Validation of the Single Tree Stand Growth Simulator Prognaus with Permanent Plot Data : Caring for the Forest: Research in a Changing World. Statistics, Mathematics and Computers.

Birmensdorf

1996

S. 36-50

19 Lit. Ang.

Bandaufführung

77332

Our objective is to examine the validation of the basal area increment model in the individual tree stand growth simulator, PROGNAUS. The simulator is designed to forecast the development of both pure even-aged and mixed-species uneven-aged stands in Austria using distance-independent individual tree methodology. The primary model is for basal area increment (BAI), which is predicted from size (diameter, crown ratio), competition (basal area of larger trees, crown competition factor), and site descriptors. Because the model must be able to simulate the development of both uneven- and evenaged stands, site index and age are intentionally not used as predictors. Available for testing is an independent data set of 22 permanent plots in mixed Scots pine (Pinus sylvestris L.) - Norway spruce (Picea abies L.) stands in the Austrian part of the Bohemian Massif, northeastern Austria. The validation plots were remeasured for three 5-yr periods between 1977 and 1992. Observed BAI increments and those projected by PROGNAUS exhibited squared coefficents of correlation (r2) of 0.31- 0.36 for Scots pine and 0.63-0.69 for Norway spruce using the validation data. This compares rather well with the multiple R2 of the original model using the Austrian Forest Inventory data, which was 0.33 for Scots pine and 0.58 for Norway spruce. A plot-specific adjustment of the model's intercept using past increment did not improve predictions ofr Norway spruce, but did improve predictions for Scots pine. For Scots pine, this means that important site- specific variation not captrued by the model can nevertheless be accounted for by using increment calibration. A time trend towards increasing underestimation of increment was detected in the later growth periods of the validation data. This trend, possibly caused by changing weather contitions and/or nitrogen deposition, only affected the intercept and thus was not site-specific. For future model development, as many growth periods as possible should be used to parameterize increment models, although even this cannot account for a long term change in site potential.