Quantifying relationships between plant functional traits and abiotic gradients is valuable for evaluating potential responses of forest communities to climate change. However, the trajectories of change expected to occur in tropical forest functional characteristics as a function of future climate variation are largely unknown. We modeled community level trait values of Costa Rican rain forests as a function of current and future climate, and quantified potential changes in functional composition. We calculated per-plot community weighted mean (CWM) trait values for leaf area (LA), specific leaf area (SLA), leaf dry matter content (LDMC), leaf nitrogen (N) and phosphorus (P) content, and wood basic specific gravity (WSG), for tree and palm species in 127 0.25 ha plots. We modeled the response of CWM traits to current temperature and precipitation gradients using generalized additive modeling. We then predicted and mapped CWM traits values under current and future climate, and quantified potential changes under a global warming scenario (RCP8.5, year 2050). We calculated the area within the multi trait functional space occupied by forest plots under both current and future climate, and determined potential changes in functional space occupied by forest plots. Overall, precipitation predicted CWM traits better than temperature. Models indicated increases in CWM SLA, N and P, and a decrease in CWM LDMC under climate change. Lowland forest communities converged on a single direction of change towards more acquisitive CWM trait values, indicating a change in forest functional composition resulting from a changed climate. Functional space occupied by forest plots was reduced by 50% under the future climate. Functional composition changes may have further effects on forests ecosystem services. Assessing functional trait spatial-gradients can help bridge the gap between species-based biogeography and biogeochemical approaches to strengthen biodiversity and ecosystem services conservation efforts.