• Open Access

Linking site and regional scales of biodiversity assessment for delivery of conservation incentive payments

Authors


  • Editor
    Andrew Knight

Correspondence
Sue Briggs, CSIRO Sustainable Ecosystems, GPO Box 284, Canberra 2601, Australia. Tel: +61 2 6242 1621; fax: +61 2 6242 1555. E-mail: sue.briggs@canberra.edu.au

Abstract

We compared seven scenarios for selecting incentive proposals, and evaluated their contribution to regional conservation outcomes. Scenarios ranged from strategic site assessment where proposals were ranked on scores from site-scale data and a priori regional value to systematic regional assessment with selection of proposals using optimization algorithms. The selection scenarios were simulated using proposals from a conservation investment program in South-eastern Australia. All selection scenarios provided a positive gain in biodiversity compared with selecting no proposals. Regional assessments provided ∼4% gain in the biodiversity index over site assessments. The performance of scenarios using optimization was similar to scenarios using rankings. A framework for integrating site and regional approaches for selecting proposals for conservation management is proposed, where proposals are assessed on their contribution to regional biodiversity outcomes rather than just on the comparative gain in biodiversity at the site, and regional conservation priorities are updated between rounds of incentive programs.

Introduction

There is an extensive literature on conservation assessment and planning (Pressey et al. 1993; Margules & Pressey 2000; Sarkar et al. 2006; Hajkowicz et al. 2007; Moilanen et al. 2009), much of which details methods for identifying priorities for conservation actions and ways of selecting areas for conservation when resources are limited. There is increasing discussion around embracing opportunism in conservation planning while being mindful of the need to base investment on systematic assessment (Knight & Cowling 2007, 2008; Pressey & Bottrill 2008). Informed opportunism is where decisions to make conservation investments when opportunities arise are informed by a priori analysis of regional priorities (Noss et al. 2002; Knight & Cowling 2008; Pressey & Bottrill 2008).

Informed opportunism can play an important role in conservation investment on private land. In such situations, conservation assessments based only on regional-scale spatial data (hereafter called “systematic regional assessment”) might be of limited value if (1) the areas they identify as high priority are not available for conservation actions, that is, opportunity and priority status do not coincide (Cowling et al. 2009), or (2) important elements of biodiversity, for example, site condition, are not well described by regional-scale data. An alternative approach is to assess biodiversity attributes at sites in the context of conservation priorities identified through systematic regional assessment. Combining site attributes with regional priorities places local data in a strategic regional framework to assess conservation opportunities in situ when they arise. This approach hereafter called “strategic site assessment” can be used to select proposals in tender programs where private landowners submit bids to undertake conservation actions for financial incentives (Stoneham et al. 2003; Windle et al. 2009).

In this study, we compared the effectiveness of systematic regional assessment and strategic site assessment for selecting proposals for conservation investment using data from a conservation incentive program in South-eastern Australia. We provide a conceptual framework for combining systematic regional assessment and strategic site assessment to facilitate informed opportunism in conservation investment.

Methods

Study area and data

Site and regional assessment methods were compared using data collected by the Southern Rivers Catchment Management Authority for a tender program for conservation management on rural land in the coastal plains of South-eastern Australia (Figure 1). The tender program invited private landowners to submit proposals to undertake conservation actions for financial incentives. We used data from the incentive proposals in the Southern Rivers Catchment Management Authority for the analyses in the study; we did not analyze the tender process per se. The study area was approximately 714 km2; freehold land covered 78% of the region, and 17% of the region was in reserves or under other conservation management (Table 1).

Figure 1.

Map of study area showing proposal sites.

Table 1.  Tenures and areas in the study area and proposal sites
TenureArea (ha)% study regionArea in proposal sites (ha)% proposal sites
National Park1108015.50.10.03
State Forest–native production forest1245 1.700
State Forest–timber reserve15  0.0200
Sate Forest–other953 1.300
Crown land—lease114  0.1600
Crown land—reserve812 1.100
Crown land—other1529 2.100
Freehold–property vegetation agreement143 0.212 5.7
Freehold–voluntary conservation agreement20  0.0319 9.2
Freehold–other5547377.717885.1
Total71383100208.6100

The assessments used surrogates for biodiversity at different scales (McElhinny et al. 2006; Gibbons et al. 2008, 2009; Ferrier & Drielsma 2010; Appendix A). Biodiversity data for the regional assessments were obtained from vegetation maps (Tozer et al. in press), supplemented by site data for some scenarios (Table 2). The biodiversity data for the site assessments were collected at the proposal sites by Catchment Management Authority staff in 2006 and 2007. Data from 30 incentive proposals ranging from 0.2 ha to 42.1 ha (median 3.8 ha) and totaling 208 ha (0.29% of the study region) were used in this study. Proposals covered six vegetation communities: South Coast Grassy Woodland (60% cleared), Illawarra Gully Wet Forest (45% cleared), Southern Lowland Wet Forest (10% cleared), Warm Temperate Layered Forest (45% cleared), Subtropical Dry Rainforest (80% cleared), and Subtropical Complex Rainforest (25% cleared) (Tozer et al., in press).

