Rabbit control in the Parwan Valley and its value for catchment

RABBIT CONTROL IN THE PARWAN VALLEY ANDITS VALUE FOR CATCHMENT MANAGEMENTSeptember 1995CENTRE FOR LAND PROTECTION RESEARCHTechnical Report No. 26Richard Hartland and Michael PapworthISBN No. 0 7306 6107 5ISSN No. 1038-216XLand and Catchment Protection Branch, Department ofConservation & Natural Resources

Hartland, Richard, 1944 –Rabbit control in the Parwan Valley and its value for catchment management.Bibliography.ISBN 0 7306 6107 5.1. Rabbits - Control - Victoria - Parwan Creek Valley. 2. Watershed management Control - Victoria - Parwan Creek Valley. I. Papworth,M.P. II. Victoria. Land Protection Branch. III. Centre for Land Protection Research. IV. Title. (Series : Technical report (Centre for LandProtection Research (Vic.)) ; no. 26 ).632.69322

CONTENTSABSTRACT...................................................................................................................................... 51. INTRODUCTION ........................................................................................................................ 52. DESCRIPTION OF THE SITE .................................................................................................... 52.1 Location .................................................................................................................................. 52.2 Instrumentation ....................................................................................................................... 52.2 Landscape ............................................................................................................................... 72.3 Soils ........................................................................................................................................ 73.1 Reclamation at Western end of White Elephant Hills ............................................................ 73.2 The fencing plan ..................................................................................................................... 83.3 The development of targetted poisoning ................................................................................ 83.4 Rabbit number surveys in the Parwan Valley....................................................................... 103.5 Measurements of biomass..................................................................................................... 104. RABBIT EXCLUSION AND HYDROLOGY .......................................................................... 134.1 Estimation of surface run-off from the north face of the White Elephant Hill..................... 134.2 Water balance model ............................................................................................................ 144.3 Results............................................................................................................................. 155. FINANCIAL ANALYSIS ...................................................................................................... 155.1 Costs................................................................................................................................ 155.1.1 Personnel time................................................................................................................ 155.1.2 Cost of poisoned carrots ............................................................................................... 165.1.3 Cost of new fences......................................................................................................... 165.1.4 Cost for repair of existing fences kg bags...................................................................... 165.1.5 Summary of costs........................................................................................................... 175.2 Benefits ................................................................................................................................. 176. DISCUSSION............................................................................................................................. 186.1 The success of the rabbit control strategy............................................................................. 186.2 The relationship of rabbits and run-off, and possible mechanisms ...................................... 196.3 Future management recommendations ................................................................................. 197. CONCLUSIONS ........................................................................................................................ 208. ACKNOWLEDGEMENTS........................................................................................................ 209. REFERENCES ........................................................................................................................... 20List of PlatesPlate 1 - Photograph taken in autumn 1989 illustrating effects of rabbit exclusion. Full details aregiven in the text............................................................................................................................... 10List of FiguresFigure 1. Location of study site showing relationship of experimental area and White ElephantReserve.............................................................................................................................................. 6Figure 2. White Elephant Reserve showing poison trails................................................................. 9Figure 3. Location of transect in Parwan Valley for rabbit surveys, and survey points................. 11Figure 4. Results of rabbit surveys, 1985 to 1989 .......................................................................... 12Figure 5. Water Balance Model. Difference between Observed and Predicted Values (mm) ....... 12Figure 6 Mean monthly differences between observed run-off and values predicted from waterbalance model ................................................................................................................................. 16

List of TablesTable 1. Some characteristics of the Parwan subcatchments............................................................ 7Table 2. Application rates of poisoned carrots on sites of rabbit activity ........................................ 8Table 3. Mean rabbit numbers observed on homogeneous landscape............................................ 13Table 4. Mean rabbit density for surveyed land class in Parwan Valley........................................ 13Table 5. Characteristics of biomass at given sites .......................................................................... 13Table 6. Components of landscape of the vegetated area on the lower slopes in 1978.................. 14Table 7. Empirical model parameters ............................................................................................. 15Table 8 Costs for supply of poisoned carrots ................................................................................. 16Table 9 Costs of fencing ................................................................................................................. 17Table 10 Summary of costs ............................................................................................................ 17

