Per Angelstam, Grzegorz Mikusifiski, Britt-Inger Ronnback, Anders Ostman, Marius Lazdinis, Jean-Michel Roberge, Wolter Arnberg and Jan Olsson Two-dimensional Gap Analysis: A Tool for Efficient Conservation Planning and Biodiversity Policy Implementation The maintenance of biodiversity by securingrepresentative and well-connected habitat networks in managed land- scapes requires a wise combination of protection, management, and restoration of habitats at several scales. We suggest that the integration of natural and social sciences in the form of "Two-dimensional gap analysis" is an efficient tool for the implementation of biodiversity policies. The tool links biologically relevant "horizontal" ecologicalissues with 'vertical" issues related to institutions and other societal issues. Using forest biodiversityas an example,we illustrate how one can combineecological and institutionalaspects of biodiversity conservation, thus facilitating environmentally sustainable regional development. In particular, we use regionalgap analysis for identification of focal forest types, habitat modelling for ascertaining the functional connectivity of "green infrastructures", as tools for the horizontal gap analysis. For the vertical dimension we suggest how the social sciences can be used for assessing the success in the implementation of biodiversity policies in real landscapes by identifying institutional obstacles while implementing policies. We argue that this interdisciplinary approach could be applied in a whole range of other environments includingother terrestrial biota and aquatic ecosystems where functional habitat connectivity, nonlinear response to habitat loss and a multitude of economic and social interests co-occur in the same landscape. INTRODUCTION Today, it is widely accepted that landscapes should be managed in a sustainable manner, meaning that both production and biodiversity conservation should be considered. The 1992 Convention on Biological Diversity has not only focused international attention on the concept of biodiversity buthas also set expectations thatthe signatory stateswill establish objectives and mechanisms for local implementation. According to these international agreements, EClegislation and national policies,the maintenance of biodiversity is currently an agreed valuein Sweden. Consequently, there is a common vision that biological diversityshould be maintained, and quantitative politicaltargets have been formulated for lakes and waterways, wetlands, forests, cultural, mountain, and urban landscapes (1). In the Swedish Forest policy, the ultimate environmental goal is eveninterpreted as the survival of viablepopulations of all naturally occurring species (2). Maintaining biodiversity requires a wise combination of protected areas,management and restoration to createand maintain representative andfunctionally connected networks of all habitats (3, 4). However, recentanalysesin northern Europe suggest that the existing quality and amount of habitat networks areinsufficient (5). Thetool that hasbeenemployed is gap analysis,which can be defined as the identification of disproportionate scarcity of certain ecological features in a management unit, relative to representation to a larger region surrounding the management unit (6). Using this approach, in Sweden Angelstam andAndersson (7) argued that a combination of conservation, restoration and even re-creation of habitats is neededto accomplish the political goals of biodiversity maintenance in areas with a long land- use history. Similarly, in Estonia, L6hmus et al. (8) identified gapsin the amount of forest of different representative types needed to provide sufficient habitat for viablepopulation of specieswith different habitat affinities. However, the total amount of habitat found available in a regional gap analysis does not take into account the level of functionality of the habitat networks (9). Habitat SuitabilityIndex (HSI) models for specialized and area- demanding focal speciesis an approach, whichtakesfurther the resultsof such regionalgap analyses(4, 10-12). HSI- modelsareusefultools for assessing the current status of the networks of different habitat types,as well as forplanning of restoration and re-creation needed in orderto accomplish the defined policy goals. Unfortunately, complete sets of parameter values for focal species for all habitats are not available to the extentneeded to coverthe variety of habitat types found in whole landscapes. However, for forest taxa the level of knowledgeis rapidly increasing. For example studies of both the quantitative habitat needs of area- demanding specialists(13, 14) and the indicator value of these species (15, 16) are appearing. Angelstamet al. (9) used quantitative knowledge about the requirements of specialized forest birds listed in the EC Birds Directiveto estimate the size of planningunits for the assessment of habitat networks aimed at maintaining biodiversity. The estimated mean minimum size of planning units where suitable habitat dominates the landscape was about 40 000 ha, while in managed landscapes witha minimum amount of habitat the unit size averaged 250 000 ha. By contrast, the size of individual conservation areassuch as woodland key biotopes and protected reserves from which habitat networks canbe built in a managed matrix was 1- 1000 ha.Tomaintain biodiversity, there is hence clearly a need to extend the spatial and temporalscale of management and planning from the stand scale to that of landscapes within large management units. Since the aboveanalyses concern quantitative and spatial dimensionsof the land cover within landscapes, the com- bination of regional gap analysis andanalysis of functionality using HSI-models can be summarized under the single expression "horizontal gap analysis". However, biodiversity conservation, and sustainable development in general, largely dependon the "vertical" dimension from policy to implementation. Defining a social-ecological system is a criticalstep on the pathtowards achieving sustainable development of that system (117, 18). Considering that not all the parts of the world are equally suitablefor use as units of sustainable Ambio Vol. 32 No. 8, Dec. 2003 ? Royal SwedishAcademy of Sciences 2003 527 http://www.ambio.kva.se