Improved monitoring of biodiversity (by CSIR)

Within the GEOBENE project, the Council for Scientific and Industrial Research (CSIR) in South Africa, aimed to contribute to three objectives:

1) Develop a framework for assessing the benefits of improved earth observation systems
2) Apply this framework to case studies from around the world
3) Participate in global efforts to implement GEOSS vision of biodiversity observation systems

Progress and outputs are reported per objective below

1) A framework for benefit estimation

Although Earth observation systems are currently inadequate, hardly anyone can say what the optimum investment might be. This level of investment would perhaps be a level that would be justified by the value of the resource and the risks of not knowing its true state. In other words, it is the ‘stopping rule’ for further investment in observing systems? Fritz et al (2008) proposed a conceptual framework for conducting such an analysis. It involves quantifying the costs of the observation system (conceptually not a difficult thing to do, but surprisingly seldom reported), and balancing this against the social benefits that are anticipated to accrue from the system.

Figure 1: Benefit chain concept.

The concept looks at incremental changes of costs and benefits with respect to the already existing observing system (e.g., national). A logical causal benefit pathway (in steps if necessary) is established and much of the analysis is semi-quantitative (’is the benefit an order of magnitude greater than the cost’) or qualitative (what is the shape of the cost-benefit curve).

2) Applying the framework

We applied the framework under the biodiversity SBA in a number of case studies. We found in general that the procedure for quantifying the economic benefits of improved environmental information is poorly developed, largely because the logic for assigning value to ecosystems and their services is quite recent. We provide worked examples of the logic outlined in Fritz et al. (2008) in a national scale study and 2 local scale studies in South Africa. We also demonstrate the application with a European example.

South African conservation planning

This case study demonstrated the benefits of replacing commonly available coarse scale global data (the non GEOSS scenario) with finer scale data in conservation decision making. These finer scale data were comparable with those expected from GEOSS and could thus be used to estimate the potential benefits of GEOSS data. We then contrasted the benefits of these data improvements with the costs of the improvements. The example uses conservation planning tools which are used to identify spatially explicit priority areas for conservation action (e.g., land acquisition, land stewardship and management, easements, finer scale planning) and feed into land use decision making processes across the country from local to national scales supported by legislation. These tools require spatially explicit data on the distribution of biodiversity (species, ecosystems), threats facing biodiversity (e.g., land conversion, alien invasive plants) and current conservation efforts. National scale data (1:250 000) were used to conduct a conservation plan which identified broad scale priority areas for national conservation action (the GEOSS scenario). A comparison was made of the outputs of this scenario with the same assessment based on the coarse global scale data (the non GEOSS scenario), in an effort to assess the benefits of improved national scale data. The coarse scale data led to a 9% overestimate of priority areas identified by the national scale data and a 10% underestimate in other areas. Calculating the costs and benefits of these 2 scenarios was complex, but using a proxy of land acquisition costs and comparing this to the estimates of data cost the study found that the improved data cost estimate (200 million Euros) was almost an order of magnitude less than the costs of not having finer scale data (1.2 billion Euros). The study demonstrated that despite high data costs under the GEOSS scenario, the investment in high quality biodiversity inventories at a local scale are a very good conservation investment and help ensure cost efficiency in the implementation of expanded protected areas and their management.

Local scale monitoring of ecosystems and their services

In a study based in the Little Karoo of South Africa (~19 000 km2); a semi-arid, intermontane basin, where vegetation associated with three globally-recognized biodiversity hotspots intersects and intermingles, we applied the benefit chain framework to improved observation data on biodiversity and ecosystem services for determining the condition and changes in biodiversity and ecosystem services. The work was published in Reyers et al. (2009) and presented at the 13th International Symposium on Remote Sensing and the Environment in May 2009.

We used data on land cover, as well as spatially explicit data on biodiversity and the ecosystem services of forage for livestock grazing, water flow regulation, carbon storage, erosion control and tourism. Using information on the ecosystem specific impacts of land cover on biodiversity and each ecosystem service, the study quantified changes in biodiversity and ecosystem services as a result of land cover change in the Little Karoo. Fine scale and accurate land cover data were used (GEOSS scenario) and compared to the results when using a more commonly available coarse scale land cover data base (non GEOSS scenario - Figure 2).

