A Preliminary Profi le of the Kavango Region in Namibia

by IRDNC Selma el Obeid and John Mendelsohn

for the Every River has its People Project

Rössing Foundation A Preliminary Profi le of the Kavango Region in Namibia

Contents

Introduction 1 The physical environment 5 People and social conditions 16 Livelihoods and environmental pressures 26 Acknowledgements, notes and sources 45

Selma el Obeid and John Mendelsohn © Research and Information Services of Namibia (RAISON) PO Box 80044 Windhoek, Namibia E-mail: [email protected]

Published by the Namibia Nature Foundation January 2001 A view of the Okavango Basin and the Cuito River, as photo- For information on the “Every River has its People Project” contact: graphed from a space shuttle in August 1991. The many fi res Namibia Nature Foundation burning in and around the region provide a background of PO Box 245, Windhoek, Namibia smoke. Courtesy of NASA. Tel +264 61 248345 Fax +264 61 248344 www.nnf.org.na E-mail: [email protected] Chapter 1: Introduction

A Preliminary Profi le of the Kavango Region in Namibia has been produced to provide background material for Equator the “Every River has its People” project1. The bulk of the report consists of maps and graphs derived largely from analyses of recent surveys, censuses and mapping projects. Brief comments are offered to help users " Luanda interpret the analyses and to highlight major issues and processes in the region, especially those relating to ANGOLA natural resources. References to notes and sources of information on page 45 are indicated by superscript numbers.

" " Bié Lobito Huambo " Lucusse The report has been organized into four chapters, and "

" the second provides information on the Okavango River " Caconda Cangamba in its entirety (Figure 1). Thereafter, almost all information concentrates on the Kavango Region within Namibia. "Lubango " Cuito Cuanavale Comparisons are sometimes made between conditions "Mocamedes " along the river in Kavango – hereafter called the “river” Caiundo Vila Nova da Armada – and those to the south and away from the river. This " is called inland Kavango, following the popular and local Katima Mulilo " "

Okongo " Bagani use of “inland” to characterize the non-riverine areas of " Ondangwa Rundu " " Shakawe the region. É É É É É É É É É ÉÉÉ É Grootfontein Tsumkwe É É É " " É É Within Namibia, the Kavango Region is one of 13 regions É É É É É " designated for purposes of administration and political É É Maun representation. The borders of the region have changed Francistown twice in recent years (Figure 2), as has its name: " " Ghanzi “Kavango” was used for several decades by the previous Gobabis " government, in 1992 it was changed to “Okavango”, and " Windhoek " BOTSWANA then back to “Kavango” in 1998. Walvis Bay Okavango River Cuito River NAMIBIA Gaborone " Dry drainage Okavango Delta The Okavango River drainage, showing its origins in Angola, its passage through Namibia and its ending in Figure 1 the Okavango Swamps or Delta in Botswana. The river Scale 1 : 16,000,000 forms the border between Namibia and Angola along a 0 200 400 600 river frontage about 415 kilometers in length.

Kilometres " Noordoewer 1 The present borders enclose an area of about 48 500 square Recent changes in the borders of the region. Before 1992, the Okavango kilometres, roughly 5.5% of the total area of Namibia. Its Figure 2 River formed the eastern border. From 1992 to 1998 the eastern border was taken as 21oEast longitude. The eastern border is now at 22o30’East. present population of about 179 thousand people makes up some 10% of Namibia’s population. Readers should be From 1998 cautioned that almost all other population estimates in other reports relate to the region’s boundaries as they existed 18˚00'E 19˚00'E 20˚00'E 21˚00'E 22˚00'E 22˚30'E between 1992 and 1998 (Figure 2). 17˚30'S

People in the region are represented by eight regional 18˚00'S councillors, one for each of eight constituencies (Figure 3). The following table provides estimates of the number of people in each constituency. The Rundu Urban constituency is smaller 18˚30'S than the town itself, and many people living in the town fall within the surrounding Rundu Rural constituency. In 2000 19˚00'S the population of Rundu consisted of about 45 000 people.

The quarantine fence is an important border and feature, and 1992-1998 its position in the region is shown in Figure 3. Areas south of the fence within the region are controlled by the Government or 17˚30'S by its parastatal, the Nambia Development Corporation (NDC).

The results presented below provide a perspective of normal 18˚00'S conditions in the region. Those conditions have been disrupted to some degree by the current confl ict, but hopefully the 18˚30'S changes are temporary, and life should return to normal once the confl ict stops. 19˚00'S

Table 1: Estimated population in each constituency in 20002. Before 1992 Constituency People

17˚30'S Kahenge 28 400 Kapako 21 200 Mashari 9 700 18˚00'S Mpungu 19 700 Mukwe 20 400 18˚30'S Ndiyona 20 100 Rundu Rural 51 000 19˚00'S Rundu Urban 8 500

18˚00'E 19˚00'E 20˚00'E 21˚00'E 22˚00'E 22˚30'E TOTAL 179 000

2 Borders of the eight constituencies, and the position of the quarantine fence in the Figure 3 south-western area, and in Namibia as a whole.

"Nkurenkuru " Mpungu Vlei "Tondoro Mupini Rundu Urban " " " Rundu Mashare Nyangana " Mpungu " Bagani Mukwe Rundu Rural Kahenge

Kapako Ndiyona Mashari Mururani "

Scale 1:2,500,000 0 50 100

Kilometres

Kapako Constituency Mururani Major towns and known places Quarantine fence

3 The distribution of schools, health facilities, N N N Figure 4 E NÑ Mbambi airfi elds and electricity lines. ku li N Nam N u N N n N g N N u N N N nd N N Nkurenkuru N No N NNÑ Ñ N N N N N N N P Nankudu Mpungu N N Ñ N e N NN Tondoro N N N Yinsu NÑNN N N N NÑ N ÑN N N Nzinze Rupara Muveve N N NÑ N N Mupini Rundu Takawasa N Nepara N N N NNNÑN N N ÑNN N N N N N N NÑN N NN N N Ñ N N P Shambyu Ñ Mabushe N ÑN N N Ekuli N N NÑN N N N NÑN N ÑN N N NÑ Kangongo N N N Bunya Kapako NNNÑÑNNNN NÑN ÑN N N ÑNN N N N N N N Ndonga Sikarosompo N N N N e N ÑN N Mbambi N N Mayara N N N N N N NÑN Nyangana N N NÑ Biro N u N MashareN N e NÑ NÑ Shadikongoro N N k N N N N NÑ N N N N Omega N N Ñ u N N N N N NP N ÑNN

NÑ Muparara N p N Mile10 NÑ N N N N NN ÑN N N ÑNN N N N M N NÑ Sharukwe N N N N N NÑ N Gcaruhwa a N N N N N NÑ N N PAndara N N N N N g N N N N N N n N N N N N N ÑN o Shinyungwe N

N N N N Ñ e NÑ d Kapupahedi N NcuncuniN BaramashoniN ÑN N N N N N N N N N N NÑ N N N N N N N ÑN KandjaraN N Bagani N NÑ Ncamagoro N N N NÑ N N N N N N N NN Hamoye Omatak e N N N o N N ukang N N N N N N R N N N N Ncaute N N N N N NÑ N N N P Hospitals ÑN Gcwatjinga N Khaudum NÑ N Ñ Health centres N N Katjinakatji N N Ñ Clinics Mangetti N e Airfields NÑ NN Mururani N N Schools Power lines Scale 1:2,000,000 N Karukuwisa Trunk roads 0 50 100 Main roads

o e ak at m N Sikereti Kilometres O Quarantine fence

Figure 4 shows the distribution of various public Electricity is supplied to many schools, health Figure 4 are gravel along the river and the Ndonga services in the region. Of the 301 schools, 250 facilities and various other government and private omuramba to Ncaute. Elsewhere in the inland are primary schools, 9 are secondary schools, and users along the Okavango River and along the main areas they are sandy tracks, but are shown as 41 are combined schools offering some primary road between Mururani and Rundu. Only 1% of all main roads because they are important access and secondary grades. There is also one technical rural households used electricity for lighting in 1991. routes. There are, of course, a great number of college in Rundu. Additional information on the sizes other sandy tracks. of schools and the phases they offer is given in Tar roads link Mururani and Rundu and Caprivi. Part Figure 27 in Chapter 3. The region’s health facilities of the road from Rundu to Mpungu is tarred, and consist of four hospitals, six health centres and 44 eventually all of that road to Okongo in Ohangwena clinics. will be tarred. Roads shown as “main roads” in

4 Chapter 2: The physical environment

The Okavango River and its tributaries, and Figure 5 Introduction average annual rainfall in the drainage area. 1 1 0 0 This chapter presents information on the region’s climate, Angola vegetation, and the Okavango River, the river being the only " Kuito surface water of note. The majority of people in the region base " Lucusse Huambo their livelihoods around the Okavango River, and Figure 1 on " page 1 provides an overview of the river’s drainage. A key

1 feature is the fact that more or less all the Okavango’s water 0 00 " Cangamba fl owing through Namibia and into Botswana has been drained " Caconda from catchment areas in Angola. Namibia thus contributes 9 almost no water to the Okavango. The length of the Okavango 00 in the Namibian section is about 415 kilometres. Capelongo "

800 Compared with other regions in Namibia and Botswana, " Cuito Cuanavale Zambia Kavango has a climate that is relatively sub-tropical. This is 70 because it receives more rain than many areas to the south and 0 Caiundo " west. Temperatures are generally high and, most importantly, 600 " frost almost never occurs. Vila Nova da Armada " Evale

5 Even though Kavango is more sub-tropical, its climate is very 00 dry from April to October. This is due to high rates of solar Katima Mulilo " radiation and evaporation, comparatively little cloud cover and " Okongo

