NeoBiota | 8: 67-8 | (20 | 3) A peer-reviewed open-access journal 7. e doi: 10.3897/neobiota. 18.3863 %) NeoBiota www.pens oft. n et/j journa | s/ neob iota Advancing research on alien species and biological invasions An assessment of the benefits of yellow Sigatoka (Mycosphaerella musicola) control in the Queensland Northern Banana Pest Quarantine Area David C. Cook'?34, Shuang Liu***, Jacqueline Edwards”®, Oscar N. Villalta**, Jean-Philippe Aurambout”’, Darren J. Kriticos?>*’, Andre Drenth"®, Paul J. De Barro*!! | Department of Agriculture and Food, Bunbury, Western Australia, Australia 2. Plant Biosecurity Cooperative Research Centre, Bruce, Australian Capital Territory, Australia 3 The University of Western Australia, Crawley, Western Australia, Australia 4 Australian Centre for Biosecurity and Environmental Economics, Canberra, Australian Capital Territory, Australia § CSIRO Ecosystem Sciences, Canberra, Australian Capital Territory, Australia 6 Department of Primary Industries, Knoxfield, Victoria, Australia 1 Department of Primary Indu- stries, Parkville, Victoria, Australia 8 Charles Sturt University, Barton, Australian Capital Territory, Australia 9 The University of Minnesota, Minneapolis, Minnesota, United States \0 The University of Queensland, Dutton Park, Queensland, Australia \\ CSIRO Ecosystem Sciences, Dutton Park, Queensland, Australia Corresponding author: David C. Cook (david.cook@agric.wa.gov.au) Academic editor: 7. Kii/n | Received 16 August 2012 | Accepted 29 April 2013 | Published 13 September 2013 Citation: Cook DC, Liu S, Edwards J, Villalta ON, Aurambout J-P, Kriticos DJ, Drenth A, De Barro PJ (2013) An assessment of the benefits of yellow Sigatoka (Mycosphaerella musicola) control in the Queensland Northern Banana Pest Quarantine Area. In: Kriticos DJ, Venette RC (Eds) Advancing risk assessment models to address climate change, economics and uncertainty. NeoBiota 18: 67-81. doi: 10.3897/neobiota. 18.3863 Abstract The banana leaf spotting disease yellow Sigatoka is established and actively controlled in Australia through intensive chemical treatments and diseased leaf removal. In the State of Queensland, the State government imposes standards for de-leafing to minimise the risk of the disease spreading in 6 banana pest quarantine areas. Of these, the Northern Banana Pest Quarantine Area is the most significant in terms of banana production. Previous regulations imposed obligations on owners of banana plants within this area to re- move leaves from plants with visible spotting on more than 15 per cent of any leaf during the wet season. Recently, this leaf disease threshold has been lowered to 5 per cent. In this paper we examine the likely impact this more-costly regulation will have on the spread of the disease. We estimate that the average net benefit of reducing the diseased leaf threshold is only likely to be $1.4 million per year over the next 30 years, expressed as the annualised present value of tightened regulation. This result varies substantially when the timeframe of the analysis is changed, with shorter time frames indicating poorer net returns from the change in protocols. Overall, the benefit of the regulation change is likely to be minor. Copyright David C. Cook et al. This is an open access article distributed under the terms of the Creative Commons Attribution License 3.0 (CC-BY), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. 68 David C. Cook et al. / NeoBiota 18: 67-81 (2013) Keywords Biosecurity, invasive alien species, banana pathogen, yellow Sigatoka Introduction The Sigatoka disease complex affects banana cultivation in many countries. The dis- ease yellow Sigatoka (Mycosphaerella musicola) is established and actively controlled in Australia through intensive fungicide treatments and diseased leaf removal (Henderson et al. 2006). Although less virulent than the malign black Sigatoka (MZ. fijiensis), M. musicola also imposes substantial costs on affected banana growers. ‘This is particularly true for Australia’s premier banana growing regions located in the State of Queensland, including the Innisfail-Tully area. In 2011, the State as a whole produced 279.1 thou- sand tonnes of bananas valued at $448.3 million from a plantation area of approxi- mately 10 100 hectares (ABS 2012). In all areas of the State where M. musicola occurs, grower costs are incurred through compliance with government-imposed standards for de-leafing to minimise the risk of the disease’s spread. The Plant Protection Regulation 2002 was put in place under the Plant Protection Act 1989 which defined 6 banana pest quarantine areas (State of Queensland 2002). Of these, the Northern Banana Pest Quarantine Area (NBPQA) is the most significant, encompassing over 80 per cent of the State’s banana production. The Plant Protection Regulation 2002 imposed an obligation on the owner of land in a pest quarantine