COS 5-6 - Model validation and sensitivity analysis of tropical dry forest response to nutrient fertilization

Background/Question/Methods

Soil nutrients, especially nitrogen (N) and phosphorus (P), have long been recognized to regulate plant growth and hence influence carbon fluxes between the land surface and atmosphere. However, how forests adjust allocation in response to N and/or P fertilization remains puzzling. Recent work in Costa Rica and the Amazon suggest that trees increase fine root production under P fertilization while N fertilization might have no effects or reduce fine root production, but the mechanisms have not been clearly identified. In order to better understand mechanisms governing nutrient effects on plant allocation, we used the nutrient-enabled ED2 model to simulate a full factorial N and P fertilization experiment conducted in a secondary tropical dry forest in Costa Rica over three years (2015-2017). We evaluated how different allocation schemes affect model performance. These schemes included static allocation to fine roots, positive correlation between soil nutrients availability and fine root production, and negative correlation between soil nutrients availability and fine root production. Contrasting allocation schemes were also investigated for their community-level influences, including the impacts on plant functional type composition and the impacts on long-term carbon stock.

Results/Conclusions

Under all schemes, the model simulated total biomass productivity and mortality in the control plots that was similar to the observations. Simulated allocation to leaves, wood and fine roots was also similar to observations in the control plots. In the P fertilized plots, observed fine root productivity was larger than in the control plots. This dynamic was only captured by the model if we assumed there is a positive correlation between soil nutrients availability and fine root production. Like the observations, the model displayed little response of productivity to N fertilization, suggesting the applied N was quickly lost to denitrification and leaching. Contrasting allocation schemes did not dramatically change plant total productivity, but had impacts on productivity of leaves, wood and fine roots. On 30-year time scale, simulations with dynamic allocation schemes also predicted lower carbon stock than simulations with static allocation. When fine root production is negatively correlated with nutrient availability the model estimated lower biomass because of lower productivity. When fine root production is positively correlated with nutrient availability, reduced biomass production was mainly driven by increased mortality. Given the sensitivity of biomass to allocation schemes, further model-data comparison of fine root production is recommended.