Rules of thumb for metapopulation conservation management

Karin Frank

The deduction of general principles on species’ survival in fragmented landscapes is one of the most important aim of the FRAGLAND project. One way towards some generalization is the deduction of rules of thumb for whole classes of ecological situations (i.e., types of species, types of landscapes). This idea was the basis of a project in which rules of thumb for different aspects of metapopulation conservation management should be derived. (1) By taking a generic, spatially realistic metapopulation model as a basis, minimum conditions on both the species’ ecology and the spatial structure of a considered habitat network could be revealed that have to be met before metapopulation can persist over a long term at all. These results give rise to rules of thumb of the following type "Measures focusing on an improvement of landscape connectivity will only lead to a significant effect on metapopulation survival, if species condition X and landscape condition Y are fulfilled". (2) Rules of thumb for an optimum habitat network design could be derived. These rules particularly reveal how a given total area of habitat ought to be distributed over the patches of a habitat network with a certain patch configuration to maximize metapopulation viability. There is some optimum allocation share that is completely determined by the spatial configuration of the patches. Although the optimum itself can not be ensured in most cases, it gives some orientation for the management (e.g., what patches should have the priority for area protection; where should scarce resources be invested to maximize the ecological effect). The rules of thumb concerning the optimum allocation share provided the basis for an interdisciplinary project together with economists and sociologists from the Department of Ecological Economics of the UFZ Leipzig on the improvement of the ecological effectiveness and social acceptance of economic instruments in nature conservation.

Karin Frank

Spatial heterogeneity in metapopulations resulting from variation in patch area or patch distance makes every spatially realistic metapopulation model quite complex, especially when stochasticity is included. This particularly impedes the deduction of general results on metapopulation viability. This problem could be widely solved. For the first time, a closed formula for the mean lifetime of metapopulations in heterogeneous landscapes could be derived. This formula reveals how spatial and species’ ecological data have to be combined to predict the viability of a considered metapopulation without having to resort to a model. The formula also provides some "landscape index" for comparing and ranking structurally dissimilar habitat networks according to their effect on metapopulation viability. The major advantage of the formula is that its relatively simple functional structure gives insight into the interplay between spatial structure, stochasticity and metapopulation survival.

For deriving the formula, a special method of model simplification has been developed. This method is mainly based on aggregation techniques that are common in Physics, but only seldom used in ecological modelling. Aggregation techniques allow the gap between model simplicity and closeness to reality to be bridged to the benefit of a better understanding.

There are a lot of implications for the FRAGLAND context: (1) In order to demonstrate that the formula is also applicable to "real" systems, it was applied to 74 sub-networks of the Melitaea cinxia system on the Aland islands (SW Finland) showing a lot of variation in the number (2-10), patch area and patch distance (the data have been provided by the Helsinki team; thank you for them). The results of this test are very satisfactory. (2) The formula is used in the Decision Analysis co-operation project between Helsinki and Leipzig (see this Website). (3) The formula also gave rise to a further co-operation project between Leipzig and Helsinki on the effect of habitat loss on metapopulation viability (see this Website).

Karin Frank in co-operation with Otso Ovaskainen and Ilkka Hanski (both Helsinki)

The effect of habitat loss on biodiversity is among the most discussed problems of recent ecology. An assessment from the perspective of metapopulation viability is particularly difficult because of spatial effects. In order to handle these spatial effects, the Helsinki as well as the Leipzig team simultaneously and independently of each other developed a landscape measure for comparing dissimilar habitat networks regarding their ability to support a viable metapopulation (Helsinki: Otso Ovaskainen and Ilkka Hanski (deterministic view); Leipzig: Karin Frank and Christian Wissel (stochastic view)). Both measures allow a threshold value for the total amount of habitat needed for long-term metapopulation persistence to be derived. This opens up a new field of collaboration. First steps have already been made during Karin Frank’s exchange visit in Helsinki in September 1999. A comparative analysis showed that both threshold values are structurally close to each other. Furthermore, an extra-effect of spatial heterogeneity under stochastic circumstances has been revealed with consequences for understanding the effect of habitat loss in the context of metapopulation survival. As a next step, both approaches shall be applied to the Melitaea cinxia system on Aland and detailed investigations on the interplay between stochasticity and spatial heterogeneity shall be made.

