a review of flooding in macuata province, fiji islands

A REVIEW OF FLOODING IN MACUATA PROVINCE, FIJI ISLANDS April 2000 Stephen Yeo* Department of Physical Geography Macquarie University

February 2001

SOPAC Technical Report 328

*Author’s contact details in 2001 Dr Stephen Yeo Department of Physical Geography Division of Environmental and Life Sciences Macquarie University, Sydney, NSW, 2109 AUSTRALIA Facsimile: + 61 2 9850 8420 Telephone: + 61 2 9850 9670 E-mail: [email protected]

ABSTRACT Flooding of the Wainikoro and Bucaisau River valleys (Vanua Levu, Fiji) in April 2000 probably reached the highest levels since at least 1950. The one-day rainfall at a site near Labasa was the highest in 37 years of record. Damages were not exceptional by urban standards – only about F$3 million in financial losses. However, a loss of even $1000 of household contents or animals can represent the loss of many years’ savings, and possibly, livelihoods. A number of options for reducing future flood losses are explored: from ensuring the adequate capacity of floodgates, to improving informal flood warning systems. Better floodplain mapping and recording of flood histories is needed as a solid foundation for floodplain management. Efforts to educate flood-liable communities should be promoted, because in the short-term at least, the isolated rural communities north-east of Labasa will need to respond to flood threats with a degree of selfreliance.

SOPAC Technical Report 328 : Yeo – ii

ACKNOWLEDGEMENTS Fieldwork/travel expenses were met by SOPAC. I am very grateful to Alf Simpson and Russell Howorth for the opportunity of further exploring the flood hazards of Fiji. An enormous thankyou to those who smoothed the trip to Labasa – Commissioner Northern (J. Ratuvuku), the District Officer (Suren) and the expert driver/interpreter (Qio). Thanks to all those who shared their experiences of the flood, including the General Manager of FSC’s Labasa Mill (A. Radrodro) and the Town Clerk (S. Nand). Virend Chand Prasad (Hydrologist, Divisional Engineer Northern) has been immensely helpful in tracking down and making available hydrological data. Thanks to Ashmita Gosai (Fiji Meteorological Service) for providing climate summaries. K.M.Cho (Principal Engineer, Land and Water Resources Management Division) calculated catchment areas and slopes. Atish (Koronivia Research Station) supplied a map of land use. Tale, Koli and Susie (Department of Town and Country Planning) provided a copy of the Labasa Town Planning Scheme. Thanks again to the CSR Company for access to archives at ANU in Canberra. Frank Siciliano (Macquarie University) provided a great deal of assistance in preparing the maps.

SOPAC Technical Report 328 : Yeo – iii

TABLE OF CONTENTS ABSTRACT

ii

ACKNOWLEDGEMENTS

iii

1. INTRODUCTION

1

2. MACUATA PROVINCE

1

2.1 Physical environment 2.2 Social environment

1 3

3. RAINFALL AND FLOODING

4

3.1 April 2000 3.2 Frequency

4 9

4. APRIL 2000 FLOOD DAMAGES 4.1 4.2 4.3 4.4 4.5

Fatalities Government infrastructure Agriculture Industrial sector Commercial sector

4.6 Residential sector 4.7 Summary

14 14 14 15 16 17 18 20

5. FLOOD DAMAGE REDUCTION STRATEGIES 5.1 Flood management 5.2 Floodplain management

20 20 22

6. CONCLUSION

25

REFERENCES

27

APPENDICES 1.1 Macuata Flood Damage Survey Form, 14/15 April 2000

28

2.1 Appendix 2.1: Population and housing data for Fiji, Macuata and Bua, 1996 3.1 Highest one-day rainfall by water year (July-June), Divisional Engineer Northern (DEN) Yard

29 31

SOPAC Technical Report 328 : Yeo – iv

List of Figures 2.1 River catchments in the flood-affected area, Macuata Province 3.1 Average monthly rainfall totals, Labasa Airfield, with 1999/2000 monthly rainfall and 1999/2000 Southern Oscillation Index

2 5

3.2 3.3 3.4 3.5

Cyclone tracks for Macuata floods Four-day rainfall isohyets for flood-affected area, Macuata Cumulative rainfall, Macuata, 14-15 April 2000 Measured flood levels, Macuata rivers

6 7 8 11

3.6 3.7 4.1 4.2 4.3

Flood rugby, Labasa, 15 April 2000 Resident indicating flood height, Daku, Wainikoro Area of crops damaged in flooding, according to Tikina Cost of lost production due to crop damages, according to Tikina Damaged tramline, near Daku, Wainikoro

11 14 15 16 17

4.4 Destroyed house, Nakelikoso village 4.5 Damaged house, Nakelikoso village 5.1 Floodgate through seawall, Korovatu

19 19 22

List of Tables 2.1 Catchment characteristics for some river basins 3.1 Daily rainfall, Macuata, 13-16 April 2000 (Cyclone Neil) 3.2 Times of flood rise and peak, 14-15 April 2000 3.3 3.4 3.5 4.1 4.2

Table 3.3: Historic rainfalls, Labasa District Oral evidence pertaining to flood levels, Macuata rivers Historic flood records, Macuata rivers Average household damages estimated by four advisory councillors Summary of financial losses

Abbreviations CSR

Colonial Sugar Refining Company

DEN FMS FSC LWRMD NLTB

Divisional Engineer Northern Fiji Meteorological Service Fiji Sugar Corporation Land and Water Resources Management Division Native Lands Trust Board

PWD SOI

Public Works Department Southern Oscillation Index

SOPAC Technical Report 328 : Yeo – v

3 5 9 10 12 13 18 20

1. INTRODUCTION The passage of a trough associated with Tropical Cyclone Neil generated heavy rainfall and severe flooding over much of Macuata Province, Vanua Levu, Fiji, on Friday 14th and Saturday 15th April 2000. Compared to the heavy financial losses associated with flooding of the urban centres of Ba and Nadi in January 1999 (Yeo, 2000), overall damages in the rural areas of Macuata were relatively light. Nevertheless, for several farming communities north of Labasa, the flood had devastating impacts. Damage and needs assessments are routinely conducted after cyclone/flood disasters in Fiji. Alltoo-often, the lessons for long-term disaster reduction are forgotten in the haste to return life to normal. This report takes a long-term view, both in endeavouring to understand the character and frequency of such phenomena, and in the development of strategies to manage the flood risk. Reports on the Macuata flooding were collected from Government departments and newspapers. Interviews with affected households or representatives were conducted over three days in May 2000 (see Appendix 1.1 for survey form), including 12 residents (some of whom were also shopkeepers), four advisory councillors (representatives of residents), the general manager of the FSC Mill at Labasa, and the Town Clerk of Labasa. Some research was also undertaken at the CSR Archives1, held at the Australian National University in Canberra. The first section of the report outlines some key features of the physical and social environment in Macuata Province. Following this is an analysis of the pattern and frequency of the rainfall and flooding event of April 2000. The damages sustained in the flooding are reviewed in the third section, and suggestions for the reduction of future damages are described in the fourth section.

