoby 1976 and references within).
Stockton's dendrohyrolologic work in the 1970s was the first to take advantage
of a suite of new quantitative methods whose routine use was made possible by the
development of high-speed computers. The methods included multivariate statistical
analysis for evaluating the climatic signal in tree-ring records and for reconstructing
climate (Fritts et al. 1971 ) , and standardized protocol for field sampling and tree-ring