Dr. Ulrich Graf

Wiederholungen von Vegetationserhebungen sind wichtig, um zeitliche Veränderungen der biologischen Vielfalt und der Umweltbedingungen zu erfassen. Vegetationserhebungen sind jedoch anfällig für verschiedene Fehlerquellen. Diese gilt es zu minimieren und alternative Ansätze für die Analyse der Daten zu finden. Anhand eines Datensatzes von 224 doppelt erhobenen Dauerflächen in der Schweiz untersuchten wir die häufigste Fehlerquelle bei Vegetationserhebungen: die Unterschiede zwischen den Bearbeitenden. Unsere Ergebnisse zeigen, dass Artenlisten von Untersuchungsflächen, die am selben Tag von verschiedenen Bearbeitenden erstellt wurden, voneinander abweichen (28,5 % Unterschied) − hauptsächlich weil Arten mit geringer Häufigkeit von jeweils einem der beiden Bearbeitenden übersehen wurden. Wir konnten zeigen, dass sich der Fehler durch Aggregierungen bestimmter Einträge in den ursprünglichen Artenlisten (z. B. Unterarten auf Artniveau zusammenfassen oder Arten zu Aggregaten zuordnen) reduzieren lässt und dass mittlere ökologische Zeigerwerte robust gegenüber Unterschieden zwischen Bearbeitenden sind. Mittels der genannten Aggregierungen in den Artenlisten und durch Verwendung mittlerer Zeigerwerte können somit zeitliche Veränderungen der Vegetation und von Umweltbedingungen verlässlich erfasst werden.

Resurveys of vegetation plots are important for capturing temporal changes in biodiversity and environmental conditions. However, vegetation
surveys are prone to several error sources. It is important to minimise these and to find alternative approaches for analysing the data. Using a
dataset of 224 double-surveyed permanent plots in Switzerland, we investigated the most common source of error in vegetation surveys, namely
differences between observers. Our results show that species lists sampled by two different observers independently on the same day differ from
each other (28.5 % difference), mainly because species with low abundance were overlooked by one or the other of the two observers. We showed
that by applying simple aggregation steps to the original species lists, the error can be reduced (e. g. setting subspecies to species level or assigning species to aggregates) and that mean ecological indicator values are robust to differences among observers. Thus, by aggregating certain entries in the original species lists and using mean indicator values, temporal changes in vegetation and environmental conditions can be recorded reliably.

Aims: To develop a consistent ecological indicator value system for Europe for five of the main plant niche dimensions: soil moisture (M), soil nitrogen (N), soil reaction (R), light (L) and temperature (T). Study area: Europe (and closely adjacent regions). Methods: We identified 31 indicator value systems for vascular plants in Europe that contained assessments on at least one of the five aforementioned niche dimensions. We rescaled the indicator values of each dimension to a continuous scale, in which 0 represents the minimum and 10 the maximum value present in Europe. Taxon names were harmonised to the Euro+Med Plantbase. For each of the five dimensions, we calculated European values for niche position and niche width by combining the values from the individual EIV systems. Using T values as an example, we externally validated our European indicator values against the median of bioclimatic conditions for global occurrence data of the taxa. Results: In total, we derived European indicator values of niche position and niche width for 14,835 taxa (14,714 for M, 13,748 for N, 14,254 for R, 14,054 for L, 14,496 for T). Relating the obtained values for temperature niche position to the bioclimatic data of species yielded a higher correlation than any of the original EIV systems (r = 0.859). The database: The newly developed Ecological Indicator Values for Europe (EIVE) 1.0, together with all source systems, is available in a flexible, harmonised open access database. Conclusions: EIVE is the most comprehensive ecological indicator value system for European vascular plants to date. The uniform interval scales for niche position and niche width provide new possibilities for ecological and macroecological analyses of vegetation patterns. The developed workflow and documentation will facilitate the future release of updated and expanded versions of EIVE, which may for example include the addition of further taxonomic groups, additional niche dimensions, external validation or regionalisation.

