Restoration of plant species diversity of ditch banks : ecological
constraints and opportunities
Blomqvist, M.M.
Citation
Blomqvist, M. M. (2005, February 3). Restoration of plant species diversity of ditch banks :
ecological constraints and opportunities. Retrieved from https://hdl.handle.net/1887/592
Version:
Not Applicable (or Unknown)
License:
Licence agreement concerning inclusion of doctoral thesis in the
Institutional Repository of the University of Leiden
Downloaded from:
https://hdl.handle.net/1887/592
Seed and (micro)site limitation in ditch banks:
germination, establishment and survival under different
management regimes
M. M. Blomqvist, W. L. M. Tamis, J. P. Bakker &
E. van der Meijden
Abstract
Chapter 5 94
Introduction
Conservation of the biodiversity of the agricultural landscape is increasingly being recognized as an important topic, both among ecologists and policy makers (e.g. Ovenden et al. 1998; Stoate et al. 2001; Benton et al. 2003). Ditch banks, along with other small-scale linear landscape elements such as hedgerows, ditches and field margins, serve as important refuges for many plant species in otherwise intensively managed agricultural landscapes (e.g. Baudry 1988; Fry 1994; Geertsema 2002; Twisk et al. 2003). Ditch banks are particularly common in the Western Peat District in the Netherlands, but similar habitats are also found elsewhere, e.g. in the UK (Kruk 1991). In this habitat we currently find many previously common wetland, grassland and hayfield species, such as Caltha palustris, Cirsium dissectum, Lychnis flos-cuculi and
Lysimachia thyrsiflora (Clausman & van Wijngaarden 1984; Melman 1991). In response
to concerns regarding declining species richness in ditch banks, ‘nature-friendly’ management was developed in the 1980s. However, subsequent implementation within agri-environment schemes has generally not been successful in increasing the plant diversity (Kleijn et al. 2001; Blomqvist & Tamis, unpublished manuscript, but see Dijkstra 1994).
Management recommendations in ditch banks were developed at a time when high fertilizer application rates were causing a decline in ditch bank species richness. As a result, they were mainly focused on reducing productivity by restricting the nutrient input (no fertilisation, no ditch sediment deposition in the ditch bank). In addition, extensive mowing and grazing regimes were often advocated to allow for reproduction (van Strien 1991; Melman & van Strien 1993; Kruk et al. 1994; LNV 1995, but see below). From an ecological viewpoint they thus focused almost exclusively on enabling long-term survival and preventing extinction. As in so many other restoration projects in grasslands, an increase in species richness was expected, yet failed to materialize (e.g. Olff & Bakker 1991; Pfadenhauer & Klötzli 1996; Berendse et al. 1999). Subsequently, many grassland studies have emphasized the importance of seed banks and dispersal (Marshall & Hopkins 1990; Bakker et al. 1996; van Dorp 1996; Prins et al. 1998; Pakeman et al. 1998; Bakker & Berendse 1999). The focus has thus shifted from the process of extinction to (re-)colonisation.
constraints currently hampering an increase in species diversity in ditch banks, and thus to establish the relative importance of seed, microsite and site limitation.
In recent studies, based on correlative relatinonships between species traits and demographic parameters, we found that (lack of) colonisation was more important for determining species increase or decline than extinction (Blomqvist et al. 2003b). High nutrient levels seemed to affect germination and establishment, indicating microsite limitation rather than site limitation. Moreover, the lack of target species in ditch bank seed banks and some indications of problems with and colonisation constraints also point in the direction of seed limitation (Blomqvist et al. 2003a; Blomqvist et al. 2003b; Blomqvist & Tamis, unpublished manuscript). However, experimental verification in the field and better guidelines for practical management are needed.
A considerable range of biomass and competition pressure exists in modern ditch banks today. In a recent grassland study (Foster 2001; Foster et al. 2004), the relative importance of seed and (micro)site limitation differed in situations with high and low productivity, suggesting the need for different management. Moreover, ditch bank and grassland studies have indicated that more intensive management (with earlier first cuts) is often better in high productive situations (Bakker 1989; Melman & van Strien 1990; Melman 1991; Bakker & Olff 1992; Ryser et al. 1995). Yet, a clear distinction has usually not been made between high and low productive situations within agri-environment schemes. The earliest ‘management agreement’ schemes (LNV 1995) demanded late cuts, while the newest schemes (DLG 2000) impose few restrictions on the mowing and grazing management, but lack recommendations with regard to the timing of cuts. Only in ‘direct payment’ schemes (Melman 1991), earlier cuts were advocated for situations with high biomass.
Chapter 5 96
Material and Methods
Site selection and experimental treatments
The research area (51°51’N - 52°07’N and 4°45’E - 5°08’E) is situated mainly in the Province of South-Holland in the Western Peat District in the Netherlands. The dominant soil types are peat and peat with clay. Ditch water levels in the permanently water-filled ditches are artificially controlled: winter levels are normally some 10-15 cm below summer levels. For a more detailed description of the study area and the vegetation see Blomqvist et al. 2003b). Our study was carried out on six modern dairy farms with ‘direct payment’ agri-environment schemes adopted since 1992 (Kruk et al. 1994). These farms differ in the intensity of the applied management and biomass (Blomqvist & Tamis, unpublished manuscript). Per farm, we selected one ditch bank, with the overall aim of including a large biomass range in the study.
Each ditch bank was divided into four mowing and grazing treatments (with five replicates per treatment) differing in date of first cut and subsequent cutting frequency (Table 1, Fig. 1). Treatment I (‘July mowing’) approximates the most extensive management recommendations with the first cut in July and the second one in September-October. Treatment III (‘May mowing’) is equally extensive in terms of number of cuts, but the first cut was much earlier, in May, removing biomass early in the season, but allowing for further undisturbed reproduction. Treatment II (‘June+ mowing’) was more intensive, with a total of three cuts (in June, August and September-October) and applied with the aim of removing as much biomass as possible. Treatments I-III were excluded from grazing. At the time of mowing the vegetation was mown to a height of 5-10 cm with a brush-cutter, approximating standard cutting with a disc mower. Treatment IV (‘control or normal management’) was open for standard cutting and grazing and thus acted as a ‘treatment control’. The management in TIV was recorded on each farm. No ditch sediment deposition was allowed in any of the four treatments.
Sowing experiment
We selected nine typical ditch bank species (Caltha palustris, Cirsium palustre, Lotus
pedunculatus, Lychnis flos-cuculi, Lysimachia vulgaris, Lythrum salicaria, Mentha aquatica, Prunella vulgaris, Valeriana officinalis) differing in ecological requirements
Table 1. Mowing (M) and grazing (G) treatments for sowing and transplanting
experiments. Treatment IV acts as a control for ‘normal’ management and therefore varied between farms. An overall indication is given for the two productivity categories (low vs. high, three ditch banks in each category, the number of farms indicated within brackets) in 2002 and 2003.
