- Email: [email protected]
Natural revegetation of indigenous roadside vegetation by propagules from topsoil
ARTICLE IN PRESS
Urban Forestry & Urban Greening 3 (2004) 29–37 www.elsevier.de/ufug
Natural revegetation of indigenous roadside vegetation by propagules from topsoil Astrid Brekke Skrindo*, Per Anker Pedersen ( Norway Department of Plant and Environmental Sciences, Agricultural University of Norway, Box 5003, 1432 As, Received 16 June 2003; accepted 2 December 2003
Abstract Road construction degrades large areas of indigenous vegetation. Better understanding of ecosystem restoration after degradation will require development and assessment of improved restoration techniques. The potential of natural revegetation based on propagules from topsoil as a restoration technique, to achieve succession towards indigenous vegetation, is not fully understood and accepted as an alternative method for vegetation establishment. The aim of this study was to evaluate this restoration technique by comparing different topsoils with respect to their potentials for revegetation and to describe the early succession on these topsoils by recording change in vegetation cover and the number of species. The study road RV23 is situated in south-east Norway and runs mainly through mixed, mostly coniferous forest. The topsoil (the upper 30 cm) was removed before road construction started, stockpiled during the construction period and then redistributed on the road verge afterwards. In 1999, 10 macroplots of 25 m2 were placed randomly on selected topsoils of different qualities, to represent the soil variation in the study area. The soil quality varied from organic-rich to organic-poor soils. Each macro plot contained four 1-m2 mesoplots in which the species composition was recorded in 2000 and 2001.Univariate statistical tests were applied to reveal change in vegetation cover between macroplots within 1 year and between years 2000 and 2001. Furthermore, statistical tests were used to reveal the change in singlespecies frequencies between the study years. Two years after deposition of topsoils, the vegetation cover was generally satisfactory from an aesthetical point of view. However, the variation in vegetation cover and species composition was considerable on different topsoils, reflecting both the local variation of the indigenous vegetation and the soil quality. A total of 121 species of vascular plants were recorded in all 10 macroplots, although not more than 59 species were found in one macroplot. Species composition and single-species frequencies changed considerably from the first year to the second. Among the 61 species that were recorded in more than five mesoplots, the frequency of 16 species increased and the frequency of 10 species decreased significantly (Po0:05). Most of the decreasing species can be considered as weed species and are not represented in the present indigenous vegetation, while most of the increasing species are. The vegetation change from 2000 to 2001 apparently represents the first steps in a succession towards an ecosystem dominated by species of the indigenous vegetation. r 2004 Elsevier GmbH. All rights reserved. Keywords: Revegetation; Indigenous vegetation; Roadside; Topsoil; Propagule bank
*Corresponding author. E-mail address: [email protected] (A. Brekke Skrindo). 1618-8667/$ - see front matter r 2004 Elsevier GmbH. All rights reserved. doi:10.1016/j.ufug.2004.04.002
ARTICLE IN PRESS 30
A. Brekke Skrindo, P. Anker Pedersen / Urban Forestry & Urban Greening 3 (2004) 29–37
Introduction Road construction degrades large areas of indigenous vegetation. When building new roads, the added verges make up a significant area to be revegetated for environmental and aesthetical reasons and to prevent erosion. Preservation and re-establishing indigenous vegetation helps to reach the official goal: to restore degraded areas and to improve road aesthetics within the natural landscape (Statens, 1992). There are four ways to re-establish indigenous vegetation: (1) Transplant larger patches of vegetation, (2) cultivate indigenous plants, (3) in situ germination of seeds collected from the surrounding areas and (4) germination from the topsoil propagule bank and the areas adjacent to the road. In Norway, seeding of non-indigenous grasses is one of the most used methods along roads resulting in a somewhat monotonous green lawn. Since the three first methods are expensive and time-consuming, natural revegetation from topsoil is an appealing alternative and the topic of this study. Topsoil is defined differently throughout literature. Hargis and Redente (1984) present three possible definitions for topsoil: soil from the A-horizon; soils from the A and E-horizon; or a specific soil depth regardless of soil layer. In this study, the latter definition is used for practical reasons. Revegetation from topsoil is based on germination from the propagule bank followed by the process of natural succession, defined as the nonseasonal, directional and continuous pattern of colonisation and extinction on a site by species populations (Begon et al., 1990). The procedure is straightforward: topsoil is removed before construction, stockpiled and then replaced. The advantages are that plant nutrients, viable propagules, mychorriza and microfauna remain in the upper soil (cf. Hargis and Redente, 1984). Restoration of indigenous vegetation from the topsoil propagule seed bank shows contradictory results. It has failed in several grasslands (Thompson and Grime, 1979), but succeeded in South African shrubland (Holmes, 2001), in a Californian (USA) meadow (Kotanen, 1996) and in woodland restoration in Western Australia (Rokich et al., 2000) and in England (cf. Warr et al., 1993). Furthermore, it has proved to be successful for revegetation of contaminated soils, such as lead/zinc mine tailings in China (Zhang et al., 2001) and fly ash and gypsum in the United Kingdom (Shaw, 1996). To our knowledge, restoration of indigenous roadside vegetation from the propagule bank in the topsoil has not been assessed before in Scandinavia. The study sites were integrated in the construction of a new road (‘‘Oslofjordforbindelsen’’), which passes through various vegetation types. The revegetation methods along the study road are a mosaic of revegetation from topsoil and seeded and planted areas, but this investigation deals only with revegetation from
redistributed topsoil. Since this study consists of data from only 2 years, we focus on the first phase of a secondary autogenic succession. The aim of the study was to evaluate the redistribution of topsoil as a restoration technique and to compare revegetation success from different topsoils. Particular attention was given to change in vegetation cover, the number of species, relationship between soil and vegetation and the frequency of species divided into four functional groups: woody plants, flowering forbs, agricultural weeds and graminoids.
Materials and methods Study area and restoration technique The study road, RV23 (Oslofjordforbindelsen), passes through 13 km of mixed forest, coniferous forest dominated by Scots pine (Pinus sylvestris) and Norway spruce (Picea abies) and bogs. The road is situated in south-east Norway in the municipalities of Frogn, Hurum and Røyken in Akershus and Buskerud counties respectively (10 440 E, 59 420 N to 10 250 E, 59 440 N). The climate is sub-oceanic. Annual mean precipitation in the years 1961–1990 (normal) was 920 mm at the Drøbak meteorological station (Førland, 1993). The annual precipitation in the study period was 1191.7 mm (1999), 1360.8 mm (2000) and 991.7 mm (2001). The precipitation during the growing season was higher or about the same as the normal with no continuous drought periods. The topsoil, defined by the Public Roads Administration as the upper 30 cm of the soil profile due to practical reasons, was excavated before road construction and stockpiled separately during the construction period (in 1998). The topsoil volume decreased during stockpiling, consequently a layer of 10 cm was redistributed on top of the subsoil on the road verge (in May and August 1999). The exact origin of the soil replaced on plots is not known due to some degree of mixing of soil during stockpiling. The vegetation composition in the original site is therefore unknown. The thickness of the humus layer in the original soil profile varied between 5 and 20 cm except in bogs.
Sampling Recording of vegetation data In 1999, 10 macroplots of 25 m2 were placed randomly on selected topsoils of different qualities representing the topsoil-variation in the study area. Each macroplot contained four 1 m2 mesoplots systematically placed in the corners by grid sampling (Økland, 1990). Each mesoplot was divided into 16 (25 25 cm2) subplots.
ARTICLE IN PRESS A. Brekke Skrindo, P. Anker Pedersen / Urban Forestry & Urban Greening 3 (2004) 29–37
In July 2000 and 2001, the frequency of all vascular plant species in the 40 mesoplots was recorded by (1) estimating percentage cover on a 0100 scale, and (2) by recording the presence/absence of all species in each subplot to calculate subplot frequency on a 016 scale (Økland, 1988). The vascular plants nomenclature followed Lid and Lid (1994).
31
weight (unit: mg/100 ml). For soil variables, skewness and kurtosis were calculated using Statgraphics, Version 5.0 (Table 1). Homogeneity of variances (Table 1) was achieved by transforming variables to zero skewness by ln (c þ x) according to Økland et al. (2001). Statistical significance of differences in total vegetation cover (%) between the macroplots in each year 2000 and 2001 and the change between the years 2000 and 2001, were tested using an analysis of variance including a multiple range test, Ryan–Einot–Gabriel–Welsh (REGWQ), Po0:05 (SAS Institute, 1987) according to the model
Recording of soil data Soil samples were collected from the upper 10 cm layer on the boundaries of each mesoplot and analysed to provide soil variables (Table 1): Readily available plant nutrients, P, K, Mg, Ca, were extracted by 0.1 M ammonium lactate and 0.4 M acetic acid according to the AL-method (Egne! r et al., 1960) and determined by inductively coupled spectroscopy (ICP). The content of organic matter was assessed by determining the loss on ignition and soil pH was measured in water suspension. The soil texture was classified (Table 2) according to Sveistrup and Njøs (1984).
Xij ¼ mi þ ai þ eij ; where Xij is the performance of the macroplot, i the ði ¼ 1; 2; :::10Þ; and due to the mesoplot, j the ðj ¼ 1; 2; :::4Þ; m the expected mean, a the expected effect of the macroplot, e the residual (variation between the mesoplots). The hypothesis tested was a ¼ 0 (no effect of the macroplot) against aa0: Relationships between the number of species and the vegetation cover in the macroplots were addressed using Kendall’s rank correlation method (Kendall, 1938). To address the relationships between the soil variables and the relationship between the soil and vegetation cover Kendall’s rank correlation method (Kendall, 1938) was
Statistical analyses All soil variables except pH were adjusted to the soil density by multiplying the variable to the soil volume Table 1. Soil variables measured in all 1 m2 plots Variable
Loss on ignition pH P K Mg Ca
Unit
Summary statistics of untransformed variable
mg/100 ml — mg/100 ml mg/100 ml mg/100 ml mg/100 ml
Transf. ln (c þ x)
Summary statistics of transformed variable
Range
S.S.