" 4 " Ondangwa little rain, of course. Humidity levels are also low during these 00 Rundu " dry months. Most rains fall from November to March, but there Bagani " Shakawe are marked fl uctuations in total falls from year to year. É É É É 3 0 É É É 0 Namibia É É Much of the vegetation consists of É É É É É É É Grootfontein Tsumkwe É É " É É deciduous woodland, with local " É É É É É É É variations in its structure and species É É " Maun É É composition being largely dependent Okavango River É 2 0 on topography and soil conditions. Cuito River 0 Clearing and repeated burning have Okavango Delta Botswana had a great impact on the region’s Dry drainage lines Scale 1 : 7,500,000 vegetation (see Chapter 4). 1 0 0 0 100 200 300 " Ghanzi

Kilometres 5 The drainage of the Okavango River Water fl ow in the Okavango River

Figure 5 provides a more detailed view of the Overall, the Cuito contributes about 45% of all the only 175 m3/s in April). The Okavango also has Okavango’s drainage3. In Angola the river is called water that the Okavango delivers to the delta every a much more variable fl ow than the Cuito. Thus, the Cubango River, and the distance from its source year. While the remaining 55% is provided by the the highest rate of fl ow (962 m3/s) ever recorded to the Cuito confl uence is 930 kilometres. The Cuito Okavango, the relative and actual amounts provided for the Okavango is about 90 times greater than River is 730 kilometres in length from its source by the two rivers vary a great deal during the year the lowest rate ever recorded (11.1 m3/s) while to the confl uence, and the Cuito is the Okavango’s (Figure 6). At Rundu, the Okavango peaks in April and the same fi gures for Cuito vary only by a factor only signifi cant tributary. Indeed, for seven months of while the Cuito also usually peaks in April, it also often of less than 10 (lowest of 64 m3/s compared the year the Cuito contributes more water than the has its highest fl ow in May. Peak fl ows reach the delta with the highest of 550 to 600 m3/s). The higher Okavango (see below). The catchment of the Cubango in Botswana about 3-4 months after passing Rundu. discharge rates of the Okavango mean that it is estimated to be 88 700 km2, while that of the Cuito carries more water than the Cuito between January is 60 600 km2. Taking these two catchment areas The Okavango brings in much more water during and May. However, it then drops so much that the together means that the river obtains its water from fl oods, with the highest discharges at Rundu being more stable Cuito contributes more water for the a total area of almost 150 000 km2. The Omatako much higher (average in April is 405 cubic metres/ remaining seven months of the year (Figure 6). Omuramba has never been known to actually fl ow into second (m3/s)) than peaks for the Cuito (average of the Okavango, but it potentially drains a substantial area (about 55 700 km2) and must have been a signifi cant tributary in much wetter periods long ago. Average volumes of water (in cubic metres) carried each month by the Figure 6 3 The length of the Omatako is 635 kilometres from its Okavango and Cuito Rivers at Rundu and Dirico, respectively . source near the Omatako hills to its confl uence at Ndonga. A number of other river courses (Ekuli, Mpuku, 1,200,000 Ndonga, and Rukange) also meet the Okavango but - like the Omatako - they are always dry. 1,000,000 The Okavango’s gradient is extremely shallow, varying between 1:5 800 and 1:7 100 along the Namibian 800,000 section up to the confl uence with the Cuito. At Kangongo, east of the Cuito confl uence, the gradient 600,000 steepens to about 1:1 350 to 1:2 250. The fl oodplain Okavango at Rundu in the Namibian section varies between 2 and 6 Cuito at Dirico Cubic metres kilometres in width. 400,000

Higher areas in Angola where both the Okavango/ Cubango and Cuito have their sources have an 200,000 average annual rainfall of over 900 millimetres per year (Figure 5). From those wet areas, the river 0 progressively fl ows through drier and drier zones until it reaches the delta in Botswana, where the average October November December January February March April May June July August September annual rainfall is less than 400 millimetres. The fl ooded area of the delta is estimated to be 8 400 km2.

6 The adjacent table provides measures of annual Table 2: Total average, minimum and maximum annual runoffs at Rundu and Mukwe in runoffs at Rundu and Mukwe, i.e. above and below million cubic metres. Each season runs from October to September. the Cuito’s confl uence. The differences between maximum and minimum volumes are substantial: Place and years of data Average Minimum (season) Maximum (season) about four times at Rundu and about three times at Mukwe. But what is perhaps more important is that Rundu (1945-1999) 5 198 2 260 (1971/72) 9 810 (1962/63) river fl ows oscillate frequently between high and low years. There are thus many years when the Mukwe (1948-1997) 9 594 5 607 (1995/96) 15 354 (1967/68) fl ow is high and many other years with little fl ow. The extent of those year-to-year changes is shown clearly in Figure 7. This graph also refl ects some Rundu longer term changes, especially the lower fl ows 20,000 during the 1980’s and 1990’s, much of which is attributable to lower rainfall in those two decades.

15,000

10,000

Millions of cubic metres 5,000

Total volumes of water (in millions of cubic metres) carried per season by the Okavango Figure 7 at Rundu and Mukwe. The orange lines are 0 5-year moving averages. Mukwe 20,000

15,000

10,000

Millions of cubic metres 5,000

0

1945/46 1949/50 1954/55 1959/60 1964/65 1969/70 1974/75 1979/80 1984/85 1989/90 1994/95 1999/00

7 Total volumes of water (in millions of cubic metres) carried by Annual variations in discharge per month in October Figure 8 and April are compared in Figure 8. It is clear that the Okavango in April and October at Rundu and Mukwe. there is relatively little variation from year to year in Rundu 3000 3000 October volumes, showing that the river subsides each year to a fairly stable rate of discharge. The 2500 2500

only signifi cant change is the lower fl ows in recent Millions ofcubicmetres years. April fl ows, by contrast, fl uctuate wildly from 2000 2000 year to year, refl ecting great year-to-year differences April in rainfall and runoff. 1500 1500

1000 1000

Water abstraction within Namibia Millions of cubic metres

The total amount of water abstracted per year from 500 500 the Namibian side was estimated to be 5.1 Mm3 in October 0 0 1996, plus another 0.26 Mm3 from boreholes close to the river. This was much less than was allowed in terms of permits issued for purposes of irrigation and domestic supply: in the early 1990s, permits Mukwe provided for an amount of just under 18 Mm3 per 3000 3000

year to be removed from the river. The greatest use April of river water was for irrigating Nambia Development 2500 2500 Millions ofcubicmetres Corporation (NDC) and government farms, mainly downstream of Rundu, and to supply the town of 2000 2000 Rundu itself. Even during years and months with low fl ows, the total amounts of water removed from 1500 1500 the river in Kavango are very small. 1000 1000 Millions of cubic metres October 500 500

0 0 1945/46 1949/50 1954/55 1959/60 1964/65 1969/70 1974/75 1979/80 1984/85 1989/90 1994/95 1999/00

8 Underground water

Most underground water lies in a variety of aquifers in Kalahari sediments that extend to depths of up to 350 metres in some places. The only other aquifers are in Damara Sequence rocks of the Nosib Group in the area of Andara and Popa Falls.

The Kalahari aquifers are largely recharged by Figure 9 Depths of water below the ground, and the general water seeping in a northerly direction from direction in which water fl ows under the ground 4. elevated areas in the south of the region (Figure 9). This means the Okavango River probably “gains” N N Nkurenkuru water from shallow aquifers, rather than the river Mpungu Rupara N Rundu water recharging aquifers. There are two quite N MashareN different zones of water depth. In the north-eastern Omega part of the region water is usually about 20 metres N below the ground, while it is generally at between 40 and 80 metres in the south-west (Figure 9). Most boreholes produce yields in excess of 1-5 cubic metres per hour, which are adequate to N Katjinakatji Direction of water flow supply small villages. In addition, the majority of borehole water is of good quality, with total Depth of water about 20 metres dissolved solid (TDS) values of less than 1 000 Depth of water 40-80 metres milligrams per litre. Scale 1:3,000,000 N Karukuwisa 0 50 100 N Sikereti Several hundred boreholes and many hand-dug wells are spread across the inland part of Kavango. Kilometres Most boreholes were drilled and supplied with pumps by the Department of Water Affairs. There are, in addition, boreholes and pumps that are privately owned by farmers, parastatals such as the Namibia Development Corporation, and others.

9 Rainfall

The Kavango region forms part of a much bigger, largely fl at landscape of Figure 11 also shows the numbers of years during which different total amounts Kalahari sands, across which there is a clear rainfall gradient (Figure 5). Total of rain fell at Andara and at Rundu. More than 500 mm fell in most years (85% of annual falls are higher to the north and east of the region, and lower to the west years at both Andara and Rundu), and more than 600 mm was recorded in about and south. It is for this reason that rainfall within Kavango is greater in the east 65% of all years at both places. By contrast, drier years with less than 400 mm were and along the river, and lower to the west and south of the river (Figure 10). recorded in about one out of every six or seven years at Rundu and Andara.

Records over the past 60 years are available for most years at Rundu and Table 3: Average, lowest and highest annual rainfalls at Andara, Rundu and Andara (Figure 11). The annual totals are extremely variable, ranging from less Nkurenkuru. Each season runs from July to June. than 400 millimetres (mm) in the driest years to over 1000 mm in exceptionally wet years. Almost every year differs from the one before and after it. Place and years of data Average Lowest (season) Highest (season)

Some cyclical changes are evident, especially the dry cycle during the 1960’s, Andara (52 years) 579 247 (1994/1995) 1204 (1977/1978) the wetter period during the 1970’s, and then a long recent period of drier years Rundu (60 years) 566 274 (1972/1973) 1120 (1977/1978) since the early 1980’s. These cycles are similar to those seen in rainfall records Nkurenkuru (36 years) 610 330 (1995/1996) 1058 (1973/1974) at places in the Caprivi, north-western Botswana and to the south-west in the Grootfontein area.