Larissa Conradt (Leeds) in co-operation with Karin Frank and Christian Wissel

The field data on the movement behaviour of the Meadow Brown butterfly (Maniola jurtina) in Cambridgeshire (experiments by Larissa Conradt) provide the basis for an interesting collaboration in modelling individual movement in fragmented landscapes and assessing its effect from different perspectives of metapopulation dynamics. A first model that has been mainly developed by Larissa Conradt with support from Karin Frank and Christian Wissel is already in progress. It investigates the success-rate and resulting colonization patterns of systematic search strategies of individuals during dispersal in various landscapes. In a second step, the metapopulation-level consequences of these colonization patterns shall be analyzed. In order to reach this goal, Meta-X, the user-friendly computer program for Metapopulation Viability Analysis developed by the Leipzig team (see this website), will be used. By combining both steps, the search strategies under consideration can be ranked according to their overall effect on metapopulation viability and the role of landscape fragmentation can be investigated.

A user friendly computer program for Metapopulation Viability Analysis (MPVA) and its application in conservation, planning, metapopulation research and teaching

Karin Frank, Volker Grimm and Christian Wissel

Metapopulation Viability Analysis represents a standard task for conservation biologists and provides an important basis for decision making in conservation management and landscape planning. But to develop a separate model for each new species or each new landscape is neither useful nor acceptable. For that reason, the Leipzig team together with the OFFIS – Oldenburg Research Institute on Computer Tools and –Systems decided to develop a computer program that enables non-modellers to perform a MPVA in a comfortable way and can be applied to a large range of situations. The result is Meta-X. The "scientific heart" of this program consists of a generic, spatially realistic, stochastic metapopulation model (a time-continuous equivalent to Hanski’s Incidence Function Model) and a standard procedure for determining the viability of a metapopulation from simulation results, both developed by the Leipzig team. In contrast to other MPVA computer tools (e.g., RAMAS, ALEX, or VORTEX), Meta-X directly supports comparative evaluations of alternative scenarios in terms of their effect on different aspects of metapopulation viability (mean lifetime; extinction risk at a certain time; probability of recovery after a landscape manipulation).

Great importance has been attached to make Meta-X really user friendly and easy to use. Amongst other things, a modern Windows user interface is available; all important steps of a metapopulation viability analysis (parameterization, simulation, evaluation, documentation) are supported by wizards that guide the user through the program; a graphical landscape editor allows habitat networks to be created per "mouse click"; the manual contains a special Guided Tour that documents the whole working process with text and figures. Under the supervision of Martin Drechsler and Karin Frank, a student, Marion Höhn, is simulating metapopulation dynamics, using the program META-X. Her results will be compared with results obtained from Karin Frank's metapopulation model and her approximation formula. This will give us a better understanding of the range of applicability of the formula. This is important for the use of the formula in conservation decision analysis. The beta test of the program is already over. The final product will be provided by a professional publisher. Meta-X is especially useful in the FRAGLAND context. It has already been used during the TMR Modelling Course in Leipzig in November 1998 and will be applied in the collaboration with Larissa Conradt (Leeds) on the effect of the movement behaviour of the Brown Argus butterfly (Maniola jurtina) on metapopulation viability (see this website). For further information on Meta-X, see also the Meta-X homepage.

The impact of dispersal distance on the persistence of a metapopulation in a dynamic landscape

Karin Johst

Dynamic landscapes with a high turnover rate of patches can be generated naturally or by man. Extinction risk in such landscapes is increased by the destruction of patches but contrary to habitat loss regeneration of the patches is possible. Efficient dispersal strategies to colonise regenerated patches should be important for long-term persistence in such landscapes. Among the various traits of dispersal the mean distance animals are able to disperse is especially interesting in this respect because it determines the spatial scale on which regenerated patches are available. Using a spatially explicit metapopulation model which includes both local population dynamics and patch turnover we therefore study the effect of the mean dispersal distance on the persistence of a metapopulation in a dynamic landscape.