RABBIT CONTROL IN THE PARWAN VALLEY ANDITS VALUE FOR CATCHMENT MANAGEMENTSeptember 1995R. HARTLAND AND M.P. PAPWORTHCentre for Land Protection Research, Department of Conservation & NaturalResources, 5/250 Victoria Pde, East Melbourne. Vic. 3002ABSTRACTThe highly eroded slopes of the White Elephant Hills provide ideal harbour for a rabbit population which can disperse readily intothe surrounding farmlands. The nature of the terrain makes eradicating rabbits from these sites extremely difficult. Rabbits areimplicated in the major soil erosion problems of the area. Rabbit numbers fell to very low levels during the 1982-83 drought, andthe opportunity was taken to intensify rabbit control measures. It was recognized that the problem would be much more difficult toaddress when rabbit numbers were high.Maintenance of very low rabbit numbers has been achieved by fencing off small areas and then targeted poisoning at sites of rabbitactivity. Consequently, there has been a dramatic increase in the regeneration of trees, shrubs and grasses, and particularly in thesurvival of tree and shrub seedlings over much of the White Elephant Reserve. Experimental equipment already in place hasidentified a 20% decrease in average annual surface run-off, attributed to the greater biomass with increased infiltration andevapotranspiration.This increase in pasture production corresponds in general terms to an improvement in the gross margin of approximately 1DSE/ha. The implications of rabbit control for catchment management both in the Parwan valley and more generally are examined.1. INTRODUCTIONSome fifty years ago, interest in the Parwan valley was initiated bytwo processes, erosion and sediment transport. It was recorded thatMelton reservoir was rapidly losing its water capacity as it filledwith sediment. The high sedimentation rate was the result ofexcessive soil erosion, predominantly from the Parwan Valley.The Parwan Creek catchment occupies only 17% of the totalcatchment area above Melton reservoir.The Parwan experimental area was established in 1953 by the SoilConservation Authority. It is located within the White ElephantReserve, now owned by the Victorian government (figure 1).The objectives in establishing the project were to gain a greaterunderstanding of the hydrological processes operating in an erodedcatchment, to determine the relationship between rainfall and runoff,and to evaluate the effects of remedial land use and pastoralpractices.While certain management factors have been deliberatelymanipulated, rabbit infestation has varied across the experimentalcatchments. The length of record enables some assessments to bemade of these changes.2. DESCRIPTION OF THE SITE2.1 LocationThe White Elephant Reserve covers some 300 ha, and islocated about 16 km south-west of Bacchus Marsh, 53 kmwest of Melbourne, Victoria (figure 1). The Parwanexperimental area covers 94.5 ha and is located within theReserve. Coordinates of the area are latitude 37°41' south,longitude 144°20' east. There are seven referencesubcatchments within the area. The orientation of thesubcatchments is shown in figure 1, and somecharacteristics are given in table 1.2.2 InstrumentationClimate data is collected at a meteorology station locatedwithin the area (figure 1). There is a continual record of thekey climatic factors (rainfall amount and intensity, windspeed and direction, temperature and humidity and panevaporation). Surface run-off from the six smallsubcatchments is recorded through a 300mm H-type flumeconnected to a stilling chamber. Surface run-off from themain channel is estimated by a 120° broad-crested weir andfloat recorder. In 1978, a sharp-crested weir was installed.Water level is recorded on L & S type A35 recorders. Amore detailed description of the instrumentation is given inWu (1980) and Wu et al. (1986). The average annualrainfall is approximately 530 mm, while average annualevaporation is 1540 mm, a deficit in the annual waterbalance of 1010 mm.

Figure 1. Location of study site showing relationship of experimental area and White Elephant Reserve

Table 1. Some characteristics of the Parwan subcatchmentsNumber 1 2 3 4 5 6 7aspect north north north south south south northarea (ha) 1.5 1.5 1.5 1.6 1.6 1.6 80.9land cover woodland permanentpasturenative pasture woodland permanentpasturenative pasture see Table 62.2 LandscapeThe landscape of the experimental area has twodominant features; the north face of the WhiteElephant Hill (27 ha), which has slopes of 20° to 30°,and the lower slopes (54 ha) which are relatively flat(5° to 10°').The development of the Rowsley fault line hasremoved any stable base point in the stream system.Consequently, the Parwan Creek continues to downcut, disturbing the side slopes of the Parwan Valley.The combination of a geomorphic instability, basaltcapping and highly unstable substrate has resulted ina range of land degradation problems, including gullyerosion, sheet erosion and mass movement.2.3 SoilsThe soils of the experimental area are of Mioceneorigin. The surface soil horizon has compactedfollowing removal of vegetation. Raindrop action hasproduced a widespread crusty surface layer. Thelower soil horizon has very poor structure and quicklyloses stability when wet (Hexter et al. 1956).Consequently, sheet and gully erosion are widespreadon the White Elephant Hills. The major problem isthe extensive tunnel erosion, which in places hasdeveloped into gullies up to 5 metres deep (Hexter etal. 1956). These sites provide harbour for rabbits.Where the surface layer of soil has been eroded, theground cover degrades until eventually, only lichenssurvive (Forbes, 1948).3. STRATEGY FOR RABBIT CONTROLWITHIN THE RESERVEThe strategy for rabbit control started when theresponse of the experimental plot to fencing wasobserved in 1982-83 (Section 3.1). This promptedmore fencing (Section 3.2) and poisoning (Section3.3). The results lead to development of the plan. Theeffects have been monitored by surveys of rabbitnumbers (Section 3.4), and by assessment of thechange in biomass following rabbit exclusion(Section 3.5).3.1 Reclamation at Western end of WhiteElephant HillsAn experimental plot of 1.6 ha at the western end ofthe White Elephant Hills was fenced for rabbitexclusion in 1958, and 344 trees of 19 species wereplanted with tree guards. However, the exclusionfence was not maintained, rabbits reinfested and thecondition of the plot deteriorated. By 1975, the fencehad been undermined to a large degree, and gave onlyminimal protection. A vegetation survey at that timeshowed that there were 250 survivors from theoriginal planting. There were 120 additional shrubs ofAcacia pycnantha (Golden Wattle), which regenerateprofusely and appear to be less palatable to rabbits,but there were no other species (Soil ConservationAuthority, unpublished data). In 1982, rabbits wereexcluded from the plot by repair of the fence, bypoisoning and by lack of food caused by the 1982-83drought. The effectiveness of attempts at revegetationwas examined, and the importance of maintaining thefences was recognised as a key factor in excludingrabbits. This lead to the development of the rabbitcontrol strategy. By 1984, an estimated 3000seedlings had germinated in the experimental plot,with most of the original 19 species being present.Since 1984, there has also been extensiveregeneration of understorey native grasses andshrubs.

