Chemical assumptions in input designation and parameter determination in the DayCent-Chem model

St. Brice, Lois

Following taken from abstract: 'Ecosystem scientists use computer models to predict the effects of atmospheric deposition on terrestrial and aquatic ecosystems. Models require a set of unique input variables which are used to predict outputs, and the relationship between the two is characterized by various parameters. Many ecological models represent heterogeneous ecosystems as momogeneous entities and so require that simplifying assumptions be made about the ecosystem being studied. Additionally, assumptions are made in determining the parameters which define the relationships between these inputs and outputs. The assumptions made about inputs and parameters affect the degree to which model preditions vary from observed values for variables in the field. Sensitivity analysis seeks to identify and quantify the effects that changes in input variables and model parameters have on model output. In this study, DayCent-Chem (a coupled, non-spatial biogeochemical model) was initialized for Hadlock Brook Watershed in Acadia National Park. Sensitivity analysis was used to explore the relationship between chemical assumptions made, both in the representation of model processes (through parameter determination) and in the designation of the model inputs, on the model's predictive capability. The model was found to be sensitive to the five categories of manipulations carried out, being most sensitive to cation exchange capacity manipulations followed by mineral denudation, texture, partial pressure of CO2 and reaction database manipulations, in that order.' Research recommendations: 'Further study of the texture dependence of hydrology represented in the model can help us determine if the addition of groundwater contributions to discharge are necessary. One important model revision may be the creation of a terrestrial biotic sink for base cations.' /STUDY LOCATIONS:/ /MOUNT DESERT ISLAND/Hadlock Brook/