BioAssess

Background to the BioAssess Project

•  The BioAssess project was the first project to use standardised protocols to measure several major elements of biodiversity across Europe (in eight countries and six biogeographical regions) to simultaneously develop methods for assessing biodiversity, or "biodiversity assessment tools", and to quantify the impact on biodiversity of land use change, a major driver of change in biodiversity in Europe and elsewhere.

•  For the list of BioAssess partners, click here.

•  The main purpose of the BioAssess project was to develop biodiversity assessment tools for inland terrestrial ecosystems, comprising sets of indicators of biodiversity, to assess the impact of policies on changes in biodiversity in Europe. "Biodiversity assessment tools" may be defined as a set of indicators, which provides information on status and trends in biodiversity for a range of stakeholders. This approach to monitoring acknowledges that a single measure of biodiversity is unlikely to satisfy most stakeholder needs, particularly those interested in trends in biodiversity at the European level.

•  The project responded to a perceived need for a better method of biodiversity assessment in the face of a global decline in biodiversity and the need to quantify both the impact of the policies instigated to address this decline and also the impact of other policies that might affect biodiversity. Many international, European and national policies and actions, notably the Convention on Biological Diversity, the European Biodiversity Strategy and an increasing number of Biodiversity Action Plans, operating at a range of scales, have been instigated to deal with this problem. However, the establishment of targets to reduce (Johannesburg Summit on Sustainable Development, 2002) or halt (EU strategy for Sustainable Development, Göteborg 2001) the loss of biodiversity have significantly added to the urgency of this task.

•  An electronic conference (http://www.gencat.es/mediamb/bioassess/) and workshop were held to consider the policy framework for the development of monitoring and indicators of biodiversity in Europe, the major biogeographical influences on biodiversity, the major key factors (or drivers) of biodiversity and to propose a provisional set of indicators for study in this project. Biological relevance was identified as being the most important criterion – biodiversity indicators must provide a measure of biological diversity. The following categories of indicator were identified: diversity measures of soil macrofauna, soil Collembola, ground beetles, butterflies, plants, lichens and birds, and landscape indices derived from remote sensing (earth observation).

1 Methods

Establishing land use transects across Europe

Land-use intensity transects were set up in eight countries (Table 1). In each country six 1km 2 sites (or "Land Use Units" (LUUs) to avoid confusion between sampling locations within each site) were established. These sites represented a land-use intensity transect of: old-growth or unmanaged forest, managed forest, forest / woodland-dominated landscape, mixed-use landscape, pasture-dominated landscape and arable crop-dominated landscape

Table 1: Biogeographic regions and countries involved in BioAssess .

Region Country Boreal Finland Atlantic UK (Scotland)   Ireland Continental France Pannonian Hungary Alpine Switzerland Mediterannean Spain   Portugal

2 Measuring biodiversity and evaluating indicators

2.1 Sites (Land-Use units)

The impact of land-use intensity on biodiversity was measured in the sites (land-use units) along each country's transect by assessing the diversity of soil Collembola, soil macrofauna, Ground beetles (Carabidae), plants, lichens, butterflies and birds. Protocols were developed for each group of plants and animals. These are summarised below. In most cases sampling of each group of plants and animals was done at the same locations on a 16-point grid in each site (Figure 1).

Figure 1 One of the BioAssess sampling sites, or Land Use Units (in Switzerland). Note the 16 sampling locations (marked red).

2.2 Sampling methods

Birds . Bird data were collected using point counts (Verner 1985). These were undertaken at each of the 16 sampling points per site. In each year a total of four different counts was made at each point, with each count lasting five minutes. The four counts were undertaken on separate occasions (visits) spread throughout the breeding season. In most countries this was in the months April, May and June, but was slightly earlier in the Mediterranean countries and slightly later in Finland. During each count, the observer recorded all birds seen or heard as long as they were deemed to be using the square (e.g. for nesting, displaying, hunting, roosting etc.). The locations of these were recorded into distance bands of 0-25 m, 25-50 m, 50-100 m and >100 m from the point. Singing birds were noted separately from birds not showing any such behaviour. All counts took place in the morning, starting as near to dawn as possible. No counts were undertaken during excessively wet or windy conditions. As far as possible, the counts for individual visits were undertaken at all 16 points within a site on the same day. Observers were asked to vary the order in which sample points were surveyed on different visits. Additional data were recorded whilst walking between the points to ensure that as large a species list as possible was compiled for each site.

Butterflies . On each of the six sites in the eight countries butterflies were counted weekly on a line-transect (Pollard and Yates 1993). The transects were divided into sections of a homogeneous vegetation structure. The total length of a transect was 1 km per km-square, which took the recorder 20-40 minutes dependent on field conditions, the season and the number of butterflies. Butterfly counts were made on a weekly basis in standardized weather conditions from April to September.

