Oak-rich Temperate Forests

Transkript

1 Göteborg University Faculty of Science 2008 Oak-rich Temperate Forests Conservation Ecology of Cryptogams and Vascular Plants at Local and Landscape Level Heidi Paltto Dissertation Akademisk avhandling för filosofie doktorsexamen i Växtekologi som enligt Naturvetenskapliga fakultetets beslut kommer att offentligen försvaras fredagen den 29 februari 2008, kl i Föreläsningssalen, Institutionen för Växt- och Miljövetenskaper, Carl Skottsbergs Gata 22 B, Göteborg. Examinator: Ulf Molau. Fakultetsopponent: Magne Sætersdal, Norsk Institutt for Skog og Landskap, Fana, Norge. 1

2 Paltto. H Oak-rich Temperate Forests: Conservation Ecology of Cryptogams and Vascular Plants at Local and Landscape Level. Doctoral thesis. Department of Plant- and Environmental Sciences, Göteborg University, Göteborg, Sweden. Heidi Paltto Department of Plant- and Environmental Sciences Göteborg University Box 461 SE Göteborg Sweden Cover photographs: The lichen Lobaria pulmonaria (upper left), the fungi Plicaturopsis crispa (upper right) and the vascular plant Lathyrus niger (lower left) are taken by Leif Stridvall and the bryophyte Porella platyphylla is photographed by Tomas Hallingbäck. ISBN: Copyright Heidi Paltto 2008 Printed by Chalmers Reproservice 2008 Published papers in the thesis are reprinted by the kind permission of Elsevier: Götmark, F., Paltto, H., Nordén, B. & Götmark, E Evaluation of partial cutting in broadleaved temperate forest under strong experimental control: Short-term effects on herbaceous plants. Forest Ecology and Management 214: Paltto, H., Nordén, B., Götmark, F. & Franc, N At which spatial and temporal scales does landscape context affect local density of Red Data Book and Indicator species? Biological Conservation 133: Nordén, B., Paltto, H., Götmark, F. & Wallin, K Indicators of biodiversity, what do they indicate? Lessons for conservation of cryptogams in oak-rich forest. Biological Conservation 135:

3 Table of Contents List of papers... 5 Abstract... 6 Sammanfattning Introduction Methods Species groups and their responses to habitat changes at different spatio-temporal scales Effects of partial cutting/biofuel harvest for different species groups Relationships among organism groups On finding good indicators of Red Data Book species Towards an approach for efficient conservation of different species groups in oak-rich forest stands References Tack/Thanks An oak in a dense oak-rich forest at Aspenäs, one of the study sites, at the start of the partial cutting in the autumn of Note the forester to the left. (Photographer: Frank Götmark) 3

4 The life as a PhD-student... Heidi drawing the measuring-tape, assisting with surveys, and coordinating the field work The Oak Forest Research Group : Frank Götmark, Björn Nordén, Ted von Proschwitz, Heidi Paltto, Niklas Franc, Bjørn Økland Andreas Karlsson measuring tree girths Ingela Sandberg surveying vascular plants Johan Dahlberg surveying bryophytes Kia Jungbark measuring tree girths The grid for vascular plant surveys. 4

5 List of papers This thesis is based on the following papers which will be referred to in the text by their Roman numerals: I Paltto, H. & Nordén, B. Which spatio-temporal scales are most important to predict occurrence of species at a local scale (Manuscript) II Paltto, H., Nordén, B., Götmark, F. & Franc, N At which spatial and temporal scales does landscape context affect local density of Red Data Book and Indicator species? Biological Conservation 133: III Götmark, F., Paltto, H., Nordén, B. & Götmark, E Evaluation of partial cutting in broadleaved temperate forest under strong experimental control: Short-term effects on herbaceous plants. Forest Ecology and Management 214: IV Paltto, H., Nordén, B. & Götmark, F. Partial cutting as a conservation alternative for oak Quercus spp. forest response of bryophytes and lichens on dead wood. (Manuscript) V Nordén, B., Götmark, F., Ryberg, M., Paltto, H., Allmér, J. Partial cutting reduces species richness of fungi on woody debris in oak-rich forests. (Manuscript under revision, preliminary accepted by Canadian Journal of Forest Research) VI Nordén, B., Paltto, H., Götmark, F. & Wallin, K. Indicators of biodiversity, what do they indicate? Lessons for conservation of cryptogams in oak-rich forest. Biological Conservation 135: My contributions to the six papers: (I) I came up with the research idea, carried out the modelling work and analyses, and I wrote most of the manuscript with assistance by BN. I got help with the construction of the GIFMmodel and with writing of the technical parts about the model. (II) I conducted part of the field work (vascular plants, forest floor bryophytes/lichens), all the analyses, and I wrote the manuscript with assistance by BN and FG. Environmental data was gathered jointly by the research group. (III) I conducted part of the fieldwork together with FG, EG and others, I conducted the ordination analyses, and gave comments on the manuscript. (IV) I conducted part of the field work, all the analyses and wrote the manuscript, with assistance by FG and BN. (V) I carried out the ordination analyses, and gave comments on the manuscript. (VI) I conducted part of the field work (forest floor bryophytes/lichens), part of the data analyses, and gave comments on the manuscript. 5

6 Abstract Oak-rich forests a species rich but threatened habitat Oak-rich temperate forest is a species-rich forest type, which mainly occurs as more or less isolated remnants. The area of oak forest has decreased dramatically during recent centuries due to overexploitation and changed land use. In addition, during the last century the few remaining oak forests have become denser and darker due to ceased hay-cutting and grazing by domestic animals. Fragmentation and succession are probably the main threats facing biodiversity in this forest type. In this thesis, I study oak-rich forests and examine how different species (organism) groups respond to habitat fragmentation, and to partial cutting with the combined goals of increased biodiversity and biofuel extraction. I also examine how well different species groups indicate one another with respect to their species richness, and how these correlations may be related to the dispersal ability and habitat requirements of the species groups. For these purposes, I study four organism groups vascular plants, bryophytes, lichens and wood-inhabiting fungi in 25 old oak-rich forest stands in southern Sweden. In addition, I carry out a theoretical landscape ecological study to complement the empirical studies. Lichens and bryophytes need large landscapes, while vascular plants and wood-inhabiting fungi responded to small landscapes Species may be affected negatively by losses of their natural habitat, both directly by the loss of habitat patches, and indirectly in the remaining habitat patches due to limited dispersal resulting in insufficient recolonisation of empty patches. However, few studies have shed light on at which spatial and temporal scales the habitat amount is of practical importance for species and their conservation. Using a metapopulation model, I showed that species superior at dispersal generally were affected by the amount of habitat in larger landscapes compared to species inferior at dispersal for which habitat in smaller landscapes had a stronger impact. This means that the minimum landscape size meaningful for conservation of a species is larger for species with long average dispersal distance compared to species with shortdistance dispersal. The empirical study showed that the density of lichens and bryophytes (of conservation concern) increases in a local forest stand with increasing amount of current deciduous forest 1 5 km from the study stands. In contrast, the current number of vascular plants and wood-inhabiting fungi (of conservation concern) increased with increasing amount of deciduous forest at a smaller spatial scale (0 1 km). These results may indicate that lichens and bryophytes are superior at dispersal compared to vascular plants and wood-inhabiting fungi. In addition, the two latter organism groups were affected more by the habitat amount 120 years ago than by the current amount of habitat. In other words, vascular plants and wood-inhabiting fungi showed a delayed response to changing landscapes, which is consistent with an extinction debt since the amount of habitat has decreased recently in the landscape. After partial cutting the vascular plants and lichens on dead wood increased, while the wood-inhabiting fungi decreased Many species in oak-rich forests are adapted to semi-open conditions, but the oak-rich forests have become darker. Large-scale restoration to grazed pastures would be desirable but may be too costly for large-scale implementation outside existing nature reserves. Therefore management with combined goals of biodiversity conservation and forest management could be a good complement to core forest conservation. A partial cutting experiment was started to evaluate the effect of sunnier conditions for species expected to gain from such cutting, as well as for other species groups. Old oaks were retained and some of the thinner trees were cut for biofuel. At each of the study sites, one experimental and one control plot, each 1 ha, were surveyed before and after the cutting. The short-term response to partial cutting varied 6

7 among organism groups: the species density of vascular plants and lichens on dead wood increased; bryophytes on dead wood were not significantly affected; and wood-inhabiting fungi decreased. If also the responses of species groups studied within the same project but not included in this thesis (forest floor bryophytes; beetles; fungus gnats) are taken into account, the majority of the organism groups were positively affected or not affected at all, while only one organism group (wood-inhabiting fungi) was negatively affected. Species with similar ecology covaried weakly Species groups with similar ecology may covary in richness across landscapes. However, the species densities of four organism groups (only species of conservation concern considered) were weakly correlated or not correlated at all (n = 25 study sites; 2 ha study plots). Weak pairwise correlations (bryophytes with lichens; and vascular plants with wood-inhabiting fungi) were related to similar substrate requirements and dispersal ecology, while no correlations were found between species with large differences in their ecology. Possible explanations to the lack of strong correlations among species groups may be that the species groups are heterogenous and rarely have same substrate requirements and dispersal ecology. Indicator species were weak predictors of Red Data Book species in oak-rich forests of high conservation values In Sweden and other Nordic countries, Indicator (or signal ) species have been used to find forest stands with Red Data Book (threatened) species. Such data could potentially also be used for prioritising forests for conservation purposes, e.g. establishment of nature reserves. I evaluated the relationships for three cryptogam groups (lichens, bryophytes and wood-inhabiting fungi) and found that the total number of Indicator species was not correlated with the total number of Red Data Book species in oakrich forests. When only deciduous forest lichens were considered, the Indicator species and Red Data Book species were weakly correlated. Thus Indicator species, when treated collectively as a group, are not very useful in prioritising oak-rich forests for conservation. Indeed, they may still work for prioritising forest of high conservation value from production forests without conservation values. Landscape scale conservation of oak-rich forests In conclusion, the Indicator species may not be useful to find the most valuable oak-rich forests for conservation among oak-rich forests of high conservation value, with the exception of lichen indicators species. The amount of habitat at landscape scale may be a better indicator of Red Data Book species than are Indicator species used at local level. I suggest that conservation in oak-rich forests should preferably be focused on landscapes rich in deciduous forests instead of selecting individual forest patches rich in Indicator species. An appropriate minimum size of a landscape suitable for conservation may be 300 km 2 if the target group is lichens, 80 km 2 for bryophytes and 3 km 2 for vascular plants and wood-inhabiting fungi. In these landscapes the aim should be to maximize the amount of oak-rich deciduous forest. My suggestion is that grazing in oak woodland pastures and restoration of dense oak-rich forests to oak woodland pastures should be concentrated to these core conservation areas, while partial harvesting in oak-rich forests may be a good complement to this type of conservation, and should be applied elsewhere in the landscape. My suggestion is to carry out partial harvesting (or pure conservation actions) in % of all oak-rich forest in southern Sweden, and leave % of the forests for natural succession. However, it is important that the partial cuttings are done carefully, and the cuttings are evaluated in longterm perspective. The recommendation to cut such a high proportion of the oak-rich forests is based on the assumption that many of the woodinhabiting fungi, that decreased due to cutting, also can be found in naturally closed deciduous forests without oak, while species confined to open oak-rich forests (many lichens and beetles) to large extent lack suitable habitat in current forests. 7