2. MACUATA PROVINCE 2.1

Physical environment

The drainage pattern of Macuata Province is characterised by a number of short, steep, subparallel river basins originating on Vanua Levu’s central spine to the east (Figure 2.1). Catchment areas, lengths, altitudes and slopes are available for only three of these river basins (Table 2.1). The nature of these rivers is conducive to rapid-rising floods following heavy rain. Another

The Colonial Sugar Refining (CSR) Company operated Labasa Mill from 1894 to its withdrawal from Fiji in 1973. Correspondence from the Mill to the Head Office in Sydney conveys reports of flooding. 1

SOPAC Technical Report 328 : Yeo – 1

Figure 2.1: River catchments in the flood-affected area, Macuata Province


important characteristic is the hydrological independence of these river valleys. Investigation of both the CSR Archives and the contemporary flood record shows several occasions where a major flood in one catchment was matched by only minor flooding in an adjacent catchment. For example, in January 1950 it was reported that ‘flooding was much more severe in the Qawa than the Labasa River’ (CSR, 1950a). This variability over short distances reflects the high spatial variability of rainfall in Fiji (Liedtke, 1989) and the comparatively small size of the catchments in view – which can therefore easily ‘miss’ the most intense rain. Table 2.1: Catchment characteristics for some river basins Source: River Engineering Section, Land and Water Resource Management Division 2

River Valley

Area (km )

Length (km)

Altitude (m)

Average slope

Qawa R.

126

33.5

450

1 in 13

Labasa R.

170

49.5

910

1 in 18

Wailevu

98

39.5

700

1 in 18

The underlying rocks are primarily volcanic. A geological map indicates volcanic centres situated near the lower Wainikoro and Bucaisau basins, and the Labasa River appears to drain a major intervolcanic basin. The predominant land uses in Macuata Province are sugarcane and rice farming. Many farmers are engaged in mixed farming. Formal forestry appears to be restricted to the Koroutari Forestry plantation situated in the upper Labasa and Wailevu catchments. 2.2

Social environment

At the 1996 Census (see Appendix 2.1), Macuata Province had a population of 80 207 people, consisting of 22 363 Fijians (28%), 56 294 Indians (70%) and 1550 others (2%). The ratio of Fijian to Indian residents (0.40) is significantly below the average for Fiji as a whole (1.16), and well below the ratio for Bua Province in the south-west of Vanua Levu (3.28). About 24 000 (30%) of the residents of Macuata Province live in and near Labasa Town, the administrative centre for the Northern Division. Nevertheless, the proportion of rural dwellers for both Fijians and Indians is higher than the national average. The census also shows that the proportion of residents engaged in paid employment and in subsistence agriculture matches the profile for Fiji (Bua, in contrast, shows a substantial divergence). The primary difference in the material used to construct the outer walls of houses is a lesser proportion of concrete/brick walls and a greater proportion of wooden walls for Macuata relative to the national average. In general, residents of Macuata are less likely to have access to


household goods like cars, refrigerators, televisions, washing machines and telephones than most (but have many more financial resources than the people of Bua). Data on main water supply indicate an extensive reliance on groundwater in Macuata. Most residents of Macuata have access to electricity (compared to only 1 in 5 households from the less developed province of Bua). Land tenure seems to reflect ethnicity. Three-quarters of households in Macuata leased their land from the State or the Native Lands Trust Board (NLTB), particularly the latter. Only 7% of households owned freehold land.

3. RAINFALL AND FLOODING 3.1

April 2000

Rainfall In association with positive values of the SOI, substantially more rain (> 600 mm) fell at Labasa in the eleven months prior to April 2000 compared to the long-term average (Figure 3.1). It is likely that soil moisture levels were near saturation even before the rains of April 2000. The importance of soil moisture levels for the generation of runoff from rainfall was previously demonstrated in a hydrological investigation of the Vanua Levu floods associated with Cyclones Martin and Raja in 1986 (Raj, 1987). A preliminary track of Tropical Cyclone Neil is shown in Figure 3.2. The heavy rain over Vanua Levu, to the right-hand side of the developing system, was the result of an associated trough (FMS, 2000). Figure 3.2 shows cyclone tracks for a number of other events known to have generated flooding in Macuata. That these cyclones came from a variety of directions demonstrates the difficulty in associating cyclone track with flood potential. A notable feature of the cyclones of 1950 and 1986 (Raja) was their slow movement as they reached turning points, allowing prolonged heavy rain (Kerr, 1976, p.76; Krishna, 1987).


12

600

10

Rain (mm)

500

SOI (5-month running mean)

700

8

400 6 300 4

200

2

100 0

0 May

Jun

Jul

Aug

Sep

1956-1999

Oct

Nov

Dec

1999/2000

Jan

Feb

Mar

Apr

SOI (1999/2000)

Figure 3.1: Average monthly rainfall totals, Labasa Airfield, with 1999/2000 monthly rainfall and 1999/2000 Southern Oscillation Index Source of data: FMS

Daily and accumulated four-day rainfall totals for the April 2000 flood are listed for sites across the affected region in Table 3.1. The highest one-day reading was 422 mm at Naseva gauge, situated in the catchment of the Wainikoro river. (It is quite possible that heavier rain fell at higher, ungauged sites). Rainfall totals decrease from the north-east to south-west, as depicted in Figure 3.3. Table 3.1: Daily rainfall, Macuata, 13-16 April 2000 (Cyclone Neil) Sorted by catchment, north-east to south-west Station

Catchment (watershed)

Naseva

th

th

th

th

13 April 2000

14 April 2000

15 April 2000

16 April 2000

Four-day total

Source

Wainikoro

26

422

85

4

537

PWD

Wainikoro

Wainikoro

5

Underwater

-

FSC

Bucaisau

Bucaisau

14

432

446

FSC

DEN Yard

Qawa

49

7

438

PWD

Mill

Qawa

50

4

372

FSC

Vunimoli

Labasa

44

230

102

31

408

FSC

AES (Korowiri)

