The present report focuses on the results of the transnational and national crown condition surveys of the International Co-operative Programme on Assessment of Air Pollution Effects on Forests (ICP Forests) of the United Nations Economic Commission for Europe (UN/ECE) and of the European Union Scheme on the Protection of Forests against Atmospheric Pollution established by Council Regulation (EEC) No 3528/86. National survey results are presented from 33 European countries. These are based on 20 345 sample plots with 434 457 sample trees. The plots cover about 130 million hectares of forests. 29 of these countries have also submitted results of the 16x16 km grid (transnational survey). The results indicate that crown condition, having been assessed as a visible expression of forest condition for 11 years, hat deteriorated over this period and continues to decline, expecially in central Europe. According to the results of the transnational crown condition survey, 25.1 % of the total sample of around 116 200 trees were defoliated by more than 25 % and were thus classified as damaged. As regards the two main species groups, 25.0 % of the total broadleaves and 25.2 % of the total conifers were damaged. The most valuable feature of the survey is to provide time series of data. Whilst forest condition has clearly deteriorated over the years in Europe as a whole there were improvements at the regional level and in different species and age groups. In order to be able to trace the development of forest condition over several years without distortion due to annual differences in the composition of samples, special tree and plot samples which were common to certain survey years were analysed. Such common samples were determined for the periods 1995-1996, 1992-1996 and 1988-1996. The share of damaged common sample trees of 1995-1996 increased only slightly within that period from 25.4 % to 25.6 %. Of the individual regions, the largest increase (from 7.5 % to 15.7 %) occurred in the Atlantic (south) region. Expressed in terms of mean defoliation, for the most regions a small, but significant worsening was found. An improvement took place only in the sub-atlantic region, where the mean defoliation dropped from 25.7 in 1995 to 24.6 in 1996. This, however, was not significant and was mainly a consequence of the recovery of Pinus sylvestris in eastern Germany and northern Poland, and partly due to the weather conditions in 1996. The greatest increase in defoliation was found for Castanea sativa and decidous Quercus spp., in which the share of damaged trees rose from 15.5 % to 22.9 %, and from 30.6 % to 35.6 %, respectively. Defoliation of Quercus suber decreased significantly (from 25.2 % to 14.8 %). Among the conifers, Abies spp. continued to deteriorate, with defoliation increasing from 31.8 % to 34.6 %. A less pronounced deterioration occurred in Picea spp., whereas Larix spp. and Pinus spp. improved slightly. In the subsamples of the common trees of the periods from 1988 and 1992 to 1996, the development of defoliation in 8 and 13 species (respectively) was analysed. In the total common tree sample of 1992-1996 an overall increase in damaged trees occurred, from 22.4 % in 1992 to a maximum of 26.9 % in 1995, followed by a slight recovery to 25.6 % in 1996. Pinus sylvestris, which represents the largest share of the common sample of 1992-1996, showed no obvious trend. The share of damaged trees lies between a maximum of 32.1 % in 1994 and a minimum of 23.0 % in 1996. Statistical testing and spatial analysis showed that crown condition worsened significantly on only 16.1 % of the Pinus sylvestris plots, whereas 28.5 % of the plots showed an improved health status between 1992 and 1996. This improvement was an effect of the already mentioned recovery in the Sub-atlantic region. Picea abies constitutes the second largest share of the common sample of 1992-1996. It shows a steady increase in the proportion of damaged trees from 27.4 % in 1992 to 33.3 % in 1996. Among the geoclimatic regions the most alarming increase in defoliation occurred in the Sub-atlantic region (from 38.6 % in 1992 to 47.0 % in 1996). Fagus sylvatica, constituting the biggest share of common trees in the Sub-atlantic and the Mountainous (south) regions, shows a clear increase in percentage of damaged trees. Between 1992 and 1996 the share of damaged trees increased rather moderately, but at the same time the proportion of healthy trees dropped rapidly from 48.4 to 34.9 %. The decrease in the share of trees with full foliage between 1992 and 1996 is confirmed by the common sample of 1988-1997. Quercus robur shows a steady deterioration of crown condition between 1992-1996. It is most represented in the Sub-atlantic region, where it also shows the highest defoliation. In prising the second largest proportion of Quercus robur, the crown condition worsened. A comparable development, i.e. increase in defoliation, was observed in the common sample of 1988-1996. Although crown condition assessment at the large-scale was not designed for cause-effect analyses, certain relationships between the temporal variation and spatial patterns of crown condition and site factors can be established. As in previous years, in 1996 biotic stressors, climatic events and air pollution were reported as the main causes for forest damage. Acute effects clearly attributable to air pollution were rarely found. However, according to the results of several national surveys, particularly in the main forest damage areas of some countries, air pollution is considered as a predisposing, accompanying and locally triggering factor weakening forest ecosystems. With respect to cause-effect relationships, knowledge of the temporal development and the spatial patterns of crown condition is indispensable for integrative studies in connection with other ecological key parameters assessed e.g. on the transnational grid. Whilst this may reveal interrelationships and pathways at the large scale (Level I), the results from the intensive monitoring on permanent observation plots (Level II) will provide more insight into the functioning at the ecosystem level. Both levels of information should be linked together by means of upscaling results from Level II to the large scale.