Average annual rainfall in Figure 10 the Kavango region.

" Nkurenkuru

Rundu

"

" Bagani

Annual average rainfall (millimetres) Less than 500 500-525 525-550 More than 550

10 Figure 11 Total annual rainfall at Rundu and Andara over the past 60 years (left) and the number of years during which different amounts of rain fell (right).

1500 Rundu 20

1200 16

900 12

600 8 Number of years Total rainfall (millimetres) rainfall Total 300 4

0 0 0 200 400 600 800 1000 1200 1940/41 1950/51 1960/61 1970/71 1980/81 1990/91 1999/00 Total rainfall (millimetres)

1500 Andara 20

1200 16

900 12

600 8 Number of years Total rainfall (millimetres) rainfall Total 300 4

0 0 0 200 400 600 800 1000 1200 1940/41 1950/51 1960/61 1970/71 1980/81 1990/91 1999/00 Total rainfall (millimetres)

11 Rainfall in the region is highly seasonal, with no rain of any signifi cance being received The number of days on average that falls of more than 1, 5 and 10 between May and September (Figure 12). Thereafter, there is a progressive increase until Figure13 millimetres of rain were recorded each month at Andara, Rundu and Rupara/ Nkurenkuru 5. January, the month in which the highest totals are recorded on average. February has slightly less rain, followed by even less in March and April. About 80% of all rain falls Andara Rundu Rupara/Nkurenkuru between December and March, with the remaining 15% being received in November and 15 April. There is a slight tendency for rains to fall earlier in the east and thus later in the west. 12

9 Average rainfall per month at Andara, Figure 12 Rundu and Nkurenkuru. 6 200 Average of the three stations 3 Andara

150 Rundu with 10 or more millimetres of rain Days 0 Nkurenkuru

100 15

12

Total rainfall (millimetres) rainfall Total 50 9

6 0 Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun 3

It is clear that rainfall is often inadequate for crop growth as a result of low rainfall, irregular with 5 or more millimetres of rain Days 0 falls, or long periods of hot and dry weather when crops planted earlier wither and die. The growth of natural pastures is likewise limited in some years as a result of poor rainfall.

While no clear measures of what might be called drought are available, graphs showing 15 the numbers of days on which different amounts of rain fell provide indications of how frequently rains of potential value for crop and pasture growth may fall (Figure 13). The 12 graphs are for daily totals of more than one, fi ve and ten millimetres respectively. 9 Based on these graphs, better falls of rain can only be expected from November onwards, and few good falls can be expected after March. January is clearly the month when rain 6 falls both most frequently and in the largest amounts each day. There are slight differences in the onset of higher falls of rain, such that good falls occur earlier in the season in the east 3 (Andara) compared with the west (Rupara/Nkurenkuru).

Days with 1 or more millimetres of rain Days 0 October November December January February March April

12 Temperatures, evaporation, and wind Average daily maximum, mean and minimum Figure 14 temperatures each month at Rundu. The region’s generally warm climate is refl ected in January Figure 14, which shows that temperatures increase 40 N very rapidly from the coldest months of June and July Mean maximum W 57 E to the warmest month of October. This is because 30 there is relatively little cloud cover to shield incoming Mean solar radiation in October, whereas increasing cloud Mean minimum S cover and rains make the remaining summer months 20 cooler. Average maximum temperatures are above April o (Celsius) Degrees 30 Celsius in all months excepting May, June and 10 N July. Temperatures during the winter months seldom W 59 E get close to freezing point, and only in June, July 0 and August are average minimum temperatures Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May June below 10o Celsius. Frost is therefore exceptionally S rare, and only ever occurs in low-lying valleys. Average rates of evaporation at Rundu compared with Figure 15 250 average rainfall. Total evaporation rates amount to about 1 950 July Evaporation millimetres per year. About three to four times more N 200 Average rainfall at Rundu water is therefore lost through evaporation than falls W 62 E as rain per year. The highest rates of evaporation are in September and October (Figure 15) when 150 temperatures are high, there is little moisture in the S 100 air and it is often more windy than at other times of Millimetres the year. October 50 N Wind speeds are generally low, rising from averages 49 W E of about 3 kilometers per hour (km/hour) in the 0 mornings to about 12 km/hour in the early afternoon, before dropping again to about 3 km/hour in the Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun evenings (Figure 16). Morning and evening winds 15 S are strongest in October and November, while 14h00 Figure 16 afternoon winds are lightest in December and 12 January. In most months, however, it is completely Average wind speeds per month at 08h00, calm for over half the time. Most winds are from 9 14h00 and 20h00 at Rundu (left), and wind the east, between north-east and south, and only in roses (above) showing the proportions of wind January is there some wind from the west. from different directions in January, April, July 6 and October. Figures at the centres of each

Kilometres per hour rose are the percentages of time that it is calm 08h00 3 during those months. 20h00

0 Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun 13 Vegetation

Much of the vegetation in the Kavango consists diffi cult. Thus, much of the vegetation is often best groupings. Firstly, much of the region consists of broad-leafed, deciduous woodlands that vary characterised as mosaics of various small units. of fairly tall woodland growing on deep Kalahari according to topography and the nature of the soils sands. This is true of the Burkea-Teak woodlands, that support them. A great deal of that variation While the vegetation types shown in Figure 17 Burkea woodland and shrublands, Woodlands of the is extremely local, with the result that classifying are mosaics consisting of many species and northern sand plain, Kalahari woodlands of Caprivi, plant communities into zones or types is often very communities of plants, there are three general Eastern drainage vegetation, and Catchment divided vegetation. In many of these areas there

Figure 17 Vegetation types in the Kavango 5.

N N Nkurenkuru Mpungu

Rundu NMusese Okavango N NMashare N Ndonga N N Muparara NOmega Gcaruhwa N Andara N Bagani N Ncamagoro

N Ncaute ko Omata N N Khaudum Katjinakatji

NMangetti N Mururani

Scale 1:2,000,000 N Karukuwisa 050100 o ak at N m Sikereti Kilometres O

11 Vegetation types Catchment divide vegetation Omatako drainage Burkea-Teak woodlands Eastern drainage vegetation Riverine forests Burkea woodland and shrubland Floodplains and open water Shrublands of the southern panveld Camelthorn-Silver Terminalia shrubland mosaic Kalahari woodlands of Caprivi Woodlands of the northern sand plain

14 are the remains of old dune systems and the Riverine forest along the river near Andara and Figure 18 provides information on vegetation vegetation varies considerably between that on Bagani should be given high conservation value cover. The scale of mapping is rather fi ne, with the sandy dunes and the more clayey soils in the because it is the only forest of its kind remaining the result that many of the local changes in interdune valleys. along the Okavango. vegetation structure are visible. This is especially true in areas where the remnants of old dune Secondly, there are vegetation types associated Thirdly, there are two vegetation types in which systems are visible, with that in the interdune with drainage systems, including the Floodplains there are many pans: Camelthorn-Silver Terminalia valleys often being more open or grassland. and open water, Riverine forest, and the Omatako shrubland mosaic and Shrublands of the southern The same is true of plant cover along the dry drainage. Vegetation along some of the other dry panveld. The soils in these areas are often drainage lines. Many of the patches of dense drainage lines is similar to that in the Omatako. shallow. bush are teak woodlands, especially those in the Mururani-Katjinakatji area, west of Andara and in the Caprivi Strip.

Figure 18 Vegetation cover in the Kavango 6.

15 Chapter 3 : People and social conditions

Introduction

This chapter presents information on the region’s population, housing, health and education. The population has grown rapidly in recent decades and is dominated by young people. Substantial movements by Approximate boundaries of the major Figure19 people into the region have also added to the population, language groups 8. but other people have moved out to seek work in other regions. Likewise, many people have moved from rural areas to Rundu, which has the only urban population in the Kavango. Most rural people live in homes built of wood and grass harvested locally, and most homes use N NNkurenkuru Mpungu wood for cooking fuel. The majority of rural people live Rupara N Rundu and farm in small settlements near the Okavango River N MashareN or along a number of omurambas. It is in these more OmegaN Mbunza Kxoe densely populated areas that most schools and health Shambyu Mbukushu facilities are to be found. Enrolments are dominated Kwangali by pupils in the lower primary grades and relatively few people have completed higher grades. The most Rumanyo N (Gciriku) important ailments are HIV/AIDS, malaria, diarrohea, Katjinakatji tuberculosis, bilharzia and malnutrition.

NKarukuwisa Language groupings Scale 1:3,000,000 0 50 100 NSikereti

The distribution of people speaking different languages Kilometres (Figure 19) was derived from the 1991 Population and Housing Census. That census provided for fi ve major languages in the region: Kwangali, Shambyu, Rumanyo (Gciriku), Thimbukushu and Bushman (primarily Kxoe) languages. Several other languages are recognised as being spoken by signifi cant numbers of people, including Mbunza in an area just west of Rundu. There are also substantial numbers of people that have Nyemba as their mother tongue. This is an Angolan language, which in 1999, was reported as being spoken by 15% of the rural population and by 30% of the urban population in Rundu7.

