Blog by March cover author Williams and Stuart-Smith

Submitted by editor on 6 March 2014.

Making good conservation decisions with limited data: improvements to simple range maps for biodiversity conservation planning

Rob Williams, University of St Andrews (UK) and Oceans Initiative (Canada)

Rick Stuart-Smith, University of Tasmania and Reef Life Survey (Australia)

Photos in this blog are taken by Rick Stuart-Smith, Reef Life Survey


There are a number of different motivations for establishing marine protected areas (MPAs), and deciding where they should be put depends on the motivation. But in the face of rapid environmental change and habitat loss, an argument can be made for prioritizing global conservation efforts by identifying ‘hotspots’ where conservation action can do the most good for the most species. Unfortunately, we lack quantitative data on global distribution and abundance of many marine wildlife species. As a practical shortcut, some efforts have identified species richness hotspots by overlaying range maps for marine species, often based on presence records in global databases such as OBIS (the Ocean Biogeographic Information System) or expert-drawn range maps. These gridded range maps can be layered to give a simple tally of the number of species expected to be found in a given area. It was noted in the terrestrial ecology literature more than a decade ago that this approach can be quite misleading (Gaston and Rodrigues 2003).


In recent months, a series of papers have noted, independently, that areas of high species richness inferred from range maps alone can miss out on aspects of the community that are important for conservation (Stuart-Smith et al. 2013, Williams et al. 2013, Dulvy et al. 2014, Edgar et al. 2014). One study appearing in Ecography that specifically addressed this issue found an inverse relationship between animal density and species richness (Williams et al. 2013). It showed that layering range maps could lead to prioritization of areas at the margins of the range of multiple species, and miss out high-density areas for most species. Simplistic presence/absence range maps essentially reduce a continuous covariate of habitat quality, density, importance or habitat suitability to a binary one. The resulting picture of biodiversity is extremely sensitive to where one places the arbitrary cutoff between “presence” and “absence”. If placed too low, one risks watering down the definition of presence to include marginal habitats with vanishingly small densities. As a result, places where many species’ ranges overlap will look superficially like biodiversity hotspots, but may actually be marginal habitats for multiple species and critical habitats for none.

To add complication, this phenomenon is likely to vary with the life history traits of the species in any particular location. Traits such as body size, mobility/migratory habits and rarity/occurrence patterns create species-specific differences in how effectively range maps characterize true occupancy patterns. We recognize that the clear shortcomings of using range maps for conservation prioritization for highly mobile marine mammals (Williams et al. 2013) represents one end of a continuous spectrum.


What are options for moving forward? The two obvious options are increased collection of appropriate, quantitative data, and improved methods for identifying and accounting for error in the use of range maps. The former would of course be the preferred situation, directing available funding to fill critical data gaps (Kaschner et al. 2012). This task would be more tractable for some taxonomic groups than others, but really requires the development of novel methods and means for maximizing resources. Well-designed citizen science offers a promising way to do this. Despite successes in this front on land, such as in the British breeding bird and butterfly surveys (Buckland et al. 2005), for example, marine programs are generally not as well developed to provide the needed quantitative data. The Reef Life Survey program ( is one of a few exceptions, with quantitative data collected on fishes, invertebrates and macroalgae at > 2000 sites worldwide through the selective inclusion and scientific direction of a team of volunteer SCUBA divers who undertake standardized underwater visual censuses. The Reef Life Survey model is well suited for filling in data gaps in coastal regions, but not for pelagic environments (i.e. areas beyond national jurisdiction).


Efforts to fill in offshore data gaps may benefit from strategic partnerships with wilderness-based tourism, which seeks to find out increasingly remote and untouched habitats. There are now statistical methods to get relative density estimates for pelagic species from dedicated observers on expedition style cruises as platforms of opportunity visiting polar regions (Williams 2006). Coordination of such an approach across larger scales is likely to represent an additional mechanism to fill the large spatial and taxonomic gaps in available data. Ultimately, providing more quantitative data on relative abundance or habitat importance will allow us to reach our conservation targets using more informative, abundance-based biodiversity metrics, rather than simply tallying the number of species.


The funding needed for such citizen science initiatives is only a small proportion of what is being spent on global conservation though, so support of those initiatives which have well-developed methods to collect data of the detail and quality needed would represent an appropriate priority for international conservation that wouldn’t mean large sacrifices to other necessary spending.


Lastly, despite our best efforts to fill spatial and taxonomic gaps, it is inevitable that some form of prediction of biodiversity into un-sampled space will be needed. In some cases, restricting range maps to area of occupancy, rather than extent of occurrence, may chop out a lot of near-zero-density habitat and may be helpful. Better understanding where and when using range maps for conservation prioritization could also be informed by research on the interplay between life history traits and occurrence patterns within species ranges, and by incorporation of species distribution models (also known as habitat modelling, predictive habitat modelling, ecological niche modelling, habitat suitability modelling and essential habitat modelling) into prioritization. We do not advocate any one solution, and indeed, it may be naïve to suggest that there is any one “correct” range map. Our main objective was to provide yet another reminder, from a few different taxonomic groups, that simple, binary (presence/absence) range maps provide a tempting but often misleading shortcut when mapping global biodiversity patterns.



Buckland, S. T. et al. 2005. Monitoring change in biodiversity through composite indices. – Phil. Trans. R. Soc. 360: 243–254.

Dulvy, N. K. et al. 2014. Extinction risk and conservation of the world's sharks and rays. – eLife 3.

Edgar, G. J. et al. 2014. Global conservation outcomes depend on marine protected areas with five key features. – Nature advance online publication.

Gaston, K. J. and Rodrigues, A. S. L. 2003. Reserve selection in regions with poor biological data. – Conserv. Biol. 17: 188–195.

Kaschner, K. et al. 2012. Global coverage of cetacean line-transect surveys: status quo, data gaps and future challenges. – PLoS ONE 7: e44075.

Stuart-Smith, R. D. et al. 2013. Integrating abundance and functional traits reveals new global hotspots of fish diversity. – Nature 501: 539–542.

Williams, R. et al. 2006. Modeling distribution and abundance of Antarctic baleen whales using ships of opportunity. – Ecol. Soc. 11.

Williams, R. et al. 2013. Prioritizing global marine mammal habitats using density maps in place of range maps. – Ecography 37: 212–220.