Keywords
DAYCENT; model assessment; bioenergy sorghum
Start Date
25-6-2018 3:40 PM
End Date
25-6-2018 5:00 PM
Abstract
Intermediate complexity biogeochemical models are usually composed of several sub-models which are interrelated. Model assessment using a single treatment or measurement may not be adequate to fully examine the correctness of model mechanisms and has the risk of over-parameterization. In order to evaluate the simultaneous performance of all sub-models in DAYCENT, the model was parameterized using field measurements of soil temperature and water, aboveground biomass carbon (C), soil organic C (SOC), and carbon dioxide (CO2),and nitrous oxide (N2O) emissions from an 8-year field trial of bioenergy (biomass) sorghum with treatments of residue return, nitrogen (N) fertilization, and tillage. An overall satisfactory fit was obtained when comparing simulated outputs to measured data, with a Nash-Sutcliffe efficiency (NSE) range of 0.2-0.8. However, future model development directions were also suggested by model limitations. First, the model does not include a mechanism to represent increases in soil water holding capacity resulting from incorporated litter. This deficiency was indicated by limited yield differences among different residue return levels with sufficient fertilization. Second, the model’s sensitivity to the impact of soil moisture content on SOC decomposition did not accurately represent field observations, demonstrating a smaller SOC drop for early harvest seasons followed by high precipitation. Third, DAYCENT does not explicitly model SOC distribution throughout the soil profile, only to a conceptual 20-cm depth. This deficiency resulted in underestimation of plant yield and greenhouse gas (GHG) emissions in treatments with N limitation by potentially overlooking N mineralization beyond the conceptual soil depth.
DAYCENT assessment under multiple treatment-measurement scenarios in bioenergy sorghum production
Intermediate complexity biogeochemical models are usually composed of several sub-models which are interrelated. Model assessment using a single treatment or measurement may not be adequate to fully examine the correctness of model mechanisms and has the risk of over-parameterization. In order to evaluate the simultaneous performance of all sub-models in DAYCENT, the model was parameterized using field measurements of soil temperature and water, aboveground biomass carbon (C), soil organic C (SOC), and carbon dioxide (CO2),and nitrous oxide (N2O) emissions from an 8-year field trial of bioenergy (biomass) sorghum with treatments of residue return, nitrogen (N) fertilization, and tillage. An overall satisfactory fit was obtained when comparing simulated outputs to measured data, with a Nash-Sutcliffe efficiency (NSE) range of 0.2-0.8. However, future model development directions were also suggested by model limitations. First, the model does not include a mechanism to represent increases in soil water holding capacity resulting from incorporated litter. This deficiency was indicated by limited yield differences among different residue return levels with sufficient fertilization. Second, the model’s sensitivity to the impact of soil moisture content on SOC decomposition did not accurately represent field observations, demonstrating a smaller SOC drop for early harvest seasons followed by high precipitation. Third, DAYCENT does not explicitly model SOC distribution throughout the soil profile, only to a conceptual 20-cm depth. This deficiency resulted in underestimation of plant yield and greenhouse gas (GHG) emissions in treatments with N limitation by potentially overlooking N mineralization beyond the conceptual soil depth.
Stream and Session
Session: F3: Modelling and Decision Making Under Uncertainty