SYMP 23-4
Tree carbon allocation dynamics using a carbon mass balance approach

Friday, August 15, 2014: 9:40 AM
Gardenia, Sheraton Hotel
Tamir Klein, University of Basel, Basel, Switzerland
Günter Hoch, Institute of Botany, University of Basel, Basel, Switzerland
Background/Question/Methods

The dynamics of carbon (C) allocation in trees are fundamental to their functioning as well as to forest C stocks, with major implications on the terrestrial C cycle. A full description of tree C allocation dynamics must account for all C fluxes and pools in a tree. These include compartment- and compound-specific C pools and the internal C fluxes between them, which are difficult to measure directly. On an annual time-scale, healthy trees maintain a C mass balance between source (assimilation) and sinks (respiration, growth, export to soil, and litter). However on shorter time-scales, storage (starch synthesis) and cash (starch degradation) can buffer source-sink imbalances, and hence must be included in the analysis. 

In this study we collected independent measurements of all tree C inflows, outflows and pools (both at the compartment and compound level) from Pinus halepensis in a semi-arid forest, into a dataset where all fluxes were converted to g C tree-1 d-1. Using this dataset, a simple flow model was created to describe and quantify the long- and short-term tree C balances.

Results/Conclusions

The annual C source of 24.5 kg C tree-1 year-1 was balanced by C sinks of 23.5 kg C tree-1 year-1, which, in turn, partitioned into 70%, 17%, and 13% between respiration, growth, and litter (plus export to soil), respectively. On the monthly time-scale, large imbalances (up to 57 g C tree-1 d-1) were observed as C excess during the rain season, and as C deficit during the dry season. Our model showed that concurrent changes in C reserves (starch) were sufficient to buffer these transient C imbalances.

The C pool dynamics simulated in the flow model were in good agreement with the observed pool sizes. This provides confidence to our estimations of the timing, magnitude, and direction of the intrinsic transport fluxes simulated by the model. These intrinsic fluxes indicated small (6 g C tree-1 d-1), yet persistent, relocation flows from roots to the stem, and from the stem to the leaves, during July-August.