8 Sammanfattning Artrikedomen i ekskogar är hotad Ädellövskogar, och i synnerhet lövskogar med ek, är oerhört artrika miljöer, men arealen av dessa har minskat till en bråkdel av vad som fanns för 300 år sedan. Traditionellt har den här typen av skogar betats av nötkreatur, får eller hästar, och ännu längre tillbaka i tiden har många av markerna skötts som slåttermarker. Under det senaste århundradet har flertalet lövskogar vuxit igen. Igenväxningen och den kraftiga arealminskningen av ekrik skog är troligen två viktiga orsaker till den utarmning av arter som observerats i ekskogar och annan ädellövskog. Mitt doktorandarbete baseras på studier av biologisk mångfald av kärlväxter, lavar, mossor och vedsvampar i 25 ekrika lövskogar i Götaland. Jag har studerat hur den biologiska mångfalden påverkas av det omgivande landskapet och utvärderat effekterna på biologisk mångfald av naturvårdsgallring. Gallringen syftar till att öppna upp skogen till gagn för de många arter som minskat pga igenväxning samtidigt som skogen kan nyttjas till biobränsleproduktion. Jag har också undersökt hur olika organismgrupper samvarierar i landskapet och hur detta kan relateras till likheter och olikheter i artgruppernas ekologi. Utöver dessa studier har jag gjort datasimuleringar i en landskapsmodell för att studera landskapets betydelse för arter med olika egenskaper. Olika organismgrupper är beroende av olika stora landskap Enligt landskapsekologiska teorier minskar en art om arealen eller kvaliteten på dess habitat (livsmiljö) minskar. Överlevnaden av en art är större i större landskap (givet samma täthet av habitat), men det finns få studier som undersökt vilken storlek på landskap som är mest betydelsefull för en viss art eller artgrupp, och hur denna storlek på landskap skiljer sig mellan olika arter eller artgrupper. I en teoretisk studie visade jag att arter med god spridningsförmåga är beroende av större landskap (givet samma täthet av habitat) jämfört med arter med sämre spridningsförmåga. Detta resultat använde jag vid tolkning av en studie där jag relaterade artrikedomen av naturvårdsintressanta arter av kärlväxter, lavar, mossor och vedsvampar i de ekrika skogarna med mängden lövskog i det omgivande landskapet. Lavar och mossor, vars artrikedom ökade med ökande mängd lövskog i stora landskap (cirkelradie 5-10 km), borde därmed vara mer lättspridda än kärlväxter och vedsvampar, vars artrikedom ökade med mängden lövskog i små landskap (cirkelradie 1 km). Utöver detta fann jag att artrikedomen av kärlväxter och ved-svampar var bättre relaterad till mängden lövskog för 120 år sedan jämfört med dagens landskap, vilket tyder på att arterna reagerar långsamt på förändringar som sker på landskapsnivå. Eftersom mängden lövskogar har minskat i landskapet under de senaste 120 åren innebär detta att arter ännu inte anpassat sig till den mindre mängden habitat som finns nu, och att det finns risk att vissa arter inte kommer att kunna anpassa sig utan dör ut. Antalet kärlväxter och vedlevande lavar ökar efter naturvårdsgallring medan antalet vedsvampar minskar Många arter som är knutna till öppna och halvöppna ekskogar och ekhagar, men under det senaste århundradet har de flesta ekrika hagar och skogar vuxit igen och blivit för mörka för dessa arter. För att bevara många utrotningshotade arter, vore det önskvärt att dessa skogar öppnades upp igen och betades med t ex nötkreatur. Detta skulle bli dyrt, och frågan är om inte det finns andra möjligheter att komplettera den relativt dyra skötseln som idag görs inom mycket begränsade arealer, inte minst i naturreservat. Därför startades det ett projekt med fokus på naturvårdsgallringar i ekskogar, och deras effekt på biologisk mångfald. Gallringarna görs i syfte att utvinna biobränsle, vilket ger en inkomst för markägaren, samtidigt som artgrupper, som missgynnas av igenväxning får en möjlighet att bevaras och/eller nyetableras. Vid naturvårdsgallringarna sparades de gamla ekarna, medan yngre träd togs ut för biobränsle. I varje ekskog inventerades två 100 x 100 m stora ytor, en gallringsyta och en referensyta, före och efter gallringen. Antalet kärlväxter och 8

9 vedlevande lavar ökade efter gallringen, antalet vedlevande mossor förändrades inte, och antalet vedlevande svampar minskade. Sammantaget ökade artmångfalden tack vare gallringen, särskilt om man räknar med andra artgrupper som inte studerades i denna avhandling (markmossor, skalbaggar och svampmygg). signalarter, om flera organismgrupper används ihop, inte kan användas för att välja ut de mest värdefulla ekskogarna till naturvårdsändamål. Möjligen skulle antalet signalarter av lavar knutna till lövskogar kunna användas för ändamålet, men sambandet var som nämndes svagt. Arter med liknande ekologi visar samma mönster i artrikedom Artrikedom av olika artgrupper kan samvariera över ett landskap. Om två artgrupper är båda två artrika på ett ställe och artfattiga på ett annat ställe, och i övrigt varierar på liknande sätt, kallar man detta för samvariation. Artrikedomen av naturvårdsintressanta arter bland de fyra studerade organismgrupperna samvarierade svagt eller inte alls bland de 25 ekrika skogarna. Svag samvariation (mellan lavar och mossor; mellan kärlväxter och vedlevande svampar) var related till likartade krav på livsmiljöer och likartad spridningsekologi, medan artrikedomen av artgrupper som hade mycket olika ekologi inte samvarierade alls. Stark samvariation mellan artgrupper verkar vara en sällsynt företeelse och en orsak kan vara att artgrupper sällan har lika krav på sina livsmiljöer och skiljer sig ofta med avseende på spridningsförmåga. Antalet signal- och rödlistade arter samvarierar inte i ekskogar med höga naturvärden I Sverige och andra nordiska länder används signalarter tillsammans med ett antal skogliga strukturer såsom död ved och gamla träd för att hitta rödlistade (hotade) arter. Om det finns ett starkt samband mellan antalet rödlistade och signalarter, skulle signalarter även kunna användas för att välja ut de mest värdefulla skogarna (med flest rödlistade arter), inför exempelvis naturreservatsbildning, bland skogar med höga naturvärden. Vid utvärdering av signal- och rödlistade arter av tre organismgrupper av kryptogamer fanns inget samband när artgrupperna (lavar, mossor, vedsvampar) testades ihop. Däremot samvarierade antalet signal- och rödlistade arter när lövskogsberoende lavar testades separat, men sambandet var svagt. Därför drar jag slutsatsen att Skoglig naturvård på landskapsnivå Eftersom signalarter förmodligen inte är tillräckligt effektiva mätare på antalet rödlistade arter i ekskogar (förutom möjligtvis lavar), är frågan om inte det finns bättre sätt att finna områden med rödlistade arter. Eftersom antalet rödlistade arter är beroende av mängden lövskogar i landskapet, föreslår jag att artbevarande snarare bör utgå från landskap som enhet i praktisk naturvård, istället för enskilda skogsbestånd på basis av antalet signalarter. Utifrån mina resultat föreslår jag att en potentiellt meningsfull skala för detta ändamål kan vara 300 km 2 för lavar, 80 km 2 för mossor och 3 km 2 för kärlväxter och vedlevande svampar. I dessa landskap bör man maximera mängden ekrika lövskogar. Jag föreslår också att restaurering av igenvuxna ekbestånd till ekhagar eller betade ekskogar görs företrädesvis i dessa landskap, och att dessa åtgärder kompletteras med naturvårdsgallringar där restaurering inte är möjlig och framförallt i övriga delar av Sverige. Jag föreslår att % av de ekrika skogstyperna naturvårdsgallras eller sköts som ekhagar/betade ekskogar och % lämnas för fri utveckling. Den höga andelen gallringar motiveras av att det finns ett flertal arter framförallt skalbaggar och lavar som är beroende av ljusöppna skogar och som idag inte har andra alternativa livsmiljöer än trädklädda hagar, medan t ex vedsvamparna även kan finna lämpliga växtmiljöer i lövskogstyper utan ek. Det är mycket viktigt att ingreppen görs med fokus på att maximera naturvärdena och på att bevara denna skogstyp för framtiden. Uttagen får inte bli ett självändamål. Det är också nödvändigt att andra typer av nyckelbiotoper utan ek, lämnas för fri utveckling och att följderna av naturvårdsgallringar i de ekdominerade nyckelbiotoperna följs upp på längre sikt och på landskapsnivå. 9