S.K.
c-value
Mean
S.D
S.K
1–400 4.2–8.6 5.7–383 5.2–65.7 2.68–45 9.93–496
1.58 0.68 2.78 0.04 2.13 4.60
1.16 1.25 0.22 1.85 0.16 3.80
0.65 1.00 1.00 0.1 — 4.23
2.25 1.95 1.95 0.99 — 2.09
1.10 0.16 0.16 0.34 — 1.00
1.51 1.17 1.17 0.53 — 0.38
Abbreviations: S.S.=standardised skewness; S.K.=standardised kurtosis, c-value=index used in transformation formula; S.D.=standard deviation, =analysed by the AL-method.
Table 2. Soil classification of the macroplots Macro plot
Soil classification
Organic components
1 2 3 4 5 6 7 8 9 10
Silt loam Silt loam Medium loamy sand Organic soil Mold rich loam Clay loam+silt loam Silt loam+loam+medium loamy sand Mold rich loam Clay loam+silt loam Mold rich loam
Litter—weakly decomposed Litter—variable degree of decomposition Litter—weakly decomposed Sphagnum—weakly decomposed Litter—weakly decomposed Sphagnum—weakly decomposed Sphagnum and litter—weakly decomposed Litter—weakly decomposed Litter—well decomposed Litter—weakly decomposed
Description of the organic components; litter originating from dry coniferous woodland, litter originating from swamp woodland.
ARTICLE IN PRESS 32
A. Brekke Skrindo, P. Anker Pedersen / Urban Forestry & Urban Greening 3 (2004) 29–37
used. The statistical difference of single species frequencies between years 2000 and 2001 was tested using Wilcoxon’s signed-rank test for two paired samples (cf. Sokal and Rohlf, 1995). The tests were against the twotailed alternatives. Testing was restricted to species recorded in 5 or more mesoplots either year.
Results Vegetation cover and the number of species Two years after the redistribution of the topsoil, six macroplots had at least 50% vegetation cover. The mean vegetation cover ranged from 1% to 95% in 2000 and from 5% to 100% in 2001. The macroplots differed statistically (Po0:001) both years, and there was no significant difference within each macroplot. In year 2000, no distinct groups of macroplots could be statistically defined (Table 3). In 2001, the macroplots could be divided into four groups: macroplots 1, 5, 8, 9, 10 had the highest vegetation cover followed by plot 6 then plots 2 and 7, and finally plots 3 and 4 (Table 3). The vegetation cover increased significantly in all the macroplots from year 2000 to 2001 (Table 3). In general, plots with high cover vegetation cover in 2001 had high cover in 2000 as well. On the other hand, in plots with initially low vegetation cover (1%) the differences between year 2000 and 2001 varied substantially. There were 121 recorded species in total and the number of species per macroplot ranged from 8 to 51 in year 2000 and from 10 to 59 in year 2001 (Table 3). Fourteen species were found only in the year 2000 and 27 species appeared for the first time in year 2001. There were strong correlations between the vegetation cover
and the number of species both years (year 2000; t ¼ 0:5004; P ¼ 0:0001; year 2001; t ¼ 0:442; P ¼ 0:0002).
Relationships between vegetation cover and soil variables The content of plant nutrients in the soil varied between plots (Table 4). In most samples the content was medium high. The organic soil in macroplot 4 was low in all elements. On the contrary, the macroplots with clay loam (6 and 9) had the highest content of all nutrients except for Ca, which was highest in the medium loamy sand of macroplot 3. Vegetation cover did not correlate strongly with most soil variables (Table 5). The only significant relationship in both years at level Po0:01 was with K, and organic matter (Loss on ignition) in year 2000. There were rather strong correlations between some of the soil variables (Table 6) indicating a more complex relationship between the soil variables and the vegetation cover than the direct correlation explained. The loss on ignition had statistically significant negative correlaTable 4. Mean values of the soil variables Macro Loss on pH P-AL K-AL Mg-AL Ca-AL Density plot ignition 1 2 3 4 5 6 7 8 9 10
6.4 7.9 2.8 14.1 15.4 7.5 4.6 11.8 6.4 23.7
5.4 5.2 8.0 4.3 5.5 7.0 7.4 6.2 7.2 5.8
2.8 2.3 2.2 1.6 0.8 3.4 4.7 1.0 5.1 1.4
9.8 7.2 11.0 3.0 9.5 15.2 11.7 26.8 29.0 9.5
15.0 3.6 21.4 8.1 12.8 31.2 29.6 13.9 28.6 23.1
63.0 30.8 1000.4 34.9 117.6 442.3 390.4 267.9 296.8 484.0
0.95 1.12 1.16 0.43 0.42 0.98 1.24 0.51 1.20 0.70
The unit for all variables (except pH) is mg/100 ml.
Table 3. The number of species and mean vegetation cover in each macro plot in years 2000 and 2001 Macroplot Number of species
1 2 3 4 5 6 7 8 9 10
Mean vegetation cover (%)
2000
2001
2000
2001
51 17 8 9 40 42 24 48 47 34
59 29 13 10 42 42 27 43 40 39
45C,D 1F 1F 1F 68B,C 34D,E 19E,F 95A 50C,D 82A,B
99A 26C 10D 5D 84A 53B 32C 100A 98A 95
The letters (A–F) indicate groups with statistically significant difference in cover (Po0:05). Mean vegetation covers indicated with the same letter within a column are not significantly different.
Table 5. Kendall’s non-parametric correlation coefficients (t) between the soil variables and vegetation cover in year 2000 and 2001 Vegetation cover (2000)
Loss on ignition pH P K Mg Ca
Vegetation cover (2001)
t
P
t
P
0.3335 0.0415 0.1722 0.3210 0.1080 0.1571
0.0037 ns ns 0.0052 ns ns
0.1842 0.0320 0.0151 0.3916 0.1554 0.0618
ns ns ns 0.0007 ns ns
Correlations significant at the level of Po0:05 in bold, ns (nonsignificant)=P>0.1.
ARTICLE IN PRESS A. Brekke Skrindo, P. Anker Pedersen / Urban Forestry & Urban Greening 3 (2004) 29–37
33
Table 6. Kendall’s non-parametric correlation coefficients (t) between the soil variables (lower triangle), with significance probabilities (upper triangle) Loss on ign. Loss on ignition pH P K Mg Ca
0.4162 0.5083 0.1578 0.1643 0.1809
pH
P-AL
K-AL
Mg-AL
Ca-AL
0.0002
0.0000 0.0208
ns 0.0001 ns
ns 0.0000 0.0024 0.0000
ns 0.0000 ns 0.0009 0.0000
0.2575 0.4484 0.5722 0.7369
0.2142 0.3338 0.1142
0.4657 0.3667
0.5555
Correlations significant at the level of Po0:05 are in bold, ns (non-significant)=P>0.1.
tion with pH and P and showed a tendency (Po0:1) of negative correlation with the other inorganic soil variables (Table 6), illustrating that the organic-rich soil had a relatively low nutrient-level and low pH. Macroplot 3 had the lowest content of organic matter (Table 4) and among the lowest vegetation cover in year 2000 and 2001 (Table 3). Macro plot 8 had one of the highest organic matter content (Table 4) and also the highest vegetation cover both years. The inorganic variables made up a group of more or less significantly positively correlating variables (Table 6). The highest increase in vegetation cover took place in the macroplots (1 and 9) with intermediate vegetation cover in year 2000, on soils relatively low in organic matter but with relatively high content of mineral nutrients (Table 4).
Species frequency Since the vegetation on the site of origin is not recorded, revegetation from topsoil might not only include indigenous species but all species in the propagule bank. This is the case in most construction areas and due to practical maintenance and aesthetical aspects, the change in frequency of species is discussed in relation to the following functional groups: woody plants, flowering forbs, agricultural weeds and graminoids. The different species showed a highly variable frequency pattern: among the 121 species, 58 were recorded in more than 5 mesoplots (Table 7) and 30 occurred in five or more of the macroplots (the bold names in Table 7): five were woody plants, four were considered to be typical agricultural weeds, 11 were flowering forbs (forbs are not considered to be weeds in this context) and 10 were graminoids (including the families: Poaceae, Juncaceae and Cyperaceae). Nine of the 30 species found in five or more macroplots, occurred in 20 or more mesoplots: four woody plants (Rubus idaeus, Picea abies, Salix caprea and Pinus sylvestris), two agricultural weeds (Persicaria maculosa and Tussilago farfara), two flowering forbs (Ranunculus
repens and Galeopsis tetrahit) and one grass species (Agrostis capillaris). Some of the most frequent species accounted for the largest vegetation cover: Ranunculus repens and Tussilago farfara had greater cover percentage than the rest of the species, and both increased in cover from 2000 to 2001, T. farfara more than R. repens. Other dominant species, Picea abies, Pinus sylvestris and Galeopsis tetrahit, were low in cover. Among the 14 species recorded only in year 2000, three species were considered to be agricultural weeds, five flowering forbs and five graminoids. Among the species appearing in year 2001, all functional groups were represented although the number of flowering forbs and graminoids increased notably (only one woody species, but five agricultural weeds, nine flowering forbs and 11 graminoids). Thirteen species showed statistically significant increase (Po0:05) of frequency from 2000 to 2001 while 10 species showed significant decline (Table 7). Pinus sylvestris was the only woody plant that increased and Picea abies was the only woody species that decreased. Other species did not change statistically significantly in frequency. Four species of flowering forbs (Cerastium fontanum, Ranunculus repens, Veronica officinalis and Veronica scutellata) increased significantly (Po0:05) and Stellaria nemorum showed a tendency of increasing (Po0:1). Only Omalotheca sylvatica declined significantly. Thirteen other species of flowering forbs did not change significantly in frequency during the 2 years. The graminoids Agrostis capillaris, Carex pallescens, Deschampsia caespitosa, Glyceria fluitans, Luzula pilosa and Juncus effusus increased significantly while Alopecurus geniculatus, Poa annua and Poa nemoralis decreased significantly. Juncus articulatus and J. bufonius showed a tendency of declining (Po0:1). Ten other graminoids did not change significantly. Among the agricultural weeds Cirsium arvense and Tussilago farfara were the only ones, which increased significantly (Po0:05) and Senecio sylvaticus showed a tendency of increasing (Po0:1). Chenopodium polyspermum, Matricaria perforata, Persicaria maculosa, Rorippa palustris and Stellaria media which are wellknown agricultural weeds, declined significantly from
ARTICLE IN PRESS 34
A. Brekke Skrindo, P. Anker Pedersen / Urban Forestry & Urban Greening 3 (2004) 29–37
Table 7. Change in frequency of the vascular species occurring in more than 5 meso plots from year 2000 to 2001 Code and species
Occur.