Aims: Ellenberg-type indicator values are expert-based rankings of plant species according to their ecological optima on main environmental gradients. Here we extend the indicator-value system proposed by Heinz Ellenberg and co-authors for Central Europe by incorporating other systems of Ellenberg-type indicator values (i.e., those using scales compatible with Ellenberg values) developed for other European regions. Our aim is to create a harmonized data set of Ellenberg-type indicator values applicable at the European scale.
Methods: We collected European data sets of indicator values for vascular plants and selected 13 data sets that used the nine-, ten-or twelve-degree scales defined by Ellenberg for light, temperature, moisture, reaction, nutrients and salinity. We compared these values with the original Ellenberg values and used those that showed consistent trends in regression slope and coefficient of determination. We calculated the average value for each combination of species and indicator values from these data sets. Based on species' co-occurrences in European vegetation plots, we also calculated new values for species that were not assigned an indicator value.
Results: We provide a new data set of Ellenberg-type indicator values for 8908 European vascular plant species (8168 for light, 7400 for temperature, 8030 for moisture, 7282 for reaction, 7193 for nutrients, and 7507 for salinity), of which 398 species have been newly assigned to at least one indicator value.
Conclusions: The newly introduced indicator values are compatible with the original Ellenberg values. They can be used for large-scale studies of the European flora and vegetation or for gap-filling in regional data sets. The European indicator values and the original and taxonomically harmonized regional data sets of Ellenberg-type indicator values are available in the Supporting Information and the Zenodo repository.

Aims: Resurveys of vegetation plots are prone to several errors that can result in misleading conclusions. Minimizing such errors and finding alternative approaches for analyzing resurvey data are therefore important. We focused on inter-observer error and excluded other sources of variation. Our main questions were: How large is the inter-observer error (i.e. pseudoturnover) in vegetation surveys, and can it be reduced by simple data aggregation approaches? Which factors are affecting pseudoturnover and does it vary between morphological species groups or change over time? Is ecological inference robust against inter-observer differences?
Location: Switzerland.
Methods: Over seven years, we double-surveyed a total of 224 plots that were marked once in the field and then sampled by two observers independently on the same day. Both observers conducted full vegetation surveys, recording all vascular plant species, their cover, and additional plot information. We then calculated mean ecological indicator values and pseudoturnover.
Results: Average pseudoturnover was 29% when raw species lists were compared. However, by applying simple aggregation steps to the species list, pseudoturnover was reduced to 17%. Pseudoturnover further varied among habitat types and declined over the years, indicating a training effect among observers. Most overlooked taxa, responsible for pseudoturnover, had low cover values. Mean ecological indicator values were robust against inter-observer differences.
Conclusions: To minimize pseudoturnover, we suggest continuous training of observers and species-list aggregation prior to analysis. As mean ecological indicator values were robust against inter-observer differences, we conclude that they can provide a reliable estimate of temporal vegetation and ecological changes.

A typical degraded mountain bog in the Swiss Alps was restored by cessation of grazing and partial rewetting. We observed the development of the vegetation over 15 years, starting immediately after the grazing cessation (4 years before rewetting). Cessation of grazing stimulated the recolonisation of bare peat and the establishment of bryophytes. While many typical bog forest species benefited from grazing cessation, the relative abundance of Cyperaceae decreased. Rewetting (in our case: closing ditches with sheet piles and filling them with sawdust) as an additional restoration measure, had a stabilising effect on most environmental factors (e.g. water availability, soil aeration). Furthermore, rewetting had a stabilising effect on stress tolerators and promoted typical species of bog hummocks and hollows. The abundances of species of nearly all groups changed far less dynamically in the rewetted area than in the area that remained drained. It turned out that cessation of grazing as the only restoration measure was insufficient to maintain the vegetation of bogs, whereas the combination of grazing cessation and rewetting directed the site towards the desired development. However, even eleven years after rewetting and 15 years after cessation of grazing, the site had not yet reached a state that was close to an undisturbed alpine bog.

