Study area and species

The study site known as Río de San José is located in the Eastern Sierra Madre, specifically in the municipality of Galeana, Nuevo León (Figure 1).

Figure 1 Location of the study area in the ejido Río de San José, in the municipality of Galeana, Nuevo León, Mexico. 

The predominant climate is semi-dry, semi-warm, characterized by cool winters. The mean annual temperature is 21.3 °C, the accumulated precipitation reaches 663 mm per year, with 58 % of this total recorded between the months of June and September (Figure 2) (^{Harris et al., 2020}). The predominant soil type in this region is Lithosol, characterized by its stony and poorly developed nature with a large amount of calcareous material (^{Schad, 2023}).

Río de San José (1 240 m). 1990 to 2022. 21.3 °C, 660 mm.

Figure 2 Walter-Lieth climate graph for precipitation and mean monthly temperature for the period 1990-2022. 

The study species was Pinus arizonica var. stormiae. This consists of medium-sized trees, 15 to 20 m tall. It forms pure stands or is commonly found mixed with Quercus L. species, Pinus engelmannii Carrière and P. strobiformis Engelm. It grows in moderately dry to mesic forests, with chalky soils and at altitudes ranging between 1 500 and 2 800 masl (^{Farjon et al., 1997}). It is mainly distributed in southern Nuevo León, as well as in some localities of southern Coahuila, Zacatecas, and San Luis Potosí (^{Farjon et al., 1997}).

Sampling and sample preparation

Two sites were selected with different northeast (NE) and northwest (NW) exposures. Five trees were selected from each site (Figure 1) with a normal diameter >20 cm (caliper, Haglöff^® Mantax Blue), good vigor, a long-standing appearance, and a straight, cylindrical stem without bifurcations, with the least possible disturbance. The coordinates (GPS, Garmin eTrex^® 25x) of each tree were recorded, and two growth cores were obtained for each tree with a Pressler increment borer (Haglöff^®, 5 mm) in opposite directions, always perpendicularly to the slope of the terrain (^{González-Tagle et al., 2023}). Radial increment cores were attached to wooden guides with white glue and allowed to dry at room temperature. Subsequently, they were sanded with sandpaper of successively finer grain (100:1 000) (^{González-Tagle et al., 2023}).

Chronology of ring width and basal area increment

The samples were visually dated using a stereomicroscope (AmScope^TM SM-4TPZ-B). Growth patterns were identified, and each ring was assigned its year of formation (^{Fritts, 2001}). The ring width was measured using the VELMEX^® measuring system (0.001 mm) and the Measure J2X software V5.03 (^{VoorTech Consoulting, 2021}). The quality of the visual dating was statistically evaluated by comparing the individual tree series with the master chronology of the species. This procedure was carried out with the COFECHA software (^{Grissino-Mayer, 2001}). A residual ring width chronology of Pinus arizonica was constructed. The ring width series were standardized to remove non-climate associated growth trends such as tree geometry and age to obtain ring width indices (RWI). An autoregressive model was applied to the generated indexes (^{Bunn, 2008}). Subsequently, the residual ring width chronology was calculated, using a robust average to determine the averages of the individual indices. Residual ring width indices were used to remove most of the temporal autocorrelation related to the previous year's growth (^{Cook and Holmes, 1986}).

The basal area increment (BAI) of Pinus arizonica was calculated using the following formula (^{Bunn, 2008}):

Where:

R = Radius of the tree without bark at the height where the core was extracted (1.3 m)

t = Year of ring formation

First, the BAI was obtained at the sample (core) level, and then the annual average BAI was calculated for each tree. The construction of the chronology and the calculation of the BAI was done with the dplR (Dendrochronology Program Library in R) package (^{Bunn, 2008}) of the R statistical software (^{R Core Team, 2021}).

Climatic data

Monthly climate data were obtained: mean, maximum, and minimum temperature, and total precipitation. The data were downloaded from the Climate Research Unit at http://climexp.knmi.nl/ (^{Harris et al., 2020}). Correlation coefficients were calculated between climate data and the residual chronology of the species for the 1920-2021 period, in order to evaluate the variation of the species radial growth sensitivity as a function of time for 10-year time windows: 1990-2000 (B1), 2001-2011 (B2), and 2012-2021 (B3).

Statistical analysis

The Mann-Whitney U test was utilized to determine whether there were significant differences in basal area increment (BAI) between the northeast and northwest exposures (^{Rubio-Cuadrado et al., 2021}). The Kruskall-Wallis test and a Dunn’s multiple comparison test (^{Hood et al., 2018}) were applied to assess the existence of significant differences between time periods. The choice of these analyses was based on the nature of the data, which did not show a normal distribution. Both were performed with the R-package ggstatsplot (^{Patil, 2021}).