Labasa

59

193

73

8

333

FSC

Nasarava

Wailevu

52

110

157

28

347

PWD

Waiqele

Wailevu

70

366

FSC

Labasa Airfield

Wailevu

47

312

FMS

Wailevu

Wailevu

49

339

FSC

Gata/Nasealevu

Dreketi

41

233

PWD

Seaqaqa

Dreketi

218

FSC

317

65 318

297 174

75

16

290 76

97 218


19

Figure 3.2: Cyclone tracks for Macuata floods Sources: Visher, 1925, p.77 (1912); Kerr, 1976, p.37 (1950 Feb, 1952 Jan); FMS (1986 Martin, 1986 Raja, 1997 Gavin, 2000 Neil). Note: Insufficient information was available to construct a track for the December 1929 cyclone (and no published track was found)


Figure 3.3: Four-day rainfall isohyets for flood-affected area, Macuata Sources of data: Divisional Engineer Northern, Hydrology; FSC, Labasa Mill; FMS, Nadi Airport. Note: Locations of FSC stations are imprecise


The highest hourly rainfall 2 was also recorded at Naseva gauge – 84 mm between 11 p.m. and midnight on Friday 14 April (Figure 3.4). This heavy rain appears to explain the earlier flooding of the Lagalaga and Wainikoro Rivers (Table 3.2), since catchments there would be expected to reach saturation and contribute runoff more rapidly than those experiencing less-intense rain.

Cumulative rain (mm)

600 500 400 300 200 100

1600-1700

1400-1500

1200-1300

1000-1100

0800-0900

0600-0700

0400-0500

0200-0300

0000-0100

2200-2300

2000-2100

1800-1900

1600-1700

1400-1500

1200-1300

0

14-15 April 2000 Naseva

DEN Yard (Labasa)

Nasarava

Gata/Nasealevu

Figure 3.4: Cumulative rainfall, Macuata, 14-15 April 2000 Source of data: Divisional Engineer Northern, Hydrology

Flooding Differences between the stated times of the floods’ rise and peak (Table 3.2) suggest that the rates of rise were not as fast as has occurred during other floods in Fiji. At one house in Coqeloa (Bucaisau catchment), it apparently took five hours for the water to rise about 1.5 m (five feet). The floods for rivers situated north-east of Labasa would still be classified as ‘flash floods’, however. Another respondent, at Korovatu (Bucaisau catchment) claimed that the water entered the house at 3.30 a.m., and rose to a height of 1.5 m in just half an hour. The duration of flooding was typically short – most areas were reported to be clear of water less than 24 hours after the peak.

2

Hourly rainfall was recorded at only four sites.


Table 3.2: Times of flood rise and peak, 14-15 April 2000 Note: Data for Lagalaga, Wainikoro and Bucaisau Rivers are median times based on several interviews River Valley

Time of rise

Lagalaga R.

2200 (Fri)

Wainikoro

2230 (Fri)

Bucaisau R.

0230 (Sat)

Qawa R.

0445 (Sat)

Labasa R. Wailevu R.

0000 (Sat)

Time of peak

Source Interviews

0300 (Sat)

Interviews Interviews

0900 (Sat)

FSC interview

1630 (Sat)

LWRMD gauge (Civic Centre)

1400 (Sat)

PWD gauge (Nakama)

Flood levels can be exacerbated in the lower reaches of rivers by high tides and/or storm surge. The Labasa River is tidal for at least 3 km upstream of the bridge in Labasa Town (Raj, 1987, p.11), and Labasa Town is thought to be situated only 1 m above the high-water mark (Sharma Nand, Town Clerk, pers. comm., 5 May 2000). North-westerly (on-shore) winds are believed to have backed up floodwaters at Labasa in the 1912 and 1929 hurricanes 3. Storm surges are known to have affected coasts north-east of Labasa during Cyclones Martin, Raja and Gavin (see Figure 3.2 for tracks). During Cyclone Gavin, storm surge breached sea-walls in ten places (Raj, 1998; Terry and Raj, 1999). There were no reports of a significant storm surge along the Macuata coastline during Cyclone Neil, but there were a few sketchy references to the influence of the high tide. Although mainly rainfall driven, the rise of rivers in the early hours was coincident with a rising tide. The late (and minor) flood peak for the Labasa River at the Civic Centre gauge (Table 3.2) appeared to be tidal driven 4.

3.2

Frequency

The local Member of Parliament reported that the flood was ‘unprecedented in its speed, intensity and damage done’. It is, however, not uncommon for particularly damaging events to be described as unprecedented when in fact historical records portray floods of similar or even higher levels (Yeo, 1999).

Rainfall Using Raj’s (1987) estimates of maximum 1-day rainfall frequency for the Divisional Engineer Northern’s (DEN) Yard, the total of 317 mm recorded on 14 April 2000 would have an average recurrence of between 50 and 100 years. However, caution must be exercised when extrapolating In 1912, the manager of the CSR Mill at Labasa reported that, ‘The rivers flooded and the hurricane from the West caused the biggest flood ever seen here’ (CSR, 1912). The Commissioner at Labasa reported that in 1929, a vessel was ‘held in stays by the opposing wind [i.e. westerly wind] and flood (Burrows, 1952, p.189). 4 Tides for Suva (45 minutes ahead of Malau, near Labasa) were as follows: 14/04 0255 1.5m, 0857 0.4m, 1511 1.6m, 2134 0.3m, 15/04 0351 1.5m, 0957 0.4m, 1608 1.6m, 2228 0.3m (Fiji Nautical Almanac, 2000). 3


from a record of only 27 years (Raj, 1987). What is clear from the assembling of annual maxima for water seasons from 1963/64 to 1999/2000 is that the daily rain at DEN Yard on 14th April 2000 was the highest in 37 years of record (Appendix 3.1), or 41 years when Raj’s (1987) data are included. Though heavy, the 422 mm recorded at Naseva on 14th April 2000 was less than the extreme oneday falls for the Fiji Group reviewed in Wither (1980)5. An incomplete list of historic rainfalls in the Labasa district is shown in Table 3.3. The highest fall in the area of interest occurred on 25-26 February 1950, when ‘during the afternoon and night of the storm 24.39 inches [620 mm] of rain were recorded’ (CSR, 1950b). Even this limited historical perspective demonstrates that the recent fall was not unprecedented. Table 3.3: Historic rainfalls, Labasa District Date

Place

1950 Jan 2 1950 Feb 25

1986 Dec 29

Rainfall (mm)

Duration

Source

Labasa

257 mm

24 hour

Wainikoro

620 mm

1929)

Record flood at time (>1929)

‚ 0.56 m

1950 Mar 30

Major

Major

Moderate (Mill not flooded)

1952 Jan 24?

Major

Major

Minor? (no mention)

ƒ Major

1986 Dec 30 (C. Raja)

Minor?

Minor?