10 1. Introduction Oak-rich temperate forest is a disproportionately species-rich forest type that in Europe mainly occurs as more or less isolated remnants. For example, in many areas in south-eastern Sweden up to 95 % of the old oaks were cut during the 18 th and 19 th century (Eliasson 2002). In addition, the forest structure of oak-rich forests has changed because of a lack of natural disturbance regimes such as flooding, fire and grazing by large mammals (Rose 1992). In general, the forests are now denser and darker than a century ago (Nilsson et al. 2005). The change in forest structure is considered as a cause of widespread species decline in oak forests (Rose 1992). Belyea & Lancaster (1999) consider three major factors that define local community assembly: dispersal constraints, environmental constraints and internal dynamics. Species dispersal is constrained by the amount of suitable habitat available in the landscape (Fahrig 2003; Hanski & Gaggiotti 2004). Thus the loss of habitat is a serious threat to species living in particular habitat types (Fahrig 2003; Meffe & Carroll 2005), but species may respond to habitat loss at different spatio-temporal scales depending on their life history traits, especially dispersal characteristics. To better understand the effects of habitat loss for biodiversity in general and if necessary, to halt its effects, it is important to understand which landscape scales are of importance for different species. Local environmental factors is one of the constraints for a community of species (Belyea & Lancaster 1999). Some environmental factors are not easily influenced by man, for example temperature and soil-ph, while other factors, e.g. forest structure may change substantially through time because of human impact. Even though the oak forest and the associated disturbance regimes are natural from a long-term perspective (Lindbladh et al. 2000), the present structure of old oak-rich forests is strongly influenced by historical factors such as wood demands for ship-building and grazing regimes. A large proportion of the species found in the Swedish oak forests depend on relatively open or semi-open conditions that were common in grazed or wind- and fire-disturbed forests. After the abandonment of grazing in forests for years ago the forests became denser, and many species have decreased and are today at risk of extinction. Increasing the light penetration to the old oaks, e.g. by cutting trees in between the oaks, could potentially reverse the current negative trend for biodiversity. Opening of the forest may also be critical for oak regeneration, as there is now widespread lack of oak seedlings in dense forest (Götmark 2007). With increasing national and international interest in biofuel extraction, i.e. harvesting trees of thinner dimensions as an energy resource, conservation efforts could potentially be combined with sustainable forestry practices. This possibility has not previously been explored experimentally. Nature conservation and forestry authorities at international, national and regional levels seek and attempt to produce long-term conservation strategies. It is important that useful scientific knowledge is translated into practice in a way that conservation efforts may be carried out in the best and most efficient way. For example, accurately locating areas of highest conservation value is of great importance. This may be done by comparing surveys of total species richness or surveys of threatened species (Red Data Book species) for various areas, but such surveys are time consuming and therefore expensive. Species indicators of high conservation value are in use, but have rarely been scientifically evaluated. My thesis focuses on species conservation at regional level. I study the richness of four organism groups, vascular plants, lichens, bryophytes and wood-inhabiting fungi, in 25 oakrich forest stands located in southern Sweden. In addition to local species richness, I use the richness of Red Data Book species as an important measure of conservation value of a forest stand, since these are key species for maintaining the regional species richness. The used Red Data Book species are nationally listed as threatened (Gärdenfors 2005) according to criteria developed by the World Conservation Union (IUCN). I also measure the richness of Indicator species as a proxy for high conservation value. The Indicator species (or signal species in Swedish) that I study are used in the Swedish Woodland Key Habitat inventory, 10

11 Explanations of ecological terms Biodiversity = a general term including diversity of species, genes, habitats, landscapes etc. Forest type = a forest with defined characteristics such as a typical tree species composition Deciduous forest = Forest mainly consisting of trees that drop their leaves in the autumn Habitat = an area that supports the needs of an organism, i.e. a suitable environment for that species Habitat patch = a patch of habitat, for example a forest stand for forest-living species Habitat loss = the decrease of habitat regardless how the configuration of habitat is affected Habitat fragmentation = the configuration of habitat is changed towards a larger number of small habitat patches, the amount of habitat is not affected; the term is also commonly used broadly for simultaneous habitat loss and fragmentation Species richness = species number or species density collectively Species density = species number/area unit Species occupancy = species presence (in contrast to species absence) Species abundance = amount/frequency of a species Indicator species = species that indicate presence of other species/species groups or environmental conditions Signal species = indicator species as used in the Key Habitat Inventory, these species are expected to indicate presence of Red Data Book species Red Data Book species = species at risk of extinction to various degree according to criteria developed by the World Conservation Union (IUCN) Population = a group of organisms of the same species living together within a common area at the same time Population dynamics = change in size of a population due to reproduction, dispersal and extinction Viable population = a population that persists in the long run Metapopulation = a number of populations interacting with one another through dispersal between distinct patches of habitat Life history trait = characteristics of a species, e.g. age, amount of dispersal units, extinction rate etc., influencing the species population dynamics Spatial scale = a scale with area as unit, e.g. a small spatial scale consists of a small area Temporal scale = time scale Spatio-temporal scale = combination of area and time scale Woodland Key Habitat (WKH) = a forest area of high conservation value, i.e. Red Data Book species occur or are expected to occur, Extinction debt = a species have an extinction debt when it still persists in a recently fragmented landscape that no longer supports its long-term existence Epiphytic = living on plants, in this thesis living on bark Epixylic = living on dead wood Saproxylic = wood-decaying Some Swedish-English translations: Lichens = lavar Bryophytes = mossor Wood-inhabiting fungi = Vedsvampar Ascomycetes = sporsäcksvampar, t e x. skåloch kärnsvampar Vascular plants = Kärlväxter Herbaceous vascular plants = örtartade kärlväxter Ruderal species = ruderatväxter, dvs växter som trivs i störda miljöer Fungus gnats = svampmygg which aims at locating forest stands with high conservation value. By definition, a Woodland Key Habitat is a forest where Red Data Book species occur or probably occur (Nitare & Norén 1992). The list of Indicator species was established by the Swedish National Board of Forestry (Norén et al. 1995) to be used together with aspects of forest structure as indicators of 11

12 valuable forests. In this thesis I address four main questions that aim at 1) describing ecological patterns of biodiversity and 2) understanding the processes underlying these ecological patterns: How are different species groups affected by habitat loss? Which species show time delays in their response to changes in the landscape, and may thus suffer from an extinction debt? What are the most important landscape scales for species differing in their life history traits such as dispersal distance and local extinction? How do different species groups respond to partial cutting with combined goals of conservation and sustainable forestry (biofuel extraction)? How well is the species richness of different species groups correlated to one another at the forest stand level? How are the correlations and absences of correlations related to the species life history traits, and environmental constraints at local and landscape scales? How well is the species density of Indicator species correlated to the species density of Red Data Book species? Are Indicator species useful in prioritising oak-rich forests for conservation? In addition, I consider how the results of this thesis can be used as a basis for recommendations and advice in practical nature conservation: Given limited resources for conservation, how do we best organize the conservation of oak-rich forests in Southern Sweden? My doctoral thesis is a part of the research conducted within the Oak Forest Research Group at the Göteborg University, which is studying conservation biology in relation to forest stand management (partial cutting or absence of it) and landscape ecology for seven different organism groups (vascular plants, bryophytes, lichens, fungi, beetles, fungus gnats (Diptera, mycetophilids), and terrestrial molluscs). From , the group mainly consisted of Frank Götmark, Björn Nordén, Heidi Paltto, Niklas Franc, Bjørn Økland and Ted von Proschwitz. In addition, many undergraduate students and assistants have contributed with valuable work and help. 2. Methods I briefly summarize the methods, for details see the Methods sections of individual papers A theoretical study of landscape effects (paper I) By using a focal species landscape study approach I systematically explored what effect habitat amount in a landscape had on 18 hypothetical species with different dispersal and extinction characteristics. The average dispersal distance, but not the amount of dispersal units per time (colonization rate), was varied among the species. Thus, I consequently refer to average dispersal distance below when I use expressions such as inferior and superior at dispersal, or mobile and less mobile species. A spatially explicit grid-based metapopulation model that simulates habitat loss was constructed, and was run 200 times for each species, one at a time. Initially the whole grid ( cells) represented suitable habitat for the focal species, but during the simulation about 1 % habitat was lost per time step (year). The habitat loss was stochastic and varied greatly among time steps and among runs. New habitat patches were never created during the simulation. In the beginning of a run the model species occupied all the grid cells, and thereafter the population dynamics was adjusted for different species by changing mean values for average dispersal distance and extinction risk. There-upon I related current species occupancy in a focal habitat patch in the fragmented land-scape (5 % habitat remaining) to the a- mount of habitat in the surrounding landscape at varying points back in time and at varying spatial scales Empirical studies (papers II VI) The empirical studies (papers II VI) were based on field surveys in 25 oak-rich temperate forests in Southern Sweden (Fig.1) or a subset thereof. The forests were situated >15 km apart from one another. The study sites were former oak woodland pastures or oak meadows abandoned 12

13 Southern Sweden N Boreonemoral zone Nemoral zone km 1. Skölvene 14. Aspenäs 2. Karla 15. Norra Vi 3. Östadkulle 16. Fröåsa 4. Sandviksås 17. Ulvsdal 5. Rya åsar 18. Hallingeberg 6. Strakaskogen 19. Ytterhult 7. Bondberget 20. Fårbo 8. Långhult 21. Emsfors 9. Bokhultet 22. Getebro 10. Kråksjö by 23. Lindö 11. Stafsäter 24. Lilla Vickleby 12 Åtvidaberg 25. Albrunna lund 13 Fagerhult Fig. 1. Map of southern Sweden and the 25 study sites with oak-rich temperate forest. Reproduced from Franc (2007) with kind permission from the author years ago, and were characterized by remnant large oaks (the oldest at each site years old) and other broadleaved/coniferous trees of smaller dimensions. They all had high conservation value for nature conservation and have been designated as nature reserves (n = 7) or Woodland Key Habitats (WKHs, n = 18). In each stand, two plots (each 1 hectare) were delimited. One of the plots was partially cut in the winter 2002/2003, while the other was treated as a non-intervention control plot. For the landscape study (paper II) and the indicator study (paper VI), species data from before partial cutting was used from six diffe- 13