Climate-radial growth association

The ring width chronology (Figure 3) shows the growth pattern of the 10 Pinus arizonica trees. The correlation between segments (0.38) was higher than the 0.328 (p>0.01) required by COFECHA, and the expressed population signal (EPS) obtained was 0.88. The first order self-correlation was 0.49, and the signal to noise ratio was 7.85. A low RWI is observed in 1996, 1998-1999, 2008-2009, 2019, and 2021, while high RWI is observed in 1993, 2007, 2010, and 2016.

The solid line represents the ring width index.

Figure 3 Ring width chronology of Pinus arizonica Engelm. var. stormiae Martínez for the 1920-to-2021 period. 

During the 1990-2021 period, significant correlations were found between climate variables and RWI. Precipitation in December prior to the growing year and in January and in April in the growing year were positively and significantly correlated with the RWI. The maximum temperature in December prior to the growing year and in January, February, and April in the growing year were negatively and significantly correlated with the RWI (Figure 4).

PP = Accumulated monthly precipitation; Tmax = Average maximum temperature; Tmin = Average minimum temperature. Significant correlations (95 %) are represented by red bars. Months in lower case correspond to the previous year of growth, and those in uppercase, to the current year of growth.

Figure 4 Growth-climate associations for the 1990-2021 period. 

Correlation analyses reveal significant temporal variations (Figure 5). Between 1990-2000, precipitation in January of the year of growth had a positive and significant influence on the RWI. In the following ten years, the RWI correlated positively and significantly with the precipitation of June in the year of growth.

PP = Accumulated monthly precipitation; Tmax = Average maximum temperature; Tmin = Average minimum temperature. Significant correlations (95 %) are represented by red bars. Months in lower case correspond to the previous year of growth, and those in uppercase, to the current year of growth.

Figure 5 Growth-climate associations between 1990-2000 (B1), 2001-2011 (B2), and 2012-2021 (B3). 

The RWI responds negatively and significantly to the Tmax of different months between periods. For the 1990-2000 period, it was to the Tmax for January and May of the year of growth. Between 2001-2011, to the Tmax of the previous December, as well as to the Tmax of February and July of the year of growth. In the period 2012-2021, to the Tmax of August of the year of growth. Tmin only negatively and significantly affected the RWI in the 2001-2011 period, particularly, the Tmin for January and February of the year of growth.

Effect of exposure on basal area increment

The Mann-Whitney U analysis revealed statistically significant differences in basal area increment between exposures (p<0.001) for the 1990-2021 period (Figure 6). Pinus arizonica trees with a northwest exposure showed a significantly higher mean BAI (224.7 mm² yr^-1) compared to trees exposed in the northeast direction (143.9 mm² yr^-1).

The yellow box shows the analyzed 1990-2021 period.

Figure 6 Comparison of the basal area increment (BAI) of Pinus arizonica Engelm. var. stormiae Martínez trees between the northeast (NE) and northwest (NW) exposures. 

The Kruskall-Wallis test showed significant differences in the basal area increment between growth time periods (p<0.001) (Figure 7). Dunn's multiple comparison test indicated that the BAI in the 2001-2011 (B2, 62.1 mm² yr^-1) and 2012-2021 periods (B3, 56.9 mm² yr^-1) did not show a significant difference between each other (p>0.05). However, both B2 and B3 exhibited significant differences (p<0.05) in BAI compared to the 1990-2000 period (B1, 51.2 mm² yr^-1).

B1 = 1990-2000; B2 = 2001-2011; B3 = 2012-2021.

Figure 7 Comparison between basal area increment (BAI) of Pinus arizonica Engelm. var. stormiae Martínez trees between growth time periods. 

Years that registered low RWI (Figure 3) agree with low site records of the Palmer drought index (PDSI) (^{Stahle et al., 2016}). They also coincided with the lowest records of the standardized precipitation-evapotranspiration index (SPEI) (^{Vicente-Serrano et al., 2010}) and the superposed epoch analysis (SEA) (Figure 8).

Figure 8 Results of the superposed epoch analysis (SEA) showing the significant influence of drought episodes on RWI of Pinus arizonica Engelm. var. stormiae Martínez. 

