SEARCH

SEARCH BY CITATION

Keywords:

  • Costs of conservation;
  • conservation planning;
  • location-allocation;
  • p-median;
  • protected areas;
  • road network

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Solution method
  6. Results
  7. Discussion
  8. Acknowledgments
  9. References
  10. Supporting Information

Conservation organizations that manage networks of protected areas commonly require staff to travel to those areas for management and monitoring purposes. We examine how conservation organizations can reduce the resulting travel costs by locating human resources effectively. Specifically, we focus on the problem of siting the home offices of management staff, in a way that minimizes the travel costs involved. We illustrate the importance of travel cost using two case study applications, the Yorkshire Wildlife Trust (YWT), U.K., and the Northwest Florida Water Management District (NWF), USA. For YWT, siting an additional office effectively could save $43,000 in annual travel costs. Optimally, siting NWF's four existing offices could save $95,000 annually. These savings are sufficient for each organization to acquire 171–360 additional hectares of protected area or to hire an additional protected area manager. We also calculated the reduction in greenhouse gas emissions made possible by optimizing office locations.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Solution method
  6. Results
  7. Discussion
  8. Acknowledgments
  9. References
  10. Supporting Information

Faced with limited funding, conservation organizations must aim to allocate what resources are available to conserve species and ecosystems as effectively as possible. Framing choices between different conservation strategies as optimization problems can enable conservation organizations to identify those choices that will provide the greatest conservation benefit per dollar invested (Groves 2003; Murdoch et al. 2007; Moilanen et al. 2009; Wilson et al. 2009; Cullen 2013). Recently, there have been calls for conservation groups to consider other organizational aspects, such as how they structure their operations, through a similarly strategic lens (Kark et al. 2009; Sutherland et al. 2009; Armsworth et al. 2012). While many authors have focused on where conservation groups should allocate funding for land acquisition or management (e.g., Meyers et al. 2000; Underwood et al. 2008), allocating available human resources to support conservation actions is also important. Here we focus on a particular version of this question—namely, where should a conservation organization locate its staff relative to its protected areas, in order to carry out conservation actions cost effectively?

Spatial prioritization studies in conservation planning have tended to focus on the efficient selection of protected areas (Cullen 2013). In addition to the spatial distribution of biodiversity, recent examples account for factors such as heterogeneous land costs (Ando et al. 1998; Armsworth 2014), heterogeneous threats (Strange et al. 2006; Pressey et al. 2007; Wilson et al. 2007; Wade et al. 2011), the use of different types of conservation activity in different locations (Murdoch et al. 2007; Wilson et al. 2007) and the optimal sequencing of land protection efforts (Costello & Polasky 2004; Strange et al. 2006). Some authors have explicitly considered distances between candidate locations for protected areas and their proximity to other features on the landscape (e.g., Önal & Briers 2002; Williams 2008; Bauer et al. 2009). For example, when choosing protected areas intended to support recreation as well as biodiversity goals, distance to population centers is important (Önal & Yanprechaset 2007). In addition, proximity between protected areas can influence the likelihood that protected species actually persist (Williams 2008; Bauer et al. 2009).

In this article, we take a spatial approach to siting human resources to reduce management-necessitated travel costs. Conservation organizations that are active in land protection commonly maintain a small number of administrative offices where staff are based, and employ a team of site managers who travel from these offices to visit protected areas to conduct management, monitoring and maintenance activities. With this organizational structure, office-to-site travel is necessary but costly in terms of manager time, as well as fuel and maintenance of organization-owned vehicles. While by no means the biggest cost component involved in protecting land, travel costs can nonetheless sum to meaningful amounts when totaled across a protected area network, as we show below. Locating offices and staff to reduce these travel costs would allow some of the funds used for travel to instead be used directly for conservation action (e.g., land acquisition, hiring additional staff).