14 rent inventories: (1) epiphytic Red Data Book and Indicator species (bryophytes and lichens), (2) forest floor bryophytes and lichens; (3) epixylic bryophytes and lichens and (4) epiphytic bryophytes and lichens on the largest oak trees; (5) wood-inhabiting fungi and (6) forest floor herbaceous vascular plants. In the landscape study the local species density of Red Data Book and Indicator species of vascular plants, lichens, bryophytes and wood-inhabiting fungi were correlated to one another, and related to landscape and local data using regression analyses (n = 22 study sites). Only species typical of deciduous forests were included in the analyses (i.e. species mainly occurring in coniferous forests were excluded). The Indicator species of vascular plants were mainly long-lived forest species with clonal growth and relatively large seeds. In the Indicator species study the density of Red Data Book species, Indicator species and total species richness of lichens, bryophytes and wood-inhabiting fungi were related to one another using correlation analyses (n = 25 study sites). For the evaluation of partial cutting (papers III, IV and V), species data from before and after cutting were used. Paper III evaluated the effect of partial cutting for vascular plants (n = 6 study sites), paper IV evaluated the effect for epixylic bryophytes and lichens (n = 15 study sites), and paper V evaluated the effect for wood-inhabiting fungi (n = 21). Throughout the thesis the term species of conservation concern refers to Indicator & Red Data Book species, as treated collectively. 3. Species groups and their responses to habitat changes at different spatiotemporal scales (Papers I and II) Widespread habitat loss is a landscape scale process that has negative consequences for species living in the habitat type in question (Fahrig 2003). The species become extinct from the lost habitat fragments, and may also decrease in the remaining habitat if the habitat quality decreases or the species is dispersal limited (Saunders et al. 1991; Eriksson & Ehrlén 2001; Hanski & Gaggiotti 2004; With 2004). Around the oakrich forests studied in this thesis the average loss of deciduous forest area was 26 % during the last 120 years (within circles of 5 km radius; paper II). Around 16 of the study sites the area of deciduous forest decreased, while it increased around six sites. An important question is how habitat changes at landscape level affect different species and species groups, and what mechanisms cause the differences among these species. Species differ in their life history traits, which probably is important for their capability of resisting the effects of habitat loss (Ewers & Didham 2006). Especially their dispersal ability is important since the population dynamics at landscape level is restricted by dispersal and by establishment at new sites. I tested empirically the landscape effects for four different organism groups (paper II) using a study design called focal patch landscape study (Brennan et al. 2002; Fig. 1, paper II). This method is used to study the relationships between species occupancy/abundance or species richness and habitat amount in a landscape. In such analyses current species occupancy/abundance or species richness in a focal habitat patch is related to the a- mount of habitat in the surrounding landscape at various spatial scales (eg. in circles of different radii around the focal patch) or at various temporal scales (current and historic amounts of habitat). However, the lack of historic data of sufficiently good quality is an obstacle for the analyses and for drawing general conclusions from such studies. This is the case especially for studies exploring the length of time delays in species responses to landscape changes, which would need long time series of historical habitat data. In addition, the interactions of spatial and temporal scales in species responses are largely unexplored in a theoretical perspective. Therefore, I complemented the empirical study (paper II) with a modeling approach exploring the effects of species dispersal and extinction characteristics on their responses to landscape (paper I). 14

15 Fig. 2. Oak-rich forest before and after partial cutting at the study site Norra Vi. The photographs are taken at the same point at the same direction. Photographer: Frank Götmark. Before partial cutting 2002 After partial cutting 2003 After partial cutting

16 3.1. Lichens and bryophytes were dependent on habitat in large landscapes, while vascular plants and wood-inhabiting fungi responded to small landscapes Dispersal influences the species capability to withstand habitat loss (Clobert et al. 2004). In the metapopulation model used in paper I, an increase in the average dispersal distance of a species enhanced its survival probability in landscapes undergoing fragmentation. When 5 % habitat remained in the landscape, the occupancy of species superior at dispersal in the focal habitat patch were better explained by the amount of habitat at large landscape scales compared to species with shorter average dispersal distance (Fig. 2, paper I). The results from the empirical landscape study dealing with species of conservation concern (paper II) can be interpreted in the light of these results: Bryophytes and lichens, which responded to the amount of habitat at large spatial scales (1 5 km or 1 10 km), are probably better at dispersal than vascular plants and wood-inhabiting fungi, which responded to the landscape at 0 1 km scale. Also other empirical studies suggest that different species or species groups respond to different spatial scales, e.g. birds and terrestrial mollusks may be most dependent on the amount of habitat at large scales (5 10 km; Bailey et al. 2002; Götmark et al. 2008, in press), beetles, and Diptera on an intermediate scale (1 5 km; Chust et al. 2004; Holland et al. 2004), and vascular plants, beetles, mycetophilids (Diptera: Sciaroidea) and Homoptera on a small scale (0 1 km; Bailey et al. 2002; Chust et al. 2004; Holland et al. 2004; Økland et al. 2005; Franc et al. 2007). The response of vascular plants to small spatial scales was expected since the species in this group mainly consist of typical forest or ancient forest species with long life span, clonal growth and relatively large seeds (Honnay et al. 1998; Hermy et al. 1999; Kolb & Diekmann 2005). Similarly, it may be reasonable that lichens and bryophytes, of which many disperse with small airborne and plentiful spores, are better at dispersal. Also the Indicator and Red Data Book lichen Lobaria pulmonaria, found at our study sites, seems to be relatively good at dispersal (Werth et al. 2006; Werth et al. 2007). However, few comparative studies of dispersal distances for different species groups have been carried out. The response at a small spatial scale for wood-inhabiting fungi was the least expected, since fungi often are considered as relatively good at dispersal (Nordén & Appelqvist 2001). On the other hand, Sverdrup-Thygeson and Lindenmayer (2003) found that the occupancy of one wood-inhabiting fungi (listed as a Red Data Book species) was explained by the a- mount of habitat in the surrounding landscape at an even smaller landscape scale than in my study. In addition, this species showed delayed response to the amount of habitat in the landscape, which indicates that the species has relatively short-distance dispersal. The possible difference in dispersal between lichens and bryophytes, on one hand, and wood-inhabiting fungi, on the other hand, might be explained by differences in their dispersal ecology. In addition to spores, many lichens and bryophytes disperse by asexual propagules much larger than spores, which may be more easily dispersed by animals compared to spores. Fungi do not have this type of dispersal. However, whether these differences can explain the differences in my study or not, remain untested. In addition, some groups of epiphytic lichens may also be inferior at dispersal, since they are better explained by the habitat amount at small landscapes and also show delayed responses to habitat changes (Ellis & Coppins 2007) Vascular plants and woodinhabiting fungi showed a time delay in their responses to landscape changes Species responses to habitat changes often occur after a time delay, especially if the species are long-lived or they are near their extinction thresholds (Ovaskainen & Hanski 2002). A longlived species can persist a long time in a local habitat patch with little interaction with other individuals of the same species in other patches. In contrast, the time delay in a species response to increasing amount of habitat in the landscape may be mainly caused by limitation in the species dispersal. But for most species, both 16

17 dispersal and extinction characteristics may be important for the length of the time delay in their response to changes in landscape. A time delay in a species response following habitat loss may also be linked with a delay in extinction, if the amount of habitat in the current landscape does no longer support the occurrence of the species, and this is called extinction debt (Tilman et al. 1994). Shortly after habitat loss, delay in local extinctions should result in oversaturation of species in habitat patches. After an increase of the habitat in the landscape, slow dispersal should result in under-saturation, until a new equilibrium is reached. It is possible to explore the presence of time delays by analysing species responses or relation to current versus historic amount of habitat. A stronger species response to historic compared to current amount of habitat (at a certain spatial scale) is an indication of their slow response to temporal changes and this is often interpreted as a delayed response, or possibly an extinction debt if the landscape has experienced habitat loss. In the metapopulation modelling study, only one species showed a clear effect of habitat a- mount in the landscape in combination with a clear time delay to changing amount of habitat (Fig. 2g, paper I). This species had an intermediate dispersal capacity and I suggest that the reason for the time delay may be a nearextinction-threshold effect, since the species declined substantially during the simulation of habitat loss. For species inferior at dispersal (Fig. 2a, b, c, paper I) the overall landscape effects were weak, but the length of the time delays increased clearly with decreasing local extinction risk, as predicted by theory (Ovaskainen & Hanski 2002). In contrast to the species showing clear time delays, the time delays, especially at the largest landscape scale, were hardly detectable due to low R 2 -values, and would probably be interpreted as lack of landscape effects in empirical studies. This type of weak landscape responses may be typical of species showing remnant population dynamics (Eriksson 1996), e.g. by long-lived, often clonal plants with large seeds. These hardly detectable time delays in species responses may possibly explain the lack of landscape effects in some empirical studies of vascular plants (Cousins et al. 2007; Öster et al. 2007). In conclusion, species with low extinction rates, e.g. due to their long life lengths, may suffer from an extinction debt that is difficult to detect in empirical studies such as focal patch landscape studies. In the empirical study (paper II) the vascular plants and wood-inhabiting fungi showed time delays in their responses to habitat change in the landscape, and thus these may be linked with a risk of an extinction debt. The reason may be a low extinction risk or a near-threshold effect for the species within the organism group. The vascular plants in this study are long-lived, thus the first explanation may be appropriate. The explanatory power of the regression model was much higher for the wood-inhabiting fungi (R 2 = 0.78 compared to R 2 = 0.40 for vascular plants; paper II) and their response also otherwise resembled the response of species (g) in paper I (Fig. 2), which could indicate that the fungi rather are near their extinction threshold than long-lived and inferior at dispersal. It is also possible that the delayed response for wood-inhabiting fungi in the current study is an effect of a delay in substrate (dead wood) production compared to the dynamics of living trees, see paper II for further discussion of this possibility. Also other studies found that species or species groups are better predicted by historical than current amount or connectivity of habitat in the landscape: the species richness of vascular plants in semi-natural grasslands (Lindborg & Eriksson 2004; Helm et al. 2006), one polypore fungi (Sverdrup-Thygeson & Lindenmayer 2003), epiphytic lichens in aspen forests (Ellis & Coppins 2007), and rainforest fauna (Graham et al. 2006). However, these studies do not explicitly discuss differences in responses of different groups of species, with respect to different spatio-temporal scales of landscape included in the studies Red Data Book species were affected by the landscape, while Indicator species were not Two species groups often used in Swedish forest conservation are Red Data Book species (i.e. threatened species) and Indicator species of wich the latter are supposed to indicate presence 17

