Your E-mail
Remember me
 Forgotten password
SCALETOOL IntroductionDriversBiodiversityPolicies and managementConnectivity and protected areas
From species traits to dispersal distances Simulation of genetic data Population Viability Sex-biased dispersal Biodiversity scaling Perspectives for landscape scale management Conservation strategies at appropriate spatial scales

Conservation strategies at appropriate spatial scales

Studies and policies aimed at conservation inevitably focus on some particular criteria of interest and some spatial scale. Conservation planning could be improved by considering how the outcomes of interest depend on these choices. We report a literature review and a simulation study designed to explore the scale-dependence of conservation science.

1. At what scales is nature conservation currently studied?

Our literature review showed that different conservation goals are often being addressed at appropriate scales (extents) [Fig. 1a], but with some exceptions. There appears to be a lack of studies addressing genetic diversity at the very local scales where small sub-populations may be at risk of drift and inbreeding, and where microevolution may be occurring, or addressing species richness at country-wide and continental scales where broad patterns in mobile taxa are important for understanding the future of global biodiversity [Fig. 1b]. We also noted that relatively few studies applied conservation planning approaches at scales broader than a local region [Fig. 1b].

Image is missing

2. Do we need different kinds of landscapes for different conservation goals?

Conservation as a multi-faceted problem
Good conservation policies should promote the genetic diversity and viability of populations and the species richness of communities. Policy-makers are also concerned about ecosystem services provided by natural systems, such as crop pollination and pest control by useful insects nesting in grassland habitats. A network of nature reserves, such as Natura 2000, should ideally address all these goals simultaneously.

We draw together the ecological components of the SCALES project by asking if different conservation objectives are closely linked, so that a strategy designed for one will generally do well for others. Clearly some types of habitat are generally better than others, but what about the spatial arrangement of habitat patches in the landscape? We used simulation models to work towards recommendations.

Modelling at multiple scales
If there are general geometrical principles for good conservation planning, they need to be scale-specific. The strategy of investing all resources in a single large patch would clearly be inappropriate at the global scale, and it would be equally naïve to interpret a good regional strategy as a global template. We therefore considered 25 landscape patterns containing diverse patch sizes (Figure 2) and interpreted them as maps at different scales, so that they represented either a 5-km or a 50-km landscape.

Figure 2: Landscape patterns assessed by each model. Note how the patterns differ in the amount of habitat (A - O: 10%, P - Y: 2% of the total area), number and size of patches, and patch shapes and connectivity.

We obtained six ecological simulation models (Table 1) from within and outside the SCALES project and used them to address four conservation objectives. Pollination service provision was assessed by two relatively simple models, while genetic diversity, population persistence and species richness were each assessed by more complex models parameterised for several contrasting types of organisms (plants, insects, mammals and birds). The results thus enable us to explore how different landscapes compare with respect to these objectives, considering a range of organisms (Fig. 3).

Figure 3: Our scheme for assessing multiple conservation goals with simulation models.

Table 1. Models used, with the conservation criteria assessed using each model (shaded cells).
Model Genetic diversity Population viability Species richness Pollination services
Spatial neutral model        
Metacommunity model        
Diffusion model        

Simulation results
The best kinds of landscapes for different conservation goals could be strikingly different. Figure 4 shows how the relative values vary strongly with the spatial scale. At the finer, 5-km scale, if 10% of a landscape could be allocated to habitat, most criteria favoured a "blocky" pattern such as in landscapes E and F - especially for organisms that need a lot of space (e.g. birds and mammals). Indeed, with only 2% habitat cover, the best solution was often to have it all in a single large patch (landscape P or Q). At the coarser, 50-km scale, more-fragmented arrangements (e.g. landscapes L and N) performed better (Figure 4), especially among the scenarios with 10% habitat cover.

Figure 5 suggests some landscape configurations that could provide a good compromise among the different conservation goals. The left-hand panel of Figure 5 (50-km scale, 10% coverage) shows a broad trend in performance across multiple criteria, from generally poorly-performing scenarios in the lower front region (scenarios A-E, with few large patches) to ones that do quite well on all four axes at the upper rear of the figure (scenarios J, K and especially L), which are fairly fragmented. On the other hand, there is less agreement among criteria in the scenarios with 2% cover (right-hand panel of Figure 5).

Figure 4: Relative values of four conservation criteria for contrasting scenarios, as assessed by simulation models for landscapes spanning 5km (top) and 50km (bottom). Results for scenarios with 10% reserve cover are shown on the left, and those with 2% coverage on the right. Each of the four panels shows performances relative to the best modelled scenario for each model. The grey area represents the value if scores from the four criteria are simply averaged.

Figure 5: Scatterplots comparing landscapes against all four criteria simultaneously, for scenarios with 10% habitat cover (left) and 2% reserve cover (right). The letters refer to the landscape patterns (Figure 2) as implemented here at the coarser scale, spanning 50km. The size of each circle indicates genetic diversity, the height of stalks indicates relative species richness and positions on the horizontal axis indicate metapopulation size - all for songbirds, while positions on the receding (diagonal) axis represent insect pollination service. The symbols are coloured such that dark shades indicate high scores on the three axes (rear upper left corner) and pale shades indicate low scores (front lower left). All four variables are expressed relative to the highest values found for each level of cover.

3. Towards policy advice

While the quality of habitat itself, in its local context, is of paramount importance for all conservation goals, our work shows how different spatial configurations of patches may be needed for different conservation objectives, with appropriate consideration of spatial scale. The relative importance given to different objectives is essentially an ethical question that must be addressed politically, not scientifically. However, assuming an answer to that question, we may make some broad conclusions:
  • Conservation studies and simulations should be tailored to the most appropriate spatial scales with respect to relevant ecological processes.
  • The ideal degree of fragmentation will depend on the amount of land that can be protected in a given area.
  • Making a single nature reserve as large as possible is unlikely to be ideal unless a very small area of land is available. There is probably no single ideal patch size.
  • Different groups of species may be conserved by reserves of differing sizes.
It would be premature to make strong recommendations from the results shown above, but ongoing work in this area should lead to some general geometric principles for conservation planning.


Baguette M., Clobert J. & Moulherat S. (2012). Population viability across scales. In: Securing the Conservation of biodiversity across Administrative Levels and spatial, temporal, and Ecological Scales.

Bocedi G. (2010). A general framework for modelling metacommunity dynamics under environmental changes. In: School of Biological Sciences. University of Aberdeen Aberdeen, p. 51pp. Natural Capital Project (2012). Natural Capital Project. URL: http://www.naturalcapitalproject.org

Pe'er G., Henle K., Dislich C. & Frank K. (2011). Breaking functional connectivity into components: a novel approach using an individual-based model, and first outcomes. PloS one, 6, e22355.

Rosindell J. & Cornell S.J. (2007). Species-area relationships from a spatially explicit neutral model in an infinite landscape. Ecology Letters, 10, 586-595.

Copyright and disclaimer: SCALES and SCALETOOL

CONDITIONS OF USE: We explicitly encourage the use of SCALETOOL. SCALETOOL is freely available for non-commercial use provided you acknowledge SCALES as source. For more extensive access to databases (e.g. for statistical analyses or if you want to contribute data), tools, or background material, please contact the SCALES coordinator (send us email).


© 2010 - 2018 SCALES. All rights reserved.