Treat-ment April End May Mid Begin June Mid / end June
Begin
July end July Mid / August End Sept / Oct I M M II M M M III M M IV - low 2002 G (2) G/M (2) G (1) G (3) IV - high 2002 G (2) G (1) M (3) G/M (3) G (3) IV-low 2003 G (3) G/M (2) - IV- high 2003 G (1) G/M (3) G/M (3) -
Fig. 1. Experimental set up of sowing and transplanting experiments (not to scale).
Table 2. Ecological characteristics and presence of the nine selected target species in the studied ditch banks. Germination = germination time(s); flowering = month of flowering onset; N-value = Ellenberg productivity value; seed bank index = seed longevity: 1 = transient, 2 = short-term persistent, 3 = long-term persistent (> 5 years); Presence vegetation = farm number where species was seen in the vegetation in the study ditch bank; Presence sb = farm number where study species had been found in the seed bank in a previous study; Natural seedlings = refers to our own observations of seedlings in different ditch banks between October 2001 - October 2003; (r) = rare sighting
Species Germination Germination in the field
Flowering
1) N-value 2) Seed bank index 3) Presence vegetation Presence seed bank 4) Natural seedlings
Caltha palustris Late spring Mid Mar 4 x 1 6(r) No No
Cirsium palustre Spring Mid Apr 6 3 2 4 No 4(r)
Lotus pedunculatus Spring End Nov 6 4 3 4,5,6(r) 4 4,5
Lychnis flos-cuculi Directly or spring End Apr 5 x 3 1(r),2(r),3(r),
4,5,6(r) 1,2,4,5,6 4,5
Lysimachia vulgaris Late spring Begin Jun 6 x 2 No No No
Lythrum salicaria Spring Begin Apr 6 x 3 No No No
Mentha aquatica Autumn or spring Begin May 7 5 3 No No No
Prunella vulgaris Autumn or spring Mid Feb 5 x 3 No No No
Valeriana officinalis Spring or summer End Feb 6 5 1 No No No
1) CBS (1997)
2) Ellenberg et al. (1992); Wiertz (1992)
3) Thompson et al. (1997), modified according to Tamis et al. (2000) 4) Blomqvist et al. (2003a)
98 Chap
ter
A pilot sowing experiment (M. Blomqvist, unpublished results) had indicated that germination and establishment is virtually non-existent in closed ditch bank vegetation. Therefore, seeds were not sown in undisturbed vegetation in this experiment. Instead, the vegetation was mown and artificial gaps (15 x 15 cm, about the size created by trampling cows) were created in the ditch bank about 5-10 cm above normal summer water levels and at about 20 cm distance (measured along the ditch bank) from the water line (Fig. 1). In each replicate (five per treatment), we created nine 2-5 cm deep gaps (one for each species), by removing the plants and disturbing the soil layer, resulting in a total of 20 gaps per species per ditch bank (four treatments x five replicates). In October 2001, batches of about 100 seeds per species (Appendix 1) were sown in each gap, totalling around 12 000 seeds per species (six ditch banks x four treatments x five replicates x 100 seeds). The order of the species was randomly determined for the five replicates, after which the same order was repeated in each treatment. The number of seedlings and the number of established plants (defined as ≥ 4 cm in height or diam.) were monitored nearly every month until the end of September 2003, totalling 23 monitoring events. Information about the seed bank was available from an earlier study (Blomqvist et al. 2003a) and ‘wild’ seedlings were recorded during the monitoring process (Table 2). Since this information was available, unsown controls were deemed unnecessary (see discussion).
Transplanting experiment
Seeds from the nine selected species were sown in May-June and allowed to grow outside during the summer 2001. At the end of August 2001, when most (trans)plants had a diameter or height of about 2-4 cm, they were planted into the field in small gaps (diam. 5 cm), four plants per replicate (= 80 plants per species per ditch bank). However, since the germination rate of the Cirsium seeds was so poor, only two plants per replicate were planted into four ditch banks (farms 1,3,4,5). Plants were transplanted in three rows. Plants requiring very moist conditions (Caltha and Mentha) were planted close to the summer water line. The seven other species were transplanted in four blocks in two rows, 20 and 40 cm from the summer water line (measured along the ditch bank, Fig. 1). The order of the species was determined at random in each block and separately for each of the five replicates, after which the same order was repeated in each treatment. Plant survival and reproduction (flowering / seed-setting) were monitored in October 2001 and monthly in April-September/October 2002 and 2003, totalling 12 monitoring events.
Above-ground biomass
Chapter 5 100
the amount) at 70ºC. Biomass is expressed as g dry weight / m2. On the six farms in this study, the mean biomass ranged from 310 to 855 g/m2 in July 2001 and from 320 to 934 g/m2 in July 2002 (Fig. 2). For the statistical analyses the ditch banks were ranked and divided into two categories with high (farms 1-3) and low (farms 4-6) biomass.
Fig. 2. Mean biomass (g dw / m2) per farm (= per ditch bank) in July 2001 before the experiment and in July 2002, the year the treatments started. Vertical bars represent standard deviation (SD). SD (per ditch bank) in 2001 is based on 15 replicates (in Treatments TI-TIII); SD in 2002 is based on 5 replicates (in TI; ‘July mowing’). The difference in biomass between the two biomass categories was tested with ANOVA (after log-transformation): 2001: F1,89 = 51.721, p < 0.0001 (original means: low = 366, high = 640 g m-2); 2002: F1,29 = 37.939, p < 0.0001 (original means: low = 432, high = 837 g m-2; corresponding to a mean annual biomass of 662 and 1210 g m-2, respectively). Different letters refer to significant differences between ditch banks with high vs. low biomass.
Data analyses
in each gap during each year. As such, it should be clear that we are not dealing with the actual maximum (cumulative) germination and establishment rates, but rather the best possible approximation (which in fact is the smallest possible maximum value in each gap). The % germination in 2002 (hereafter referred to as germination 2002) was expressed as the maximum number of germinated seedlings / mean number of sown seeds per species (Appendix 1). No correction for germinability was needed, since the data was analysed per species and no bias is to be expected within or between ditch banks. Germination in 2003 was negligible (Figs. 3a,b) and not included in the analyses. The % establishment in 2002 and 2003 (establishment 2002 and 2003) were expressed as the maximum number of established seedlings / the maximum number of germinated seedlings in 2002. Due to changes in water level regimes and (occasional) flooding some gaps were frequently filled with water. Since we were not interested in the effects of water levels in this study and to reduce the variation in the dataset, gaps that were inundated more than 30 % of the time (≥ 7 inundated recording events out of a total of 23) were excluded from our analyses. This resulted in an unbalanced design.