Change
Ma
Me
n+
n
(w)Pinus sylvestris (g)Luzula pilosa (g)Juncus effusus (g)Glyceria fluitans (g)Deschampsia cespitosa (g)Carex pallescens (g)Agrostis capillaris (f)Veronica scutellata (f)Veronica officinalis (f)Stellaria nemorum (f)Ranunculus repens (f)Cerastium fontanum (a)Tussilago farfara (a)Senecio sylvaticus (a)Cirsium arvense
7 5 4 6 7 6 8 4 6 5 7 6 9 4 4
24 8 10 17 18 11 22 8 17 19 26 11 29 15 10
13 8 10 12 11 8 16 7 11 13 14 11 24 8 9
(w)Picea abies (g)Poa nemoralis (g)Poa annua (g)Juncus bufonius (g)Juncus articulatus (g)Alopecurus geniculatus (f)Omalotheca sylvatica (f)Galeopsis tetrahit (a)Stellaria media (a)Rorippa palustris (a)Persicaria maculosa (a)Matricaria perforata (a)Chenop. polyspermum
9 3 4 8 5 7 4 8 8 4 9 6 3
32 9 11 18 12 17 13 20 15 7 22 10 6
4 0 2 6 3 6 1 5 3 0 4 0 0
Species P
I
7 0 0 4 4 1 3 1 3 4 2 0 2 7 1
0.0418 0.0095 0.0064 0.0161 0.0157 0.0208 0.0027 0.0280 0.0185 0.0886 0.0318 0.0028 0.0001 0.0832 0.0080
+ + + + + + + + + + + + + + +
26 8 8 12 8 11 12 13 12 6 18 7 6
0.0000 0.0095 0.0249 0.0641 0.0559 0.0092 0.0016 0.0611 0.0381 0.0210 0.0029 0.0142 0.0210
(w)Salix caprea (w)Rubus idaeus (w)Betula pubescens (g)Poa pratensis (g)Luzula multiflora (g)Juncus filiformis (g)Juncus conglomeratus (g)Festuca rubra (g)Descha. flexuosa (g)Carex flava (g)Carex echinata (g)Carex canescens (g)Agrostis stolonifera (f)Viola riviniana (f)Trifolium repens (f)Stellaria graminea (f)Solanum dulcamara (f)Silene dioica (f)Oxalis acetosella (f)Lythrum salicaria (f)Impatiens noli-tangere (f)Galium palustre (f)Epilobium sp. (f)Epilo. angustifolium (f)Anemone nemorosa (a)Polygonum aviculare (a)Plantago major (a)Filipendula ulmaria (a)Chenopodium album (a)Urtica dioica
Occur.
Change
Ma
Me
n+
n
P
I
8 9 5 5 3 2 4 4 4 3 3 5 2 8 4 3 3 5 3 3 4 5 5 2 5 3 2 2 5 4
25 33 12 15 10 6 7 11 11 6 6 10 7 16 5 5 5 11 6 7 13 11 14 5 13 5 5 5 11 11
12 15 7 8 3 2 5 5 7 4 4 7 4 7 4 4 1 4 1 4 4 8 7 3 8 2 3 3 5 5
9 14 2 7 6 4 1 6 3 2 2 2 3 7 1 1 2 5 3 2 9 3 5 2 5 3 2 1 6 5
ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns
+ + + + + + + + + + + + + + + + + + +
The codes: (w)=woody plant, (f)=flowering forbs except (a)=agricultural weeds, (g)=graminoids, total number of occurrences in 10 macroplots (Ma) and 40 mesoplots (Me) in either 2000 or 2001; n+, n=number of mesoplots with decreasing and increasing subplot frequencies, respectively; P=statistical significance (Po0:05 in bold), Wilcoxon one-sample test, (ns=non-significant); I=indication of decreasing () or increasing (+) frequency. Species in bold were recorded in more than 50% of the mesoplots.
year 2000 to 2001. Five other weeds showed no significant change. There was no clear frequency pattern for the different functional groups related to the soil. Woody species occurred in all plots, but more species were recorded in macro plots 5–10 than in the rest. The agricultural weeds occurred in all macro plots except macroplot 3 but only a few were present in plots 2, 4 and 7. The flowering forbs and the graminoids occurred in all macroplots, but the highest number of species was recorded in macroplots 1, 8, 9 and 10 for the forbs and 1, 2, 5 and 6 for the graminoids.
Discussion When driving on the study road 2 years after the topsoil was redistributed, the vegetation appears to fit
well into the surrounding indigenous vegetation even though the species composition is far from identical to the surrounding and the vegetation cover varies from less than 10% to more than 90%. The variation between the macro plots from year 2000 to 2001 could not be explained by the soil variables only. The significant relationship between the vegetation cover and organic matter (loss on ignition) in year 2000 is well documented from other studies. Bradshaw (1989) found increased growth yield in topsoils with high content of organic matter. Soils with a high content of organic matter are likely to have a larger seed bank than soils with lower organic matter content (cf. . Granstrom, 1982). Somewhat surprisingly there was no significant relation between the vegetation cover and the content of organic matter in year 2001. However, the subsoil with high clay content may have provided more essential nutrients than the organic soil and thus support good growth. This is a possible explanation why some
ARTICLE IN PRESS A. Brekke Skrindo, P. Anker Pedersen / Urban Forestry & Urban Greening 3 (2004) 29–37
macro plots with low content of organic matter, but high clay content have rather high vegetation cover (e.g. macroplots 6 and 9). Other possible explanations could lie in processes during stockpiling, uneven distribution of organic soil on the sites, the shorter growth period at macro plots 1–4 and other ecological factors not recorded in this study . The quality of the organic soil is probably of major importance as illustrated in macroplot 4 with high content of moderately decomposed Sphagnum sp. Slow mineralisation leads to limited amount of available nitrogen, resulting in low vegetation cover. Nitrogen is usually considered to be the most important limiting resource in boreal forests (cf. Hesselmann, 1937; Tamm, 1991). The thickness of the excavated topsoil influences the availability of propagules. The number of viable seeds from different depths has been studied in different . (1982) found most ecosystems. For instance, Granstrom viable seeds in the thin humus layer in five different boreal forest stands. The humus—mineral soil ratio varies with the thickness of the humus layer. Consequently, number of seeds is likely to be higher in macroplots with topsoil rich in organic matter. The quantity of redistributed topsoil also affects the revegetation. The topsoil layer in this study after redistribution was supposed to be 10 cm thick, but there are patches of thinner and thicker topsoil layer due to the inaccuracy of the distribution method. Several scientists working in the field of reclamation have accepted the hypothesis that the deeper the humus layer (up to the natural depth), the more productive the plant community (Power et al., 1976; McGinnies and Nicholas, 1980; Redente and Hargis, 1985). However, Redente et al. (1997) found no difference in vegetation cover after 10 years although the thickness of the humus layer varied from 15 to 60 cm. When comparing seed bank studies and revegetation from stockpiled topsoil (including the propagule bank, the micro fauna, flora and the nutrition), the effects of stockpiling on the soil quality have to be taken into account. Hargis and Redente (1984) found decreased germination and revegetation success and Rives et al. (1980) found less mycorrhiza-infected roots in stockpiled topsoil. The soil used in this study had been stockpiled for 1–1.5 years but good revegetation results are obtained from different soiltypes up to 5 years of storage according to Stark and Redente (1987). The revegetation process in macroplots 1–4 was 2 months shorter than in the other plots due to the delayed redistribution of topsoil. The variation between the macro plots, however, cannot be explained by this delay alone because the difference between the macroplots with the same growth period was significant as well. In spite of good climatic conditions for germination and growth throughout the summer some macro
35
plots still had a vegetation cover less than 10% in year 2001. The propagule bank size does influence the vegetation cover demonstrated by the correlation between the number of species and the vegetation cover. In this study, some of the most frequent species had the largest cover percentage (i.e. Ranunculus repens, Tussilago farfara and Rubus idaeus). However, other frequent species did not contribute much to the vegetation cover (Picea abies, P. sylvestris and Galeopsis tetrahit). The change in species composition over time is demonstrated by the 14 species recorded only in the first year and 27 new species appearing the second year. This considerable alteration is typical in the first phase of a secondary succession due to variation in the microclimate, variation in the propagule bank and due to seed dispersal from the surroundings (Connell and Slatyer, 1977). The 121 species recorded in this study are dominated by pioneer species but species expected to establish throughout the whole succession period occurred as well. Connell and Slatyer (1977) found that individuals of any species that happened to fulfil their germination requirements could establish in this early stage. Among the species with altered frequency from 2000 to 2001 some grasses (Agrostis capillaris, Deschampsia caespitosa and Glyceria fluitans) increased and others (Alopecurus geniculatus, Poa annua and P. nemoralis) decreased and other graminoids (Carex pallescens, Luzula pilosa and Juncus effusus) increased. All of these species are often found in moist places, but the species that decreased are known to loose in competition for light (cf. Lid and Lid, 1994). As woody plants may represent a maintenance problem and influence traffic safety, their slow invasion is good news. Only P. sylvestris increased significantly, whereas Picea abies decreased. Rubus idaeus was among the most frequent species, which was expected because of the viability of the seeds in comparable seed banks . (cf. Granstrom, 1982). Salix caprea was also among the most frequent species even without a persistent seed bank, because it disperses and germinates easily (cf. Brinkman, 1974). The weeds are neither an aesthetic nor a management problem yet. Among the 16 species that increased significantly in frequency, only two can be considered as weeds (Cirsium arvense and Tussilago farfara). If they continue to increase, they might dominate some areas completely and prevent other indigenous species to grow. Six of the nine species that decreased were also typical weeds (Chenopodium polyspermum, Matricaria perforata, Persicaria maculosa, Rorippa palustris and Stellaria media). All of these species are annuals or at least short-lived (Korsmo, 1986; Lid and Lid, 1994) and might be expected to decrease. There were several weeds that did not change in frequency, and their future presence remains unknown.
