1. Controversies about successional dynamics of woodland communities have a long history, dating back to the classical debates between Clements and Gleason and continuing into the present. These debates have largely concerned the predictability or convergence of forest developmental trends as well as the relative importance of different mechanisms and forces driving forest succession. However, opportunities for rigorous testing of competing hypotheses are limited, mainly because plot-based studies of forest vegetation spanning more than a decade are scarce and even fewer concern late-successional stands.

2. We exploit a unique long-term data set from mesic temperate forests of eastern Poland, spanning ca. 80 years (1936-2012) in strictly protected, late-successional woodlands assigned to seven different ‘structural types’. We use non-metric multidimensional scaling to assess stability of species composition over the study period. We examine predictability of composition and change trajectories over time using Mantel statistics, and we examine changes in distributions of dissimilarity indices to assess convergence or divergence at the examined time and spatial scales.

3. Tree communities in Białowieża Forest have changed substantially over the last eight decades. Several species (aspen, birch, pine, oak, ash, maple and spruce) exhibited large decreases in density, while few other species (especially hornbeam and lime) have increased in importance across a wide range of initial compositional types. Forest types recognized in earlier periods have become much less distinguishable. Metrics do not yet show clear successional convergence, mainly due to a) declines in the previously broadly distributed spruce and b) persistence of large individuals of intermediate, long-lived species (even though these species lack significant regeneration).

4. Synthesis. Late-successional woodland communities of Białowieża Forest are clearly dynamic and do not show quasi-equilibrial properties often assumed of old-growth forests. Forest types previously recognized as distinct have become progressively less differentiated. Plausible explanations invoke alteration in competitive relationships due to complex changes in environment, including climate, N deposition, natural and human disturbance, and ungulate herbivory over 70 years. Hornbeam and lime have been favored over spruce, pine, oak and birch, although the resulting successional convergence is far from complete at this time. In the absence of major disturbance, we expect future decades to show continued and successional homogenization.

Paper

Over 80 years without major disturbance, late-successional Białowieża woodlands exhibit complex dynamism, with coherent compositional shifts towards true old-growth conditions.  Brzeziecki B., Woods K., Bolibok L., Zajączkowski J., Drozdowski S., Bielak K., Żybura H.

Data scope

This data file contains numbers for tree species basal area (m²), calculated for all live trees with DBH ≥5cm, occurring on quadrats 20 m x 20 m. Quadrats were selected in a chess-board pattern within the permanent research plots located in Białowieża National Park (cf. the paper). Calculations were performed separately for 11 tree species, 5 research plots, and 7 census dates.

Research plots

The permanent research plots are transects of different width (40-60m) and length (200-1380m) (total area = 15.44 ha), located in the core section of the Białowieża National Park, the so called „Strict Reserve” (see Appendix 1 in the paper).

Census dates and scope of measurements

Transects have been sampled on a total of 7 major occasions: in 1936, 1955-1959 (1957), 1968-1972 (1970), 1981-1983 (1982), 1991-1993 (1991), 2001-2003 (2002) and 2011-2013 (2012). During the first survey, all trees with a DBH (diameter at breast height) of at least 5 cm were identified and their spatial locations were mapped. During subsequent censuses, similar measurements were made. Additionally, recruits, i.e. trees which had increased in diameter by up to 5 cm in the period between consecutive censuses, were identified, as well as trees which had died since the time of the last record (losses). Scientific species names and numbers of live trees measured during the first and last census are given in Appendix 2 in the paper.