Previous PageTable Of Contents

The role of urban landscape irrigation in inland New South Wales in changing the growth potential of Queensland fruit fly

H.S. Mavi, B.C. Dominiak and H.I. Nicol

NSW Agriculture, 161 Kite Street, Orange, NSW 2800, Australia.
Phone 02 6391 3637 Fax 02 6391 3767
Email: harpal.mavi@agric.nsw.gov.au

Abstract

In exploring the impact of town environments on fruit fly potential, water use and other town statistics were obtained from 17 towns west of the Dividing Range. Based on these data, models were developed that explain the water use in landscape irrigation during different months and within different towns. Landscape or garden water use remains more or less unchanged during winter. A significant change away from the static winter irrigation occurs in October, with a steep rise in November and a peak at the height of summer in January and February.

Many factors - like rainfall, evaporative demand, town size, cost of water, source of water supply and affluence of the towns - were examined to assess why some towns have high water use compared with others. Water cost was the primary determinant, with environmental factors such as rainfall and evaporation demand being less important. Contrary to traditional perceptions, town sizes and affluence played a small role in landscape water use.

Introduction

The Queensland fruit fly Bactrocera tryoni (Froggatt)(QFF) is a severe pest of horticultural areas in eastern Australia. The Gosford-Sydney area in New South Wales was considered by Froggatt (1909) to be the natural southern limit of the pest. The southern limits of distribution are set by thermal restrictions while western limits are set by insufficient rainfall (Fletcher, 1987). The Fruit Fly Exclusion Zone (FFEZ) was developed in 1994 as a trade zone based on being demonstrability free from QFF (Anon. 1993). The FFEZ and the nearby Risk Reduction Zone (RRZ) are both are outside the ecological limit of the range of QFF (Yonow and Sutherst, 1998).

The primary determinant from the abundance of QFF is moisture (Bateman 1972), particularly towards the southern fringe of its permanent distribution. Mature larvae and newly emerged adults are most susceptible stages to desiccation. QFF fecundity was drastically reduced to about 17% of normal in drought conditions (Bateman, 1968). Reduced environmental moisture may also indirectly affect survival and fecundity of QFF by reducing the abundance of bacteria (Drew et al.1984) which is the main source of protein for QFF. Inadequate moisture will also limit host tree growth and fruiting. In drought conditions, less fruit was available and the shrivelled fruit on trees dropped prematurely; this resulted in 89% of eggs perishing before reaching the pupal stage (Bateman, 1968).

While the dependence of QFF on temperature and moisture is relatively well researched in laboratory situations, the understanding of how laboratory results translate into field survival is relatively poorly understood. Mavi and Crichton (1998) demonstrated the regular nature of summer droughts in southern NSW. Mavi and Dominiak (1999) noted the increased humidity and temperature found in modified town environments, compared with the surrounding unmodified rural environment.

This paper further examines the possible reasons for the localised effects of the modified urban environment. The differences in the unmodified and modified environments are used to examine the survival potential of QFF and how this information might be used in control strategies in the RRZ.

Methods

Assessment 1.

We wanted to develop a model which would best identify changes in the monthly landscape irrigation pattern. Seventeen local councils (Map 1) were contacted and information gathered covering water supplied to the town, sewerage output, number of connections, block size, and the source of water supply for the 1996/97 year (Table 1).

Map. 1. Map of New South Wales showing the towns covered in the analysis and their locations relative to the Fruit Fly Exclusion Zone

The monthly amount of landscape irrigation for each town was estimated by subtracting the sewerage output from the water pumped into each town. This monthly figure was divided by the number of connections to create an estimated monthly landscape water use (kilolitres per connection). Any losses of water in pipes and volumes used by industry were assumed to be equal for all towns on a per connection basis.

A model, which best accounted for the calculated landscape irrigation, was developed using rainfall, water deficit, evaporation, months of the year, and any interaction between these factors. Climatological data were obtained from MetAccess package (Donnelly et al., 1997). Stepwise regression was performed using Genstat 5 (Genstat 5 Committee, 1998).

