Choosing the battles: The economics of area wide pest management for Queensland fruit fly

Food Policy 38 (2013) 203–213

Contents lists available at SciVerse ScienceDirect

Food Policy journal homepage: www.elsevier.com/locate/foodpol

Choosing the battles: The economics of area wide pest management for Queensland fruit fly Veronique Florec a,b, Rohan J. Sadler a,b, Ben White a,b,⇑, Bernie C. Dominiak b,c a

School of Agricultural and Resource Economics, The University of Western Australia, M089, 35 Stirling Highway, Crawley, WA 6009, Australia Cooperative Research Centre for National Plant Biosecurity, LPO Box 5012, Bruce, ACT 2617, Australia c Plant Biosecurity Risk Management, Industry and Investment, Locked Bag 21, Orange, NSW 2800, Australia b

a r t i c l e

i n f o

Article history: Received 16 May 2012 Received in revised form 24 October 2012 Accepted 27 November 2012 Available online 10 January 2013 Keywords: Queensland fruit fly Area wide management Surveillance Eradication Roadblocks Invasive species Biosecurity

a b s t r a c t ‘‘Surely the best way to meet the enemy is head on in the field and not wait till they plunder our very homes’’ Oliver Goldsmith (1730–1774). Area-wide management (AWM) of crop pests is an alternative strategy for pest control to reliance on the uncoordinated control decisions of farmers. Relative to uncoordinated pest control, AWM has been shown to be cost-effective and, by reducing pesticide use, environmentally beneficial. The fact that AWM schemes provide imperfect public goods and are prone to free-riding means that most successful schemes depend on government funding, regulation, coordination and management. The economics of AWM concerns the economics of information and time in complex bioeconomic settings. This paper explores the economics of AWM in relation to Queensland fruit fly (Qfly), Bactrocera tryoni (Frogatt), a damaging pest and a major barrier to Australian trade in horticultural produce. We analyse the economics of roadblocks, surveillance and eradication. The results show that returns from tighter roadblocks are greater than returns from increased surveillance and enhanced eradication capacity. These results depend on market access rules, the spatial extent of the pest free area, the horticultural commodities at risk, and pest ecology. Ó 2012 Elsevier Ltd. All rights reserved.

Introduction The growth of world trade and increased movement of people and goods has raised the risks to agriculture and natural environments from invasive organisms (Lichtenberg and Lynch, 2006; Mumford, 2002). Further, food safety standards have low tolerances for pesticide residues in food and markets requiring low pesticide and low pest risks are increasing (Hendrichs et al., 2005; Mumford, 2005). To export, Australian producers have to adopt pest management methods that satisfy increasingly stringent food safety and pest-free requirements. One alternative to intensive pre-harvest spraying and post-harvest treatments is the establishment of an area-wide management (AWM) scheme, where controls apply over a region (Devorshak, 2007; Faust, 2008). AWM has the potential to contribute towards increased food safety and reduce the environmental effects of pesticide use. With reference to Queensland fruit fly (Qfly) management in Australia, this paper explores what constitutes an economically efficient AWM scheme.

⇑ Corresponding author at: School of Agricultural and Resource Economics, The University of Western Australia, M089, 35 Stirling Highway, Crawley, WA 6009, Australia. Tel.: +61 8 6488 4634. E-mail address: [email protected] (B. White). 0306-9192/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.foodpol.2012.11.007

There are numerous studies of the scientific principles underpinning AWM.1 However, detailed theoretical and empirical economic analyses of AWM design are rare. Exceptions include Mumford (2000, 2005), Enkerlin (2005), and Mau et al. (2007). Without economic analysis, it is difficult to assess the value of different management strategies, technologies and scale of operations of an AWM program. Furthermore, government departments managing AWM schemes are expected to be cost efficient and demonstrate that the benefits of AWM exceed the costs (Mumford, 2005). Queensland fruit fly (Qfly), Bactrocera tryoni (Frogatt), is economically the most significant horticultural pest in Australia (Clarke et al., 2011; Dominiak et al., 2000; Gilchrist et al., 2006; Sutherst et al., 2000).2 In regions where Qfly is endemic,

1 See for instance Koul et al. (2008), Tan (2000), Vreysen et al. (2007), Hendrichs et al. (2005), Klassen (2005), Lloyd et al. (2010), Jessup et al. (2007), and papers from the joint FAO/IAEA international conferences on area-wide control of insect pests. 2 Qfly has more than 200 hosts, including stone fruit, citrus and table grapes, among others (see http://www.pestfreearea.com.au for a complete host list). Table grapes have been previously classified as a marginal host, but the fruit can sustain the insect from larval to adult stage in the laboratory (Clarke et al., 2011). Even though larvae do not always develop to maturity in the field, Qfly outbreaks have been reported to cause significant damage to grape producing areas (Dominiak, 2011). Citrus tends to ripen during winter periods when fruit fly activity is low, and thus experience a lower rate of infestation in general. Stone fruit are a preferred host.

V. Florec et al. / Food Policy 38 (2013) 203–213

horticultural producers are either excluded from or have restricted access to export and domestic markets, and their profitability is reduced by pest control costs, yield reduction and post-harvest disinfestation costs (Lloyd et al., 2010).3 Qfly is viewed by Australia’s trading partners as a serious biosecurity threat to their horticultural industries. To reduce the risk of Qfly outbreaks strict import regulations are placed on Australian produce, for instance, requiring exported fruit and vegetables to be below a Probit 9 likelihood of infestation.4 This paper presents an economic analysis of the design of AWM schemes for Qfly in Australia. The next section reviews the economic literature. Section ‘‘Biological and environmental characteristics of Queensland fruit fly’’ outlines aspects of Qfly biology relevant for an economic analysis of AWM design. Section ‘‘Key elements for the economic analysis of area-wide management for Queensland fruit fly’’ introduces the empirical bioeconomic model. Section ‘‘Case study for the Fruit Fly Exclusion Zone (FFEZ)’’ presents a case study of the Fruit Fly Exclusion Zone (FFEZ). Section ‘‘Results’’ gives results for optimal surveillance, roadblock activities and eradication effort. Section ‘‘Conclusion’’ concludes. Background AWM of Qfly is a form of public good (Burnett, 2006; Hennessy, 2008; Hinchy and Fisher, 1991). The principle behind AWM is that economies of scale, and a degree of non-rivalry and non-exclusivity in benefits means that ‘‘pests can be effectively managed using an organised and coordinated attack on their population over large areas rather than by using a field-by-field approach’’ (Koul et al., 2008, p.1). Indeed, the uncoordinated effort of individual producers is generally insufficient for the effective management of mobile pests (Klassen, 2005; Perrings et al., 2002). Such schemes are typically not incentive compatible and are undermined by free riding and sub-optimal pest control by producers. Further, technologies and actions available to government, namely roadblocks, surveillance grids and sterile insect technology (SIT), would either be illegal for a producer group (due to land access issues) or would be infeasible due to the range of scientific expertise required. Therefore AWM schemes are provided by government, although producer groups are often required to share the costs. AWM of fruit flies commonly involves monitoring of traps over large areas (surveillance), the control of movements of host produce (roadblocks), and the use of the sterile insect technique for the eradication of outbreaks. Investments in surveillance may reduce the time to detection of the pest population and investments in eradication capacity may reduce the time to eradication of a detected population (Fig. 1). Early detection means the population at detection is smaller and eradicated more rapidly. In Australia, five market access rules apply to AWM of fruit flies: (1) Biosecurity rule: in the absence of fruit flies, exported produce must be below the required Probit 9. (2) Capture rule: corresponds to the spatiotemporal rules for the declaration of an outbreak, that is, the threshold number of fruit flies captured within a period and distance to each other that leads to outbreak declaration. Since the size of

3

Qfly damages horticultural produce when females sting host fruit or vegetables to lay their eggs below the skin of the fruit. Larvae feed within the fruit, reducing its quality and rendering it unmarketable. It is estimated that without control, an infestation of Qfly can damage between 80% and 100% of host fruit and vegetable grown in the infested area (Sutherst et al., 2000). 4 Probit 9 corresponds to 3.2  105 survival rate after treatment.