Moderate (Mill not flooded)

‚ Worst since

2000 Apr 15 (C. Neil)

Major

Major (Mill flooded)

Minor

1950 Feb 26

Major

below 1929

Major

1929 Moderate

The flooding of April 2000 was clearly a rare event on the Lagalaga, Wainikoro, Bucaisau and Qawa Rivers (Figure 3.7 illustrates the height at one site). Records are insufficient to determine whether the flood was higher than the 1950 flood on the Wainikoro and Bucaisau Rivers. Though probably the highest event in 50 years at Labasa Mill (on the Qawa River), it was not the record event there. For the Labasa River, the flood was a reminder of the potential for much more serious inundation.

The mention of flooding at Labasa in 1952 by a 76-year old after the Raja flood (Fiji Times, Wednesday 31 December 1986, pp.10-11), by Blong (1994) (who may have used the Fiji Times article), and by two respondents suggests that there may indeed have been a flood at Labasa in 1952. Moreover, Kerr (1976, p.36) tracked a tropical depression moving from the north-west to south-east, crossing the coast of Vanua Levu at about noon on 24 January (different from the devastating hurricane of 28 January). 8


April 2000

Figure 3.7: Resident indicating flood height, Daku, Wainikoro (taken 4 May 2000)

While the record of floods shown here is not complete (e.g. Blong, 1994, lists five other (minor) Labasa floods from the 1980s), it is thought to be a true reflection of the frequency of major flooding in the region. Relative to Ba Town, flooding at a given site in Macuata is not common. This probably reflects differences in catchment area and configuration – the Ba Valley is larger and is rather round, capturing runoff from many different quarters, whereas apart from the Dreketi, the Macuata river valleys are small and often elongated, requiring ‘direct hits’ for serious flooding.

4. APRIL 2000 FLOOD DAMAGES 4.1

Fatalities

One Fijian man in his mid-30s drowned as he attempted to cross the river near Nayarabale village (Bucaisau catchment). In addition, there were a number of ‘near-misses’, as families living near the Wainikoro and Bucaisau Rivers were forced to climb into ceilings and even onto the external roof as the waters rose, before being rescued by boat. 4.2

Government infrastructure

A significant component of the tangible damages was to infrastructure. The Divisional Engineer Northern estimated a cost of $883 000 to restore infrastructure to its original condition, consisting


of $280 000 for Water Supply ($195 000 for pipework and $85 000 for road access) and $603 000 for Roads. Interestingly, only $120 000 of the latter was for the Roads East area, where flooding was most severe. This suggests that, following months of above-average rains, the rain of April 1415 experienced elsewhere in Vanua Levu was still sufficient to cause much road damage. 4.3

Agriculture

Damages to agriculture were also significant. Figure 4.1 indicates the area damaged for different crop types according to Tikina (administrative units within a Province), and Figure 4.2 shows the corresponding estimated cost of that damage. Government surveys suggest that a total of 121 hectares of root crops and vegetables was damaged, at a total cost of $752 128. The most affected Tikinas were Nadogo (encompassing the Nubu, Lagalala and Wainikoro River Valleys) and Labasa (encompassing the Bucaisau, Qawa and Labasa River Valleys). Areas of rice farming were extensively affected, particularly in Nadogo Tikina – 26 farmers from Taganikula alone reported the loss of crops that were ready for harvesting. Cassava was also widely damaged by the event. Although little affected in terms of area, yaqona (kava) formed a major component of total dollar losses, due to its high value per hectare. For example, the cost from damaged yaqona in Dogotuki Tikina reached almost $100 000. These costs are ultimately borne by the farmer, although from time to time the Government assists in rehabilitation by providing cuttings for replanting.

90 80

Area damaged (ha)

70

Vegetables 60

Rice

50

Yaqona

40

Yams Dalo

30

Cassava

20 10 0 Dogotuki

Namuka

Nadogo

Labasa

Wailevu

Dreketi

Tikina

Figure 4.1: Area of crops damaged in flooding, according to Tikina Source of data: Principal Agricultural Officer Northern. Note: Tikina sorted north-east to south-west


$450,000

Cost of lost production

$400,000 $350,000

Vegetables

$300,000

Rice Yaqona

$250,000

Yams

$200,000

Dalo

$150,000

Cassava

$100,000 $50,000 $0 Dogotuki

Namuka

Nadogo

Labasa

Wailevu

Dreketi

Tikina

Figure 4.2: Cost of lost production due to crop damages, according to Tikina Source of data: Principal Agricultural Officer Northern. Note: Tikina sorted north-east to south-west

Government surveys also recorded dead and missing animals for Nadogo and Labasa Tikinas. In all, 84 cattle (at $400/head), 7 horses (at $300/head), 290 goats (at $50/head) and 471 poultry (at $6/head) were counted, yielding a damage bill of $53 026.

FSC reported that 5% of the sugarcane crop had ‘lodged’ (fallen over) in the Wainikoro and Bucaisau areas, but anticipated full recovery. Waterlogging would not be expected for such a short-duration flood. A farmer at Naleba near the Bucaisau River did describe a heavy deposit of silt that he suggested would reduce the cane harvest. Excessive rain and flooding late in the growing season can also promote growth at a time when the cane should be maturing, resulting in decreased sugar content. (See Yeo, 1998, pp.76-77, for a discussion of the effects of flooding on sugarcane in the Ba Valley). 4.4

Industrial sector

Damages sustained by FSC were estimated at $259 391, consisting of $184 042 damages to the Factory, $51 600 sustained by the Civil Department (including $40 000 to tramlines – see Figure 4.3), $22 360 to the Field Department and $1389 to Stores. Relative to typical flood damages sustained at Rarawai Mill in Ba (Yeo, 2000, p.7), a total damage bill of $259 391 sounds remarkably low. This probably reflects lesser flood depths and silt deposits at Labasa – ‘as if someone had come through with a bucket and cleaned it’ (Alipate Radrodro, GM Labasa Mill, pers. comm., 5 May 2000).


Figure 4.3: Damaged tramline, near Daku, Wainikoro (taken 4 May 2000) The timing of the flood only a month before the scheduled start of crushing caused problems – necessitating hard work to catch up on routine maintenance. 4.5

Commercial sector

A Cabinet Paper reported that six businesses were flooded in the Navoalevu and Taganikula areas of the Wainikoro Valley. Two of those businesses each reported damages of $250 000, though it is difficult to verify the accuracy of these estimates. Items damaged at one (in Navoalevu) included shop stock ($80 000), various machines and transport, generators, office equipment, stored rice (4 tonnes), stored timber, fertiliser (200 bags) and several beehives ($1600). Two other shops reported losses of $20 000 and $10 000. In contrast to the December 1929 or December 1986 floods, Labasa Town was little affected by floodwater in April 2000. The Town Clerk reported that the Fish Market and public toilets were inundated, but the cost was minimal because the flood came and went on the Saturday, and the Council routinely cleans the facilities every Sunday. An estimate of total losses sustained by the commercial sector is $560 000.