Assessment 2

In the same 17 towns, the calculated monthly landscape water use figure were summed to determine the annual use and compared with population size, cost of water, source of water and average rate notice. The cost of water was requested from councils, however, the answers did not have a common base; some councils have a free initial allocation while others charge different rates for different volumes. Previous experience had suggested that population size and community affluence had an impact of the amount of water used, along with the climate (water deficit). Subsequently the population size, water average account ($ per connection), sewerage water average account ($ per connection) and average rate notice per residence was obtained from Anon (1999) for the 1996/97 year.

Table 1. Annual data used in the development of the model to explain landscape irrigation.

Town

Landscape irrigation per connection (kL)

Rainfall /Evaporation (mm)

Town population

Rates notice per property ($)

Water cost per connection ($)

Albury

762

629/1748

42,382

704

271

Bathurst

369

664/1322

29,355

591

579

Condobolin

492

326/1968

3,500

215

505

Coonamble

696

460/2191

4,970

173

293

Cowra

543

595/1360

12,524

220

461

Forbes

799

538/1935

10,331

337

505

Gilgandra

455

630/1951

4,862

116

392

Griffith

1172

298/1767

22,462

426

413

Hay

1312

299/2149

3,795

244

384

Lake Cargelligo

488

393/2077

1,300

215

@

Leeton

863

380/1976

11,394

395

447

Narrandera

992

386/992

7,074

264

418

Narromine

388

513/2032

6,714

296

281

Orange

229

848/1494

35,050

731

320

Trangie

517

500/2032

951

*79

@

Wagga

703

530/1644

56,188

394

@

Young

411

756/1360

11,385

378

497

@ no data available * estimate only.

The average rate notice per residence was used as an estimate of town affluence. The sewerage water costs were not available for 3 out the 17 towns and an analysis was conducted on the remaining 14 towns. Sewerage water was then removed from the array of variables and Dubbo excluded because no figures were available leaving 16 towns in the subsequent development of the second model.

Table 2. Summary of terms fitted to the model in regression and the percentage of variance accounted for by the model.

Fitted term

Estimate

Standard error

T probability

% variance accounted for by each additional term

a. month

     

42.4%

- July

-9.1

24.4

0.710

 

- August

14.3

28.1

0.613

 

- September

21.2

27.2

0.438

 

- October

38.9

29.2

0.185

 

- November

44.9

30.6

0.145

 

- December

70.0

30.0

0.021

 

- January

52.2

32.4

0.109

 

- February

49.7

27.5

0.073

 

- March

25.3

26.4

0.339

 

- April

40.9

24.8

0.100

 

- May

31.7

30.8

0.304

 

- June

7.4

27.6

0.789

 

b. Evaporation

0.3035

0.0669

<0.001

52.0%

c. Rainfall

0.137

0.266

0.603

57.1%

d. Rain.month

     

62.2%

- July

0.137

0.262

0.603

 

- August

-0.233

0.348

0.503

 

- September

-0.330

0.326

0.313

 

- October

-0.699

0.366

0.058

 

- November

-0.979

0.425

0.022

 

- December

-1.251

0.334

<0.001

 

- January

-0.859

0.396

0.032

 

- February

-0.954

0.325

0.004

 

- March

0.032

0.414

0.939

 

- April

1.86

1.34

0.167

 

- May

-0.305

0.439

0.488

 

- June

0.208

0.484

0.668

 

Results

Assessment 1

Table 1 shows the 17 towns and their different characteristics. We looked to develop a model which best accounted for the landscape irrigation use. Table 2 shows the fitted terms and the percentage of variance accounted for by the model. The model that accounted for 62.2 percent of variance was

Use = a.month + b.evaporation + c.rain + d.rain.month

When individual months were analysed, landscape irrigation for the months of July, August, September, May and June did not differ significantly. October, November, December, January, and February had significantly higher landscape irrigation than the cooler months. April was intermediate. However April was an unusually dry month in all sites except Griffith. When the Griffith/April data was removed from the data and the model was re-run, the best fit remained with the above model (61.9 percent of variance). For each mm increase in evaporation, the landscape irrigation increased by 0.3035 kL/connection.

