INNES and BARKER: ECOLOGICAL CONSEQUENCES OF TOXIN USENew Zealand Jour - PDF document

111 INNES and BARKER: ECOLOGICAL CONSEQUENCES OF TOXIN USENew Zealand Journal of Ecology (1999) 23(2):111-127 ©New Zealand Ecological Society ECOLOGICAL CONSEQUENCES OF TOXIN USE FORMAMMALIAN PEST CONTROL IN NEW ZEALAND„ AN OVERVIEW 112NEW ZEALAND JOURNAL OF ECOLOGY, VOL. 23, NO. 2, 1999ScaleFrequencyObjective: Tuberculosis eradicationlargeinfrequent(to 50 000 ha)(repeat 4…8 yrs)Objective: Restoration of off-shore islandssmall-mediumonce only(average 180ha)(eradication)Objective: Maintenance of mainland ecosystemslargevariable(to 60 000 ha)(repeat 1…8 yrs)Objective: Restoration of mainland ecosystemsmedium-largefrequent (annual)Objective: Environmental protectionmedium-largevariableObjective: Protecting productionsmall-largevariablesmall-scale conservationsmall-largefrequentObjective: Forestry protectionlargeinfrequent 113 INNES and BARKER: ECOLOGICAL CONSEQUENCES OF TOXIN USEconservation management will inevitably focus onthe control of these mammals. Non-toxin techniques,such as immunocontraception, may be available forsome species such as possum within 5…10 years(Cowan and Bayliss, 1998), but toxins will remainthe key control method for many other species,land area subjected to possum control has increasedfor both the Animal Health Board and the DOC inrecent years (Fig. 2). An increasing proportion ofoperations for both agencies is maintenance ratherthan initial control. For example, in 1993…94, allcontrol operations performed by the Animal HealthBoard were initial control, but this decreased toabout 10% by 1996…97 (N. Hancox, pers. comm.),conducted by the DOC increased from 21% to 32%between 1993 and 1995 (Parkes, Baker andEricksen, 1997).Ecological consequences oftoxin useTwo contexts for ecological consequencesUsing toxins in preference to traps, fences, shooting,or immunocontraception as the method of pestcontrol has particular ecological consequences, suchoften mentioned as impacts or costs (e.g., Spurr,1994a, b; Eason and Spurr, 1995). Otherconsequences may occur due to the reduction oreradication of pest mammals. These are likelyregardless of the pest control method used, and arecost and benefit are terms of value, not science.In fact all pest control methods have both directecological consequences which may be seen as costsor benefits, depending on the objectives of theparticular pest control operation.Consequences at different ecological levelsindividualsmay die, or can suffer sub-lethal effects. Sub-lethalthemselves as altered behaviour (e.g., increasedsusceptibility to predation), or decreased breedingthepopulation level, where definite effects onindividuals (e.g., death, decreased number ofas density, sex ratio, age structure, mortality rates).In one sense, a population could be said to bemake differing contributions to key populationprocesses. Population impacts are especiallydetermined by the proportion of individuals in theeffective breeding population killed or sublethallyaffected by the toxin, and the rate of populationpopulations of K- than r-selected species, since theformer will take longer to make up the losses (Spurr,1979).populations which occur together in space and timeet al., 1990). Communities have specificattributes such as species diversity, stability,complexity, and succession. Nutrients cycle throughusually referred to as an ecosystem. New Zealandresearchers and managers are increasingly aware ofcan have unforeseen repercussions such as prey-switching (Murphy and Bradfield, 1992; Norburyand Heyward, 1996; Murphy et al., 1998).Pests, and control methods such as toxin use,can have ecosystem-level effects by influence onproperties emergent from the interaction of the biotaand the physical environment. These ecosystem-level properties include litter decomposition rates,primary productivity.There are two powerful reasons why betterunderstanding of communities in New Zealand isrequired. First, managers have to manage whole Figure 2: Area of possum control operations by the AnimalHealth Board and the Department of Conservation, 1993…in litt.;DOC Annual Reports to its Minister). 114NEW ZEALAND JOURNAL OF ECOLOGY, VOL. 23, NO. 2, 1999framework?(1)Can be used to trace toxin movement(2)Can be used to identify species and functional(3)Can be used to integrate disparate information(4)Can serve as conceptual foci for research.