Although responses of aerial tree parts to wind excitation are of crucial importance for deciduous broadleaved tree stability against wind loading, only a few field studies have been conducted that explore wind tree part interactions. Therefore, responses of aerial parts of an individual Norway maple tree (Acer platanoides) to wind excitation were measured in a field study near the city of Freiburg (Southwest Germany) between July 2009 when the tree was in leaf and March 2010 when the tree was out of leaf. Using Fourier analysis, for most of the sampled tree parts six frequencies associated with the first three vibration modes in two motion directions were identified. In agreement with previous studies, vibration frequencies were lower when the tree was in leaf compared to the leafless period. The results indicate the importance of the foliage for the seasonal variation in aerial tree part motion damping. For the analyzed wind speed range, it is shown that the importance of higher order vibration modes localized on the sampled tree parts for total tree part response to wind excitation decreases with increasing wind speed. Results from wavelet analysis demonstrate that tree part response strongly covaried in the time frequency-space only with wind components having frequencies much lower than the frequencies associated with the dominant tree part vibration modes. (C) 2013 Elsevier B.V. All rights reserved.

Wind-induced motion of 29 neighbouring trees growing in a Scots pine plantation was measured over the period 14 March to 24 March 2008. The bi-orthogonal decomposition (BOD) of the complex tree motion field into a limited number of spatio-temporal modes provided the basis for the analysis of the response behaviour to wind excitation of the group of sample trees. It is shown that the first BOD-mode was the most energetic and differed from all other BOD-modes. The BOD-results as well as the results from wavelet analysis of the temporal eigenvectors of the BOD-modes demonstrate that two concomitant low-frequency components in the streamwise wind velocity component stimulated coherent response of the sample trees at the tree group level. However, in the range of the wind speed measured close to the top of the Scots pine forest canopy (hourly mean wind speed values lower than 6 m s−1), the wind loads associated with these low-frequency airflow structures were too low to harmonise the motion of the sample trees completely. It is shown that instantaneous single tree responses to wind excitation were highly irregular in magnitude and direction. Results from Fourier and wavelet analysis demonstrate that sway in the first mode dominated the wind-induced sway behaviour at the tree level.

Data on storm damage attributed to the two high-impact winter storms ‘Wiebke’ (28 February 1990) and ‘Lothar’ (26 December 1999) were used for GIS-based estimation and mapping (in a 50 × 50 m resolution grid) of the winter storm damage probability (PDAM) for the forests of the German federal state of Baden-Wuerttemberg (Southwest Germany). The PDAM-calculation was based on weights of evidence (WofE) methodology. A combination of information on forest type, geology, soil type, soil moisture regime, and topographic exposure, as well as maximum gust wind speed field was used to compute PDAM across the entire study area. Given the condition that maximum gust wind speed during the two storm events exceeded 35 m s-1, the highest PDAM values computed were primarily where coniferous forest grows in severely exposed areas on temporarily moist soils on bunter sandstone formations. Such areas are found mainly in the mountainous ranges of the northern Black Forest, the eastern Forest of Odes, in the Virngrund area, and in the southwestern Alpine Foothills.

The wind damage probability (P (DAM)) in the forests in the federal state of Baden-Wuerttemberg (Southwestern Germany) was calculated using weights of evidence (WofE) methodology and a logistic regression model (LRM) after the winter storm 'Lothar' in December 1999. A geographic information system (GIS) was used for the area-wide spatial prediction and mapping of P (DAM). The combination of the six evidential themes forest type, soil type, geology, soil moisture, soil acidification, and the 'Lothar' maximum gust field predicted wind damage best and was used to map P (DAM) in a 50 x 50 m resolution grid. GIS software was utilised to produce probability maps, which allowed the identification of areas of low, moderate, and high P (DAM) across the study area. The highest P (DAM) values were calculated for coniferous forest growing on acidic, fresh to moist soils on bunter sandstone formations-provided that 'Lothar' maximum gust speed exceeded 35 m s(-1) in the areas in question. One of the most significant benefits associated with the results of this study is that, for the first time, there is a GIS-based area-wide quantification of P (DAM) in the forests in Southwestern Germany. In combination with the experience and expert knowledge of local foresters, the probability maps produced can be used as an important tool for decision support with respect to future silvicultural activities aimed at reducing wind damage. One limitation of the P (DAM)-predictions is that they are based on only one major storm event. At the moment it is not possible to relate storm event intensity to the amount of wind damage in forests due to the lack of comprehensive long-term tree and stand damage data across the study area.