Assessment 2

Data pertaining to water used in irrigation, open pan evaporation recorded, affluence of the towns, source of water supply and cost of water were tabulated and charted (Fig.1) to examine the emerging patterns and correlations. There are four interesting relationships.

1) Evaporative demand and water used. Annual water evaporated at each of the towns was plotted against water used (Fig 1a). It revealed that more the western location of a town, greater is the amount of evaporation recorded. However, no similar trend was noticed in water used and geographical location. The correlation between evaporation and water used was 0.5

2). Town affluence, town size and water used. Assessment rates of the towns were plotted against water used. No clear trends or pattern emerged (Fig. 1b). For some of the towns, assessment rates were high while water used was low. For some others, assessment rates were comparatively low but water used was high. The correlation between the two factors was just below – 0.21. A similar poor correlation was found between town size and water use in landscape irrigation.

3). Water source and water used. Towns were grouped according to the source of water supply. There appears to be a broad and generalised pattern (Fig. 1c). Towns supplied water from rivers use much more landscape irrigation than towns supplied water from dams while towns supplied from underground fall between the two extremes. Within the towns supplied water from underground, there is big difference in the water used. The depth from where water is pumped reflects on the water used. Water used is less at towns where water is pumped from greater depths (Gilgandra bores are at 100 m depth, Coonamble is at 400 m depth) while more is used where water is pumped from shallow depth (Narrandera town bores pump from a 36 m depth). We have no explanation for Narromine, which has a low water use, and town bore depths of only about 30 m.

4) Water supply rates and water used. For a majority of towns an opposite relationship existed between water supply cost and water used (Fig. 1d). Higher the cost of water, lesser was the amount of water used and vice versa. A correlation of – 0.83 was established between water supply cost and water used.

Finding these trends and correlations, we looked to develop a model which best accounted for the annual landscape irrigation use. Table 3 demonstrates the fitted

terms and the percentage of variance accounted for by the model.

Table 3. Summary of terms fitted to the model in regression and the percentage

of variance accounted for by the model.

Fitted terms

Estimate

Standard error

T probability

Variance accounted for by
the model

Constant

2450

486

<0.001

 

a.water cost

-0.612

0.133

<0.001

66.9%

b.rainfall

-0.943

0.340

0.017

71.8%

c.evaporation

-0.458

0.191

0.033

77.6%

The model that accounted for 77.6 percent of variance was:

Water Use = constant + a.water cost + b.rainfall + c.evaporation

For each $ increase in water cost per connection, average landscape irrigation decreased by 0.582 kL/connection. For each mm increase in rainfall, landscape irrigation

decreased by 0.697 kL/connection and each mm increase in evaporation decreased

landscape irrigation by 0.340 kL/connection.

Discussion

Moisture remains a critical determinant of QFF populations and correlations between summer rainfall and QFF numbers are significant (Bateman 1968). If

other environmental factors were not limiting, a relatively high QFF population would survive with a mean monthly rainfall greater than 48 mm. In inland NSW, this rainfall scenario is poorly fulfilled. Rainfall data for the 17 studied towns show that monthly rainfall ranged between 4 and 30 mm in at least one month of the summer season. At many of the towns, there are three to four months with this range of natural rainfall suggesting that much of the summer season is too dry to sustain the fruit fly population, unless the environment is modified.

Assessment 1

The lack of impact of the landscape irrigation in the winter months of May, June, July, August and September may have several explanations. Landscape irrigation is calculated by the difference between the water pumped into the town and the amount pumped out via sewerage. This formula does not take into account any leakage in pipes and the apparent flat rate of water use may merely be a reflection of pipe leakage. Also basic functions such as the use of water for drinking, watering pot plants and other uses that do not end up in the sewerage system, are not accounted for by our calculations. Given the T value for February, it might be assumed that February is the most important month for water use and possibly for fruit fly survival. The January figure is unexpected low for the peak of the season; our explanation is that a proportion of people leave their residences during the main holiday period of January and that this lowers the T value - we have no other obvious explanation. None of these towns are noted tourist destinations and so we have no chance to see the possible reverse trend for January in tourist towns.