Table 2. Application rates of poisoned carrots on sites of rabbit activityYear 86 87 88 89 90 91 92 93 94Application rate (kg/ha) 7 7 7 7 4 4 4 2 1The complexity of the challenge is well illustrated byan example. An eroded area with extensive gullies atthe west end of the White Elephant Hill wasearmarked as a problem area, and was fenced off andpoisoned in 1986 (figure 1). At the time, the Reserveperimeter fence was not in good condition, andneeded repair. Rabbits quickly returned following thepoisoning, because the fence was low in places andrabbits were able to jump over it. The low sectionsoccurred for two reasons. Firstly, cows from theadjoining farm were pushing between the rabbit-proofwire and the barbed wire to graze the phalaris, andwere bending the rabbit-proof wire. Secondly, therewere low spots across an eroding gully. An extraheight of 30 cm of rabbit-proof wire was fitted in thelow sections, which stopped the cows but not therabbits. A guard of barbed wire 10 cm from the top ofthe fence was needed to finally stop the rabbitsentering the area. Overall, approximately 12 monthselapsed between when the area was fenced off andwhen there were visible signs of recovery by thevegetation. The condition of the area in autumn 1989is shown to the right of the fence in Plate 1. The areato the left of the fence is the experimental plotdescribed in the previous paragraph. The photographwas taken facing south-east, approximately 10 mfrom the fence junction. The site is marked in figure1.3.2 The fencing planWhen the Reserve was established in the 1950s theexisting exclusion fence was renewed around theperimeter. This fence linked with the boundary fenceof the experimental catchments (figure 1). Fencesaround the lower slopes section of the experimentalarea were maintained from 1975. Rabbits wereexcluded from subcatchments 1 and 4 from 1978.Little maintenance was carried out on the Reservefences until 1982, when all fences were repaired byrenewing rusted sections along the bottom.The fence around the 1.6 ha experimental plot wasrepaired in 1982 (figure 2). At the time, it was notedthat requirements for fencing maintenance slowlyreduced after a period of continual breaches of thenetting fence, which indicated that rabbits weredispersing under the harsh conditions.further areas of the White Elephant Hill. In 1985 and1986, an additional 14 ha of the worst infested siteswere fenced off into exclusion plots. This is markedin figure 2. Fenced areas were then incorporated intodesign for detailed poisoning described in Section3.3.3.3 The development of targetedpoisoningPoisoning with sodium monofluoroacetate (1080) incarrots is the preferred method of rabbit control. Areview of the properties of this compound is given inAtzert (1981). Up to 1985, the standard procedure forpoisoning rabbits in the White Elephant Reserve wasto lay a trail along a defined track using a bait layer,as shown in figure 2. Two poison-free feeds weregiven with a three day interval, and then poison waslaid. From 1978 to 1984, an additional trail was laidalong the ridge of the White Elephant Hill (figure 2),because it was recognised that rabbits were using theeroded face of the White Elephant Hill to reinfest thelower slopes. Poison was laid twice along the trailover the summer of 1982-83. In 1984, there was someadditional poisoning by hand along the gullies. In1985, poisoning along the standard trail wasaugmented with aerial poisoning, with furtherpoisoning by hand along some gullies.From 1986 to 1993, there was only hand poisoning atall sites of rabbit activity (scratches, dung hills andfeeding sites) throughout the Reserve. The intervalbetween hand applications was increased to sevendays. Application rates were progressively reduced asrabbit numbers declined from levels present prior tofencing. Application rates on sites of rabbit activityare shown in table 2.Fencing and targeted poisoning together reducedrabbit numbers dramatically while rabbits were inplague proportions throughout the Parwan Valley.Spotlight surveys made within the Reserve oncompletion of targeted poisoning in 1989 and 1990located only two rabbits.As noted above, the experimental plot regeneratedrapidly following repair of the fences. This had twoconsequences. Firstly, a high priority was placed onkeeping the fences in good condition by regularpatrolling and maintenance. Secondly, the findingsindicated what could be expected by fencing out

Figure 2. White Elephant Reserve showing poison trails

Plate 1 - Photograph taken in autumn 1989 illustrating effects of rabbit exclusion. Full details are given in the text.3.4 Rabbit number surveys in the ParwanValleyRabbit numbers in the Parwan Valley have beensurveyed by the Department (D. McPhan l ,unpublished data). Thirteen surveys were made indifferent seasons from 1985 to 1989. The data fromthese rabbit surveys have been used to find where therabbits are located on the landscape, and theirdensity.Surveys were made in the early evening, startingafter dusk. Rabbit numbers were observed byspotlight from a vehicle driving slowly along a 9 kmtransect in the Parwan valley. Estimated range of thespotlight is 50 metres. The location of the transect isshown in figure 3. Rabbits were counted on bothsides of the vehicle path. Counts were recorded at 0.5km intervals from a stationary vehicle. The meannumber of rabbits along the transect over the period1985 to 1989 is shown in figure 4. To illustrate theextreme situation, results are also shown of thesurvey on 18 January 1988, when rabbit numberswere high. Rabbit numbers tend to be higher nearpreferred habitat or where it is difficult to poison,e.g., at the foot of steep slopes.Valley. Results are shown in table 4.3.5 Measurements of biomassTwo sets of plots were studied to quantify thechanges in vegetation following rabbit exclusion. Thefirst set was in pastured land of the main catchment;the second on the ridge of the White Elephant Hill.The site in the pasture had a north-west aspect andwas near rabbit activity. Grasses present are mostlyDanthonia species. The site on the ridge had littlerabbit activity since 1986. Grasses are Danthoniaspecies, with a small amount of Poa labillardieri.The method was the same at both sites. An area of 1m 2 was harvested from the exclusion plot and fromthe adjacent area. Material was dried at 45°C for 4hours and weighed to determine biomass. Results aregiven in table 5. Because the vegetation is mostlynative grasses and mosses, the overall ground coverhas been subdivided to show their relativecontributions. These values are included in table 5.Science Officer, Department of Conservation andNatural Resources, Bacchus Marsh Numbers arelower where maintenance is done and conditions areharsher, e.g., near the road with the rabbit prooffence. Survey points 15 to 18 give the closestapproximation to the landscape within theexperimental area.Rabbit numbers have been averaged along the surveytransect for a given land use classification, as definedin Hexter et al. (1956). Results are shown in table 3.Simple arithmetic then gives the expected density ofrabbits for that land use class throughout the Parwan