4.6

Residential sector

Early newspaper reports indicated a figure of 537 people in urgent need of food rations (Fiji Times, Wednesday 19 April 2000, p.1). Government reports recorded a total of 1202 flood victims from 221 households, defined as those whose houses had actually been submerged. Four advisory councillors provided estimates of the average damage per house for their areas of responsibility (Table 4.1). The advisory councillor for Coqeloa (in the Bucaisau catchment) differentiated between cane-farmers who may have lost an average of $10 000, and casual labourers who may have lost an average of between $1000 and $1500. In the knowledge that some houses in the Naodamu Housing area of Labasa were inundated by only a few centimetres, it seems reasonable to assume an average loss for all houses of $1500, which based on 221 households yields a total damage figure of $331 500. This coarse estimate excludes losses from the agricultural sector. Table 4.1: Average household damages estimated by four advisory councillors Settlement (and River Valley)

Number of houses flooded

Average damage per house

Total damage

Navoalevu (Wainikoro)

16

$1000

$16 000

Taganikula (Wainikoro)

10

$1000

$10 000

Coqeloa (Bucaisau)

17

10@$10000

$108 750

Naleba (Bucaisau)

16

7@$1000-$1500 $1500

$24 000

An early article reported that 15 homes had been ‘washed away’ (Fiji Times, Tuesday 18 April 2000, p.1), but a subsequent Government report of only two houses ‘washed away’ in the Wainikoro area suggests that the early report was exaggerated. Figure 4.4 shows one of these houses, believed to be from Nakelikoso village, adjacent to the Lagalaga River (Figure 2.1). A number of bamboo-walled houses and a meeting hall were also badly damaged at this village (Figure 4.5). A respondent from Daku (near the Wainikoro river) reported a sunken wooden floor. Interviews revealed that commonly damaged items included clothes, cooking utensils, floor mats and furniture (especially chipboard). The loss of school books and school uniforms represented significant losses for a number of families. Some respondents reported damaged mattresses, televisions and stereos. A substantial loss for the village of Nakelikoso was the damage to $15 000 worth of musical equipment and instruments that had been stored at the pastor’s house. Most respondents sustained damages to paddy rice stored in ‘bakhaars’9, and several saw bags of expensive fertiliser dissolve in the water. Some respondents reported damage to vehicles. Another effect of the flood was the contamination of drinking wells (Fiji Times, Saturday 29 April 2000, p.11) – a significant effect given that groundwater is the main water supply for 23% of households in Macuata (Appendix 2.1).


Figure 4.4: Destroyed house, Nakelikoso village Source: District Office, Labasa

Figure 4.5: Damaged house, Nakelikoso village (taken 4 May 2000) Emergency food supplies for a 14-day period were distributed to flood victims, at a cost of $8137.

9

A ‘bakhaar’ is a small air-tight structure built in the compound to keep paddy rice after harvesting.


4.7

Summary

The estimated financial losses associated with the flooding of Macuata Province in April 2000 are summarised in Table 4.2. Overall damages approached $2.85 million. Table 4.2: Summary of financial losses Sub-sector

Sector

Water Supply

$280 000

Roads

$603 000

Crops (excl. sugarcane)

$752 128

Animals

$53 026

Infrastructure (Government) Agriculture

Estimated costs $883 000 $805 154

Industrial (FSC)

$259 391

Commercial

$560 000

Residential

$331 500

Food ration (Government)

$8137

TOTAL

~$2.85 million

Broad financial statistics do not adequately capture the effects of the flood on an individual level. A loss of $1000 to a family supported only by casual labour can represent the savings of many years, and can take many years to recover. Other items may have little monetary value but great personal value, such as family photographs.

5. DAMAGE REDUCTION STRATEGIES Broadly speaking, strategies for flood damage reduction are two: flood management and floodplain management (Yeo, 2000: Figure 1.2). Strategies for flood management include dredging of rivers to allow the easier passage of floodwater. Floodplain management encompasses everything from land use and building regulations to warning systems and emergency-response planning. Existing and potential damage-reduction strategies for the affected areas are briefly discussed in this section. 5.1

Flood management

Dredging Dredging at the mouth of the Qawa River began in 1998 or 1999 (conflicting reports). The estuaries of the Labasa and Wailevu Rivers are scheduled for dredging, and the flood of April 2000 prompted some to call for dredging at the mouth of the Wainikoro river. By May 2000, 800 m of the planned length of 1300 m channel had been dredged, at a width of 100 m and depth of 5 m. The


manager of the FSC Mill suggested that the dredging had eased pollution problems in the lower Qawa River through enhanced tidal flushing. Nevertheless, a policy of flood damage reduction consisting largely of dredging has several limitations (Yeo, 1997; Yeo, 1998). The design flood for dredging activities in Fiji has been (only) the 1-in-10-year flood, though with limited hydrologic data, even the estimation of such must contain significant uncertainties. Dredging is expensive, short-term (given high sediment loads), and of limited effect in tidal areas. Its other effect – again apparent from conversations in Macuata – is to create a false sense of security, consequently increasing damage potential. Seawalls Seawalls have been constructed along much of the coastline in the flood-affected areas. They are intended to prevent saline intrusion into arable farmland during high tides (Raj, 1998). Damage to seawalls was reported for many of the events listed in Table 3.5. Seawalls were breached in 10 places as a result of the storm surge associated with Cyclone Gavin in 1997 (Terry and Raj, 1999). An interesting finding of the field survey was not the damage to seawalls from flooding or storm surge, but the damage to houses that might be (partly) attributed to seawalls. It seems likely that during major floods, Yanucari Creek 10 acts as a natural distributary channel for the Bucaisau River. Construction of a seawall near the mouth of Yanucari Creek has blocked the floodway. The floodgate constructed through the seawall in 1985 (Figure 5.1) was not of sufficient capacity during the flood of April 2000 – the farmer living upstream of the floodgate reported a ‘head’ of 1-2 m from the landward to the seaward sides of the floodgate. A backing-up of floodwater may have contributed to the over-ground depth of about 1.6 m at the farmer’s house. The capacity of the Korovatu floodgate, and others, should be checked and if necessary increased in view of the experience of April 2000.

