PARENT SESSION
Poster Session 22: Biogeography
Wednesday, August 10, 5:00 PM - 6:30 PM, Exhibit Hall 220 A-E, Level 2, Palais des congrès de Montréal

Carbon accumulation at the Harvard Forest: A comparison of methods for measuring tree biomass for regional extrapolation of the eddy-flux tower footprint.

McKain, Kathryn^*,1, Hammond Pyle, Elizabeth², Wofsy, Steven ², Bubier, Jill ¹, ¹ Mount Holyoke College, South Hadley, MA² Harvard University, Cambridge, MA

ABSTRACT- Forests in the Northern Hemisphere are estimated to sequester 1 - 2 PgC/yr, about 25% of the carbon that is emitted by fossil fuel combustion each year. The eddy-flux tower at the Harvard Forest has produced the longest running record of net-ecosystem carbon exchange in a North American forest. Though the tower carbon data allows for extensive temporal extrapolation, its spatial relevance depends on our ability to scale the 1 km² tower footprint to the surrounding region. To examine the larger area around the tower, we investigated two sets of previously installed plots for measuring tree biomass: fixed-radius tower footprint plots and more extensively placed variable-radius BigFoot plots. An initial comparison of the BigFoot and tower plots revealed that while the tower plots yielded an average of 108±33 MgC/ha as of 2002, the BigFoot plots yielded an average of 72±26 MgC/ha. The different methods employed by the two groups confounded this result, and it was unclear if the measured differences reflect true spatial differences in forest composition. A resurvey of the Big Foot plots in 2004 using both variable- and fixed-radius plot designs reveals that the two methods did not yield equivalent biomass estimates. In 2004, the variable-radius plots yielded a biomass estimate of 85±33 MgC/ha, a significant increase over the 2002 values (p-value = 0), yet still significantly lower than the 2004 fixed-area estimate of 105±38 MgC/ha (p-value = 0). The large increase in estimated biomass with the variable-radius sampling method is attributed to a 26% recruitment rate between 2002 and 2004, which may result from differences in investigator judgment about tree inclusion or exclusion during variable-radius plot sampling. The disparity in biomass estimates from fixed- and variable-radius plots appears to result from different tree density estimates: While variable-radius plots showed an average of 951±380 trees/ha, fixed-radius plot gave a significantly larger average of 1016±330 trees/ha (p-value = 0.0353). Comparisons of the extensively placed fixed-radius plots with those in the tower footprint indicate that the forest composition of the tower footprint is representative of the region in terms of species composition, stem count, basal area, and biomass.

Key words: carbon, forest, region, methods