2011 IFQRG-9-19
Geospatial intelligence as a method of predicting risk of ship contamination by gypsy moths from Asia (Lymantria spp.; Lepidoptera: Lymantriidae)
John E. Nielsen Grains and Forestry Branch, Plant Division,
Australian Government Department of Agriculture, Fisheries & Forestry
Abstract
Geospatial intelligence is based on imagery of the environment, and can be a powerful tool to both predict risk and develop risk management strategies. It was used as a risk assessment tool for gypsy moths (Lymantria spp.) utilising maritime vessel pathways. Analysis based on proximity of individual seaports to Lymantria-susceptible forest types allowed the highest risk ports to be recognised and vessels departing these ports to be targeted for inspection. A trial program conducted by the Australian Quarantine and Inspection Service successfully intercepted Lymantria dispar egg masses from previously unrecognised risk ports in Japan and China. The potential for this method to produce targeted inspection regimes for vessels is discussed, along with potential future applications.
Introduction
Lymantria spp., commonly known as gypsy moths, are potentially invasive pests of trees grown for forestry and amenity (Pogue and Schaefer 2007). Larvae of these moths cause damage by consuming the leaves of as many as 1600 plant taxa (DAFF, unpublished), including numerous plantation timber species. The most damaging species specialise in attacking the spring growth flushes of deciduous and semi-deciduous trees, although evergreen species may also be attacked (Pogue and Schaefer 2007). Damage inflicted by Lymantria larvae can cause significant economic loss to plantation forestry operations, with some industry estimates reaching US$400 million per plantation per cropping cycle (Maryland Cooperative Extension 2007). An ongoing suppression program for the flightless European gypsy moth (L. dispar dispar Linnaeus) is estimated to provide a direct return to taxpayers of at least 3:1 (Sills 2008).
Invasive Lymantria taxa present a management challenge in that pathway entry is non-selective relative to most other plant pests. Females of invasive Lymantria taxa are attracted to any light source emitting attractive (mostly UV) light spectra (Wallner et al. 1995). As the majority of commercially available light sources emit attractive spectra (General Electric 2009), cities and industrial areas (including seaports) are highly attractive environments for gypsy moths. However, the ability of Lymantria moths to penetrate these environments is limited by their very attraction to light. In a recent study, Liebhold et al. (2008) demonstrated that populations of one gypsy moth (L. dispar japonica) had settled at lights within 2000 metres of entering urban environments. Once settled, female Lymantria begin to produce egg masses within a few hours and die once oviposition is completed (Koshio 1996). This behaviour allows gypsy moths access to a range of quarantine pathways, including maritime vessels, road vehicles, shipping containers and break-bulk cargo. Of these, maritime vessels are generally the most frequently contaminated.
Historically, quarantine operations internationally have focused on the quarantine risk posed by Lymantria dispar asiatica. This situation is the result of numerous interceptions of L. dispar asiatica egg masses from maritime vessels departing seaports in the Russian Far East. However, recent trade proliferation with Asia has seen maritime vessel traffic enter seaports in areas occupied by additional Lymantria taxa with similar biological attributes. Despite this situation, the sheer volume of trade received from Asia has made it difficult to use inspection to reliably identify specific seaports that expose visiting vessels to contamination with Lymantria egg masses. Other attempts to identify specific risk pathways are probably limited because they rely on different attractants (eg. male pheromones; Munson 2008) to those responsible for pathway contamination (lights).
Geospatial intelligence is based on imagery of the environment, and can be a powerful tool to both predict risk and develop risk management strategies. It is appropriate for use in informing quarantine decision making for invasive Lymantria taxa because their favoured host plants (eg. Quercus, Larix and Pinus spp.) (Pogue and Schaefer 2007) tend to dominate and define forest types in temperate Eurasia (Globcover 2008). Imagery is also useful in being able to identify infrastructure supporting Lymantria-susceptible trade pathways, such as seaport facilities. This paper describes Australia’s experience in using geospatial intelligence to inform border policy for the management of invasive Lymantria taxa. Future applications of this technology to the management of risk Lymantria taxa are also discussed.
Materials and Method
Pest categorisation
A review of the biosecurity risk posed by exotic Lymantria taxa travelling via maritime vessels was performed using standard pest categorisation methods (FAO 2004). These included consideration of the presence/absence of the taxa in Australia, their ability to enter, establish and spread and their potential for economic and environmental consequences in Australia.
Egg diapause requirements
A literature review was also conducted on the egg diapause of Lymantria taxa identified by pest categorisation. Because Australia is counter-seasonal to Eurasia, the effect of dormancy on egg masses travelling on maritime vessels also needs to be understood and has implications for quarantine operations. The effect of climate during trans-shipping on egg diapause was also considered, with data from sea container temperature studies (Leinberger 2006) used as a surrogate for vessel temperature.
