Introduction

Agri-environment schemes (AES) are increasingly used in agricultural landscapes, with primary aims to enhance biodiversity and to maintain or promote attractive or historic landscapes (e.g., Whittingham 2007). Implementation of these schemes is subsidised with public money, from regional, national or European funds. Since AES might not always be effective in reaching conservation goals (Kleijn and Sutherland 2003; Kleijn et al. 2006; Blomqvist et al. 2009), evaluation should be an important aspect to confirm the value or to propose improvements of the management involved. A common AES is the creation and management of all types of semi-natural vegetation. Frequently established semi-natural areas on arable lands are field margin habitats (e.g., de Snoo 1999). These margins can have a range of nature conservation functions, including the provision of wildlife habitat, refuges and corridors (e.g., Asteraki et al. 2004; Walker et al. 2007). In addition, these margins may provide other useful functions, such as the reduction of manure and pesticide drift into adjacent ditches (de Snoo and de Wit 1998; Marshall and Moonen 2002).

The creation of field margins is a widely applied AES in the Netherlands (Manhoudt and de Snoo 2003). In this article, we focus on the margins created in the most south-western part of this country: the Province of Zeeland. Although the primary goal of these particular margins is to provide habitat for animals, the composition of the vegetation is regarded as highly important. First, because the vegetation composition determines to a large extend the animal composition (Schaffers et al. 2008). Second, plant species richness has decreased dramatically in agricultural landscapes, and field margins are expected to act as refuges for plant species richness (Kiss et al. 1997; Moonen and Marshall 2001). Third, some plant species are unfavourable to farmers, as seeds or roots easily intrude the arable fields (de Cauwer et al. 2008; Musters et al. 2009). And finally, plant species composition determines the general appearance of the margins, which holds importance to public opinion and acceptance by farmers of these margins (Marshall and Moonen 2002).

The first goal of this study was to gain insight in vegetation development of newly sown field margins under a widely applied AES which prescribes that these elements have to be maintained for 6 years. We analysed species richness and composition as a function of time after establishment. We expect the total species number to increase in time, due to the stop of the manure applications and possibilities for the colonisation of not-sown plants. However, we expect the sown species to decrease, because many of them are annuals that probably will not be able to maintain. In the same area, we performed a second study, surveying plant species richness and composition on ditch banks bordered by field margins of different ages. It is expected that the presence of field margins has beneficial effects on species richness on these ditch banks, since they are fenced off from negative influences from the arable fields, like manure and pesticides, and that this effect increases over time.

Our research questions are:

  1. 1

    Does vascular plant species richness increase in arable field margins sown as habitat for animal species?

  2. 2

    Does the age of field margins affect species richness of adjacent ditch banks?

The results are used to formulate management recommendations that could improve plant species richness in these agricultural habitats with time.

Field sites and methods

Field sites

For our first study, we selected 31 arable farms with sown field margins in the province of Zeeland, a marine clay area in the most south-western part of the Netherlands. Of 18 farms, two margins were selected. Not all 49 margins were visited both years: 44 margins in 2006 and 41 margins in 2007 were surveyed, and from 34 margins, data were available from both years. All selected farms had contracts under the AES ‘Faunarand’ (=fauna margin), and we questioned the farmer on the history of the margins. A contract under this particular scheme lasts for 6 years and prescribes the creation and sowing of a field edge that is at least 6 m wide and 50 m long. The seed mixes used by farmers differed and were not always known. As far as we could find out, at least 13 different mixtures were used. Of these, 12 were dominated by herb species, of which again 6 included some grasses. The average number of species was 13.4 (max: 43, min: 2). Table 2 in “Appendix” gives information on the species sown and the frequencies in which the species were encountered in the field. Some farmers implemented this scheme on an already existing margin, but scarification and sowing (and thus the start of vegetation development) was always done at the start of the contract. The age of the margins in our figures therefore starts at ‘1’ in the year of sowing. Mowing or mulching is permitted once a year between 15th July and 14th September, and regularly done, but the removal of the cuttings is not required, and in practice never done. The application of manure or pesticides on the margins is prohibited, but local control of Cirsium arvense and Rumex obtusifolius is allowed.