Like other spatial prioritization studies in conservation planning, we draw on techniques from location science, an active area of operations research (Hale & Moberg 2003; ReVelle & Eiselt 2005; Daskin 2008). Reserve site selection has adapted covering problems from the field of discrete location science to maximize species representation in protected area networks (ReVelle et al. 2002). Similarly, we find the p-median problem to be a well-studied analogue to the efficient siting of conservation staff (Reese 2006; Daskin 2008). The p-median problem places p supply points (e.g., warehouses) into a network to serve n demand nodes (e.g., retail outlets) in a way that minimizes the total distance between each demand node and its nearest supplier (Hakimi 1965; Daskin 2008). Here, conservation offices represent supply points of the management effort demanded by the organization's protected areas.

A full accounting of the travel costs involved in protected area management should also include the resulting greenhouse gas (GHG) emissions. Many conservation organizations involved in managing protected areas are also involved in policy and advocacy campaigns seeking to persuade other sectors of society to reduce their greenhouse gas emissions. It is imperative therefore that conservation organizations are seen to be taking whatever steps possible to reduce their own emissions (e.g. National Park Service 2010). What studies are available to date suggest that many conservation organizations do not have a climate mitigation plan (Lemieux et al. 2011ba), but that changes to their operations can reduce GHG emissions substantially. For example, Parks Canada reduced emissions from its operations by 10.5% within a decade (Lemieux et al. 2011bb).

To explore the savings made possible by reducing management-related travel, we apply a heuristic office-siting algorithm to two real-world case studies. For each, we find travel-minimizing locations for their existing offices, and for an additional office, given the existing configuration. We compare the resulting annual travel savings with the cost of (a) land acquisition and (b) hiring additional managers. We also calculate the differences in travel-related GHG emissions produced in each case, and estimate their social cost.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Solution method
  6. Results
  7. Discussion
  8. Acknowledgments
  9. References
  10. Supporting Information

Case studies

To demonstrate the benefit of using optimization approaches in conservation office siting, we analyzed two case studies. These differ in organizational structure, regional extent and shape, road density and settlement patterns. For each case study, we explored the advantages of (a) optimally re-siting all existing offices and (b) optimally siting an additional office with the current configuration. For both formulations, we assume that the protected areas and their management burdens are already established.

Yorkshire Wildlife Trust

The Yorkshire Wildlife Trust (YWT) is a regional conservation nonprofit with a land trust like business model. YWT manages 84 small protected areas covering 2,067 Ha distributed across the county of Yorkshire (U.K.) (Figure 1a). Staff involved in site management are based out of two offices, one in York and one located onsite at Potteric Carr, their most intensively managed protected area. Levels of management effort applied to each site were obtained using a questionnaire survey of site managers (see Armsworth et al. 2011).The median annual management cost per protected area is $3,204 in 2008 equivalent U.S. dollars (Armsworth et al. 2011).

image

Figure 1. (a) Yorkshire roads (gray lines), and YWT's protected areas (circles) and offices (stars). The size of each circle denotes the number of manager-days are spent annually at that site. A travel-minimizing third office (diamond) would be placed at Spurn, the largest and second most intensively managed protected area. (b) The travel-minimizing configuration of two offices (diamonds) is similar to their existing locations.

Download figure to PowerPoint

Northwest Florida Water Management District

The Northwest Florida Water Management District (NWF) is one of five districts covering the state of Florida (USA), which enact water-quality projects, water-use permitting, and the acquisition and management of wetlands, floodplains and uplands for water resource protection. NWF (Figure 2a) manages 12 protected areas and 19 conservation easements totaling 89,396 Ha. Properties managed by NWF are larger, more distant from each other and more elongated than those managed by YWT. NWF is subdivided into three management regions, each served by a regional office. Conservation easements are monitored once a year, from the NWF headquarters. Management effort is determined from time sheets spanning January 2008 to May 2012.

image

Figure 2. (a) NWF's road network (gray lines) and managed lands. Four offices carry out NWF's management and monitoring activities (stars). A travel-minimizing fifth management office (diamond) would be located in the middle of the region. (b) The travel-minimizing locations of four offices (diamonds).

Download figure to PowerPoint

Solution method

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Solution method
  6. Results
  7. Discussion
  8. Acknowledgments
  9. References
  10. Supporting Information

We based our approach to minimizing total annual travel between management offices and protected areas on the p-median problem (Hakimi 1965; Daskin 2008). In our formulation, management offices “supply” conservation effort to a set of protected areas that each demands some level of management. We use the number of manager-days spent annually at each protected area as a measure of demand. Because the spatial distribution of management effort may not always be known, we explore the consequences of simpler weighting schemes in the supporting information (SI).