(5)Offer an excellent start-point for all 115 INNES and BARKER: ECOLOGICAL CONSEQUENCES OF TOXIN USEMost research has been on direct (primary)impacts of 1080; recently some studies havemammal predators, but none has looked for impactsat tertiary level or beyond. This may be becauseusually rapidly eliminated from live animals, andrapidly broken down by microbial activity in baits, et al., 1993b; Eason,Gooneratne and Rammell, 1994; Parfitt et al., 1994,and references therein). Nonetheless, the linksdepicted in the foodweb (Fig. 3) clearly suggest thatsuch third order impacts are possible if the rate ofspecies interactions exceed the rate of toxin decay.Very few accounts describe population-levelimpacts of aerial application of 1080, for example(for vertebrates) with more than half of monitoredet al. (1999) describe such a level of loss for anytaxon. The North Island robin (Petroica australis)population which he observed at Pureora inSeptember 1996 recovered to pre-application levelswithin a year, due to improved nesting success afterRattus rattus) werekilled by the 1080. Further, the robin kill was causedby excess chaff in the carrot bait, which can bethis outcome is unlikely to characterise more routineaerial applications.However, most attempts to quantify impacts onnon-target species level are very simplistic andshort-term. Commonly, changes in the abundance inreference sites, are interpreted and extrapolated tosuggest that the management practice is adverse.Little consideration is given to the distribution ofmortality through the life cycle and littleappreciation that marked changes in mortality in aimpact on the population trend of the animal ifvariation in that stage is not critical to population Figure 3: Descriptive food web for a model New Zealand podocarp-broadleaved forest. The web has been constructed toprovide an example community setting for analysis of ecological consequences of toxins (see Figs. 4…6). Note some boxes 116NEW ZEALAND JOURNAL OF ECOLOGY, VOL. 23, NO. 2, 1999 117 INNES and BARKER: ECOLOGICAL CONSEQUENCES OF TOXIN USEafter toxin poisoning. An example of how food webscan be used to generate hypotheses for testing isshown in Fig. 7, demonstrating food web responses tothe near-eradication of possum and rodents in theFig. 3. The number of trophic interactions (linkagearrows) declines by one third with the removal ofare omnivores which feed on vegetation,invertebrates and vertebrates. More food may beformer prey (source boxes) of possum and rodents,and predators which formerly ate rodents and possumorder responses to possum and rodent removal inboth source and sink directions are shown. Clearlythe actual community response will be morecomplex, depending on the relative importance oftop-down and bottom-up controls on individualdoes, however, present many hypotheses which couldbe tested in the field. One of these - prey-switchingby mustelids - has already been verified (Murphy andBradfield, 1992; Murphy et al., 1999a).When all introduced mammals are removedfrom the web, the number of trophic group linksdeclines by two thirds, and predatory and scavengingbirds (and their parasites) are identified as topin New Zealand.Ecological relationships other than trophicNot all interactions between species are trophic, oreven biotic. Examples of non-trophic relationshipsare pollination and seed dispersal, or competition fornest or roost sites. Parasitic New Zealand mistletoescollecting, habitat clearance, and browsing bypossum (Ladley et al., 1997). However, Ladley et al.(1997) suggest also that they suffer a shortage of tuiand/or bellbird that pollinate flowers and disperseseeds. Each Peraxilla flower is dependent on a tui(Prosthemadera novaeseelandiae) or bellbird Figure 5: Documented routes of toxin occurrence after application of 1080 in bait stations on the two large islands (Northand South) of New Zealand, shown in the food web format of Fig. 3, derived from data in the literature. See caption to Fig.4 for explanations and limitations of data. 118NEW ZEALAND JOURNAL OF ECOLOGY, VOL. 23, NO. 2, 1999long-term maintenance of kokakowill be from 119 INNES and BARKER: ECOLOGICAL CONSEQUENCES OF TOXIN USEeffort, each lasting perhaps 3…5 years (Innes et al.,1999). This regime allows fledged young to mature in1…2 years and then breed. Kokako adults are long-pest control. However, this regime may be ineffectivefor increasing populations of species such as wetavulnerable to predation in the inter-pulse periods.Net outcome at community level?The ecological consequences of toxin use for pestmammal control are complex. Toxins kill manytargets directly but non-target individuals may also belethally or sublethally poisoned. Secondary or evenoccur (Eason et al., 1999 b; Gillies and Pierce, 1999;Murphyet al.,1999; Murphy et al., 1998 a, b; otherpapers, this proceedings). Large reductions in pestmammal populations also have many consequences(Fig. 7). The cascading community outcomes of pesttoxicity, bait formulation, application rate, timing,weather, quality control), control history (e.g., toxin orbait aversion, diet, population structure), and theparticular ecological community where the operationoccurs. Finally, whether the ecological outcome will beparticular operation was, which attributes of whatspecies were measured, and how long after poisoningthese measurements were taken.We suggest that large-scale use of toxinscontinues in New Zealand despite these largesuggests that the harmful effects of pest mammals areoverwhelmingly greater than those of the toxins used.One way to examine this is to picture pests as toxins.Table 2 compares attributes of 1080, the most widely Figure 7: Food web responses to the near-eradication of possums and rodents in the mainland podocarp-hardwood forestdepicted in Fig. 3. All source boxes previously giving food to possums and rodents may now be more available for other 120NEW ZEALAND JOURNAL OF ECOLOGY, VOL. 23, NO. 2, 1999 November. Ship rat and possumrecovery rates are hypothetical; kokako nesting success data are from Innes et al., 1999. 