Assessment 2

The model that best accounted for annual landscape water use is primarily based on water cost. This is next modified by rainfall and then evaporation. The model did not support the theories suggesting the role of town size or population (and the associated increase in affluence as estimated by rate notices). Towns, where water costs are high and landscape irrigation is likely to be low and hence fruit fly populations not supported by considerable landscape irrigation might be rapidly identified by their high cost of water. Based on this assumption, Carrathool ($693), Condobolin ($505), Forbes ($505) and Jerilderie ($502) are less likely to have fruit fly problems. Conversely towns with low water costs (Culcairn $212, Albury $271, Corowa $356, Cootamundra $396, Hay $384 and Gundagai $442) are likely to be supportive of fruit fly and control programs are more likely to fail. The FFEZ towns of Griffith ($413), Narrandera ($418) and Leeton ($447) seem to be in the middle.

Overview

Earlier researchers have reported the preference for urban areas by QFF. Fletcher (1974b) reported that 78% of flies were caught in urban and creek areas with only 3% in pasture and lightly wooded areas. Fletcher (1974a) and MacFarlane et al. (1987) both noted that higher trap catches along watercourses compared with the surrounding dry sclerophyll bushland. Greater numbers of QFF were found in suburbia (Raghu, 1998) as compared to their natural rainforest habitats.

This paper further supports previous findings showing that towns create favourable environments for QFF. How might eradication programs benefit from this information? While the concept of finding QFF favourable towns based on the cost of water seems easy, the answers are not always straightforward. Not all towns in the FFEZ and RRZ have the same criteria for fixing the water rates (Anon., 1999). Some towns have yet to install water meters. Other towns have a complex costing system and it is difficult to calculate the average cost of water.

Some of our assumptions may be incorrect. We assumed that industrial water use was the same per capita across the region. We assumed that water losses are uniform in each town; there are probably higher water losses from clay or porcelain pipes compared with newer connections using plastic pipes.

However, the principle of town “oasis’ and rural “desert” remains (Dominiak et al., 2000b, Mavi and Dominiak, 1999). Where the cost of water is easily determined, eradication programs should concentrate on towns with high water costs, and thus likely dry environments. Conversely, eradication programs in towns with a low water cost and predicted high landscape water use are more likely to be unsuccessful because high landscape irrigation assists fruit fly survival.

QFF are unlikely to survive in the unmodified rural environment. Sterile release programs should concentrate on urban areas (Dominiak et al., 2000a,b) and not waste released flies in the dry rural environment; sterile QFF have a much better chance of survival in the urban situation. Minimising the movement of potentially infested fruit (Dominiak et al, 1998, 2000c,d) from one town to another would also reduce the chance of QFF reinfestation after a local population has been eradicated. These findings also support early reports identifying the southern natural boundary of QFF – there were no towns with modified environments in southern NSW to assist survival and spread of QFF prior to European colonisation. However, the long-term survival of the FFEZ seems based on a sound climatic premise while the oasis and desert principle remains.

References

Anon. (1993). Proposal to establish a Fruit Fly Exclusion Zone in southeast Australia. Agriculture Victoria. 33 pp.

Anon. (1999). Comparative Information on New South Wales Local Government Councils 1997/98. NSW Department of Local Government. 186 pp.

Bateman, M.A. (1968). Determinants of Abundance in a population of the Queensland Fruit Fly. Symposia of the Royal Entomological Society of London: Number Four. 119-131.

Bateman, M.A. (1972). The ecology of fruit flies. Annual Revision of Entomology. 17, 493-518.

Dominiak, B.C., Rafferty, T.D. and Barchia, I. (1998). An analysis of travellers carrying fruit near Griffith, NSW, during Easter 1996 to assess the risk of Queensland fruit fly (Bactrocera tryoni Froggatt). General and Applied Entomology. 28, 13-19.

