Spatial heterogeneity of soil CO2 efflux after harvest and prescribed fire in a California mixed conifer forest

• We assessed the effects of forest management on spatial variability of soil CO2 efflux.
• We compared a control, a harvested and a burned stand in a mixed conifer forest.
• Spatial variability was increased by disturbance, especially by the clear-cut.
• Harvest reduced the ability to quantify soil CO2 flux and detect treatment effects.

Spatial variability is a key factor when quantifying soil CO2 efflux and punctual measurements need to be extended to larger stand, ecosystem, or regional scales. Spatial variation also affects comparisons among ecosystems, as when quantifying effects of disturbances on ecosystem carbon dynamics. However, spatial variability of soil CO2 efflux is still unknown and difficult to predict. We quantified the effects of silvicultural practices (prescribed fire and harvesting) on spatial variability of soil CO2 efflux in a mixed conifer forest from the central Sierra Nevada in California, USA. Soil CO2 efflux was measured using a portable chamber system, on 20–29 locations in four treatment sites: an untreated control, a prescribed fire site (burned in 2002 and 2009) and two clear cut sites harvested in 2010. In one of the harvested sites the soils were mechanically ripped to reduce soil compaction, a common practice done on industrial timber forest lands in the Sierra Nevada. Results showed that disturbance increased spatial variability of soil CO2 efflux. Coefficient of variability increased from an annual average of 32% at the control site to 37% at the burned site, and 49–51% at the harvested sites (without and with soil ripping, respectively), mirroring post-disturbance increases in spatial variability of soil temperature and soil water content. Because of the post-harvest increase in spatial variability, the ability to detect differences became lower, and the number of samples needed to obtain a value representative of the full population mean (within a 10% range) increased by 100%, from 60 to 120 samples. To reduce uncertainty in our soil CO2 efflux treatment estimates, more than 10–15 randomly selected locations per study site were necessary. Spatial variability of soil CO2 efflux at our sites was not affected by distance between measurement locations, was correlated to fine root and litter biomass at the control site, negatively correlated to soil bulk density at the fire site, and un-correlated to soil temperature and water content at all sites. The increase of spatial variability in soil CO2 efflux after disturbance and the requirement for a sufficient number of measurement locations should be considered when quantifying carbon dynamics of disturbed ecosystems, or assessing effects of different forest management practices.