Aims: Classification of European bog vegetation (Oxycocco-Sphagnetea class); identification of diagnostic species for the class and vegetation subgroups (orders and alliances); development of an expert system for automatic classification of vegetation plots; and production of distribution maps of the Oxycocco-Sphagnetea class and its alliances.
Location: Europe.
Methods: A data set of vegetation-plot records was compiled to include various bog types over most of the European continent. An unsupervised classification (beta-flexible linkage method, Sørensen distance measure) and detrended correspondence analysis (DCA) ordination were applied. Formal definitions of syntaxa based on species presence and covers, and respecting the results of the unsupervised classification, were developed and included in a classification expert system.
Results: The Oxycocco-Sphagnetea class, its two orders (Sphagno-Ericetalia tetralicis and Sphagnetalia medii) and seven compositionally distinct alliances were formally defined. In addition to the syntaxa included in EuroVegChecklist, three new alliances were distinguished: Rubo chamaemori-Dicranion elongati (subarctic polygon and palsa mires); Erico mackaianae-Sphagnion papillosi (blanket bogs of the northwestern Iberian Peninsula); and Sphagno baltici-Trichophorion cespitosi (boreal bog lawns). The latter alliance is newly described in this article.
Conclusions: This first pan-European formalized classification of European bog vegetation partially followed the system presented in EuroVegChecklist, but suggested three additional alliances. One covers palsa and polygon mires, one covers Iberian bogs with endemics and one fills the syntaxonomical gap for lawn microhabitats in boreal bogs. A classification expert system has been developed, which allows assignment of vegetation plots to the types described.

1. Evaluation of restoration activities is indispensable to assess the extent to which targets have been reached. Usually, the main goal of ecological restoration is to restore biodiversity and ecosystem functioning, but validation is often based on a single indicator, which may or may not cope with whole-ecosystem dynamics. Network analyses are, however, powerful tools, allowing to examine both the recovery of various biotic and abiotic properties and the integrated response at community and ecosystem level.
2. We used restoration sites where topsoil was removed from former intensively managed grassland and seeds were added. These sites were between 3 and 32 years old. We assessed how plants, soil biota, soil properties and correlation-based interactions between biotic communities and their abiotic environment developed over time and compared the results with (i) intensively managed (not restored), and (ii) well-preserved targeted semi-natural grasslands.
3. Plant, nematode, fungal and prokaryotic diversity and community structures of the restored grasslands revealed clear successional patterns and followed similar trajectories towards targeted semi-natural grasslands. All biotic communities reached targeted diversity levels no later than 18 years post-restoration.
4. Ecological networks of intensively managed and short-term (~4 years) restored grasslands were less tightly connected compared to those found in mid- and long-term (~18–30 years) restored and target grasslands. Restoration specifically enhanced interactions among biotic communities, but reduced interactions between biotic communities and their abiotic environment as well as interactions among abiotic properties in the short- and mid-term.
5. Synthesis and applications: Overall, our study demonstrated that topsoil removal and seed addition were successful in restoring diverse, tightly coupled and well-connected biotic communities above- and below-ground similar to those found in the semi-natural grasslands that were restoration targets. Network analyses proved to be powerful in examining the long-term re-establishment of functionally connected biotic communities in restored ecosystems. Thus, we provide an approach to holistically assess restoration activities by notably considering the complexity of ecosystems, much in contrast to most traditional approaches.