16 Population Figure 20 Growth of the population between 1951 and 1999 and Estimates of the number of people in the region projected growth to 2020 9. are available from 1951 onwards (Figure 20). These 250,000 fi gures suggest that the population grew very rapidly Recorded population Projected population Rundu population from 1951 until 1981, and especially so between 1970 and 1981 when the population grew at about 200,000 7.5% per year. The growth rate between 1960 and 1970 was 6.0%. While it is seems certain that growth rates were indeed high, the rates might have been 150,000 infl ated if the coverage of people in 1960 and 1970 was less than complete. In addition, some of the growth has been due to the migration of people 100,000 into the region from Angola (see below), and it is Number of people possible that rates of migration have been greater in some periods and lower in others. 50,000 The most recent population estimate based on a survey is from 1999, when the total number of people in the region was put at about 176 6007. Population 0 projections from then onwards estimate the total 1951 1960 1970 1981 1991 1999 2010 2020 population to be 179 400 in 2000. Subsequent growth over the next two decades suggests that Rundu has grown at a very rapid rate, from about that the distribution of people, at least in that the population will amount to about 216 000 in 1 500 people in 1971 to roughly 45 000 in 2000. part of the region, has been strongly infl uenced 2020 (Figure 20). These projections assume rates Those changes mean that the proportion of the by the presence of good road access. Secondly, of population growth to be about 1.5% over the next region’s urban and rural population has changed to there is a large area consisting of many small, few years, before dropping to below 1% as a result a great degree. About 97% of the population was scattered settlements south of the river and west of the increasing effects of AIDS mortality. rural and 3% urban in 1971, while in 2000 about of the Mururani-Rundu road. Many of these small 75% is rural and 25% of the population is urban. settlements are clustered along some of the dry Average life expectancies in 1996 were 55 years drainage lines or omurambas. Densities around for men and 62 years for women, while the most The distribution and density of people can be the settlements generally range between 1 and recent estimate for 1999 by the United Nations classed into three zones (Figure 21). The fi rst is 10 people per km2. It is clear that many of the Development Programme is 49 years. This reduction the very densely populated ribbon along the river, settlements were established quite recently. The is attributed to high rates of HIV infection in Kavango where densities are generally higher than 40 people third zone consists of the very sparsely populated (see below), and predictions by the Central Statistics per km2 and exceed 100 people per km2 in some remainder of the region where there are very few or Offi ce and Ministry of Health and Social Services places. About 114 000 people, or about two thirds no people. This zone, with population densities of suggest that life spans will drop to 38 years for men of the region’s population, live close to the river. less than 1 person per km2, makes up about 78% and 40 years for women by 2011. Thereafter, they Forty-fi ve thousand of these people are in Rundu. of the total area. Most large, open areas are in the are expected to rise to 48 and 50 years in 2021 for There are also very high densities along certain south-east, south west and in the Caprivi Strip. men and women, respectively. sections of the Mururani-Rundu road, indicating

17 Table 4: Areas and percentages of the region with different population densities.

Density (people per km2) Area (km2) Percentage

Less than 1 38 037 78 1-10 7 678 16 Figure 21 The estimated density of people in 2000 10. 10-40 1 935 4 40-100 610 1 More than 100 213 1

18 The age and sex structure of the Figure 22 The age and sex structure of the 1991 population in 1999 (Figure 22) population in 1991 and 1999. reveals several noteworthy features. Firstly, young people make up a 75+ Females Males huge proportion of the population: 70-74 75% of all people are less than 65-69 30 years old and 43% are less 60-64 55-59 than 15. 50-54 45-49 Secondly, there are signifi cantly 40-44 fewer 0-4 year-olds than older 35-39 children, and this reduction has Age group 30-34 happened since 1991. The fall 25-29 suggests that population growth 20-24 may be slowing down, perhaps 15-19 as a result of increased mortality 10-14 due to HIV transmitted by infected 5-9 mothers. Part of the reduction is 0-4 also probably due to a decline in fertility. The average fertility rate in 1991 was 7.1 children while in 1996 it had fallen to 6.6 children. 1999

Thirdly, there are roughly equal 75+ Females Males proportions of adult males and 70-74 females in the region, unlike some 65-69 other regions in Namibia where 60-64 there are far fewer males as a result 55-59 50-54 of migration. 45-49 40-44 35-39

Age group 30-34 25-29 20-24 15-19 10-14 5-9 0-4

15,000 12,000 9,000 6,000 3,000 0 3,000 6,000 9,000 12,000 15,000

Number of people

19 1960 While large numbers of men remain resident in the region, it is clear that more males than females have indeed left to live 65+ Females Males and work elsewhere (Figure 23). Moreover, these graphs indicate 55-64 that men started to leave as migrants before women, and that the proportions of men and women now living elsewhere have 45-54 increased over the past few decades. 35-44

These graphs refl ect migration away from Kavango, but large 25-34 numbers of people have also moved into the region, especially 15-24 from Angola. In 1991, 8% of all people counted during the Figure 23 census in Kavango had been born in Angola, 2% in Ohangwena, 0-14 Oshikoto, Oshana or Omusati, and 2% in other places8. These The age and sex structure of fi gures probably underestimate the number of immigrants since a the population in 1960, 1970 and 1991, and the number signifi cant number of people may have been reluctant to declare of people speaking Kavango 1970 that they were born in a foreign country. That conclusion is languages resident inside the supported by the observation that 15% of all people reported region (green bars) and else- 65+ 9 Nyemba (generally considered to be an Angolan language) as where in Namibia (red bars) . 55-64 their home language in 19997. 45-54 In addition to migration in and out of Kavango, there have been 35-44 large-scale movements of people within the region. One signifi cant Age group movement is of people from rural areas to the urban area of 25-34 Rundu, thus contributing to the substantial growth of the town. 15-24 In 1999, 24% of residents in Rundu had moved there within the past 10 years. Two-thirds of people had moved to Rundu from 0-14 elsewhere in Kavango, while 12% had come from Ohangwena, Oshikoto, Oshana or Omusati, 9% from Angola, 5% from Caprivi and the remainder from elsewhere7. 1991

Another signifi cant movement is of people between settlements, 65+ and 2.8% of people were reported in 1999 as having arrived to live 55-64 in new villages within the last three years. Many such movements are from homes along the river to small villages in the inland. 45-54 At least some of these movements are apparently initiated by wealthier people who establish cattle posts in unsettled areas, 35-44 and then recruit family members and other people to live at the 25-34 cattle posts. These new settlements become more permanent and developed as the new settlers clear land for crops, and boreholes 15-24 are drilled nearby. However, it is also evident that many such small 0-14 settlements do not last permanently, their residents either moving 40,000 30,000 20,000 10,000 0 10,000 20,000 30,000 40,000 back to the river or to other new villages in the interior. Number of people

20 Households and housing conditions Health

The majority of people live in rural households consisting of about 6-7 people, A comprehensive set of reliable statistics to refl ect on health issues while urban homes in Rundu are smaller on average, consisting of 5-6 people is currently not available. However, it is clear that many people in the (Figure 24). However, there are also substantial numbers of bigger homes of 10 region suffer from diseases and conditions commonly associated with and more people. Female-headed households are smaller (average of 5 people) a poor, rural and subtropical environment. The most important of these than male-headed homes (7 people). About 70% of households are headed by are malaria, diarrohea, tubercolosis, malnutrition and bilharzia, the men, both in rural areas and Rundu. Most households have about one active latter ailment being common along the river. HIV infection rates have person for each dependant in the home. A dependant is someone under the age risen dramatically in recent years (Figure 25), as refl ected by rates of of 15 or older that 64, and dependency ratios are calculated as the total number infection among pregnant women attending four hospitals. The fi gures, of dependants as a percentage of the total population. In 1991 the dependency which are reliable indicators of infection rates among all sexually active ratio was 51%, while an estimate in 1999 gives the ratio as 46%. adults, suggest that rates of infection may have stabilised at about 15-20%. As might be expected, the great majority of homes rely on locally available natural resources for fuel and building materials. Thus, 97% of homes use wood Mortality rates among children are high, Caprivi being the only region for cooking, and 92% have the walls of their houses made of wood or mud. in Namibia with higher death rates. Of 1 000 children born in the Eighty four percent of houses have grass roofs, most others being of iron (9%) Kavango, 84 die before the age of one, 109 will not reach the age of or wood or sticks. A slight change in sources of domestic water is shown in the fi ve, and 122 will die before the age of ten years. following table.

Table 5 : Percentages of homes using different sources of domestic water in 1991 and 19997,8. Piped water Wells Boreholes River water 1991 14% 12% 24% 48% Percentage of pregnant women infected with HIV tested in four 1999 18% 7% 24% 41% Figure 25 hospitals11.

50% 20% 1994 Rural 1996 40% Urban Figure 24 1998 15% 2000

30% Percentages of rural and urban households of 10% 20% different sizes 8. Percent of households Percent

10% 5%

0% 1-3 4-6 7-9 10-12 13-15 >15 0% Andara Rundu Nyangana Nankudu Number of people per household

21 Figure 26 provides an indication of the breadth of coverage of health facilities. This is based on the assumption that people within 5 kilometres of these facilities have good access to health services. Using the underlying density of people, estimates suggest about 67% of people live within 5 kilometres of a health facility, leaving about 33% of the population beyond 5 kilometres of health services.

Areas within 5 kilometres of health facilities Figure 26 compared with the density of people.

22 Education

Compared with other regions in Namibia, the In the middle of 1999, there were some 58 000 areas are forced to either drop out of school after education system in Kavango is extremely defi cient. children enroled at 301 schools (Figure 27). The completing the fi rst few grades or to leave home Teachers are less qualifi ed than elsewhere, there many schools along the river and the road between and move to a school offering higher grades are few secondary schools, physical facilities at Mururani and Rundu are comparatively large, while along the river. schools are often in bad shape, most adults have those elsewhere in the inland areas are much received little formal education, and relatively few smaller. Almost all the small, inland schools offer children complete their schooling. only lower primary grades, so most children in those

The distribution and sizes of schools offering primary grades and secondary grades, and combined Figure 27 schools that offer both primary and secondary grades. Note that many schools offer only a number of the primary or secondary grades, for example Grades 1-5 or Grades 8-10.