Flooding in September 2001 and February 2002 killed many plants, before the actual start of the experiment. Therefore, to correct for this loss only ‘plants present at the beginning of the experiment’, i.e. that were recorded alive prior to the start of the treatments in May 2002 or at least once after this date, were included in the subsequent analyses of the transplant data. For the analyses, transplanted plants were divided into two groups: all surviving and those reproducing. Again, since the number and percentage of surviving and reproducing plants varied in time among species, treatment and years, we selected the maximum recorded value per replicate (highest possible value per species = 4, Fig. 1). The % survival in 2003 (survival 2003) was expressed as a percentage (maximum number of surviving plants / number of plants present at the beginning of the experiment). Likewise, the % reproducing plants 2002 and 2003 (reproduction 2002 and 2003) were calculated as number of reproducing plants / number of plants present at the beginning of the experiment.
Chapter 5 102
ditch bank nested within treatment were included as random factors to account for the split-plot design (Fig. 1). Differences between management regimes were tested with Wald test. Secondly, to investigate whether the selected nine ‘target’ species showed a general preference for the same management, we performed nonparametric (rank) tests on all species together. Differences in means of the six dependent variables in ditch banks with low and high biomass (again including TI, II, III, but excluding TIV) were tested with Wilcoxon signed rank test (Siegel & Castellan, Jr. 1988). Differences between treatments TI-III (for both biomass categories together and per biomass category (low, high)) were tested with Friedman two-way analysis of variance. To quantify the effects of TIV within ditch banks with high and low biomass, the last analyses were also performed for TI-IV.
Results
Variation in ’normal management’
The ‘normal management’ (TIV) differed between ditch banks with high and low biomass. Although management also differed between farms within the two biomass categories, as a rule, management was more intensive (earlier grazing and more cuts) in ditch banks with high biomass than with low biomass (Table 1). Differences were largest in 2002 when TIV was the most intensive management in ditch banks with high biomass (more intensive than ‘May mowing’ TIII) and least intensive in ditch banks with low biomass (similar to ‘July mowing’ (TI)). In 2003, TIV was between ‘June+’ (TII) and ‘July mowing’ (TI) in terms of intensity.
Seedling germination and establishment
Species started to germinate at different times (Table 2), with an overall peak in May 2002 (Figs. 3a,b), both with low and with high biomass. A small new cohort of seedlings emerged at the low biomass sites in the second year (Figs. 3a,b), while very few seedlings emerged at the high biomass sites.
Fig. 3. Mean (%) germination (per no. of sown seeds) and establishment (per maximum
as the rank analyses also indicated, germination and establishment were still higher in ditch banks with low biomass for most species. Germination and establishment are thus hampered by high biomass values.
When high and low biomass sites were tested together for the effects of ‘May, June+ and July mowing’ (TI-III) on germination and establishment, no particular treatment was significantly better overall than another one (Table 3A). This indicates that species differ in the treatment they ‘prefer’ (see Appendix 2A). However, within ditch banks with high biomass, establishment 2002 was significantly higher with ‘May and June+ mowing’ (TIII, II) than with ‘July mowing’ (TI), while no differences were found in ditch banks with low biomass (Table 3B, Figs. 4a-c). Moreover, when TIV was added to these analyses, both germination and establishment was highest with this, ‘normal management’ at high biomass sites (Fig. 3), suggesting that more intensive management will enhance germination and initial establishment under such conditions. At species-level, very few species (Caltha, Lythrum, Prunella) showed (near-)significant differences in germination and establishment in response to treatments TI-III (GLMM: Appendix 2). No significant interactions were found.
Plant survival
At species-level, most species showed higher survival in low than in high biomass ditch banks. For five species (Lotus, Lysimachia, Lythrum, Mentha, Prunella) this difference was (near-)significant (Appendix 2). When all species were lumped, the maximum survival in 2003 (as % of plants present at the beginning) was 60-80% in low and 50-70% in high biomass ditch banks (Fig. 4d). When all treatments were lumped, survival was significantly higher in ditch banks with low biomass values (Table 3A, Fig. 5). This was also the case within most treatments (except with July mowing (TI)), indicating that high biomass values hamper survival.
Different species showed considerable variation in the optimal treatment for survival when all ditch banks were lumped, hence no significant differences between ‘May, June+ and July mowing’ (TI-III; Table 3A, Fig. 4d). Neither were there any significant differences between treatments TI-III within ditch banks with low or high biomass (Table 3B). At species-level, three species (Lychnis, Mentha, Prunella) showed (near-)significant treatment effects, but only showed two significant interactions
Fig 4. (opposite page). Mean % (‘maximum’) (a) germination 2002 (per no. of sown
seeds), (b, c) establishment 2002-2003 (per maximum germination 2002), (d) survival 2003 and (f, g) reproduction 2002-2003 (per plants present at the beginning), for nine ‘target’ species in different management situations (low vs. high biomass and four mowing/grazing treatments: TI = ‘July mowing’, TII = ‘June+ mowing’, TIII = ‘May mowing’, TIV = ‘normal management’, Table 1). Significant differences between TI-IV
within productivity categories are indicated with an asterisk. Other rank tests in Table 3,
Table 3. Differences for nine species in mean (‘maximum’ %) germination 2002 (per no. of seeds), establishment 2002-2003 (per max. germination 2002), survival 2003 and reproduction 2002-2003 (per no. of plants present at the beginning). A. Results of analyses for three ‘controlled’ experimental treatments (TI = ‘July mowing’, TII = June+ mowing, TIII = May mowing): ranking of species in ditch banks with low (L) and high (H) biomass (the treatments together and per treatment) and overall ranking of the treatments (low + high biomass). B. Ranking of treatments within ditch banks with low and high biomass for TI-III and TI-IV. Wilcoxon signed rank test was used to test differences between biomass categories; Friedman two-way analysis of variance by ranks was used to test for differences between treatments. P-values significant at 0.05-level are indicated in a bold letter type, n = 9. See also Fig 4.