ARTICLE IN PRESS 36
A. Brekke Skrindo, P. Anker Pedersen / Urban Forestry & Urban Greening 3 (2004) 29–37
The macro plots are situated in a landscape with different vegetation history and ecology influencing the propagule bank as argued by Baskin and Baskin (1998): The size and the content of the propagule bank varies with plant age, species composition, disturbance level, predispersal seed predation and plant seed production. There is no clear relationship between the soil properties and the species composition except in macroplot 4 with organic soil that had few weeds, woody plants and graminoids, and in plots 2, 3 and 7 with loamy soil that host none or only a few weeds. Several studies show a lack of correspondence between the species composition of the seed bank and the associated plant community in surrounding forests (e.g. Oosting and Humphreys, 1940; Thompson and . Grime, 1979; Granstrom, 1982). The species composition in the early phase of succession seems hard to predict. That the species composition in the early phase of succession is hard to predict is illustrated by the variation between the macroplots in this study. In general, however, both the vegetation cover and species composition were satisfactory already 2 years after the soil was redistributed. Natural revegetation from redistributed topsoil is therefore recommended in comparable ecosystems.
Acknowledgements This project was partly financed by the Public Roads Administration, Buskerud county and the Directorate of Public Roads in Norway.
References Baskin, C.B., Baskin, J.M., 1998. Seeds. Ecology, Biogeography, and Evolution of Dormancy and Germination. Academic Press. San Diego, USA. Begon, M., Harper, J.L., Townsend, C.R., 1990. Ecology. Blackwell Scientific Publications, Oxford. Bradshaw, A.D., 1989. The quality of topsoil. Soil Use Management 5, 101–108. Brinkman, J.A., 1974. Salix. In: Young, J.A., Young, C.G. (Eds.), Seeds of Woody Plants in the United States Agricultural Handbook No. 450. Forest Service, US Department of Agriculture, Washington, DC, pp. 746–749. Connell, J.H., Slatyer, R.O., 1977. Mechanisms of succession in natural communities and their role in community stability and organisation. American Naturalist 111, 1119–1144. Egn!er, H., Riehn, H., Domingo, W.R., 1960. Untersuchnungen uber . die chemische Bodenanalyse als Grundlage fur . der . Beurteilung des N.ahrstoffzustandes der Boden. II. Chemische Extractions Methoden zur Phosphor und Kalium. ok, . bestimmung. Statens Jordbruksfors S.artryck och sm(askrifter, Uppsala (in German).
Førland, E.J., 1993. Nedbørnormaler, normalperioden 1961– 1990. (Mean precipitation, 1961–1990) Norske meteorologiske institutt. Rapport Klima 9, 1–63 (in Norwegian). . Granstrom, A., 1982. Seed banks in five boreal forest stands originating between 1810 and 1963. Canadian Journal of Botany 60, 1821–1855. Hargis, N.E., Redente, E.F., 1984. Soil handling for surface mine reclamation. Journal of Soil and Water Conservation 39, 300–305. Hesselmann, H., 1937. Om humust.ackets beroende av besta( ndets a( lder och sammans.attning i den nordiska granskogen av bla( b.arsrik Vaccinium-typ och dess inverkan pa( . skogens foryngring och tillv.axt. (Effects of tree stand age and species composition on the humus layer, forest regeneration, and forest yield in a bilberry-dominated Norway spruce forest) Medd. Statens Skogsforskningsanstalt, Vol. 30, pp. 529–716 (in Swedish). Holmes, P.M., 2001. Shrubland restoration following woody alien invasion and mining: effects of topsoil depth, seed source, and fertilizer addition. Restoration Ecology 9, 71–84. Kendall, M.G., 1938. A new measure of rank correlation. Biometrika 30, 81–93. Korsmo, E., Vidme, E., Fykse, H., 1986. Korsmos ugrasplansjer (Korsmos weed illustrations) Landbruksforlaget, Oslo (in Norwegian) Kotanen, P.M., 1996. Revegetation following soil disturbance in a California meadow: the role of propagule supply. Oecologia 108, 652–662. Lid, J., Lid, D.T., 1994. Norsk flora (Norwegian flora) Det norske samlaget, Oslo (in Norwegian). McGinnies, W.J., Nicholas, P.J., 1980. Effect of topsoil thickness and nitrogen fertilization in the revegetation of coal mine spoils. Journal of Environmental Quality 9, 681–685. Økland, T., 1988. An ecological approach to the investigation of a beech forest in Vestfold, SE. Norway. Nordic Journal of Botany 8, 375–407. Økland, R.H., 1990. Vegetation ecology: theory, methods and applications with reference to fennoscandia. Sommerfeltia 2 (Suppl.), 1–233. Økland, R.H., Økland, T., Rydgren, K., 2001. A Scandinavian perspective on ecological gradients in north-west European mires. Journal of Ecology 89 (3), 481–486. Oosting, H.J., Humphreys, M.E., 1940. Buried viable seeds in a successional series of old field and forest soils. Bulletin of the Torrey Botanical Club 67, 253–273. Power, J.F., Ries, R.E., Sandoval, F.M., 1976. Use of soil materials on spoils—effects of thickness and quality. North Dakota Farm Research 34, 23–24. Redente, E.F., Hargis, N.E., 1985. An evaluation of soil thickness and manipulation of soil and spoil for reclaiming mined land in north-west Colorado. Reclamation and Revegetation Research 4, 17–29. Redente, E.F., McLendon, T., Agnew, W., 1997. Influence of topsoil depth on plant community dynamics of a seeded site in Northwest Colorado. Arid Soil Research and Rehabilitation 11, 139–149. Rives, C.S., Bajwa, M.I., Liberta, A.E., Miller, R.M., 1980. Effect of topsoil storage during surface mining on the viability of VA mycorrhiza. Soil Science 129, 253–257.
ARTICLE IN PRESS A. Brekke Skrindo, P. Anker Pedersen / Urban Forestry & Urban Greening 3 (2004) 29–37
Rokich, D.P., Dixon, K.W., Sivasithamparam, K., Meney, K.A., 2000. Topsoil handling and storage effects on woodland restoration in Western Australia. Restoration Ecology 8, 196–208. SAS Institute Inc., 1987. SAS/STAT user’s Guide. Version 6, Fourth Edition, Volume 2, SAS Institute Inc., Cary, NC, USA. Shaw, P.J.A., 1996. Role of seedbank substrates in the revegetation of fly ash and gypsum in the United Kingdom. Restoration 4, 61–70. Sokal, R.R., Rohlf, F.J., 1995. Biometry: The Principles and Practice of Statistics in Biological Research. Freeman, New York. Stark, N.M., Redente, E.F., 1987. Production potential of stockpiled topsoil. Soil Science 144, 72–76. Statens V., 1992. Veg-og gateutforming. (Road and street construction) Vegvesenets h(andbøker nr. 017. Oslo, Norway (in Norwegian).
37
Sveistrup, T., Njøs, A., 1984. Kornstørrelsesgrupper i mineraljord (Textural classes in mineral soils.). Jord og Myr 8, 8–15 (in Norwegian). Tamm, C.O., 1991. Nitrogen in terrestrial ecosystems. Questions of productivity, vegetational changes and ecosystem stability. Ecological Studies 81, 1–116. Thompson, K., Grime, J.P., 1979. Seasonal variation in the seed banks of herbaceous species in ten contrasting habitats. Journal of Ecology 67, 893–921. Warr, S.J., Thompson, K., Kent, M., 1993. Seed banks as a neglected area of biogeographic research—a review of literature and sampling techniques. Progress in Physical Geography 17, 29–347. Zhang, Z.Q., Shu, W.S., Lan, C.Y., Wong, M.H., 2001. Soil seed bank as an input of seed source in revegetation of lead/zinc mine tailings. Restoration Ecology 9, 378–385.
