Vulnerability of biodiversity in the agro-ecosystem as influenced by green veining and land-use intensity (GREENVEINS)

Twelve novel polymorphic microsatellite loci have been developed for the dark bush cricket Pholidoptera griseoaptera. All loci are polymorphic, with up to 37 alleles per locus. These molecular marekrs will be useful tools to study the influence of landscape structure and land-use intensity in agricultural landscapes on genetic diversity within and among populations of Pholidoptera griseoaptera. The markers have been applied within the project for the study of Pholidoptera griseoaptera populations in two LTSs (Landscape Study Sites) in Switserland and Belgium (Diekötter et al., submitted). In addition, the markers are being published: Arens P, JH Wernke-Lenting, T Diekötter, C. Rothenbühler, M. Speelmans, F Hendrickx, MJM Smulders (2005) Isolation and characterisation of microsatellite loci in the dark bush cricket, Pholidoptera griseoaptera (Tettigoniidae). Molecular Ecology Notes 5 (in press)

The intensity of agricultural production was assessed in 25 landscape test sites (LTS) across temperate Europe using a standardised farmer questionnaire. The intensity indicators nitrogen input (to arable crops and to permanent grassland), density of livestock units, number of pesticide applications (herbicides, insecticides, fungicides, retardants) were recorded and integrated into an overall intensity index. All three components were needed to characterise the intensity of agricultural management. Four hypotheses were tested. - A low diversity of crops is related to higher intensity. The contrary was observed, namely because diverse crop rotations contained a higher share of crops which are more demanding in terms of nitrogen and of plant protection. -Intensity decreases when there is more permanent grassland. This was confirmed by our study. - Large farms are managed more intensively. There was no relation between farm size and intensity. - Large fields are managed more intensively. There was a tendency towards higher nitrogen input and livestock density in landscapes with larger fields but only a few of the results were statistically significant. The aggregated overall intensity index was of limited usefulness mainly because of limitations in interpretability.

To present the results of Greenveins project in public there was a need to create a suitable demonstration tool. Basic requirements, which the demonsration tool should meet, were: - Presentation of data of all LTS's together with scenarios which will show how the change of key factors affects the biodiversity for each LTS; - Ability to use user-defined parameters in the model to simulate relation of species richness to their changes; - To use such technological solution which will allow public availability of the tool and its easy use. To fulfill these requirements the Internet-based solution seems to be the best approach. The functionality of created internet application can be divided into two groups: - LTS-oriented: The user can view basic information of each LTS and there are also available graphs showing predicted change of species richness as a result of independent change of every single key factor in predefined steps. - Simulation with user defined parameters: The user can enter his own changes of key factors (effects) - either as absolute or relative values and the application will compute the predicted changes of species counts in all species groups. Technological solution: The developed Internet application is cross-platform, built only by using basic HTML, pregenerated images and client-side scripting. This allows simple integration to existing WWW presentations without any special requirements for the WWW server (operating system etc.). Since there is no server sripting used the application can be also used in offline mode (e.g. from CD) without any restrictions of functionality.

In the Greenveins project, the relationship between Land Use Intensity, Landscape Structure and several aspects of farmland biodiversity was explored in a large scale landscape ecological experiment covering 25 test sites spread over 7 countries across the temperate zone of Europe. The gathered data covered multiple species groups and several spatial scales. Data has been analysed and results have been published for several different research questions and species groups. In this paper, all these results are used to generate an integral picture of the relationship between farmland biodiversity LUI and LS. We discuss the possible application of results, the role and consequences of geographic variation and indicate directions for further research.