18 of Red Data Book species. The densities of Red Data Book and Indicator species were low, and therefore I pooled data for four organism groups (vascular plants, lichens, bryophytes and woodinhabiting fungi) to be able to test their dependence on the amount of habitat at landscape level. The density of Red Data Book species per study site (n = 22 study sites) was affected by the a- mount of deciduous forest in the landscape at the 1 5 km scale, while the Indicator species were not affected by the landscape factors at all. One reason for the lack of landscape effect for Indicator species may be the difference in the ecological amplitude between these two groups: The Red Data Book species are narrower in their substrate requirements while the Indicator species include a broader range of species with contrasting ecology (paper II). See further discussion on how species ecology and dependence of landscape may be related in section 5. The scale of importance for Red Data Book species was 1 5 km, i.e. a scale intermediate of the scale important for lichens (1 10 km) and wood-inhabing fungi (0 1 km), which sounds reasonable since both groups are included in the Red Data Book species in about equal frequencies. (Only one Red Data Book species of vascular plants and bryophytes each was found) Landscapes and ecological continuity Long ecological continuity of forests has been suggested to be an important factor for biodiversity, and indicator species for long ecological continuity has been used by conservationists to find valuable forests for conservation. However, the term ecological continuity remains poorly defined, and includes often aspects of both forest age and true continuity. Ecological continuity is often referred to as local continuity at stand level, and can be defined as the length of time a forest have had a typical old-growth forest structure (Nordén & Appelqvist 2001). Landscape level continuity has also been discussed (Nordén & Appelqvist 2001) and in this case, the continuity of habitat at landscape level is addressed regardless of the local continuity of individual forest stands. In a dynamic land-scape with forest fires or other disturbance regimes, landscape continuity may be more decisive for species long-term persistence than local continuity. The historical landscape was important to explain the richness of vascular plants and wood-inhabiting fungi of conservation concern, which can be interpreted as these organism groups depend on ecological continuity at some landscape level. In contrast, local ecological continuity seems not to be important for richness of the studied groups (Röstell 2006). In contrast, landscape continuity did not determine species richness of lichens and bryophytes of conservation consern at the spatial scales measured in paper II. Although it is probable that the continuity at larger spatial scales could be important for these species groups. 4. Effects of partial cutting/ biofuel harvest for different species groups (Papers III, IV and V) Oak forest is a disproportionately species rich forest type in southern Sweden as compared to other forest types (Ranius et al. 2001). Reasons for this may be the large size and the high age of oaks compared to other trees, hence they provide a wide range of microhabitats such as bark structures, large hollows and special qualities of dead wood (Rose 1974; Pihlgren 2000). Many insects, lichens, fungi and birds are dependent on these structures for their survival (Thor 1998; Ranius et al. 2001). But the biodiversity of oak forests is seriously threatened due to habitat loss and decreased structural diversity in the remaining forest remnants. Oak was one of the most common trees in southern Sweden and elsewhere in Europe during the postglacial time (Lindbladh et al. 2000). The openness of the primeval temperate forests is debated, with arguments for closed forests (Ellenberg 1988; Peterken 1996; Mitchell 2005), and for more open forest structure (Andersson & Appelqvist 1990; Nilsson 1997; Vera 2002). Regardless of the primeval conditions, humans were the cause of oak forest openness in southern Sweden and in Europe in historical 18

19 time, i.e. during the last about years, which should have had a positive effect on species dependent on oaks standing in light or semi-light conditions. Due to ceased grazing and secondary succession during the last years, the oak forests have become significantly denser, which seems to be a serious threat for the biodiversity of oak forests (Nilsson et al., 2005). Traditionally, conservation of biodiversity has been a matter for the state in Sweden, and it has been a political question how expensive nature conservation is allowed to be and who should pay for short-term costs. Better knowledge about species ecology is important, but it is also important to evaluate new conservation concepts including significant benefits for human being. If biodiversity in oak forests benefits from increased canopy openness, some cutting in the forests could be desirable. Further, this could potentially be combined with extraction of woody material for biofuel production. An income from some cutting should be a better motivation for forest owners to conduct conservation action compared to traditional handsoff state conservation. This motivation has potential to result in cuttings in regionally much larger scales than pure conservation actions financed by the limited budgets of conservation authorities. The interest in biofuel extraction has increased during the last years. Today about 20 % of the energy consumption in Sweden consists of biofuels, of which 85 % comes from the forest. The amount of biofuel is predicted to increase with about 80 % until year 2020 (Skogs-Eko 1997). However, except for the extraction of tree tops and branches of coniferous trees, little research has been conducted on the effects of forest fuel harvest on biodiversity. If conducted with care and evaluated in the longterm, biofuel extraction in dense oak forests may result in benefits for both conservation and forest production (Skogsstyrelsen 2001). Although species richness of epiphytic lichens and wood-depending beetles may be expected to respond positively to partial cutting, it is important to evaluate the effects on a broad range of organism groups. It is possible that some groups would be disfavoured by opening of canopies even if it is done carefully. I quantified short-term effects of partial cutting on vascular plants (paper III), lichens and bryophytes living on dead wood (paper IV) and wood-inhabiting fungi (paper IV) Vascular plants and woodinhabiting lichens increased after partial cutting, while the woodinhabiting fungi decreased The species richness of vascular plants (paper III, Table 2) and lichens on stumps (paper IV, Fig. 2b) increased due to the partial cutting. Lichens and bryophytes on logs, bryophytes on stumps, and wood-inhabiting ascomycetes were not significantly affected (paper IV, Fig. 2a and 2b; paper V; bryophytes on stumps showed a near significant decrease, p = 0.058). The species richness of wood-inhabiting basidiomycetes decreased due to cutting (paper V), though the decrease was smaller than between-year variation in the reference plot. Several ruderal species among the vascular plants increased in abundance, but the most of the changes occurred in grassland and forest species. The epixylic species composition shifted towards a flora typical for dryer dead wood. For vascular plants and lichens on stumps, the two groups that clearly increased due to cutting, the increase was caused by increased colonisation, while the extinction rates were not affected. For the wood-inhabiting fungi, the decrease was significant for species living on fine woody debris, while species richness on coarse woody debris was unaffected. The number of Red Data Book species was too low to be tested one organism group at a time. Also the animal groups studied at the same oak-rich study sites varied in their responses to partial cutting. Wood-inhabiting and herbivorous beetles were favoured in species richness (Franc 2007; Franc & Götmark 2008); the species richness of Red Data Book species of woodinhabiting beetles was unaffected, but regression analyses suggest that opening of forest canopies by more than in the current cutting would disfavour this group of species. Species richness of fungus gnats was not affected (Økland et al. 2008, in press). In conclusion, the richness of three organism groups (vascular plants, lichens and wood-in- 19

20 habiting/herbivorous beetles) clearly increased due to the partial cutting as expected. One group declined (wood-inhabiting basidiomycetes) and three groups were not significantly affected (wood-inhabiting bryophytes, wood-inhabiting ascomycetes and fungus gnats). The main trend is thus an increase in species richness after partial cutting. 5. Relationships among organism groups (Papers II and VI) An Indicator species is a species indicating presence of another species, a set of species, a species community, or environmental condition (Ferris & Humphrey 1999). Indicators of biodiversity may be useful in conservation, if the target group of species or species richness is difficult or time-consuming to investigate in comparison with the Indicator species or Indicator species group. For example, well-known species groups may act as indicators of less well-known species groups, or conspicuous species may act as indicators of small or cryptic species. In this thesis I have evaluated the relationships among the species groups studied and their usefulness as indicators for selection of conservation areas Correlations among species groups are weak or absent The species richness of three organism groups in oak-rich forests was tested pair-wise with correlation analyses, using sites as statistical sample units (paper VI). The number of bryophytes increased with increasing number of wood-inhabiting fungi (non-parametric bootstrapping r = 0.37; p = 0.021; n = 25), while species richness of lichens was not related to the richness of the other two groups. Correlation analyses among four organism groups with respect to the number of species of conservation concern (deciduous forest Indicator and Red Data Book species pooled; only species confined to deciduous forests included) revealed no significant correlations (paper II). Two of the correlations were near significant (bryophytes lichens: Spearman r = 0.41; p = 0.060; n = 22; and vascular plants woodinhabiting fungi: Spearman r = 0.37; n = 22; p = 0.094). Overall, the correlations between different types of richness-measures were weakly positive or absent (see above, and papers II and VI). This corresponds well to the general trend in correlation tests between different species groups (reviewed by Wolters et al. 2006). The correlation coefficients in the reviewed studies are on average (both significant and nonsignificant correlations included), which corresponds to a degree of explanation as low as 14 % (Wolters et al. 2006). The correlation coefficients for the significant correlations found in this thesis were of about the same magnitude ( ). Only 1/3 of all correlations in the review (n = 152) were significant, which corresponds well with the results in this thesis. The question is why the correlations among different species groups often are weak, and consequently, if Indicator species are useful in conservation work. The richness of different organism groups may be correlated for several reasons: (1) random coincidence, (2) interactions between species in the various organism groups, (3) similar response to common factors such as soil-ph, substrate type or precipitation, and (4) response to different environmental factors that covary spatially (Gaston 1996). Here I will discuss the third explanation with respect to Indicator and Red Data Book species of vascular plants, lichens, bryophytes and wood-inhabiting fungi Species groups with similar ecology were weakly correlated to one another Bryophytes and lichens (Indicator species and Red Data Book species pooled; paper II) had similar habitat requirements (favoured by high precipitation) and they were related to the habitat amount in the landscape at about the same spatial scale (1 5 for bryophytes and 1 10 km for lichens). The species density of lichens was predicted by the amount of Woodland Key Habitats while the bryophytes were predicted by the amount of deciduous forests in general. I suggest that the similarity in the spe- 20

21 cies requirements explains the weak, nearly significant correlation between bryophytes and lichens. Similarly, both vascular plants and woodinhabiting fungi were mainly favoured by the same environmental factors, high soil-ph and the amount of habitat at the same spatial scale (0 1 km), thus a weak, nearly significant, correlation was found between these two groups. and species with different ecology were not correlated at all In contrast to the correlations presented above, no other pair-wise correlations were found among these four organism groups. This may be explained by larger differences in ecology between the other possible comparisons of species, e.g. lichens vascular plants. In conclusion, an explanation to the lack of strong correlations among species groups may be the fact that two species never have exactly the same substrate requirements or dispersal ecology, thus a large but not easily explained variation is always included in the measures of species richness. In addition, stochasticity in external factors such as weather conditions affecting dispersal and extinction processes probably cause additional variation in species densities. The question is if the weak correlations among organism groups can be useful for practical conservation. Different organism groups may need different conservation actions, and may only marginally be pooled to gain efficiency in conservation. 6. On finding good indicators of Red Data Book species (Paper VI) 6.1. What kind of species should be in focus in conservation? An area of high conservation value, a hot spot for conservation, is often defined as an area with high local species richness or presence of rare or threatened species. High local species richness may arise due to several factors. One such factor is the naturalness of an area: a high diversity of natural micro-habitats favours a wide array of species. For example, an old-growth forest may be species rich because it contains a high number of different tree species, and dead wood of different qualities including large stumps and logs. High species richness may also arise due to other factors, such as high diversity of habitats surrounding a habitat patch. For example, a forest stand surrounded by several habitat types may be more species rich than a forest surrounded solely by one type of forest, even though the forest stands are very similar. Thus, the species in the species richer habitat may be easily dispersed ones from many types of habitats, especially ruderal species common in the region. To conserve species diversity at a regional (for example national) level, it may be inefficient to conserve common species with viable populations. Therefore nature conservation in the Nordic countries often focuses on regionally rare species, especially Red Data Book species, i.e. species considered threatened (Gärdenfors 2005). One rational for this conservation strategy is: If we can protect many of these species, much of the regional species pool may also be conserved. Ideally, we would conserve all areas suitable for Red Data Book species, and areas that already host Red Data Book species. However, many Red Data Book species are small, cryptic, rare and not well known by people other than taxonomic experts. Therefore it would be easier if indicators of Red Data Book species could be used instead to prioritise areas for conservation. A good indicator species for old-growth forests should be easy to find, easy to identify, have an even distribution across the region, and be restricted to semi-natural or old-growth forest of conservation values (Ferris & Humphrey 1999; Nitare 2000). In the Woodland Key Habitat Inventories in Nordic countries, a pre-selected group of Indicator species (Signal species) have been used, in combination with indicators of forest structure, to find potential forests with presence of Red Data Book species. Forests delineated in the inventory are called Key Habitats. The selection of Indicator species was based on expert judgments, without thorough scientific evaluation. This course of action enabled the inventory to start very quickly, but still very few eva- 21