Post-harvest treatment period

Loss of areafreedom status

Reinstatement of area-freedom status

Infestation intensity

204

Surveillance investment Detection

Recertification Rule

Eradication investment Eradication

Time Fig. 1. Detection, eradication and reinstatement of area-freedom status, after Kompas and Che (2009).

the infestation is not known with certainty it is assumed that under this threshold the risk of infestation is low.5 (3) Area rule: the area that loses area-freedom status for trade purposes when an outbreak is declared, called the suspension zone. This area has a radius between 15 km and 80 km from the outbreak origin. Produce grown in the suspension zone has to undergo post-harvest treatments before export to fruit fly free markets during eradication and for a period after eradication activities cease. The size of the suspension zone is established according to the intended market and the distance between consecutively trapped flies. (4) Treatment rule: an outbreak zone is delimited around sites where the flies are captured. Eradication activities are carried out within the outbreak zone. (5) Recertification rule: After eradication activities cease, access to restricted markets can only be restored after a period has elapsed during which no more flies are captured. The time to area freedom recertification in the suspension zone is prolonged until the risk of spread and establishment of a second generation is very low (see Fig. 1). This time is defined by the generational time span of Qfly, which is temperature dependant.6 The regulator’s problem is one of designing the AWM program so that resources are allocated efficiently across surveillance, roadblocks and eradication. In implementing AWM programs the regulator faces the ‘‘weaker link’’ problem (Burnett, 2006), whereby sub-optimal prevention and/or control in a part of the AWM system lowers the returns to AWM. For instance, if one inspection roadblock is not effective in detecting infested fruit, the fact that all other roadblocks are effective is irrelevant. Burnett (2006) demonstrated that the incentive structure resulting from the weaker link public good problem causes contributors to under invest in pest management, emphasising the importance of the careful design of AWM schemes. To date the only assessments of AWM schemes for Qfly in Australia have been aggregated benefit-cost analyses (BCAs). Such 5 The size of an infestation is determined by the number of flies present and the extent of the area they occupy. 6 For a summary of restrictions regarding fruit fly outbreaks per importing country, see Appendix A in Florec et al. (2010a). Updates can be found per country in PHYTO (AQIS, 2008), Australian Quarantine and Inspection Service’s plant and plant product export conditions database (available on http://www.aqis.gov.au/phyto/asp/ ex_home.asp).

V. Florec et al. / Food Policy 38 (2013) 203–213

analyses indicate that the scheme as a whole represents a potential welfare gain but do not include interactions between the technologies used, the spatial characteristics of the region studied, the ecology of the pest, and market access rules. The earliest comprehensive BCA was PricewaterhouseCoopers (2001). This study estimated an annual benefit of $14.9 million (growers and exporters captured most of the benefits) for the TriState Fruit Fly Strategy, and a benefit cost ratio (BCR) of 2.5:1. For Qfly management in Queensland, Franco-Dixon and Chambers (2009) presented a BCA for the proposed application of AWM in the Central Burnett District. The authors considered the probability of the Australian Pesticides and Veterinary Medicines Authority (APVMA) banning the use of dimethoate as a post-harvest treatment of fruit, and the probability of the Interstate Certification Assurance (ICA-28) scheme being extended to include four Australian States. The estimated net present value for the Central Burnett District AWM project over 10 years was $5.2 million, with a BCR of 2.27:1. For Qfly AWM in Victoria, DPI Victoria (2010) compared three different AWM options and estimated a larger benefit cost ratio for the option that encompassed the establishment of areas of low pest prevalence, BCR of 2.35:1. This study presented three options that included different production areas, but did not analyse changes in the technologies used, the environmental conditions of the areas studied, the market access rules, or the availability of produce throughout the year. All the BCAs cited showed that public funding of AWM schemes is justified, but there is little analysis of their design. Mumford (2000) discerned four major questions that need to be answered when designing an AWM program: (i) should the pest be controlled locally or area-wide; (ii) over which area; (iii) what is the most efficient form of control; and (iv) what level of organisation should be used. In relation to AWM the specific design questions relate to: the intensity of inspections and spatial distribution of roadblock sites, the intensity and spatial distribution of surveillance, eradication capacity, and the spatial extent of the protected area. This paper focuses on what needs to be known about AWM of Qfly in order to answer these questions to optimally design an AWM program. The economic literature on invasive species has shown that the design of pest management policies requires an understanding of the biological features of the pest and accurate taxonomic, geographic, and temporal data concerning the pest and the region studied (Stohlgren and Schnase, 2006). The optimal choice of policy depends ‘‘on the characteristics of the exotic pest, its costs, how it spreads, how easy it is to detect and eradicate, and the extent of the infestation’’ (Acquaye et al., 2007, p.102). A variety of invaders have been considered in the literature: weeds (Cacho et al., 2007; Chalak-Haghighi et al., 2008), insects (Bogich and Shea, 2008; Ceddia et al., 2009), plant disease vectors (Brown et al., 2002), plant diseases (Acquaye et al., 2005), fungi, and vertebrate animals (Bomford and O’Brien, 1995). These studies, which often present optimisation models that minimise the cost of management, potential damages and the risk of introductions, demonstrate the importance of incorporating all the key spatial and temporal aspects of the biophysical system and the determinants of cost and benefits. The next section focuses on the biological characteristics of Qfly that are relevant for its management and may significantly affect cost-benefit estimates. Biological and environmental characteristics of Queensland fruit fly Qfly possesses a number of biological and environmental characteristics that are critical in determining costs and benefits of alternate AWM designs. First, temperature and moisture regulate its life cycle (it thrives in warm and moist places) (Sutherst et al.,

205

2000; Yonow and Sutherst, 1998), with the time to maturity of a generation changing seasonally and spatially with climate. This has a number of consequences: (i) outbreaks affect producers differently in temperate and tropical climatic areas; (ii) seasonal climate drives seasonal peaks and troughs in Qfly populations (Muthuthantri et al., 2010); (iii) time to area freedom reinstatement (that is, the time until export markets accept untreated fruit anew) varies throughout the year; and (iv) the spatial extent of a management option may be shaped by climatic boundaries. Second, population abundance follows phenological patterns (Muthuthantri et al., 2010). Namely, seasonal changes in temperature and moisture availability affect fly abundance and distribution. Hence, the fly’s invasive pressure to pest free areas (PFAs) varies throughout the year. During the warmer and wetter times of the year, the fly presents a greater threat to horticultural production regions located at the periphery of its endemic range (for information on the current distribution of Qfly in Australia, see Dominiak and Daniels, 2012). Third, although there is some evidence for a ‘fat-tailed’ dispersal distribution (see Sadler et al., 2011), it is generally agreed that Qfly does not disperse far (Dominiak et al., 2003; Meats and Edgerton, 2008). Despite this, Qfly can be found in locations a long way from its endemic range, suggesting road transportation through infested produce. In this context, it is important to consider the costs and benefits of activities that reduce the risk of entry, such as inspection roadblocks at the border of the AWM program, quarantine measures, and buffer zones and activities that minimise the period between the arrival of the pest and its discovery (surveillance) at different times of the year. Finally, the extent and characteristics of a region can have considerable effects on benefit-cost estimates (Mumford, 2000). The results of an analysis can differ substantially depending on relative production value among Qfly’s wide range of hosts, areas of noncrop hosts (i.e. backyard trees), and type of control suitable for the landscape. The availability of susceptible produce varies throughout the year, so the costs of Qfly incursions during harvest are potentially higher due to greater potential yield damage (that is there is more fruit available for the fly to lay eggs) and larger quantities of fruit that need to undergo post-harvest treatments. For instance, Florec et al. (2010a) conducted an economic analysis of Qfly incursions in the FFEZ showing that revenue loss associated with Qfly incursions vary greatly with the time of the year the outbreak occurs, due to host availability and the time taken to reinstate area freedom. Since the characteristics of the pest targeted and the environment studied determine the optimal choice of policy and design, the next section identifies key elements in an economic analysis of AWM for Qfly. Key elements for the economic analysis of area-wide management for Queensland fruit fly Here the aim is to determine the approximately optimal allocation of AWM tools (roadblocks, surveillance and eradication). The counterfactual (benchmark) is the current application of AWM. The benefits of implementing more efficient AWM schemes stem principally from reduced post-harvest costs and reduced time to eradication. Therefore in this section, we look first at the estimation of revenue losses due to post-harvest treatments and we examine the costs of each AWM tool and potential benefits of changes in their implementation. Post-harvest treatments Post-harvest treatments are a significant component of the costs of Qfly incursions and these costs arise whilst the outbreak

