Modelling diameter increment in a lowland evergreen rain forest in French Guiana
Since 1984, the diameter increment of nearly 50 000 trees 10 cm DBH or over has been monitored every year in 12 long-term experimental 6.25 ha plots of the Paracou experimental station, French Guiana. In order to study and compare forest dynamics in disturbed and undisturbed stands, contrasted silvicultural treatments were performed between 1986 and 1988 in these plots. Taking advantage of this experimental design and of this large database, a forest dynamics modelling project was started in 1992. This paper aims at presenting the growth component of the single-tree, distance-dependent model that was developed within the framework of this project. This individual diameter growth model was built according to the 'potential x modifier' paradigm: it links tree annual diameter increment to variables describing tree size and instantaneous competition as well as the recent past evolution of the competition. Species information was also taken into account and resulted: (i) in the classification of tree species into 15 functional groups that are homogeneous for diameter growth; and (ii) in a definite improvement of model efficiency. Attempts to further reduce the residual variance through the use of soil and topographical information remained unsuccessful.
| Summary: | Since 1984, the diameter increment of nearly 50 000 trees 10 cm DBH or over has been monitored every year in 12 long-term experimental 6.25 ha plots of the Paracou experimental station, French Guiana. In order to study and compare forest dynamics in disturbed and undisturbed stands, contrasted silvicultural treatments were performed between 1986 and 1988 in these plots. Taking advantage of this experimental design and of this large database, a forest dynamics modelling project was started in 1992. This paper aims at presenting the growth component of the single-tree, distance-dependent model that was developed within the framework of this project. This individual diameter growth model was built according to the 'potential x modifier' paradigm: it links tree annual diameter increment to variables describing tree size and instantaneous competition as well as the recent past evolution of the competition. Species information was also taken into account and resulted: (i) in the classification of tree species into 15 functional groups that are homogeneous for diameter growth; and (ii) in a definite improvement of model efficiency. Attempts to further reduce the residual variance through the use of soil and topographical information remained unsuccessful. |
|---|