A. Ranking of two biomass categories & treatments TI-III
Analysed factor: Biomass (low vs. high) Treatment (TI vs. TII vs. TIII)
Situation: Rank TI-III TI TII TIII Low & high biomass 3
2 p p p p rank p Germination 2002 1 L > H 0.008 0.008 0.214 0.011 3 > 2 > 1 0.124 Establishment 2002 1 L > H 0.008 0.015 0.021 0.021 3 > 2 > 1 0.972 Establishment 2003 1 L > H 0.008 0.011 0.110 0.012 2 > 1 > 3 0.895 Survival 2003 L > H 0.021 0.173 0.038 0.008 3 > 2 > 1 0.121 Reproduction 2002 L > H 0.674 0.237 0.310 0.917 3 > 1 > 2 0.905 Reproduction 2003 L > H 0.110 0.767 0.327 0.017 1 > 2 > 3 0.368 B. Ranking of treatments (TI-III & TI-IV) within the two biomass categories
Analysed factor: Treatment (TI vs. TII vs. TIII) Treatment (TI vs. TII vs. TIII vs. TIV)
Situation: Low biomass 3 High biomass 3 Low biomass 3 High biomass 3
rank p rank p rank p rank p
Germination 2002 1 3 > 1 > 2 0.124 3 > 2 > 1 0.175 3 > 1 > 4 > 2 0.233 4 > 3 > 2 > 1 0.027 Establishment 2002 1 2 > 1 > 3 0.972 3 > 2 > 1 0.016 2 = 3 > 1 > 4 0.903 4 > 3 > 2 > 1 0.016 Establishment 2003 1 2 > 1 > 3 0.895 2 > 3 > 1 0.276 2 > 1 > 3 > 4 0.508 2 > 4 > 3 > 1 0.293 Survival 2003 3 > 2 > 1 0.121 3 > 1 > 2 0.972 3 > 2 > 1 > 4 0.012 4 > 1 > 3 = 2 0.756 Reproduction 2002 1 = 3 > 2 0.905 3 > 1 > 2 0.468 1 > 3 > 2 > 4 0.893 4 = 3 > 1 > 2 0.615 Reproduction 2003 3 > 2 = 1 0.368 1 > 3 = 2 0.703 3 > 4 > 2 > 1 0.012 4 > 1 > 2 > 3 0.016
1 excluding Lychnis and Lotus (A. Biomass) yields very similar results, except that the difference between ditch banks with high and low biomass in TII
were also significant for establishment 2003.
2 means were always higher in ditch banks with low than with high biomass (for all treatments together and within treatments) 3 Note that the order in this table is based on (weighted) ranks. Some treatments deviate in order from that seen in Fig 4.
106 Chap
ter
(Lychnis, Valeriana; Appendix 2), suggesting that survival cannot be substantially enhanced by applying different management in situations differing in biomass. However, when ‘normal management’ (TIV) was included in the analyses, survival was lowest with TIV in ditch banks with low biomass. No significant differences were found within high biomass ditch banks, yet, overall, most plants survived with (intensive) ‘normal management’ (Figs. 4d, 5).
Fig. 5. Mean % surviving plants (sum all species) through time of those present at the
Chapter 5 108
Plant reproduction
In terms of reproduction, species differed considerably in the optimal site biomass and treatment (Appendix 2). When all species were lumped, the % reproducing plants was similar at low and high biomass sites (Figs. 4e,f). In 2002, the % reproduction was 15-20% in low and 10-15-20% in high biomass ditch banks. In 2003 these percentages were 25-50% in both situations. When all treatments were tested together and within treatments, reproduction was slightly higher in ditch banks with low biomass, but the differences were not significant (Table 3A). All in all, it seems that, although high biomass tends to have a negative effect on reproduction, it is far less pronounced than for germination, establishment and survival.
At species-level, five species (Cirsium, Lychnis, Lythrum, Mentha, Prunella) showed significant treatment effects (Appendix 2), indicating that management had considerable effects on flowering and seed-set of individual species. When all species were lumped, reproduction did not differ significantly between ‘May, June+ and July mowing’ (either including both or within the two biomass categories; Table 3A,B), suggesting that the preferred treatment differed between species. However, within high biomass ditch banks, many species showed highest reproduction in plots with ‘normal management’ (TIV) and ‘May mowing’ (TIII) in 2002 and ‘normal management’ in 2003, the difference being significant in 2003 (Table 3B, Figs. 4e,f). In low biomass ditch banks, reproduction was highest with ‘May mowing’ (TIII) and ‘normal management’ (TIV) in 2003.
Discussion
Seed and microsite limitation in ditch banks differing in productivity
We found that an increase in ditch bank ‘target’ species is hampered by a combination of seed and microsite limitation and that the relative importance of microsite limitation is greater in ditch banks with high biomass. The introduction of seeds always resulted in increased germination (seedling emergence) and initial establishment (seedling survival) indicating that an increase in species richness is being impeded by the lack of seeds (Fig. 3). In most ditch banks the investigated target species were absent both from the vegetation and from the seed bank (Table 2; Blomqvist et al. 2003a), which clearly shows that dispersal is a problem for many species in ditch banks, corresponding with the findings of van Dorp (1996) and Geertsema (2002).
banks with high biomass (Table 3A). The relative importance of microsite limitation is thus greater in ditch banks with high productivity. This is in contrast with the findings of Kitajima & Tilman (1996) that suppression of seed germination, rather than seedling survival (establishment), is the main mechanism reducing colonisation in high productive situations. The discrepancy is probably explained by the small biomass range (200-500 g/m2) in the study by Kitajima & Tilman (1996), while the overall mean range in our study was about 350-850 g/m2. Our results are, however, in accordance with the results of Foster (2001) and Foster et al. (2004) and provides further evidence for the ‘shifting limitations hypothesis’, which postulates that the major factor regulating species richness shifts from regional to local processes with an increase in productivity.
Since the seeds were sown in gaps rather than in undisturbed vegetation, one could argue that microsite limitation is always much more important than seed limitation in the ditch bank system. Certainly this is the case, if we consider the fact that ‘natural’ germination and germination of sown seeds is minimal in undisturbed ditch bank vegetation (M. Blomqvist, unpubl. results). However, gaps are a natural part of this ecosystem, whether created by trampling cows and sheep, or by water rats, field mice and moles. The importance of gaps was, for example, also demonstrated in an introduction experiment on a salt marsh, where the seedling emergence was negatively related to the height of the canopy and positively to the amount of bare soil created by trampling cattle (Bakker & de Vries 1992, see also Watt & Gibson 1988). Thus, our experiments reflect natural conditions encountered by seeds and as such do give an indication of the relative importance of seed and microsite limitation.
As for a possible effect of a seed bank, we are confident that most species were not present in the seed bank at all, or (as in the case of Cirsium), present only so infrequently that they cannot have influenced our results (Table 2). Although care should be taken in interpreting the germination data on Lotus and Lychnis at species level, a possible bias would only be relevant for the effects of biomass on germination (no effect is to expected between treatments within ditch banks). However, exclusion of these two species in the all-species analyses did not affect our conclusions (Table 3A).
