Assessing the Impact of Parameter Uncertainty on Modeling Grass Biomass Using a Hybrid Carbon Allocation Strategy

Abstract

Grasslands play an important role in feeding the growing human population by supporting dairy and livestock production. The purpose of this paper was to improve the representation of carbon (C) allocation processes in ecosystem models, and to understand the sensitivity of simulated grass biomass to model parameters, providing guidance for future environmental change research. We used the Regional Hydro-Ecologic Simulation System to compare estimates of above and belowground biomass for C allocation strategies based on plant growth, resource use, and a newly developed HYBRID approach that integrates the first two allocation schemes. Observed values of biomass at seven grassland sites are used to compare performance of these three allocation strategies. Partial ranked correlation coefficients are then calculated to assess parameter importance and sensitivity under current and future climate conditions. We found that both resource use and plant growth play a role in simulating dynamic C allocation in grass biomass over time, indicating the need for the HYBRID approach to capture grassland C allocation. Parameter importance was site-specific under increased temperatures, enhanced carbon dioxide, and changed precipitation. Site-specific parameter importance also yielded insight into limiting resources (i.e., energy, moisture) at sites and subsequent effects on ecological processes like photosynthesis and allocation. This work increases our understanding of C allocation in ecosystem models related to grasses, and guides what model parameters may require further research.