The factors constraining the productivity of tropical dry forests remain a subject of debate. During the wet season, these forests may be limited by nitrogen or phosphorus. To evaluate the effects of nutrient limitation on productivity, we initiated a factorial nitrogen and phosphorus fertilization experiment in 2015. The experiment is located in diverse, mature forest stands in Guanacaste, Costa Rica. In parallel to this field experiment, we have been carrying out mechanistic model simulations to assess our process-level understanding of the biogeochemistry. In this talk, we identify the model processes that seem most important for capturing the field observations. As points of comparison, we target short-term (2-year) observed and simulated wood growth, root dynamics, arbuscular mycorrhizal fungi colonization, nodule count, soil respiration, microbial biomass, soil inorganic nitrogen, and soil inorganic phosphorus. For each of these metrics, we evaluate the importance of model process representations including phosphorus dynamics, symbiotic nitrogen fixation, microbial dynamics, and microbial functional groups. We then carry out multi-decadal simulations to understand whether the processes affecting the short-term dynamics also influence the longer-term dynamics.
Overall, our results underscore the importance of phosphorus dynamics for simulating the short-term response to fertilization. Observed increases in fine root biomass were only replicated in simulations that included phosphorus dynamics and that allowed for dynamic adjustment of plant allocation to fine roots. We also find that simulations including facultative symbiotic nitrogen fixation to be consistent with the observed increase in nodule count in phosphorus-fertilized plots. In the short term, we find that microbial dynamics do not strongly affect the comparison of simulation to experiments. However, in the longer-term simulations, our preliminary results indicate that microbial dynamics controlled the type of nutrient limitation that the ecosystem encountered. Simulations without microbial dynamics were ultimately subject to nitrogen limitation, but simulations including microbial dynamics were ultimately subject to phosphorus limitation. Our results indicate that tropical dry forest productivity can be constrained by both nitrogen and phosphorus, and that multiple processes dynamically regulate the net impact of nutrient limitation on productivity. Furthermore, processes affecting the short-term response to fertilization may be different from those affecting the longer-term response.