22 luations have been done on the efficiency of the inventory. In addition to using Indicator species for delineating Key Habitats from surrounding production forests, the Indicator species themselves could potentially be used to prioritize areas for conservation among forests of high conservation values. In this study, I have evaluated the Indicator species capability of indicating Red Data Book species among Key Habitats and other forests of high conservation values, i.e. nature reserves Indicator species are not correlated with Red Data Book species Correlation analyses were carried out between the number of Indicator species and Red Data Book species (bryophytes, lichens and woodinhabiting fungi pooled; paper VI). Regardless of whether all species or only species confined to temperate deciduous forests were included, the number of Red Data Book species recorded was essentially independent of the number of Indicator species (Fig. 3a and 3b, n = 25 oakrich forests). These results are consistent with previous conclusions that among sites, species richness for species groups with different ecology may not be correlated to one another. The Indicator species had generally a broader ecological niche compared to Red Data Book species (paper II, Fig. 2a.), which may weaken the applicability of them as indicators, at least if all Indicator species, regardless of organism group, are pooled Indicator species of deciduous forest lichens are weakly correlated with Red Data Book species As the next step, the organism groups were tested separately. When both species typical of coniferous and deciduous forests were included, the number of Indicator species of lichens was not significantly correlated to the Red Data Book species (Fig 3c: Non-parametric bootstrapping: r = 0.37; n = 25; p = 0.067), but when species confined to temperate deciduous forests were considered separately, the number of Red Data Book species more clearly increased with in-creasing number of Indicator species (Fig. 3d: Non-parametric bootstrapping: r = 0.42; p = 0.039). For wood-inhabiting fungi, no such Red Data Book species Red Data Book species a) b) Indicator species (signalarter) c) 3 p = p = Indicator species (signalarter) Red Data Book species Red Data Book species Indicator species (signalarter) d) Indicator species (signalarter) Fig 3. The relationship between the density of Red Data Book species and Indicator species (n = 25 oakrich forests): a) all bryophytes, lichens and wood-inhabiting fungi, b) bryophytes, lichens and woodinhabiting species characteristic for deciduous forest, c) all lichens, and d) lichens characteristic for deciduous forest. 22

23 correlation was found. The bryophytes and vascular plant species were too few for meaningful tests The Indicator species are not very useful in prioritising oak-rich forests of conservation values Deciduous forest lichens seem to be the only species group among the studied ones that may be useful as Indicators for Red Data Book species, although also this correlation was relatively weak (paper VI). The correlation is reasonable since naturally dry semi-open oak forests host a disproportionately large number of both Indicator and Red Data Book lichens, and many of these lichens depend on old, large oaks and dead wood thereof. A high density of Key Habitats with old oaks or other deciduous trees of high conservation quality at the landscape level seems necessary for locally high diversity of Indicator and Red Data Book species of lichens (paper II). Thus the number of these two species groups could be expected to covary. In contrast, the number of indicator species and Red Data Book species of wood-inhabiting fungi were not correlated, and I suggest one main reason for this: The Indicator species are confined to fine woody debris (often of hazel) while the Red Data Book species grow on both fine and coarse woody debris, i.e. the two groups differ in their ecological requirements more than do the Red Data Book and Indicator species of lichens. In addition, the fungal species may be inferior at dispersal compared to lichens (paper II, see also section 3 in this thesis), and thus the delay in their responses to habitat changes at landscape level may weaken their relationship to their substrate at landscape level. A third explanation could be that the rare wood-inhabiting fungi are disfavoured by light and dry conditions (paper V) and may therefore have arrived lately to the forests due to secondary succession. The adjustment of species distribution may vary a- mong species within the group of wood-inhabiting fungi, and if the Red Data Book species differ from the Indicator species in this respect, the correlation between these two groups may be weak. In other words, compared to woodinhabiting fungi, the Indicator and Red Data Book species of lichens may have more similar habitat requirements, and similar immediate response to changing substrate amounts in the current landscape, explaining the stronger correlation between them. The number of Red Data Book species of vascular plants and bryophytes was very low in the inventories carried out at the 25 study sites (only one species of each group was found). The reason for this may be the preference of Red Data Book bryophytes to moist forests, hence oak forests may be suboptimal for these. Generally, few forest-dependent vascular plants in Sweden are currently red-listed. Notwithstanding, many of the long-lived forest vascular plants inferior at dispersal may be doomed to extinction because of long time delays in their responses to large-scale habitat loss at landscape level. These species may not been included in the Red Data Books, since they have not yet decreased to alarming low levels. In conclusion, the indicator species were not very efficient at identification of oak-rich forests rich in Red Data Book species. A possible exception may be the group of deciduous forest lichens, although only weak correlations were discovered. The indicator species concept for selecting very rich forests among a set of relatively rich sites seems too vague. But, it is still possible that individual species may be useful as good indicators of Red Data Book species a- mong forests of high conservation value, e.g. the lichen species Lobaria pulmonaria (Nilsson et al. 1995). 7. Towards an approach for efficient conservation of different species groups in oak-rich forest stands (Papers I VI) Belyea & Lancaster (1999) listed three major factors defining local community assembly: dispersal constraints, environmental constraints and internal dynamics. The dispersal constraints of species is linked to the species dependence on habitat at landscape level, and the environmental constraints for cryptogam species and vascular plants consists mainly of precipitation, 23

24 forest structure, i.e. presence of large trees and dead wood of different qualities, presence of boulders, and for vascular plants edaphic factors such as soil-ph. One conclusion from this thesis, relevant for conservation, is that vascular plant and cryptogam species may depend on much larger landscapes than have been supposed by researchers or conservationists. In addition, the four studied organism groups have different ecological requirements at local and landscape level. Hence, one conservation strategy may not satisfy the requirements of all species groups. The concept of Indicator species for indicating presence of Red Data Book species has been extensively used in the Swedish Key Habitat Inventory, together with information about structural qualities of the forests. The aim has been to separate sites of special conservation value from those without such value. As shown in the previous section and in paper VI, the indicator species are not very precise and useful in separating oak-rich forests of highest conservation value from those of lower or intermediate conservation value. Nevertheless, the Key Habitat Inventory identified a large number of potential sites for species dependent upon old trees and dead wood. In addition, many oldgrowth forests on calcareous ground have been identified, which may be important for Red Data Book species of fungi Finding high priority landscapes for conservation Landscapes rich in deciduous forests are predicted to be rich in Red Data Book species, and landscapes rich in WKH:s, i.e. old-growth deciduous forests, are predicted to be rich in Red Data Book and Indicator species of lichens (paper II). Based on my results, I suggest that an efficient strategy would be to use the Key Habitat Inventory, satellite data, IR-photographs and other map data for identification of landscapes rich in deciduous forests for landscape level conservation. Currently, the selections of conservation areas are often done at stand level, based on information about individual Key Habitats or other surveys The minimum area of a high priority landscape The focus of conservation is currently shifting from local stands as main units of conservation to a broader landscape perspective (Groom et al. 2005; SEPA & NBF 2005). Multiscaled biodiversity planning, matrix management (Lindenmayer & Franklin 2002; Lindenmayer et al. 2006) and conservation priorities for landscapes with large proportion of semi-natural habitats (Andersson 2002; SEPA & NBF 2005) are increasingly emphasized, but in most cases recommendations about minimum landscape sizes are vague or lacking. My thesis has contributed to this work by pointing out some potentially important scales for long-term conservation planning that preferably should span over several centuries, since oaks may be very long-lived. For lichens (Indicator species and Red Data Book species pooled) landscapes of a minimum size of 300 km 2 (corresponding to circle radius 10 km) may be an appropriate recommendation, for bryophytes the minimum landscapes could be 80 km 2 (corresponding to circle radius 5 km) and for vascular plants and wood-inhabiting fungi 3 km 2 (corresponding to circle radius 1 km). In addition, the largest landscape scale may be appropriate for terrestrial molluscs (Götmark et al. 2008, in press), and the smallest landscape scale may be appropriate for animal groups such as wood-inhabiting beetles (Franc et al. 2007) and fungus gnats (Økland et al. 2005). For beetles, the area of oakrich Woodland Key Habitat was important, while Woodland Key Habitats with mixed (deciduous-coniferous) forest was important for fungus gnats Partial cutting in oak-rich forests as a tool for restoring forest structural qualities Some valuable species, e.g. wood-inhabiting beetles and lichens, are restricted to semi-open or open oak forests. One probable reason for their recent decrease in Sweden is changed forest structure following ceased forest grazing by domestic animals. The partial cutting experiment showed positive short-term effects on species richness of four species groups 24