The speckled wood Pararge aegeria (L.) was selected as a model organism occupying natural and semi-natural habitats in order to analyse the relative impact of environmental variables subjected to anthropogenic alteration such as climate, land use and habitat quality. Its distribution was recorded at 23 test sites of 5 x 5 km in agricultural landscapes across seven European countries (France, Belgium, The Netherlands, Switzerland, Germany, Czech Republic, Estonia). Environmental predictors were mapped at a local (250 m) and a regional scale (5 x 5 km). We developed logistic regression models for two environmental scenarios. - The High abundance scenario was characterized by beneficial environmental and weather conditions coinciding with high local abundance of P. aegeria. - The Low abundance scenario reflected environmental stress and adverse weather conditions during larval development coinciding with low local abundance. The high abundance scenario revealed a low but equal effect of local and regional factors. Hence, P. aegeria was predicted to occur nearly anywhere under beneficial conditions. The low abundance scenario revealed totally different patterns. The effects of local and regional factors were high but climate dominated. P. aegeria was restricted to high quality patches and landscapes under adverse conditions. As both scenarios resulted in entirely different models, our study showed that the sensitivity of P. aegeria to local and landscape features might change, and alleged less important factors could turn into limiting factors. This stressed the importance of high quality landscape conditions at both local and regional scales even for species that appear to be relatively tolerant, and sounds a note of caution when predicting population response for management purposes based on just a single (or a few) year(s) of observation.

The 25 test sites (LTS) were mapped according to a common protocol. LTSs are 5X5km. To set up the protocol we run test of interpretation of the various types of greenveining from air photo. This gives a first product on method to map greenveining. Mapping was done from air photos, then field checked to ground truth the maps. The various landscape elements are characterized in two manners: herbaceous or woody greenveining vs. non greenveining and with the EUNIS habitat codes. All maps are stored on Arc-View ® in a vector and a raster format, the latter is restricted to greenveining information. Rasterization for done at a very fine scale to insure that linear elements (road and road verges) are distinguishable. Landscape analysis were performed within a 4X4 km, the remaining 1km around permit to assess the surrounding of every point with the 4X4 square, avoiding edge effects. Two types of analysis were done. The first one using landscape metrics as found in the Fragstat ® software; those metrics (mean distance between patches etc.) are at the LTS scale. The second set of measure is with LTSs analysis of connectivity for species having different requirement both in terms of range of movement and amount of greenvening within this range. Both sets of measures permit to make structural difference between LTSs.

In each of the seven GREENVEINS partner countries, three to four homogeneous agricultural landscapes (LTS) were selected for study. The study used altogether 25 landscape test sites of 16 km2 (LTS), with three or four LTSs being located in each of seven countries (France, Belgium, the Netherlands, Germany, Switzerland, the Czech Republic and Estonia). Together they covered wide ranges in land-use intensity and in the proportion of natural and semi-natural habitat types The biodiversity of each LTS was assessed by recording the presence and abundance of species in plants, birds and arthropods for the whole 16km2. Plants were surveyed by growth-form strata herbaceous layer, shrubs and trees. The arthropods studied were Apoidea (bees), Heteroptera (bugs), Carabidae (carabid beetles), Syrphidae (hoverflies) and Araneae (spiders). According to Greenveins project field observations, the temperate Europe agricultural landscape fragment of 16km² consists in average of 258 herbaceous species, 32 shrub and 26 tree species. In these landscape units, in average 53 bird species nest or forage. In field edges with (semi-)natural habitats, in average 51 bee species, 58 species of bugs, 69 species of carabids, 29 hoverfly species and 85 species of spiders can be found. However, we also observed considerable variation in species numbers among sites. The pan-European range in species number was sometimes up to eight fold (e.g. Bees min=15 in Belgium-KAP and max=125 in Germany-FRI). Particularly large variation in the number of species between LTS was observed in the case of bees (St.Dev is 55% of average value), bugs (42%) and hoverflies (38%). In the case of of Carabids (14%), Birds (16%) and Spiders (17%), the species richness per landscape unit varied to the smallest extent. Diversity index of Simpson, which takes into account the abundance, evenness and dominance of species, reveals a bit different view onto diversity of various groups in agricultural landscape. The most evenly distributed species groups within test sites are trees and shrubs. Their average Simpson diversity value is almost twice as any other group observed. Frequently trees and shrubs are planted or human managed and this explains the high homogeneity within an agricultural landscape. From insects, hoverflies have the highest score of diversity, which can be expected because of their ability to fly, resulting in good dispersal in this range of scale. However, between site variation of hoverflies diversity is remarkable. The most satellite-species rich taxonomic groups are birds and carabids. If carabids are ground living and connected to certain landscape elements, then the low Simpson diversity of birds is really surprising. The low diversity of birds is common feature for all test sites in all countries. Observed variation in diversity indices (number of species and Simpson diversity index) shows strong correlations between groups. The first two PCA axes describe 45% of total variation. The geographical gradient is main global factor of biodiversity of European agricultural landscapes, showing correlations with the first and second PCA axis. The land-use intensity index (LUI) is also correlated to them, but perpendicularly to geographical effects. We also found that there are strong between country differences in large-scale diversity, specific for each country. It means, even if the variation of land-use intensity and landscape structure are kept at maximum in selection of sites variation, the general biogeographical effect overwhelms anthropogenic effects and therefore in further analyses, the country effect always should be taken into account.