23 In addition to the lack of schools offering higher Numbers of male and female pupils in Figure 28 primary grades in the inland areas, there is also a Grades 1-12 in 1999 12. lack of schools providing secondary grades along 5000 the river. This is one reason why there are so many Females more pupils in primary than in secondary grades Males 4000 (Figure 28). There are also more boys than girls in secondary grades, largely because many girls drop out of school as a result of teenage pregnancies. 3000

Levels of education amongst adults are extremely low, especially for women and older people Enrolments 2000 (Figure 29). In 1996, only 4% of men and 1.5% of women aged 15 and older had completed Grade 11 or any higher level of education. Taking the 1000 completion of Grade 4 as an indicator of literacy, most men older than about 50 years and most women older than about 40 cannot read or write. 0 1 2 3 4 5 6 7 8 9 10 11 12

Grade

Figure 29 Percentages of people aged 15 and older that have completed various levels of education 13.

100% Females Males 100%

80% 80%

Higher than Grade 10 60% 60% Grades 8 -10 Grades 4 - 7 Lower than Grade 4 40% 40% No education

20% 20%

0% 0% 15-24 25-34 35-44 45-54 55-64 65-74 >75 15-24 25-34 35-44 45-54 55-64 65-74 >75 Age group Age group

24 Pupils enrolled in Grades 1, 4, 7, 10 and 12 Total enrolments in the 12 grades Figure 30 have increased from about 41 100 between 1992 and 1999. pupils in 1992 to 58 100 in 1999. 12,000 Part of that increase has been due to population growth, but enrolments 10,000

also increased because more s il 8,000 children now remain at school for Grade 1 pup

f Grade 4 longer to complete higher grades 6,000 er o Grade 7 (Figure 30). For example, numbers b Grade 10 um of Grade 4 pupils almost doubled N 4,000 Grade 12 from 4200 to 1992 to 8000 in 1999, and there were only 212 Grade 12’s 2,000 in 1992 compared with 600 in 1999. By contrast, numbers of Grade 1’s 0 dropped because of policy changes 1992 1993 1994 1995 1996 1997 1998 1999 introduced to prevent pupils from failing and repeating the same grades; the lower grades were previously over-enrolled with many repeaters before these changes.

Areas within 5 kilometres of schools Figure 31 provides an indication of Figure 31 compared with the density of people. the breadth of coverage of schools. This is based on the assumption that people within 5 kilometres of schools have good access to education. Using the underlying density of people, estimates suggest about 94% of people live within 5 kilometres of a school, and 6% live beyond a distance of 5 kilometres of schools.

25 Chapter 4 : Livelihoods and environmental pressures

Introduction Land uses and ownership

Information on land uses and ownership, household The two maps provide information on the uses of Shadikongoro, Shitemo, Vungu Vungu and Mashare, economies and structures, farming and pressures land (Figure 32) and the people or organisations that and along the Mururani-Rundu road at Ncamagoro on natural resources is provided in this fi nal chapter. own the land (Figure 33). In some circumstances, and north of Katjinakatji. The 44 Mangetti farms were Good recent information is available for some aspects “ownership” refl ects legal custody over land, while established to promote commercial farming in the but rather little appears known or available for others. in others it refers to the people or bodies that have region, but apparently not all of them are currently The following sets of data provided most of the results control over land. Areas used for different purposes allocated to farmers. Each of the 44 farms is about presented here: the Lux-Development population and owned/controlled by different groups are shown 500 hectares in size. survey done in 1999; Agricultural Surveys in 1996/97, in Tables 6 and 7 on the next pages. 1997/98 and 1998/99; and the 1994 Income and The other group of large-scale farms are those used Expenditure Survey. The greatest part (about 74%) of the region is by private individuals who have apparently either laid used for communal grazing, rural settlements and claim to the land or have been allocated large tracts It is clear that many inland households were only crop cultivation. Areas immediately in and around of land by tribal authorities. Most of those farms are established relatively recently, and agricultural settlements are treated as a mix of settlements and along the Mururani-Rundu road, but there are also practices and economic activities in rural households small-scale farming in Figure 32. The area actually several new farms that have been established to the along the river and those in inland areas differ in a cleared for cultivation amounts to about 4% of the west of Katjinakatji and east of Ncaute. The bounda- number of ways. Some analyses compare conditions region (see Figure 48 on page 42). ries of these new farming ventures are not known, in households along the river with those in inland and the areas are thus roughly indicated in Figures areas, while other results compare riverine homes A total of 15% of the region is used for nature 32 and 33. It is quite possible that many other farms with those in three inland zones in the inland: western, conservation in the Khaudum, Mangetti, Popa Falls are being claimed elsewhere in the inland areas. central and eastern. This is because some data and Caprivi nature reserves, and those reserves are can only be grouped into riverine and inland areas, owned and controlled by the government through the Other than communal or tribal authorities that control whereas other sets can be divided into the three Ministry of Environment and Tourism. about 74% of the region, the other major owners or inland areas. controllers of land are the government (17%), NDC There are two kinds of large-scale farms. The fi rst (5%) and large-scale private farmers (5%). The most important pressures on the environment consist of surveyed farms used by the government reported here are the clearing of natural vegetation or the Namibia Development Corporation (NDC), or for crop cultivation and the widespread bush fi res that which have been allocated by government to private are such a prominent feature of the winter months. individuals. Two large farms in the Mangetti area and Few data appear to be available on other pressures, one at Omega are used by the NDC, while the Minis- such as the effects of fi shing in the river, confl icts try of Agriculture, Water and Rural Development runs between people and wildlife, and the harvesting of the quarantine farm at Mangetti. Other large govern- selected trees for commercial timber production. ment or NDC farms are along the river, at Musese,

26 Table 6: Approximate areas and percentages of the region used for various purposes.

Use Area (km 2) % of total area Rundu town 7 Less than 1 Quarantine 588 1.2 Large-scale farming 4 840 10.0 Nature conservation 7 319 15.1 Communal 35 703 73.7 Total 48 457 100.0

N Nkurenkuru N Mpungu NTondoro Musese N O Rundu kavango Vungu Vungu Mashare N N N Shitemo N Shadikongoro N Muparara Gcaruhwa Omega N NBagani N Ncamagoro N Hamoye

NcauteN Land uses N Khaudum N Katjinakatji Communal grazing Nature conservation

Mangetti Quarantine farm N Settlement and small-scale farming N Mururani Large-scale farming Developing large-scale farming Scale 1:2,000,000 N Karukuwisa Trunk roads

0 50 100 o Main roads ak at m O N Sikereti Quarantine fence Kilometres

Figure 32 Areas used for different purposes.

27 Table 7: Approximate areas and percentages of the region owned or controlled by different groups.

"Owner" Area (km 2) % of total area Local government (Rundu) 7 Less than 1 Namibia Development Corporation 2 222 4.6 Private farmers 2 401 5.0 Central government 8 124 16.8 Communal authority 35 703 73.7 Total 48 457 100.0

N Nkurenkuru N Mpungu Rundu Tondoro N townland Musese N O kavango Vungu Vungu Mashare N N N Shitemo

u

N k Shadikongoro N Muparara Gcaruhwa u Omega N p M NBagani N Ncamagoro N Hamoye

NcauteN Land ownership N Khaudum N Katjinakatji Communal authority Government

Mangetti Local government N Namibia Development Corporation N Mururani Private farms Developing large-scale farming Scale 1:2,000,000 N Karukuwisa Trunk roads

0 50 100 o Main roads ak at m O N Sikereti Quarantine fence Kilometres

Figure 33 Areas owned or controlled by different people or organizations.

28 Farming

There are several good reasons why much of this in November and December. Batches of crops are and pumpkins are also grown in small quantities, chapter is devoted to farming. First, a great proportion often planted at different times, each planting session generally by intercropping with mahangu. of people depend on farm produce for much of their following a period of good rainfall. This improves the food intake. About 96% of all households are reported chances of at least some of the crop being successful Proportions of households growing different crops as having some kind of farming activity, and 71 % if some plants do not survive periods of hot, dry vary slightly between the river zones and the three of rural households depend upon farming as their weather. Planting at different times also means that inland zones of the region (Figure 35). Thus, almost all main source of income. Second, a fair number of the harvest can be done gradually over a longer time households in the inland areas grow both mahangu people derive cash incomes by selling farm products. so less labour is needed than if the whole crop had and maize whereas about 15% of riverine households Third, farming provides opportunities for people who to be harvested in a short period. The chances of the do not grow these crops. These homes may lack have substantial cash incomes from wages or other whole crop being eaten by birds are also reduced. access to fi elds with suitable soils, a constraint that is sources to invest surplus cash, perhaps by acquiring more severe along the densely populated river than additional cattle and larger fi elds. Fourth, of all human Mahangu is the dominant crop in the region, and elsewhere. In addition, alternative incomes are more activities in the region, farming has a much greater about 75% of the cultivated area is used for its available along the river, for example from retail trade impact on the natural environment than any other. production. Smaller areas are used to grow maize and jobs with wages. Five percent of households Finally, because agricultural activities and methods and sorghum. Most of the mahangu and maize along the river have neither crops nor livestock, and have been practiced over many generations, an harvest is consumed domestically, while sorghum is 1% cultivate only vegetables and fruits. understanding of farming provides a useful view used to produce beer. Groundnuts, beans, melons of how people have adapted to the environmental conditions that surround them. Likewise, an The crop farming calender in Kavango, showing the usual periods when appreciation of the established values that people Figure 34 attach to farming provides perspectives on how different activities are undertaken, and average rainfalls per month at Rundu. people may face changing circumstances in the 200 future. Intense working periods Less intense working periods The section is divided into two parts covering crop 150 Threshing and livestock farming, respectively. However, it is important to remember that most households both grow crops and keep livestock. Harvesting 100 Crops Weeding The great majority of crops are grown on dryland

fi elds, which means that they are not irrigated. This (millimetres) Rainfall Ploughing and planting 50 also means that crop production is entirely dependent upon rainfall. Fields are often cleared before the fi rst Clearing fields rains, but all other events during the crop calender (Figure 34) follow the onset of the rains, generally 0 Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug

29 Percentages of households planting different crops in four zones. Figure 35 Areas coloured green are those in which there are few or no crops 7.