206

V. Florec et al. / Food Policy 38 (2013) 203–213

is declared active (during the eradication campaign and until area freedom is restored according to the recertification rule). Produce harvested within the suspension zone around each declared Qfly outbreak requires post-harvest treatment before export to sensitive markets. Requirements differ significantly between markets: some sensitive markets only accept untreated fruit when pest-free status is maintained but require post-harvest treatments when an outbreak is declared; other markets attach little importance to area-freedom status and require all fruit to be treated; and a few markets accept fruit without restrictions. On this basis the revenue lost due to post-harvest treatments, from outbreak declaration to reinstatement of area freedom is differentiated according to market by

cH ðtÞ ¼

M M;T I a ðmÞ X X X wiH Am qim

ð1Þ

the outbreak is declared (see NFFWG, 2008; White et al., 2011 for a table indicating the minimum duration of outbreaks depending on the time of declaration). Here we assume that ce(Td, Te) is uniform across the outbreak area. In reality, ce(Td, Te) may vary depending on terrain, accessibility, the number of backyard trees with ripe fruit and distance from operations centre. However, since the eradication campaign follows a standard protocol we do not incorporate spatial variation in the variable costs of eradication. An increase in eradication effort may lead to a reduction in the time to eradication and time to area freedom reinstatement. Hence, the period during which produce grown within the suspension zone has to undergo post-harvest treatments may be shortened. Combining Eqs. (1) and (2), the potential benefits of a change in costs due to an increase in eradication effort is

m¼1m¼1;t¼T d i¼1

wiH

with the unit cost of post-harvest treatment of product i; Am is the area of the suspension zone for market m; qim ðtÞ the quantity of product i harvested per ha at time t going to sensitive market m requiring post-harvest treatment during an outbreak Td the time of outbreak declaration; and Ta(m) the time of area-freedom reinstatement for market m. As the costs cH(t) of post-harvest treatments are expressed on a per outbreak basis, for the annual expected cost, they are multiplied by the probability of a Qfly outbreak in the area. Eradication Eradication is initiated immediately after outbreak declaration. The protocol for eradication however is standard, both in terms of equipment used and extent of the area treated, irrespective of the size of the infestation. Eradication campaigns generally start with 2 weeks of intensive bait spraying followed by a minimum of 10 weeks of sterile Qfly release and continuing until no more flies are found. Since eradication effort is assumed constant per unit of time, the invasion size only affects cost through the time taken to eliminate Qfly. The expected variable costs of an eradication campaign may be written as

ce ðT d ; T e Þ ¼

Te X

aðtÞ

ð2Þ

t¼T d

where Td is the time of outbreak declaration; and Te when the eradication campaign finishes. Parameter a(t) denotes the average cost of control activities per unit time, and may be expressed as

aðtÞ ¼ wLe Le þ wx x þ wQ Q e

ð3Þ

where wLe is the salary of staff conducting the eradication campaign in time t; Le the staff necessary to conduct the eradication campaign; x the equipment necessary to eradicate the pest; and wx the price of this equipment. In the case of an eradication campaign, x may include the amount of chemicals, the sprayers to apply pesticides, pupae of sterile insects, trucks for the release of sterile insects. Qe is the number of supplementary traps deployed and wQ the cost of increased surveillance per trap. Supplementary traps are included in the eradication costs because, when the pest is first detected, a decision may be made to increase surveillance activities and set up more traps around the discovery point. Typically a minimum of 16 supplementary Lynfield traps are installed in the outbreak zone and remain deployed for 9 weeks after the last Qfly detection. These traps are inspected twice weekly during the time they are deployed, thus increasing surveillance activities (see NFFWG, 2008). The expected variable costs of an eradication campaign ce(Td, Te) has also been shown to depend on the time of year

T e1 X

Be ¼

a1 ðtÞ þ

t¼T d



T e2 X t¼T d

! M M;T I a1 ðmÞ X X X wiH Am qim ðtÞ P m¼1m¼1;t¼T d i¼1

! M M;T I a2 ðmÞ X X X i i a2 ðtÞ þ wH Am am ðtÞ P

ð4Þ

m¼1m¼1;t¼T d i¼1

where a1(t) denotes the average cost of control activities per unit time for the current level of eradication effort; Te1 the time when the eradication campaign is finished with the current level of eradication effort; Ta1(m) the time of area-freedom reinstatement for market m with the current level of eradication effort; P the average probability of outbreak in the area studied; a2(t) the average cost of control activities per unit time for a new level of eradication effort; Te2 the time when the eradication campaign is finished with a new level of eradication effort; Ta2(m) the time of area-freedom reinstatement for market m with a new level of eradication effort. Roadblocks The purpose of roadblocks is to reduce the likelihood that an exotic pest enters a region (Beare et al., 2005). Given that human transportation of infested fruit is the main entry vector for most Qfly into a PFA (Clift and Meats, 2001), placing permanent roadblocks at the border of the protected area may reduce the probability of outbreaks. Roadblock operators inspect vehicles entering the PFA and control the movements of host fruit through a permit system (Dominiak and Coombes, 2010). Road signage close to roadblock site is also used as a strategy to reduce the amount of infested fruit entering the PFA (Dominiak and Coombes, 2009; Jessup et al., 2007). The effect of roadblocks on outbreak probability depends on how rigorously roadblocks are operated, the number of roadblock sites and the invasive pressure of the pest (Mumford, 2000). The variable costs of roadblocks for a given period t can be written as

C R ðtÞ ¼ ðwLR þ GL ÞLR þ D þ K þ F

ð5Þ

where wLR is the salary of staff operating roadblocks; GL the cost of accrediting roadside staff and the provision of logistical support such as transport; LR the staff necessary to operate roadblocks; D overhead costs; K the cost of road sign maintenance; and F the cost of the permit system for the movement of fruit through the PFA. More rigorous inspections or an increased number of roadblocks may lead to a reduction in the probability of outbreak. A reduction in this probability can result in decreased costs of eradication and post-harvest treatments. Modifying Eq. (4), we incorporate the costs of the current level of roadblock protection and the associated probability of outbreaks and estimate the difference with a new level of roadblock protection that results in a new

207

V. Florec et al. / Food Policy 38 (2013) 203–213

(Florec et al., 2010b). The number of traps deployed in a given area can be written as

probability of outbreak. The potential benefits of increased roadblock protection can then be written as Te X

BR ¼

! ! M M;T I a ðmÞ X X X i i aðtÞ þ wH Am qm ðtÞ P1 þ cR1 ðtÞ

t¼T d



Te X t¼T d

Qk ¼

m¼1m¼1;t¼T d i¼1

! ! M M;T I a ðmÞ X X X i i aðtÞ þ wH Am qm ðtÞ P2 þ cR2 ðtÞ

Ak

a2k

k ¼ r ðruralÞ; u ðurbanÞ

where Ak is the size of the area; and a2k the squared distance between grid points. The number of inspectors needed to monitor the trapping grid for a given time period is

ð6Þ

m¼1m¼1;t¼T d i¼1

where P1 is the probability of outbreak under the current level of roadblock protection; cR1(t) the variable costs of roadblock operations for the current level of protection; P2 the probability of outbreak for the new level of roadblock protection; and cR2(t) the variable costs of roadblock operations for the new level of protection. As for eradication costs, we assume that the variable costs cR(t) of roadblock operations and the inspection effort are uniform across the whole PFA. However, since the availability of host produce, the time to area freedom reinstatement and the probability of outbreaks vary seasonally and spatially, the most efficient roadblock protection level may also vary with time of year and location. It may be more efficient to increase the rigour of inspections or the number of roadblocks when either the consequences of an outbreak (potential loss) and/or the risk of an outbreak (probability of loss) are relatively high.