area to treat every banana plant by removing every leaf from the plant that has visible symptoms of M/. musicola (and another endemic disease, banana leaf speckle (/Z. musae)) on more than 15 per cent of any leaf at any time between 1 November and 31 May or on more than 30 per cent of any leaf at any time between 1 June and 31 October. An amendment to the Plant Protection Regulation 2002, the Plant Protection Amendment Regulation (No. 4) 2003, was subsequently put in place in response to concerns that the de-leafing standards initially imposed were too permissive. In par- ticular, during wet season conditions in the NBPQA the 15 per cent de-leafing thresh- old was deemed insufficient to prevent MZ. musicola and M. musae from spreading. Moreover, the 30 per cent action level in the dry season was thought to be far too high for wet weather conditions highly conducive to disease spread (State of Queensland 2003). The Amendment imposed a lower threshold of 5 per cent throughout the year in the NBPQA. As deleterious as these amended regulations appear to be in terms of the foliage carried by commercial banana plants, the impact on production volume is likely to be minimal. During their life, individual banana plants may produce 30 or more leaves, which is surplus to their phosynthetic needs. ‘The oldest leaves are shed at a rate of ap- proximately 1 leaf every 10 to 12 days so that when the fruit bunch emerges from the top of the pseudostem the plant has an average of 15 leaves. After the bunch shoots no An assessment of the benefits of yellow Sigatoka (Mycosphaerella musicola)... 69 new leaves are produced. The oldest leaves of the plant continue to fall until, at harvest, between 6 and 8 leaves remain (Ostmark 1974). While the incidence of leaf disease is likely to be reduced if stricter thresholds are implemented and maintained over time, additional costs to banana growers in the NBPQA will apply. These include substantial increases in chemical treatment and ap- plication costs in addition to more rigorous de-leafing cycles. In this paper we estimate the likely change in net returns to the banana industry in the NBPQA from adopting the new 5 per cent de-leafing threshold. Methods The stochastic simulation model used in this assessment determines total expected (or probability-weighted) damage from M. musicola in the NBPQA over a 30-year period under both a 15 per cent and a 5 per cent de-leafing threshold. Uncertain or variable parameters are specified as probability distributions. 10 000 model iterations are run using values randomly sampled across the range of each distribution using a Latin hy- percube sampling algorithm. The total damage banana producers in the NAPQA experience because of the dis- ease in time period ¢ (d) is estimated by: 4,=YPA+VA, (1) where: Y is the mean change in banana yield resulting from infection (assumed 100 per cent) in year 4 P is the prevailing domestic price for bananas in year #; V, is the increase in variable cost of production per hectare induced by MM. musicola on- plantation management methods in year 4 and A_ is the area infected with VM. musicola in year ¢. A stratified diffusion model combining both short and long distance dispersal pro- cesses is used to predict A. Parameter estimates for this model appear in Table 1, and are explained below. Note that due to the uncertainty surrounding some of these parameters, they are specified using a range of distributional forms, rather than simple point estimates. Types of distributions used in the table include: (a) pert — a type of beta distribu- tion specified using minimum, most likely (or skewness) and maximum values often preferred when parameters are reliant a number of sources (or expert opinions) since the mean is relatively insensitive to minimum and maximum values compared to the most likely value; (b) uniform —a rectangular distribution bounded by minimum and maximum values used to highlight the fact that there is little known about a parameter (Vose 2008). The dispersal model is derived from the reaction diffusion models originally de- veloped by Fisher (1937) which have been shown to provide a reasonable approxima- tion of the spread of a diverse range of organisms (Cook et al. 2011a; Dwyer 1992; 70 David C. Cook et al. / NeoBiota 18: 67-81 (2013) Table |. Parameter values Description 15% de-leafing threshold 5% de-leafing threshold Detection probability (%). 100 100 Infection diffusion coefficient, D (m7/yr). * Pert(2.0x10°,3.5x10*,5.0x10%) | Pert(0.0,1.0x107,2.0x 107) Percentage of total NBPQA plantation area infected in the first time step (%). * Pere) D5 1952-0) Bent. 22) Minimum area infected, A™” (m7). 1.0x10° 1.0x10° Maximum area infected, A™™ (m7). $ 9.8x107 9.8x10’ Intrinsic sate OR IniectiOn and density Pert (0.00,0.01,0.02) Pert (0.00,0.01,0.02) increase, r(yr'). * Minimum infection density, N™'" (#/m/”). 