Table 2.  Selection scenarios and assessment procedures for selecting proposals for the scenarios. See text and Appendix A for further details
Selection scenarioDescription
Reference scenarios
 None selected scenarioNo proposals selected for conservation management. Regional biodiversity index calculated without the proposals by regional assessment
 All selected scenarioAll 30 proposals selected for conservation management. Regional biodiversity index (upgraded with site data for the proposal sites) calculated with all proposals included by regional assessment
Site assessment scenarios
 Site assessment scenario–rank selectedSite assessment. Proposals selected on ranked site-biodiversity scores, that is, (site score per ha x proposal area)/cost
 Site assessment + landscape measure scenario–rank selectedSite assessment with enhanced neighborhood habitat measure for landscape value. Proposals selected on ranked site-biodiversity scores (incorporating enhanced landscape measure), that is, (site score with landscape measure per ha x proposal area)/cost
 Site assessment + landscape measure scenario–optimization selectedSite assessment with enhanced neighborhood habitat measure for landscape value. Proposals selected by optimization of the regional biodiversity index using site-biodiversity scores (incorporating enhanced landscape measure) for the set of proposals within the available funds
Regional assessment scenarios
 Regional assessment scenario–rank selectedRegional assessment. Proposals selected on ranked improvement in regional biodiversity indices, i.e., (improvement in biodiversity index for the proposal)/cost
 Regional assessment scenario–optimization selectedRegional assessment. Proposals selected by optimization of the regional biodiversity index for the set of proposals within the available funds
 Regional assessment + site data scenario– optimization selectedRegional assessment upgraded with site data. Proposals selected by optimization of the regional biodiversity index (upgraded with site data for the proposal sites) for the set of proposals within the available funds
 Regional assessment + site data + landscape measure scenario– optimization selectedRegional assessment upgraded with site data and enhanced neighborhood habitat measure for the site. Proposals selected by optimization of the regional biodiversity index (upgraded with site data and enhanced landscape measure for the proposal sites) for the set of proposals within the available funds

Selection scenarios

A total of nine scenarios was run (Table 2). We simulated regional conservation outcomes from seven scenarios for selecting incentive proposals for conservation management, comprising three scenarios using strategic site assessment and four scenarios using systematic regional assessment (Table 2). We also calculated regional conservation outcomes when no proposals were selected for conservation management and when all proposals were selected (Table 2). We compared scenarios with proposals selected by ranked scores with those selected by optimization.

We simplified the financial aspects of the incentive proposals to focus on comparisons between the selection scenarios. We allocated a fixed cost of implementing conservation management of $1,000 per hectare for all proposals. This figure was based on incentive costs of $100 per hectare per year over 10 years (Briggs 2009). Thirty incentive proposals were included in our analysis (Appendix B). The overall budget was set at $125,000, or 60% of the total cost of all proposals, to select realistic numbers of proposals for testing the scenarios, before funds were exhausted. Sixty percent of the total cost was just above the mean proportion of proposals funded in other studies (Stoneham et al. 2003; Gole et al. 2005; Crowe et al. 2008).

Strategic site assessment

Strategic site assessment was undertaken using the incentive component of the BioMetric biodiversity assessment tool (Gibbons et al. 2005). BioMetric combines regional-, landscape-, and site-scale measures of biodiversity to provide a site-biodiversity score (Gibbons et al. 2005; Gibbons et al. 2009; Appendix A). The regional-scale measure in the site-biodiversity score provides regional context for site assessment. Management that improves site-scale condition and landscape-scale connectivity of native vegetation, especially of heavily cleared vegetation communities, improves the site-biodiversity score. The site-biodiversity score with the proposed conservation management actions was calculated for individual proposals in the site-assessment scenarios.

One of the selection scenarios for site assessment used an enhanced landscape measure, namely neighborhood habitat area from Drielsma et al. (2007), instead of landscape value from Gibbons et al. (2009). Neighborhood habitat area is a measure of the connectivity of the proposal site (see Appendix A).