10

GR700713 on the Labasa (Edition 1, 1989) 1:50000 Topographical Map


Figure 5.1: Floodgate through seawall, Korovatu (taken 3 May 2000)

5.2

Floodplain management

Floodplain development Damages from flooding occur because people live and work on floodplains. While flood heights at Korovatu may have been exacerbated by the backing-up of floodwater behind the seawall, several houses are situated on land that was reclaimed from tiri (mangrove), so flooding of this low-lying area should not surprise. However, relocation of these households is not feasible because the Indian occupants enjoy Crown leases, which have a greater security of tenure than the expiring Native leases. They claim to have ‘nowhere else to go to’. In contrast, despite 60 years at the same site, residents of Nakelikoso village were giving serious consideration to shifting the entire village to a flood-free site on mataqali land to the north. Relocations of populations due to flooding have precedent: the CSR Company shifted its labour lines to high ground at Batinikama sometime after flooding in 1938 (FSC, 1994, p.32). Labasa (then Nasea) had major stores in the 1890s, and a Post Office from 1892. Major development was carried out in 1922, and the business centre was proclaimed a Township in 1939 and a Town in 1972 (FSC, 1994, p.53). The Labasa Town Planning Scheme was revised in 1982, with a map indicating areas and depths for the 10-year flood (even though it was acknowledged


that ‘there are no comprehensive records of flooding in Labasa’) (Directorate of Town and Country Planning, 1981-82). According to the Scheme (Paragraph 7.22), future development proposals should be based on an assessment of flood frequency and depth, and development controls such as height restrictions should be imposed to avoid damages to life and property. Insufficient time has been spent at Labasa to assess the implementation of this provision. Certainly, the floodplain is under pressure from developers, including the low-lying Naiyaca Peninsula (Sharma Nand, pers. comm., 5 May 2000). One conspicuous need is for a more-rigorous collection of flood ‘intelligence’, to enable planners to make informed (not misinformed) development decisions. A detailed contour map (at 0.25 m resolution) of the floodplain at Labasa Town is required. Even though 14 years have elapsed, it would also be sensible to map the extent and depths of the flood associated with Cyclone Raja – undoubtedly higher than the ‘10-year flood’ presently attached to the Scheme. In the absence of detailed flood information, planners need to adopt a precautionary approach to development of the floodplain. A means of reducing agricultural losses would be to plant the most valuable crop – yaqona – above potential flood levels. House design Most house walls in Macuata are constructed from timber or corrugated iron; relatively few are made from concrete, which is more flood resistant (Appendix 2.1). Of the houses surveyed, one had a floor raised 3 feet (0.90 m) above the ground, while most were raised only 1 ft (0.30 m) or 2 ft (0.60 m) – and not necessarily as a response to floods. A casual inspection of Labasa’s shops indicated that few had raised floors (the Post Office being an exception) – in marked contrast to Ba (Yeo, 1998). Probably this is a reflection of flood frequency. A shortcoming of the recovery stage of a flood disaster is that affected households – certainly the poorer of those – tend to replace damaged furniture with whatever is affordable or donated. Too often, chipboard furniture is replaced by chipboard furniture, which will inevitably need replacement after the next flood. Governments and NGOs would do well to educate flood-liable residents about the need to break the ‘chipboard cycle’, and where possible, to provide funds to enable them to purchase superior materials. One shopkeeper told a familiar story of putting all goods on the counter only to find the counter tipped over. The securing of counters to the floor would reduce losses (Yeo, 2000: Figure 5.2).


Warning, evacuation and education Formal warning of the flooding of low-lying areas was first mentioned at about 5 a.m. on the 15th April (FMS, 2000), well after the onset of flooding in most river valleys (Table 3.3). Interviews with residents indicated that most saw the water rising, while some were awoken by neighbours11. At the FSC Mill, rain fell from about 9 p.m. on the 14th, but was thought to be ‘normal Labasa weather, so no one was really concerned’ (FSC, 2000). At 4.30 a.m. on the 15th there was ‘no indication that river water would rise’, but by 5 a.m. the flood had risen. The rapid rise of the water meant that even those aware of the rise were not able to save many possessions – goats were rescued at two houses, and a religious book and video recorder at another. This was true also for the FSC Mill, where it was concluded that ‘this was a flash flood … there would have been very little chance of removing/dismantling any plant or electrical motors to avoid the rapidly rising floodwaters … there was very little that anyone could have done under the circumstances’ (FSC, 2000). It is, however, difficult to imagine flooding of the affected rivers that is not ‘flashy’ in character. To be sure, an unsaturated catchment would allow a longer lag between rainfall and runoff. But the short and steep nature of the valleys, and propensity for intense rain, suggest that available warning times will always be a matter of hours. With present resources, it would be difficult for the FMS or PWD to issue flood warnings of any certainty. And the benefit-cost ratio for the development of a highly technical flood warning system would not be favourable for the affected rural areas. The best course of action may be to enhance the capacity of local communities to assess the likelihood of flooding. As recommended for Ba (Yeo, 2000), Police Posts could be furnished with rain gauges and Police trained to take hourly (or even half-hourly) readings during heavy rain. Warnings would be issued locally if pre-determined thresholds were exceeded. More important still is education of the at-risk communities. Research at Ba indicated that flood warnings are as much perceived as received. Despite a higher level of flooding (Table 3.5), damages at Labasa Mill in February 1950 were less than those in January 1950 because of more warning and better preparation (CSR, 1950b). Had FSC been more aware of the flood risk, it is likely that they would have perceived the danger a little earlier. In response to the April 2000 flood, Labasa Mill has now prepared a Flood Contingency Plan to sit alongside the existing Cyclone Contingency Plan (FSC, 2000). The training of Heads of Departments should be scheduled near the onset on each cyclone season, to maintain a state of readiness.

The Town Clerk warned residents of Naodamu Housing area in Labasa, where fortunately the flood peak occurred during the day (Ta ble 3.3) (Sharma Nand, pers. comm., 5 May 2000). 11


Education is also needed at a national level to remind people of the danger of water. Evidence from Ba (Yeo, 2000) suggests that a majority of recent flood fatalities have been associated with dangerous actions, such as attempts to cross flooded watercourses. Evidence also suggests that Fijian residents are as susceptible to drowning in this manner as Indian residents, contrary to longheld notions that nothing could trouble a Fijian in water. The Fiji Red Cross Society’s efforts to promote water safety should be encouraged, with a focus on education of adults as well as of children. Education is also needed to bring the past to the present. National newspapers should be encouraged to include a historical column, which could focus regularly on a disaster. Memories are short. The flooding associated with Cyclone Neil was certainly rare for river valleys north-east of Labasa, but not without historical precedent. A corollary of this recommendation is that greater efforts are needed to document and understand contemporary disasters, as a basis for future education and planning. It is hoped that this report contributes to that end.