Urban Forestry & Urban Greening 3 (2004) 29–37 www.elsevier.de/ufug
Natural revegetation of indigenous roadside vegetation by propagules from topsoil Astrid Brekke Skrindo*, Per Anker Pedersen ( Norway Department of Plant and Environmental Sciences, Agricultural University of Norway, Box 5003, 1432 As, Received 16 June 2003; accepted 2 December 2003
Abstract Road construction degrades large areas of indigenous vegetation. Better understanding of ecosystem restoration after degradation will require development and assessment of improved restoration techniques. The potential of natural revegetation based on propagules from topsoil as a restoration technique, to achieve succession towards indigenous vegetation, is not fully understood and accepted as an alternative method for vegetation establishment. The aim of this study was to evaluate this restoration technique by comparing different topsoils with respect to their potentials for revegetation and to describe the early succession on these topsoils by recording change in vegetation cover and the number of species. The study road RV23 is situated in south-east Norway and runs mainly through mixed, mostly coniferous forest. The topsoil (the upper 30 cm) was removed before road construction started, stockpiled during the construction period and then redistributed on the road verge afterwards. In 1999, 10 macroplots of 25 m2 were placed randomly on selected topsoils of different qualities, to represent the soil variation in the study area. The soil quality varied from organic-rich to organic-poor soils. Each macro plot contained four 1-m2 mesoplots in which the species composition was recorded in 2000 and 2001.Univariate statistical tests were applied to reveal change in vegetation cover between macroplots within 1 year and between years 2000 and 2001. Furthermore, statistical tests were used to reveal the change in singlespecies frequencies between the study years. Two years after deposition of topsoils, the vegetation cover was generally satisfactory from an aesthetical point of view. However, the variation in vegetation cover and species composition was considerable on different topsoils, reflecting both the local variation of the indigenous vegetation and the soil quality. A total of 121 species of vascular plants were recorded in all 10 macroplots, although not more than 59 species were found in one macroplot. Species composition and single-species frequencies changed considerably from the first year to the second. Among the 61 species that were recorded in more than five mesoplots, the frequency of 16 species increased and the frequency of 10 species decreased significantly (Po0:05). Most of the decreasing species can be considered as weed species and are not represented in the present indigenous vegetation, while most of the increasing species are. The vegetation change from 2000 to 2001 apparently represents the first steps in a succession towards an ecosystem dominated by species of the indigenous vegetation. r 2004 Elsevier GmbH. All rights reserved. Keywords: Revegetation; Indigenous vegetation; Roadside; Topsoil; Propagule bank
*Corresponding author. E-mail address: [email protected] (A. Brekke Skrindo). 1618-8667/$ - see front matter r 2004 Elsevier GmbH. All rights reserved. doi:10.1016/j.ufug.2004.04.002
ARTICLE IN PRESS 30
A. Brekke Skrindo, P. Anker Pedersen / Urban Forestry & Urban Greening 3 (2004) 29–37
Introduction Road construction degrades large areas of indigenous vegetation. When building new roads, the added verges make up a significant area to be revegetated for environmental and aesthetical reasons and to prevent erosion. Preservation and re-establishing indigenous vegetation helps to reach the official goal: to restore degraded areas and to improve road aesthetics within the natural landscape (Statens, 1992). There are four ways to re-establish indigenous vegetation: (1) Transplant larger patches of vegetation, (2) cultivate indigenous plants, (3) in situ germination of seeds collected from the surrounding areas and (4) germination from the topsoil propagule bank and the areas adjacent to the road. In Norway, seeding of non-indigenous grasses is one of the most used methods along roads resulting in a somewhat monotonous green lawn. Since the three first methods are expensive and time-consuming, natural revegetation from topsoil is an appealing alternative and the topic of this study. Topsoil is defined differently throughout literature. Hargis and Redente (1984) present three possible definitions for topsoil: soil from the A-horizon; soils from the A and E-horizon; or a specific soil depth regardless of soil layer. In this study, the latter definition is used for practical reasons. Revegetation from topsoil is based on germination from the propagule bank followed by the process of natural succession, defined as the nonseasonal, directional and continuous pattern of colonisation and extinction on a site by species populations (Begon et al., 1990). The procedure is straightforward: topsoil is removed before construction, stockpiled and then replaced. The advantages are that plant nutrients, viable propagules, mychorriza and microfauna remain in the upper soil (cf. Hargis and Redente, 1984). Restoration of indigenous vegetation from the topsoil propagule seed bank shows contradictory results. It has failed in several grasslands (Thompson and Grime, 1979), but succeeded in South African shrubland (Holmes, 2001), in a Californian (USA) meadow (Kotanen, 1996) and in woodland restoration in Western Australia (Rokich et al., 2000) and in England (cf. Warr et al., 1993). Furthermore, it has proved to be successful for revegetation of contaminated soils, such as lead/zinc mine tailings in China (Zhang et al., 2001) and fly ash and gypsum in the United Kingdom (Shaw, 1996). To our knowledge, restoration of indigenous roadside vegetation from the propagule bank in the topsoil has not been assessed before in Scandinavia. The study sites were integrated in the construction of a new road (‘‘Oslofjordforbindelsen’’), which passes through various vegetation types. The revegetation methods along the study road are a mosaic of revegetation from topsoil and seeded and planted areas, but this investigation deals only with revegetation from
redistributed topsoil. Since this study consists of data from only 2 years, we focus on the first phase of a secondary autogenic succession. The aim of the study was to evaluate the redistribution of topsoil as a restoration technique and to compare revegetation success from different topsoils. Particular attention was given to change in vegetation cover, the number of species, relationship between soil and vegetation and the frequency of species divided into four functional groups: woody plants, flowering forbs, agricultural weeds and graminoids.
Materials and methods Study area and restoration technique The study road, RV23 (Oslofjordforbindelsen), passes through 13 km of mixed forest, coniferous forest dominated by Scots pine (Pinus sylvestris) and Norway spruce (Picea abies) and bogs. The road is situated in south-east Norway in the municipalities of Frogn, Hurum and Røyken in Akershus and Buskerud counties respectively (10 440 E, 59 420 N to 10 250 E, 59 440 N). The climate is sub-oceanic. Annual mean precipitation in the years 1961–1990 (normal) was 920 mm at the Drøbak meteorological station (Førland, 1993). The annual precipitation in the study period was 1191.7 mm (1999), 1360.8 mm (2000) and 991.7 mm (2001). The precipitation during the growing season was higher or about the same as the normal with no continuous drought periods. The topsoil, defined by the Public Roads Administration as the upper 30 cm of the soil profile due to practical reasons, was excavated before road construction and stockpiled separately during the construction period (in 1998). The topsoil volume decreased during stockpiling, consequently a layer of 10 cm was redistributed on top of the subsoil on the road verge (in May and August 1999). The exact origin of the soil replaced on plots is not known due to some degree of mixing of soil during stockpiling. The vegetation composition in the original site is therefore unknown. The thickness of the humus layer in the original soil profile varied between 5 and 20 cm except in bogs.
Sampling Recording of vegetation data In 1999, 10 macroplots of 25 m2 were placed randomly on selected topsoils of different qualities representing the topsoil-variation in the study area. Each macroplot contained four 1 m2 mesoplots systematically placed in the corners by grid sampling (Økland, 1990). Each mesoplot was divided into 16 (25 25 cm2) subplots.
ARTICLE IN PRESS A. Brekke Skrindo, P. Anker Pedersen / Urban Forestry & Urban Greening 3 (2004) 29–37
In July 2000 and 2001, the frequency of all vascular plant species in the 40 mesoplots was recorded by (1) estimating percentage cover on a 0100 scale, and (2) by recording the presence/absence of all species in each subplot to calculate subplot frequency on a 016 scale (Økland, 1988). The vascular plants nomenclature followed Lid and Lid (1994).
31
weight (unit: mg/100 ml). For soil variables, skewness and kurtosis were calculated using Statgraphics, Version 5.0 (Table 1). Homogeneity of variances (Table 1) was achieved by transforming variables to zero skewness by ln (c þ x) according to Økland et al. (2001). Statistical significance of differences in total vegetation cover (%) between the macroplots in each year 2000 and 2001 and the change between the years 2000 and 2001, were tested using an analysis of variance including a multiple range test, Ryan–Einot–Gabriel–Welsh (REGWQ), Po0:05 (SAS Institute, 1987) according to the model
Recording of soil data Soil samples were collected from the upper 10 cm layer on the boundaries of each mesoplot and analysed to provide soil variables (Table 1): Readily available plant nutrients, P, K, Mg, Ca, were extracted by 0.1 M ammonium lactate and 0.4 M acetic acid according to the AL-method (Egne! r et al., 1960) and determined by inductively coupled spectroscopy (ICP). The content of organic matter was assessed by determining the loss on ignition and soil pH was measured in water suspension. The soil texture was classified (Table 2) according to Sveistrup and Njøs (1984).
Xij ¼ mi þ ai þ eij ; where Xij is the performance of the macroplot, i the ði ¼ 1; 2; :::10Þ; and due to the mesoplot, j the ðj ¼ 1; 2; :::4Þ; m the expected mean, a the expected effect of the macroplot, e the residual (variation between the mesoplots). The hypothesis tested was a ¼ 0 (no effect of the macroplot) against aa0: Relationships between the number of species and the vegetation cover in the macroplots were addressed using Kendall’s rank correlation method (Kendall, 1938). To address the relationships between the soil variables and the relationship between the soil and vegetation cover Kendall’s rank correlation method (Kendall, 1938) was
Statistical analyses All soil variables except pH were adjusted to the soil density by multiplying the variable to the soil volume Table 1. Soil variables measured in all 1 m2 plots Variable
Loss on ignition pH P K Mg Ca
Unit
Summary statistics of untransformed variable
mg/100 ml — mg/100 ml mg/100 ml mg/100 ml mg/100 ml
Transf. ln (c þ x)
Summary statistics of transformed variable
Range
S.S.