Systematic regional assessment

Regional assessment was undertaken using the Biodiversity Forecasting Toolkit, which is a spatial analysis tool that calculates regional biodiversity outcomes using regional- and landscape-scale measures of biodiversity (New South Wales Department of Environment and Conservation 2006; Appendix A). The Biodiversity Forecaster calculates a regional biodiversity index from “effective habitat area,” which represents the status of biodiversity in the region as a proportion of the pristine biodiversity status of the region (Westaway et al. 2004; New South Wales Department of Environment and Conservation 2006; Ferrier & Drielsma 2010; Appendix A). Management that improves condition and connectivity of native vegetation, especially of heavily cleared or fragmented vegetation communities, improves the biodiversity index. The regional biodiversity index was calculated with and without the proposed conservation management actions at the proposal sites to measure improvement in biodiversity at regional scale (Appendix A).

Selection by ranked benefit-cost

Proposals were ranked on their benefit-cost ratios to provide the sets of proposals for the ranked selection scenarios. Benefit-cost for each proposal was calculated as the site-biodiversity score or improvement in the biodiversity index with conservation actions at the proposal site, divided by the financial cost of the proposal. Proposals were selected from greatest to least benefit-cost ratio until the available budget was exhausted. Where the inclusion of the proposal with the next lowest benefit-cost resulted in the total cost exceeding the available budget the proposal was excluded, and successive proposals with a total cost less than or equal to the remaining budget were included.

Selection by optimization

Optimization using a genetic algorithm was applied to four of the scenarios (Table 2) to find the optimal or near-optimal set of proposals that provided the greatest improvement in biodiversity within the budgetary constraints (Moilanen & Cabeza 2002). The algorithm assessed the contribution of each combination of proposals (with their proposed conservation actions) to improvement in the regional biodiversity index, within the available budget. A tournament-selection technique (Miller & Goldberg 1995) was applied to select the set of proposals that gave the greatest improvement in the regional biodiversity index within the available budget.

The improvement in the biodiversity index measured the regional conservation outcome for the proposal or set of proposals. The contributions of the selection scenarios (Table 2) to regional conservation outcomes were compared by calculating the difference between the regional biodiversity index for the sets of proposals selected by the scenario for conservation management and the regional biodiversity index when no proposals were selected for conservation management. The differences between the regional biodiversity index for the sets of proposals selected by the scenario for conservation management and the regional biodiversity index where all proposals were selected for conservation management (Table 2) were also calculated for the scenarios.

Results

All of the selection scenarios provided a higher overall biodiversity index than the “None selected” scenario, noting that the magnitudes of the differences were small, with a mean improvement in the assessment scenarios of 0.000547 over the “None selected” scenario (Table 3). All the selection scenarios achieved more than 85% of the maximum possible gain in the biodiversity index, that is, the gain if funds were not limited and all 30 scenarios were selected for management. The greatest gain in the biodiversity index (89.2% of maximum possible under the “All selected” scenario) was under the rank selected, regional assessment scenario (Regional–ranked; 24 proposals) and the lowest gain (85.5%) was under the rank selected, site assessment with enhanced landscape score scenario (Site + landscape–ranked; 24 proposals) (Figure 2, Table 3). Site assessment with rank selection (Site–ranked) was the simplest selection scenario and achieved 86.4% of the maximum gain in the biodiversity index with 21 of the 30 proposals selected (Figure 2, Table 3).

Table 3.  Biodiversity indices for each selection scenario. “None selected” is where no proposals were selected for conservation management. “All selected” is where all 30 proposals were selected for conservation management. The biodiversity index measures the regional conservation outcome from each scenario, and is scored between zero and one. See Table 2, Appendix A, Appendix B, and text for further details
Selection scenarioBiodiversity indexNumber of proposals selected
Reference scenarios
 None selected0.854962 0
 All selected0.85558830
Site assessment
 Site assessment–rank selected0.85550321
 Site assessment + landscape measure–rank selected0.85549724
 Site assessment + landscape measure–optimization selected0.85550321
Regional assessment
 Regional assessment–rank selected0.85552024
 Regional assessment–optimization selected0.85551217
 Regional assessment + site data–optimization selected0.85551417
 Regional assessment + site data + landscape measure–optimization selected0.85551421
Figure 2.

