Altitude and community traits explain rain forest stand dynamics over a 2370‐m altitudinal gradient in Costa Rica

Abstract

AbstractStand dynamic rates in Earth's vitally important tropical mountain rain forests are a central component of the response of these forests to global change, but their relationships to environmental gradients are poorly understood. We worked in mature forests on a 440–2810 m asl altitudinal gradient on Costa Rica's Talamanca Cordillera, making five censuses of 29 0.25‐ha permanent sample plots during 2012–2019. We determined mortality and recruitment rates, as well as basal area (G) increments, for individuals ≥10 cm diameter at breast height. Our main hypothesis was that stand dynamic rates decrease with altitude (therefore temperature); we also tested the hypotheses that rates increase with plot community‐weighted mean specific leaf area (CWM SLA) and decrease with CWM wood specific gravity (WSG). We developed regressions using generalized additive models to test our hypotheses. Mortality and recruitment rates decreased with altitude in line with our main hypothesis, though the strongly non‐linear mortality trend may be driven by extreme weather and temperature. Also, the best model for recruitment rates included plot CWM SLA with a negative relationship. Gross basal area increment ∆Ggross, the annual basal area increment of trees that survived the study period, had a hump‐shaped relationship to altitude, possibly related to the low CWM WSG of mid‐altitude forests. ∆Ggross was indeed negatively related to CWM WSG. However, net basal area growth (∆Gnet, the annualized difference between initial and final plot basal areas) was positively related to altitude due to low mortality in montane forests, in a model with CWM WSG with a negative relationship. ∆Gnet was negative in nine of the 29 plots at <1500 m asl. Patterns of rain forest stand dynamics on this long altitudinal gradient go beyond a direct response to temperature, and further work is required to improve scenarios of forest response to climate change. CWM traits should be measured to improve understanding. The effects of storms and lightning on low‐ and middle‐altitude forest and the potentially high resilience of montane Fagaceae‐dominated forests require attention. Diversity, not generality, should be expected on comparing tropical forest altitudinal transects.