Microsite vs. site limitation and long-term survival
Chapter 5 110
limitation is relatively more important than site limitation in the ditch bank system, i.e. that overall, in the life-cycle of a plant, seedling germination and establishment are more limiting for an increase in species richness than survival. This confirms earlier findings of Blomqvist et al. (2003b) that colonisation is more important than extinction in ditch banks.
In addition, our results show that the percentage of reproducing target species (of the transplants) was high; by 2003 up to 50% of the plants present at the beginning of the experiment were reproducing (Figs. 4e,f). Therefore, it should be possible for target species to establish a viable population. This is further corroborated by observations in the field. In 2003, we occasionally observed seedlings of some transplanted species (Cirsium, Lychnis, Prunella) in natural gaps in the transplant section of the ditch banks, also where these species were absent from or rare in the natural vegetation (Table 2). Moreover, all species, except Caltha and Lysimachia, also produced some flowering plants in 2003 from the seeds sown in the autumn of 2001. All in all, long-term establishment is therefore possible for most species, as long as natural gaps are present and proper management is applied.
Methodology: statistical limitations
Effects of biomass and management regime on initial establishment
The optimal management for increasing germination, establishment, survival and reproduction differed considerably for the nine target species (Appendix 2). Still, when all species were treated as a group, certain patterns emerged in terms of the ‘best’ management in high and low productive situations. Hardly any seedlings emerged at the sites with high biomass (means 650-850 g/m2 in July) in the second year (2003), in contrast to the sites with low biomass (350-450 g/m2), where a small second cohort of seedlings emerged. This is in contrasts with a study in moist grasslands (Bakker et al. 1980), where new cohorts of seedlings emerged until four years after the introduction of seeds, despite the high biomass (600-900 g/m2). However, these seedlings did not survive, which is in agreement with the present experiment.
Our results indicate that for ‘initial’ establishment, intensive management is more beneficial in ditch banks with high biomass than with low biomass, confirming the findings of Melman (1991). In high biomass ditch banks, germination and ‘initial’ establishment in 2002 were significantly higher with early grazing or first cuts (‘normal management’ (TIV) > ‘May mowing’ (TIII) and with more cuts (‘June+ mowing (TII)) and lowest with the latest first cut (‘July mowing’ (TI), Tables 1 and 3, Figs. 4a,b). Although differences were not significant, the highest establishment in 2003 was found in the treatments with the most cuts, i.e. with ‘June+ mowing’ (TII) and ‘normal management’ (TIV). The beneficial effects ‘early’ and ‘intensive’ management must be linked to the reduction of (light) competition (early) in the growing season providing safe sites both for germination and initial establishment and corresponds to earlier findings by e.g. (Bullock et al. 1994; Oomes & van der Werf 1996; Jones & Hayes 1999; Isselstein et al. 2002). In low biomass ditch banks, the vegetation is more open, and indeed, germination and establishment, did not differ much between treatments. Unfortunately, we cannot compare ‘intensive normal management’ (TIV) between ditch banks with low and high biomass, since ‘normal management’ was very extensive at sites with low biomass (Table 1). However, considering the similarity the results of different treatments in low biomass ditch banks it is reasonable to assume that earlier grazing or mowing would not be beneficial. This all goes to show that, for initial establishment, management does matter in high productive ditch banks, while the type of applied management is less important in low productive situations (confirming the findings of Melman (1991)). However, it should be remembered that this applies as long as a late cut is present in the autumn; without this late cut fewer seedlings may have established (Bakker et al. 1980; Oomes & van der Werf 1996; Jones & Hayes 1999).
Effects of biomass and management regime on survival and reproduction
Chapter 5 112
enhanced by choosing one of these three treatments. However, survival within low biomass ditch banks was significantly lower with ‘normal management’ (TIV). The fact that this treatment was accidentally (partially) covered with ditch sediment on one farm (F5) in the autumn 2002, probably partly explains this result (and also the low survival of established seedlings with ‘normal management’ in 2003, Fig. 4c). Within ditch banks with high biomass, although no significant differences were found in survival 2003, most plants survived with (intensive) ‘normal management’ (Figs. 4d, 5). In an introduction experiment at a salt marsh, Bakker et al. (1985) found that early establishment was much higher at cattle-grazed than at abandoned sites, which was attributed largely to higher light levels at the grazed sites. The higher survival in the intensive ‘normal management’ in this study was therefore probably the result of the beneficial effects of low (light) competition when the (small) transplants were growing fast in 2002.
As for reproduction, most species benefited especially from ‘May mowing’ (TIII) in 2002 and 2003, but also from ‘normal management’(TIV) within low biomass ditch banks in 2003 (Table 3, Fig. 4f, Appendix 2). The positive effect of ‘May mowing’ (TIII) is probably related to the early removal of competing plants followed by a long undisturbed period allowing species to flower and set seed. As for ‘normal management’ (TIV), we know that, in 2003, ‘normal management’ was somewhere between ‘June+ and July mowing’ (TII, III) in terms of intensity (Table 1). Yet, reproduction was lower with ‘July’ and with ‘June mowing’. The beneficial effects of ‘normal management’ (TIV) could be related to the fact that these plots were grazed rather than mown, allowing for more small-scale variation and thus opportunities to ‘escape’. Contrary to our expectations, ‘July mowing’ (TI), which is presumed to allow for the best reproduction opportunities (Melman 1991), had lower reproduction rates than May mowing (TIII). This indicates that extensive management recommendations advocating late first cuts should be changed, and that an earlier first cut is beneficial even in ditch banks with low biomass.
Management implications
Ditch bank target species - and thereby species richness - is limited more by colonisation (seed and microsite limitation) than by extinction (site limitation). Since target species are often lacking from the seed bank, restoration of species richness in ditch banks should focus on improving dispersal. If target species are absent from neighbouring farms, introduction should be considered.
Conditions for germination and initial establishment also need to be improved. Since gaps are essential for the germination of species in ditch banks, a certain amount of trampling by cows and / or sheep will be beneficial. Moreover, the vegetation should be kept short early in the growing season (April-May) to reduce competition in this initial establishment phase. Such measures are crucial especially in high productive ditch banks with a peak biomass in July above 600 g/m2, where microsite limitation is more severe. Overall, early grazing or cutting, followed by light grazing in July (thinning out the vegetation, but giving plants some chance to flower and set seed) and a final total cut in the autumn appears to be the best approach in these situations. In low productive ditch banks (300-600 g/m2) mowing in May and autumn or grazing in June, possibly followed by light grazing later in the summer (plus a final cut in the autumn) are better strategies for increasing survival and reproduction than management strategies with late first cuts in July. Current management recommendations in agri-environment schemes (DLG 2000) should therefore be revised and separately defined for high and low productive ditch banks. Since the recommendations in this study are founded on ecological principles and processes, our results are also applicable for (restoration)management in other similar systems, including grasslands in nature reserves.