Dominiak, B.C., McLeod, L.J. and Cagnacci, M. (2000a). Review of suppression program using three ground release methods of sterile Queensland fruit fly Bactrocera tryoni (Froggatt) at Wagga Wagga, NSW, in 1996/97. General and Applied Entomology. 29, 49-57.

Dominiak, B.C., McLeod, L.J., Landon, R. and Nicol, H.I. (2000b). Development of a low-cost pupal release strategy for Sterile Insect Technique (SIT) with Queensland fruit fly and assessment of climatic constraints for SIT in rural New South Wales. Australian Journal of Experimental Agriculture. 40, 1021-1032.

Dominiak, B.C., Rafferty, T.D. and Barchia, I. (2000c). A survey of travellers carrying host fruit near of Queensland fruit fly (Bactrocera tryoni (Froggatt)), into a fruit fly free area in 1996/97. General and Applied Entomology 29, 39-44.

Dominiak, B.C, Campbell, M., Cameron, G. and Nicol, H.I. (2000b). Review of vehicle inspection historical data as a tool to monitor the entry of hosts of Queensland fruit fly Bactrocera tryoni (Froggatt) (Diptera: Tephritidae) into a fruit fly free area. Australian Journal of Experimental Agriculture. 40, 763-771.

Donnelly, J.R., Moore, A.D. and Freer, M. (1997). GrazPlan: Decision Support System for australian grazing enterprises. 1. Overview of GrazPlan project and a description of MetAccess and LambAlive DSS. Agricurtural Systems. 54, 57-76.

Drew, R.A.I, Zalucki,M.P. and Hooper, G.H.S. (1984). Ecologoical studies of eastern Australian fruit flies (Diptera: Tephritidae) in their endemic habitat. I. Temporal variation in abundance. Oecologia. 64, 267-272.

Fletcher, B.S. (1974a). The ecology of a natural population of the Queensland fruit fly, Dacus tryoni V. The dispersal of adults. Australian Journal of Zoology. 22, 189-202.

Fletcher, B.S. (1974b). The ecology of a natural population of the Queensland fruit fly, Dacus tryoni. V. Seasonal changes in fruit fly numbers in the areas surrounding the orchard. Australian Journal of Zoology. 22, 353-363.

Fletcher, B.S. (1987). The biology of Dacine fruit flies. Annual Review of Entomology. 32, 115-144.

Froggatt. W.W. (1909). Report on parasitic and injurious insects. Department of Agriculture, Sydney. 117 pp.

Genstat 5 Committee (1998) Genstat for Windows, 4th Edition Numerical Algorithms Group Ltd., Oxford UK

MacFarlane, J.R., East, R.W., Drew, R.A.I. and Betlinski, G.A. (1987). Dispersal of irradiated Queensland fruit flies, Dacus tryoni (Froggatt) (Diptera:Tephritidae), in south-eastern Australia. Australian Journal of Zoology. 35, 275-281

Mavi, H.S. and Crichton, J. (1998). Trends in the incidence of drought exceptional circumstances in NSW. 12th ANZ Climate Forum. Perth. Western Australia.

Mavi, H.S. and Dominiak, B.C. (1999). Influence of urban climate and hydrology on the growth potential of Queensland fruit fly in south-eastern New South Wales, Australia. Proceedings of the 15th International Congress of Biometeorology and International Conference on Urban Climatology. November 8-12, 1999. Sydney, Australia. ICUC7.3.

Raghu, S. (1998). Geographical distribution, seasonal abundance and habitat preference of fruit fly species (Diptera: Tephritidae) in South-East Queensland, with special reference to Bactrocera tryoni (Froggatt) and Bactrocera neohumeralis (Hardy). Master's Thesis. Griffith University. Queensland.

Yonow T & Sutherst RW. (1998). The geographical distribution of the Queensland fruit fly, Bactrocera (Dacus) tryoni, in relation to climate. Australian Journal of Agricutural Research. 49, 935-53.

Previous PageTop Of Page