Selecting conservation reserves using species-covering models: Adapting the ARC/INFO GIS

Conflicts between human development of the landscape and conservation of biodiversiry will continue to grow. Given this reality, there have been a number of attempts to model the optimal selection of conservation reserve sites such that maximum biodiversity protection can be attained within a limited budget for land acquisition. Here we adapt the Location-Allocation module of ARC/INFO to solve the problem of representing, or covering, as many species as possible in a fixed number of selected reserve sites. Resident ARC/INFO solution routines are applied to an innovative 'logical' network that converts the problem of optimal reserve selection into a problem of optimal facility placement, which the Location-Allocation module can recognize and solve. Use of this unique logical network structure as input to ARC/INFO's internal solvers makes possible, compared to previous methods, a much tighter integration of spatial optimization tools with mapping and database tools, all of which are internal to the GIS and accessed via a menu-driven interface. The main advantage is that users of public domain data (such as the US Gap Analysis data) can conduct their own explorations of possible reserve systems without having to acquire and master optimization packages and reformat model output data for GIS display and post-analysis of solutions. Our sample application uses species data from south-western California. We also present a second major form of species-covering model grounded in the same logical network. This enhanced modcl accommodates weighting of species by their conservation importance, thus allowing reserve systems to be designed around the protection of the most threatened or vulnerable biota.     

Spatial Scale and the Geography of Tobacco Smoking in New Zealand: A Multilevel Perspective

Smoking in New Zealand is more common in deprived areas and in areas with a significant Maori population. Despite its status as a major health problem there has been little work investigating this apparent geography of smoking Data from the 1996 Census is used to construct a multilevel ‘proportions‐as‐responses’ model of smoking prevalence. This enables an exploration of the geography of smoking at different spatial scales. Levels within the model distinguish contextual variation between local authorities, census area units and meshblocks. Particular account is taken of the influence of deprivation and ethnicity on smoking. Results confirm the importance of ethnicity and deprivation and indicate that cross‐level interaction between meshblock and census area unit measures is significant. They also challenge crude stereotypes about the apparent geography of smoking and suggest that, while levels of smoking may be high in parts of North Island, they are less high than might be expected given the socio‐demographic composition of the areas concerned. Conversely, smoking is more prevalent than expected in parts of South Island. The paper notes the health policy implications of these emergent geographies.     

Time flies: the geological record of insects

Two recent databases show that the fossil record of insects is better than commonly supposed, comprising over 1000 families and 5000 genera. Insects (hexapods) first appeared in the Devonian, but by the Permian the number of orders present was similar to that of the present day. However, family data suggest an unsteady increase in diversity during the Phanerozoic, with four main peaks and troughs but no sudden change (supported by genera) at the Cretaceous/Tertiary transition.