Lichens. Lichen surveys were carried out within a circular area of 1 ha (56.41 m radius) around the 16 sampling plots of each site. Within each sampling plot 12 collecting points were selected randomly and lichen relevés were carried out on three different substrates, namely soil, rocks and trees. All lichens which occurred inside a 50x40 cm frequency grid (mesh size 10 cm) of the relevés on soil or inside four frequency ladders (each 50x10 cm, mesh size 10 cm) of the relevés on rocks or relevés on trees were considered, unless smaller than 5 mm. Lichenicolous fungi and non-lichenised fungi were not included. For soil relevés, all lichens growing on the ground, on rotten wood, on shrubs and small trees (= 12 cm Ø, up to a height of 150 cm above ground), on pebbles and rocks (size <50x40 cm) or on any other substrate were considered. For rock relevés, starting from the centre of the collecting location, the nearest saxicolous object within the border of the sampling plot with a size larger than 50x40 cm was selected. Four frequency ladders were placed on the object in a way that the epilithic and terricolous (if the saxicolous object is partially or completely covered with soil) lichen species diversity were sampled as completely as possible. For tree relevés, sampling was done to ensure representative coverage of tree species with acidic bark and trees more or less neutral bark, and both small (young) trees and thick (old) trees. The four sectors of the compass were marked on the trunk, 150 cm above ground, and four sampling ladders were fixed on the trunk. In the case of cork oaks where the bark of the trunks was harvested in regular intervals, the ladders were set up on the main branches above the upper limit of the cork harvest.

Plants. At each of the 16 sampling points within each site, a rectangular plot of 100 m2 (20 x 5 m), inscribed within a circular plot of 25 m radius around the central point of the sampling point, was laid down and sampled. The rectangular plot comprised two sets of nested areas of 1, 5, 12.5, 25 and 50 m 2 that were sampled successively for species richness. Cover of each species, cover of the different vegetation layers and litter and stone cover were estimated within the 12.5 m 2 quadrats; tree crown cover was also estimated over the area of the rectangular plot. The diameter of living and dead trees was measured both in the rectangular and the circular plots. An exhaustive sampling of woody species richness, including low shrubs, was also done in the circular plots.

Carabidae . At each sampling plot, four pitfall traps (8 cm in diameter, 10.5 cm in depth) were placed 4-5 m apart in a regular 2 × 2 grid. The traps were partly filled with propylene or ethylene glycol, and a plastic roof was placed a few centimetres above the trap to prevent flooding from rain. In Spain, large stones were placed above the traps to minimize either flooding and damages from wild animals. Sampling was carried out from May to October, according to beetle activity.

Soil macrofauna comprise all the invertebrates groups visible to the naked eye. They contain a great variety of taxa, including Mollusca, Arthropoda and Annelida. The extraction of the soil macrofauna was performed once at each sampling location with an application of 1.5 l of 0.2% formalin solution followed by hand sorting of soil dug down to 15 cm. In some places where formalin application was not efficient (e.g. damp soils) or legal, soil was hand sorted down to 25cm.

Soil Collembola . At each sampling point Collembola were collected with a sample core (5cm diameter) including the organic horizon (when present) plus 5cm of the mineral soil. From each core sample, the organic and mineral fractions were separated into two samples. Collembola were extracted from the soil in the laboratory using Berlese or MacFadyen funnels.

Remote sensing . Landscape composition in each of the 1km 2 sites was assessed based on remote sensing data. Two types of satellite images were fused: Landsat ETM 7 images and IRS images. Landsat ETM 7 was chosen because of its good spectral resolution (seven channels) good availability in an archive database, affordability, and continuous processing and IRS because of good spatial resolution (5.8m) and low cost when compared to other high spatial resolution remote sensing data.

3 BioAssess- results summary

•  These data were also used to quantify the impact of biogeography, land use, and a range of environmental factors on biodiversity. Birds, butterflies, lichens, plants, ground beetles (Carabidae), soil macrofauna and springtails (Collembola) responded markedly to land use in terms of abundance, diversity and species composition. Some taxa responded in similar ways, others behaved in contrasting ways. The nature of the response of each taxon and the factors underlying their response are summarised below.

•  A total of 111 breeding bird species, 135 butterfly species ( 30,445 individuals ), 757 lichen species, 1467 plant species and subspecies, 301 ground beetle (Carabidae) species (152,866 individuals) , 908 soil macrofauna species ( 14,166 individuals) and 281 springtail (Collembola) species ( 47,774 individuals ) were recorded.