Gain in the biodiversity index under each selection scenario as a percentage of the maximum gain in the biodiversity index, that is, the gain in the index calculated using regional assessment where all 30 proposal sites were selected for conservation management. The gain in the biodiversity index is measured as: (biodiversity index for the selection scenario−biodiversity index for none selected)/(biodiversity index for all selected−biodiversity index for none selected). Note the scale. See Table 2 for explanations of scenarios and Table 3 for the biodiversity indices for the selection scenarios.

The regional-assessment scenarios performed slightly better than the site-assessment scenarios, with a mean gain in the biodiversity index of 88.4% of the maximum possible under the “All selected” scenario compared with a mean increase of 86.1% for the site assessments (Figure 2). There was little overall difference between the performance of scenarios using rankings (87.0% of gain in the biodiversity index with “All selected”) compared with using optimization (87.7% of gain in the biodiversity index with “All selected”) (Figure 2).

There was a strong relationship (R2= 0.86) between rankings of proposals using site assessment and rankings using regional assessment (Figure 3). Notwithstanding this, site assessment with ranked selection performed less well than regional assessment with ranked selection (Figure 2).

Figure 3.

Relationship between rankings of proposals using their site-biodiversity scores from strategic site assessment and rankings of proposals using their regional biodiversity indices from systematic regional assessment. Further details are in Appendix B.

Nine proposals (30%) were selected by all seven selection scenarios, four were selected by only one scenario, and one proposal was not selected by any of the scenarios (Appendix B). The nine proposals selected by all selection scenarios had higher site scores (mean = 32.1, SE = 1.5) than the site scores for all proposals (mean = 28.2, SE = 1) and higher site scores than proposals selected once or not at all (mean = 21.7, SE = 0.7); and they had higher improvements in the biodiversity index (mean = 0.0041, SE = 0.0024 compared with 0.0017, SE = 0.0008 for all proposals and 0.0009, SE = 0.0004 for proposals selected once or not at all) (Appendix B).

Discussion

This study has highlighted several salient points for conservation investment. Regional assessment performed better than site assessment as measured by overall improvement in regional biodiversity status, but the difference was slight (less than 4%). Both approaches achieved over 85% of the maximum conservation gain for 60% of the total cost of all proposals. There was no difference between the outcomes of selection scenarios with proposals ranked according to individual scores versus the more time-consuming and computationally complex, optimization.

The rankings of proposals using site and regional approaches were highly correlated. The site approach we used, BioMetric, assesses biodiversity at sites in a regional context (Gibbons et al. 2005; Gibbons et al. 2009). BioMetric draws on similar information (percent cleared of vegetation communities) as the regional approach in the Biodiversity Forecaster, which considers the remaining effective habitat area (vegetation community, vegetation condition, and fragmentation) (New South Wales Department of Environment and Conservation 2006). If a regional approach had been compared with a site approach without regional information, there would be a much bigger difference in their performances. This similarity between the two types of assessment provides the opportunity for integration between strategic site assessment and systematic regional assessment (see below).

Optimization did not produce improved biodiversity outcomes compared with ranked selection scenarios. In our study, the size and number of the proposals was small compared with the scale of the region (0.29% of the area), and the proposals were dispersed. The effects of individual proposals on regional priorities probably were not large enough to alter the priorities much during optimization, and thus the priorities remained the same as additional proposals were selected. There was little opportunity for complementarity to operate (Margules & Pressey 2000), as additional proposals did not affect priorities for regional conservation very much because most proposals were small and there were few proposals overall. In addition, many of the proposals were dispersed and did not complement each other spatially, that is, they did not provide reciprocal benefits for each other. Consequently, there was little need for optimization that selects proposals that complement each other, and optimization provided little or no improvement on ranking proposals to select those with higher benefit-cost ratios until the budget was exhausted.

The results of this study, showing lack of difference in conservation outcomes between selecting proposals by strategic site assessment or systematic regional assessment, or by ranking or optimization, were similar to the results in Gole et al. (2005). Using data from a tender scheme in the wheatbelt of Western Australia, these authors found little difference in conservation outcome between site assessment using ranking compared with systematic planning with optimization. Gole et al. (2005) attributed the lack of difference between the two approaches to the small number and area of proposals in their study, and the large area of land required to achieve regional conservation goals.