Figure 3. Location of transect in Parwan Valley for rabbitsurveys, and survey points

Figure 4. Results of rabbit surveys, 1985 to 1989Figure 5. Water Balance Model.Difference between Observed and Predicted Values (mm)

Table 3. Mean rabbit numbers observed on homogeneous landscapeLandscape unit Basalt plateau Scarp(steep slopes)Paddocks onvalley floorRoad withrabbit prooffencePaddocksaround base ofhillLand use class 2A 5 4B 4B 5Mean 11 26 18 9 59Table 4. Mean rabbit density for surveyed land class in Parwan ValleyLand class Area (km 2 ) Number/km 22A 67 7004B 37 4005 42 1500Table 5. Characteristics of biomass at given sitesSite Year established Dry biomass(kg/m 2 )Ground cover(%)Proportion:native grassesProportion:moss and litter1954 0.958 100 0.5 0.5Pasture (a) 1991 0.386 90 0.5 0.4No exclusion 0.098 50 0.35 0.15Ridge 1954 0.428 80 0.4 0.4No exclusion (b) 0.133 60 0.3 0.3(a)(b)Note that plant numbers are similar in the three plots.little rabbit activity was observed at this site.4. RABBIT EXCLUSION ANDHYDROLOGYPrevious sections have shown that the rabbit controlstrategy was successful in controlling rabbit activity,and consequently allowing the vegetation toregenerate. It would be valuable for managementpurposes to know if there has been a detectablechange in run-off associated with this rabbitexclusion. The north face of the White Elephant Hillis of particular interest, because it is themost degraded part of the landscape. The estimationof run-off from the north face is described in section4.1. A simple water balance model is applied to thenorth face in section 4.2 to test for changes in run-off.Findings are presented in section 4.3.4.1 Estimation of surface run-off from thenorth face of the White Elephant HillThe stream gauge in the main catchment (figure 1)measures combined drainage from the north face ofthe White Elephant Hill and from eroded andvegetated areas. This can be expressed as -Q mc = Q nf + Q ea + Q ls(I)

Table 6. Components of landscape of the vegetated area on the lower slopes in 1978LandscapeRepresentativesubcatchmentProportion of lowerslopes vegetated area(A)Savannah woodland 1 0.29Native pasture north aspect 3 0.42Native pasture south aspect 6 0.29where,Q mc is run-off from main catchment,Q nf is run-off from the north face,Q ea is run-off from the eroded areas,Ql s is run-off from the lower slopes.Run-off from the north face and eroded areas isestimated by subtracting the contribution due to thelower slopes from the total run-off, because there isno stream gauging at the break of slope. In turn, thecontribution from the lower slopes is estimated by theweighted sum of run-off from the individualreference subcatchments, thus –Ql s = A 1 Q 1 + A 3 Q 3 + A 6 Q 6 (2)where Q i is run-off from subcatchment i,A i is proportion of the main catchment represented bysubcatchment i,i = 1, 3, 6. Values of A are given in table 6.Equation (2) is evaluated to give the estimated run-offfrom the lower slopes section. This is substituted inequation (1) to give estimated run-off from the northface and eroded areas by difference. Possible changesin runoff due to rabbit exclusion can now be testedusing a water balance model. Details are given inSection 4.2.The above derivation assumes that partitioning theland use of the main catchment is valid, and thatsummation of the surface run-off (weighted for area)from the subcatchments is a reasonableapproximation of the actual processes. Routing ofsurface run-off and channel absorption effectsbetween the break of slope and the gauging stationare considered to be small and have not been includedin the analysis. It is also assumed that the relationshipof the reference subcatchments and the lower slopessection of the main catchment remains constantduring the course of the study. This appearsreasonable as the targeted poisoning design operatedon both subcatchments and the main catchment.4.2 Water balance modelA simple water balance model has been applied to thenorth face of the White Elephant Hill to quantify anychange in surface run-off.The daily water balance is expressed as:P = I + ET + Q (3)where P is rainfall (mm),I is interception (mm),ET is evaporation/transpiration (mm),Q is streamflow (mm).For modelling purposes, this equation has beenrearranged as:Q = P - I - ET (4)Variables on the right of equation (4) can beestimated. Interception has been estimatedempirically as the loss in rainfall before run-offcommences. Daily values applied for a given monthare shown in table 7. An empirical surrogate for theET term has been developed, defined as the time indays needed to dry out the catchment. Rainfall isdepleted daily by an amount equal to the number ofdays since rain fell multiplied by 0.64 (units: mm).The depletion continues until either the day value intable 7 is reached, or rain falls. If the day value isreached, a daily maximum of 14x0.64 = 8.96 mm issubtracted. When rain falls, the day counter restartsfrom unity. Output from the model is predictedstreamflow (mm).