•  Each protocol used to assess biodiversity was satisfactorily applied in the range of biogeographical regions, countries, habitats and land uses sampled. Therefore, applied together, they would comprise a thorough “biodiversity assessment toolkit”, measuring four major components of above-ground biodiversity (birds, butterflies, lichens, plants) and three major components of soil and soil-surface dwelling biodiversity (ground beetles, soil macrofauna, springtails). Such an approach, applied in a network of sites across Europe would complement the assessment of biodiversity through the use of one or few indicators, such as an index of bird abundance.

•  Each of the potential indicators was evaluated by analysing their ability to predict other elements of biodiversity because such biological relevance had been identified as the most important criterion for a biodiversity indicator. Other factors are important in selecting indicators and although a single project cannot consider all the purposes that biodiversity indicators may be used for, the following general conclusions emerged from the BioAssess project:

•  Birds were found to be useful indicators of biodiversity; they significantly predicted the species richness of butterflies, lichens and plants. However, they were not found to be a good indicator of soil and soil-surface dwelling biodiversity. Birds are also suitable indicators of biodiversity for a number of other reasons including the ease with which they can be identified, the existence of ample ecological information and bird monitoring schemes and the fact that they are more threatened than most other taxa. For a more detailed summary, click here.

•  Butterflies were found to be useful indicators of biodiversity; they significantly predicted the species richness of birds, lichens and plants. However, they were not found to be a good indicator of soil and soil-surface dwelling biodiversity. Butterflies are also suitable indicators of biodiversity because they are relatively easy to identify, are more threatened than most other taxa and there are butterfly monitoring schemes, using well-tested protocols, in many countries. For a more detailed summary, click here.

•  Plants were found to be useful indicators of biodiversity; they significantly predicted the species richness of birds, butterflies, and lichens. However, they were not found to be a good indicator of soil and soil-surface dwelling biodiversity. Plants are also suitable indicators of biodiversity because they are relatively easy to survey and identify, as primary producers they play a critical role in supplying ecosystem goods and services, and because they are the single most important group of organisms in shaping the habitats and determining the physical environments for other species. For a more detailed summary, click here.

•  Lichens were found to be useful indicators of biodiversity; they significantly predicted the species richness of birds, butterflies and plants, although a poorer predictor of the richness of other groups of species than birds, butterflies and plants. They were not found to be a good indicator of soil and soil-surface dwelling biodiversity. Lichens are also suitable indicators of biodiversity because they are easy to survey and many species- the macrolichens- are relatively easy to identify. In addition, their particular sensitivity to a wide range of anthropogenic factors and the length of time they tend to take to recover from their impacts make them a unique taxon. Macrolichen diversity is a good predictor of total lichen diversity. For a more detailed summary, click here.

•  Macrofauna were found to be the most promising of the three groups of soil (or soil-surface) dwelling organisms as an indicator of the richness of other taxa, showing weak correlations with butterflies and carabids and stronger correlations with plants. However, only two of the many invertebrate groups that comprise soil macrofauna- soil Coleoptera and earthworms- were evaluated at species level, leaving the potential of this taxon least well understood in this project. A rapid assessment of soil macrofaunal could be done through combining measures of ant and earthworm diversity with macrofaunal family diversity. For a more detailed summary, click here.

•  Carabidae (ground beetles) were found to be a poor indicator of other elements of biodiversity, only showing a weak correlation with soil macrofauna. Carabids are, however, potentially useful indicators of biodiversity because they are a very easy group of invertebrates to survey and are relatively easy to identify. For a more detailed summary, click here.

•  Soil Collembola were found to be a poor indicator of other elements of biodiversity, only showing a weak correlation with lichens. Collembola are, however, potentially useful indicators of biodiversity because they are an easy group of soil invertebrates to survey. It is also possible to compare samples at a higher taxonomic level (genus) thus decreasing identification costs. For a more detailed summary, click here.

•  Of the landscape indices derived from remote sensing , several indices were shown to be potentially useful indicators of the richness of single taxa and although no single index was correlated with the diversity of all components of biodiversity studied, a few indices correlated with more than one taxon. Total core area or disjunct core area density correlated with the richness of lichens, butterflies and ground beetles. Patch richness correlated with the richness of birds and Collembola. Landscape evenness correlated with the richness of birds and butterflies. Intensive study of one set of sites (Switzerland) provided more information on the potential use of landscape indices derived from remote sensing and is summarised below. For a more detailed summary, click here.

•  Despite these promising results, the project identified some of the crucial limits within which these potential indicators should be used. The presence of these limits implies that indicators of biodiversity must be cautiously applied. A network of biodiversity observation sites where detailed information on biodiversity was regularly collected would, amongst other things, permit the testing of indicators and, therefore, greatly increase the confidence in their application.

For complete project reports, click here.