1.0x10+ 1.0x10“ Maximum infection density, K (#/m/?). * Pert(100,550, 1000) Pert(100,550,1000) Minimum number of satellite sites generated 1 1 in a single time step, S™” (#). Maximum number of satellite sites generated Pert(0,5,10) Pert(0,5,10) in a single time step, S™™* (#). * Intrinsic rate of new foci generation per unit area of infection, y (#/m?’). Demand elasticity. | Uniform(-1.1,-1.0) Uniform(-1.1,-1.0) Prevailing market price of bananas in the first Pert(1.0x10°,3.0x107,5.0x 107) | Pert(1.0x107,3.0x107,5.0x 107) 1 time step ($/T). § ey a: Maximum area considered for eradication, 0 0 Aerad (ha). tia costs upon detection — chemical | (8 95.103,1.1«108,1.3x104) | Pert(1.6%104,5.0x104,6.6x104) Treatment costs upon detection — de-leafing : : ‘ , 2 7 ($/ha). * Pert(1.4x10%,2.1x10°,2.8x10°) | Pert(2.1x10*,3.1x10°,3.2x10°) Yield reduction despite control (%). Pert(0.0,2.5,5.0) Pert(0.0,0.5, 1.0) Discount rate (%). ™ 5 5 " Specified with reference to Cook (2003) and Waage et al. (2005). * Derived from Peterson et al. (2005). S ABS (2012), Note lha = 10 000m’. | Ulubasoglu et al. (2011). ‘ Assumes: (i) average density of planting of 2 000 stems/ha and removal, (ii) control of M. musicola in the NBPQA involves applications of dithane (at 3kg/ha or $21.60/ha) and oil (at 3L/ha or $8.85/ha) at week- ly intervals during the wet season (Cook 2003); (iii) it is desirable for growers to rotate the use of dithane and oil with propiconazole (at 0.3L/ha or $22.20/ha) to manage resistance (Cook 2003); (iv) 15 to 25 cycles of fungicides are used for control of M. musicola in the NBPQA to comply with a 15% de-leafing threshold; (v) an additional 5 to 10 spray cycles are needed to comply with a 5% de-leafing threshold. * De-leafing plantations to control M. musicola to a 15% threshold occurs up to 15 times per season. Assume an additional 5 to 10 de-leafing cycles are necessary to achieve a 5% threshold at a cost of $140/ha each. Commonwealth of Australia (2006). Holmes 1993; McCann et al. 2000; Okubo and Levin 2002). These models assert that an invasion diffusing from a point source will eventually reach a constant asymptotic radial spread rate of 2, / rD, in all directions, where r describes a growth factor for MV. An assessment of the benefits of yellow Sigatoka (Mycosphaerella musicola)... 71 musicola per year in the NBPQA (assumed constant over all infected sites) and D is a diffusion coefficient for an infected site with an age index ; (indicating the time step in which the site was originally formed) in the NBPQA (Cook et al. 201 1a; Hengeveld 1989; Lewis 1997; Shigesada and Kawasaki 1997). Hence, we assume that an original infection in a previously unaffected area takes place in a homogenous environment within the NBPQA and expands by a diffusive process such that area infected at time t, a,, can be predicted by: ay = r(2tv rD ‘al = 4D, art’. (2) We assume D. is constant across all sites with an age index j, so ignore demo- graphic stochasticity and consequent non-uniform invasion. Since the two control strategies we are considering (i.e. 5 per cent and 15 per cent de-leafing regulations) are very similar, many of the parameters remain the same for both scenarios. But, D is as- sumed to be lower under the 5 per cent de-leafing threshold due to increased chemical suppression limiting local dispersal opportunities for the disease. The density of M. musicola infection within a, influences the control measures required to counter the effects of infection and thus partially determines the value of A. We assume that within each site with age index / affected, the infection density, NV. grows over time period ¢ following a logistic growth curve until the carrying Space of the host environment, K, is reached: a K, Nee rt jt — K, + NV" (e" - —1) : (3) Here, Ne is the size of the original infection at site with age index j and r is the intrinsic rate of density increase (assumed to be the same as the intrinsic rate of in- fection increase) (Cook et al. 2011b). Note that for simplicity we assume NV, and Y are independent of one another. In addition to a, and N,, the size of A, depends on the number of nascent foci or satellite infection sites in year t, 5, which can take on a maximum value of sin any year (Moody and Mack 1988). ‘These sites result from events external to the initial outbreak itself, such as weather phenomena, animal or human behaviour, which periodically jump the expanding infection beyond the infec- tion front (Cook et al. 201 1b). We use a logistic equation to generate changes in s_ as an infectious outbreak continues: max min _ ut s S @€@ MS ot a ae t as aly ie (e” —1) (4) where y is the intrinsic rate of new foci generation (assumed constant over time) ands . is the minimum number of satellite sites generated. Given equations (1)-(4), we can express A as: A, = Qala) where OS A. < Ae 6) Fs 72 David C. Cook et al. / NeoBiota 18: 67-81 (2013) Spread area, infection density and the number of foci are combined with the prob- ability of entry and establishment in an expression of probability-weighted, or ex- pected damage over time. Assuming a discount rate 6, the present value of expected damage after ¢ time periods (TC’) is: r =f TC’ = pal 7 6) ao (6) t= This expression provides us with an estimate of infection-induced producer losses over time. It therefore provides an indication of the economic significance of M. musi- cola over time given a de-leafing protocol. If we denote the total expected damage un- der a 15 per cent and a 5 per cent de-leafing protocol TC’,.,, and TC’.,., respectively, we can determine the likely change in expected damage (ATC?) from adopting the new 5 per cent protocol as: ATC? = ATC? —ATC!,, (7) If indeed the 5 per cent de-leafing protocol is more effective than the previous 15 per cent protocol at reducing MZ. musicola prevalence and impact over time, we would expect ATC” >0. Results M. musicola is assumed to be present within the NBPQA at the beginning of time period 1. Therefore, the resultant expected spread area values calculated from 10 000 iterations of the model are positive, as revealed by Figure 1. Spread is predicted to be very slow in both the 5 per cent and 15 per cent de-leafing protocol scenarios due to the effectiveness of chemical and de-leafing treatments applied simultaneously. The box-whisker plot used in this Figure (and in Figures 2 to 4 to follow) shows the extent of uncertainty in the model predictions dictated by the uncertainty in param- eter specification. The box-whisker plot shows the 25™ percentile of the frequency distribution of model outcomes, the median (i.e. the 50™ percentile), the 75" per- centile and remaining values up to and including the 5° and 95" percentiles of the frequency distribution of model outcomes. Figure 2 illustrates how the resultant TC’, so and TO™,, (i.e. see equation (6)) are expected to change over the 30-year period of the simulation. Here, the mean values of TC? .,,and TC”... predicted by the model in each year are plotted with 10" and 50" percentiles of the frequency distribution of model outcomes. All projected costs are discounted at 5 per cent per annum. By the 30" year, TC’. ..., is expected to average just under $30 million per year, and TC’,,, just under $15 million per year. Note that despite the area affected by the disease remaining relatively constant in both control scenarios, the erosive effects of the discount rate lead to a gradual decline in present value of future expected annual industry damage. 600 a (=) (=) Area Affected (ha) aS oO oO w [=] oO 200 An assessment of the benefits of yellow Sigatoka (Mycosphaerella musicola)... 73 95% 75% Mean 25% 5% 15% Threshold i ‘| reshold T_T T rr —T : T_T ~-~ WN Oo rN OnrwmoedoeaeoeredaAa OvwrTrnonooaoernrwdmdoeoeoe-r- N SSS ES eS Re oe) Col on Die Sy Oe 9 i Pi gh i 9 i gS 9 Dn 9 9? 9? 9 2 9 9? 9 9 9? 9? a a 0 Oo'-o ©: (oo. 2 ao oa Oo oc &' & Ov aS o oO & 6. © & 86 Oo Oo oOo 2 O&O © NNNNNNNNNNNNNNNNNNNNNNNNNNNN NN EN Year Figure |. Expected area of commercial banana plantations affected by yellow Sigatoka in Australia under different management guidelines. 80 70 60 50 40 30 Values in Millions ($) 20 95% 75% Mean 25% 5% 5% era a SS SS SS Se 2021 2022 2023 - 2024 2025 - 2026 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 Figure 2. Predicted industry losses from yellow Sigatoka in Australia under different management guidelines. Figure 3 illustrates how the difference between TC’ and TC? 15% (i.e. ATC? in 5% equation (7)) is expected to change over time, and therefore the relative merit in the banana industry choosing a 5 per cent de-leafing protocol over a 15 per cent protocol 74 David C. Cook et al. / NeoBiota 18: 67-81 (2013) on Oo NO yr WwW oO F NN OO OO Oo Om om © oo Co oo 000 9 NN SN NNN NN AN Figure 3. Predicted gross benefit of adopting a 5 per cent de-leafing threshold for yellow Sigatoka sup- 95% 75% Mean 25% 5% Ee T T T is) = fot Oo i] Oo N SS SE Se Values in Millions ($) o oa 8 a 8 & ——————————s 3 a 4 eee 5 ey 6 lia 7 SS, } 9 Sy ay 0, 2021 -n 2022 ae 2023 Sy 2024 Sy LT 2025 a ———— 2031 2040 - 2041 pression in the NBPQA relative to a 15 per cent protocol. in the NBPQA. Over the 30 years simulated by the mode, the annualised present value of benefit to producers is $11.3 million. But, as Figure 3 shows there is a large amount of uncertainty associated with predicting this far into the future (i.e. the standard de- viation of the benefit estimate is $9.3 million). Discussion Economic research in the area of invasive species has grown substantially in the last 20 years from a modest base (Born et al. 2005; Colautti et al. 2006). Most research has in- volved retrospective assessments of invasive species impacts and management (Born et al. 2005; Naylor 2000). Where predictive models have been used, analyses