Table 7. Empirical model parametersMonth Jan Feb Mar Apr May June July Aug Sept Oct Nov DecInterception loss (mm) 6 6.2 3.5 4 3 2 2 3 4 5 8 6Time to dry thecatchment (days)6 6 8 10 10 12 12 10 10 8 8 6The data set consists of daily rainfall and run-off recordsfrom January 1978 to December 1993 for the six referencesubcatchments and the main catchment (figure 1). Duringthe calibration period (January 1978 to December 1980),model parameters have been estimated by successiveapproximation to minimise differences between observedand predicted run-off. There is a period of no change fromJanuary 1981 to December 1984. During the test period(January 1985 to December 1993) parameter values fromthe calibration period have been used in the model topredict run-off. For a null hypothesis that there is nodifference in run-off in the test period, only smalldifferences between observed and predicted run-off areexpected. Differences have been tested for statisticalsignificance. Any significant differences are due to theeffects of rabbit control.4.3 ResultsThe difference between estimated run-off and the valuepredicted by the water balance model has been plotted frothe length of record considered in this study. This isshown in figure 5.The aim of the analysis is to identify any major changes insurface run-off associated with rabbit control. With this inmind, the differences between estimated and predictedrun-off have been averaged at monthly intervals over thecalibration and test periods. Specifically, over thecalibration period, the January 1978, 1979 and 1980differences have been summed, and the mean has beenplotted in figure 6. This procedure has been applied forthe remaining months in the calibration period. Resultsare shown in figure 6.The same procedure has been used during the test period(January 1985 to December 1993). Average monthlydifferences for the interval with rabbits excluded is shownin figure 6. Total monthly decrease in surface run-off overthe 9 year test period amounts to 87 mm (figure 6). Theaverage annual decrease is 87/9 ~10 mm, which isapproximately 20% of mean annual run-off. This will bediscussed further in section 6.5. FINANCIAL ANALYSISThe increased water use by the vegetation on site isreflected in increased biomass. The objective here is toquantify this change in practical financial terms. In theabsence of any firm information, estimates have beenmade of the costs involved and benefits received toidentify the major financial factors.The analysis has aimed to normalize costs to the farmlevel, so that travel times and associated costs fromMelbourne to the Parwan area have been excluded.Similarly, the off-site costs of sediment export (forexample, costs incurred in removal of silt from MeltonReservoir) have not been considered. Costs of rabbitsurveys have not been included.The only benefit assessed in this analysis is that due toincreased biomass using the data from the exclusion plots(table 5). Other benefits which have not been quantifiedinclude maintenance of productive land due to reducedsoil erosion an improvement in native habitat on nonagriculturalsites.5.1 CostsCosts arise in allocated time and in materials. Allocatedtime includes time spent by personnel in the field, i.e., forsurveys, labour and supervision. Materials costs occur inpurchase of fencing, repair of existing fences, transportand purchase of poisoned carrots.5.1.1 Personnel timeOne day per week is spent in the Reserve Tasks varyduring the year. New fences were installed over thesummer months, while repairs are a continual task. It isestimated that half the working day is spent on rabbitcontrol measures. Time involved is 5 hours per day at$10.00 per hour (including on-costs) for 40 weeks peryear. Total annual cost for personnel is 5 x $10.00 x 40 =$2000.00.

Figure 6 Mean monthly differences between observed run-off and values predictedfrom water balance modelTable 8 Costs for supply of poisoned carrotsYear 86 87 88 89 90 91 92 93 94Number of 20 kg bags 60 60 60 60 47 47 47 30 18Cost – free feed ($) 260 260 260 260 202 202 202 130 78Cost 1080 feed ($) 180 180 180 180 144 144 144 90 54Total Cost ($) 440 440 440 440 346 346 346 220 132wire, and a gate with5.1.2 Cost of poisoned carrotsCarrots were purchased at bulk feed stores in20 kg bags at prices varying from $2.50 to$6.00 per bag depending on season. Mean costwas $4.00 per bag. The number of 20 kgpurchased are given in table 8. TheDepartment charges $2.00 per 20 kg bag for1080 poison, plus $0.50 Approximate cost of a100 metre roll of rabbit-proof wire for asuitable bag to hold the poisoned carrots.There is a cutting fee of $2.50 per 20 kg bag.All up cost for supply of cut and poisonedcarrots is $9.00 per 20 kg bag. The cost for afree feed is purchase and cutting only; a totalof $6.50 per 20 kg bag. There are two freefeeds given, and then poison.Total annual costs for supply of carrots aregiven in table 8.5.1.3 Cost of new fencesItems for new fences are wooden corner postsand supports, steel posts at 5 m intervals,rabbit-proof wire, support wire and barbedsupport. An approximate all inclusive cost fornew fences including labour is $10.00 permetre.Lengths of new fences and associated annualcosts are given in table 9.5.1.4 Cost for repair of existing fences kgbagsApproximate cost of a 100 metre roll ofrabbit-proof wire is $250.00 for standardwidth (1 m) and $80.00 for narrow width (30cm). Standard width is used for major repairs.Narrow width is used to repair breaches closeto ground level, gaps where cows pushthrough, and to increase fence height in lowspots. Labour includes pegging down thewire, and the carting of stones for use asweights. An approximate all inclusive cost forrepair of These existing fences includinglabour is $3.00 per metre.Lengths of existing fences which wererepaired annually and associated costs are