For the second study, we selected 32 ditch banks in the same region as the first study, all of which were separated from the arable fields by a field margin. All of the locations coincided with the location of our first study. All ditch banks are in all probability constructed in the 1950s. Regional water boards own and manage these ditch banks, generally by mulching once a year.

Vegetation and site characterisation

In week 26–27 of 2006 and 2007, vascular plant species composition on the field margins was recorded in 1 × 25 m. A recording was positioned in the middle of each margin, never done in margins mowed before within the same year and at least 10 m from field corners or disturbances such as tire tracks. For our second study, we made in 1 × 25 m vegetation recording in week 26–27 of 2006 in the middle of ditch banks which are separated from the arable fields by a field margin. Again, we kept at least 10 m away from field corners or obvious vegetation disturbances. Plant species occurrence was noted using nomenclature according to van der Meijden (2005), and their abundance was estimated and directed into classes following an adapted Braun-Blanquet method (cf. Barkman et al. 1964). In addition, we made a distinction between sown and not-sown species from information supplied by local agri-environmental farmer collectives (see Table 2 in “Appendix”). In the sown mixtures, a number of indigenous plants are included; these were excluded from the ‘not-sown plant species richness’ analysis, to give the most reliable impression of species richness of colonising plants. Species considered ‘sown’, but that may also colonise these margins, include Achillea millefolium, Daucus carota, Symphytum officinale, Tanacetum vulgare, Trifolium pratense, Trifolium repens and Tripleurospermum maritimum.

For determining the total nitrogen concentration, soil samples were taken from a depth of 10 cm at five sites within each plot. The five samples were pooled and thoroughly mixed. Ten gram of fresh weight soil was extracted for 1 h with 1 M KCl on a rotary shaker, and water content was determined gravimetrically. Nitrogen content was measured according to Keeney and Nelson (1982).

Analyses

An explorative principal component analysis (PCA) was performed on all plant species and their log-transformed cover data (in Canoco 4.5). The data were centred by species, not by samples. In order to detect the important clusters of sites based on plant species occurrence, the species and margins shown in the graph are those with a fit higher than 13% in the analysis.

We used Hierarchical Generalised Linear Models (HGLM), a generalized mixed model procedure of GenStat 12.0, to calculate the relationship between age of the field margin and plant species richness, given the fact that we chose certain farms and years for sampling (Royle and Dorazio 2008). In our models, age of the margin was the main fixed factor. Since we sampled usually two field margins per farm over 2 years, farm, year of sampling, and their interaction were included as random factors. In case of the ditch bank, that we sampled only 1 year, farm was included as random factor. For analysing confounding factors, we included the covariates nitrogen content and width of the field margin, and the factor crop in the field in the fixed part of the model. Plant species richness was log-transformed, ln(species richness + 1), to achieve a normal distribution of our response variable. After this transformation, we could use the identity link function both for the fixed and the random part of the model. Age of the field margin was regarded as a categorical variable; nitrogen content was square root transformed to get it close to a normal distribution and crop was in eight categories. The Wald test for testing the change in likelihood between the full model and the reduced model when taking out a variable was used for testing the significance of the fixed variables.

For making the graphs of the relationship between age of the field margin and plant species richness, we applied the complete HGLM’s for estimating the mean species richness (±s.e.) per age category. In one case, species richness of all species in the field margins where age of the field margin had a significant contribution to the model, we also estimated the regression line of age as scaled variable on species richness, again including all other factors in the model. We tested whether the regression coefficient deviated significantly from zero by applying a Wald test again.

For testing whether the age of the field margin affected the nitrogen content of the soil, we again applied a HGLM on square rooted nitrogen content as response variable with the scaled variable age as fixed factor and farm, year of sampling, and their interaction were included as random factors.