To simplify the optimization problem, we restrict our search for office locations to road junctions and endpoints. Similarly, we assume that conservation demand occurs at the centroid of each protected area rather than across its spatial extent (but see SI for sensitivity tests on the number and location of centroids). This discretized problem always has at least one optimal solution (Hakimi 1965). Discretizing the solution space reduces the number of possible office locations to 670,543 in YWT and 150,634 in NWF. To find travel-minimizing sets of office locations, we used ArcGIS 10 (ESRI) to calculate the road distance between the centroid of each protected area and every possible office location. We then imported these distances to MATLAB 2011a (Mathworks), and applied an interchange algorithm (Teitz & Bart 1968) to find an optimal office placement (see SI). This algorithm is a commonly used heuristic for solving p-median problems (Reese 2006). When choosing multiple office locations, we ran the algorithm 100 times with a random initial solution set.

The office siting algorithm generates a set of travel-minimizing office locations, as well as the total annual distance that managers would travel from those offices to maintain all protected areas at present effort levels. In both cases, we assume that each protected area is managed by its nearest office. Distance is a natural metric to express travel savings, but because we are interested in how those savings compare to the magnitude of other conservation costs (e.g., land acquisition, cost of employing site managers), we convert our results into monetary values. To calculate a time cost, we use average speed data (USDOT 2009; Wang et al. 2009) and staff pay rates (provided by YWT and NWF). We use government reimbursement rates to estimate the costs of fuel and vehicle wear. In calculating costs, we accounted for the average number of managers per vehicle (see SI for details). All monetary amounts are presented in 2008 equivalent U.S. dollars.

We also estimated GHG emissions that would result from travel. We converted annual travel distances to tonnes of CO2e emissions using The Nature Conservancy's carbon footprint calculator (http://www.nature.org/greenliving/carboncalculator/index.htm, accessed on March 14th, 2013) for mid-size (20–30 mpg) and large (<20 mpg) vehicles. We convert our CO2e estimates to dollar values using $12 per tonne, the mean value of social cost estimates collected by the IPCC 4th Assessment (IPCC 2007), as well as the more conservative estimate of $85 suggested in the Stern Review on the Economics of Climate Change (Stern 2007). Converted to 2008 equivalent dollars, these values are $12.46 and $88.26, respectively.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Solution method
  6. Results
  7. Discussion
  8. Acknowledgments
  9. References
  10. Supporting Information

Office placement

Table 1 details our estimates of management-related travel costs in terms of monetary value (Table 1, first column) and associated GHG emissions (second column). Costs are shown for (a) the current office configuration for YWT and NWF, (b) for an optimized placement of existing offices, and (c) the optimized placement of an additional office. These results are underpinned by our use of existing management effort data to estimate conservation demand at each protected area. Other weighting schemes based on more easily obtained data, such as the size of each protected area, are possible. However, using size of protected areas as a proxy for management effort can result in inefficient recommendations because effort scales differently with area for different conservation networks (see SI for a comparison between YWT and NWF).

Table 1. Estimate of annual cost of management-related travel (2008 dollars) and annual GHG output from management trips (CO2e) for both YWT and NWF
 YWTNWF
 Travel cost (2008 dollars)GHG emissions (tonnes CO2e)Travel cost (2008 dollars)GHG emissions (tonnes CO2e)
  1. Numbers are given for the current configuration of offices in the region (first row), a travel-minimizing configuration with the same number of offices (second row), and for a configuration of existing offices with a travel-minimizing additional office.

  2. a

    20–30 miles per gallon (47–71 km/L).

  3. b

    Less than 20 miles per gallon (47 km/L).