Karin Johst, Katrin Schöps

Field observations showed that the monophagous, flightless weevil Hadramphus spinipennis causes frequent local extinction of its host plant, Aciphylla dieffenbachii, through overexploitation. The weevils tended to stay in their host plant patch as long as food was available and dispersal was only observed after local extinction of the plant patch. Therefore, it was suggested that long-term persistence of this locally unstable herbivore-plant system is only possible via a metapopulation structure. Using a spatially explicit simulation model we test this hypothesis and search for key factors of long-term persistence. This work is performed in cooperation with Katrin Schoeps, previously from the Helsinki lab.

Rosemary Setchfield

Dispersal is a crucial process influencing the dynamics and persistence of the spatially-structured populations characteristic of fragmented landscapes. Dispersal rates can strongly relate to the isolation and density of individual populations, but there may be an important behavioural component in the decision to leave a patch. If we do not understand how behaviour interacts with habitat and environmental quality, we may lose the ability to predict the fate of metapopulations in a world where ecological processes and land-use strategies are adapting to, for instance, global climate change. There is empirical evidence that habitat quality can influence dispersal between populations. A particular case for host-specific butterflies is the abundance and suitability of available host plants. Rosemary is developing a model to examine the role of host quality, abundance and spatial distribution on the potential emigration rates of female host-specific butterflies. The model is spatially-explicit and individual-based, and determines the number of patch boundary encounters during female searching behaviour prior to oviposition. Initial investigations will focus on the species Melitaea cinxia, and it is planned that the model will be applied to Plebejus argus and other focal species of the FRAGLAND TMR project when data becomes available.

Thomas Stephan

Thomas is working on the development of EXI (user-friendly PVA software) in cooperation with OFFIS Oldenburg. Work on the software is almost finished, it can already be used for PVA’s. Interested users will find a information about EXI and a first program version on the OFFIS‘ web page soon. A handbook has to be written, it should be finished within the next four months. The program is currently being used by a diploma student from the University of Halle/Saale for an analysis on a population of wall lizard Podarcis muralis. Matthias Wichman from the UFZ uses EXI to analyse the viability of a population of tawny eagle aquila rapax situated in the Kalahari. Some results of earlier analyses are given below.

i) Natterjack Toad Bufo calamita

Small populations of natterjack toad observed in the region of Halle/Saale (Saxony-Anhalt, Germany) are actaully at a high risk of extinction, whereas a population near List (Schleswig-Holstein, Germany) has a good chance of long-term survival. Our results indicate that the mortality rates of juveniles and the availability of spawning ground have the greatest influence on a population’s survival. Natterjack toad populations observed in the field follow a strategy with three spawning phases. This strategy supports the survival of the population better than strategies with less phases, which are more likely to result in a complete breeding failure during one season.

ii) Sand Lizard Lacerta agilis

Habitat areas of sand lizard populations observed near Halle/Saale (Saxony-Anhalt, Germany) are too small to guarantee 95% survival probability within 100 years. Our results show that a doubling of actual areas is necessarry to achieve this goal if the habitat quality is optimal, whereas under suboptimal habitat quality the area should be ten times as large.

iii) Great Grey Shrike Lanius excubitor

A population of shrikes has been observed in the district of Neustadt/Aisch (Bavaria, Germany). Different habit quality zones play a decisive role in detrermining the extinction risk. Working with only one zone overestimates the extinction risk because the overwhelming influence of optimal habitats is neglected. Removing one optimal habitat from the investigated site has about the same effect on the extinction risk as removing ten suboptimal habitats. Optimal habitats have to be saved in order to guarantee long-term survival of the population.