6. CONCLUSION The total tangible loss sustained in the flooding associated with Cyclone Neil in April 2000 was less than F$3 million. This was significantly less than the damage caused by flooding of Ba and Nadi in January 1999, and reflects the predominately rural nature of the April 2000 flood. However, a raw damage figure fails to convey the very substantial losses incurred by individual families. Heavy damages were sustained by infrastructure and agriculture, for which measures to mitigate losses may be limited to minor, site-specific adjustments in the construction and placement of roads and water intakes, and in the placement of valuable crops like yaqona. There are more extensive possibilities for the mitigation of industrial, commercial and residential losses, though some Indian residents appear to be tied to their house-sites by the paucity of available land of suitable tenure. A reassessment of the capacity of floodgates is required. Future development of the Labasa floodplain should be tightly regulated. Consideration should be given to means of breaking the ‘chipboard cycle’. There is some scope for improving flood warning systems. Business managers in Labasa should be encouraged to learn from the ‘near-miss’, and follow the example of Labasa Mill in preparing a flood response plan. Measures to maintain community flood readiness should be instituted. A greater monitoring and analysis of each flood event (hydrology and damage) is prerequisite for more effective planning and education.


Heavy rain was recorded on 14-15 April – 422 mm at Naseva gauge over 24 hours, including 84 mm in one hour. The daily rainfall at DEN Yard was the highest in at least 37 years of record, estimated at less frequent than the 1-in-50-year event. The ensuing flood was the highest known to most res pondents in the Lagalaga, Wainikoro and Bucaisau Valleys. However, extraction of data from the CSR Archives suggests that it was at least rivaled by a flood in 1950. The flood of April 2000 should certainly not be regarded as a ‘freak of nature’. It will occur again.


REFERENCES Blong, R.J. 1994. Natural Perils and Integrated Risk Assessment in Fiji, Blong and Associates, Denistone. Bureau of Statistics 1998. 1996 Fiji Census of Population and Housing: General Tables, (Parliament of Fiji, o Parliamentary Paper N 43 of 1998), Bureau of Statistics, Suva. Burrows, W. 1952. ‘Hurricanes in Fiji’, Corona, May, 186-189. CSR 1912. Head Office Correspondence, 142/2078, Labasa In, Letter Number 944A, 29/01/1912. CSR 1950a. Head Office Correspondence, M85 Reel 143, Labasa In, Letter Number 693, 06/01/1950. CSR 1950b. Head Office Correspondence, M85 Reel 143, Labasa In, Letter Number 708, 06/03/1950. Directorate of Town and Country Planning 1981-82. Labasa Town: Town Planning Scheme Revision 198182, Draft Report of Survey. FMS 2000. Tropical Cyclone Neil (15-17 April 2000): Preliminary Report, Fiji Meteorological Service, Nadi Airport. FSC 1994. Labasa Mill Centennial Anniversary Publication. FSC 2000. Internal Memorandum from General Manager Labasa Mill to Group General Manager Finance and Group General Manager Operations, 20 April 2000. Hasan, M.R. 1986. Fiji Watershed Management Study: Hydrology of Rewa and Ba Watersheds, FIJ/86/001 Field Document 4, UNDP/FAO, Suva. Kerr, I.S. 1976. Tropical Storms and Hurricanes in the Southwest Pacific: November 1939 to April 1969, New Zealand Meteorological Service Miscellaneous Publication 148, Wellington. Krishna, R. 1987. Tropical Cyclone Raja 22 Dec 1986 - 01 Jan 1987, Tropical Cyclone Report 87/1, Fiji Meteorological Service, Nadi Airport. Liedtke, H. 1989. ‘Soil erosion and soil removal in Fiji’, in Hohnholz, J.H. (ed.) Applied Geography and Development: A Biannual Collection of Recent German Contributions, Volume 33, Institute for Scientific Cooperation, pp.68-92. Raj, R. 1987. Cyclone Raja - Floods and Rainfall in Vanua Levu, PWD Hydrology, Suva. Raj, R. 1998. ‘Hydrological aspects of tropical cyclones’, WMO Bulletin, 47(4), 345-354. Terry, J.P. and Raj, R. 1999. ‘Island environments and landscape responses to 1997 tropical cyclones in Fiji’, Pacific Science, 53(3), 257-272. Visher, S.S. 1925. Tropical Cyclones of the Pacific, Bernice P. Bishop Museum Bulletin 20, Honolulu. Wither, B. 1980. ‘Fiji’s Easter present from ‘Wally’’, Soil and Water (NZ), 16(4), 19-21. Yeo, S.W. 1997. ‘Disasters, dredging and development: approaches to flood management in the Ba River Valley, Fiji’, in Planitz, A. and Chung, J. (eds.) Proceedings of the Disasters and Development Session, VIII Pacific Science Inter-Congress, 13-19 July 1997, The University of the South Pacific, Suva, UNDHA-SPPO, Suva, pp.84-103. Yeo, S.W. 1998. Natural and Human Controls on Flood Damages in the Ba River Valley, Fiji, unpublished PhD thesis, Natural Hazards Research Centre, School of Earth Sciences, Macquarie University, Sydney. st

Yeo, S.W. 1999. ‘A review of Australian flood disasters in 1998’, in Disaster Prevention for the 21 Century: Proceedings of the Australian Disaster Conference 1999, Emergency Management Australia, pp.287292. Yeo, S.W. 2000. Ba Community Flood Preparedness Project: Final Report, SOPAC Technical Report 309, SOPAC, Suva.


APPENDIX 1.1 Macuata Flood Damage Survey Form, 14/15 April 2000 BACKGROUND 1) Today’s date: __ / __ / __ 2) Your name: _________________________________________________ 3) House/shop address: ________________________________________ [Grid Ref: __________ ] 4) What is [was] the primary wall material in your house/shop? _____________________________ 5) How high is [was] the main floor level of your house/shop above ground level? ____ m, or ____ ft 6) Do you own or lease the affected property? Own [ ] Lease [ ] 7) How long have you been living at this address? Since ____, or ____ years 8) a) Have you experienced other floods (before April 2000) at this address? No [ ] Yes [ ] b) If yes, in what years? ______________________ c) How did they compare to the recent flood? ______________________ FLOOD DETAILS 9) At about what time of day or night did the recent flood enter your house/shop? 10) At about what rate did the water rise? ___________________________ 11) What was the maximum height of floodwater over the main floor? ____ m, or ____ ft 12) About when did the water fall below the floor of your house/shop? _____________ WARNING AND EVACUATION 13) a) Did you have any forewarning of the flood? No [ ] Yes [ ] b) If yes, how? Radio broadcast [ ] Contacted by family/friends [ ] Saw heavy rain [ ] Saw river rise [ ] 14) a) Did you lift or shift any property? No [ ] Yes [ ] b) If yes, were your actions effective in saving property? No [ ] Yes [ ] Partly [ ] 15) a) Did you or your family evacuate the house? No [ ] Yes [ ] b) If yes, where did you go? ___________________________ FLOOD DAMAGES 16) Please indicate whether the following items were damaged, detail any damage, and estimate the cost: Item

Damage?

