Colonization processes

Contrary to other studies which report delayed colonization processes in restoration (Bakker et al. 1996; Coulson et al. 2001; Verhagen et al. 2001), more than 50% of the species found in the surrounding area of 30 km2 were able to colonize the investigated mining sites north and south of Halle (Table 1). Depending on the landscape structure of the surroundings, up to 40% of the species already present in the mining sites were found more than 3 km away (long-distance dispersal). This phenomenon can be observed mainly at hospitable mining sites within a species-poor surround­ing landscape (examples “southern mining sites”).

High migration rates from larger distances can be explained by extraor­dinary events like gales, thermally induced turbulences or zoochory. These events can be compared to the migration processes after the last ice age (Clark 1998). Even rare species can accumulate at the mining sites which act as large “seed traps” in the cultural landscape (Tischew and Kirmer 2003). The availability of large-scale competition-free space in the mining sites supports the establishment of these species. Most of the species that migrated from distances greater than 3 km represent species from open landscapes and only 25% are woodland species. Most of these woodland species are trees and shrubs, woodland herbs are rather rare. However, woodland herbs are suitable for assessing successional dynamics in wood­lands (Wolf 1989). Further analysis, therefore, refers particularly to this.

The relationship of distance of diaspore sources to the number of herb species in dump woodlands was analyzed in detail on four spoil dumps of similar age and similar site conditions (Fig. 1). In the case of the site with nearby diaspore sources (A), more than twice as many species migrated into the site than in the sites with more distant diaspore sources (sites B – D). The average number of species with animal dispersal (epizoochory, endozoochory) and self-dispersal (autochory) is much higher for the nearby site, while the average number of wind-dispersed (anemochorous) woodland herbs is nearly the same.

Fig. 1. Effect of the dis­tance of diaspore sources on the species number of woodland herbs in pioneer woodlands on mining sites (plot size 400-525 m2).

A Mining site Goitzsche – Tagesanlagen, B mining site Goitzsche Halde 1035, C mining site Rohbach, D mining site Kayna-Sud

More detailed investigations with respect to time-dependent coloniza­tion by woodland herbs were carried out based on data from all woodland areas investigated in eastern German post-mining landscapes. Table 2 shows a relatively high number of species that can already occur on young spoil dumps (10-60 years old). The seeds of these species were effectively dispersed by wind or animals over long distances (long-distance dispersal). Species with less effective dispersal mechanisms (autochory or dispersal by ants) mainly occur only on spoil dumps with pioneer woodlands, which are older than 60 years and characterized by an accumulation of intermedi­ate and climax tree species in the shrub or tree layer (late-successional pioneer woodlands).

Indicator species of “ancient woodlands” (Peterken 1994; Wulf 1995) have a low frequency in general. They only occur in woodlands directly bordering “ancient woodlands” or on so-called “handmade dumps” (top­soil containing humus). It will probably take more time for isolated sites to be colonized by woodland species. Therefore, maintaining “ancient wood­lands” in the surroundings of spoil dumps is an important basis for devel­oping species-rich woodlands in post-mining landscapes (Benkwitz et al. 2002). In spontaneously developed woodlands, neophytes have rather a small share, with three species per plot. They comprise on average 2.8% of the herbaceous layer and 3.4% of the tree layer. Only the North American red oak (Quercus rubra) occurs more frequently in some regions. Quercus rubra was frequently planted in plantation forests, which provide a source for dispersal by jays.

Table 2. Frequency of selected woodland herbs in early and late successional pio­neer woodlands with reference to their dispersal capacity and occurrence of indi­cator species for recent and ancient woodlands. A: number of occurrences (nto- tal = 104), B: frequency in pioneer woodlands < 60 years (%, ntotal = 86), C: frequency in woodlands > 60 years (%, ntotal = 18), D: dispersal modes (Muller – Schneider 1986; Bonn et al. 2000): a – anemochorous, zepi – epizoochorous, zendo – endozoochorous, m – myrmecochorous, s – autochorous, ? – dispersal mode un­known, E: Long-distance dispersal capacity (Frey and Losch 1998; Verheyen et al. 2003), long – present; no – not present, F: indicator species for r: recent wood­lands and a: ancient woodlands (Wulf 1995, 2003). A B C D E F Aegopodium podagraria 2 1.2 % 5.6 % s no r Anemone nemorosa 1 0 % 5.6 % s, m no a Avenella flexuosa 27 26.7 % 22.2 % a, zepi long r Brachypodium sylvaticum 26 12.8 % 83.3 % a, zepi long a Calamagrostis arundinacea 1 1.2 % 0 % a, zepi long Campanula trachelium 2 0 % 11.1 % s no a

