|
EXHIBIT B
<br /> 102 D.J.Lee et at/Ecological Economics 111(2015)100-110
<br /> increased with the size of the infestation by Gutrich et al.(2007);as a damage,and start producing new colonies that can be transported to
<br /> quadratic function by Burnett et al.(2007);and as a non-linear function other locations.
<br /> by Olson(2006)and Haight and Polasky(2010).
<br /> Leung et al. (2002), Burnett et al. (2007),and Eiswerth and van 2.1.4.Original Contribution
<br /> Koote.n(2002)modeled invasive species population growth up to a car- In previous models of invasive ants,spatial spread was forecast using
<br /> rying capacity over discrete time using a logistic function. Lee et al. radial and cellular specifications. Our approach is novel in that we
<br /> (2007)modeled invasive species growth and spread overtime using a model LFA spread within and across economic sectors over time.On
<br /> transition matrix to forecast the probability of uninfested locations be- the island of Hawaii,LFAs are transported unintentionally by humans
<br /> coming invaded,and then tracking that status of the infestation from in- with the movement of infested soil,produce,and other goods;mecha-
<br /> cubating,to propagating,to spreading to other locations.Carrasco et al. nisms which do not follow a radial or cellular pattern. Further,LFA
<br /> (2010),Leung et al. (2002),and Burnett et al.(2007) simulated short nests are tiny compared with the area they can impact; several tiny
<br /> distance dispersal via colony budding using a reaction-diffusion walnut-sized nests can disrupt the activities of an entire household or
<br /> model. Suarez et al. (2001), Souza et al. (2008), and Wilson et al. farm.Thus rather than units of length or area,we use discrete locations
<br /> (2009)modeled unlimited carrying capacity and long distance human as our unit measure of"space",e.g.a home,a school,a farm,and a park.
<br /> mediated dispersal.Nathan et al.(2003)used gravity models to quantify One unit represents one location.This pseudo-spatial approach pro-
<br /> human-mediated long distance dispersal.Hastings et al. (2005),and vides us with a compact way of specifying and simulating the joint rela-
<br /> Bossenbroek et al. (2001) used commerce and traffic flows to model tionships between economic activity, LFA movement, LFA impacts
<br /> human-mediated dispersal pathways.Carrasco et al. (2010)assumed (economic and social),and management response.Our second contri-
<br /> outward dispersal via a random walk process and used diffusion bution is an accounting of the number of LFA stings and a comparison
<br /> models. Eiswerth and van Kooten (2002), Kot and Schaffer (1986), of the Pareto tradeoff between economic impacts and stings.
<br /> Hastings et al.(2005),and Law et al.(2003)employed a probabilistic
<br /> transition function3 to model dispersal. Scanlan and Vanderwoude 2.2.Model Scope and Detail
<br /> (2009)modeled dispersal using a stochastic cellular automata.
<br /> Our model includes ninety thousand locations on the island of
<br /> 2.Materials and Methods Hawaii within seven economic sectors i E(nursery,agriculture,lodging,
<br /> residential,parks,schools,and all others).Of the ninety thousand loca-
<br /> 2.1.Model Framework lions,4581 locations are infested initially.Our model simulates infesta-
<br /> tion 35 years t E(0...35)into the future.The number of locations per
<br /> We simulate future spread and impacts from LFA on the island of sector and initial LFA infestation is shown in Table 1.
<br /> Hawaii using a bioeconomic model comprised of three integrated sub-
<br /> models:impact,management,and spread.Control variables determine 2.2.1.Impacts
<br /> the level of effort allocated toward LFA detection,spread prevention, Impacts from LFA comprise economic damages,management costs,
<br /> and mitigation treatment and human and pet sting incidents.
<br /> 2.2.1.1.Economic Damages.Economic damages are sector-specific and
<br /> 2.1.1.Impacts vary with the size and extent of the LFA infestation.For example,in
<br /> The impact sub-model quantifies economic impacts (economic the residential sector we include the impact of LFA on property values
<br /> damage and management expenditure)and social impacts(the number
<br /> when homes are sold.In the lodging sector we include reduced reve-
<br /> of human and pet LFA sting incidents)per sector over time.Economic nues from decreased room occupancy and cheaper room rates.The eco-
<br /> damages are sector dependent and vary with the size and extent of nomic damage per sector location is based on estimated mean economic
<br /> the infestation.Management expenditures are based on best manage- impacts from LFA and is assumed to increase with the number of
<br /> ment practices and current technology and vary with management infested locations and overall level of infestation.The economic damage
<br /> goals,management effort,and the cost of labor and materials.Sting in- in sector i at time t is:
<br /> cidents are based on the number of infested locations in each sector,
<br /> human population,pet population,demographics,and employment in
<br /> each sector.A sting"incident"may involve multiple LFA stings. D c amRg,Ne`rabtish2
<br /> = i Nmax (1)
<br /> 2.1.2.Management
<br /> The management sub-model quantifies the effect of management
<br /> decisions on LFA survival,growth and dispersal.Management activities Table 1
<br /> include detection,prevention,and mitigation treatment.Detection al- Little Fire Ant infested locations on the island of Hawaii in 2012.
<br /> lows new infestations to be treated before they become established. Sector %Infested Infested locations Total locations
<br /> Prevention reduces the likelihood that LFA will be transported to anoth-
<br /> Nursery 22.5% 170 757
<br /> er location by humans.Mitigation treatments reduce the intensity and Agriculture 4.0% 186 4650
<br /> extent of infestations. Lodging 0.2% 1 468
<br /> Residential 7.0% 3648 52,216
<br /> Parks 3.9% 6 152
<br /> 2.1.3.Spread Schools 1.2% 1 84
<br /> The spread sub-model simulates LFA survival,growth and dispersal Other 1.7% 568 32,547'
<br /> as follows.When LFAs are initially introduced to a new location their Total 5% 4581 90,874
<br /> chance of surviving is low.If they survive,they go unnoticed for several From Motoki et al.(2013'.b
<br /> years during which they have time to establish and increase in number. ' Hawaii Island is 2.58 million acres.With our 6 major sectors we account for 23 million
<br /> The first year after introduction,LFAs are comparatively easy to eradi- acres.Our sector"other"consists of 0.28 million acres and 81,556 parcels(according to
<br /> tate.Once they establish,they are difficult to eradicate,begin causing 2010 tax records).To scale the model,we represented the"other"sector with 32,547
<br /> locations.
<br /> b Using data from the Hawaii Ant Lab;information from the 2007 Census of Agriculture,
<br /> 3 A transition matrix is a kernel without a functional form,matrix elements denote the the 2011 Visitor Plant Inventory.City-datacom,and the State of Hawaii Data Book,and
<br /> probability of transitioning between states or spatial locations. 2013 PCSU Technical Report#186.
<br />
|