Geospatial analysis
The known distributions of Lymantria taxa identified as potential quarantine risks were used to define geographic boundaries for this study. Any seaport within these boundaries was considered for analysis. Seaports supporting existing (known) maritime pathways to Australia were identified from vessel inspection records maintained by the Australian Quarantine and Inspection Service (AQIS) for the years 2008 and 2009. Vessels arriving during the Lymantria dispar flight period recognised by Australia (July – September) were considered; October data was also included to consider vessels departing Eurasia during late September. Future and other potential pathways were also considered by examining a range of maritime resources (World Port Source 2009), or by a dedicated search for seaport facilities using satellite and aerial imagery (Google 2010).
Each port considered for analysis was assigned a reference set of geo-coordinates (longitude and latitude in decimal degrees) based on data obtained from World Port Source (2009) and Google Earth (2010). A set of spectral data for world forest cover, which also identified forest types (Globcover 2008), was then used to determine which seaports were within potential flight range of Lymantria risk taxa in built-up areas. While Liebhold et al. (2008) identified that Lymantria dispar japonica could penetrate human habitation for as far as 1500 metres, this study used a maximum penetration distance of 2000 metres to allow for less brightly lit urban environments than Japan. Individual risk maps were then created for each seaport within 2000 metres of identified areas of vegetation. The vegetation cover classes used in the analysis were based on their suitability for use as host plants by Lymantria risk taxa identified by pest categorisation (Appendix 1). Each map displayed an area of 30 square kilometres surrounding the seaport’s geo-coordinates (Figure 1). Radii of 2, 4, 8 and 10 kilometres from the seaport reference location were also added. The map for each seaport was then overlaid on Google Earth (2010) for analysis of seaport infrastructure and site topography. The location of vessel berthing sites were also manually added using a marker with a 2000 metre risk radius (Figure 1). Each seaport site was also viewed in 3-dimensions using Google Earth (2010) topography data, allowing consideration of the effect various landforms may have on the dispersal Lymantria taxa. Additional areas of forest cover identified via Google Earth were manually added to the map under a separate vegetation category. Similarly, cleared areas represented as forest were manually removed from the maps.
Surveillance
Following identification of potential risk ports from satellite analysis, the Seaports program in AQIS conducted a pilot surveillance study for Lymantria taxa on maritime vessels. The surveillance methodology was designed using ISPM guidelines as their basis (FAO 1997), meeting the definition of specific surveillance. All Lymantria taxa identified during pest categorisation were targeted for surveillance. Surveillance design was streamlined by targeting high-risk areas on the vessel (eg. areas surrounding light sources) where egg masses are most likely to be present. Data regarding the location of egg masses on inspected vessels was also captured, allowing future inspections to increasingly focus on the highest risk areas of the vessel. Egg mass collection techniques were designed to allow both morphological and molecular-based identifications to be made.
The pilot study using these surveillance criteria is being conducted from 1 July – 30 September 2011. The trial modified the surveillance criteria by restricting inspections to vessels arriving at the four Australian first ports expecting to receive the highest number of vessels eligible for surveillance (Gladstone and Brisbane, Queensland; Newcastle, New South Wales and Port Headland, Western Australia). Inspections were also conducted at the discretion of individual seaports based on staffing and resource availability. The trial also commenced two weeks earlier than the inspection period (15 July – 30 September) defined by Australia for L. dispar asiatica.
Results
Pest categorisation
Pest categorisation identified a total of six Lymantria taxa that had the potential to arrive at Australian ports on maritime vessels, establish and spread and cause economic and environmental consequences in Australia (Appendix 1).
Egg diapause requirements
A literature review found egg diapause data was available for only two taxa, L. dispar sensu lato and L. xylina. The data was most comprehensive for L. dispar. Generalised requirements for diapausing egg masses of L. dispar are a minimum of 60 days exposure to cold temperatures, with 5°C being optimal. Upper and lower thermal limits for egg viability are 40°C (substrate temperature) and -30°C, respectively. A study by Tauber et al. (1990) also found that L. dispar dispar egg masses may produce larvae without cold exposure. Egg masses collected during autumn were found to produce larvae if they were moved to conditions replicating spring or summer (longer days and/or warmer temperatures). The time to emergence was as little as 6-9 weeks (Tauber et al. 1990) under conditions likely to be experienced during a sea voyage from Asia transiting the Pacific Ocean (Leinberger 2006). Hatch rates for individual egg masses under these conditions were significant, with between 8-15% of eggs hatching (Tauber et al. 1990). Assuming an average egg mass contains approximately 600 eggs (Koshio 1996), 50-90 larvae could feasibly emerge per egg mass, with emergence likely to coincide with the arrival of at least some vessels in Australian seaports.