Thirteen novel polymorphic microsatellite loci were developed for Geum urbanum (Rosaceae). The microsatellites will be useful tools to analyse the influence of landscape structure and land-use intensity in agricultural landscapes on genetic diversity within and among populations of Geum urbanum. Transferability was tested in 19 other Geum species and in two Waldsteinia species. In most species PCR products of the expected range were obtained, therefore the markers reported here appear to be applicable across the whole genus. The markers have been used for studying Geum urbanum populations in LTSs (Landscape Test Sites) in the project. In addition, they have been published: Arens P, W Durka, JH Wernke-Lenting, MJM Smulders (2004) Isolation and characterisation of microsatellite loci in Geum urbanum (Rosaceae) and their transferability within the genus Geum. Mol Ecol Notes 4: 209-212

All biodiversity data collected during the greenveins project, as well as all landscape and landuse date, has been stored in the computer program FieldMap (IFER). This program contains a relational database and the information of the three components (biodiversity, landscape and land use) can be linked with each other. The database can be used for further in-depth analysis and research as well as the foundation for new research projects. The database is open to the greenveins-members only for five years. After this period the database will be open to the public to use.

The project aimed to explore the relationship between biodiversity, landscape structure and land use intensity in the agricultural landscape. The primary objective of the project was to define vulnerability of biodiversity in this landscape type: at what combinations of landscape structure and land use intensity does it occur. A threshold of Vulnerability is indicated if slight changes e.g. in landscape structure cause significant changes at certain levels of biodiversity. The Landscape Test Sites (LTS) were selected according to two main criteria: they should represent a gradient in the amount of natural and semi-natural habitats (green veining structures) and in the intensity of land use (e.g. input of fertilizer, pesticides). Within each of the 7 participating countries 3-4 LTS (4x4 km) were selected. The LTS were situated in France, Netherlands, Belgium, Germany, Switzerland, Czech Republic and Estonia. Thus they covered a geographical gradient too. For the survey of the species and species group levels, a set of species and/or species groups was selected that was representative for - The proportion of the regional species pool present in local species pool and - The presence of functional or taxonomic groups (so must cover as many ecosystems, trophic levels and dispersal range scales as physically possible). Furthermore, the selection of species groups depended on additional criteria like availability of specialists for determination; state of taxonomic knowledge in each country and the selected groups should be trapped using a minimum of different methods. Selected groups: vascular plants (primary producers), Araneae (spiders: predators), Carabidae (carabid beetles: mostly predators), Apoidea (bees: pollen and nectar feeders), Heteroptera (true bugs: phytophages and predators), Syrphidae (hoverflies: pollen and nectar feeders as adults) and as most mobile group birds. In order to investigate which species may be missed in the actually registered species lists of each LTS we compiled regional and local species pools of our target groups. The methods applied for the different species groups are presented in detail in a special report containing all protocols, which will be available from www.greenveins.nl. Plants were recorded on up to 250 randomly chosen plots per LTS. Arthropods were trapped by pitfall traps or combined flight traps (a combination of window and yellow water pan traps). Birds were surveyed three times on each LTS in April, May and June by visiting 20 sampling points per LTS.

