Wood Anatomy of Douglas-Fir in Eastern Arizona and Its Relationship With Pacific Basin Climate

Abstract

Dendroclimatic reconstructions, which are a well-known tool for extending records of climatic variability, have recently been expanded by using wood anatomical parameters. However, the relationships between wood cellular structures and large-scale climatic patterns, such as El Niño-Southern Oscillation (ENSO) and the Pacific Decadal Oscillation (PDO), are still not completely understood, hindering the potential for wood anatomy as a paleoclimatic proxy. To better understand the teleconnection between regional and local climate processes in the western United States, our main objective was to assess the value of these emerging tree-ring parameters for reconstructing climate dynamics. Using Confocal Laser Scanning Microscopy, we measured cell lumen diameter and cell wall thickness (CWT) for the period 1966 to 2015 in five Douglas-firs [Pseudotsuga menziesii (Mirb.) Franco] from two sites in eastern Arizona (United States). Dendroclimatic analysis was performed using chronologies developed for 10 equally distributed sectors of the ring and daily climatic records to identify the strongest climatic signal for each sector. We found that lumen diameter in the first ring sector was sensitive to previous fall-winter temperature (September 25^th to January 23^rd), while a precipitation signal (October 27^th to February 13^th) persisted for the entire first half of the ring. The lack of synchronous patterns between trees for CWT prevented conducting meaningful climate-response analysis for that anatomical parameter. Time series of lumen diameter showed an anti-phase relationship with the Southern Oscillation Index (a proxy for ENSO) at 10 to 14year periodicity and particularly in 1980-2005, suggesting that chronologies of wood anatomical parameters respond to multidecadal variability of regional climatic modes. Our findings demonstrate the potential of cell structural characteristics of southwestern United States conifers for reconstructing past climatic variability, while also improving our understanding of how large-scale ocean-atmosphere interactions impact local hydroclimatic patterns.

Keywords: El Niño/Southern Oscillation; Pacific Decadal Oscillation; Pseudotsuga menziesii; paleoclimate; quantitative wood anatomy; western United States.

Figure 1

(A) Geographical location of the two study sites. (B) Walter-Lieth climatic diagram for the period 1981–2015 based on PRISM data.

Figure 2

Microscopic composite (merged) image of Pseudotsuga menziesii at 300x magnification with brightness correction.

Figure 3

Annual (gray solid lines) and mean (black dashed lines) standardized tracheidograms of (A) lumen diameter and (B) cell wall thickness (CWT) for the five sampled trees at Nutrioso and Black River for the years 1966–2015. Roman numerals indicate tree-ring sectors, from the beginning (sector I) to the end of the ring (sector X).

Figure 4

Standardized sector chronologies for the period 1966–-2015 for lumen radial diameter (A) and CWT (B) moving from the beginning (sector I) to the end (sector X) of the ring.

Figure 5

Correlations of sector chronologies of lumen diameter with daily (A) maximum temperature (Tmax) and (B) precipitation for the period 1981 to 2015. The horizontal length of each bar corresponds to the strongest optimal period identified by the moving window correlation analysis. All correlations, whose value is given by the color, are significant at p<0.05.

Figure 6

Normalized time series of radial lumen diameter chronologies for sector I and for the average between sectors I-V plotted against (A) temperature and (B) precipitation. Climate data are the average daily Tmax from September 25^th to January 23^rd, and the sum of daily precipitation from previous October 27^th to current February 13^th. The sector I chronology of lumen diameter is shown with opposite sign for easier comparison with the temperature time series.

Figure 7

Spatial field correlation maps between selected radial lumen diameter sector chronologies and climate parameters for the period 1981–2015. Sector I chronology was correlated against (A) mean October to January temperature (C) October to January one-month Standardized Precipitation and Evaporation Index, and (D) October to January sea surface temperature. Mean chronology for sectors I-V was correlated against (B) November to January precipitation. The location of the study site is identified by a black dot. Colored areas show correlations significant at p<0.05.

Figure 8

(A) Normalized time series of earlywood lumen diameter (sectors I and I-V) plotted against previous October-current January Southern Oscillation Index (SOI) for the period 1966 to 2015 (bold lines are 5-year cubic smoothing splines). The sign of SOI has been inverted for easier interpretation. (B) Cross-wavelet transform of lumen diameter (sectors I and I-V) and SOI index. White contours represent 95% significance level. The relative phase relationship is shown as arrows (in-phase pointing right, anti-phase pointing left, and SOI leading lumen diameter by 90° pointing down). Results falling outside the cone of influence (white shaded area) might be distorted by edge effect.