Ecological stress memory in wood architecture of two Neotropical hickory species from central-eastern Mexico

Drought periods are major evolutionary triggers of wood anatomical adaptive variation in Lower Tropical Montane Cloud Forests tree species. We tested the influence of historical drought events on the effects of ecological stress memory on latewood width and xylem vessel traits (vessel density, vessel grouping index, hydraulic diameter, and percent conductive area) in two relict hickory species (Carya palmeri and Carya myristiciformis) from central-eastern Mexico. We hypothesized that latewood width would decrease during historical drought years, developing correlations between growth and hydric stress conditions, and that during past tree growth, moisture stress would imprint on wood anatomical traits between successive drought events. We analyzed latewood anatomical traits that developed during historical drought and pre- and post-drought years in both hickory species. We found that repeated periods of hydric stress left climatic signatures for annual latewood growth and xylem vessel traits that are essential for hydric adaptation in tropical moist tree species. Our results demonstrate the existence of cause‒effect relationships in wood anatomical architecture and highlight the ecological stress memory linked with historical drought events. Thus, combined time-series analysis of latewood width and xylem vessel traits is a powerful tool for understanding the ecological behavior of hickory species.

Methods
The Ring-Width Index (RWI), earlywood-width (EWw) and latewood-width (LWw) of hickory species were measured separately to obtain intra-annual climate signals from the tree-ring parameters by the presence of distinct narrow to wide earlywood conduits in a single intermittent row, and medium to narrow solitary and radial multiples of latewood conduits from two to three (https://www.wood-database.com) were used as phenological indicators to determine the EWw and LWw boundaries. Growth rings (EWw and LWw) were measured under a stereoscopic microscope (Olympus SZ61, Olympus Corporation, Center Valley, PA, USA) and with a Velmex Tree Ring Measuring System (Velmex, Bloomfield, NY, USA) with 0.001 mm accuracy using TSAP-Win v. 4.67c. The ring-width time series were visually and statistically cross-dated using the software TSAP-Win and COFECHA. To obtain the average of the detrended RWIs, EWw, and LWw, we standardized the raw ring-width series using the ARSTAN program to remove non-climatic trends. A cubic smoothing spline with a 50% frequency cutoff at 30-year intervals was used to perform detrending on each series, which preserved high variance at a frequency equal to two-thirds of the length of each series. The individual detrended tree-ring series were then averaged to create mean site chronologies by computing the bi-weighted robust mean. To assess the reliability of the site chronologies, we used the expressed population signal (EPS > 0.85) and interseries correlation.