Current$128,00054 (mid-size cara)$199,00074 (mid-size cara)
  62 (large carb) 86 (large carb)
Optimized (all offices)$126,00053 (mid-size cara)$104,00039 (mid-size cara)
  61 (large carb) 45 (large carb)
Optimized (additional office)$85,00035 (mid-size cara)$139,00052 (mid-size cara)
  41 (large carb) 61 (large carb)

Yorkshire Wildlife Trust

For YWT, our analysis identified travel-cost minimizing office locations very close to the organization's existing offices. These are located near protected areas whose management demands are heavily weighted (Figure 1b). The optimized office configuration offers a slight reduction in travel costs, of about $2,000 annually, and would also prevent the emission of about 1 tonne of CO2e.

Both existing offices are located within the main cluster of protected areas, but both are relatively far from the site with the second-highest management demands. A third office, optimally sited in or near this protected area, would save $43,000 in travel costs and 19–21 tonnes of CO2e annually. This reduction in emissions from adding a third office to YWT could reduce the social cost of its operations by $237–$262 annually when using the IPCC's estimate of social cost of CO2e emissions, or $1,677–$1,853 when using the Stern Review's estimate.

Northwest Florida Water Management District

Optimizing the locations of all four management offices across NWF results in a configuration (Figure 2b) in which three of the offices are near their current locations and the fourth is quite different. This revised configuration could reduce annual management-related travel cost by $95,000. Expansion of the existing office configuration by adding an optimally placed fifth office (Figure 2a) would save $60,000 per year.

In terms of GHG emissions, optimally locating four offices would save 35–41 metric tonnes of CO2e emissions per year with an estimated social cost of $436–511 using the IPCC estimate and $3,089–3,619 using the Stern Review's estimate. Taking the current office locations as given but expanding by adding a fifth optimally sited office would reduce travel-related emissions by 22–25 tonnes of CO2e per year with a social cost savings of $274–312 (IPCC estimate) or $1,942–2,207 (Stern Review estimate).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Solution method
  6. Results
  7. Discussion
  8. Acknowledgments
  9. References
  10. Supporting Information

Organizations active in conservation management necessarily face travel costs in managing their protected areas. We examine how organizations can reduce these costs by allocating their staff resources effectively. Our case studies reveal that some conservation organizations may have arrived near a travel-minimizing configuration without ever having relied on a formal algorithm and careful travel cost accounting, while others may not. Our approach can also be used to estimate where the greater savings lie—whether in a reconfiguration of existing offices or in the addition of a new office.

Here we seek to put the magnitude of the possible savings in better context for conservation. The savings from a travel-minimizing additional YWT office is 13 times the annual median cost of managing one of its protected areas. These savings are also more than enough to employ an additional site officer at current salary rates. Reorganizing the offices of NWF into a travel-minimizing configuration also saves nearly twice the salary of a management officer. Alternately, travel savings could be used to acquire new protected areas. The addition of a travel-minimizing third office to YWT yields savings with a net present value (5% discount rate) of $860,000. This would be enough to acquire 360 hectares of comparable protected land in Yorkshire using the median of past acquisition costs paid by YWT. Similarly, the savings resulting from a reconfiguration of NWF's four existing offices have a net present value of $1,900,000, enough to acquire 171 hectares in Northwest Florida at a median price of $11,085 per Ha (USDA 2007). The savings that we detail here are increased slightly (by 0.5%–4.5% annually) by including the estimated social costs of carbon, although there is no means for these externalized costs to be recovered by the organizations.

We have developed this analysis from the perspective of expanding conservation networks, but organizations facing budget shortages may also consider closing offices. In these cases, our analysis could identify which office closure affects travel cost the least. As a hypothetical example, an office closure would increase NWF's annual travel costs by $11,000–$229,000. Similarly, were YWT to close an office, its travel costs would increase by $56,000-$72,000 annually. Although office closures reduce other operating expenses, the increase in travel costs may be substantial.

Our results illustrate the importance of considering office-to-site distance when choosing office locations. In practice, travel cost would be included alongside many other considerations. The cost of opening and closing offices varies spatially, affected by land values, zoning, and whether the new office must be rented, bought or built. Quality-of-life considerations also figure in to many business relocation decisions (Love & Crompton 1999), because factors like housing availability, commuting time and school quality affect organizations’ ability to attract and retain employees. Our model can be extended to address such concerns, for example, by restricting candidate sites to those within a specified driving distance of the nearest population center. A related consideration is that longer home-to-work commutes could increase GHG emissions from employees’ vehicles, potentially offsetting some of the carbon savings associated with an office location that only minimizes management-related travel. Moreover, regional offices have other benefits and costs that should be considered in actual siting decisions, such as providing conservation organizations with an on-the-ground presence through which to connect to local communities.