given in table 9.$16.00/ha. Second, there is the currentmaintenance strategy, where annual runningcosts are approximately $2700/300ha =$9.00/ha. The major item is the time and.effort made to ensure that rabbits areexcluded. costs have been stable for sevenyears; there is a slightTable 9 Costs of fencingOperation 1982 1983 1984 1985 1986 1987 to 1994Repair of existing fenceLength (km) 2.0 1.0 0.5 0.2 0.1 0.1Cost ($) 6000 3000 1500 600 300 300New fencingLength (km) 0.3 0.2Cost ($) 3000 2000Table 10 Summary of costsYear Personnel Poisoned carrots Fence repairs New fences Rounded total($) ($) ($) ($) ($)82 2000 n.a. 6000 >800083 2000 n.a. 3000 >500084 2000 n.a. 1500 >350085 2000 n.a. 600 3000 >560086 2000 440 300 2000 470087 2000 440 300 270088 2000 440 300 270089 2000 440 300 270090 2000 346 300 270091 2000 346 300 270092 2000 346 300 270093 2000 220 300 250094 2000 132 300 24005.1.5 Summary of costsThe individual items described in the aboveparagraphs have been summarised in table 10 togive an indication of overall costs.Prior to 1986, rabbit poisoning was done byregional Departmental staff, and included aerialpoisoning. Costs were not available.Table 10 shows two distinct phases. First, thereis an interval of five years with relatively highcosts due to requirements of new fences andrepairs to existing fences. Costs in this periodare greater than $4800/300ha or $16.00/ha.Second, there is the current maintenancestrategy, where annual running costs areapproximately $2700/300 ha = $9.00/ha. Themajor item is the time and effortmade to ensure that rabbits are excluded. Thesecosts have been stable for seven years; there is aslight decrease with time as the quantity ofcarrots required for rabbit control has decreased.5.2 BenefitsConsider that rabbits are controlled throughoutthe Parwan Valley in one year, and that theeconomic benefit becomes available in the sameyear. It is assumed that the increase in biomasscorresponds to the values in table 5, (i.e., 0.386 -0.098 - 300g/m 2 ). This has occurred over a threeyear period, so the benefit is -100g/m 2 annually.

In the absence of rabbit survey informationwithin the experimental area (land use class 5), itis assumed that grazing pressure is similar to thevalues surveyed in the Parwan Valley in thesame land use class (i.e., 1500/1= 2 , table 4).Mean grazing pressure on pastures on the floorof the Parwan Valley (land use classes 2A and4B) has been estimated from table 4 as 550/km 2 .On the basis of relative grazing pressure byrabbits, an average improvement in pasturebiomass would be expected of 550/1500 timesthe results observed in the exclusion plots, i.e.,37 g/m (= 370kg/ha) increase in biomass. For anominal weight of 25 kg for a bale of dry feedcosting $4.00, this converts to approximately 14bales/ha of dry feed per annum, or $56.00/ha perannum. The financial benefit of this additionalbiomass exceeds the total estimated costs($9.00/ha per annum, Section 5.1.5), so themethod is financially feasible.This gain can also be expressed in more practicalterms. On the assumption that a wether eatsapproximately 1 kg of dry feed per day tosurvive (D Perry, pers. comm.), its feedrequirements for a year amount to 365 kg. This isapproximately equal to the annual improvementin biomass determined in the previous paragraph.Therefore, the gain in biomass due to rabbitcontrol corresponds in general terms to animprovement in the gross margin ofapproximately 1 DSE/ha.This is a broad indication only as no allowancehas been made for variations in productivity ofdifferent pasture species or for varying siteproductivity. Preferential grazing by rabbits hasnot been considered. Nevertheless, these figuresdo indicate the value of a thorough strategy forrabbit control. This analysis has ongoingimplications for the local industries of lucernehay, dairy farming, beef and fine woolproduction.6. DISCUSSION6.1 The success ofstrategythe rabbit controlThere was little vegetation on the WhiteElephant Hill when the experimental area wasestablished in the early 1950s. Rabbit numberswere high enough to escalate land degradation,and their grazing pressure effectively suppressedregeneration (Section 3.1). The White ElephantHill provided ideal rabbit habitat because thesoils are highly erodible and the landscape isbroken. Forbes (1948, page 29) described thestudy area and the upper Parwan Valley ashaving "fantastic numbers of the vermin". Rabbitnumbers surveyed at this latter site are includedin column 5, table 3.The standard poisoning plan made little impacton the source areas of rabbits where vehicleaccess was impossible. Short migration distancesfrom the deep gullies ensured that any depletionin rabbit numbers by the standard poisoning planwas only temporary. Rabbit migrations were alsoassisted by the lack of maintenance of fences.An attempt was made to keep rabbit numberslow during the drought of 1982-83 by repairingthe fences in the experimental area. The resultsfrom the experimental plot (Section 3.1)indicated that rabbit exclusion was an importantfactor in regeneration. Consequently, other areaswere fenced off, in particular at the western endof the White Elephant Hill (figure 2) whereerosion was severe. Fenced areas were thenincorporated into a poisoning strategy which wastargeted at sites of rabbit activity (Section 3.3).The major advantage of this development is in itseffectiveness. Results show that low applicationrates at sites of rabbit activity maintain adequatecontrol (table 2). The benefits are that thequantity of applied poison is minimised and thatcosts are reduced.When this study was initiated, time constraintsdictated that free feeding and poisoning couldonly be done at 7-day intervals. Results haveshown that this interval was effective; both thetechnique and the findings have been validated inthe literature. Cowan et al (1987) studiedpercentage of rabbits consuming baits at variousdistances from the warren. They found that 90%of rabbits consumed baits up to 25 metres fromwarrens up to 8 days after baiting commenced.This shows the importance of firstly applyingbaits close to warrens, and secondly of allowingsufficient time for all rabbits to browse the baits.Cowan et al (1987) conducted their study inEngland where the climate is cooler than inVictoria and the bait is likely to stay fresh for alonger period. In this study, much of the handpoisoning was carried out in autumn and winterwhen cooler conditions prevail.The overall result of the fences and targetedpoisoning has been that rabbit numbers withinthe Reserve have dropped from high values(figure 4) to close to zero. Surveys in 1989 and1990 gave only two sightings over a 1.5 kmcourse. This contrasts with numbers surveyed inthe surrounding Parwan Valley (table 3). In theabsence of any control measures, rabbit numberssimilar to those in the valley would be expected,i.e., in the range 4001500/km 2 (table 4).The maintenance of declining rabbit numbersfrom this point forwards can be related to theimproved technique of bait feeding at sites ofactivity with an extended period between baitruns. Poisoning complements proper fence