Climate-radial growth association

In this study, a significant correlation was found between the ring width index (RWI) of Pinus arizonica and climatic conditions, specifically precipitation and maximum temperature. A larger radial growth of the species was related to higher levels of precipitation in the months of December prior to the growth year and January and April of the growth year, while the maximum temperature in those same months and February reduced the RWI (^{González-Cásares et al., 2017}).

In the Western Sierra Madre, Pinus arizonica was found to be highly sensitive to climatic variables, particularly to the maximum temperature over a longer period, from the previous October to June of growth. Also, the minimum temperature in April was shown to have a significant limiting effect on the growth of this species (^{González-Cásares et al., 2017}). These differences in sensitivity are attributed to both the specific characteristics of the variety studied and the conditions of the site, which is characterized as a dry area throughout the year (^{Harris et al., 2020}).

In addition, the ring width indices of this species in northwestern Mexico has been found to exhibit a close connection with large-scale circulatory phenomena such as El Niño-Southern Oscillation (ENSO), the North Pacific Oscillation (NPO), and El Niño in region 3.4 (^{González-Cásares et al., 2017}).

Furthermore, differences were detected in the response of the RWI to climatic variables by time periods. During the 1990-2000 period, the January precipitation significantly influenced radial growth, while in the 2001-2011 period, it was the June precipitation that had a significant impact due to the occurrence of hurricanes and tropical storms in those years (^{NHC and CPH, 2023}).

As for maximum temperatures, a significant decrease in the radial growth of Pinus arizonica was observed during the 1990-2000 and 2001-2011 periods in response to three-month maximum temperatures. However, during the last study period (2012-2021), only the maximum temperature in August had a limiting effect on the RWI, indicating that the increase in temperature is affecting the growth of the species. On the other hand, the minimum temperature exerted a significant influence only during the period 2001-2011, exhibiting a negative association between the minimum temperature in January and February and the RWI of Pinus arizonica. These variations in radial growth sensitivity were related to changes in local temperatures over the three periods (^{Harris et al., 2020}).

Effect of exposure and time periods on the BAI

A greater basal area increment (BAI) occurs in response to topographic exposure; such behavior is documented in several studies as a consequence of the complex interaction of such factors as solar radiation and microclimate, which together generate diverse microhabitats (^{Vitali et al., 2017}). These microsites harbor conditions that result in differences in the basal area increment of the species. Elements such as temperature and the availability of water and nutrients (^{Lévesque et al., 2016}) impact dendrochronological indicators. However, the genus Pinus can develop well in poor and dry sites. It is important to consider the soil composition of the Regosol, the stoniness, and the low water storage capacity (^{Eilmann et al., 2013}). These conditions are accentuated by the topographic variation.

Different exposures result in specific conditions in limited geographical spaces. Although elevation and slope were not addressed in this research, they are topographic variables that influence radial climate-growth relationships (^{Pompa-García et al., 2017}). Moreover, another study conducted in a dry temperate forest with conifers documented that the heterogeneity and sparse vegetation in this type of forest increases soil temperature, favoring a high rate of evapotranspiration (^{Breshears et al., 1998}). Comparatively, the area in the Río de San José ejido region is sparsely vegetated and has a high spatial heterogeneity of individuals. These factors can lead to variations in tree growth and BAI across different topographic exposures within the study area.

A significant increase in the BAI of Pinus arizonica trees was observed in time periods B2 and B3 compared to period B1. This increase in BAI could be related to changes in environmental conditions during these periods. This is reflected in the climatic variables, especially during the 1990-2000 period, when less precipitation was documented in contrast to the other two time segments. In that decade, a mean annual precipitation of 556 mm was recorded, i. e., 159 and 167 mm less than during the periods B2 and B3, respectively. Furthermore, it was the only period that presented a dry summer month (Figure 7) (^{Harris et al., 2020}).

Limitations and prospects

The limited distribution of a vulnerable species in a drought-prone area reduced the availability of trees for sampling. Consequently, the limited number of trees analyzed could compromise the adequate representation of the intrinsic variability of the population. This implies that the small sample may not fully encompass the diversity of individual responses to climatic and topographic conditions. Moreover, the ability to identify temporal patterns and trends across time periods may be compromised with a reduced number of trees. This could affect the ability of the study to detect short-term responses to climatic events or disturbances. In order to strengthen the robustness of future analyses, it is recommended to increase the number of trees sampled, considering various exposures, as well as to evaluate other environmental variables that may influence the growth of the species, such as drought indices. However, it is important to note that, despite the limitations in the number of trees sampled, this research contributes significantly to the general knowledge of the species.