25 (lichens on stumps, forest floor bryophytes, vascular plants and wood-inhabiting/herbivorous beetles). Three species groups were not significantly affected (wood-inhabiting bryophytes, wood-inhabiting ascomycetes and fungus gnats) and one species group (wood-inhabiting basidiomycetes) was negatively affected (though not much). Although partial cutting increased species richness of many species groups, this management option will not be suitable for all Red Data Book species. Especially some light demanding species may still need a degree of openness not available in partially cut oak-rich forests. It may be a good suggestion to concentrate grazing in oak woodland pastures and restoration of dense oak-rich forests to oak woodland pastures, to the high priority landscapes suggested in the previous subsection (7.3.), while partial harvesting in oak-rich forests may be a good complement to this type of conservation, and could be applied elsewhere in the landscape. My suggestion is to carry out partial harvesting (or pure conservation actions) in % of the oak-rich forest area in southern Sweden, and leave % of the forests for natural succession. The reason for the recommendation of a relatively high proportion of partial cutting is based on the assumption that many of the wood-decaying fungi found in our study areas also can be found in naturally closed deciduous forests without oak, while species confined to open oak-rich forests to large extent lack suitable habitat in other types of forests. The suggestions above are tentative, since the effect of the surrounding landscape on species composition in restored local stands remains unclear, as well as the impact of the local restored stands to species composition in similar habitats in the surrounding landscape. Dispersal of species is potentially a slow process, and it may take long time for species gaining by partial cutting to actually establish in the cut forests. In landscapes changing rapidly the drawbacks of extinctions must not be outweighed by recolonisations of threatened species. The evaluation performed in this thesis is based on short-term observations that cannot fully cover the total potential immigration of new species. It is of utmost importance that the effect of partial cutting for species richness is continued in a long-term perspective, and that the effect of the surrounding landscape is included in the evaluations. 8. References Andersson, L Mapping nature protection values a habitat-wise presentation of regional variation in rare and vulnerable species richness (in Swedish; English summary). Svensk Botanisk Tidskrift 96: Andersson, L., and T. Appelqvist Istidens stora växtätare utformade de nemorala och boreonemorala ekosystemen En hypotes med konsekvenser för naturvården. Svensk Botanisk Tidskrift 84: Angelstam, P., and L. Andersson I vilken omfattning behöver arealen skyddad skog i Sverige utökas för att biologisk mångfald ska bevaras? (How much does the forest area for conservation in Sweden need to be increased to be able to conserve the biological diversity?) in Miljövårdsberedningen, editor. Skydd av skogsmark: Behov och kostnader. Bilagor. Statens Offentliga Utredningar SOU, Stockholm, Sweden. Bailey, S.-A., R. H. Haines-Young, and C. Watkins Species presence in fragmented landscapes: modelling of species requirements at the national level. Biological conservation 108: Belyea, L. R., and J. Lancaster Assembly rules within a contingent ecology. Oikos 86: Brennan, J. M., D. J. Bender, T. A. Contreras, and L. Fahrig Focal patch landscape studies for wildlife management: Optimizing sampling effort across scales in J. Liu, editor. Integrating Landscape Ecology into Natural Resource Management. Cambridge University Press, Port Chester, NY, USA. Chust, G., J. L. Pretus, D. Ducrot, and D. Ventura Scale dependency of insect assemblages in response to landscape pattern. Landscape Ecology 19: Clobert, J., R. A. Ims, and F. Rousset Causes, mechanisms and consequences of dispersal. Pages in I. Hanski, O.E. Gaggiotti, editors. Ecology, genetics and 25

26 evolution of metapopulations. Elsevier, Burlington, USA. Cousins, S. A. O., H. Ohlson, and O. Eriksson Effects of historical and present fragmentation on plant species diversity in seminatural grasslands in Swedish rural landscapes. Landscape Ecology 22: Eliasson, P Skog, makt och människor: En miljöhistoria om svensk skog Kungl. Skogs- och Lantbruksakademien. Durö Grafiska, Malmö, Sweden. Ellenberg, H Vegetation ecology of central Europe. Cambridge University Press, Cambridge, UK. Ellis, C. J., and B. J. Coppins th century woodland structure controls standscale epiphyte diversity in present-day Scotland. Diversity and distributions 13: Eriksson, O Regional dynamics of plants: a review of evidence for remnant, sourcesink and metapopulations. Oikos 77: Eriksson, O., and J. Ehrlén Landscape fragmentation and the viability of plant populations. Pages in J. Silvertown, and J. Antonovics, editors. Integrating ecology and evolution in a spatial context. Blackwell Science, Oxford, UK. Ewers, R. M., and R. K. Didham Confounding factors in the detection of species responses to habitat fragmentation. Biological Reviews 81: Fahrig, L Effects of habitat fragmentation on biodiversity. Annual Review of Ecology, Evolution and Systematics 34: Ferris, R., and J. W. Humphrey A review of potential biodiversity indicators for application in British forests. Forestry 72: Franc, N Conservation ecology of forest invertebrates, especially saproxylic beetles, in temperate successional oak rich stands. Doctoral thesis 30 pp. Department of Zoology, Göteborg University, Göteborg, Sweden. Franc, N., and F. Götmark Openness in management: hands-off versus partial cutting in conservation forests, and the response of beetles. Biological Conservation (manuscript under review). Franc, N., F. Götmark, B. Nordén, and H. Paltto Factors and scales potentially important for saproxylic beetles in temperate mixed oak forest. Biological Conservation 135: Gaston, K. J Spatial covariance in the species richness of higher taxa. Pages in M. E. Hochberg, and J. Clobert, editors. Aspects of the genesis and maintenance of biological diversity. Oxford University Press, New York, US. Graham, C. H., C. Moritz, and S. E. Williams Habitat history improves prediction of biodiversity in rainforest fauna. Proceedings of the National Academy of Sciences of the USA 103: Groom, M. J., G. K. Meffe, and C. R. Carroll, editors Principles of conservation biology. Sinauer associates, Sunderland, Massachusetts, USA. Gärdenfors, U., editor The 2005 Red List of Swedish species. ArtDatabanken, Uppsala, Sweden. Götmark, F Careful partial harvesting in conservation stands and retention of large oaks favour oak regeneration. Biological conservation 140: Götmark, F., T. von Proschwitz, and N. Franc. 2008, in press. Are small sedentary species affected by habitat fragmentation? Local vs. landscape factors predicting species richness and composition of land molluscs in Swedish conservation forests. Journal of Biogeography, doi: /j x. Hanski, I., and O. E. Gaggiotti, editors Ecology, genetics and evolution of metapopulations. Elsevier, Burlington, USA, San Diego, USA, and London, UK. Helm, A., I. Hanski, and M. Pärtel Slow response of plant species richness to habitat loss and fragmentation. Ecology letters 9: Hermy, M., O. Honnay, L. Firbank, C. Grashof- Bokdam, and J. E. Lawesson An ecological comparison between ancient and other forest plant species of Europe, and implications for forest conservation. Biological conservation 91:9 22. Holland, J. D., D. G. Bert, and L. Fahrig Determining the Spatial Scale of Species Response to Habitat. BioScience 54:227 26

27 233. Honnay, O., B. Degroote, and M. Hermy Ancient-forest plant species in western Belgium: a species list and possible ecological mechanisms. Belgian Journal of Botany 130: Kolb, A., and M. Diekmann Effects of life-history traits on responses of plant species to forest fragmentation. Conservation Biology 19: Lindbladh, M., R. Bradshaw, and B. H. Holmqvist Pattern and process in south Swedish forests during the last 3000 years, sensed at stand and regional scales. Journal of Ecology 88: Lindborg, R., and O. Eriksson Historical landscape connectivity affects present plant species diversity. Ecology 85: Lindenmayer, D., and J. F. Franklin Conserving forest biodiversity a comprehensive multiscaled approach. Island press, Washington, DC, USA. Lindenmayer, D. B., J. F. Franklin, and J. Fischer General management principles and a checklist of strategies to guide forest biodiversity conservation. Biological conservation 131: Margules, C. R., and R. L. Pressey Systematic conservation planning. Nature 405: Meffe, G. K., and C. R. Carroll, editors Principles of conservation biology. Sinauer associates, USA. Mitchell, F. J. G How open were European primeval forests? Hypothesis testing using paleoecological data. Journal of Ecology 93: SEPA & NBF (Swedish Environmental Protection Agency and National Board of Forestry) National strategy for the formal protection of forest. Swedish Environmental Protection Agency, Stockholm. Nilsson, S. G Biodiversity over the last thousand years in the cultural landscape of southernmost Sweden (in Swedish with English summary). Svensk Botanisk Tidskrift 91: Nilsson, S. G., U. Arup, R. Baranowski, and S. Ekman Tree dependent Lichens and beetles as indicators in conservation forests. Conservation Biology 9: Nilsson, S. G., M. Niklasson, J. Hedin, P. Eliasson, and H. Ljungberg Biodiversity and sustainable forestry in changing landscapes Principles and southern Sweden as an example. Journal of Sustainable Forestry 21:11 43 (available online at Nitare, J Indicator species for assessing the nature conservation value of woodland sites a Flora of selected cryptogams (in Swedish with English summary. Skogsstyrelsen, Jönköping. Nitare, J., and M. Norén Woodland key habitats of rare and endangered species will be mapped in a new project of the Swedish National Board of Forestry (in Swedish; English summary). Svensk Botanisk Tidskrift 86: Nordén, B., and T. Appelqvist Conceptual problems of ecological continuity and its bioindicators. Biodiversity and Conservation 10: Norén, M., B. Hultgren, J. Nitare, and I. Bergengren Instruktion för datainsamling vid inventering av nyckelbiotoper (Directions for data collection in Key Habitat Inventories). Swedish National Board of Forestry, Jönköping, Sweden. Ovaskainen, O., and I. Hanski Transient Dynamics in Metapopulation Response to Perturbation. Theoretical Population Biology 61: Peterken, G. F Natural woodland Ecology and Conservation in Northern Temperate regions. Cambridge University Press, Cambridge. Pihlgren Red-listed lichens on Quercus robur trees at Biskops-Arnö. Page 32. Institutionen för naturvårdsbiologi. Uppsala university, Uppsala. Ranius, T., K. Antonsson, N. Jansson, and J. Johannesson Inventering och skötsel av gamla ekar i Eklandskapet söder om Linköping. Fauna och Flora 96: Rose, F The epiphytes of oak. Pages in M. G. Morris, and F. H. Perring, editors. The British oak. Its history and natural history. Rose, F Temperate forest management: Its effects on bryophyte and lichen floras and habitats. Pages in J. W. Bates, and 27