Introducing greater detail into the analysis clearly improves its applicability. At the same time, the cost of data collection and the human effort needed to conduct an analysis increases as detail is added. Because conservation planning is itself costly (Groves 2003; Bottrill & Pressey 2012), one important extension of our work would be an evaluation of just how much detail is needed in different aspects of office location decisions in order to offer a reasonable level of efficiency gain. For example, the assumptions we make about the distribution of management effort across protected areas affect the cost of our solutions. In our illustrative examples, we started from the current protected area networks and had detailed information on the allocation of recent management effort across these protected areas. For some applications, such information may be costly to collate. In order to test the importance of management effort to travel cost, we recreated two simple approaches to data-poor scenarios (assuming that all reserves receive equal management effort, or that management effort scales linearly with protected area size) and calculated their travel cost under current management regimes (SI). In YWT, optimizing office locations under the assumption that all areas require equal management effort gives an office configuration that would increase annual travel costs by $20,000 compared to the current office locations when the actual distribution of management effort is applied. In NWF, using the size of protected areas instead as a proxy for management demand increases annual travel cost by $100,000 compared to the existing office configuration. This suggests that realistic management effort data may play an important role in this type of analysis. It is also possible that travel costs already affect how management effort is distributed across sites, in which case changing the office configuration could also change the amount of effort directed at each site. Organizations may even prioritize areas for protection differently, something we will examine in future work.

Conservation organizations, like other businesses, stand to benefit from increasing the efficiency of their operations. Much could be gained by broadening application of techniques from the field of operations research, beyond their customary application to reserve site selection problems to other decisions that conservation organizations take. This is something we have sought to illustrate here by considering where conservation organizations should locate the human resources they have available for management.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Solution method
  6. Results
  7. Discussion
  8. Acknowledgments
  9. References
  10. Supporting Information

This study was funded by NSF's SCALE-IT IGERT (NSF Award 0801540) and the University of Tennessee. We thank several colleagues, including R. Fovargue, G. Iacona, and A. Milt, as well as A. Ando, H. Grantham, and one anonymous referee, for helpful comments and suggestions.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Solution method
  6. Results
  7. Discussion
  8. Acknowledgments
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Solution method
  6. Results
  7. Discussion
  8. Acknowledgments
  9. References
  10. Supporting Information

Disclaimer: Supplementary materials have been peer-reviewed but not copyedited.

FilenameFormatSizeDescription
conl12115-sup-0001-SupMat.pdf336K

Figure S1: Workflow to find travel-minimizing office locations, given data on 11 existing protected areas and roads.

Figure S2: Left: current office locations. Center: optimal office locations, when travel is weighted by the number of days spent at each protected area over the course of a year. Right: optimal office locations, when sites are not weighted.

Figure S3: Top: four travel-minimizing office locations, when the solution is weighted by the area of each site. Bottom: four travel-minimizing offices when travel is weighted by management effort.

Figure S4: For the NWF regression of management effort against site area (left), n = 31, P = 1.166 x 10–7, r2 = 0.626 and the slope is 0.94 with a standard error of 0.13. For the YWT regression (right), n = 74, P = 8.943 × 10–12, r2 = 0.4736, and the slope is 0.45 with a standard error of 0.06.

Figure S5: Escambia River Water Management Area, split along the river (left) and into shorter segments (right).

Figure S6: Current office locations (stars) and travel-minimizing locations (diamonds) in NWF. Top row: travel-minimizing locations for an additional office (left) and four offices (right), for both sensitivity tests. Bottom row: travel-minimizing office locations as reported in the main text.

Table S1: Estimated annual travel costs for both sensitivity test cases, with the current office locations, a travel-minimizing fifth office, and a travel-minimizing configuration of four offices. The “Office Locations” column denotes the ID of the road junctions where offices are located.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.