In the end, however, we should not forget that all these management recommendations are based on short-term research. Long-term structural changes may create situations where other management is needed. Moreover, since the optimal management differs among species and different stages in the life-cycle, variation in management both in space and time is likely to give the best results. Continuous monitoring (by the farmer or the supervising authorities) and adaptation of the implemented management to fit each individual situation are therefore necessary.
Acknowledgements
Chapter 5 114
Reference List
Bakker, E. S. 2003. Herbivores as mediators of their environment: the impact of large and small species on vegetation dynamics. PhD thesis, Wageningen University, NL.
Bakker, J. P. 1989. Nature management by grazing and cutting. Kluwer Publishing, Dordrecht, NL.
Bakker, J. P. & Berendse, F. 1999. Constraints in the restoration of ecological diversity in grassland and heathland communities. Trends. Ecol. Evol. 14: 63-68.
Bakker, J. P. & de Vries, D. M. 1992. Germination and early establishment of halophytes and glycophytes on a grazed and mown salt marsh. J. Veg. Sci. 3: 274-252.
Bakker, J. P., Dekker, M., & de Vries, Y. 1980. The effect of different management practices on a grassland community and the resulting fate of seedlings. Acta Bot. Neerl. 29: 469-482. Bakker, J. P., Dijkstra, M., & Russchen, P. T. 1985. Dispersal, germination and early
establishment of halophytes and glycopgytes on a grazed and abandoned salt-marsh gradient. New Phytologist 101: 291-308.
Bakker, J. P. & Olff, H. 1992. Restoration management in moist grasslands in the Netherlands. Norddeutsche Naturschutzakademie NIA Berichte 5: 42-48.
Bakker, J. P., Poschlod, P., Strykstra, R. J., Bekker, R. M., & Thompson, K. 1996. Seed banks and seed dispersal: important topics in restoration ecology. Acta Bot. Neerl. 45: 461-490. Baudry, J. 1988. Hedgerows and hedgerow networks as wildlife habitat in agricultural landscapes.
In: Park,J.R.(ed.) Environmental management in agriculture. European Perspectives, pp. 111-123. Belhaven Press, London, UK.
Bazzaz, F. A. 1975. Plant species diversity in old-field successional ecosystems in Southern Illinois. Ecology 56: 485-488.
Benton, T. G., Vickery, J. A., & Wilson, J. D. 2003. Farmland biodiversity: is habitat heterogeneity the key? Trends. Ecol. Evol. 18: 182-188.
Berendse, F., Oomes, M. J. M., Altena, H. J., & Elberse, W. Th. 1999. Experiments on the restoration of species-rich meadows in The Netherlands. Biol. Conserv. 62: 59-65. Blomqvist, M. M., Bekker, R. M., & Vos, P. 2003a. The potential for restoration of plant species
richness of grassland ditch banks from the soil seed bank. Appl. Veg. Sci. 6: 179-188. Blomqvist, M. M. & Tamis, W. L. M. The effects of agri-environment schemes on productivity
and plant species diversity in ditch banks. (unpublished manuscript, Chapter 4)
Blomqvist, M. M., Vos, P., Klinkhamer, P. G. L., & ter Keurs, W. J. 2003b. Declining plant species richness of grassland ditch banks - a problem of colonisation or extinction? Biol. Conserv. 109: 391-406.
Bullock, J. M., Clear Hill, B., Dale, M. P., & Silvertown, J. W. 1994. An experimental study of the effects of sheep grazing on vegetation change in a species-poor grassland and the role of seedling recruitment into gaps. J. Appl. Ecol. 31: 493-507.
Burke, M. J. & Grime, J. P. 1996. An experimental study of plant community invasibility. Ecology 77: 776-790.
CBS 1997. Register Biodiversiteit (BIOBASE). Centraal Bureau voor de Statistiek, Voorburg / Heerlen, NL (in Dutch).
Clausman, P. H. M. A. & van Wijngaarden, W. 1984. Het vegetatie onderzoek van de provincie Zuid-Holland: Verspreiding en ecologie van wilde planten in Zuid-Holland. Ia waarderingsparameters, 64 pp. Provinciale Planologische Dienst Zuid-Holland, Den Haag, NL.
Crawley, M. J. 1993. GLIM for ecologists. Blackwell Scientific Publications, Oxford etc.
Dijkstra, A. A. 1994. Vegetatie-onderzoek relatiegebied Nieuwkoop 1990-1994. Publicate no. 76: Provincie Zuid-Holland / Consulentschap Natuur, Bos, Landschap en Fauna, / Directie beheer landbouwgronden, Ministerie Landbouw, Natuurbeheeer en Visserij, 's Gravenhage, NL.
Ellenberg, H., Weber, H. E., Düll, R., Wirth, V., Werner, W., & Paulissen, D. 1992. Zeigerwerte von Pflanzen in Mitteleuropa. Scripta Geobotanica 18: 1-258.
Eriksson, O. & Ehrlén, J. 1992. Seed and microsite limitation of recruitment in plant populations. Oecologia 91: 360-364.
Fenner, M. 1985. Seed Ecology. Chapman & Hall, London, UK.
Foster, B. L. 2001. Constraints on colonisation and species richness along a grassland productivity gradient: the role of propagule availability. Ecology Letters 4: 530-535.
Foster, B. L., Dickson, T. L., Murphy, C. A., Karel, I. S., & Smith, V. H. 2004. Propagule pools mediate community assembly and diversity - ecosystem regulation along a grassland productivity gradient. J. Ecol. 92: 435-449.
Foster, B. L. & Tilman, D. 2003. Seed limitation and the regulation of community structure in oak savanna grassland. J. Ecol. 91: 999-1007.
Fry, G. L. A. 1994. The role of field margins in the landscape. In: Boatman, N.(ed.) Field Margins: integrating agriculture and conservation, pp. 31-40. BCPC (British Crop Protection Council), Surrey, UK.
Geertsema, W. 2002. Plant survival in dynamic habitat networks in agricultural landscapes. PhD thesis. Wageningen University, Wageningen, NL.
Grubb, P. J. 1977. The maintenance of species richness in plant communities: The importance of the regeneration niche. Biological Reviews of the Cambridge Philosophical Society 52: 107-145.
Hodgson, J. G. 1986. Commonness and rarity in plants with special reference to the Sheffield flora. Part II: the relative importance of climate, soils and land use. Biol. Conserv. 36: 253-274.
Isselstein, J., Tallowin, J. R. B., & Smith, R. E. N. 2002. Factors affecting seed germination and seedling establishment of fen-meadow species. Rest. Ecol. 10: 173-184.