Table 2. (cont.) Convallaria majalis 4 1.2 % 16.7 % s zendo no & long a Carex brizoides 2 1.2 % 5.6 % a zepi long Carex montana 5 3.5 % 11.1 % 3? ZepU m no & long Carex pilulifera 10 10.5 % 5.6 % 3? ZepU m no a Circaea lutetiana 2 0 % 11.1 % zepi long a Clematis vitalba 5 2.3 % 16.7 % a, zepi long Dactylis polygama 6 5.8 % 5.6 % a, zepi long Dryopteris carthusiana 3 2.3 % 5.6 % a long a Dryopteris filix-mas 4 2.3 % 11.1 % a long Epilobium montanum 5 4.7 % 5.6 % a long Epipactis atrorubens 22 18.6 % 33.3 % a long Festuca gigantea 4 0 % 22.2 % a, zepi long a Festuca heterophylla 1 0 % 5.6 % a, zepi long Fragaria vesca 37 29.1 % 66.7 % zendo long Hedera helix 3 1,2 % 11.1 % zendo long Hieracium lachenalii 54 55.8 % 33.3 % a long a Hieracium laevigatum 40 39.5 % 33.3 % a long Hieracium murorum 20 19.8 % 16.7 % a long Hieracium sabaudum 76 74.4 % 66.7 % a long Hieracium umbellatum 1 1.2 % 0 % a long Holcus mollis 1 1.2 % 0 % a, zepi long Listera ovata 6 7.0 % 0 % a long Luzula pilosa 2 0 % 11.1 % m no a Maianthemum bifolium 3 0 % 16.7 % zdyso no a Melampyrum pratense 9 8.1 % 11.1 % m no a Melica nutans 2 0 % 11.1 % m no a Milium effusum 1 0 % 5.6 % m? zepi no & long a Moehringia trinervia 4 0 % 22.2 % m no a Monotropa hypophegea 3 2.3 % 5.6 % ? ? Orthilia secunda 16 12.8 % 27.8 % a long Poa nemoralis 27 25.6 % 27.8 % a, zepi long a Polygonatum multiflorum 1 0 % 5.6 % zepi no a Pyrola chlorantha 1 1.2 % 0 % a long Pyrola minor 18 15.1 % 27.8 % a long Pyrola rotundifolia 1 0 % 5.6 % a long Scrophularia nodosa 1 0 % 5.6 % s no Solidago virgaurea 10 10.5 % 5.6 % a, zepi long Stachys sylvatica 1 0 % 5.6 % zepi long a Sanicula europaea 2 0 % 11.1 % zepi long a Stellaria holostea 2 0 % 11.1 % zepi no? a Vacinium myrtillus 10 5.8 % 27.8 % zendo no & long a Vacinium vitis-idea 7 3.5 % 22.2 % zendo no & long Viola riviniana/V. reichenba- 15 7.0 % 50.0 % s, m no a chiana

Site-dependent and chronological woodland differentiation

Based on multivariate analysis, three successional series (A, B, C) could be derived for woodland development on different site conditions (Fig. 2). Only woodlands not directly influenced by groundwater were included in the analysis.

Woodland development proceeds very slowly on sites with a high pro­portion of Tertiary substrates (A). These are characterized by extremely acid pH-values (minimum 3,0), high to very high coal contents (25 % on average in the 60-100 years old stages) and hydrophobic features. The carbon-to-nitrogen ratio as one parameter for nutrient availability was ex­tremely high, even on the older dumps (1:55 in a soil depth of 0-10 cm in the 60-100 years old stages). These substrates are hardly suitable for colo­nization. Several pioneer woodland stages have to be passed through fre­quently until the development to later woodland stages is possible.