Building structure

No [ ] Yes [ ]

Contents

No [ ] Yes [ ]

Animals

No [ ] Yes [ ]

Crops

No [ ] Yes [ ]

Other

No [ ] Yes [ ]

Details of any damage

Cost ($)

17) What is the total direct damage bill from the flood? $ _________ 18) How many days did it take to clean the house / restore trade or production? _________ days 19) Were you insured against flood? No [ ] Yes [ ] Partly [ ] COMMENTS 20) Do you have any comments about the causes and solutions to the flooding problem? ___________________________________________________________________________________ _____________________________________________________________________________ 21) Is there anything the Government can do better to manage the problem? ________________________________________________________________________________ 22) What lessons have you learned from the event? ________________________________________________________________________________ CONTACT DETAILS 23) May we contact you about this survey? No [ ] Yes [ ] Phone: ____________ Postal address: ____________


APPENDIX 2.1 Population and housing data for Fiji, Macuata and Bua, 1996 Source of data: Bureau of Statistics (1998) – pages listed with attribute

Fiji

Macuata Province

Bua Province

ETHNICITY (pp.87-88) and LOCALITY (pp.30-38) Fijian

393 575

51%

22 363

28%

10 992

73%

232 240 161 335

59% 41%

17 168 5195

77% 23%

10 473 519

95% 5%

338 818

44%

56 294

70%

3356

22%

170 783 168 035

50% 50%

37 818 18 476

67% 33%

3311 45

99% 1%

42 684

6%

1550

2%

640

4%

775 077

100%

80 207

100%

14 988

100%

415 582 359 495

54% 46%

55 718 24 489

69% 31%

14 396 592

96% 4%

ECONOMIC ACTIVITY (p.136) Employed only Subsistence only Employed and subsistence Unemployed Not in labour force

130 752 61 191 88 562 17 265 203 143

26% 12% 18% 3% 41%

12 204 5495 9859 1414 22 364

24% 11% 19% 3% 44%

193 4318 1618 378 2316

2% 49% 18% 4% 26%

Total aged 15+

500 913

100%

51 336

100%

8823

100%

48 760 29 002 50 342 5445 6975 2618 1475

34% 20% 35% 4% 5% 2% 1%

1578 6335 5126 628 1038 257 139

10% 42% 34% 4% 7% 2% 1%

54 670 1422 261 169 64 70

2% 25% 52% 10% 6% 2% 3%

144 617

100%

15 101

100%

2710

100%

24 027 9763 67 211 66 491 29 722 81 519 39 347 4215 2244

17% 7% 46% 46% 21% 56% 27% 3% 2%

2146 1613 5211 5377 1527 8450 2385 393 279

14% 11% 35% 36% 10% 56% 16% 3% 2%

15 74 116 154 32 706 30 150 33

1% 3% 4% 6% 1% 26% 1% 6% 1%

144 617

-

15 101

-

2710

-

Rural (% of total Fijian) Urban (% of total Fijian) Indian Rural (% of total Indian) Urban (% of total Indian) All others Total Rural Urban

OUTER WALLS (p.255) Concrete, brick Wood in good or sound condition Tin or corrugated iron Traditional bure material Wood in poor condition Makeshift or improvised materials Other materials Total households ITEMS AVAILABLE TO HOUSEHOLD (p.258)12 Car Carrier/truck Refrigerator Video/TV Washing machine Gas/electric stove Telephone Outboard motor Water pump Total households

12

Each household could nominate more than one item.


Fiji

Macuata Province

Bua Province

7620 2727 145 3502 632 475

279 1448 40 542 316 85

MAIN WATER SUPPLY (p.255)

Metered Communal standpipe Rooftank Well River or creek Other

86 684 29 944 3992 13 404 5887 4706

Total households

60% 21% 3% 9% 4% 3%

144 617 100%

50% 18% 1% 23% 4% 3%

15 101 100%

10% 53% 1% 20% 12% 3%

2710 100%

ELECTRICITY (p.258)

Yes No

96 756 47 861

Total households

67% 33%

144 617 100%

8590 6511

57% 43%

15 101 100%

545 2165

20% 80%

2710 100%

LAND TENURE (p.258)

Owned outright freehold Leased from State Leased from NLTB Occupied without legal arrangement, state or freehold land Occupied Native Land with informal arrangement Occupied through traditional village tenure Other Total households

26 594 27 849 37 722 5770

18% 19% 26% 4%

995 3168 8171 435

7% 21% 54% 3%

304 28 444 127

11% 1% 16% 5%

4338

3%

365

2%

80

3%

35 647 6697

25% 5%

1692 275

11% 2%

1641 86

61% 3%

144 617 100%


15 101 100%

2710 100%

APPENDIX 3.1: Highest one -day rainfall by water year (July-June) Divisional Engineer Northern (DEN) Yard Source: DEN, Hydrology

Year

Month

Day

Rainfall (mm)

1963/64

Mar

26

135

1964/65

Jan

13

130

1965/66

Mar

16

29

1966/67

Apr

9

145

1967/68

Jun

14

108

1968/69

Jan

4

39

1969/70

Nov

9

85

1970/71

Dec

18

186

1971/72

Dec

25

177

1972/73

Oct

23

182

1973/74

Feb

24

170

1974/75

Feb

9

116

1975/76

Jan

20

173

1976/77

Dec

27

50

1977/78

Nov

28

97

1978/79

Jan

10

269

1979/80

Apr

18

81

1980/81

Jan

28

165

1981/82

Jan

24

139

1982/83

Feb

25

72

1983/84

Mar

17

147

1984/85

Sep

17

128

1985/86

Feb

12

133

1986/87

Dec

29

258

1987/88

Nov

24

135

1988/89

Feb

9

116

1989/90

Mar

19

235

1990/91

Feb

19

75

1991/92

Dec

7

84

1992/93

Apr

11

105

1993/94

Feb

12

92

1994/95

Mar

13

247

1995/96

Feb

22

108

1996/97

May

3

140

1997/98

Jun

4

99

1998/99

Jan

17

153

1999/00

Apr

14

317