maintenance as a successful tool in rabbitcontrol, in that the fence minimizes thedispersion of high rabbit populations andmaintains areas of low rabbit density as lessfavoured sites.6.2 The relationship of rabbits andrun-off, and possible mechanismsThe broad scale damage caused by grazingrabbits and the consequences has been welldocumented in the literature (Stead, 1935;Forbes, 1948; Hexter et al. 1956; Hills, 1975;Williams et al. 1995).The native grasses of the area (Danthonia, Stipaand Poa species) have a tussock habit of growth.As a result of an extended period of high rabbitpopulations these tussocks were closely croppedand continually harvested. They resembled aseries of pads with low ground cover (30 to 40%)and low biomass(-0.1kg/m 2 , table 5).When the grazing pressure was removed, treeseedlings, understorey and grasses began toregenerate. The native grasses recovered to atussock formation with 80 to 90% cover, biomassincreased by a factor of approximately 4 times in3 to 4 years, with the prospect of further longterm improvement (table 5). It appears that thereis minimal increase in numbers of native grasses,with the emphasis being in growth of existingplants. As the native grasses recover, interceptionincreases and the microclimate under thevegetative canopy becomes less severe. Theseconditions favour growth of mosses, which isshown by an increase in ground cover (from 0.15to 0.4, column 6, table 5). The growth of mossesretains surface moisture on site, andconsequently increases infiltration. The growthpattern of mosses is seasonal, as they dry outover the summer months.T he reductions in run-off go through a cycle, withmaximum reduction in mid-winter and minimumreduction over mid-summer (figure 6). In theabsence of direct observation, this may be due toa combination of the increase in biomass andelevated infiltration. This would be consistentwith the findings of Dunin and Downes (1962)and Dunin (1965), when poor native pasture wasconverted to annual improved pasture. May is thetime of the autumn flush of growth, whentemperatures are still elevated. In the calibrationperiod, the rabbits would have continuallycropped this biomass. However, the reduction inrabbit numbers has allowed the development ofthe understorey and a large number of acacias,which were not present in the exclusion plots.This has increased interception and providedshading and windbreak for the development ofground cover (table 5). Evaporation rates arereduced in the shade, and consequently surfacesoil moisture tends to be higher (Chow, 1964).These conditions remain favourable over thewinter months for further infiltration and plantgrowth in the Spring. However, conditionschange dramatically over the summer months.Low summer rainfall and high temperatures putall vegetation under soil water stress, andevaporation and transpiration are at low levels.High rainfall intensities still tend to compact anyareas of exposed soil surface and produce surfacerun-off.Although this explanation describes the overallfindings, the details are tentative and requirefurther study to clarify the underlying processes.In particular, experiments on the seasonalvariation of infiltration and its relationship toground cover and rainfall intensity wouldprovide further insight into best practice.As part of ongoing work, surveys of the maincatchment have shown that the sheet erodedareas are rehabilitating very slowly, so that theseareas now yield the greater proportion of theobserved surface run-off. Therefore, the increasein interception and infiltration has occurred inareas with higher ground cover (pasture and/oracacias), due to exclusion of rabbits. Furtherdecreases in run-off are expected as the sheeteroded areas rehabilitate. The possibility orfeasibility of accelerating this rehabilitationwarrants further investigation.6.3 Future managementrecommendationsThe landscape of the White Elephant Reserve istypical of the slopes of the Parwan valley. Deepgullies and sheet erosion on solodic soils occurthroughout the Reserve, in the Parwan valley, inthe adjacent Werribee River catchment and nearAnthony’s cutting. The management practiceswhich have been implemented in the Reservehave been effective in (i) reducing rabbitnumbers to very low levels, (ii) enablingregeneration of vegetation to take place and (iii)retaining water on site and reducing run-off. Asmore areas have been fenced out and poisoned,rehabilitation of eroded areas has beencommenced, mostly by extensive tree plantingwith straw mulch to retain moisture.In principle, it should be straightforward toextend these management practices to thebroader scale. Rabbit haven is made unattractivein two ways, by fences selected areas and bypoisoning sites of rabbit activity. The method isfinancially feasible for the 300 ha reserve. Themajor costs are in the initial purchase of fencesto exclude rabbits, and in ongoing maintenance:the benefits are in greatly increased biomass,reduced run-off and higher productivity from theselected area. These are important considerations