Table 1 Ports found to be within 2000 metres of forested areas likely to support populations of Lymantria risk taxa
Country / Port Name / Country / Port Name /China / Dandong / Russia (Far East) / Aleksandrovsk-Sakhalinsky
Dongshan / De-Kastri
Lianyungang / Kholmsk
Shanwei / Lazarev
Yangjiang / Magadan
Japan / Aioi / Nakhodkha
Fukuyama / Nevelsk
Hiroshima / Nikolaevsk on Amur
Imari / Olga
Kikuma / Petropavlovsk -Kamchatsky
Matsushima / Poronaysk
Mizushima / Posyet
Moji / Sovgavan
(Sovetskaya Gavan)
Naha / Uglegorsk
Niigata / Ust Kamchatsk
Shimotsu / Vanino
Susaki / Vladivostok
Tachibana / Vostochny
Tamano / South Korea / Busan (Pusan)
Tsukumi / Donghae
North Korea / Chonjin / Gwangyang
Haeju / Masan
Hungnam / Mokpo
Nampo / Okgye
Sonbong / Pyeongtaek
Sonjin / Samcheok
Taiwan / Taichung / Ulsan
Figure 1 Example risk assessment maps generated using geospatial intelligence techniques for (A) Lianyanguang, China, (B) Fukuyama, (C) Mizushima and (D) Tachibana, Japan
1
2011 IFQRG-9-19
Geospatial analysis
A total of 121 Asian seaports within the geographic distribution of risk Lymantria were included in the geospatial analysis. Of these, risk maps were produced for a total of 54 seaports found to be within 2000 metres of vegetation identified by satellite imagery (Table 1). Examples of these maps are shown in Figure 1.
Surveillance
The surveillance trial inspected a total of 53 vessels departing 16 potential risk ports between 1 July and 26 August 2011. Surveillance activities successfully recovered a total of 66 egg masses from eight vessels (Table 2), with contaminated vessels representing 15% of all inspected vessels.
Table 2 Lymantria egg mass interceptions from the geospatial surveillance program
Risk port / Vessel type / Australian port / Summary / Identification1Fukuyama, Japan / Bulk carrier / Gladstone, QLD / 12 egg masses ex superstructure
(adjacent to light sources on bridge, poop deck and lifeboat deck) / Lymantria dispar
Mizushima, Japan / Bulk carrier / Gladstone, QLD / 10 egg masses ex superstructure
(unspecified locations) / Lymantria dispar
Fukuyama, Japan / Bulk carrier / Gladstone, QLD / 5 egg masses ex superstructure
(gangway and near light sources) / Lymantria dispar
Fukuyama, Japan / Cargo vessel / Newcastle, NSW / 1 egg mass ex ladder
(starboard side) / Lymantria dispar
Fukuyama, Japan2 / Bulk carrier / Gladstone, QLD / 1 egg mass ex superstructure (wall adjacent to light source) / Lymantria dispar
Tachibana, Japan / Bulk carrier / Newcastle, NSW / 1egg mass ex superstructure (bridge) / Lymantria dispar
Mizushima, Japan / Bulk carrier / Port Headland, WA / 30+ egg masses ex superstructure (walls on bridge and multiple deck layers) / Lymantria dispar
Lionyunggong, China / Bulk carrier / Port Headland, WA / 6 egg masses ex superstructure / Lymantria dispar
1 The diagnostic protocol used for these identifications (deWaard et al. 2010) is unable to distinguish L. dispar asiatica from L. dispar japonica.
2 The last port of call for this vessel was Niihama/Ehime, Japan, but it had previously spent three months in dry dock at Fukuyama following repairs sustained during the 2011 tsunami.
Discussion
The trial inspection program successfully identified four new risk ports for Lymantria taxa, which had not been recognised or regulated by any country. The frequency of vessels contaminated with egg masses, and numbers of egg masses intercepted, immediately justify the need for ongoing vessel inspections against Lymantria taxa. As of August, over 66 egg masses had been recovered from 8 vessels. This data contrasts sharply with interceptions from the regulated pathway for maritime vessels from the Russian Federation, with 10 contaminated vessels intercepted since 2005. These results clearly demonstrate the potential for geospatial intelligence to inform quarantine border operations and quarantine policy.
A larger scale trial is needed to fully validate the technique’s efficacy. As only four Australian seaports participated in the trial, 36 vessels out of an expected 183 were considered nationally (20% of eligible vessels). A larger trial would also be able to identify patterns based on different vessel types and light-source types. For example, evidence from inspections conducted during the trial suggest that the structure of some vessels may make them less vulnerable to contamination. Roll-on roll-off (RORO) vehicle carriers in particular may be less vulnerable to contamination due to their reduced superstructure and large slab sides, which may be difficult for moths to land on, and reduced lighting on the vessel itself. In addition, the high-intensity lighting required for vehicle movements onto and off RORO vessels may not necessarily direct the bulk of emitted light onto the vessel itself. The type of lighting used on individual vessels may also play a role in contamination, given adult Lymantria spp. in the Russian far east are known to favour light sources emitting UV spectra (Wallner et al. 1995). Some light sources used on maritime vessels (eg. sodium vapour, halogen and “warm white” type incandescent bulbs) emit few attractive spectra (General Electric 2009) and are less likely to attract Lymantria moths. Future vessel surveys will record the type of on-vessel illumination types used, allowing any association between vessel contamination and light source type to be identified and investigated further.