PARENT SESSION
Poster Session #42: Carbon Storage.
Thursday, August 9, 2001. Presentation from 3:30 PM to 5:00 PM. Exhibition Hall

Scaling up estimates of tree biomass from plot measurements of forest inventories to regional domains using a process-based forest ecosystem model .

Pan, Yude¹, Birdsey, Richard¹, Hom, John¹, McCullough, Kevin¹, ¹

ABSTRACT- Forest inventory-based studies and process-based ecosystem models are two important approaches for estimating regional forest biomass. However, the inventory studies have difficulties in distinguishing the amount of biomass accumulation attributable to individual mechanisms or processes. The process-based ecosystem models can predict effects of climate or atmospheric chemistry on biomass accumulation, but often result in higher estimates because the parameterization in the models relies on limited field data from less disturbed sites. In this study, we tried to combine plot estimates of tree biomass from forest inventories (FIA) and synthesis via process-based forest ecosystem model PnET-CN. The comparison between the FIA data and the model predictions had shown a systematic bias and overestimate of biomass by the model. To solve the problem, we chose a large number of FIA data from relatively mature sites as target values for tuning the two most sensitive parameters in the model, i.e. foliage nitrogen and water holding capacity. A statistical equation was developed based on calibrated foliage N values and environmental conditions and replaced the constants previously used in the model. A simple linear equation accounting for altitudinal effect on runoff was applied in the model to adjust the values affected by water holding capacity data. The modified model significantly improves the accuracy of the biomass prediction, verified by the independent FIA data. The model can scale up tree biomass from plot measurements to regional domains and makes estimate of forest biomass more realistic at regional scale.

KEY WORDS: forest biomass, forest inventories, process-based ecosystem model, scaling up plot measurements