It is generally assumed that restoring biodiversity will enhance diversity and ecosystem functioning. However, to date, it has rarely been evaluated whether and how restoration efforts manage to rebuild biodiversity and multiple ecosystem functions (ecosystem multifunctionality) simultaneously. Here, we quantified how three restoration methods of increasing intervention intensity (harvest only < topsoil removal < topsoil removal + propagule addition) affected grassland ecosystem multifunctionality 22 yr after the restoration event. We compared restored with intensively managed and targeted seminatural grasslands based on 13 biotic and abiotic, above- and belowground properties. We found that all three restoration methods improved ecosystem multifunctionality compared to intensively managed grasslands and developed toward the targeted seminatural grasslands. However, whereas higher levels of intervention intensity reached ecosystem multifunctionality of targeted seminatural grasslands after 22 yr, lower intervention missed this target. Moreover, we found that topsoil removal with and without seed addition accelerated the recovery of biotic and aboveground properties, and we found no negative long-term effects on abiotic or belowground properties despite removing the top layer of the soil. We also evaluated which ecosystem properties were the best indicators for restoration success in terms of accuracy and cost efficiency. Overall, we demonstrated that low-cost measures explained relatively more variation of ecosystem multifunctionality compared to high-cost measures. Plant species richness was the most accurate individual property in describing ecosystem multifunctionality, as it accounted for 54% of ecosystem multifunctionality at only 4% of the costs of our comprehensive multifunctionality approach. Plant species richness is the property that typically is used in restoration monitoring by conservation agencies. Vegetation structure, soil carbon storage and water-holding capacity together explained 70% of ecosystem multifunctionality at only twice the costs (8%) of plant species richness, which is, in our opinion, worth considering in future restoration monitoring projects. Hence, our findings provide a guideline for land managers how they could obtain an accurate estimate of aboveground-belowground ecosystem multifunctionality and restoration success in a highly cost-efficient way.

Die Wirkungskontrolle Biotopschutz Schweiz (WBS) wurde 2011 vom Bundesamt für Umwelt BAFU und der Eidgenössischen Forschungsanstalt WSL initiiert, um Veränderungen in den Biotopen von nationaler Bedeutung – Kerngebiete der Ökologischen Infrastruktur – zu untersuchen. Die Ersterhebung der Wirkungskontrolle Biotopschutz Schweiz endete 2017, die ersten Resultate zu Zustand und Veränderung der Biotope von nationaler Bedeutung liegen nun vor.

En 2011, l'Office fédéral de l'environnement (OFEV) et l'Institut fédéral de recherches sur la forêt, la neige et le paysage (WSL) ont donné le coup d'envoi au Suivi des effets de la protection des biotopes en Suisse dans le but d'étudier l'évolution des biotopes d'importance nationale, aires centrales de l'infrastructure écologique. Le premier cycle de relevés s'est achevé en 2017 et permet de livrer aujourd’hui des conclusions sur l'état des biotopes d’importance nationale et les transformations qui ont marqué ces derniers.

Die Wirkungskontrolle Biotopschutz Schweiz untersucht, ob Biotope von nationaler Bedeutung in ihrer Qualität erhalten bleiben. Trotz Schutzbemühungen verschlechterte sich jedoch der Zustand des artenreichen Grünlands in den vergangenen Jahren. Nutzungsaufgabe und zunehmende Verbuschung werden als Ursache vermutet. Besonders ausgeprägt ist dieser Trend in hohen Lagen. Abhilfe kann ein Früherkennungssystem leisten. Eine Auswertung bestätigt einmal mehr, dass auch modernste GPS-Systeme die permanente Markierung von Untersuchungsflächen nicht ersetzen können.

