You are here
Harvard Forest REU Symposium Abstract 2023
- Title: Validation of SIPNET Model’s Soil Respiration Forecasts using observations from Harvard Forest
- Author: Fernando Miguelena (University of Florida)
- Abstract:
Decades of research indicate that ongoing global climate change has been largely caused by anthropogenic effects, which can be attributed to methane and carbon dioxide emissions. Temperate forests such as the Harvard Forest, sequester immense amounts of carbon within the woody biomass and the soil. Thus, understanding the terrestrial carbon cycle is becoming increasingly important, as the components are responsible for the sequestration and flux (net movement over a specific area) of carbon. The implementation of short-term iterative forecasts allows for projections of the carbon fluxes across an ecosystem through a 35-day lead time. Compared to a centennial scale forecasts that are often used to predict long-term anthropogenic climate change, short-term forecasts can be readily validated and adjusted with concurrent data. In this study, we focused on validating the SIPNET model of soil respiration forecasts across Harvard Forest using soil collar data collected by the Picarro Mobile Gas Concentration Analyzer (Picarro G4301; Santa Clara, CA). Uncertainty was calculated using root mean square error (RMSE), mean absolute error (MAE), and bias. Preliminary results indicate that the SIPNET model’s instantaneous forecasts underpredict soil respiration by several magnitudes. With given preliminary results, we conclude that the SIPNET forecast can be improved through increased soil respiration observations to improve daily forecasts.
- Research Category: Soil Carbon and Nitrogen Dynamics; Ecological Informatics and Modelling
