Assessing different sampling methods to develop forest carbon budgets

A complete forest carbon budget is built up from an accurate estimates of carbon pools in the forests. However, the methods for determining belowground carbon pool vary considerably in forest ecosystems. Here, we present results of several methods used to develop carbon budgets for four forests in the Sierra Nevada of northern California. Soil carbon is calculated as the product of three separate measures: 1) the proportional volume of the sample that is actually fine soil (not rocks, roots, or woody debris), 2) the fine soil bulk density, and 3) the fine soil carbon concentration. As carbon concentration is negatively correlated with bulk density, using the product of mean forest values of carbon concentration and bulk density produced estimates 10 to 13 % higher than using the product from individual samples and then averaging. We sampled soils using mini pits 15-cm in diameter and these larger samples reduced site to site variability of soil carbon estimates by 20 % over core samples. The pits also produced fine soil bulk density estimates that were 24 % higher than core samples. In the lab, we separated soil components from a single sample to obtain rock content, live roots, dead roots, woody debris, and fine soil bulk density and fine soil carbon content. We found that the commonly used LiCor 6200 or 6400 with soil respiration chamber produced annual estimates of soil CO2 efflux that were 15 % higher than what we believe our carbon budgets can support. Reducing fan speeds reduced measured efflux rates by over 25 % in our highly porous soils. Input of woody debris was efficiently estimated using transects on the forest floor that are cleared of woody debris and periodically resampled. These results suggest that without an appropriate method, forest carbon pools could be over- or under-estimated.