I¼

  Qk z bk

ð8Þ

where bk is the average number of traps an inspector can check in a given time period; and z the frequency of inspections in the given period of time (i.e. how often the grid is inspected, usually weekly or fortnightly). An estimate of the costs of surveillance c, for a given time period is

cS ðtÞ ¼ ðQ k Þg þ ðwI þ MÞI

ð9Þ

where g is the cost of trap maintenance per year per trap; wI the average salary of an inspector in time t; and M the average cost of a vehicle and any other overhead costs during time t. The estimation of surveillance costs for AWM is straightforward. However, the benefits of surveillance, though potentially large, are difficult to estimate, since it involves assessing the value of early detection of a pest. In general, the earlier the pest is detected the lower is the economic damage and the cost of pest management (Myers et al., 1998). There is therefore a trade-off between earlier detection and the costs of increased surveillance. Costs associated with an incursion can vary significantly depending on the time of the year when the outbreak occurs (Florec et al., 2010a). When long reinstatement periods coincide with peak production intervals, the benefits of earlier detection are potentially higher. Fig. 2 illustrates this by showing a hypothetical example of different times to detection of the pest population, different area freedom reinstatement periods and fruit production throughout the year. In Fig. 2, the arrival of Qfly in a pest free area tA may be detected at time tD1 or at time tD2 depending on the level of surveillance effort. Area-freedom status can be restored at time tE1 if the incursion is detected at tD1 because the no-capture period required for Qfly is shorter at that point in time due to the seasonally dependent recertification rule. In contrast, area freedom may not be reinstated until tE2 if the outbreak is detected at tD2 because the no-capture period required is considerably longer at that time.

Surveillance The main objective of surveillance is to minimise the time between the arrival of a pest and its discovery (Kompas and Che, 2009; Lindner and Macleod, 2009). Surveillance intensity determines the expected time to detection of a pest population. The trapping grid of a surveillance system also provides a means to prove pest-free status to regulators in export markets. Therefore surveillance has a marginal benefit and a fixed benefit: the marginal benefit relates to reducing the expected time to detection and thus eradication; the fixed benefits occur when minimum levels of surveillance apply that are consistent with market rules. Surveillance costs change quadratically in relation to grid point spacing and linearly in the size of the area trapped (Florec et al. 2010b). Since it is a labour intensive technology, surveillance costs largely depend on the number of inspectors needed to check traps, which in turn depends on the number of traps deployed, the distance between traps and how frequently they are monitored

Production (tonnes)

Time to eradication of the fly population + Time to area freedom reinstatement

Jan

Feb

tA

Mar

t D1

Apr

t D2

May

ð7Þ

Jun

Jul

Aug

Sep

Oct

t E1

Time Fig. 2. Time to detection, recertification rule and hosts availability.

Nov

Dec

t E2

208

V. Florec et al. / Food Policy 38 (2013) 203–213

An increased level of effort in surveillance may reduce the time to detection. Earlier detection can result in shorter time to eradication and, depending on the time of the year, in a shorter area freedom reinstatement period. Thus, eradication and post-harvest treatments costs may be reduced. Altering Eq. (4), we incorporate the costs of the current level of surveillance and estimate the difference with a new level of surveillance effort. The potential benefits of increased surveillance effort can be written as T e1 X

Bs ¼

! ! M M;T I a1 ðmÞ X X X i i aðtÞ þ wH Am qm ðtÞ P þ cs1 ðtÞ

t¼T d



T e2 X t¼T d

m¼1m¼1;t¼T d i¼1

aðtÞ þ

!

M M;T I a2 ðmÞ X X X wiH Am qim ðtÞ P þ cs2 ðtÞ

larger number of captures in a denser surveillance grid does not necessarily imply a greater biosecurity threat. Since the rule for outbreak declaration represents the threshold that indicates the presence of an incursion of a viable density, this threshold might be different for varying configurations of the trapping grid. Consequently, a change in surveillance effort may require a change in the rules for declaring an outbreak. This is, however, beyond the scope of this paper. Using the methods described above we analysed the case of the Fruit Fly Exclusion Zone (FFEZ). Case study for the Fruit Fly Exclusion Zone (FFEZ)

! ð10Þ

m¼1m¼1;t¼T d i¼1

where Te1 is the time when the eradication campaign is finished with the current level of surveillance effort; Ta1(m) the time of area-freedom reinstatement for market m with the current level of surveillance effort; cs1(t) the costs of surveillance for the current level of effort; Te2 the time when the eradication campaign is finished with a new level of surveillance effort; Ta2(m) the time of areafreedom reinstatement for market m with a new level of surveillance effort; cs2(t) the costs of surveillance for a new level of effort. There may be a point where increased surveillance intensity may actually lead to an increase in declared outbreaks. A denser array of surveillance points may lead to a greater number of captures, and hence an increase in the frequency of declared outbreaks if rules for outbreak declaration remain unchanged. However, a

The FFEZ, which covers about 185,000 km2 in south-eastern Australia, was established to facilitate exports and interstate trade. The region includes the majority of horticultural production for southern New South Wales, northern Victoria and eastern South Australia (Hendrichs et al., 2005; Jessup et al., 2007). Embedded within the FFEZ is the Greater Sunraysia Pest Free Area (GSPFA), which includes the high value production areas of Mildura and Swan Hill (Fig. 3). Since both eradication and post-harvest treatments are obligatory within the FFEZ, the region acts as a buffer to the dispersal of Qfly from Australia’s east coast into the GSPFA. The GSPFA is recognised as free from pest fruit flies by other Australian states and some export markets including New Zealand and the USA. The three main horticultural crops grown in the GSPFA that are subject to Qfly market access restrictions are citrus, table grapes and stone fruit. In fact grapes are the most significant horticultural

Fig. 3. The Fruit Fly Exclusion Zone. Note: In light green: The Fruit Fly Exclusion Zone; in blue: the Greater Sunraysia Pest Free Area. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Source: http://www.fruitfly.net.au.

209

250 200 150

0.6 0.4

0

0.2

Scaled production

0.8

Table Grapes Stone Fruit

100

Days to area freedom reinstatement

Citrus

50

1.0

V. Florec et al. / Food Policy 38 (2013) 203–213

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 0.0

Date last fly caught Fig. 5. One generation and 28 days or 12 weeks (whichever is longer) in the GSPFA. Note: After the last fly is caught, area freedom may be reinstated when the agreed no-capture period has elapsed. Source: NFFWG (2008).

