Stand stability is an important criterion for evaluating the condition of forests, but it is also quite difficult to assess. The objective of this study was to evaluate the assessment technique of stand stability applied in the Swiss National Forest Inventory (NFI). The Swiss NFI recorded various tree, stand and site attributes on sample plots at two inventory occasions, between 1983 and 1985 and between 1993 and 1995. The NFI defined the attribute "stand stability" as the probability that severe damage will occur to the considered stand within a fixed interval of time. So, the assessment of stand stability in NFI can be considered a risk evaluation. The assessment technique is based on evaluations performed by experts following specific assessment guidelines. First, the risk factors listed in the guidelines were rated regarding the potential hazard to the stand and then, the listed stand attributes were rated considering the potential resistance of the stand. Finally, the attribute "stand stability" was derived from these rates. Since the rating of both risk factors and stand attributes relies on the opinion of experts, this technique runs a risk of resulting in a rather subjective evaluation. This circumstance affects the interpretation of stand stability and in particular, makes the verification of changes of stand stability very uncertain. An examination of the assessment technique, thus, is especially relevant as regards the attribute "stand stability".
The study was focused on the reproducibility and relevance of this attribute. Ln particular, it investigated the concordance between stand stability as evaluated at the first inventory occasion and the damage occurred afterwards, between the first and the second occasion. It also explored whether stand stability can be related to more objective and reproducible variables. The variables providing the best explanation of stand stability were selected by means of regression models.
As regards reproducibility, the evaluations of stand stability resulting from two independent surveys did not differ significantly, but stand stability proved to be an unprecise variable. On the other hand, stand stability as evaluated at the first inventory occasion resulted to accord with the level of damage occurred afterwards to the stands. Some of the variables assessed in NFI were related to the attribute stand stability, but only weak correlations were found. The results suggest that the attribute stand stability can be considered a valid indicator for the risk of damage, but should be used and interpreted only on certain conditions.