We planted 30 seedlings of 13 shade-tolerant tropical montane cloud forest tree species in eight forest sites (3120 seedlings in total) along an elevation gradient (1250 to 2429 m a.s.l.) in central-eastern Mexico. We recorded sapling survival and relative growth rate in height (RGRh) and diameter (RGRd) after eight years. We tested the survival and growth response for all species together to the effect of Climatic Transfer Distance (CTD; the difference between the historical climate at the seed source and the current climate at translocation sites) in terms of mean annual temperature (CTD_MAT), maximum and minimum annual temperatures (CTD_Tmax and CTD_Tmin), mean annual precipitation (CTD_MAP), and climate moisture deficit (CTD_CMD). We also tested the sapling response to the climatic variables at the translocation site. We found evidence to support the decline, although slight, in tree sapling survival and growth across tree species with increasing CTD, thus supporting the hypothesis of a reduction in performance with increasing distance between the climate to which tree species are adapted (historic climate of seed origin) and that at the new translocation sites. Our results support a higher mortality risk caused by increasing CTD in MAP and MAT and a decline in growth caused by the increasing CTD in MAT, Tmax, MAP, and CMD. Although high variation occurred among species, this general pattern still emerged and was consistent for all the climatic variables.

A total of eight forest sites were selected to establish the plantings in field common gardens along an elevation gradient (1250 to 2429 m a.s.l.) in Veracruz, Mexico. In each site, a 50 × 55 m plot was delimited, fenced, and 30 seedlings of each of 13 tropical montane cloud forest tree species were transplanted in May-June 2015. Species were distributed at random within each plot and the individuals were planted ~2.6 m apart. A ~1 m radius around each seedling was weeded at planting time, and again after three, six and 12 months. Survival, height, and diameter at the base of all plants were recorded immediately after planting and after 8 years. The relative growth rates in height and diameter (RGRh and RGRd, respectively) were calculated following Hunt (1982). 

To test the climate sensitivity of the young tree stages, we selected the following five climate variables: mean annual temperature (MAT), mean annual precipitation (MAP), mean minimum temperature in the coldest month (Tmin), mean maximum temperature in the warmest month (Tmax), and Hargreave’s Climate Moisture Deficit (CMD). The climate data for each seed source (average 1961-1990) and translocation site (average 2015-2022) were downloaded from the ClimateNA data portal (https://climatena.ca/), based on Wang et al. (2016). For each georeferenced location of a tree species´ seed source (seed trees), we extracted the climatic variables for the reference period 1961 – 1990, assuming that this represents the climate to which the seed trees were adapted. We refer to this 1961 –1990 period as the “historic climate”. For the field translocation sites, climate data were obtained for the period in which the experimental transplants were actually grown (2015–2022), and this is referred to as the “contemporary climate”. To measure the impact of the difference between the historic and contemporary climates at each planting site, we estimated the Climatic Transfer Distance (CTD), sensu Leites et al. (2012) as: CTD = (contemporary climate at the test site) – (historic climate at the seed source).

To evaluate the power of CTD as predictor of tree survival (binomial response variable), a Generalized Linear Mixed Model (GLMM; binomial family and logit link function) was used. CTD was included as a fixed-effects term and tree species as a random-effects term. In each full model, we also included the quadratic term for CTD considering that the optimum (highest survival) was expected to occur at the CTD with the value of zero (or closest to zero), and a decline would occur in either direction from the zero, describing a downward facing parabola. We included the climatic variables at the translocation site that were least correlated with the CTD variable (R < 0.7) to fit the complete models. To fit the GLMM, we ran the function glmer with the package lme4 in R Statistical Environment (version 4.2.2; (2022-10-31)(R Core Team, 2022). To evaluate the relative growth rate in height (RGRh) and diameter (RGRd) as a function of CTD, a Linear Mixed Model (LMM) with a normal error distribution was used. We also included the quadratic term for the CTD. The full models also included the climatic variable at the translocation site that was least correlated with the CTD variable. The function lmer of the package nlme was used to fit the LMM (Cayuela Delgado & De La Cruz, 2022).

References

Cayuela Delgado, L., & De La Cruz, R. O. T. (2022). Análisis de datos ecológicos en R. México: Ediciones Mundi-Prensa.

Hunt, R. (1982). Plant growth curves. The functional approach to plant growth analysis. London, UK: Edward Arnold Ltd.

Leites, L. P., Rehfeldt, G. E., Robinson, A. P., Crookston, N. L., & Jaquish, B. (2012). Possibilities and limitations of using historic provenance tests to infer forest species growth responses to climate change. Natural Resource Modeling, 25(3), 409-433. doi:10.1111/j.1939-7445.2012.00129.x

R Core Team, R. (2022). R: A language and environment for statistical computing. R Foundation for Statistical Computing. Vienna, Austria.URL https://www.R-project.org/.

Wang, T. L., Hamann, A., Spittlehouse, D., & Carroll, C. (2016). Locally Downscaled and Spatially Customizable Climate Data for Historical and Future Periods for North America. Plos One, 11(6), 17. doi:10.1371/journal.pone.0156720