Table 1 Annual budget for operating the FFEZ. Source: Vijaysegaran (2008) Annual budget

7,000,000

Eradication (40%) Surveillance (30%) Roadblocks (15%) Inspection, awareness programs and R&D (15%)

2,800,000 2,100,000 1,050,000 1,050,000

Results Roadblocks

0.0030

The potential gain of an increased level of resources allocated to roadblocks corresponds to reductions in the probability of Qfly incursions. The benefits of a lower probability of outbreaks repre-

0.0020

crop in the FFEZ, corresponding to 70% of the region’s horticultural production, but only a small proportion (15%) is consumed and traded as table grapes. Dried and wine grapes are not a biosecurity risk. Fig. 4 presents the average annual scaled production of citrus, table grapes and stone fruit. Peak production for citrus occurs between May and September, therefore if an outbreak occurs during this period, a larger quantity of produce is at risk of being located within the suspension zone. Similarly, table grapes are at risk between February and June and stone fruit between December and March. Following the National Code of Practice for the Management of Queensland fruit fly (NFFWG, 2008), surveillance in the FFEZ consists of a grid of traps spaced at 400 m in towns and 1 km in horticultural production areas. An outbreak is declared when five male flies are trapped within one km and within 14 days (or three male flies if no supplementary traps are deployed), or when one larvae or one gravid adult female are detected. An eradication campaign is launched the moment an outbreak is declared. The annual budget for operating the FFEZ is $7 million, allocated between different AWM activities (Table 1). In the GSPFA, where temperature and irrigation provide a moderate climatic suitability for Qfly development (Sutherst et al., 2000), a generational cycle may take between 4 weeks in summer and up to 29 weeks in winter. In contrast, a full Qfly life cycle takes between 3 weeks in summer and 8 weeks in winter in warmer areas (e.g. Tennant Creek, Northern Territory; see Annex 3 in NFFWG, 2008). The recertification rule in the GSPFA, which accounts for this climatic variation, is usually one generation and 28 days or 12 weeks (whichever is longer) (Fig. 5). Consequently, relatively long reinstatement periods are required before pest-free

status can be re-established in the suspension zone when the last Qfly capture of a declared outbreak occurs between March and August (Fig. 5). The invasive pressure of Qfly also changes throughout the year. Fig. 6 presents the current mean probability of outbreaks per grid cell over all the FFEZ. We analyse the data presented above to estimate the benefits of changes in AWM tools in the FFEZ. A full description of the spatial and dynamic bioeconomic QFAWM model can be found in White et al. (2011) and Sadler et al. (2011). We used the current application of AWM in the FFEZ as the counterfactual.

0.0010

Fig. 4. Citrus, table grapes and stone fruit annual scaled production in the Sunraysia. Note: scaled production is calculated by dividing production by the maximum quantity per month across the three crops (i.e., citrus in July). Sources: Australian Bureau of Agricultural and Resource Economics (ABARE), Citrus Australia, Murray Valley Citrus Board (MVCB), Australian Table Grape Association Inc (ATGA) and Summerfruit Australia.

Probability of outbreak (weekly)

Time

0.0000

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Time Fig. 6. Mean weekly probability of outbreaks per grid cell over the entire FFEZ. Source: Industry and Investment, NSW and Sadler et al. (2011).

210

V. Florec et al. / Food Policy 38 (2013) 203–213

sent cost savings in terms of post-harvest treatments and eradication expenditures. To estimate the net benefits however, it is necessary to calculate the benefits obtained less the investment that was necessary to generate those benefits. Hence, the relationship between the level of roadblocks expenditure and the associated probability of outbreak for each level of expenditure has to be known. However, in the Qfly literature, there are no studies that examine this relationship. Though some work has been done on the risk of Qfly introductions posed by road transport (e.g. Dominiak et al., 2000), there are no models that link the level of roadblock protection to the probability of outbreaks. Hence, in this paper, we estimate the expected benefits over a 20 year period of an improvement to the probability of Qfly outbreaks and expressed them as a breakeven value, i.e. a point where the additional cost of roadblocks and benefit (cost saved) are equal. To this end, we calculate the current average annual costs of incursions (with existing roadblock protection) and compared them with the average annual costs of incursions that would be incurred if the probability of a Qfly incursion was reduced by 10%. Table 2 shows the results for an entire year. Effective roadblocks resulting in a 10% reduction in the overall probability of outbreak directly reduces eradication costs and post-harvest costs (fewer outbreaks). The value of improved roadblocks was estimated as $0.5 million per annum in total (Table 2). As shown in Section ‘‘Key elements for the economic analysis of area-wide management for Queensland fruit fly’’, the probability of Qfly outbreak, the availability of host fruit and the recertification rule for the FFEZ vary through the year. Hence, the benefits of more intensive roadblock protection also vary seasonally. Table 3 illustrates seasonal differences. The maximum gain in the expected net benefit for a 10% reduction in the probability of outbreaks in the FFEZ is $0.42 million in summer and $0.61 million in winter. Although the mean weekly probability of outbreaks per grid cell over the entire FFEZ is lower in winter, an outbreak during winter is more costly. Despite the probability of outbreak being greater during the summer months, outbreak durations are shorter in summer, leading to lower post-harvest costs overall. Indeed, during winter, longer reinstatement periods (Fig. 5) envelope a greater proportion of the peak citrus production period (Fig. 4). A 10% reduction in the probability of outbreaks results in higher expected benefits during winter months than during summer months, as on average individual outbreaks over winter incur greater costs per outbreak. The same process can be repeated for any proportional reduction in the probability of outbreaks and estimate a net benefit for each reduction. Although with QFAWM it is possible to estimate the potential benefits of reducing the average probability of outbreaks at any time of the year, research is currently not available to assess if this reduction in the probability of outbreak is feasible, how much it would cost, and the spatial variability of these factors. Hence, further work needs to be done on the effectiveness of roadblocks. However, as a comparison, when compared with the costs given

in Table 1, the cost savings represent 48% of current roadblock costs and this is some indication that the additional expenditure to reduce outbreaks by 10% would be justified. With QFAWM, it is also possible to estimate the expected benefits of a proportional reduction of the probability of outbreaks for each level of surveillance. In Fig. 7, as the spacing between traps increases in the x axis and the intensity of surveillance falls, the number of traps diminishes (the size of the area where the traps are deployed remains unchanged). Fig. 7 shows that the potential benefits of reducing the probability of outbreaks by 10% increase with augmented surveillance (i.e. reduced spacing between traps). Eradication An increase in the level of eradication effort may result in a reduction of the time to elimination of the pest population. However, it is difficult to establish a relationship between eradication effort and time to eradication. In the economics of pest management, a substantial body of literature examines if eradication is the optimal control strategy (Burnett et al., 2007; Eiswerth and Johnson, 2002; Fraser et al., 2006; Kotani et al., 2009; Myers et al., 1998; Sharov and Liebhold, 1998) and a few studies analyse the optimal eradication effort as a function of the size of the infestation (Olson and Roy, 2002, 2008), but there are no studies that provide insights into the relationship between eradication effort and time to eradication. Moreover, in AWM of Qfly in the FFEZ the eradication of outbreaks follows a standard process, regardless of the size of the infestation. Notably, only the size of the suspension zone changes with the size of the infested area. Hence, we have also expressed a proportional reduction in the time to eradication as a breakeven value. The lower solid line in Fig. 7 shows avoided post-harvest treatment costs following a 10% reduction in the duration of each simulated outbreak for varying spacing distance between traps. Under the current surveillance system (i.e. 1 km between traps in horticultural production areas), the average annual expenditure on post-harvest treatments is reduced by around $0.43 million per year when the average duration of outbreaks is reduced by 10% (which corresponds to point b in Fig. 7). Point a in Fig. 7 indicates the counterfactual (i.e. no changes to the current system). This cost saving represents 15% of the costs given in Table 1, indicating that an increase in eradication capacity is likely to give a lower return than an increase in expenditure on roadblocks. The breakeven value of a 10% reduction in the probability of outbreak (upper dotted line in Fig. 7) is always greater than the breakeven value of a 10% reduction in outbreak duration (lower grey line in Fig. 7) for any trap spacing. This is because there is a perverse effect in which increased surveillance may lead to a greater number of outbreaks declared. If the spacing between traps is reduced, more flies might be caught; but if the threshold number of captures that triggers the declaration of an outbreak remains un-

Table 2 Estimated benefits of a change in the probability of Qfly outbreak. Roadblock protection level

For an entire year Current 10% Reduction in the probability of outbreaks

Probability of outbreaks (mean per week per grid cell over all FFEZ)

0.00028 0.00025

Average annual costs of incursions ($ millions) Average eradication costs per season