Optimization has an advantage where the proposals modify regional priorities as objectives are achieved and/or where one proposal affects and is affected by other proposals (see Hajkowicz et al. 2007). These do not occur where the proposals cover a small area of the region and are dispersed, and hence priorities do not change as more proposals are funded (Gole et al. 2005). In these circumstances, ranking provides equally good conservation outcomes as optimization, because complementarity does not operate and proposals do not complement each other spatially. If the area of proposals increases and approaches the objective of including all vegetation types, optimization will select proposals with additional vegetation communities that require conservation management, and proposals that complement each other spatially. Conservation of larger and better connected remnants proportionally increases as incentive funding and area of land under conservation management increases (Chomitz et al. 2006). Optimization is likely to provide a more efficient outcome than ranking where conservation proposals comprise a larger proportion of the region than in the current study (see Gole et al. 2005).

The cost of running assessments for conservation incentive programs is mostly staff time. This study did not analyze the time required to undertake site assessments or run regional and site analyses. This is a fruitful area for future research. The ratio of conservation outcome to cost of staff time to collect field data and undertake analyses is likely to vary as conservation investment in a region increases. In the initial stages of an incentive program, most proposals selected on the ranked benefit-cost ratios of their site scores (including regional context) are likely to improve regional conservation outcomes. As the program proceeds, optimization becomes important to avoid investing in proposals that suboptimally contribute to regional conservation outcomes. Analyses of staff time and conservation outcomes of systematic regional assessment and optimization compared with strategic site assessment and ranking should be undertaken for the different phases of incentive programs.

The information used by strategic site assessment and systematic regional assessment overlaps. This overlap can be exploited to integrate the site and regional approaches by exchanging biodiversity information across spatial scales using existing assessment methods. Integrating conservation actions at site scales with planning at regional scales enhances conservation outcomes by targeting investment at sites that meet regional priorities (Briggs 2001; Knight et al. 2008). There are four potential points of data interchange between site and regional assessment methods: (1) using regional conservation priorities from regional assessment for site assessment, (2) using spatial analytical techniques to assess landscape value for site assessment, (3) upgrading regional spatial data by field validation of vegetation communities and/or condition at proposal sites, and (4) upgrading regional spatial data for areas surrounding proposal sites using field validation of vegetation communities and/or condition.

Site and regional assessments can be run in a nested fashion. An initial analysis to provide regional priorities can be run on regional spatial data, such as mapped vegetation communities and modeled vegetation condition, with an algorithm, such as neighborhood habitat measure, to assess landscape value (Figure 4). Regional conservation priorities and landscape value then provide context for site assessments. This facilitates informed opportunism in conservation investment decisions (Noss et al. 2002; Knight & Cowling 2008; Pressey & Bottrill 2008), and assists with conserving functional landscapes and sites across scales (Poiani et al. 2000). Field verification of data on vegetation communities and/or condition is then incorporated into regional data layers to enhance future regional and site assessments (Figure 4).

Figure 4.

Integrating scales of conservation assessment by linking and exchange of biodiversity data from regional, landscape, and site assessments. The timing of site, landscape, and regional assessments can be tailored to suit the conservation objectives and available resources. For example, an initial regional assessment run on spatial data provides regional conservation priorities using measures such as conservation status of vegetation communities. One or more investment rounds may then be run, comparing management proposals using site-scale assessments informed by the regional priorities. Site data including the type and condition of vegetation communities can be fed back to update the accuracy of regional- and landscape-scale data for subsequent regional assessments, perhaps after a number of investment rounds.

The advantage of using systematic regional assessment to inform site assessment is illustrated by comparing the measure used in site assessment to assign regional value, namely the proportion of original extent of the vegetation community that is cleared, with the proportion of original effective habitat area that remains derived from regional assessment (New South Wales Department of Environment and Conservation 2006). Effective habitat area incorporates pattern of clearing as well as impacts of land use on vegetation condition. Historical clearing patterns often leave some vegetation communities more fragmented than others, which is not reflected in the simple proportion of original extent cleared.

Integrating site and regional scales of biodiversity assessment enhances informed opportunism, because proposals are evaluated on their contribution to regional conservation outcomes rather than simply on their comparative contribution at site scales. Regional conservation priorities can be updated between sequential rounds of incentive programs. Programs run this way accommodate the voluntary nature of conservation opportunities on private land and ensure that funded proposals meet regional conservation priorities.

Acknowledgments

We thank the staff from the Southern Rivers Catchment Management Authority for providing data from their conservation incentives program, CSIRO Sustainable Ecosystems for providing intellectual support, and the referees whose comments helped improve the article. Funding was provided by the NSW Government through its Environmental Trusts and the Australian Government through the Natural Heritage Trust and the National Action Plan for Salinity and Water Quality. The contents of this article do not represent the official policy of the NSW Government, the NSW Department of Environment, Climate Change and Water or any other agency or organization.

Ancillary