Results

Succession of plant composition in field margins

We recorded 164 vascular plant species in the field margins, of which 68 species are considered to be sown (Table 2 in “Appendix”). The PCA analysis, of which both axes together accounted for 23.4% of the variation in species composition, showed clear clusters of field margins in relation to their age (Fig. 1). The lower-right part of the graph is dominated by the margins in their first year and the abundant species are many sown species and several ruderal annual species, for example, Chenopodium spp., Persicaria spp. and Solanum nigrum. The upper cluster consists of margins that are 2 years old. Characteristic plants for the margins in this age class are predominately sown biennial or perennial species, for example, Achillea millefolium, Daucus carota, Silene latifolia, etc. The cluster on the lower-left part of the graph represents mostly margins that are 5 and 6 years old. Only a few plant species typify these margins, three grass species, Elytrigia repens, Dactylis glomerata and Poa trivialis, and the nitrophilous species Urtica dioica (Fig. 2). From the sown species, only Centaurea jacea seems well established here.

Fig. 1
figure 1

PCA graph (first two axes) showing the relations between the plant species (arrows and abbreviations) and the field margins, for both only if there is a fit of 13% or more in the analysis. Plant species are abbreviated by the first five letters of their genus name and the first three letters of their species name. Sown species are underlined. The field margins of different ages are given different symbols: filled triangle margin in the first year, filled gray square margin in second year, filled square margin in third year, open square margin in fourth year, filled circle margin in fifth year, open square margin in sixth year

Full size image
Fig. 2
figure 2

Example of a field margin in its first year and one after sixth year. Both were sown with a mixture dominated by herbs. Species composition—and general appearance—has changed dramatically (Photos: B. Kruijsen)

Full size image

Succession of plant species richness in field margins

We found the age of the field margin to significantly affect overall species richness (Table 1) and that species richness decreases with age (Wald stat. = 11.889, df = 1, approx. pr. = 0.001, Fig. 3). The age of the field margin did not affect species richness of not-sown plants, so the decrease is a result of the decrease of sown plants (Table 1; Fig. 3). These results were not affected by any of the possible confounding factors, although nitrogen content might had a marginal significant, but unexpectedly positive effect on not-sown species richness (Table 1). No changes in nitrogen content of the soil in relation to the age of the field margin could be found (Wald-stat. = 0.0004, df = 1; P = 0.983).

Table 1 Results of the complete HGLM
Full size table
Fig. 3
figure 3

Estimated mean diversity (±s.e.) and linear regression of all and not-sown plant species in relation to the age of field margin, based on HGLM (Table 1). Age 1 reflects the first growing season after the margins were sown in winter. The decrease in species richness of all species in field margins is significantly different from zero (P < 0.001)

Full size image

Plant species richness on ditch banks bordered by field margins

We recorded 53 vascular plant species in the 32 ditch banks (Table 3 in “Appendix”); only one potentially sown species was recorded: Symphytum officinale. However, this species is highly characteristic for ditch banks and it cannot be excluded that it grows here spontaneously. A PCA analysis did not reveal any apparent clusters (not shown here). In addition, we found no significant effect in plant species richness on ditch banks of the age of the field margin (Table 1). Several grass species (Agrostis stolonifera, Arrhenatherum elatius, Dactylis glomerata and Phragmites australis), including the weed Elytriga repens and Urtica dioica, are the most frequently found species, both in ditch banks next to the recent and next to the older field margins.