Average post-harvest treatments costs per season

1.31 1.18

5.24 4.87

Estimated average annual costs of incursions ($ millions)

Estimated annual benefits of a 10% improvement in the probability of outbreaks ($ millions)

6.55 6.05

0.50

211

30

Benefits of Reducing Outbreak Probability Benefits of Reducing Outbreak Duration

Costs post-harvest treatments Costs surveillance Costs eradication

25

0.5

c b

Total costs

0

1000

2000

3000

4000

20 15 10

-1.0

-0.5

Costs (AUD millions)

0.0

a

-1.5

Cost Savings (AUD millions)

1.0

V. Florec et al. / Food Policy 38 (2013) 203–213

5000

Fig. 7. Benefits (cost savings) from a 10% reduction in the probability of outbreaks due to improved roadblocks and a 10% reduction in the duration of outbreaks due to increased eradication rates against surveillance grid spacing. Note: the distance ca gives the cost saving for a 10% reduction in outbreaks relative to the counterfactual of the current probability of outbreaks and surveillance on a 1 km grid. The distance ba is the cost saving for a 10% reduction in the duration of outbreaks relative to the counterfactual.

changed, some of the declared outbreaks could be redundant since they would be declared even though the risk of spread was very low. Conversely, the number of unnoticed outbreaks for which the risk of establishment of a second Qfly generation is very high increases with less surveillance. Hence, while the duration of each outbreak is 10% shorter, the number of outbreaks declared for varying surveillance grids makes the benefits of reduced outbreak duration less significant. The message from this analysis is that changes in eradication capacity and/or surveillance should be coupled with changes in the market access rules (including size of the suspension zone to its ecological optimum, number of flies that trigger the declaration of an outbreak, and time to area-freedom reinstatement), which could be a complex and lengthy process since it requires trading partners to accept such changes. Surveillance Increased surveillance may lead to a shorter time to detection. A reduction in the current expected time to detection of Qfly in the FFEZ may result in avoided post-harvest treatments and possibly shorter eradication campaigns. Following on previous studies of Qfly surveillance (e.g. Meats, 1998; Meats et al., 2003), we developed a relationship between surveillance effort and the expected time to detection of a Qfly population (see Sadler et al.,

0

5

Spacing of traps (m)

15.2

15.4

15.6

15.8

16.0

16.2

16.4

Expected time to detection (weeks) Fig. 8. Estimated effects of a change in costs of post-harvest treatment, eradication and surveillance arising from a change in the expected time to detection.

2011) and integrated it into the QFAWM model. Hence, it is possible to estimate potential net benefits of different levels of surveillance effort. Although benefit estimates vary throughout the year, for simplicity we have focused on yearly average costs of the different components of AWM. Yearly average costs of post-harvest treatment, eradication and monitoring vary with changes in the expected time to detection (Fig. 8). The expected time to detection with a 1 km grid is approximately 16.3 weeks, which, over the range of surveillance grid densities analysed, minimises costs. Thus an increased investment in more intensive surveillance is not warranted. A decrease in the expected time to detection, while greatly increasing surveillance costs, does not result in a significant reduction in post-harvest treatments and eradication expenditures. In our example, reductions in post-harvest treatments and eradication costs are less than the corresponding increase in surveillance cost. Conclusion In this paper we have evaluated what changes in the implementation of existing AWM tools (roadblocks, surveillance and eradication) have the potential to generate the greatest benefits. Reductions in the probability of outbreaks and the time to eradica-

Table 3 Estimated benefits of a change in the probability of Qfly outbreak with seasonal differences. Roadblock protection level

Summer Current 10% Reduction in the probability of outbreaks Winter Current 10% Reduction in the probability of outbreaks

Probability of outbreaks (mean per week per grid cell over all FFEZ)

Average annual costs of incursions ($ millions)

Estimated average annual costs of incursions ($ millions)

Estimated annual benefits of a 10% improvement in the probability of outbreaks ($ millions)

Average eradication costs per season

Average post-harvest treatments costs per season

0.00039 0.00035

1.79 1.62

3.71 3.46

5.50 5.08

0.42

0.00012 0.00011

0.59 0.53

7.39 6.84

7.98 7.37

0.61

212

V. Florec et al. / Food Policy 38 (2013) 203–213

tion of a Qfly population were expressed as breakeven values. We found that the breakeven value for a 10% reduction in the probability of outbreaks in the FFEZ is greater in winter than in summer, totalling $0.61 million in winter and $0.42 million in summer. We also found that the breakeven value of a 10% reduction in the probability of outbreak is always greater than the breakeven value of a 10% reduction in outbreak duration for all spacing in the trapping grid. A relationship was found between the level of surveillance effort and the expected time to detection of a Qfly population. We found that increases in surveillance effort greatly increased surveillance costs but did not result in significant reductions in post-harvest treatments and eradication expenditures. The majority of the benefits justifying the ongoing funding of AWM programs stem from avoided post-harvest disinfestation costs. Hence assumptions in cost estimation regarding the quantities of produce treated and treatment prices greatly influence the results. Consequently, our estimates of potential benefits and net benefits of changes to AWM tools are sensitive to per unit postharvest treatment costs. This suggests that the design of AWM schemes can be enhanced with the aid of accurate and complete information on the region’s land uses, production per unit area throughout the year, surveillance effectiveness, probability of outbreaks per unit area, post-harvest treatment costs and market distribution per commodity at different times of the year. Our study considered changes in the effort of roadblock protection, surveillance, and eradication. Changes in production or consumption that could potentially reduce the economic significance of a Qfly AWM scheme have not been analysed. Rather than emphasising a benefit maximising mixes of different AWM tools the invasive species problem may be formulated as a risk-benefit trade-off, where the policy maker may reduce the risk of worstcase scenarios by mitigation (e.g., with quarantine restrictions, trade and transport regulations to reduce the risk of introduction, and control efforts to reduce the pest population in case of an outbreak), adaption (i.e., changing production and consumption decisions to reduce the damage caused by the pest) or combining both mitigation and adaptation (Shogren, 2000). Moreover, some of the assumptions we have made could be relaxed. For instance, we assumed that the variable costs of eradication are uniform across the outbreak area, but a more spatially explicit version could be put forward in future study. Likewise, roadblock costs are also uniform across the FFEZ. In future work this could be modified by developing a model where roadblock costs vary depending on location. Our model may also be developed to deal with multiple outbreaks subject to distinct market access rules and for which the time to eradication and size of the area affected would be different. Based on scientific data on Qfly dispersal, new outbreak thresholds and quarantine distances have been proposed (Dominiak, 2012; Dominiak et al., 2011). Also, this paper indirectly contributes to a debate about how to integrate economics and biology initiated by Clark (1976) and recently reviewed in relation to invasive species by Finnoff et al. (2010). In this case the use of a fully specified dynamic and spatial ecological model and a large amount of empirical data on climate, topography, the road system and, trapping data led, we believe, to a qualitatively different analysis of the problem. The advantage of analysing an endemic pest, treated as a biosecurity threat by other countries, is that there is abundant data on its ecology. The model illustrated that the market access rules are all important; seasonality, even down to the week in which an outbreak occurs is critical; and similarly the precise spatial location of an outbreak is also critical. Our approach complements theoretical contributions such as Olson and Roy (2002) which do not address the importance of the duration of qualitative states such as market closure and do not make explicit assumptions about the role of surveillance. In invasive and endemic biosecurity management what you know,