Riverine 100

80

60

40

20

Western Inland of households Percentage 0 100 Maize Mahangu Sorghum Groundnuts Vegetables

80

60

40

20

Percentage of households Percentage 0 Maize Mahangu Sorghum Groundnuts Vegetables

Eastern Inland 100

80

60

Central Inland 40 100 20 80 Percentage of households Percentage 0 Maize Mahangu Sorghum Groundnuts Vegetables 60

40

20

Percentage of households Percentage 0 Maize Mahangu Sorghum Groundnuts Vegetables

30 Fields are concentrated along the river, in inter-dunes valleys Labour is the most valuable input to crop growing, and most labour is provided by members of the and in the dry omuramba valleys (see Figure 48 on page 42). family. However, it is common practice to hire labourers or to exchange labour between households, Each household is allocated land to cultivate by the local especially during busy periods of weeding. Women provide most of the work (about 62% of worked headman in consultation with the village inhabitants. New fi elds days), while men provide 33% and children under 15 years old and older people above sixty years are cleared on an on-going basis, both by new households old contribute 5% of worked days14. Women are generally more engaged in cultivation and men and as a result of shifting cultivation. A survey in 1992/93 are more involved in the clearing and preparation of land. Hired labour consists either of groups of showed that 48% of all households had cleared new land workers or individuals, and hired labour is usually paid in-kind, for example with mahangu, meat or within fi ve years, and 33% of households had given up fi elds beer. Most hired labour is provided by women. and started new ones within the past fi ve years14. Weeding, as the most time consuming component of crop production, is done two or three times Different surveys have produced quite different average areas during the season. Harvesting starts around April and is done manually. Threshing is done by cultivated by each household, with the highest estimate of 6.7 groups of women. hectares being more than four times bigger than the lowest of 1.4 hectares15. Nevertheless, it is clear that sizes of cultivated Draft power is used to prepare most fi elds, and almost all other work to produce crops is done areas vary in relation to several factors. First, they are related manually. Although only half of all households have their own ploughs, 85% of households have their to rainfall, and the survey in 1992/93 (a bad rainy season) fi elds ploughed using ploughs drawn by oxen or donkeys. Ploughs are thus frequently borrowed showed that while 70% of households intended to cultivate or hired, and 30-40% of all households report hiring ploughs. Households that do not use ploughs more than one plot, only 44% ended up using more than one with draft power either use hoes or tractors for ploughing. plot14. Second, inland households cultivate about 30% more than those along the river. Third, male-headed households The ownership of oxen and ploughs varies in relation to a number of factors (Table 9), so that more cultivate 25% more than female-headed households. Fourth, households in inland areas have ploughs and oxen than those along the river. More male headed cultivated areas vary in relation to a household’s assets and households have ploughs and oxen than female headed-ones. Households with wage incomes are wealth. Thus, homes with cash incomes cultivate areas 25% more likely to have oxen and ploughs than those with smaller or no cash incomes. bigger than those lacking any cash income, and families having their own oxen or plough cultivate double the area Table 9: Percentages of households having oxen and ploughs16. of those having no draft power and equipment. Similarly, cultivated areas vary in relation to the number of livestock as Zone: Own plough Own oxen shown in the table below. Inland 61 58 Riverine 45 40 Table 8: Average numbers of cattle per household compared 16 with areas cultivated by the same households . Head of household: Male 51 47 Area cultivated (hectares) Average number of cattle Female 41 38 Less than 1 5 1-3 10 Cash incomes: 3-5 16 No income 40 38 5-7 22 Remittances 49 43 7-9 25 Pensions 53 48 More than 9 25 Wages 60 56

31 The use of chemical fertilizers and compost is relatively late in the season, and in many years there Changes in total numbers of cattle, goats and Figure 37 very limited. Along the river, only 2% and 8% are good chances of long periods of hot and dry donkeys over the past nine years 17. of households apply fertilizers and compost, weather occurring during the growing period. Many Cattle, rate of increase 4% respectively. The use of fertilizers on inland fi elds crops, especially of younger plants, then wither and 150,000 is non-existent, while 8% of all households report die. Little good information is available on yields, but using compost16. average fi gures of 100 to 300 kilograms of mahangu 120,000 per hectare are probably harvested regularly. Other 90,000 Seeds for planting mahangu are generally selected than a lack of rain, attacks by pest insects and birds from the harvest of the previous year. This is are the main threat to cereal harvests. 60,000 particularly true for families living at the riverside where 45% of all households use traditional seeds. 30,000 About one third of all households also plant improved Percentages of households Figure 36 planting different types of seeds (Okashana-1) and 28% use a mixture of mahangu seed 16. 0 both traditional and improved seeds (Figure 36). Goats, rate of increase 8% Okashana-1 seeds are usually planted late in the 75,000 season as they have a shorter growing period.

60,000 Mahangu, maize and sorghum grain are stored Mixed 28% in different types of containers. Threshed grain is Traditional 45,000 41% stored in 50-70 kilogram bags, traditional shiietes or in 200 litre sealed drums. About 20% of households 30,000 Improved store mahangu without threshing and the heads are Number of animals gathered in bundles of stems to be stored inside 31% 15,000 or outside the house. Households do not mix the 0 harvest of the previous year with the current one. Mahangu is stored for an average of two years, and Donkeys, rate of increase 6% most households report no losses during storage. 1000 Livestock 800 Crop production in the region is evidently a risky Even though some households own donkeys, pigs activity. While some farmers may sell surpluses and sheep, livestock farming is really dominated 600 in some years (see page 37), it seems likely that by cattle and goats. In 1998 a total of about domestic production is often too low to meet all 125 000 cattle was estimated to be in the region, 400 the cereal needs of households. The greatest factor and by 2000 the number would probably have risen affecting yields is rainfall. Good years are those to at least 135 000. This is based on an annual 200 during which rainfall is both suffi cient in quantity rate of increase of 4%, being the rate at which 0 and well-timed, such that there are regular falls cattle numbers increased over the past nine years 1990 1991 1992 1993 1994 1995 1996 1997 1998 of productive rain throughout the growing period (Figure 37). Earlier fi gures are hard to fi nd and (see page 12). But the unpredictable nature of probably often unreliable, but a total of 26 400 cattle rainfall means that productive rains often only start was reported in 1938.

32 The number of goats in the region was estimated cattle, and there are quite a number of farmers Relationships between household size and cattle to be about 64 000 in 1998, having doubled and with herds of 100 and more cattle. For the region ownership. Greater percentages of bigger Figure 38 increased at an annual rate of growth of 8% over as a whole, 12% of all households have 30 or households have cattle, while smaller households 7 the past nine years. A total population of less more cattle and these farmers jointly own about are less likely to be cattle owners . than 1 000 donkeys was reported in 1998. Taking 65% of all cattle. 30 cattle and donkeys as single large stock units and Cattle owners 25 eight goats as equivalent to one such unit, cattle Secondly, livestock ownership is related to the No cattle represent 93% of all large stock units in the region, main source of income of the head of the goats 6% and donkeys less than 1%. household. Thus, those with wages have about 20 double the number of livestock compared to those The average number of cattle and goats per that have no income, as shown in Table 11. 15 household is 15 and 11, respectively7. These averages include the 40% of households that have Table 11: Average numbers of cattle and goats 10 no cattle and the 48% that do not have goats. in relation to the head of the household’s main That almost half of all households have no cattle or source of cash income16. of households Percentage 5 goats is indicative of the considerable variation in Income source Cattle Goats 0 livestock ownership (Table 10). 1-2 3-4 5-6 7-8 9-10 11-12 13-15 >15 No income 7 4 Number of people per household Table 10: Percentages of households in the whole Pensions 7 5 region owning different numbers of livestock7. Remittances 10 6 Wages 14 8 Percentage of households Herd size Cattle Goats (Note, these fi gures come from the three agricultural surveys 16. The averages are lower than those reported Average herd sizes of cattle and goats in None 40 48 Figure 39 7 above (perhaps because of a smaller sample, or different households of different sizes . 1-5 12 13 method of requesting information), but the trend in owner- 35 6-10 11 10 ship pattern is clear) 11-15 8 8 30 Cattle 16-20 6 4 Goats 21-25 5 3 25 Thirdly, large households are more likely to 26-30 6 3 More than 30 12 11 be cattle owners than those with fewer family 20 members (Figure 38). In addition, bigger households also have larger herds (Figure 39). 15 That variation relates to a number of factors. Firstly, Finally, ownership varies in relation to the gender 10 patterns of ownership vary in the different zones of the head of the household, male-headed

in the region, with only about 23% of households homes having about 30% more cattle and goats of livestock number Average 5 in the western inland not having cattle compared on average than those headed by females. with about 50% of those along the river (Figure 40). Surprisingly, there is little difference in herd sizes 0 1-2 3-4 5-6 7-8 9-10 11-12 13-15 >15 Average herd sizes along the river are 9 cattle and between male and female-headed households Number of people per household 9 goats, compared with 26 cattle and 16 goats in that own livestock, so the 30% difference is the western inland. Many farmers in the western largely due to the fact that more female-headed inland also have large herds of more than 30 households do not own livestock.