Discussion

Vegetation development

Sown species dominated the recently established field margins. Farmers prefer to sow species in their margins instead of allowing spontaneously colonising vegetation, since they expect weedy annual and rhizomatous species can be prevented this way (van der Meulen et al. 1996; West et al. 1997; Smith et al. 1999). However, the seed mixtures used in Zeeland provide barely protection to both plant groups. Weedy annuals are abundantly found in the margins in their first year; Persicaria and Chenopodium species which have many small seeds that may easily intrude arable fields (de Cauwer et al. 2008). Sequentially, the field margins readily turn into strips with grass species and few herbs (Fig. 2). Both the sown and the annual species do not have many opportunities here and disappear quickly (see also Smith et al. 1999; de Cauwer et al. 2005). The three grass species in the cluster representing mainly the old margins (in their fifth and sixth years) are typical for very nutrient-rich situations. A few margins with a younger age can be found in this cluster as well, probably those which have severe nitrogen stress, and the vegetation thus developed quicker into the types with grass dominance. The characteristic and very abundant species in this cluster Elytriga repens and Urtica dioica are considered noxious weeds, because they are rhizomatous (de Cauwer et al. 2008). These species render the margins less accessible to the farmers and roots might spread into the arable field. Most farmers who had such old margins noticed these undesirable effects and some actually stated they would not enter a similar AES again. Generally, the vegetation succession follows the ones described by Hodgson (1989) in the UK and by de Cauwer et al. (2005) in Belgium. It would be relevant to know whether other seed mixtures would result in the same succession. More research is needed here.

Although the composition of the vegetation changes dramatically, the species richness of not-sown species remains stable (see also de Cauwer et al. 2005). Our expectation that in the margins species richness would increase with age could therefore not be affirmed. Old field margins are dominated by few plant species characteristic for very nutrient-rich situations. The nitrogen levels in the field margin soil did not decrease with the age of the margin. One would expect that margins that are avoided during manure applications would loose nutrients due to leaching from the soil and as a consequence decrease in fertility. However, in the clay soils of our study area, leaching is probably very slow or even absent. And since mowing of the margins does not include the removal of plant material here, nutrients stay available in the field margins.

Although some studies show short-term positive effects of adjacent field margins on plant species richness of the ditch banks, due to the fencing off from pollution and disturbances from the arable fields (e.g., Kleijn and Snoeijing 1997; de Snoo and van der Poll 1999; Musters et al. 2009), the banks studied in Zeeland did not develop into more species-rich sites within the studied 6 years. This is probably due the mulching management, which also does not include removal of plant material, resulting in remaining nutrient availability. In all the cases studied by Muster et al. (2009), plant removal took place.

In conclusion, these field margins initially created for fauna, offer poor opportunities for plant species richness conservation, probably due to a management inappropriate for this group. An alternative explanation could be that the sown seed mixture is inappropriate.

Management recommendations

In order to obtain higher plant species richness on field margins and ditch banks, several management and landscape characteristics are important, such as a reduction of fertilizer inputs from the fields (Kleijn and Verbeek 2002) and the presence of nearby species-rich areas (Kohler et al. 2008; Leng et al. 2009). However, from our study it is clear that an increase in plant species richness may be hampered by the current management, both in the field margins as on the ditch banks. In the two linear elements, the lack of biomass removal may hamper a reduction of nutrients. Besides, litter accumulation may hinder the colonisation or growth of many plant species. Pinpointing appropriate management therefore seems a very important first step towards biodiversity gain in these habitats (Manhoudt et al. 2007). The implementation of a hay-making management (mowing with the sequential removal of the cutting) seems most urgent to enhance plant species richness. Various researches recommend this as well for such nutrient-enriched field margins, for example, Marshall and Nowakowski (1995), Smith et al. (1999), de Cauwer et al. (2005), Hovd and Skogen (2005) and Manhoudt et al. (2007). In addition, to avoid nutrient leaching from the mown vegetation, the quick removal of the cuttings is necessary (Schaffers et al. 1998). Not only is biomass removal associated with nutrient removal, it also prevents litter accumulation (Schaffers 2002). Dense litter and vegetation layers affect seed germination and seedling establishment (Schaffers 2002). Thus, the creation of open situations, either by hay-making or by locally opening up the vegetation, i.e. by creating gaps of bare ground, may offer opportunities for botanically diverse vegetation (Schaffers 2002; Blomqvist et al. 2006).

Farmers probably welcome a different management prescription of the AES, since a hay-making and opening management fits their perception of well-maintained fields and would result in fewer noxious weeds. However, this management requires extra costs to be made.