when you know it and for which spatial point is critical. For this reason AWM schemes, such as the GSPFA, remains relevant to a wide range of pest problems. References Acquaye, A.K.A., Alston, J.M., Hyunok, L., Sumner, D.A., 2005. Economic consequences of invasive species policies in the presence of commodity programs: theory and application to citrus canker. Review of Agricultural Economics 27, 498–504. Acquaye, A.K.A., Alston, J.M., Lee, H., Sumner, D.A., 2007. Hurricanes and invasive species: the economics and spatial dynamics of eradication policies. In: Oude Lansink, A.G.J.M. (Ed.), New Approaches to the Economics of Plant Health. Springer, Dordrecht, The Netherlands, pp. 101–116. AQIS, 2008. PHYTO: Plant and Plant Product Export Conditions Database. Australian Quarantine and Inspection Service. . Beare, S., Elliston, L., Abdalla, A., Davidson, A., 2005. Improving Plant Biosecurity Systems: A Cost-Benefit Framework for Assessing Incursion Management Decisions. ABARE eReport 05.10. Bogich, T., Shea, K., 2008. A state-dependent model for the optimal management of an invasive metapopulation. Ecological Applications 18, 748–761. Bomford, M., O’Brien, P., 1995. Eradication or control for vertebrate pests? Wildlife Society Bulletin 23, 249–255. Brown, C., Lynch, L., Zilberman, D., 2002. The economics of controlling insecttransmitted plant diseases. American Journal of Agricultural Economics 84, 279–291. Burnett, K.M., 2006. Introductions of invasive species: failure of the weaker link. Agricultural and Resource Economics Review 35, 21–28. Burnett, K.M., Kaiser, B., Roumasset, J., 2007. Economic lessons from control efforts for an invasive species: Miconia calvescens in Hawaii. Journal of Forest Economics 13, 151–167. Cacho, O.J., Hester, S., Spring, D., 2007. Applying search theory to determine the feasibility of eradicating an invasive population in natural environments. Australian Journal of Agricultural and Resource Economics 51, 425–443. Ceddia, M.G., Heikkilä, J., Peltola, J., 2009. Managing invasive alien species with professional and hobby farmers: insights from ecological-economic modelling. Ecological Economics 68, 1366–1374. Clark, C. W. 1976. Mathematical Bioeconomics. Wiley Interscience, Chichester, UK. Chalak-Haghighi, M., Van Ierland, E.C., Bourdôt, G.W., Leathwick, D., 2008. Management strategies for an invasive weed: a dynamic programming approach for Californian thistle in New Zealand. New Zealand Journal of Agricultural Research 51, 409–424. Clarke, A.R., Powell, K.S., Weldon, C.W., Taylor, P.W., 2011. The ecology of Bactrocera tryoni (Diptera: Tephritidae): what do we know to assist pest management? Annals of Applied Biology 158, 26–54. Clift, A., Meats, A., 2001. Why model Tephritid fruit fly incursions in agricultural production areas? In: Modsim 2001: Proceedings of the International Congress on Modelling and Simulation, Canberra, ACT, pp. 1835–1840. Devorshak, C., 2007. Area-wide integrated pest management programmes and agricultural trade: challenges and opportunities for regulatory plant protection. In: Vreysen, J.B., Robinson, A.S., Hendrichs, J. (Eds.), Area-Wide Control of Insect Pests: From Research to Field Implementation. IAEA, Springer, Dordrecht, The Netherlands, pp. 407–415. Dominiak, B., 2011. Review of grapes Vitis sp. as an occasional host for Queensland fruit fly Bactrocera tryoni (Froggatt) (Diptera: Tephritidae). Crop Protection 30, 958–961. Dominiak, B., Westcott, A.E., Barchia, I.M., 2003. Release of sterile Queensland fruit fly, Bactrocera tryoni (Froggatt) (Diptera: Tephritidae), at Sydney, Australia. Australian Journal of Experimental Agriculture 43, 519–528. Dominiak, B.C., 2012. Review of dispersal, survival, and establishment of Bactrocera tryoni (Diptera: Tephritidae) for quarantine purposes. Annals of the Entomological Society of America 105, 434–446. Dominiak, B.C., Campbell, M., Cameron, G., Nicol, H., 2000. Review of vehicle inspection historical data as a tool to monitor the entry of hosts of Queensland fruit fly Bactrocera tryoni (Froggatt) (Diptera: Tephritidae) into a fruit fly free area. Australian Journal of Experimental Agriculture 40, 763–771. Dominiak, B.C., Coombes, N., 2009. Fruit carrying characteristics of travellers into a quarantine zone in New South Wales in 1999/2000. Plant Protection Quarterly 24, 14–19. Dominiak, B.C., Coombes, N., 2010. Review of the impact of the TriState community fruit fly awareness program on road travellers – 1999/2000. Plant Protection Quarterly 25, 2–8. Dominiak, B.C., Daniels, D., 2012. Review of the past and present distribution of Mediterranean fruit fly (Ceratitis capitata Wiedemann) and Queensland fruit fly (Bactrocera tryoni Froggatt) in Australia. Australian Journal of Entomology 51, 104–115. Dominiak, B.C., Daniels, D., Mapson, R., 2011. Review of the outbreak threshold for Queensland fruit fly (‘Bactrocera tryoni’ Froggatt). Plant Protection Quarterly 26, 141–147. DPI Victoria, 2010. Benefit-Cost Analysis of Options for Managing Queensland Fruit Fly in Victoria, Evaluation Report Series No. 11. Economics and Policy Research Branch, Policy and Strategy Group, Department of Primary Industries, Melbourne, VIC.

V. Florec et al. / Food Policy 38 (2013) 203–213 Eiswerth, M.E., Johnson, W.S., 2002. Managing nonindigenous invasive species: insights from dynamic analysis. Environmental and Resource Economics 23, 319–342. Enkerlin, W.R., 2005. Impact of fruit fly control programmes using the sterile insect technique. In: Dyck, V.A., Hendrichs, J., Robinson, A.S. (Eds.), Sterile Insect Technique: Principles and Practice in Area-Wide Integrated Pest Management. IAEA, Springer, Dordrecht, The Netherlands, pp. 651–676. Faust, R.M., 2008. General introduction to area-wide pest management. In: Koul, O., Cuperus, G., Elliot, N. (Eds.), Area-Wide Pest Management: Theory and Implementation. CAB International, Wallingford, UK, Cambridge, MA, USA, p. 590. Florec, V., Sadler, R., White, B., 2010a. An Economic Analysis of Pest Incursions in an Area-Wide Management of Fruit Flies, Working Paper 1010. School of Agricultural and Resource Economics, University of Western Australia, Crawley, WA. Florec, V., Sadler, R., White, B., 2010b. Surveillance in Fruit Flies Free Areas: An Economic Analysis, Working Paper 1009. School of Agricultural and Resource Economics, University of Western Australia, Crawley, WA. Finnoff, D., McIntosh, C., Shogren, J. F., Sims, C., Warziniack, T., 2010. Invasive Species and Endogenous Risk. In: Rausser, G.C., Smith, V.K., Zilberman, D., (Eds.), Annual Review Of Resource Economics, vol. 2. pp. 77–100. Franco-Dixon, M.A., Chambers, S., 2009. Benefit-cost analysis of area wide management (AWM) of fruit flies in the central Burnett District of Queensland, Selected Paper Prepared for the Presentation at the Agricultural and Applied Economics Association 2009, AAEA and ACCI Joint Annual Meeting, Milwaukee, WI, USA. Fraser, R.W., Cook, D.C., Mumford, J.D., Wilby, A., Waage, J.K., 2006. Managing outbreaks of invasive species: eradication versus suppression. International Journal of Pest Management 52, 261–268. Gilchrist, A.S., Dominiak, B., Gillespie, P.S., Sved, J.A., 2006. Variation in population structure across the ecological range of the Queensland fruit fly, Bactrocera tryoni. Australian Journal of Zoology 54, 87–95. Hendrichs, J., Vreysen, J.B., Enkerlin, W.R., Cayol, J.P., 2005. Strategic options in using sterile insects for area-wide integrated pest management. In: Dyck, V.A., Hendrichs, J., Robinson, A.S. (Eds.), Sterile Insect Technique: Principles and Practice in Area-Wide Integrated Pest Management. IAEA, Springer, Dordrecht, The Netherlands, pp. 563–602. Hennessy, D.A., 2008. Biosecurity incentives, network effects, and entry of a rapidly spreading pest. Ecological Economics 68, 230–239. Hinchy, M.D., Fisher, B.S., 1991. A Cost-Benefit Analysis of Quarantine, Technical Paper No. 91-3, Project 1254.101. Australian Bureau of Agricultural and Resource Economics, Canberra, ACT. Jessup, A.J., Dominiak, B., Woods, B., De Lima, C.P.F., Tomkins, A., Smallridge, C.J., 2007. Area-wide management of fruit flies in Australia. In: Vreysen, J.B., Robinson, A.S., Hendrichs, J. (Eds.), Area-Wide Control of Insect Pests: From Research to Field Implementation. IAEA, Springer, Dordrecht, The Netherlands, pp. 685–697. Klassen, W., 2005. Area-wide integrated pest management and sterile insect technique. In: Dyck, V.A., Hendrichs, J., Robinson, A.S. (Eds.), Sterile Insect Technique: Principles and Practice in Area-Wide Integrated Pest Management. IAEA, Springer, Dordrecht, The Netherlands, pp. 39–68. Kompas, T., Che, T.N., 2009. A Practical Optimal Surveillance Measure: The Case of Papaya Fruit Fly in Australia. Australian Centre for Biosecurity and Environmental Economics, Crawford School of Economics and Government, Australian National University, Canberra, ACT. Kotani, K., Kakinaka, M., Matsuda, H., 2009. Dynamic economic analysis on invasive species management: some policy implications of catchability. Mathematical Biosciences 220, 1–14. Koul, O., Cuperus, G., Elliot, N., 2008. Area-Wide Pest Management: Theory and Implementation. CAB International, Wallingford, UK, Cambridge, MA, USA, p. 590. Lichtenberg, E., Lynch, L., 2006. Exotic pests and trade: when is pest-free status certification worthwhile? Agricultural and Resource Economics Review 35, 52– 62. Lindner, B., Macleod, P., 2009. ACIAR’s 25 year investment in fruit-fly research, Contributed Paper Presented to the 53rd Annual Conference of the Australian Agricultural and Resource Economics Society. Cairns, QLD. Lloyd, A.C., Hamacek, E.L., Kopittke, R.A., Peek, T., Wyatt, P.M., Neale, C.J., Eelkema, M., Gu, H., 2010. Area-wide management of fruit flies (Diptera: Tephritidae) in the Central Burnett district of Queensland, Australia. Crop Protection 29, 462– 469. Mau, R.F.L., Jang, E.B., Vargas, R.I., 2007. The Hawaii area-wide fruit fly pest management programme: influence of partnerships and a good education