S.K.
c-value
Mean
S.D
S.K
1–400 4.2–8.6 5.7–383 5.2–65.7 2.68–45 9.93–496
1.58 0.68 2.78 0.04 2.13 4.60
1.16 1.25 0.22 1.85 0.16 3.80
0.65 1.00 1.00 0.1 — 4.23
2.25 1.95 1.95 0.99 — 2.09
1.10 0.16 0.16 0.34 — 1.00
1.51 1.17 1.17 0.53 — 0.38
Abbreviations: S.S.=standardised skewness; S.K.=standardised kurtosis, c-value=index used in transformation formula; S.D.=standard deviation, =analysed by the AL-method.
Table 2. Soil classification of the macroplots Macro plot
Soil classification
Organic components
1 2 3 4 5 6 7 8 9 10
Silt loam Silt loam Medium loamy sand Organic soil Mold rich loam Clay loam+silt loam Silt loam+loam+medium loamy sand Mold rich loam Clay loam+silt loam Mold rich loam
Litter—weakly decomposed Litter—variable degree of decomposition Litter—weakly decomposed Sphagnum—weakly decomposed Litter—weakly decomposed Sphagnum—weakly decomposed Sphagnum and litter—weakly decomposed Litter—weakly decomposed Litter—well decomposed Litter—weakly decomposed
Description of the organic components; litter originating from dry coniferous woodland, litter originating from swamp woodland.
ARTICLE IN PRESS 32
A. Brekke Skrindo, P. Anker Pedersen / Urban Forestry & Urban Greening 3 (2004) 29–37
used. The statistical difference of single species frequencies between years 2000 and 2001 was tested using Wilcoxon’s signed-rank test for two paired samples (cf. Sokal and Rohlf, 1995). The tests were against the twotailed alternatives. Testing was restricted to species recorded in 5 or more mesoplots either year.
Results Vegetation cover and the number of species Two years after the redistribution of the topsoil, six macroplots had at least 50% vegetation cover. The mean vegetation cover ranged from 1% to 95% in 2000 and from 5% to 100% in 2001. The macroplots differed statistically (Po0:001) both years, and there was no significant difference within each macroplot. In year 2000, no distinct groups of macroplots could be statistically defined (Table 3). In 2001, the macroplots could be divided into four groups: macroplots 1, 5, 8, 9, 10 had the highest vegetation cover followed by plot 6 then plots 2 and 7, and finally plots 3 and 4 (Table 3). The vegetation cover increased significantly in all the macroplots from year 2000 to 2001 (Table 3). In general, plots with high cover vegetation cover in 2001 had high cover in 2000 as well. On the other hand, in plots with initially low vegetation cover (1%) the differences between year 2000 and 2001 varied substantially. There were 121 recorded species in total and the number of species per macroplot ranged from 8 to 51 in year 2000 and from 10 to 59 in year 2001 (Table 3). Fourteen species were found only in the year 2000 and 27 species appeared for the first time in year 2001. There were strong correlations between the vegetation cover
and the number of species both years (year 2000; t ¼ 0:5004; P ¼ 0:0001; year 2001; t ¼ 0:442; P ¼ 0:0002).
Relationships between vegetation cover and soil variables The content of plant nutrients in the soil varied between plots (Table 4). In most samples the content was medium high. The organic soil in macroplot 4 was low in all elements. On the contrary, the macroplots with clay loam (6 and 9) had the highest content of all nutrients except for Ca, which was highest in the medium loamy sand of macroplot 3. Vegetation cover did not correlate strongly with most soil variables (Table 5). The only significant relationship in both years at level Po0:01 was with K, and organic matter (Loss on ignition) in year 2000. There were rather strong correlations between some of the soil variables (Table 6) indicating a more complex relationship between the soil variables and the vegetation cover than the direct correlation explained. The loss on ignition had statistically significant negative correlaTable 4. Mean values of the soil variables Macro Loss on pH P-AL K-AL Mg-AL Ca-AL Density plot ignition 1 2 3 4 5 6 7 8 9 10
6.4 7.9 2.8 14.1 15.4 7.5 4.6 11.8 6.4 23.7
5.4 5.2 8.0 4.3 5.5 7.0 7.4 6.2 7.2 5.8
2.8 2.3 2.2 1.6 0.8 3.4 4.7 1.0 5.1 1.4
9.8 7.2 11.0 3.0 9.5 15.2 11.7 26.8 29.0 9.5
15.0 3.6 21.4 8.1 12.8 31.2 29.6 13.9 28.6 23.1
63.0 30.8 1000.4 34.9 117.6 442.3 390.4 267.9 296.8 484.0
0.95 1.12 1.16 0.43 0.42 0.98 1.24 0.51 1.20 0.70
The unit for all variables (except pH) is mg/100 ml.
Table 3. The number of species and mean vegetation cover in each macro plot in years 2000 and 2001 Macroplot Number of species
1 2 3 4 5 6 7 8 9 10
Mean vegetation cover (%)
2000
2001
2000
2001
51 17 8 9 40 42 24 48 47 34
59 29 13 10 42 42 27 43 40 39
45C,D 1F 1F 1F 68B,C 34D,E 19E,F 95A 50C,D 82A,B
99A 26C 10D 5D 84A 53B 32C 100A 98A 95
The letters (A–F) indicate groups with statistically significant difference in cover (Po0:05). Mean vegetation covers indicated with the same letter within a column are not significantly different.
Table 5. Kendall’s non-parametric correlation coefficients (t) between the soil variables and vegetation cover in year 2000 and 2001 Vegetation cover (2000)
Loss on ignition pH P K Mg Ca
Vegetation cover (2001)
t
P
t
P
0.3335 0.0415 0.1722 0.3210 0.1080 0.1571
0.0037 ns ns 0.0052 ns ns
0.1842 0.0320 0.0151 0.3916 0.1554 0.0618
ns ns ns 0.0007 ns ns
Correlations significant at the level of Po0:05 in bold, ns (nonsignificant)=P>0.1.
ARTICLE IN PRESS A. Brekke Skrindo, P. Anker Pedersen / Urban Forestry & Urban Greening 3 (2004) 29–37
33
Table 6. Kendall’s non-parametric correlation coefficients (t) between the soil variables (lower triangle), with significance probabilities (upper triangle) Loss on ign. Loss on ignition pH P K Mg Ca
0.4162 0.5083 0.1578 0.1643 0.1809
pH
P-AL
K-AL
Mg-AL
Ca-AL
0.0002
0.0000 0.0208
ns 0.0001 ns
ns 0.0000 0.0024 0.0000
ns 0.0000 ns 0.0009 0.0000
0.2575 0.4484 0.5722 0.7369
0.2142 0.3338 0.1142
0.4657 0.3667
0.5555
Correlations significant at the level of Po0:05 are in bold, ns (non-significant)=P>0.1.
tion with pH and P and showed a tendency (Po0:1) of negative correlation with the other inorganic soil variables (Table 6), illustrating that the organic-rich soil had a relatively low nutrient-level and low pH. Macroplot 3 had the lowest content of organic matter (Table 4) and among the lowest vegetation cover in year 2000 and 2001 (Table 3). Macro plot 8 had one of the highest organic matter content (Table 4) and also the highest vegetation cover both years. The inorganic variables made up a group of more or less significantly positively correlating variables (Table 6). The highest increase in vegetation cover took place in the macroplots (1 and 9) with intermediate vegetation cover in year 2000, on soils relatively low in organic matter but with relatively high content of mineral nutrients (Table 4).
Species frequency Since the vegetation on the site of origin is not recorded, revegetation from topsoil might not only include indigenous species but all species in the propagule bank. This is the case in most construction areas and due to practical maintenance and aesthetical aspects, the change in frequency of species is discussed in relation to the following functional groups: woody plants, flowering forbs, agricultural weeds and graminoids. The different species showed a highly variable frequency pattern: among the 121 species, 58 were recorded in more than 5 mesoplots (Table 7) and 30 occurred in five or more of the macroplots (the bold names in Table 7): five were woody plants, four were considered to be typical agricultural weeds, 11 were flowering forbs (forbs are not considered to be weeds in this context) and 10 were graminoids (including the families: Poaceae, Juncaceae and Cyperaceae). Nine of the 30 species found in five or more macroplots, occurred in 20 or more mesoplots: four woody plants (Rubus idaeus, Picea abies, Salix caprea and Pinus sylvestris), two agricultural weeds (Persicaria maculosa and Tussilago farfara), two flowering forbs (Ranunculus
repens and Galeopsis tetrahit) and one grass species (Agrostis capillaris). Some of the most frequent species accounted for the largest vegetation cover: Ranunculus repens and Tussilago farfara had greater cover percentage than the rest of the species, and both increased in cover from 2000 to 2001, T. farfara more than R. repens. Other dominant species, Picea abies, Pinus sylvestris and Galeopsis tetrahit, were low in cover. Among the 14 species recorded only in year 2000, three species were considered to be agricultural weeds, five flowering forbs and five graminoids. Among the species appearing in year 2001, all functional groups were represented although the number of flowering forbs and graminoids increased notably (only one woody species, but five agricultural weeds, nine flowering forbs and 11 graminoids). Thirteen species showed statistically significant increase (Po0:05) of frequency from 2000 to 2001 while 10 species showed significant decline (Table 7). Pinus sylvestris was the only woody plant that increased and Picea abies was the only woody species that decreased. Other species did not change statistically significantly in frequency. Four species of flowering forbs (Cerastium fontanum, Ranunculus repens, Veronica officinalis and Veronica scutellata) increased significantly (Po0:05) and Stellaria nemorum showed a tendency of increasing (Po0:1). Only Omalotheca sylvatica declined significantly. Thirteen other species of flowering forbs did not change significantly in frequency during the 2 years. The graminoids Agrostis capillaris, Carex pallescens, Deschampsia caespitosa, Glyceria fluitans, Luzula pilosa and Juncus effusus increased significantly while Alopecurus geniculatus, Poa annua and Poa nemoralis decreased significantly. Juncus articulatus and J. bufonius showed a tendency of declining (Po0:1). Ten other graminoids did not change significantly. Among the agricultural weeds Cirsium arvense and Tussilago farfara were the only ones, which increased significantly (Po0:05) and Senecio sylvaticus showed a tendency of increasing (Po0:1). Chenopodium polyspermum, Matricaria perforata, Persicaria maculosa, Rorippa palustris and Stellaria media which are wellknown agricultural weeds, declined significantly from
ARTICLE IN PRESS 34
A. Brekke Skrindo, P. Anker Pedersen / Urban Forestry & Urban Greening 3 (2004) 29–37
Table 7. Change in frequency of the vascular species occurring in more than 5 meso plots from year 2000 to 2001 Code and species
Occur.
