ATMOSPHERIC AND OCEANIC INDICES SINCE AD 1000 RECONSTRUCTED FROM TREE RINGS

1
ATMOSPHERIC AND OCEANIC INDICES SINCE AD 1000
RECONSTRUCTED FROM TREE RINGS IN THE AMERICAN
SOUTHWEST
Fenbiao Ni, Malcolm K. Hughes and Gary Funkhouser
( fenbiao_at_ltrr.arizona.edu mhughes_at_ltrr.arizona
.edu gary_at_ltrr.arizona.edu ) Laboratory of
Tree-Ring Research, University of Arizona, Tucson
AZ, 85721This work has been supported by the
U.S. National Oceanic and Atmospheric
Administration's Office of Global Programs
through cooperative agreement NA87GP0061 and by
U.S. National Oceanographic and Atmospheric
Administration's Paleoclimatology program,
through grant NA66GP0311 to Professor
M.K.Hughes.
Reconstruction Procedure
Introduction
To test if there is an SOI, PDO and SWTROF signal
in this tree-ring data set, we used multiple
linear regression (LR) with atmospheric/oceanic
indices as predictand and the 19 tree-ring
chronology as predictors. To obtain measures of
accuracy, the stepwise LR model is
cross-validated with a leave-one-out procedure.
That is, cross-validation is carried out using
developmental data sets of size n-1, and
verification data sets containing the remaining
single observation of predictand. The regression
model is recalculated for each of the n distinct
partitions of the data. The result is n similar
transfer functions, each computed without one of
the observations of the predictand. The
verification statistics were emphasized to ensure
that the calibrated relationships were preserved
in the independent estimates. Also, the
reconstruction was compared to other independent
records.
Time series of tree-ring width serve as
proxy records of past weather and climate beyond
the short period covered by instrumental data.
The purpose of this research is to link
atmospheric/oceanic indices that affect local and
regional climate with the long tree-ring
chronologies in the American Southwest.  
Data

• Monthly atmospheric/oceanic indices were obtained
  from the Joint Institute for the Study of the
  Atmosphere and Ocean (JISAO) at
  http//tao.atmos.washington.edu/ and were
  averaged over the cool-season (Nov-Apr).
• Tree-ring growth indices from 19 sites in AZ, NM,
  UT, CO, CA, and OR (Fig. 1) (Ni et al., 1999),
  selected from a set of 76 on the basis of
  dendrochronological quality (between-tree
  correlations, high sample depth back through AD
  1000) and last date of AD 1988 or later.