have tended to be deterministic, thereby neglecting uncertainties in species behaviour and environmen- tal interactions. For example, State of Queensland (2003) includes a cost benefit analysis of the change in disease threshold in the NBPQA which reveals little about the flow of producer benefits over time. Point estimates of the net producer impact of a 5 per cent de-leafing threshold (-$50 000 in the first year, -$20 000 in the second year, $0 in the third year and $400 000 by the end of the fourth year) are provided (State of Queens- land 2003), but no indication is given about possible variation around these estimates. Moreover, information about their derivation in the Amendment text is not provided. In contrast, the predictive model presented in this paper provides a more open and transparent means of summarising complex interactions between natural processes and An assessment of the benefits of yellow Sigatoka (Mycosphaerella musicola)... FD land managers over time for a policy audience. Policy-makers face a difficult challenge because invasive species impacts change with respect to time, space and other variables in ways that are difficult to predict (Regan et al. 2002). They therefore require tools that are explicit about uncertainty and management options that are both precaution- ary and adaptive (Doak et al. 2008; Simberloff 2005). Rather than over-simplified point estimates, the model we have developed takes into account system dynamics and conveys the natural variation of the system to decision-makers when analysing the intertemporal effects of changing de-leafing protocols. Several studies have integrated established ecological models (including reaction- diffusion, stratified diffusion and predator-prey models) with economic management frameworks for invasive species using comparable approaches (Barbier 2001; Cacho et al. 2008; Carrasco et al. 2010; Hyder et al. 2008; Sharov and Liebhold 1998). Others have emphasised spatially explicit approaches using stochastic simulations combining environmental variables and dissemination behaviours to characterise uncertainty in spread patterns over time (Rafoss 2003; Yemshanov et al. 2009). Other recent analyti- cal spatial studies of invasive species spread and control have relied on metapopulation models (Albers et al. 2010; Sanchirico et al. 2010). At the cost of not producing spatially explicit outputs, our model provides a more accurate estimation of the economic impacts of invasive species by incorporating par- tial equilibrium models. ‘This approach allows a detailed examination of changes in producer (and consumer) welfare in domestic (e.g. Cook 2008; Heikkila and Peltola 2004; Paarlberg et al. 2003) and export markets (e.g. Acquaye et al. 2005; Cook and Fraser 2008) induced by invasive species. Flow-on effects of incursions to other areas of the economy can also be estimated using general equilibrium models (Wittwer et al. 2005). However, these assume perfect coordination of trading mechanisms and maximise a common sector utility function making it difficult to integrate them with ecological processes (Carrasco et al. 2012). Similar problems arise with partial equilibrium models due to their aggregated and compact nature, and their integration with ecological spread requires the use of exogenous assumptions regarding the effect that an invasive species will have on the supply curve of the host commodity (Acquaye et al. 2005; Paarlberg et al. 2003). In this analysis we have been explicit about these assumptions with the notes accompany- ing Table 1 providing technical details of the way we assume grower behaviour is likely to change with a 5 per cent disease threshold compared to a 15 per cent threshold. Specifically, we assume: (i) An average density of planting of 2 000 stems per hectare and removal; (ii) Control of M. musicola in the NBPQA involves applications of dithane (at 3 kg per hectare or $22 per hectare) and oil (at 3 L per hectare or approximately $9 per hectare) at weekly intervals during the wet season (Cook 2003); (iii) | Growers rotate the use of dithane and oil with propiconazole (at 0.3 L per hec- tare or $22 per hectare) to manage resistance (Cook 2003); 76 David C. Cook et al. / NeoBiota 18: 67-81 (2013) (iv) 15 to 25 cycles of fungicides are used for control of M. musicola in the NBPQA to comply with a 15 per cent de-leafing threshold, and an additional 5 to 10 spray cycles are needed to comply with a 5 per cent de-leafing threshold; (v) De-leafing plantations to control MZ. musicola to a 15 per cent threshold occurs up to 15 times per season, and an additional 5 to 10 de-leafing periods are nec- essary to achieve a 5 per cent threshold at a cost of $140 per hectare each. Extrapolating across the entire NBPQA, these assumptions imply that producer costs will rise by approximately $43.8 million under the 5 per cent de-leafing thresh- old (Table 