Change
Ma
Me
n+
n
(w)Pinus sylvestris (g)Luzula pilosa (g)Juncus effusus (g)Glyceria fluitans (g)Deschampsia cespitosa (g)Carex pallescens (g)Agrostis capillaris (f)Veronica scutellata (f)Veronica officinalis (f)Stellaria nemorum (f)Ranunculus repens (f)Cerastium fontanum (a)Tussilago farfara (a)Senecio sylvaticus (a)Cirsium arvense
7 5 4 6 7 6 8 4 6 5 7 6 9 4 4
24 8 10 17 18 11 22 8 17 19 26 11 29 15 10
13 8 10 12 11 8 16 7 11 13 14 11 24 8 9
(w)Picea abies (g)Poa nemoralis (g)Poa annua (g)Juncus bufonius (g)Juncus articulatus (g)Alopecurus geniculatus (f)Omalotheca sylvatica (f)Galeopsis tetrahit (a)Stellaria media (a)Rorippa palustris (a)Persicaria maculosa (a)Matricaria perforata (a)Chenop. polyspermum
9 3 4 8 5 7 4 8 8 4 9 6 3
32 9 11 18 12 17 13 20 15 7 22 10 6
4 0 2 6 3 6 1 5 3 0 4 0 0
Species P
I
7 0 0 4 4 1 3 1 3 4 2 0 2 7 1
0.0418 0.0095 0.0064 0.0161 0.0157 0.0208 0.0027 0.0280 0.0185 0.0886 0.0318 0.0028 0.0001 0.0832 0.0080
+ + + + + + + + + + + + + + +
26 8 8 12 8 11 12 13 12 6 18 7 6
0.0000 0.0095 0.0249 0.0641 0.0559 0.0092 0.0016 0.0611 0.0381 0.0210 0.0029 0.0142 0.0210
(w)Salix caprea (w)Rubus idaeus (w)Betula pubescens (g)Poa pratensis (g)Luzula multiflora (g)Juncus filiformis (g)Juncus conglomeratus (g)Festuca rubra (g)Descha. flexuosa (g)Carex flava (g)Carex echinata (g)Carex canescens (g)Agrostis stolonifera (f)Viola riviniana (f)Trifolium repens (f)Stellaria graminea (f)Solanum dulcamara (f)Silene dioica (f)Oxalis acetosella (f)Lythrum salicaria (f)Impatiens noli-tangere (f)Galium palustre (f)Epilobium sp. (f)Epilo. angustifolium (f)Anemone nemorosa (a)Polygonum aviculare (a)Plantago major (a)Filipendula ulmaria (a)Chenopodium album (a)Urtica dioica
Occur.
Change
Ma
Me
n+
n
P
I
8 9 5 5 3 2 4 4 4 3 3 5 2 8 4 3 3 5 3 3 4 5 5 2 5 3 2 2 5 4
25 33 12 15 10 6 7 11 11 6 6 10 7 16 5 5 5 11 6 7 13 11 14 5 13 5 5 5 11 11
12 15 7 8 3 2 5 5 7 4 4 7 4 7 4 4 1 4 1 4 4 8 7 3 8 2 3 3 5 5
9 14 2 7 6 4 1 6 3 2 2 2 3 7 1 1 2 5 3 2 9 3 5 2 5 3 2 1 6 5
ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns
+ + + + + + + + + + + + + + + + + + +
The codes: (w)=woody plant, (f)=flowering forbs except (a)=agricultural weeds, (g)=graminoids, total number of occurrences in 10 macroplots (Ma) and 40 mesoplots (Me) in either 2000 or 2001; n+, n=number of mesoplots with decreasing and increasing subplot frequencies, respectively; P=statistical significance (Po0:05 in bold), Wilcoxon one-sample test, (ns=non-significant); I=indication of decreasing () or increasing (+) frequency. Species in bold were recorded in more than 50% of the mesoplots.
year 2000 to 2001. Five other weeds showed no significant change. There was no clear frequency pattern for the different functional groups related to the soil. Woody species occurred in all plots, but more species were recorded in macro plots 5–10 than in the rest. The agricultural weeds occurred in all macro plots except macroplot 3 but only a few were present in plots 2, 4 and 7. The flowering forbs and the graminoids occurred in all macroplots, but the highest number of species was recorded in macroplots 1, 8, 9 and 10 for the forbs and 1, 2, 5 and 6 for the graminoids.
Discussion When driving on the study road 2 years after the topsoil was redistributed, the vegetation appears to fit
well into the surrounding indigenous vegetation even though the species composition is far from identical to the surrounding and the vegetation cover varies from less than 10% to more than 90%. The variation between the macro plots from year 2000 to 2001 could not be explained by the soil variables only. The significant relationship between the vegetation cover and organic matter (loss on ignition) in year 2000 is well documented from other studies. Bradshaw (1989) found increased growth yield in topsoils with high content of organic matter. Soils with a high content of organic matter are likely to have a larger seed bank than soils with lower organic matter content (cf. . Granstrom, 1982). Somewhat surprisingly there was no significant relation between the vegetation cover and the content of organic matter in year 2001. However, the subsoil with high clay content may have provided more essential nutrients than the organic soil and thus support good growth. This is a possible explanation why some
ARTICLE IN PRESS A. Brekke Skrindo, P. Anker Pedersen / Urban Forestry & Urban Greening 3 (2004) 29–37
macro plots with low content of organic matter, but high clay content have rather high vegetation cover (e.g. macroplots 6 and 9). Other possible explanations could lie in processes during stockpiling, uneven distribution of organic soil on the sites, the shorter growth period at macro plots 1–4 and other ecological factors not recorded in this study . The quality of the organic soil is probably of major importance as illustrated in macroplot 4 with high content of moderately decomposed Sphagnum sp. Slow mineralisation leads to limited amount of available nitrogen, resulting in low vegetation cover. Nitrogen is usually considered to be the most important limiting resource in boreal forests (cf. Hesselmann, 1937; Tamm, 1991). The thickness of the excavated topsoil influences the availability of propagules. The number of viable seeds from different depths has been studied in different . (1982) found most ecosystems. For instance, Granstrom viable seeds in the thin humus layer in five different boreal forest stands. The humus—mineral soil ratio varies with the thickness of the humus layer. Consequently, number of seeds is likely to be higher in macroplots with topsoil rich in organic matter. The quantity of redistributed topsoil also affects the revegetation. The topsoil layer in this study after redistribution was supposed to be 10 cm thick, but there are patches of thinner and thicker topsoil layer due to the inaccuracy of the distribution method. Several scientists working in the field of reclamation have accepted the hypothesis that the deeper the humus layer (up to the natural depth), the more productive the plant community (Power et al., 1976; McGinnies and Nicholas, 1980; Redente and Hargis, 1985). However, Redente et al. (1997) found no difference in vegetation cover after 10 years although the thickness of the humus layer varied from 15 to 60 cm. When comparing seed bank studies and revegetation from stockpiled topsoil (including the propagule bank, the micro fauna, flora and the nutrition), the effects of stockpiling on the soil quality have to be taken into account. Hargis and Redente (1984) found decreased germination and revegetation success and Rives et al. (1980) found less mycorrhiza-infected roots in stockpiled topsoil. The soil used in this study had been stockpiled for 1–1.5 years but good revegetation results are obtained from different soiltypes up to 5 years of storage according to Stark and Redente (1987). The revegetation process in macroplots 1–4 was 2 months shorter than in the other plots due to the delayed redistribution of topsoil. The variation between the macro plots, however, cannot be explained by this delay alone because the difference between the macroplots with the same growth period was significant as well. In spite of good climatic conditions for germination and growth throughout the summer some macro
35
plots still had a vegetation cover less than 10% in year 2001. The propagule bank size does influence the vegetation cover demonstrated by the correlation between the number of species and the vegetation cover. In this study, some of the most frequent species had the largest cover percentage (i.e. Ranunculus repens, Tussilago farfara and Rubus idaeus). However, other frequent species did not contribute much to the vegetation cover (Picea abies, P. sylvestris and Galeopsis tetrahit). The change in species composition over time is demonstrated by the 14 species recorded only in the first year and 27 new species appearing the second year. This considerable alteration is typical in the first phase of a secondary succession due to variation in the microclimate, variation in the propagule bank and due to seed dispersal from the surroundings (Connell and Slatyer, 1977). The 121 species recorded in this study are dominated by pioneer species but species expected to establish throughout the whole succession period occurred as well. Connell and Slatyer (1977) found that individuals of any species that happened to fulfil their germination requirements could establish in this early stage. Among the species with altered frequency from 2000 to 2001 some grasses (Agrostis capillaris, Deschampsia caespitosa and Glyceria fluitans) increased and others (Alopecurus geniculatus, Poa annua and P. nemoralis) decreased and other graminoids (Carex pallescens, Luzula pilosa and Juncus effusus) increased. All of these species are often found in moist places, but the species that decreased are known to loose in competition for light (cf. Lid and Lid, 1994). As woody plants may represent a maintenance problem and influence traffic safety, their slow invasion is good news. Only P. sylvestris increased significantly, whereas Picea abies decreased. Rubus idaeus was among the most frequent species, which was expected because of the viability of the seeds in comparable seed banks . (cf. Granstrom, 1982). Salix caprea was also among the most frequent species even without a persistent seed bank, because it disperses and germinates easily (cf. Brinkman, 1974). The weeds are neither an aesthetic nor a management problem yet. Among the 16 species that increased significantly in frequency, only two can be considered as weeds (Cirsium arvense and Tussilago farfara). If they continue to increase, they might dominate some areas completely and prevent other indigenous species to grow. Six of the nine species that decreased were also typical weeds (Chenopodium polyspermum, Matricaria perforata, Persicaria maculosa, Rorippa palustris and Stellaria media). All of these species are annuals or at least short-lived (Korsmo, 1986; Lid and Lid, 1994) and might be expected to decrease. There were several weeds that did not change in frequency, and their future presence remains unknown.