for efficient management of low rain fall areas.In these situations, it is advisable to treat water asa resource which should be retained on site foruse at critical times. Suggested measures toretain moisture include mulching, growingspecies which provide litter and by providing theenvironment to facilitate distribution of moisturethrough the soil profile.The current practice of rabbit control haslimitations because laying a poison trail does notaddress the source areas of reinfestation. Largekills are made in the vicinity of the trail, butrabbits breed in unthreatened areas and thendisperse back into the farmlands. A continualcycle of depopulation and repopulation is almostguaranteed.An alternative practice is to firstly appraise theland and appreciate that different land classesrequire different intensities of land management.The eroded slopes of the White Elephant Hillneed fences to separate them from the flatterpastured areas. The preferred poisoning scheduleis to poison at sites of rabbit activity with lowapplication rates, and to allow sufficient time forbaits to be accepted.7. CONCLUSIONSThis study has shown that the policy ofconstructing fences around rocky or eroded hillsand leaving these areas to the rabbits is noteffective in minimising rabbit numbers, becausethese areas serve as sources of reinfestation. Aneffective method of rabbit control is to fence outthe hillslopes from the paddocks, and ifnecessary, to even fence out critical areas onhillslopes. All areas are then carefully poisonedwith attention being given to sites of rabbitactivity. It is important to give rabbits sufficienttime to take the poisoned baits. This studyconfirms findings that seven days should beallowed for high levels of bait acceptance. Rabbitnumbers can be kept low only by continuedfence maintenance and poisoning where breachesoccur, so that rabbits are prevented fromrecolonizing a site.In fact, this study has given rise to the interestingsituation where low rabbit numbers have beenmaintained for approximately 10 yearsthroughout an area of some 300 ha, and this issurrounded by an extensive area of farmlandwhere rabbits are numerous.The benefits of this method of rabbit controlhave lead to a decrease in run-off and an increasein pasture production. The increased pastureproduction corresponds to an approximateimprovement of 1 DSE/ha in the gross margin.Management to decrease surface run-off is apositive step in erosion control and has specialsignificance in low rainfall areas.Simultaneously, these practices increaseinterception and infiltration, and maximize wateruse on site. This involves the completevegetative layer, including trees, shrubs, grasses,ground litter, macropores and soilmicroorganisms.This study was initiated during a drought, whenrabbit numbers were low. Psychologically, this isa stronger starting point than when rabbits are inplague proportions. The challenge now is toundertake this strategy at other sites where rabbitnumbers are high and deep gullies and sheeterosion occur on solodic soils.The results of this study have clear relevance toland of similar nature, e.g., the Parwan valley,the Werribee River catchment, the Pentland Hillsarea and near Anthony’s cutting. However, thebasic principle of exclusion fences where there isready reinfestation to replace poisonedpopulations holds in all parts of Victoria, whilethe relationship between streamflow, plantgrowth and rabbit control should hold for at leastall dry (up to 800 mm annual rainfall)environments where there is significant surfacerun-off.8. ACKNOWLEDGEMENTSFencing repairs, upgrading and maintenancefrom 1990 to 1993 were funded by the NationalSoil Conservation Program. Devan McPhan wasinvolved in the Departmental rabbit poisoningplan and contributed enthusiasm for the studyand information on rabbit numbers. CraigClifton, David Cummings and Les Russellprovided valuable comments on drafts of thisreport.9. REFERENCESAtzert, S.P., 1981. A Review of SodiumMonoflouroacetate (Compound 1080), its Properties,Toxicology and Use in Predator and Rodent Control.U.S. Fish and Wildlife Service Special ScientificReport - Wildlife No. 146.Chow, V.T., 1964. Handbook of Applied Hydrology.McGraw-Hill Book Company. New York.Cowan, D.P., Vaughan, J.A. and Christer, W.G., 1987.Bait consumption by the European rabbit in southernEngland. Journal of Wildlife Management 51, 386-92.Dunin, F.X., and Downes, R.G., 1962. The effect ofSubterranean Clover and Wimmera Ryegrass incontrolling surface run-off from four-acre catchmentsnear Bacchus Marsh, Victoria. Jour. Exp. Agr. andAnim. Husb. 2(6), 148-152.

Dunin, F.X., 1965. The effect of vegetative changes onparameters for estimating run-off near Bacchus Marsh,Victoria. Civil Eng. Trans. Inst. Eng. Aust., 7, 16-22.Forbes, LG., 1948. Erosion in the Melton ReservoirCatchment. State Rivers and Water SupplyCommission., Victoria.Hexter, G.W., Leslie, T.I. and Pels, S., 1956. A LandUse Survey of the Parwan Valley, Victoria. SoilConservation Authority, Victoria.Hills, E.S., 1975.Physiography of Victoria: anintroduction to geomorphology. Whitcombe & TombsPty. Ltd., Australia. 5th edition.Stead, D.G., 1935. The rabbit in Australia. Winn andCo. Sydney.Williams, K. , Parer, I., Coman, B., Burley, J andBraysher, M, 1995. Managing Vertebrate Pests:Rabbits. Australian Government Publishing Service.Canberra.Wu, A.Y.K., 1980. Parwan Hydrological ExperimentalArea. Soil Conservation Authority, Victoria.Wu, A.Y.K., Hartland, R. and Papworth, M. P., 1986.Parwan Hydrological Experimental Area. LandProtection Division. Department of Conservation,Forests and Lands, Victoria.