213

programme. In: Vreysen, J.B., Robinson, A.S., Hendrichs, J. (Eds.), Area-Wide Control of Insect Pests: From Research to Field Implementation. IAEA, Springer, Dordrecht, The Netherlands, pp. 671–683. Meats, A., 1998. The power of trapping grids for detecting and estimating the size of invading propagules of the Queensland fruit fly and risks of subsequent infestation. General and Applied Entomology 28, 47–55. Meats, A., Edgerton, J.E., 2008. Short- and long-range dispersal of the Queensland fruit fly, Bactrocera tryoni and its relevance to invasive potential, sterile insect technique and surveillance trapping. Australian Journal of Experimental Agriculture 48, 1237–1245. Meats, A.W., Clift, A.D., Robson, M.K., 2003. Incipient founder populations of Mediterranean and Queensland fruit flies in Australia: the relation of trap catch to infestation radius and models for quarantine radius. Australian Journal of Experimental Agriculture 43, 397–406. Mumford, J.D., 2000. Economics of area-wide pest management. In: Tan, K.H. (Ed.), Area-Wide Control of Fruit Flies and Other Insect Pests. Penerbit Universiti Sains Malaysia, Pulau Pinang, Malaysia. Mumford, J.D., 2002. Economic issues related to quarantine in international trade. European Review of Agricultural Economics 29, 329–348. Mumford, J.D., 2005. Application of benefit/cost analysis to insect pest control using the sterile insect technique. In: Dyck, V.A., Hendrichs, J., Robinson, A.S. (Eds.), Sterile Insect Technique: Principles and Practice in Area-Wide Integrated Pest Management. IAEA, Springer, Dordrecht, The Netherlands, pp. 481–498. Muthuthantri, S., Maelzer, D., Zalucki, M.P., Clarke, A.R., 2010. The seasonal phenology of Bactrocera tryoni (Froggatt) (Diptera: Tephritidae) in Queensland. Australian Journal of Entomology 49, 221–233. Myers, J.H., Savoie, A., van Randen, E., 1998. Eradication and pest management. Annual Review of Entomology 43, 471–491. NFFWG, 2008. National Code of Practice for the Management of Queensland Fruit fly, Bactrocera tryoni (Froggart), Draft, November 2008, Version 18. National Fruit Fly Working Group, Australia. Olson, L.J., Roy, S., 2002. The economics of controlling a stochastic biological invasion. American Journal of Agricultural Economics 84, 1311–1316. Olson, L.J., Roy, S., 2008. Controlling a biological invasion: a non-classical dynamic economic model. Economic Theory 36, 453–469. Perrings, C., Williamson, M., Barbier, E., Delfino, D., Dalmazzone, S., Shogren, J.F., Simmons, P., Watkinson, A., 2002. Biological invasion risks and the public good: an economic perspective. Conservation Ecology 6. PricewaterhouseCoopers, 2001. Benefit Cost Analysis of the TriState Fruit Fly Strategy, Report Prepared for the Department of Agriculture, Fisheries and Forestry, Canberra, ACT. Sadler, R.J., Florec, V., White, B., Dominiak, B.C., 2011. Calibrating a jump-diffusion model of an endemic invasive: Metamodels, statistics and Qfly, Paper Presented at the International Congress on Modelling and Simulation, MODSIM 2011, Perth, WA. Sharov, A.A., Liebhold, A.M., 1998. Bioeconomics of managing the spread of exotic pest species with barrier zones. Ecological Applications 8, 833–845. Shogren, J.F., 2000. Risk reductions strategies against the ‘‘explosive invader’’. In: Perrings, C., Williamson, M., Dalmazzone, S. (Eds.), The Economics of Biological Invasions. Edward Elgar, Northhampton, MA, USA. Stohlgren, T.J., Schnase, J.L., 2006. Risk analysis for biological hazards: what we need to know about invasive species. Risk Analysis 26, 163–173. Sutherst, R.W., Collyer, B.S., Yonow, T., 2000. The vulnerability of Australian horticulture to the Queensland fruit fly, Bactrocera (Dacus) tryoni, under climate change. Australian Journal of Agricultural Research 51, 467–480. Tan, K.H., 2000. Area-wide control of fruit flies and other insect pests: joint proceedings of the International Conference on Area-Wide Control of Insect Pests, May 28–June 2, 1998 and the Fifth International Symposium on Fruit Flies of Economic Importance, June 1–5, 1998, Penang, Malaysia. Penerbit Universiti Sains Malaysia, Pulau Pinang, Malaysia, p. 782. Vijaysegaran, S., 2008. Establishment of Fruit Fly Pest-free Areas. Final Report project CP/2007/002. Canberra, ACT, Australian Centre for International Agricultural Research (ACIAR). Vreysen, M.J.B., Robinson, A.S., Hendrichs, J., 2007. Area-Wide Control of Insect Pests: From Research to Field Implementation. IAEA. Springer, Dordrecht, The Netherlands, p. 792. White, B., Sadler, R.J., Florec, V., Dominiak, B.C., Morey, K., Buetre, B., Abougamos, H., 2011. Optimal Investment in R&D for Plant Biosecurity, Final Report Project CRC70100, Submitted to the Cooperative Research Centre for National Plant Biosecurity. Yonow, T., Sutherst, R.W., 1998. The geographical distribution of the Queensland fruit fly, Bactrocera (Dacus) tryoni, in relation to climate. Australian Journal of Agricultural Research 49, 935–953.