33 Percentages of households in the four zones that own Figure 40 different numbers of cattle and goats. Areas shaded green are those in which there are few or no livestock 7. Riverine 75 Cattle 60 Goats 45

30

15 Percentage of households Percentage Western Inland 75 0 0 1-5Number6-10 of11-15 livestock16-20 21-25 26-30 >30 60

45

30

15

Percentage of households Percentage 0 0 1-5 6-10 11-15 16-20 21-25 26-30 >30 Number of livestock 75 Eastern Inland

60

45

30 75 Central Inland 15 60 Percentage of households Percentage 0 45 0 1-5 6-10 11-15 16-20 21-25 26-30 >30 Number of livestock 30

15

Percentage of households Percentage 0 0 1-5 6-10 11-15 16-20 21-25 26-30 >30 Number of livestock

34 Figures 41 and 42 provide perspectives on the This is because cattle are moved quite widely, cattle per square kilometre often lead to overgrazing density of cattle and goats, respectively. These maps often going out some distance from households and and damage to pastures, especially if these stocking largely refl ect the distribution of people (see Figure then returning in the evening to drink at the river. rates continue throughout the year. Measurements 21 on page 18), with the great majority of goats Nevertheless, concentrations of animals along the of areas with 10 or more cattle in Figure 41 indicate concentrated along the river. Cattle are perhaps river, in western Kavango, along the Mururani-Rundu that about 7 900 square kilometres or 16% of the more widely distributed than the map suggests, and road, and along the omurambas are noteworthy. It region has such high stocking rates. densities may be slightly greater in inland areas. is generally accepted that densities of more than 10

Estimated densities of cattle, shown as the Figure 41 number of animals per square kilometre 18.

N Nkurenkuru N Mpungu NNankudu Musese N Okavango Rundu MashareN N NNdonga N N Muparara Omega Gcaruhwa N Andara N N Bagani N Ncamagoro N Hamoye

NcauteN

N Khaudum Cattle per square kilometre N Katjinakatji Less than 1 1 - 5 Mangetti N 5 - 10 N Mururani 10 - 20 20 - 30 Scale 1:2,000,000 N Karukuwisa More than 30

0 50 100 o ak at m O N Sikereti Kilometres

35 Estimated densities of goats, shown as the number of animals per Figure 42 square kilometre. The fi gures are in groups of eight because eight goats are taken as equivalent to one large stock unit or one cow 18.

N Nkurenkuru N Mpungu NNankudu Musese N Okavango Rundu MashareN N NNdonga N N Muparara Omega Gcaruhwa N Andara N N Bagani

N Ncamagoro N Hamoye

NcauteN

N Khaudum Goats per square kilometre N Katjinakatji Less than 8 8 - 16 Mangetti N 16 - 24 N Mururani 24 - 32 32 - 40 Scale 1:2,000,000 N Karukuwisa More than 40

0 50 100 o ak at m O N Sikereti Kilometres

36 Livelihoods

Different surveys treat income sources in various heads are older men that have not managed to fi nd Most of the harvest of mahangu, maize and other ways. While the commonest method is to record the jobs in the town. Other important sources of income crops is used for domestic consumption, as are milk main source of income of the head of the household, are cash wages (31% in Rundu and 19% in rural and meat from cattle and goats. But some households it is clear that households obtain incomes from a households), business activities (13% in Rundu) and sell the remainder of harvests and livestock to add variety of sources. Some are “in-kind” or material pensions (6% of heads of households in the region to their sources of income. These sales probably incomes, such as mahangu harvests or milk from as a whole). contribute little to most households, since 62% of all cattle kept at home, while others are cash incomes, homes do not sell any farm products. Cattle sales for example from wages, trading activities, pensions As might be expected, most households do not have are perhaps a more important source of income or remittances. Many households thus have several any wage earners (Figure 43). About 25% of rural than crops, since 32% of households trade cattle different incomes, and individual people often have homes have a single wage earner, and another 15% compared with only 11% that sell mahangu. This different incomes as well. All this makes it diffi cult to have two or three family members earning salaries. would at least be true of those households that have assess the value of different activities to household In Rundu, over 40% of homes have one wage both cattle and mahangu, but cash incomes from economies, and these issues should be borne in earner and about 20% have two or more people with crops sales could be signifi cant for those that do not mind when considering the information provided salaries. own cattle. below. The great majority of households (87%) do not The most obvious reason for these low percentages Table12: Percentages of heads of households have any member of the family having a business is that most households have little or no surplus reported as having different main sources of activity. But among the remaining homes that do produce to sell. But access to customers is probably income19. have someone in business, by far the commonest also an important factor, since more households kind of business activity is informal retail trade, as along the river and the main Mururani-Rundu road reported by 88% of such households. sell produce than elsewhere (Figures 44 and 45). Income source Rundu Rural Subsistence farming 54 71 70 Wages and salaries 31 19 Urban households Business activities 13 0 60 Pensions 3 7 Rural households Cash remittances 0 2 50 Figure 43 40 Numbers of wage earners in rural and Rundu households 7. About three quarters of household heads in rural 30 areas report subsistence farming as their main 20 sources of income. Surprisingly, over half of heads of households in Rundu also have farming as their main of households Percentage 10 incomes. This may be due to these urban dwellers having farms just outside the town or perhaps 0 None 1 2 3 More than 3 elsewhere. In addition, farming may be reported as Number of wage earners per household the main source of income because many household

37 Percentages of households in four zones that report selling various amounts Figure 44 of cattle 7. Areas in green are those in which there are few or no cattle. Riverine

Many

Some Don't sell

Few Western Inland

Many

Some Don't sell

Few

EasternSome ManyInland

Few

Don't sell

Many Central Inland Some

Don't sell Few

38 Percentages of households in four zones that report selling different crops 7. Figure 45 Areas in green are those in which there are few or no cattle.

Riverine 25

20

15

10

5 Western Inland

Percentage of households Percentage 0 25 Maize Mahangu Sorghum Groundnuts Vegetables

20

15

10

5

Percentage of households Percentage 0 Maize Mahangu Sorghum Groundnuts Vegetables Eastern Inland 25

20

15 Central Inland 10 25 5 20

Percentage of households Percentage 0 Maize Mahangu Sorghum Groundnuts Vegetables 15

10

5

Percentage of households Percentage 0 Maize Mahangu Sorghum Groundnuts Vegetables

39 Food produced at home or fi sh caught in the The welfare of a household depends both on income In rural areas, 48% of households use unsafe river contribute signifi cantly to the food needs of a and its access to resources: capital resources (such sources (such as the river) to obtain water for household. However, this is not true for all homes as oxen, ploughs, livestock) and natural resources drinking and other domestic purposes. By contrast, because a good proportion of homes also rely (grazing, fi elds and fi sh, for example). Patterns of almost all residents in Rundu have access to entirely on cereals, meat and fi sh that they have ownership and access to some of these assets are safe water provided through pipes and taps. Ten been bought or perhaps bartered. Thus, about 28% presented elsewhere. But it is useful to summarize percent of households in rural areas have to walk of households do not produce their cereal needs, some key aspects which, as is true for incomes, for more than one hour to the nearest source of 50% do not produce the meat they eat, and 59% show that there is a high degree of variation between water19. do not catch the fi sh they consume (Figure 46). households. About 60% of households have cattle By contrast, about 13% of households produce all and some 48% have goats, leaving 40% and 52% the cereals they eat, while 7 and 10%, respectively, without these livestock. Poultry is owned by 71% of produce all their meat and fi sh requirements. In households. Some 79% of households have fi elds, between these extremes are those families that both while 12% use fi elds belonging to other people. Only produce and purchase some of the cereals, meat half of all households have their own ploughs, and and fi sh that they eat. about 32% of people living near the river catch fi sh.

Sources of food, showing percentages of households that largely eat food that they did not produce at home, those that consume a mix of home Figure 46 produce and food obtained elsewhere, and those that rely entirely on food that they produce themselves 19.

Home Home Home produce produce produce

Not home produced Not home Not home Mixed produced Mixed produced Mixed

Cereal Meat Fish

40 Pressures on the environment

year (Figure 47). Most of that increase probably Areas cleared for crop farming in 1972 and People place many demands on the natural Figure 47 1996. Parts shown in green are those in environment, and some of these lead to degradation occurred after the construction of the main Mururani- which there are few or no fi elds 21. of the region’s natural resources. About a third of all Rundu road. In the western and eastern inland areas 40 households along the river are reported as catching land was cleared at an annual rate of 5% and 6%, 35 fi sh in the river on a regular basis, and the effects respectively, between 1972 and 1996 (Table 13). 1972 30 of 20-30 thousand people20 harvesting fi sh along a 1996 stretch of river about 400 kilometres long may affect Most of the inland fi elds in 1972 were relatively close 25 their abundance and diversity. About 92% of homes to the river and people who chose to move inland 20 are built of wood, and locally-harvested fi rewood probably selected the closest and most accessible 15 is used for cooking by 97% of households. The places. Over the years, suitable places close to 10 large commercial timber industry for the production the river would have been become more and more of land cleared Percentage 5 of furniture has now come to an end, but great limited, thus forcing new settlers to move even further 0 Central Western Eastern Riverine numbers of kiaat and false mopane trees were felled inland. That would explain why there are fewer Inland Inland Inland for timber. These and other trees are also used cleared areas further away from the river. for the production of craft and curios. There are confl icts between people and wildlife, especially in Areas cleared for cultivation along the river increased and around Mahangu and the Caprivi Strip. Some of at only 3% per year, such that the proportion of the these involve the poaching of wildlife, while lions and whole riverine area enlarged from about 20% in 1972 elephants damaging livestock and crops cause other to almost 40% in 1996 (Figure 47). Since many parts confl icts. of the riverine area consist of areas not suited to crops, for example, fl oodplains that are often fl ooded, All of these pressures are important to a greater or reed beds, patches of very sandy soils, and the river lesser degree, but the clearing of land for cultivation itself, it seems certain that little land now remains and burning perhaps have the greatest impacts on available for the clearing of new fi elds. the region’s natural resources.