ARTICLE IN PRESS 36
A. Brekke Skrindo, P. Anker Pedersen / Urban Forestry & Urban Greening 3 (2004) 29–37
The macro plots are situated in a landscape with different vegetation history and ecology influencing the propagule bank as argued by Baskin and Baskin (1998): The size and the content of the propagule bank varies with plant age, species composition, disturbance level, predispersal seed predation and plant seed production. There is no clear relationship between the soil properties and the species composition except in macroplot 4 with organic soil that had few weeds, woody plants and graminoids, and in plots 2, 3 and 7 with loamy soil that host none or only a few weeds. Several studies show a lack of correspondence between the species composition of the seed bank and the associated plant community in surrounding forests (e.g. Oosting and Humphreys, 1940; Thompson and . Grime, 1979; Granstrom, 1982). The species composition in the early phase of succession seems hard to predict. That the species composition in the early phase of succession is hard to predict is illustrated by the variation between the macroplots in this study. In general, however, both the vegetation cover and species composition were satisfactory already 2 years after the soil was redistributed. Natural revegetation from redistributed topsoil is therefore recommended in comparable ecosystems.
Acknowledgements This project was partly financed by the Public Roads Administration, Buskerud county and the Directorate of Public Roads in Norway.
References Baskin, C.B., Baskin, J.M., 1998. Seeds. Ecology, Biogeography, and Evolution of Dormancy and Germination. Academic Press. San Diego, USA. Begon, M., Harper, J.L., Townsend, C.R., 1990. Ecology. Blackwell Scientific Publications, Oxford. Bradshaw, A.D., 1989. The quality of topsoil. Soil Use Management 5, 101–108. Brinkman, J.A., 1974. Salix. In: Young, J.A., Young, C.G. (Eds.), Seeds of Woody Plants in the United States Agricultural Handbook No. 450. Forest Service, US Department of Agriculture, Washington, DC, pp. 746–749. Connell, J.H., Slatyer, R.O., 1977. Mechanisms of succession in natural communities and their role in community stability and organisation. American Naturalist 111, 1119–1144. Egn!er, H., Riehn, H., Domingo, W.R., 1960. Untersuchnungen uber . die chemische Bodenanalyse als Grundlage fur . der . Beurteilung des N.ahrstoffzustandes der Boden. II. Chemische Extractions Methoden zur Phosphor und Kalium. ok, . bestimmung. Statens Jordbruksfors S.artryck och sm(askrifter, Uppsala (in German).
Førland, E.J., 1993. Nedbørnormaler, normalperioden 1961– 1990. (Mean precipitation, 1961–1990) Norske meteorologiske institutt. Rapport Klima 9, 1–63 (in Norwegian). . Granstrom, A., 1982. Seed banks in five boreal forest stands originating between 1810 and 1963. Canadian Journal of Botany 60, 1821–1855. Hargis, N.E., Redente, E.F., 1984. Soil handling for surface mine reclamation. Journal of Soil and Water Conservation 39, 300–305. Hesselmann, H., 1937. Om humust.ackets beroende av besta( ndets a( lder och sammans.attning i den nordiska granskogen av bla( b.arsrik Vaccinium-typ och dess inverkan pa( . skogens foryngring och tillv.axt. (Effects of tree stand age and species composition on the humus layer, forest regeneration, and forest yield in a bilberry-dominated Norway spruce forest) Medd. Statens Skogsforskningsanstalt, Vol. 30, pp. 529–716 (in Swedish). Holmes, P.M., 2001. Shrubland restoration following woody alien invasion and mining: effects of topsoil depth, seed source, and fertilizer addition. Restoration Ecology 9, 71–84. Kendall, M.G., 1938. A new measure of rank correlation. Biometrika 30, 81–93. Korsmo, E., Vidme, E., Fykse, H., 1986. Korsmos ugrasplansjer (Korsmos weed illustrations) Landbruksforlaget, Oslo (in Norwegian) Kotanen, P.M., 1996. Revegetation following soil disturbance in a California meadow: the role of propagule supply. Oecologia 108, 652–662. Lid, J., Lid, D.T., 1994. Norsk flora (Norwegian flora) Det norske samlaget, Oslo (in Norwegian). McGinnies, W.J., Nicholas, P.J., 1980. Effect of topsoil thickness and nitrogen fertilization in the revegetation of coal mine spoils. Journal of Environmental Quality 9, 681–685. Økland, T., 1988. An ecological approach to the investigation of a beech forest in Vestfold, SE. Norway. Nordic Journal of Botany 8, 375–407. Økland, R.H., 1990. Vegetation ecology: theory, methods and applications with reference to fennoscandia. Sommerfeltia 2 (Suppl.), 1–233. Økland, R.H., Økland, T., Rydgren, K., 2001. A Scandinavian perspective on ecological gradients in north-west European mires. Journal of Ecology 89 (3), 481–486. Oosting, H.J., Humphreys, M.E., 1940. Buried viable seeds in a successional series of old field and forest soils. Bulletin of the Torrey Botanical Club 67, 253–273. Power, J.F., Ries, R.E., Sandoval, F.M., 1976. Use of soil materials on spoils—effects of thickness and quality. North Dakota Farm Research 34, 23–24. Redente, E.F., Hargis, N.E., 1985. An evaluation of soil thickness and manipulation of soil and spoil for reclaiming mined land in north-west Colorado. Reclamation and Revegetation Research 4, 17–29. Redente, E.F., McLendon, T., Agnew, W., 1997. Influence of topsoil depth on plant community dynamics of a seeded site in Northwest Colorado. Arid Soil Research and Rehabilitation 11, 139–149. Rives, C.S., Bajwa, M.I., Liberta, A.E., Miller, R.M., 1980. Effect of topsoil storage during surface mining on the viability of VA mycorrhiza. Soil Science 129, 253–257.
ARTICLE IN PRESS A. Brekke Skrindo, P. Anker Pedersen / Urban Forestry & Urban Greening 3 (2004) 29–37
Rokich, D.P., Dixon, K.W., Sivasithamparam, K., Meney, K.A., 2000. Topsoil handling and storage effects on woodland restoration in Western Australia. Restoration Ecology 8, 196–208. SAS Institute Inc., 1987. SAS/STAT user’s Guide. Version 6, Fourth Edition, Volume 2, SAS Institute Inc., Cary, NC, USA. Shaw, P.J.A., 1996. Role of seedbank substrates in the revegetation of fly ash and gypsum in the United Kingdom. Restoration 4, 61–70. Sokal, R.R., Rohlf, F.J., 1995. Biometry: The Principles and Practice of Statistics in Biological Research. Freeman, New York. Stark, N.M., Redente, E.F., 1987. Production potential of stockpiled topsoil. Soil Science 144, 72–76. Statens V., 1992. Veg-og gateutforming. (Road and street construction) Vegvesenets h(andbøker nr. 017. Oslo, Norway (in Norwegian).
37
Sveistrup, T., Njøs, A., 1984. Kornstørrelsesgrupper i mineraljord (Textural classes in mineral soils.). Jord og Myr 8, 8–15 (in Norwegian). Tamm, C.O., 1991. Nitrogen in terrestrial ecosystems. Questions of productivity, vegetational changes and ecosystem stability. Ecological Studies 81, 1–116. Thompson, K., Grime, J.P., 1979. Seasonal variation in the seed banks of herbaceous species in ten contrasting habitats. Journal of Ecology 67, 893–921. Warr, S.J., Thompson, K., Kent, M., 1993. Seed banks as a neglected area of biogeographic research—a review of literature and sampling techniques. Progress in Physical Geography 17, 29–347. Zhang, Z.Q., Shu, W.S., Lan, C.Y., Wong, M.H., 2001. Soil seed bank as an input of seed source in revegetation of lead/zinc mine tailings. Restoration Ecology 9, 378–385.