Aim: The EUNIS Habitat Classification is a widely used reference framework for European habitat types (habitats), but it lacks formal definitions of individual habitats that would enable their unequivocal identification. Our goal was to develop a tool for assigning vegetation-plot records to the habitats of the EUNIS system, use it to classify a European vegetation-plot database, and compile statistically-derived characteristic species combinations and distribution maps for these habitats.
Location: Europe.
Methods: We developed the classification expert system EUNIS-ESy, which contains definitions of individual EUNIS habitats based on their species composition and geographic location. Each habitat was formally defined as a formula in a computer language combining algebraic and set-theoretic concepts with formal logical operators. We applied this expert system to classify 1,261,373 vegetation plots from the European Vegetation Archive (EVA) and other databases. Then we determined diagnostic, constant and dominant species for each habitat by calculating species-to-habitat fidelity and constancy (occurrence frequency) in the classified data set. Finally, we mapped the plot locations for each habitat.
Results: Formal definitions were developed for 199 habitats at Level 3 of the EUNIS hierarchy, including 25 coastal, 18 wetland, 55 grassland, 43 shrubland, 46 forest and 12 man-made habitats. The expert system classified 1,125,121 vegetation plots to these habitat groups and 73,188 to other habitats, while 63,064 plots remained unclassified or were classified to more than one habitat. Data on each habitat were summarized in factsheets containing habitat description, distribution map, corresponding syntaxa and characteristic species combination.
Conclusions: EUNIS habitats were characterized for the first time in terms of their species composition and distribution, based on a classification of a European database of vegetation plots using the newly developed electronic expert system EUNIS-ESy. The data provided and the expert system have considerable potential for future use in European nature conservation planning, monitoring and assessment.

Die Wirkungskontrolle Biotopschutz Schweiz (WBS) ist ein seit 2011 laufendes Programm des Bundesamtes für Umwelt BAFU und der Eidgenössischen Forschungsanstalt WSL. Sie untersucht mittels Luftbildinterpretation sowie floristischen und faunistischen Felderhebungen Veränderungen in den Biotopen von nationaler Bedeutung, den wichtigsten Knoten der Ökologischen Infrastruktur: Trockenwiesen und -weiden, Hoch- und Flachmoore, Auen (inklusive alpine Schwemmebenen und Gletschervorfelder) und Amphibienlaichgebiete. Erste Ergebnisse zeigen für die Biotope von nationaler Bedeutung aufgrund der berechneten Indikatoren meist negative Entwicklungen auf. Die Hochmoore wurden in den letzten 20 Jahren nährstoffreicher und trockener. Die Flachmoore sind ebenfalls trockener geworden, die Deckung durch Gehölze nahm zu und der Anteil an typischen Moorarten ging zurück. Weiterhin aktive Entwässerungen in Hoch- und Flachmooren und bei den Flachmooren zusätzlich die Nutzungsaufgabe erscheinen als grösste Probleme. In den Trockenwiesen und -weiden ist ebenfalls eine Zunahme der Gehölzdeckung, insbesondere auf der Alpensüdflanke, zu verzeichnen. Zudem zeigt die Vegetation nun nährstoffreichere, feuchtere, dichtere und schattigere Verhältnisse als bei früheren Untersuchungen an. Die nationalen Amphibienlaichgebiete haben im Durchschnitt mindestens eine Amphibienart pro Gebiet verloren. Kreuzkröte und Geburtshelferkröte nahmen weiterhin stark ab. Auch wenn die negativen Entwicklungen in den Biotopen von nationaler Bedeutung überwiegen, so gibt es doch regional oder in einzelnen Biotopen auch positive Entwicklungen. In den Hochmooren des Mittellandes nahm zum Beispiel die Gehölzdeckung ab und in den Amphibienlaichgebieten schwächten sich die Verluste der häufigen Amphibienarten (Erdkröte, Bergmolch, Grasfrosch) durch Neuentdeckungen ab oder wurden sogar ganz aufgehoben. Aber auch seltene Amphibienarten zeigten teilweise eine Stabilisierung der Bestände. Diese positiven Entwicklungen sind wahrscheinlich auf Renaturierungs- und Naturschutzmassnahmen der Kantone und weiterer Akteure sowie die angepasste Bewirtschaftung der Objekte zurückzuführen. Dies zeigt, dass diese Massnahmen wirksam sind. Es ist somit wichtig, dass diese fortgesetzt und mitunter verstärkt angewendet werden, um die Biodiversität in den Biotopen von nationaler Bedeutung, den eigentlichen Kronjuwelen des Schweizer Naturschutzes, langfristig zu erhalten. Mit den laufend erhobenen Daten im Rahmen der WBS wird es in Zukunft möglich sein, über die weitere Entwicklung der Biotope von nationaler Bedeutung Auskunft zu geben und den Erfolg der Massnahmen zu beurteilen.

