© 2008 Valerie Hickey

Made by Serif

                                    Theoretical Background                        Research Application

Theoretical Background

 

If nature constitutes a portfolio of natural resources, and conservation is the sum of all the investments and interventions that together manage that portfolio, then measuring the health of the portfolio is critical. Successful conservation is about saving biodiversity. But very little information exists as to whether this is happening or not (Jenkins et al, 2003). Instead, failure to evaluate has led to the adoption of dogma that can be wrong (Sutherland et al., 2004) and to an unwavering faith in the connection between each silver bullet and a conservation outcome (Ferraro and Pattanayak, 2006).

 

Measurement of the portfolio’s losses and gains is essential for measuring society’s return on its investments – how much loss is avoided per dollar spent – and for providing society a basis to make efficiency trade-offs with other priorities. Rather than focus on the effectiveness of conservation actions however, many assessments that attempt to inform the allocation of conservation funding focus on prioritizing sites for investment, irrespective of which actions, if any, may be successful at the site. For example, the latest matrix for investment effectiveness in Africa is based on prioritizing protected areas based on their irreplaceability and precariousness (Hartley et al., 2007). Nowhere does this matrix include successful biodiversity conservation as a measure of effectiveness; instead, effectiveness is about choosing the right place to protect, not about choosing the right actions. Millions of dollars have been spent mapping hotspots, ecoregions and wild places. Quite a bit less money has been spent adjudicating the relative worth of different interventions in these places.

 

Along with prioritizing sites for conservation, effectiveness has long been measured as the percentage of land under protection. Protected areas (hereafter, parks) now cover well over ten percent of the planet’s land surface (Rodrigues, 2004; WCPA, 2006). They have become a staple of conservation efforts and investments. Yet once again we do not know if these are successful: Have they saved biodiversity?

 

Though parks range from strict wilderness areas to multiple-use zones, most have a biodiversity covenant, even if it is one among many objectives. But even for those parks (IUCN categories I – IV) for which biodiversity conservation is a priority, measuring the status of biodiversity remains elusive. This is the case even in parks in developed countries, where standardization of data, ease of collection and accuracy of analysis also fall prey to the limited dollars and sense that plague parks in developing countries (Olsen et al., 1999; Danielsen et al., 2000). In addition, most efforts that do exist fail to integrate their findings into decision-making structures (Danielsen et al., 2003). And too often measures are used as substitutes for program evaluation, rather than as inputs to it.

 

Parks that do measure their biodiversity portfolio use one of two general methods to assess its status for management purposes: monitoring can be defined as quantitative data sampling which is repeated at certain time intervals; scoring is the application of qualitative sampling protocols to capture expert opinion, again repeated at certain intervals. Both focus on depicting condition assessments and estimating changing status.   Measuring happens at the community and ecosystem level as well as at the population and species level (World Bank, 1998). Perhaps the most basic part of any measuring program is its choice of surrogates to capture trends; in monitoring, indicator species are used as surrogates to estimate environmental health and species abundance and richness for an area; umbrella species are used as surrogates to measure habitat and community health; flagship species are used as surrogates to attract public attention (Caro, 1999). The best surrogates tend to be those that are well known at ecological and taxonomic levels, and those that are easy to monitor (Caro, 1999). In scoring, surrogates usually comprise management and relationship indicators.

 

Species monitoring is done using capture-recapture methods, distance sampling, and harvest models, and takes advantage of design-based or model-based inference (Stevens, 1994; Buckland et al., 2000). Species richness and abundance are key to monitoring biodiversity at any one site. Protocols for measuring richness can vary from being individual-based to sample-based (Gotelli and Colwell, 2001). Overall abundance of a particular species, while a useful indicator, is not always as telling as demographic trends in age structure for example (World Bank, 1998). As with all statistical efforts, several sources of error have been identified as inherent to the approach, including detection error, spatial variation, survey error, and observer error (Olsen et al., 1999; Yoccoz et al., 2001). The various choice of judgment or probability sampling when designing monitoring protocols can also introduce different errors. For example, when sentinel sites are chosen for tracking because of their sensitivity to change, they can introduce bias; while this can affect scientific reckonings of change in an area, it may still be useful for managers as an early warning system for abnormal change (Edwards, 1998). Moreover, the biodiversity in any park will include species and ecological systems that are poorly known, if at all (Parrish et al., 2003), so the choice of surrogate is especially important.