2). This includes both de-leafing and chemical application costs required in areas affected by M. musicola. Since the market is closed to imports, the domestic price of bananas will increase as producers pass these cost increases on to consumers. ‘The extent of the price change is predicted in the model using the elasticity of demand (i.e. the ratio of percentage quantity changes over percentage price changes). Thus, our producer behavioural assumptions combine with the spread model to simulate changes in the market over time. Note that the costs indicated in Table 2 are annualised average costs estimated across the whole NBPQA attributable to compliance with the leaf disease thresholds. The right hand column labelled B-A represents the annualised increase in chemical and de-leafing costs imposed by the lower threshold of 5 per cent. By comparing the present value (i.e. the discounted or real) value of these annual cost increments to the predicted benefits derived from Figure 3, we can estimate the likely change in net returns (i.e. benefits minus costs) to the NBPQA from adopting this lower standard. While Table 2 summarises our analysis, it does not show the extent to which likely net benefits change depending on the time frame of the analysis. Net returns on a year-by-year basis are shown in Figure 4, which reveals the full extent of uncertainty surrounding possible net returns to the region in each time step. Initially, due to the increased cost of compliance to the 5 per cent leaf disease threshold, net costs (i.e. a surplus of costs over benefits) are likely to result in the short term. However, after a period of time (between 7 and 14 years) the benefits generated by lower MZ. musicola prevalence and impact begin to outweigh compliance costs. By the end of the simula- tion period, net benefits are likely to be over $13 million per annum. On average, over the 30-year model simulation period, the annualised present val- ue of net benefit to the banana industry in the NBPQA from the adoption of the more stringent leaf disease threshold is estimated to be $1.4 million. Considering this benefit accrues over an area of approximately 10 100 hectares, the impact of the change in dis- ease thresholds appears to be marginal. If we calculate average net returns over a 20-year period, we find that a net cost of the order of -$3.4 million per annum is likely to result. As Figure 4 clearly shows, this is due to the large net costs concentrated in the early years of adopting the new threshold. The further forward in time we project, the larger the likely returns to the banana industry of imposing the stricter leaf disease threshold. Given the intertemporal nature of cost accrual, our model clearly communicates the importance of the timeframe being considered for a policy choice to decision- An assessment of the benefits of yellow Sigatoka (Mycosphaerella musicola)... 77 Table 2. Annualised cost of adopting a 5 per cent de-leafing threshold for yellow Sigatoka suppression in the NBPQA relative to a 15 per cent protocol aggregated across the region. Description 15% de-leafing threshold (A) | 5% de-leafing threshold (B) | B-A Chemical treatment costs (§ million) 31.3 De-leafing costs (§ million) 12.5 Total ($ million) 43.8 ml 20 - | 10 | ma + w 6 = -10 - 95% = 75% £ w ~20 + Mean a | — 25% o > -320 - 5% -40 - -50 + -60 +——_— Te be ee a ee ee ee ee ee ae ee NO wrnwonrwvodoe&#e#eunNnovwreoworowoooqoue3ewrinNnNt MOvwvFuodgenre~edoODo= Se oe ee, Se ee NNN NDennnenanananaenononmnnn mo mn gn mom OT + oo ooooo oooo ogo o oooiaocoooo ogo ooooooo o NNN NNNN NNN NNN DNDN DNDN NNN NNN NNN ON ON ON ON Year Figure 4, Predicted net benefit of adopting a 5 per cent de-leafing threshold for yellow Sigatoka suppres- sion in the NBPQA relative to a 15 per cent protocol. makers. Assuming they prefer to consider a 30-year time period, the annualised average present value of benefits expected to result from tightening the de-leafing threshold is likely to be small, but positive. Shorter time frames suggest the net benefits will be smaller, and (if less than a 10-year time frame is considered) possibly negative. We should also point out that if decision-makers apply a higher (personal) discount rate of 10 per cent to the mean or average model calculations, as opposed to a public/social discount rate of 5 per cent, the mean net benefit to the banana industry would fall to -$1.6 million over 30 years. This highlights the importance of both the choice of time frame and the choice of discount rate. While the modelling framework we have developed provides a solid foundation over which other comprehensive economic analyses of invasive species effects can be performed, future extensions to the model may be warranted in some situations. These could include the adoption of an ecosystems approach within the bioeconomic model to capture interactions between invasive