Jansen, F. A., Kruk, M. & ter Keurs, W. J. 1989. Natuurproductiebetaling in het veenweidengebied: een eenvoudige bepaling van de natuurwaarden van slootkantvegetaties. Landbouwkundig tijdschrift 101 (10): 5-9.
Jones, A. T. & Hayes, M. J. 1999. Increasing floristic diversity in grassland: the effects of management regime and provenance of species introduction. Biol. Conserv. 87: 381-390. Kitajima, K. & Tilman, D. 1996. Seed banks and seedling establishment on an experimental
productivity gradient. Oikos 76: 381-391.
Kleijn, D., Berendse, F., Smit, R., & Gillisen, N. 2001. Agri-environment schemes do not effectively protect biodiversity in Dutch agricultural landscapes. Nature 413: 723-725. Kruk, M. 1991. Somerset Levels & Moors: een veenweidegebied in Engeland. Veenweide 4:
21-22.
Kruk, M., Twisk, W., de Graaf, H. J., & ter Keurs, W. J. 1994. An experiment with paying for conservation results concerning ditch bank vegetations in the western peat district in the Netherlands. BCPC Monographs no. 58: Field margins: integrating agriculuture and conservation, pp. 397-402.
LNV 1995. Regeling beheersovereenkomsten en natuurontwikkeling. Ministerie LNV (Ministry of Agriculture, Nature Management and Fisheries), Landinrichting en beheer landbouwgronden, Utrecht, NL.
Marshall, E. J. P. & Hopkins, A. 1990. Plant species composition and dispersal in agricultural land. In: Bunce,R.G.H. & Howard,D.C.(eds.) Species dispersal in agricultural habitats, pp. 98-116. Belhaven Press, London.
Melman, C. P. & van Strien, A. J. 1990. Slootkanten in Veenweidegebieden: mogelijkheden voor natuurgerichte inrichting en beheer. CML mededeling no. 64, mei 1990, Leiden, NL. Melman, D. 1991. Slootkanten in het veenweidegebied: mogelijkheden voor behoud en
Chapter 5 116
Melman, D. & van Strien, A. 1993. Ditch banks as a conservation focus in intensively exploited peat farmland. In: Vos, C. & Opdam, P.(eds.) Landscape ecology of a stressed environment, pp. 122-141. Chapman & Hall, London, UK.
Olff, H. & Bakker, J. P. 1991. Long-term dynamics of standing crop and species composition after the cessation of fertilizer application to mown grassland. J. Appl. Ecol. 28: 1040-1052. Oomes, M. J. M. & van der Werf, A. 1996. Restoration of species diversity in grasslands: the
effect of grassland management and changes in ground water level. Acta Botanica Gallica 143: 451-461.
Oude Voshaar, J. H. 1995. Statistiek voor onderzoekers. Wageningen Pers, Wageningen, NL. Ovenden, G. N., Swash, A. R. H., & Smallshire, D. 1998. Agri-environmental schemes and their
contribution to the conservation of biodiversity in England. J. Appl. Ecol. 35: 955-960. Pakeman, R. J., Attwood, J. P., & Engelen, J. 1998. Sources of plants colonizing experimentally
disturbed patches in an acidic grassland, in eastern England. J. Ecol. 86: 1032-1041. Pfadenhauer, J. & Klötzli, F. 1996. Restoration experiments in middle European wet terrestrial
ecosystems: an overview. Vegetatio 126: 101-115.
Prins, A. H., Dijkstra, G. A., & Bekker, R. M. 1998. Feasibility of target communities in a Dutch brook valley system. Acta Bot. Neerl. 47: 71-88.
Ryser, P., Lagenhauer, R., & Gigon, A. 1995. Species richness and vegetation structure in a limestone grassland after 15 years management with six biomass removal regimes. Folia Geobot. Phytotax. 30: 157-167.
Siegel, S. & Castellan, N. J., Jr. 1988. Nonparametric statistics for the behavioural sciences. McGraw-Hill Book Company, New York etc.
Stampfli, A. & Zeiter, M. 1999. Plant species decline due to abandonment of meadows cannot easily be reversed by mowing. A case study from the southern Alps. J. Veg. Sci. 10: 151-164.
Stoate, C., Boatman, N. D., Borralho, R. J., Rio Carvalho, C., de Snoo, G. R., & Eden, P. 2001. Ecological impacts of arable intensification in Europe. J. Environ. Manage. 63: 337-365. Tamis, W. L. M., van 't Zelfde, M., van Elk, R. & Witte, J. P. M. 2000. Modellering van de
kansrijkdom van het biologisch herstel van natte en vochtige vegetaties. CLM & RIZA, Lelystad, NL.
Thompson, K., Bakker, J. P. & Bekker, R. M. 1997. The soil seed banks of North West Europe: methodology, density and longevity. Cambridge University Press, Cambridge, UK. Turnbull, L. A., Crawley, M. J., & Rees, M. 2000. Are plant populations seed-limited? A review
of seed sowing experiments. Oikos 88: 225-238.
Twisk, W., Noordervliet, M. A. W., & ter Keurs, W. J. 2003. The nature value of the ditch vegetation in peat areas in relation to farm management. Aquatic Ecology 37: 191-209. van Dorp, D. 1996. Seed dispersal in agricultural habitats and the restoration of species-rich
meadows. PhD thesis, Wageningen Agricultural University, Wageningen, NL.
van Groenendael, J. M., Bekker, R. M. & van Heijst, M. 1989. The accessibility of a former pasture for new species: implications for management. Brighton Crop Protection Conference - Weeds - 1989: 1075-1080, Brighton, UK.
van Strien, A. 1991. Maintenance of plant species diversity on dairy farms. PhD thesis, Leiden University, Leiden, NL.
Watt, T. A. & Gibson, W. D. 1988. The effect of sheep grazing on seedling establishment and survival in grassland. Vegetatio 78: 91-98.
Wiertz, J. 1992. Schatting van ontbrekende vocht- en stikstofindicatiegetallen van Ellenberg (1979). Instituut voor Bos- en Natuuronderzoek (IBN-DLO), Wageningen, NL.
Appendix 1. Seeds and plants used in the sowing and transplanting experiments
Seeds that had been collected in the wild were obtained from Biodivers bv (www.biodivers.nl). Most seeds had been collected in 2001 (year) Approximately 100 seeds per species were sown in individual gaps and control treatments, totalling around 12 000 seeds per species (6 ditch banks x 4 treatments x 5 replicates x 100 seeds). However, since it was not possible to obtain enough seeds for all species and since the seeds were counted by machine (CABO Contador: Pfeuffer Mess- und Prüfgeräte), numbers varied per gap. We calculated mean, minimum and maximum numbers per gap by hand-counting 20 replicate seed samples per species (no. of seeds: mean, min, max). These seeds were also weighed to establish mean seed weights of individual seeds (seed
wt in mg).
