Environmental controls on carbon dioxide flux from black spruce coarse woody debris

Abstract

Carbon dioxide flux from coarse woody debris (CWD) is an important source of CO2 in forests with moderate to large amounts of CWD. A process-based understanding of environmental controls on CWD CO2 flux (R CWD) is needed to accurately model carbon exchange between forests and the atmosphere. The objectives of this study were to: (1) use a laboratory incubation factorial experiment to quantify the effect of temperature (T CWD), water content (W C), decay status, and their interactions on R CWD for black spruce [Picea mariana (Mill.) BSP] CWD; (2) measure and model spatial and temporal dynamics in T CWD for a boreal black spruce fire chronosequence; and (3) validate the R CWD model with field measurements, and quantify potential errors in estimating annual R CWD from this model on various time steps. The R CWD was positively correlated to T CWD (R 2=0.37, P<0.001) and W C (R 2=0.18, P<0.001), and an empirical R CWD polynomial model that included T CWD and W C interactions explained 74% of the observed variation of R CWD. The R CWD estimates from the R CWD model excellently matched the field measurements. Decay status of CWD significantly (P<0.001) affected R CWD. The temperature coefficient (Q 10) averaged 2.5, but varied by 141% across the 5-42°C temperature range, illustrating the potential shortcomings of using a constant Q 10. The CWD temperature was positively correlated to air temperature (R 2=0.79, P<0.001), with a hysteresis effect that was correlated to CWD decay status and stand leaf area index . Ignoring this temperature hysteresis introduced errors of -1% to +32% in annual R CWD estimates. Increasing T CWD modeling time step from hourly to daily or monthly introduced a 5-11% underestimate in annual R CWD. The annual R CWD values in this study were more than two-fold greater than those in a previous study, illustrating the need to incorporate spatial and temporal responses of R CWD to temperature and water content into models for long-term R CWD estimation in boreal forest ecosystems.