Figure 2: Land cover of the GEOSS & non-GEOSS assessments

The results of the GEOSS scenario assessment indicate that the Little Karoo is currently comprised of 38% natural vegetation cover with another 10% cultivated or urban areas; the remainder is made up of moderately (37%) and severely (14%) degraded areas (Figure 2). The non GEOSS scenario shows the Little Karoo has 93.6% of its areas still covered with natural vegetation and only 5.8% in cultivated or urban areas; the remainder of 0.7% is classified as degraded (with no distinction between severe or moderate levels of degradation). When applied to biodiversity and ecosystem service data the GEOSS scenario demonstrates that erosion control shows the largest declines (44%), followed by forage production, carbon storage and tourism viewsheds (25, 27 and 28% reductions); water flow regulation shows the smallest decline of 18% in potential volume of the sustained flows. The non GEOSS scenario finds < 10% declines in most ecosystem services and a 15% decline in the service of erosion control. Rouget et al. (2006) demonstrate similar difference in the assessment of biodiversity condition with the Biodiversity Intactness Index (BII), providing a score of 65.4% for the GEOSS scenario, while the non GEOSS provides a score of 86.5%.

Figure 3. Changes in ecosystem service supply in the Little Karoo based on GEOSS and non GEOSS scenario databases. Change is reflected as a percentage of the potential supply (nominally that of the pre-colonial period).

Improvements in wildlife census effort

This study is based on the abundance of information in a particular field of earth observation – wildlife census. Not only are the costs well understood, but in the particular context of South Africa, benefits can be reliably quantified as well. This is because a large and well-functioning market exists in South Africa for wildlife. Furthermore, many wildlife-based enterprises count their game stocks as part of their asset register. Therefore, knowing what the stock is has a quantifiable economic benefit.

Wildlife conservation areas in Africa and elsewhere are typically periodically censused by flying a light aircraft carrying several observers at low speed and low altitude over the terrain, counting the animals observed within defined strips on either side. Generally such a census is a partial sample rather than a complete count, in that when the aircraft returns for the next transect, the strips are not contiguous, but have a gap between them. As the sampling effort increases (i.e., the number of transects flown increases), the uncertainty of the estimate decreases asymptotically. In this study we try to determine at what sampling effort the costs of the increased effort are equal to or less than the value delivered by that effort.

The example developed is based on data from the Kruger National Park (KNP), South Africa. This national park covers 18989 km2 and the large herbivores have been counted annually in the way described above from 1998 to the present time. The costs of sampling effort was determined from the costs of aircraft, fuel and maintenance costs, as well as the time of the pilot, biologist-observers and data analysts, which scales with the number of transects flown. The value can be equated with the number of animals that can be confidently claimed to exist in the area, multiplied by their market value per head. For this example, we selected the lower 99% confidence limit as the marker. The ‘marginal value’ of an increment of sampling effort is the increase in value that results from it because a more intense sample has a lower error range, and therefore the number of animals that can be claimed increases. Based on this method we found that the sample effort could be increased 95 times before the increase in costs exceeds the increased value.

We also compared this cost of survey effort with the management costs of animal populations (e.g. culling, translocations and contraception) and the valuing of knowing their numbers more accurately.

Improved management decisions in Europe

GEOSS can play a role in assisting the managers of biodiversity to make better informed decisions about where to act in order to ensure the most effective and efficient use of limited conservation resources. These decisions about where to conserve biodiversity are the concern of the discipline of conservation planning which deals with the identification of priority areas for conservation action based on data on biodiversity, threats facing biodiversity and more recently, the costs of proposed conservation action.

GEOSS can contribute to these decisions by providing more detailed and accurate information on the distribution of biodiversity (species, habitats) and threats facing biodiversity (land cover change). It can also contribute through the provision of data on the costs of conservation, which could include the costs of buying and managing the land or the opportunity costs associated with setting the land aside for conservation rather than some other form of land use. This study was based in Europe where data poor (non GEOSS), data medium and data rich (GEOSS) scenarios were constructed for the region.

Figure 4 demonstrates the difference between the GEOSS and non GEOSS scenarios when it comes to assessing change in ecosystem extent. The GEOSS scenario demonstrates more change and ecosystem loss than the non GEOSS scenario.

