An automatic smart measurement system with signal decomposition to partition dual-source CO2 flux from maize silage

Abstract Carbon dioxide (CO2) is a principal byproduct of various chemical, biological and biochemical processes. CO2 has been measured using various advanced sensors, but a limiting technical challenge has been resolving independent streams of CO2 produced by processes occurring simultaneously. The magnitude and kinetics of each stream may be chemically, biologically or/and physically informative, but such partitioning has received little attention and successful case studies remain rare. In silage production, CO2 flux, an important indicator of aerobic deterioration, microbial activity or oxidative rate of silage, derives from two different pools: gas accumulated in the pores during the early anaerobic phase, and current real-time production of CO2 as oxygenated air enters the silage across the exposed face after opening for feed-out. The former is regarded as a noise confounding the signal and the latter reflects current degradation of the silage. Using a self-developed automatic sensor system with novel signal decomposition, we successfully partitioned CO2 flux into two independent streams derived from the two distinct pools in maize silage, following two independent processes: a physical venting of stored gas through a tortuous diffusive pathway, and a biochemical process generating gas in real time. Three silage samples, treated with a chemical or a biological additive, or left untreated, were tested. The signal decomposition found two best-fit functions (0.8034 ≤ R2 ≤ 0.9036), a quadratic CO2 discharge function and an exponential CO2 production function, for characterizing these distinct processes. These results demonstrate chemical sensor with powerful data-processing capability to resolve the complexity of dual-pool CO2 emission.