The clearing of land to plant crops has obvious Table 13: Areas of the four zones that had been cleared immediate benefi ts in providing people with food for crop farming in 1972 and 1996. and income. Figure 48 shows areas that had been cleared for cultivation in 1972 and those cleared in Cleared land (hectares) 1996. Measurements of the sizes of these areas indicate that about 70 000 hectares had been cleared Region 1972 1996 Annual rate of clearing in 1972. By 1996 that area had grown to over Central Inland 209 11 817 18% 192 000 hectares at an annual rate of increase of Western Inland 14 607 51 214 5% about 4%. The greatest rate of increase was in the Eastern Inland 7 897 34 258 6% inland areas, especially in the central inland zone Riverine 47 805 94 684 3% Total 70 518 191 973 4% where cleared land increased at a rate of 18% per

41 Areas cleared for crop farming, shown Figure 48 in orange, in 1972 and 1996 21.

42 Satellite images received at the Etosha Ecological Research Unit in Okaukuejo were processed to detect Areas (in red) that were burnt Figure 49 in 1996, 1997, 1998 and 1999 areas that had been burnt during four years: 1996 to in and around eastern 1999. While results are only available for the eastern Kavango. half of the region, it is clear that very large parts of the region burn each year (Figure 49). The fi res occur throughout the winter months and the great majority are set by people. Some are probably started deliberately to stimulate the growth of new pastures, while others run-away accidentally when farmers burn small areas to clear land for cultivation or to remove vegetation from waterholes, for example. The apparent absence of fi res around the upper reaches of the Okavango Swamps is probably an artefact of the methods used to detect fi re scars, and many more fi res actually burn in that area than Table 14: Estimates of the the maps suggest. proportion of the rectangular block that burnt in each year. While no quantitative studies have apparently been done to assess the effects of fi res in the region, it is clear that such frequent and widespread burning has Year Percentage burnt several impacts. First, important grazing pastures are lost, especially when fi res rage over large areas. Second, 1996 45 many young trees are killed and there are almost no young trees of several valuable timber species in some 1997 53 areas. Third, large timber trees (and other species) are 1998 51 either killed directly by fi res or their trunks are gradually 1999 65 burnt away over several years until they die or fall over. Fourth, large areas become bush-encroached by shrubs of Terminalia sericea, Dichrostachys cinerea and Combretum species. One effect of this is that grazing for wildlife and livestock is reduced. Finally, domestic animals and wildlife may be burnt and killed by fi erce fi res.

There are few natural barriers to fi res in the region and in the absence of fi rebreaks it is clear that fi res can spread rapidly and widely. Figures in Table 14 suggest that at least half, if not more of the region burns each year, so the loss of half of the region’s pastures for part of each year must have a substantial impact on livestock farming.

43 Figure 50 provides a perspective on pressures of people and livestock are found. However, there is All of these areas are likely to have relatively little on natural resources in the region. The index of quite a measure of variation from place to place, and natural vegetation as a result of heavy grazing pressure was obtained by adding together densities demands for natural resources are relatively low pressures, extensive clearing of land for crops, of people, cattle and goats since these are the major in some stretches of the riverine belt. Other areas and the collection of wood to build homes and for consumers of resources such as natural vegetation, where pressures are high are along the omurambas cooking. water, fi sh and wildlife. By far the greatest pressures and the main road from Mururani to Rundu, and are along the river where the highest concentrations in a broad area between Mpungu and Musese.

A measure of pressure on natural Figure 50 resources in the Kavango 22.

44 Acknowledgments, notes and sources

1. This work was done using a variety of sets of data that 6. Based on results from InterConsult. 2001. Natural resource 16. Based on analyses of the Annual Agricultural Surveys done were kindly made available to us and we are especially mapping of the Kavango. Report for the Directorate of in 1996/97, 1997/98 and 1998/99 by the Central Statistics grateful to all the organizations that collected these Environmental Affairs, Windhoek. Offi ce. data: Central Statistics Offi ce, Directorate of Environmental 7. Based on analyses of the Population Survey of the 17. From the annual reports of the Directorate of Veterinary Affairs, Directorate of Surveys and Mapping, Etosha Eco- Kavango Region by Lux-Development in 1999. This survey Services. logical Research Institute, Ministry of Agriculture, Water delimited a number of enumeration areas, which we 18. Cattle and goat densities were estimated by fi rst calculating and Rural Development, Ministry of Basic Education, grouped into four zones to analyse possible differences average numbers of animals per household from the Lux- Sport and Culture, Ministry of Health and Social Services, between households and farming practices in different Development 1999 population survey data in four zones. and the Namibian Meteorological Services. The role of areas. Annual Agricultural Surveys used different enumera- These averages were then attached to point data for all Lux-Development in supporting so many data collecting tion areas, which could only be grouped into riverine and rural households (as available from the 1996 Directorate of and mapping activities in Kavango needs to be specially inland areas. Surveys and Mapping 1:50 000 mapping project). Densities acknowledged. We are also thank Yves Baudot, Chris 8. Based on analyses of the 1991 Population and Housing of cattle were then estimated by “spreading” the number Brown, Brian Jones, Mbaye Kébé, Nils Odendaal, Sylvia Census. at each household over 10 kilometers. For goats, the corre- Thompson, Tom Tolmay and Alvis Weerasinghe for their 9. Demographic information has been derived from several sponding area was 5 kilometres. ArcView’s Spatial Analyst help during the compilation of this document. sets of data: population censuses in 1960, 1970, 1981 Density function was used with the kernel method. and 1991, the 1996 InterCensal Demographic, and Lux- 19. Analyses of data collected by the 1994 Income and Expen- All the work that went into producing this book took place Development surveys of Rundu (1998) and rural Kavango diture Survey by the Central Statistics Offi ce. in Windhoek, and budget constraints meant that it had to (1999). 20. We estimate the number of people along the river as about be done within a very short time. The book is thus limited 10. Population densities were estimated by linking to each 67 000 with another 45 000 in Rundu. The sum of these by two important factors that need to be borne in mind. The household the average number of people per household in numbers can be multiplied by 32% as the proportion of fi rst is that it was not possible to spend time in the region so each enumeration area, as analysed from the 1991 Popula- people who fi sh along the river estimated by Van der Waal, as to obtain a clear understanding of issues and processes tion and Housing Census data. Positions of households BCW. 1991. A survey of the fi sheries in Kavango, Namibia. that make Kavango the way it is. Second, there was little were mapped by the 1996 Directorate of Surveys and Madoqua 17: 113-122. time to review the many publications on the region. These Mapping 1:50 000 mapping project and by the Caprivi 21. Areas cleared in 1972 were derived from the 1:50 000 limitations should be corrected when a complete profi le of Environmental Profi le project (Mendelsohn, JM & Roberts, maps produced on the basis of aerial photographs taken the region is compiled. CS. 1997. An environmental profi le and atlas of Caprivi. in 1972, while those for 1996 were obtained from aerial 2. Estimates of numbers of people in each constituency were Directorate of Environmental Affairs, Windhoek). Densities photographs taken in 1996, as available from the Director- obtained by overlaying their borders with a grid set of data were then estimated by “spreading” the number of people ate of Surveys and Mapping. on population density. at each household over 1 kilometer. Finally, the estimated 22. Units of pressure were derived by adding together the 3. Much of the information on the Okavango’s drainage density was increased by a growth factor by 6.3% to number of people, cattle and goats per square kilo- was gleaned from Crerar, S. 1997. The hydrology of the account for population growth between 1996 and 2000. metre, except that the number of goats was divided Okavango River upstream of the Okavango Delta. In Feasi- The growth factor was based on population projections by eight to obtain an assumed fi gure of eight goats cor- bility study on the Okavango River to Grootfontein link of compiled by the National Planning Commission. responding to one large stock unit. The fi ve categories are the Eastern National Water Carrier. Department of Water 11. Based on Sentinel Surveys by the Ministry of Health and as follows: very low = less than 1; low = 1-10; medium = Affairs, Ministry of Agriculture, Water and Rural Develop- Social Services. 10-40; high = 40-80; very high = more than 80 units. ment. The graphs of water fl ows were derived from analy- 12. Based on data collected by the Annual Education Census ses of hydrological data kindly provided by Sylvia Thomp- of the Ministry of Basic Education, Sport and Culture son in the Department of Water Affairs. 13. Analysis of the 1996 InterCensal Demographic Survey by 4. Information on underground water was largely summarized the Central Statistics Offi ce. from Namibian Groundwater Development Consultants, 14. Survey by Keyler, S. 1995. Economics of the pearl millet 1991. Groundwater investigations in Kavango and Bush- subsector in northern Namibia. A summary of baseline manland. Report for Ministry of Agriculture, Water and data. International Crops Research Institute for the Semi- Rural Development. Arid Tropics. Working Paper 95/03. 5. Graphs of numbers of days with different amounts of rainfall 15. Results from Annual Agricultural Surveys done in 1996/97, were derived from analyses of the daily rainfall records. 1997/98 and 1998/99 give estimates of average cultivated Figures from Rupara were used to fi ll in missing data over areas of between 2 and 3 hectares per household, whereas a number of years at Nkurenkuru (Rupara is approximately other surveys generally report averages of between 4 and 60 kilometres south-east of Nkurenkuru). 7 hectares.

45