From these 20 seed samples we randomly selected 6 samples per species for the determination of the germinability. Three samples were placed in on filter paper in petri dishes 15/25˚C (16/8 h light/dark) for 22 days (warm treatment). The three remaining samples received the same treatment, after they had first been stored moist and dark in +4˚C for 4 weeks (cold treatment). The filter paper was kept moist throughout the trial. Germinated seeds (visible dicotyledons) were counted on a regular basis. After 22 days, we checked how many of the rest of the seeds were viable by examining the endosperm under a microscope. Seeds with a firm, white endosperm were classified as viable. Here we report the % of germinated seeds per treatment (% germ. seeds: warm and cold): only Caltha had significantly higher germination rates in the cold treatment (F1,5 = 18.892, p = 0.012). Since the number of viable seeds in the cold and warm treatment did not differ from one another (not shown) we further report average proportion of viable seeds in both treatments (% viable seeds mean) and the maximum proportion of viable seeds (% viable seeds max).
Sp year No. of seeds Seed % Germ. seeds % Viable seeds
wt warm cold
mean min max (mg) mean mean mean max
CAL 2001 94 86 102 0.835 14.9 30.0 47.8 52.1 CIR 2001 76 58 97 0.972 14.0 16.4 15.4 19.8 LOT 2001 102 92 112 0.498 19.3 21.1 89.7 91.2 LYC 2001 127 112 139 0.124 80.8 -1 84.0 85.7 LYS 2001 91 81 99 0.280 23.4 25.0 38.3 46.4 LYT 2000 127 112 143 0.069 92.2 95.0 93.9 100.0 MEN 2000 101 81 115 0.117 53.9 58.5 59.2 63.9 PRU 2000 93 65 112 0.687 73.7 66.4 71.3 83.8 VAL 2001 88 78 103 0.680 67.0 72.3 70.0 76.1
Chapter 5 118
Appendix 1. contd.
The transplanted plants were grown in germination trays. Due to practical reasons only a limited number of trays could be transported at any given time. This could cause two types of biases: 1) the size of the transplanted plants might differ between farms and 2) the size of the plants may systematically vary within each ditch bank. To check this, we measured the size (height. diameter or no. leaves) of 24 plants of each species before transplanting (8 at the beginning. 8 in the middle and 8 at the end of the ditch bank). There were no significant systematic differences in any species within ditch banks (not shown). In other words, the treatment effects (I-IV) were not caused by systematic differences in plant size. However, we did find significant differences in the mean sizes of the transplanted plants between farms. Overall plant size can be ranked among the farms as follows: 4 < 5 < 1 < 3 ≤ 2 < 6. Since the size of the transplanted plants did not differ systematically between farms with high (farms 1, 2, 3) and low (farms 4, 5, 6) biomass, and since the differences usually did not exceed 1-2 cm in absolute sizes, we do not believe that transplant size affected our results.
Appendix 2. (opposite page). A. Mean ‘maximum’ germination 2002 (per no. of sown
seeds), establishment 2002 and 2003 (per maximum germination 2002) and survival 2003, reproduction 2002 and 2003 (per plants present at beginning) for nine species. Means are given for ditch banks with low and high biomass (combined for treatments TI, II, III), for three treatments (TI = ‘May mowing’, TII = ‘June+ mowing’, TIII = ‘July mowing’; combined for low and high biomass) and for biomass-treatment groups (TIV = ‘normal management’). The best management situation is indicated per group in bold letter type, worst situation is indicated in italics. Overall sum of best and worst management is indicated at the end. B. Results from GLMM species-level analyses (see methods) for the same six dependent variables: Bio = Biomass effect, Treat = Treatment effect, B x T = Biomass*treatment interaction. P < 0.05 indicated with bold letter type, p < 0.1 indicated with bold italics. Shaded fields in A. refer to significant differences in the GLMM analyses. Species acronyms (sp) refer to the following nine species: Caltha
Appendix 2. A. Means
Dependent Biomass Treatment Low biomass High biomass
variable sp low high TI TII TIII TI TII TIII TIV TI TII TIII TIV
Appendix 2. contd. A. Means
Dependent Biomass Treatment Low biomass High biomass
variable sp low high TI TII TIII TI TII TIII TIV TI TII TIII TIV
Appendix 2. contd.
B. Results GLMM-analyses
Sp Germination 2002 Establishment 2002 Establishment 2003
Bio Treat B x P Bio Treat B x P Bio Treat B x P
CAL 0.584 0.446 0.202 0.685 0.878 0.935 0.344 0.093 0.809 CIR 0.073 0.185 0.444 0.073 0.540 0.833 0.443 0.494 0.453 LOT 0.002 0.594 0.312 0.147 0.632 0.143 0.139 0.751 0.161 LYC 0.130 0.281 0.753 0.493 0.892 0.660 0.363 0.668 0.571 LYS 0.966 0.808 0.357 0.667 0.736 0.975 0.690 0.994 0.993 LYT 0.785 0.855 0.583 0.176 0.071 0.157 0.016 <0.001 0.488 MEN 0.273 0.797 0.609 0.624 0.662 0.882 0.535 0.502 0.935 PRU 0.431 0.472 0.002 0.069 0.166 0.727 0.118 0.033 0.455 VAL 0.376 0.080 0.573 0.583 0.428 0.674 0.924 0.242 0.930
Survival 2003 Reproduction 2002 Reproduction 2003
Bio Treat B x P Bio Treat B x P Bio Treat B x P
CAL 0.411 0.365 0.203 0.422 0.661 0.920 0.668 0.588 0.909 CIR 0.811 0.931 0.376 0.889 0.080 0.901 0.901 0.581 0.096 LOT 0.013 0.788 0.185 0.222 0.735 0.526 0.061 0.123 0.062 LYC 0.116 0.002 0.030 0.862 0.004 0.175 0.472 0.019 0.507 LYS 0.005 0.826 0.432 0.8451 0.1491 0.1771 - - - LYT <0.001 0.514 0.442 0.163 0.227 0.746 0.002 <0.001 0.442 MEN 0.064 0.091 1.000 0.017 0.006 0.878 0.581 <0.001 0.754 PRU 0.024 <0.001 0.315 0.107 <0.001 0.396 0.337 <0.001 0.244 VAL 0.236 0.494 0.042 - - - 0.235 0.185 0.986
1 Due to a large number of zeros, the iterative process did not converge when both random factors (ditch bank and ditch bank*treatment) were included.
We therefore report estimates from a model including only ditch bank as random factor.