28 A. M. Farmar, editors. Bryophytes and lichens in changing environment. Clarendon Press, Oxford, UK. Röstell, Å Effects of stand continuity on biodiversity of seven organism groups in temperate deciduous forest. Page 22. Department of Plant and Environmental Sciences. Göteborg University, Sweden, Götebor. Saunders, D. A., R. J. Hobbs, and C. R. Margules Biological consequences of ecosystem fragmentation: A review. Conservation Biology 5: Sverdrup-Thygeson, A., and D. B. Lindenmayer Ecological continuity and assumed indicator fungi in boreal forest: The importance of the landscape matrix. Forest Ecology and Management 174: Thor, G Red-listed lichens in Sweden: habitats, threats, protection, and indicator value in boreal coniferous forests. Biodiversity and Conservation 7: Tilman, D., R. M. May, C. L. Lehman, and M. A. Nowak Habitat destruction and the extinction debt. Nature 371: Vera, F. W. M Grazing ecology and forest history. CABI Publishing, Oxon, UK. Werth, S., F. Gugerli, R. Holderegger, H. H. Wagner, D. Csencsics, and C. Scheideregger Landscape level gene flow in Lobaria pulmonaria, an epiphytic lichen. Molecular Ecology 16: Werth, S., H. H. Wagner, F. Gugerli, R. Holderegger, D. Csensics, J. M. Kalwij, and C. Scheidegger Quantifying dispersal and establishment limitation in a population of an epiphytic lichen. Ecology 87: With, K. A Metapopulation dynamics: Perspectives from landscape ecology. Pages in I. Hanski, and O. E. Gaggiotti, editors. Ecology, genetics and evolution of metapopulations. Elsevier Academic Press, London, UK. Wolters, V., J. Bengtsson, and A. S. Zaitsev Relationship among the species richness of different taxa. Ecology 87: Økland, B., F. Götmark, and B. Nordén. 2008, in press. Oak woodland restoration: testing the effects on biodiversity of mycetophilids in Sweden. Biodiversity and Conservation, DOI: /s Økland, B., F. Götmark, B. Nordén, N. Franc, O. Kurina, and A. Polevoi Regional diversity of mycetophilids (Diptera: Sciaroidea) in Scandinavian oak-dominate forests. Biological Conservation 121:9 20. Öster, M., S. A. O. Cousins, and O. Eriksson Size and heterogeneity rather than landscape context determine plant species richness in semi-natural grasslands. Journal of Vegetation Science 18: Tack/Thanks... Min handledare Björn Nordén, för all den kunskap om lavar, svampar, mossor, kärlväxter och fåglar som han delat med sig av under min grundutbildning, på gemensamma exkursioner, och under mitt examensarbete (som handlade om kärnsvampar) där han också var min handledare, och under doktorandarbetet. Han har stöttat mig i mitt doktorandarbete med kunskap, och många intressanta naturvetenskapliga och filosofiska diskussioner. Jag har också uppskattat att han vid strategiska tillfällen alltid påminnt mig om att man måste bli färdig också, och inte bara leta efter mer kunskap och ännu bättre sätt att hantera redan bearbetade data. Min handledare Frank Götmark, som har kompletterat Björn på ett fantastiskt sätt. Jag har uppskattat hans stora entusiasm för vårt gemensamma Ekprojekt, där han varit drivande. Jag har uppskattat den gemenskap Frank, Björn, Niklas Franc och jag haft (och ibland även Ted von Proschwitz och Bjørn Økland), med gemensamma luncher, möten, fältarbeten, manusläsning och andra kontakter. Det har varit spännande att jobba ihop med Frank för hans forskningsplanering är oftast preliminär och öppen för förändringar, och där jag efter kloka förslag från honom och Björn ibland fått tänka om, vilket inte alltid varit helt enkelt. Min handledare Ulf Molau för allmänna diskussioner och för granskning av manuskript. Niklas Franc som doktorerat nyligen inom vårt Ekprojekt, och alla examensarbetare inom samma projekt (Mattias Lindholm, Bettina 28

29 Olausson, Martin Ryberg (som numera är doktorand), Johan Dahlberg, Camilla Niklasson, Ingela Sandberg, Åsa Pålsson, Åsa Berglund, Anna Bergqvist, Daniel Johansson, Åsa Röstell, Therese Ludvigsson, Ingrid Thomasson, Anna Nordberg) och assistenter (Andreas Karlsson, Kia Jungbark, Ellen Rube, Anders Malmsten, Elin Götmark, Therese Helgesson) som hjälpt till vid inventeringar och bearbetning av data under min doktorandtid, och/eller bidragit till att höja stämningen på våra gemensamma inventeringsrundor med många gemensamma middagar och vandrarhemsnätter. Paul Lazar, min mentor som delat med sig av sitt vetenskapsfilosofiska perspektiv på forskning och diskuterat forskning på ett vidare plan på ett sätt som hjälpt mig att strukturera och fokusera mitt skrivande, inte minst i slutfasen av avhandlingsskrivandet. Föreningen Fältbiologerna och alla de aktiva fältbiologer, som inspirerat och delat med mig härliga naturupplevelser, kunskaper om växter och djur och som smittat av sig sitt engagemang för att bevara de hotade arter som numera har svårt att överleva i de små fragment av natur som finns kvar. I denna förening grodde det intresse som så småningom ledde mig till detta doktorandarbete. Mitt intresse har även stötts av ett flertal andra föreningar, bl a Mossornas Vänner, och Svensk Lichenologisk Förening, efter att jag lämnat min aktiva roll i Fältbiologerna. Stiftelsen Pro Natura med Leif Andersson, Thomas Appelqvist, Tomas Fasth, Mattias Lindholm, Ola Bengtsson, Mikael Finsberg, Artur Larsson, Richard Gimdal, Camilla Finsberg, Vikki Forbes, Bettina Olausson och Martin Ryberg, som jag fått äran att arbeta ihop med innan jag började doktorera (och lite parallellt med doktorandstudierna också), och där jag förvärvat en stor del av min kunskap om naturinventeringar, praktisk naturvård, arter, signalarter, rödlistade arter, GIS, databaser etc. som jag haft stor nytta av som doktorand, inte minst i undervisningen. Statistikern Kerstin Wiklander som tålmodigt diskuterat statistiska spörsmål med mig. Tomas Hallingbäck som ställde upp med sin fältkunskap om mossor (speciellt fältbestämning av gräsmossor) genom att under två dagar åka runt med mig och examensarbetaren Johan Dahlberg till våra inventeringsytor. Ett stort tack för framsidesbilden av mossan Trädporella som Tomas tagit. Leif Stridvall, som delat med sig av sina fotografier på kärlväxter, svampar och lavar till mina powerpoint-presentationer, postrar och nu till framsidesbilder på avhandlingen. Tack också Anita Stridvall för dina exkursionstips till Hunneberg. Bo Nielsen, som utvecklat Ekprojektets databas där vi samlat fältdata för hela projektet. Tina Granqvist-Pahlén som hjälpt mig med satellitdata, och nyckfulla GIS-program. Atte Moilanen, who supported my research idea in paper I by providing a metapopulation model for the study. Tord Snäll, Otso Ovaskainen, Atte Moilanen, Kjell Wallin and David Ratkowsky for discussing study designs and statistical issues. Jörgen Rudolphi, Thomas Ranius and Robert Björk, for reading early drafts of manuscripts. Kim With and all kind researchers and PhDstudents at Kansas State University, who guided me during my one-week long visit. It was wonderful to follow you for bird watching and fieldwork at the tall-grass prairies with its lovely vascular plant flora, bisons, birds, lizards and the oaks (and other trees) along the small rivers. Lena Gustafsson för den 3 år långa Forskarskolan om skog och biologisk mångfald, som hon anordnade under min doktorandtid och där jag under doktorandkurser och övriga träffar fått många givande kontakter med företrädare för det svenska skogsbruket, journalister, svenska och internationella forskare och inte minst andra doktorander som sysslar med forskning relaterad till naturvård i skogen. Kollegorna på min institution, som jag delat luncher och fikaraster med, men också festat tillsammans med, och alla forskare, doktorander och studenter som jag mött på de otaliga doktorandkurser och andra aktiviteter jag fått deltaga inom och utanför Sverige. 29

30 Doktoranderna i Fakultetets Doktorandråd, ledamöterna i Forskarutbildningsberedningen och Fakultetetsnämnden där vi diskuterat många intressanta viktiga saker. Markägare som ställt upp med sin mark för detta projekt: Anette Karlsson, Göte Isaksson med familj, Bo Karlsson, Anders Heidesjö, Dan Ekblad, Robert Ekman, Nils-Olof och Bengt Lennartsson; Skara och Linköping stift; Sveaskog, Boxholms skogar, Holmen skog; Borås, Jönköpings, Oskarshamns, Växjö kommun; Länsstyrelserna i Östergötlands och Kalmar län. Länsstyrelserna i Västra Götalands, Östergötlands, Kalmar, Jönköpings och Kronobergs län samt Skogsstyrelsen för kartmaterial, flygbilder, information om naturreservat och nyckelbiotoper och annat som behövts till mitt doktorandarbete. Min f d sambo Mats Rosengren, som delat mitt stora naturintresse ända sedan tiden som vi båda två var aktiva fältbiologer. Min kunskap om kärlväxter har jag förvärvat genom och tillsammans med honom på de otaliga naturäventyr vi gjort genom åren. Jag har också uppskattat våra spännande diskussioner vi haft om fåglar, växter, praktisk naturvård och kulturhistoria. Han har delat en stor del av mitt, och våra barns gemensamma vardag under min doktorandtid. Jag har uppskattat det 100 % stöd för mitt doktorerande, inte minst genom att ta hand om barnen när jag haft perioder av fältjobb, doktorandkurser, utlandsresor och hjälpen med layouten av avhandlingen. Mina barn Rasmus och Ola, de är de underbaraste killar som finns i hela världen och världens bästa avkoppling från skrivarbetet. Min nuvarande kärlek Richard Larsen för att han finns och för hans intresse för djur och natur som återinspirerat mig för naturutflykter och lockat mig tillbaka från skrivbordsbiologin som upptagit en allt större andel av mitt naturintresse nuförtiden. Richard har betytt ofattbart mycket för mig det sista halvåret jag skrivit på min avhandling. Slutligen vill jag tacka alla stiftelser och fonder som stött mig ekonomiskt under doktorandtiden: M. C. Cronstedt, L. Hierta, H. Ax:son Johnson, E. and E. Larsson and T. Rignell (Tranemålafonden), C. Stenholm, Adlerbertska, W. and M. Lundgren, K. and A. Binning, P. and M. Berghaus, O. och L. Lamm, P. A. Larsson, T. Krok, H. E. Ahrenberg, och Hierta-Retzius. Illustration: Ingela Sandberg 30