Figure 4: Ecosystem change under GEOSS and non GEOSS data scenario assessments

Table of European data scenarios

Data poor Data medium Data rich
Species Species atlas data
Ecosystems WWF ecoregions WWF ecoregions HRU environmental diversity
Land cover GLC (2000) GLC (2000) CORINE 2000
Opportunity costs National level yields in euros per ha National level yields in euros per ha 1km2 yield data from HRUs
Planning unit resolution 50 x 50 km 50 X 50 km 10 x 10 km*

Using the software MARXAN (Ball and Possingham 2000) and the GIS interface CLUZ (Smith 2007) a conservation assessment was conducted for the study region using all 3 data scenarios. For each scenario 1000 000 iterations were run 100 times each to find the best solution to conserving biodiversity while minimising opportunity costs to agriculture. Figure 5 below shows the outputs of the 3 scenarios, using the best solution from 100 runs.

Figure 5: The outputs of conservation assessments on the a) data poor; b) data medium and c) data rich scenarios

When it comes to comparing these solutions using the benefit chain concept 2 main problems arose. The first is the absence of cost data for most of the datasets used. The second is measuring the benefits of these solutions in a useful way. The first challenge remains to be addressed as we attempt to collate cost data – for now we express the costs as estimated order of magnitude. The second challenge we address by comparing the outputs of the conservation plans on 2 axes – one that measures the biodiversity conserved and the other that measures the opportunity costs of each solution. Both axes were calculated using the data rich scenario’s data on biodiversity (HRU surrogates of environmental diversity) and costs (using the 1km2 yield data from FASOM). The figure below (Figure 6) illustrates these 2 axes of benefit assessment (improvements in biodiversity conservation and reductions in opportunity costs). The costs are not yet expressed. However the figure shows the improvements in both biodiversity conservation and reductions in opportunity costs with better data.

Figure 6: Benefit assessment of improved data for conservation decision making

Implementing the GEOSS vision of biodiversity observation systems

Under the GEOBENE umbrella, Dr RJ Scholes of the CSIR became involved in the development and establishment of a Global Biodiversity Observation Systems. Dr Scholes attended the 1st GEO BON meeting, Defining User Needs for a Global Observation System for Biodiversity, held at the GEO Secretariat (Geneva) on 23-25 October 2006. That meeting resulted in general agreement amongst participants to form the Biodiversity Observation Network. He attended the 1st meeting of the GEO BON interim steering committee (Geneva, 14-16 January 2008). He attended the 2nd GEO Biodiversity Observation Network Workshop, Potsdam (8-10 April 2008) where he gave the plenary paper. The meeting led to the acceptance of the GEO BON Concept Document, edited by RJ Scholes and W Turner (or NASA), and an agreed process by which it is to be turned into an Implementation Plan in the course of 2008. Dr Scholes attended a pre-CoP meeting on Biodiversity Observation Systems in Bonn 14-16 May 2008, giving another keynote talk on the subject. As a result of these various activities, the GEOBON was noted in a formal decision of the UNCBD (UNEP/CBD/COP/9/L.19. Monitoring, Assessment and indicators, Follow up to the Millennium Ecosystem Assessment. Draft decisions). A first meeting of the GEOBON steering group took place in Washington DC in January 2009. This process was defended and documented in Scholes et al. (2008) published in Science. At the first steering committee meeting of GEO BON Dr Scholes was appointed the first chair of the newly established steering committee of the Biodiversity Observation Network (GEOBON).

  • Fritz, S., Scholes, R.J., Obersteiner, M., Bouma, J. & Reyers, B. 2008. A conceptual framework for assessing the benefits of a Global Earth Observation System of Systems. IEEE 2 (3): 338-348
  • Reyers, B., P. J. O’Farrell, R. M. Cowling, B. N. Egoh, D. C. Le Maitre and J. H. J. Vlok 2009. Ecosystem services, land-cover change, and stakeholders: finding a sustainable foothold for a semiarid biodiversity hotspot. Ecology and Society 14(1): 38. [online] URL:
  • Scholes, R.J., G. M. Mace,W. Turner, G. N. Geller, N. Jürgens, A. Larigauderie, D. Muchoney, B. A. Walther, H. A. Mooney. 2008. Toward a Global Biodiversity Observing System. Science 321: 1044-1045