In a large-scale study we investigated how biodiversity is influenced by landscape structure and the intensity of agricultural production. Species richness was assessed for selected plant and animal taxa in 25 agricultural landscapes distributed across seven European countries. The landscape structure was mapped using aerial photographs and the landuse intensity assessed by interviewing the farmers of the selected sites. The collected data was ued to describe the relationship between landscape structure, land use intensity and biodiversity. In general, species diversity increased with increasing area of semi-natural landscape elements and decreased with increasing land-use intensity. The existence of such general relationships demonstrates that characteristics of the agricultural landscape can be used as reliable indicators of species richness. These indicators provide a convenient method for large scale monitoring of biodiversity in agricultural landscapes, which is a prerequisite for better environmental policy and management. The results of this analysis have been written up in a scientific publication and will be published soon. Further publications on more detailed aspects are currently being written and will be published in scientific journals. One study on the athropods only has been accepted for publication.

In the metapopulation part of GREENVEINS, the intention was to survey the effect of within LTS spatial structure on a number of species in detail and try to derive general trends in the relationship between spatial habitat structure and the spatial scale of population structure of species. Species selected for the analysis were the insects Dark Bush Cricket (Pholidoptera griseoaptera) Ringlet (Aphantopus hyperanthus), Orange Tip (Anthocharis cardamines) and Speckled Wood (Pararge aegeria), the plants Geum urbanum, Silene dioica and Teucrium scorodonia and the amphibians Tree Frog (Hyla arboraea) and Crested Newt (Triturus cristatus). Not all species were found to be present in enough LTSes to allow proper analysis. For Pararge aegeria, Aphantopus hyperanthus and Geum urbanum enough observations were present to run a complete analysis including point observations within landscape test sites, using presence/absence on spots as dependent variable and spot-specific independent variables next to variables known only at test site level. To this data a linear mixed model was fitted. For all three species, habitat cohesion ( measured as the exponentially distance weighted habitat amount in a 250m circle around the observation points) was a significant factor in explaining presence. For Aphantopus hyperanthus significance was lowest, and it was the only factor in the model. For Geum the LTS level amount of herbaceous GV was also a significantly positive factor, while the presence of Pararge aegeria was negatively affected by insecticide. Single species surveys and analysis: For the crested newt, an analysis of the relationship between terrestrial habitat, pond density and presence of the species in reproduction ponds was carried out, using existing data gathered for an area just outside one of the Dutch LTS. The resulting model was used to try to link the presence of the species in the three LTSes where it was found to the combination of model results and LUI. The same was done for the Tree frog, using an existing model. Results were so far inconclusive. For Pholidoptera griseoaptera, a Capture-Mark-Recapture study was carried out at four locations within an LTS in Switzerland. The locations differed in LUI and LS. The study revealed that movement patterns were influenced by the structure of both habitat and the matrixx of unsuitable habitat in between. P. griseoaptera was able to cover unexpectedly large distances, especially in landscapes with a large share of unsuitable habitats which they were found to cross very fast. The hypothesis that the observed rates of dispersal result in frequent inter-patch movement is supported by a genetic survey on P. griseoaptera in the same area (Diekötter et al., abstract xx). At a local scale, with 2.4 km separating the most distant sampling locations, no significant population structuring was observed. Yet, a different pattern was found in an agricultural landscape showing a different landscape structure, more recent fragmentation of woody elements, and higher land-use intensity. Overall genetic diversity was lower and local populations of P. griseoaptera were genetically different from each other, possibly indicating a lack of inter-patch dispersal (Diekötter et al., abstract xx). For Pararge aegeria, the available data was split up in two scenarios, - The High abundance scenario was characterized by beneficial environmental and weather conditions coinciding with high local abundance of P. aegeria. - The Low abundance scenario reflected environmental stress and adverse weather conditions during larval development coinciding with low local abundance. In the high abundance scenario, P. aegeria was predicted to occur nearly anywhere under beneficial conditions. In the low abundance scenario P. aegeria was restricted to high quality patches and landscapes. This showed that the sensitivity of P. aegeria to local and landscape features might change under adverse conditions, where alleged less important factors could turn into limiting ones. This result is in line with the interaction between LUI stress factors and LS that was found in species our species richness and community similarity analysis. It stresses the importance of high quality landscape conditions even for species that appear to be relatively tolerant, and sounds a note of caution when predicting population response for management purposes based on just a single (or a few) year(s) of observation.

Deliverables

Share this page

Download