 

Remote sensing and the use of geographical information systems (GIS) comprise another popular form of monitoring. These methods allow measurements of habitat loss, which can be converted into quantitative estimates of biodiversity loss using the species–area relationship (Jaccard, 1912; Cain, 1938; Pimm et al., 1995). Remote sensing and GIS are used for many ends, among them to identify and detail the biophysical characteristics of species’ habitats, predict the distribution of species and spatial variability in species richness, and detect natural and human-caused change at scales ranging from individual landscapes to the entire world (Kerr and Ostrovsky, 2003). As with other forms of monitoring, remote sensing and GIS come with their own share of statistical and practical problems. Atmospheric contamination, the vagaries of weather and the resolution and scale of any satellite image can decrease the accuracy of remote sensing. The species-area relationship is of little value in the case of empty forests (Redford, 1992). Moreover, most animal populations cannot be detected from space. Increases in computational power are driving down the costs of necessary computer hardware while the costs of remote sensing and GIS software are also declining. Nevertheless, these costs are not negligible. Nor are the costs for imagery and other data products (Turner et al., 2003).  

 

Scoring is used ubiquitously to measure the effectiveness of parks, both informally in park reports and more formally in academic journals (Bruner et al., 2001)  and funding circles (Stolten et al., 2003; WWF, 2004). Because of its ease of application – usually asking questions – its costs are small relative to monitoring. So too are its scientific bona fides. Indeed, its reliance on subjective expert and democratic reckonings of effectiveness are rarely tested against quantitative data, and as such open to error. Stakeholder bias, hoped-for rewards and investment incentives are only some of many pitfalls that may introduce bias into the answers, and hence the conclusions.

 

†back

Research Application

 

To measure the success of our management of the planet’s portfolio of natural assets in protected areas, we need to move beyond creating an inventory of those assets to actively analyzing the growth or contraction of the portfolio. This analysis promotes accountability and is a powerful tool for learning (Christensen, 2003). To date there exists no measures against which to gauge the success of interventions and investments meant to conserve biodiversity. But there does exist an administrator’s trap (Campbell, 1969): conservationists have become so aligned with the policy of establishing protected areas that there is little incentive to measure their worth. Moreover, the principle of increase as the sole measure of success is leading to a call for more protected areas, without a concomitant consideration of any evidence of what does and does not work. But are protected areas more robust at conserving biological diversity than other areas? Without such evidence-based conservation, we will forever remain uncertain as to whether we are conserving what we say we are.

 

The purpose of measuring is to help managers make better decisions, to better get to ABC. My research will assess various measures and identify those that are both effective and efficient. My research will also take into account that while data is the primary product of any measure, how it is transformed into information, managed and communicated determines its efficacy in the policy process (Davis, 1993). While identifying this measure can tell us if a park’s biodiversity is succeeding or not, it cannot tell us why. This is where program evaluation of actual actions must be implemented.

Ultimately, for program evaluation to have policy impact, it must marry an ecological measure of output to a particular action and then to a measure of the costs of implementing that action. Cost utility approaches (CUA) have been touted as one route to achieving this. CUA allows the effectiveness of unlike activities to be compared (Hughey et al., 2003). It takes several distinct factors into account in its analysis: species or habitat scarcity; qualitative change in species or habitat status; and, recognition of species charisma (Cullen et al., 2001; Haddock at al., 2007). The conservation output protection year (COPY) measure is one such attempt at CUA (Cullen et al., 2001). While my research does not look at cost factors, its analysis of effective and efficient ecological measures will further the ability to apply cost analysis to these and hence to program evaluation.

 

†back