and native species (Hulme 2006). It 78 David C. Cook et al. / NeoBiota 18: 67-81 (2013) may also be necessary to use more complex biophysical modelling of susceptibility and resilience to infection where polyphagous species are concerned (e.g. Hester and Cacho 2003). In these cases the importance of potential non-market (e.g. impacts on native biota, environmental costs due to the use of fungicides, etc.) and indirect market (e.g. impacts on input markets as industries are affected by invasive species) effects may also be critical factors that policy-makers wish to consider. Conclusion In a plant biosecurity context, it is often difficult to predict policy benefits over time due to complex biophysical interactions between invasive species, their hosts and the environment. In this paper, we have demonstrated how a bioeconomic analysis can help decision-makers using the example of M. musicola. We have developed an analyt- ical approach using a stratified diffusion spread model to simulate the likely benefits of adopting a 5 per cent de-leafing threshold for MZ. musicola suppression in the NB- PQA relative to a 15 per cent protocol. Using Monte Carlo simulation to generate a range of possible incursion scenarios over a 30-year time period, we predict that in an average year this new protocol will only reduce the net impact of the disease by $1.4 million. However, this result varies substantially depending on the timeframe over which benefits and costs are projected, with shorter time frames indicating poorer net returns from the change in protocols. Compounding this issue is the uncertainty involved in predicting policy impacts further into the future. It is therefore difficult to evaluate the 5 per cent de-leafing protocol. But, if we take the 30-year timeframe presented in our analysis as a guide, the effects of the policy are likely to be positive, but relatively minor. Acknowledgements The authors would like to acknowledge the support of the Australian Plant Biosecurity Cooperative Research Centre, established and supported under the Australian Gov- ernment’s Cooperative Research Centres Program. We would also like to thank two anonymous reviewers for their helpful comments on the paper. References ABS (2012) Agricultural Commodities 2010-11, Australian Bureau of Statistics, Canberra, 36 pp. Acquaye AKA, Alston MJ, Hyunok ML, Sumner AD (2005) Economic consequences of invasive species policies in the presence of commodity programs: Theory and applica- An assessment of the benefits of yellow Sigatoka (Mycosphaerella musicola)... 7) tion to citrus canker. Review of Agricultural Economics 27: 498. doi: 10.1111/j.1467- 9353.2005.00250.x Albers HJ, Fischer C, Sanchirico JN (2010) Invasive species management in a spatially hetero- geneous world: Effects of uniform policies. Resource and Energy Economics 32: 483-499. doi: 10.1016/j.reseneeco.2010.04.001 Barbier EB (2001) A note on the economics of biological invasions. Ecological Economics 39: 197-202. doi: 10.1016/S0921-8009(01)00239-7 Born W, Rauschmayer F, Brauer I (2005) Economic evaluation of biological invasions - a sur- vey. Ecological Economics 55: 321-336. doi: 10.1016/j.ecolecon.2005.08.014 Cacho OJ, Wise RM, Hester SM, Sinden JA (2008) Bioeconomic modeling for control of weeds in natural environments. Ecological Economics 65: 559-568. doi: 10.1016/j. ecolecon.2007.08.006 Carrasco LR, Baker R, MacLeod A, Knight JD, Mumford JD (2010) Optimal and robust con- trol of invasive alien species spreading in homogeneous landscapes. Journal of the Royal Society Interface 7: 529-540. doi: 10.1098/rsif.2009.0266 Carrasco LR, Cook D, Baker R, MacLeod A, Knight JD, Mumford JD (2012) Towards the integration of spread and economic impacts of biological invasions in a landscape of learning and imitating agents. Ecological Economics 76: 95-103. doi: 10.1016/j. ecolecon.2012.02.009 Colautti R, Bailey S, Overdijk CA, Amundsen K, Maclsaac H (2006) Characterised and projected costs of nonindigenous species in Canada. Biological Invasions 8: 45-59. doi: 10.1007/s10530-005-0236-y Commonwealth of Australia (2006) Handbook of Cost-Benefit Analysis. Department of Fi- nance and Administration, Canberra, 180 pp. Cook DC (2003) Prioritising Exotic Pest Threats to Western Australian Plant Industries. Dis- cussion Paper. Government of Western Australia - Department of Agriculture, Bunbury, 185 pp. Cook DC (2008) Benefit cost analysis of an import access request. Food Policy 33: 277-285. doi: 10.1016/j.foodpol.2007.09.002 Cook DC, Carrasco LR, Paini DR, Fraser RW (201 1a) Estimating the social welfare effects of New Zealand apple imports. Australian Journal of Agricultural and Resource Economics 55: 1-22. doi: 10.1111/j.1467-8489.2011.00558.x Cook DC, Fraser RW (2008) Trade and invasive species risk mitigation: Reconciling WTO compliance with maximising the gains from trade. Food Policy 33: 176-184. doi: 