121 1080POS- First use19541858- FrequencyRepeatedOnce only- NotificationYesNo- Rate7.5 g ha Coverage „ area of NZ9%92%PersistenceDays to monthsYearsSelectivityFaunaFauna and floraLevel of ImpactIndividual - populationCommunity - ecosystemHumaneYes?No 122NEW ZEALAND JOURNAL OF ECOLOGY, VOL. 23, NO. 2, 1999 123 INNES and BARKER: ECOLOGICAL CONSEQUENCES OF TOXIN USEThe three key tools, namely an experimentalapproach, systems models, and good field observation,need to be applied by teams of researchers who workwhich toxins are used. In the first instance, emphasiscould be on a North Island podocarp-broadleavedNothofagus forestas at Eglinton or Craigieburn, and a high altitudepastoral system as in the Mackenzie Basin.Working at each site, the teams should describefood webs for both the currently and previouslyexisting communities. The webs should be regardedThis development would be assisted by the use of anationally co-ordinated food web database applied ateach web box and trophic link could be displayedupon request. Persistent toxins such as brodifacoumcan be viewed as chemical tracers which assist foodweb model development.Compiling similar food webs for offshoreislands, especially those subjected to mammaleradications, could also greatly sharpen managementobjectives on the mainland by providing examples ofparticular restoration outcomes.Priority or focus species in constructing thesewebs should be those which are threatened,indicators of ecosystem processes.Reduce toxin useReducing the amount of toxin used is a sensible wayto minimise expense and unwanted ecologicalconsequences of toxins, regardless of the outcome ofresearch examining those consequences. Necessaryparticular pest mammals limit prey populations,spread Tuberculosis, or cause unacceptable damage,which control is necessary. This research ensuresthat poisoning operations are accurately targeted,and that the minimum control effort to achieve thedesired outcome is applied.Much valuable research has been undertaken inthe last three decades to reduce application rates ofboth carrot and pollard baits in aerial poisoningoperations against possum. Pollard bait applicationc. 20 to c. 7 kg ha-1,and further reductions to 2 kg ha-1 may be possibleat some sites (Morgan et al., 1997).More research should be directed towardsalternative toxin-free methods of pest control.. Theseinclude fences, fertility control, traps, andtaste aversion, Cowan et al., in press).Maintenance operationsAfter successful initial pest control, maintenancecontrol is needed to sustain the threatened resource.The required frequency, intensity and scale ofmaintenance pest control determine the tactics andmethods which managers should use, given theconstraints (e.g., risk, non-target effects, baitshyness, cost, topography) associated with eachcontrol technique. Pest densities may or may notneed to be maintained at low levels to sustain theresource (Choquenot, Parkes and Norbury, 1998).Generally, maintenance operations can be moredifficult than knock-down ones. Resistance oraversion to baits, toxins or traps may occur, andmore expensive control methods (e.g., station-basedinstead of aerial poisoning) may be required. Due tolower pest density, some beneficial secondarysupport can be more difficult to maintain once themajor pest damage has disappeared. Finally,managers and field workers may tire of workingrepeatedly at one location. Pulsed pest control effortmay overcome some of these problems.Good communicationIf ecosystem restoration is the task that pestmanagers are undertaking in New Zealand, thenmost biological researchers here are engaged on thisto that of previously distant colleagues such asbotanists, anthropologists, landscape ecologists, andecosystems. Land managers routinely undertakelarge-scale manipulations of mainland and islandecosystems, and these are most valuable when theyare regarded as experiments and treated as such, inco-operation with scientists. Policy-makers can alsothat pest control programmes can equally be framedas testing different policy stances in differentlocations, to guide a rational choice between difficultalternatives. Finally, pest mammal management isincreasingly undertaken in the public eye, and thedecision-making on pest management issues. Goodcommunication between these parties is needed nowannual dedicated ecosystem managementconference, which explicitly welcomed managers,from the many relevant disciplines.Ecological researchers should focus on netoutcomes of toxin use for mammalian pest controlrather than just on one aspect such as non-target 124NEW ZEALAND JOURNAL OF ECOLOGY, VOL. 23, NO. 2, 1999Acknowledgements 125 of Ecology23:Journal of Ecology23:Journal of Ecology19: 126NEW ZEALAND JOURNAL OF ECOLOGY, VOL. 23, NO. 2, 1999Ecology23:Biological Conservation84:New Zealand Journal of Ecology23: 127 Biological Conservation83