Results and Discussion
A predicted residual sum of squares (PRESS) was
used to cross-validate the reconstructing
capability of models from an independent data
set. Fig. 3 shows the tree-ring sites retained in
the best model in each case. These include
juniper in Oregon, bristlecone pine, sequoia and
foxtail pine in eastern California, limber,
ponderosa pine and douglass fir in Arizona/New
Mexico.
Fig. 5. Top correlation fields between
observed SOI and sea surface temperature (left),
500 mb geopotential height ( right) for the
calibration/verification period. Bottom same but
for reconstructed SOI. These figures further
indicate the success in using tree-rings to
reconstruct the SOI Index.
Robust quasi-periodic features in the 1000-year
atmospheric/oceanic time series were also
revealed by the multiple-taper spectrum estimate
method (MTM). For example, we have detected a
strong 73-year oscillation in the PDO. Villalba
et al. (1996) and Cook et al. (1996) and many
others also reported such century-scale
oscillations. Sudden reversals with significant
environmental impact from dry to wet were not
uncommon throughout the millennium such as around
1600 and 1900 (Ni et al., 2001). For example, the
1610s wet conditions following the 16th century
mega drought had a strong impact on tree growth
(Figure 6 left). The most recent post-1976
reversal has also been linked to regional-scale
tree growth surge in the Southwest (e.g., Swetnam
and Betancourt 1998). The instrumental records
indicate that dry to wet reversals in the 20th
century are associated with strong shifts from
cold to warm ENSO events and from negative to
positive PDO. This physical link might also be
the cause of the strong reversals during the last
millennium.
Figure 3. Tree-ring sites in the best models.
Fig. 1. Left Sea-surface temperature anomalies
(C) associated with the Pacific Decadal
Oscillation (PDO) pattern and circulation action
centers for the definitions of Southern
Oscillation Index (SOI) and Southwestern Trough
(SWTROF) Rightlocation of the 19 tree-ring
sites.
Table 2. Calibration/validation statistics
The reconstructions capture a moderate proportion
of global scale atmospheric/oceanic indices
variations (30) and do a even better job for
the regionally tailored circulation indices based
on circulation aloft (for example, 65 variance
explained for SWTROF). We have produced
cross-validated reconstructions of PDO, SOI and
SWTROF back to AD 1000 from tree rings. The
reconstructions provide a decadal to multicentury
perspective on circulation variability for work
based on the instrumental record. For example,
the amplitude and suddenness of the 1977 PDO
reversal is unusual but not unique on a 1000-year
time scale. Features like that after 1976 are
inferred for the 1890s, 1580s, 1000, etc.. The
reversal near 1890 is also noted from
instrumental data by Minobe (1997), Mestas-Nunes
and Enfield (1999). Regionally tailored
circulation indices based on circulation aloft
such as SWTROF also reveal a distinct pattern
over the last millennium but with a smaller
amplitude of fluctuation.
Table 1. Atmospheric/oceanic indices used in
this study
The trees that make up the chronologies for a
millennial network are relatively isolated and
have limited interaction with their neighbors.
Many exhibit particularly clean climatic signals.
For example, this same network has been used to
produce annual maps of summer PDSI to investigate
the decade to century scale variability in soil
moisture in the Southwestern United States
(Hughes et al., 1999). Atmospheric/oceanic
indices, such as the PDO and the SOI, are
important features that directly influence
western US climate variations. Correlation fields
in Fig.2 shows that PDO/- SOI/-SWTROF give
increased Cool-season Precipitation in the study
area.
Fig. 6. Left top cumulative severity of AD
1561-1600 growth reduction in moisture-limited
trees (Biondi et al., 2000) Left bottom
cumulative excess of AD 1601-1640 growth in
moisture-limited trees (Biondi et al., 2000)
Right top Using MTM to evaluate the 1000-year
reconstructions of cool-season PDO. Right bottom
Waveform of the detected 73-year PDO
oscillation.
References
Biondi, F., A. Gershunov, and D.R. Cayan, 2001,
North Pacific decadal climate variability since
AD 1661, Journal of Climate, Volume 14, Number 1,
January 2001. Cook, E.R., Buckley, B.M.,
D'Arrigo, R.D., 1996, in Jone, Bradley and Jouzel
(editors) Climatic Variations and Forcing
Mechanisms of the Last 2000 Years. Springer,
Berlin. DArrigo, R. D. and G. C. Jacoby, 1991.
Holocene, 1, 195-101. Fenbiao Ni, Cavazos, T., M.
Hughes, A. Comrie and G. Funkhouser, 2001. Cool
Season Precipitation in the Southwestern United
States since AD 1000 Comparison of Linear and
Nonlinear Techniques for Reconstruction. To be
published in International Journal of
Climatology. Hughes MK., G. Funkhouser and
Fenbiao Ni, 1999, PAGES-CLIVAR Workshop on
Climate of the Last Millennium. Venice, Nov.
8-12, 1999. Mantua, N.J., S.R. Hare, Y. Zhang,
J.M. Wallace, and R.C. Francis,(1996), Bulletin
of the American Meteorological Society,78,
1069-1079. Mestas-Nunez A.M., Enfield D.B., 1999
, J.Clim. 12(9), 2734-2746. Minobe, S., 1997,
Geophys. Res. Lett., 24, 683-686. Ropelewski,
C.F. and Jones, P.D., 1987, Monthly Weather
Review 115, 2161-2165. Stahle, D. W., E. R.
Cook, M. K. Kleaveland, M. D. Therrell, D. M.
Meko, H.D. Grissino-Mayer, E. Watson, and B.H.
Luckman, 2000. EOS, AGU, 81, 12 March
2000. Swetnam TW, Betancourt JL. 1998. Mesoscale
disturbance and ecological response to decadal
climatic variability in the American Southwest.
Journal of Climate. 11 3128-3147. Villalba, R.,
Boninsegna, J.A., Lara, A., Veblen, T.T., Roig,
F.A., Aravena, J.-C., Ripalta, A., 1996, in Jone,
Bradley and Jouzel (editors) Climatic Variations
and Forcing Mechanisms of the Last 2000 Years.
Springer, Berlin. Woodhouse, C.A., 1997, Climate
Research 8(2), 117-127. Cavazos, T., 1997. Int.
J. Climatol., 17, 1069-1082.
Fig. 2. Top Correlation fields between PDO and
US climatic divisional precipitation Middlesame
but for SOI Bottom same but for SWTROF
(http//www.cdc.noaa.gov/USclimate/Correlations).
Fig. 4. Cool season PDO (top), ENSO patterns
(center) and SWTROF anomaly index (bottom) from
1000 to 1987. The reconstructions have been
smoothed with cubic splines to highlight decadal
variability.