Resurveys of historical plots are important to study vegetation changes over time. However, there is a high potential of relocation error, as historical plots were usually not permanently marked, which might result in misleading interpretations of changes. Quasi-permanent plots can relatively well be relocated, because of available information on their position (e.g. GPS coordinates). This minimizes relocation error. Quasi-permanent plots have thus been suggested to be suitable for vegetation resur-veys. Land-use abandonment and intensification are the major drivers of habitat degradation and there-by cause biodiversity loss. To counteract the ongoing decline of biodiversity in Swiss grasslands, the dry grasslands of national importance, comprising a total of 3631 sites across Switzerland (0.5% of its total area) were legally protected in 2010. These grassland sites represent various vegetation types and cover a wide elevational gradient. From 1995 to 2006, several thousand quasi-permanent vegetation plots of 28 m² were surveyed in the nationally important dry grassland sites. For each plot, GPS coordi-nates and a metric measure of post-processed GPS inaccuracy were recorded. In a set of 384 sites, 538 of the historical plots were resurveyed between 2011 and 2017. We tested the effects of GPS inac-curacy and elevation on temporal changes of species richness, species turnover and mean ecological indicator values. Overall, mean indicator values for light and stress significantly decreased, and those for moisture, nutrients and ruderality significantly increased within only one decade. However, the temporal change of the mean indicator values for nutrients and moisture was more pronounced at higher than at lower elevation, indicating a particularly strong decrease of habitat quality at high elevation sites. While changes of in mean indicator values were largely robust to relocation error, the temporal turnover of key species significantly increased with increasing GPS inaccuracy. These results indicate that quasi-permanent plots are suitable to study changes in mean indicator values, but probably not to compare species composition and turnover across time.

In Switzerland, steppe vegetation is restricted to inner alpine dry valleys in the West (Valais) and the East (Grisons). They harbour many specialist and threatened species. In 2011, a long-term program was established to monitor changes in Swiss habitats of national importance. Here, we studied patterns of vascular plant diversity in a subset of 148 steppe plots of 10 m² from this program. Plots were either grazed or abandoned. On average, we found 23.3 species per plot. The proportion of steppe specialists and of threatened species out of the total species richness was 29.3% and 12.2%, respectively. The total number of species increased with elevation and peaked at intermediate productivity levels. In contrast, the proportion of threatened and steppe specialist species was generally negatively related to higher productivity and wood cover – two factors considered as indicators of habitat degradation. Abandonment vs. pasture had no effects on total species richness. Productivity was positively related to ecological indicator values for moisture, competition and ruderality, as well as the community mean traits of plant height and specific leaf area, but negatively with the indicator value for light. Similarly, wood cover was positively associated with higher ecological indicator values for productivity, moisture and competition, as well as the community mean trait of plant height, but negatively with the indicator value for light and ruderality. The negative response of threatened and steppe specialist vascular plant species to habitat degrading factors suggests their proportional richness to be a better indicator of habitat quality than the total species richness. We conclude that management regimes which guarantee low productivity and low wood cover and avoid intensification by fertilization and irrigation are essential for the conservation of steppe habitats. Beside periodic clearing of woody species, low-intensity grazing might prevent shrub encroachment.

Wie ist der Zustand der Biotope von nationaler Bedeutung? Wie entwickeln sie sich über die Zeit? Hat sich ihre Qualität verändert? Diese Fragen beantwortet die Wirkungskontrolle Biotopschutz Schweiz. Vegetationsaufnahmen und Luftbildanalysen sind dabei eine wichtige Grundlage, um Veränderungen aufzuzeigen.