10.1016/j.foodpol.2007.07.001 Cook DC, Fraser RW, Paini DR, Warden AC, Lonsdale WM, Barro PJD (2011b) Biosecurity and yield improvement technologies are strategic complements in the fight against food insecurity. PLoS ONE 6: e26084. doi: 10.1371/journal.pone.0026084 Doak DF, Estes JA, Halpern BS, Jacob U, Lindberg DR, Lovvorn J, Monson DH, Tinker MT, Williams TM, Wootton JT, Carroll I, Emmerson M, Micheli F, Novak M (2008) Under- standing and predicting ecological dynamics: Are major surprises inevitable? Ecology 89: 952-961. doi: 10.1890/07-0965.1 80 David C. Cook et al. / NeoBiota 18: 67-81 (2013) Dwyer G (1992) On the spatial spread of insect pathogens - theory and experiment. Ecology 73: 479-494. doi: 10.2307/1940754 Epanchin-Niell RS, Hastings A (2010) Controlling established invaders: integrating economics and spread dynamics to determine optimal management. Ecology Letters 13: 528-541. doi: 10.1111/j.1461-0248.2010.01440.x Fisher RA (1937) The wave of advance of advantageous genes. Annual Eugenics 7: 353-369. doi: 10.1111/j.1469-1809.1937.tb02153.x Heikkila J, Peltola J (2004) Analysis of the Colorado potato beetle protection system in Fin- land. Agricultural Economics 31: 343-352. doi: 10.1111/j.1574-0862.2004.tb0027 1.x Henderson J, Pattemore JA, Porchun SC, Hayden HL, Brunschot SV, Grice KRE, Peterson RA, Thomas-Hall SR, Aitken EAB (2006) Black Sigatoka disease: new technologies to strengthen eradication strategies in Australia. Australasian Plant Pathology 35: 181-193. doi: 10.1071/AP06017 Hengeveld B (1989) Dynamics of Biological Invasions. Chapman and Hall, London, 160 pp. Hester SM, Cacho O (2003) Modelling apple orchard systems. Agricultural Systems 77: 137-154. doi: 10.1016/S0308-521X(02)00106-3 Holmes EE (1993) Are diffusion-models too simple - a comparison with telegraph models of invasion. American Naturalist 142: 779-795. doi: 10.1086/285572 Hulme PE (2006) Beyond control: wider implications for the management of biological inva- sions. Journal of Applied Ecology 43: 835-847. doi: 10.1111/j.1365-2664.2006.01227.x Hyder A, Leung B, Miao ZW (2008) Integrating Data, Biology, and Decision Models for In- vasive Species Management: Application to Leafy Spurge (Euphorbia esula). Ecology and Society 13: 1-12. Lewis MA (1997) Variability, patchiness, and jump dispersal in the spread of an invading pop- ulation. In: Tilman D, Kareiva P (Eds) Spatial Ecology: The Role of Space in Population Dynamics and Interspecific Interactions. Princeton University Press, New Jersey, 46-74. McCann K, Hastings A, Harrison S, Wilson W (2000) Population outbreaks in a discrete world. ‘Theoretical Population Biology 57: 97-108. doi: 10.1006/tpbi.1999.1444 Moody ME, Mack RN (1988) Controlling the spread of plant invasions: the importance of nascent foci. Journal of Applied Ecology 25: 1009-1021. doi: 10.2307/2403762 Naylor RL (2000) The economics of alien species invasions. In: Mooney HA, Hobbs RJ (Eds) Invasive Species in a Changing World. Island Press, Washington, D.C., 241-259. Okubo A, Levin SA (2002) Diffusion and ecological problems: modern perspectives. Springer, New York, 467 pp. Ostmark HE (1974) Economic Insect Pests of Bananas. Annual Review of Entomology 19: 161-176. doi: 10.1146/annurev.en.19.010174.001113 Paarlberg PL, Lee JG, Seitzinger AH (2003) Measuring Welfare effects of an FMD Outbreak in the United States. Journal of Agricultural and Applied Economics 35: 53-65. Peterson R, Grice K, Goebel R (2005) Eradication of black leaf streak disease from banana growing areas in Australia. InfoMusa 14: 7-10. Rafoss T (2003) Spatial stochastic simulation offers potential as a quantitative method for pest risk analysis. Risk Analysis 23: 651-661. doi: 10.1111/1539-6924.00344 An assessment of the benefits of yellow Sigatoka (Mycosphaerella musicola)... 81 Regan HM, Colyvan M, Burgman MA (2002) A taxonomy and treatment of uncertain- ty for ecology and conservation biology. Ecological Applications 12: 618-628. doi: 10.1890/1051-0761(2002)012[0618:ATATOU]2.0.CO;2 Sanchirico J, Albers H, Fischer C, Coleman C (2010) Spatial Management of Invasive Species: Pathways and Policy Options. Environmental and Resource Economics 45: 517-535. doi: 10.1007/s10640-009-9326-0 Sharov AA, Liebhold AM (1998) Bioeconomics of managing the spread of exotic pest spe- cies with barrier zones. Ecological Applications 8: 833-845. doi: 10.1111/j.0272- 4332.2004.00486.x Shigesada N, Kawasaki K (1997) Biological Invasions: Theory and Practice. Oxford University Press, Oxford, 205 pp. Simberloff D (2005) The politics of assessing risk for biological invasions: the USA as a case study. Trends In Ecology & Evolution 20: 216-222. doi: 10.1016/j.tree.2005.02.008 State of Queensland (2003) Plant Protection Amendment Regulation (No. 4). Queensland Parliament, Brisbane, 27 pp. State of Queensland (2002) Plant Protection Regulation. Queensland Parliament, Brisbane, I l