Navigation mit Access Keys

Main menu


Tending measures in the mixed broadleaved forest Buechberg, Diessenhofen


Each year, about 50 Millions CHF are invested in tending interventions in Swiss forestry. In a field experiment of 7.3 ha size, we study how four tending variants affect tree species richness and stem quality in the long term. Particular interest is given to the question how minority species present in mixed stands can be maintained against the highly competitive beech.


What are tending measures?

Tending measures are all silvicultural interventions in the development stages young growth, thicket and small pole stand. In a young growth, the 100 largest trees per ha are smaller than 1.3 m, in the subsequent thicket, the diameter at breast height (dbh) of these trees is smaller than 10 cm and in the small pole stand, between 10 and 20 cm.

Tending measures are aimed at influencing the tree species composition according to the management goals, improving the stem quality and enhancing the trees‘ resistance against mechanical damage (e.g., through snow load).

When a stand is tended, trees with desirable properties (healthy, fast growing, desirable species, stable, straight stem throughout, small branches, see Fig. 1) are promoted by removal or weakening of their competitors („positive selection”), or trees with undesirable properties are cut (“negative selection”). The promoted trees („candidate trees“) are selected in a spatial distribution which corresponds to the distribution of the trees in an old stand („final distance“). It is also possible to promote more candidate trees.

The project „tending measures in mixed broadleaved forests“ aims at investigating how effective tending measures are and which interventions can be abstained from without significant disadvantage.

Project goals

The project goals are,

  1. to test different tending variants in mixed broadleaved forests and to propose cost-efficient and in the same time effective variants;
  2. to establish a demonstration plot of such tending variants, in cooperation with forest managers.

The four tending variants

    1. No tending intervention (control)
    2. 1st intervention 2014 at 12 m dominant height, candidate trees at final distance.
    3. 1st intervention at 5 m dominant height, candidate trees at final distance, secondary measures in between, interventions 2007 and 2014.    
    4. Conventional tending on the whole area, 500-700 candidate trees per ha, interventions 2003, 2007, 2011 and 2014.

    Research methods

    The project is based on a field experiment established in a storm-felled area which was created by the hurricane Lothar in 1999. The storm area is some 40 ha in size and is located in a mixed broadleaved forest in Diessenhofen (Canton Thurgau). The young forest originates largely from natural regeneration. The test site is located in a mesic beech forest (site associations 7a, 7f, 7g according to Schmider et al. 2003) at an altitude of 425-450 m and with a gentle slope with north and west aspect. The forest is owned by the civic community Diessenhofen and the Canton of Thurgau. The area was chosen because its size allowed for the establishment of a field trial under operational conditions and because it has a great diversity of tree species.

    The test site is 7.3 hectares in size. An experimental block design with four treatments (tending variants) and 4 replicates was applied (Figure 2). The 16 experimental units (subplots) are 0.34 to 0.69 ha in size. In each subplot 30 candidate trees were marked and followed. The variables recorded periodically include dbh, tree height and stem form. In addition, a sampling inventory of the entire experimental area was conducted in 2003 on 480 circular sample plots (10,000 saplings recorded in total).

    The tending measures are agreed between WSL and the forest enterprise and executed by the enterprise, and the time needed for tending each subplot is recorded.


    Results and practical relevance

    Between 2003 and 2014, stem numbers of trees of at least 20 cm in height decreased by 50% from 31’000 trees/ha to 15’900 trees/ha. The average basal area amounted to 19.6 m2/ha in 2014. The 474 candidate trees had 2014 an average height of 10.7 m (± 0.2 m standard error) and a dbh of 11.1 cm (± 0.3 cm). When compared to all the trees of the storm area, the dbh of candidate trees was bigger in 2014 (Fig. 3). In 2014, the storm area was dominated by beech (58% of all stems), while the other tree species contributed with proportions of: 14% hornbeam, 8% Norway spruce, 6% ash and 4% silver fir to total stem number. In total more than 20 tree species were found.


    Tending measures were found to have the following effects in 2014: 1) reduction of the basal area by ca. 25%; 2) increase in probability that a candidate maintained his status, also when regarding all tree species other than beech and hornbeam; 3) probably slight reduction of loss of tree species richness; and 4) slight reduction in tree heights with tending variant D. The costs of the measures amounted to 413 CHF/ha for variant B, to 2180 CHF/ha for variant C and to 5002 CHF/ha for variant D. We estimated the costs for maintaining a candidate between 2003 and 2004 to 133 CHF with variant C and to 188 CHF with variant D.

    This study confirms that early tending measures in mixed broadleaved forests can help to sustain other broadleaved tree species besides beech. Such tending measures are however expensive. Therefore, they should be used with caution, in a spatially concentrated manner and restricted to sites where maintaining species mixtures is important.