ABSTRACT
This chapter demonstrates how mathematical models and data-model fusion can be used to identify the processes controlling carbon (C) dynamics. A suite of process-based models was built to describe the C dynamics in a temperate lake. The models were built so as to create a factorial modeling experiment aiming to identify the processes that contributed most to explaining the observed variation in the observed data. All models were calibrated using the Bayesian Markov Chain Monte Carlo algorithm and their fit indices were used to identify the processes that contributed most to model improvement. Although the example used in this lecture focuses on C dynamics in a temperate lake, the presented hypothesis-testing algorithm is transferable to other response variables and ecosystems.