Considerably faster development compared to species-rich woodlands was recorded on more hospitable substrates depending on diaspore sources in the surroundings (B, C). Sites of series B are characterized by Tertiary and Quarternary mixed substrates. They have low to moderate pH-values (3,9-5,9; all data relate to the upper soil layers (0-10 cm depth) in the 60­100 years old woodland stages) and low to moderate nutrient availability (C/N ratio 1:23 to 1:40). An initial accumulation of intermediate and cli­max tree species in the tree and shrub layer as well as first woodland herbs in the herbaceous layer can be observed in the age-class 60 – 100 years. The quickest successional progress was recorded in the woodlands of the mesophilic to nutrient rich sites of series C. The Tertiary and Quarternary mixed substrates show moderate to high pH-values (4.7 to 6.3) and mod­erate to high nutrient availability (C/N ratio 1:17 to 1:29). Compared to se­ries B, an even greater accumulation of particular woodland herbs in the herbaceous layer as well as intermediate and climax tree species in the tree and shrub layer may be recorded after a 60-100 years development.

Fig. 2 (next page) Successional series (A, B, C) in lignite mining areas of eastern Germany

series A extreme sites
successional series
– tertiary sand or silt with extremely acid pH-value – high to very high coal contents, very low nutrient availability
characteristics
C-1: Initial pioneer birch woodlands with simultaneous development of a herbaceous layer
B-1: Initial pioneer birch woodlands by skipping or quickly passing the stage of herbaceous layer
10-30 years
A-2: Very thin pioneer wood­lands with birch and/or pine	B-2: Pioneer birch woodlands without accumulation of intermediate/ climax tree species	C-2: Pioneer birch woodlands with initial accumula­tion of wood­land herbs and intermediate/ climax tree species in the shrub layer
30-60 years
B-3: Pioneer birch woodlands with some wood­land herbs and initial accumula­tion of interme – diate/climax tree species in the shrub/tree layer	C-3: Pioneer birch woodlands with accumula­tion of woodland herbs and inter – mediate/climax tree species in the shrub/ tree layer
A-3: Persisting thin pioneer wood­lands with birch and/or pine
60-100 years

In addition to site conditions, successional rate depends on availability of diaspore sources in the surroundings and substrate-modifying influences (e. g. fly-ash coating). Non-linear developments, related to general individ­ual conditions are possible, i. e. successional stages can be passed through more quickly or slowly.

Fig. 3. Development of stand structure (bare soil, herbaceous, shrub and tree lay­ers) of the successional series A, B and C

As shown in Fig. 3, bare soils and open woodland structures persist on the less hospitable sites of series A over a considerably longer period, in contrast to series B and C. This development is not necessarily negative with respect to biodiversity as these woodlands remain as refuges for the less competitive, rare species which prefer open woodland structures. This differentiated development also leads to a varied and aesthetically attrac­tive landscape.

Fig. 4. Development of species number in herbaceous, shrub and tree layers of the successional series A, B and C

Fig. 4 presents differences in the development of species diversity in separated vegetation layers. High numbers of woodland species in shrub and tree layers, as well as herbaceous layers, mainly occur on the oldest sites of series C. Many of these spoil dumps represent “handmade dumps.” Here, woodland species can establish themselves faster on account of the topsoil coverings, which accelerate soil formative processes and provide a diaspore source of the woodland species. Therefore, slower development processes are expected on isolated younger (non-handmade) sites of this successional series in the future. However, the younger and less developed age-classes of succession are also of high value for nature conservation and for experiencing nature. For example, a total of 245 species are found on a 20 ha, 30-year-old dump with different stages of birch and pine pio­neer woodlands on the Goitzsche mining site. Fourteen species alone are to be found on Red Lists of Saxony-Anhalt (Frank et al. 1992) and Germany (Korneck et al. 1996). Therefore, young woodland stages in post-mining landscapes contribute substantially to the maintenance of biological diver­sity in post-industrial landscapes.

Gradual migration of woody species onto spoil dumps was evaluated in detail for successional series C (Fig. 5). Young stages are characterized by the pioneer tree-species silver birch (Betula pendula) and Scots pine (Pinus sylvestris). Common oak (Quercus robur) can migrate onto young sites, whereas most intermediate and climax tree species become estab­lished only slowly in the shrub and tree layer in the second or third stage. Rejuvenation of pioneer tree species is already considerably restricted on the oldest sites. Here, a general change in stand structure is beginning as pioneer tree species are also dying off.