Water Conflict and Cooperation/Incomati River Basin

From WaterWiki.net

Jump to: navigation, search
edit  ·  ToolkitWater Conflict and Cooperation
History of Humanity and Water | Shared Waters and Ethics | Legal and Institutional Approaches | Alternative Dispute Resolution Mechanisms
Emerging Trends | Public Participation | Indicators | Means and Tools | Basin Management | Lessons Learned
Organisations and Projects | Publications | Glossary | Bibliography
Case Studies: Incomati | Nile | Columbia | Lempa | Lake Titicaca | Mekong | Aral Sea | Rhine | Jordan I | Jordan II
Events | Contacts

This article is based on Sharing the Incomati Waters: Cooperation and Competition in the Balance, by Alavaro Carmo Vas and Pieter van der Zaag as part of a UNESCO-IHP, PCCP Series Publication (2003).



The Incomati river basin is situated in the southeastern corner of the African continent.[1] The basin is shared between South Africa, Swaziland, and Mozambique, and discharges in Maputo Bay in the Indian Ocean. Two adjacent basins, the Umbeluzi and Maputo basins, also discharge in Maputo Bay, and are also shared by the three countries.[2] In fact, the three basins cover the entire territory of Swaziland (see Figure 2 in original publication).

The Incomati basin in Southern Africa
The Incomati basin in Southern Africa

The Incomati basin is small compared to some other international rivers. Yet its history of water sharing commands attention for two main reasons:

1) Water use is intense, with 50 percent of the water generated in the basin being withdrawn. Water scarcity has been evident since the mid-1980s, and has become more severe in the last decade. Competition over water is real, and water abstractions are fast approaching the limits of sustainability. The effects of droughts, but also floods, become ever more pronounced.

2) The basin is situated in a part of Africa that over the last forty years has experienced a dynamic, sometimes turbulent and volatile, political history. The basin has seen a change from colonial rule to independent one-party states and subsequently a transition to democratic rule as well as the end of the apartheid regime. These changes were frequently accompanied by violent interventions.

Either ingredient might have been sufficient for the emergence of confrontations over water. Yet these did not occur. Tensions between Mozambique, South Africa, and Swaziland over Incomati waters existed but never escalated. This is the major interest of the case of the Incomati: the tensions translated into agreements and deepened the level of cooperation between the riparian countries. This is not to say that the road towards cooperation was easy and self-evident. It was a difficult and often frustrating process. But the realization by the three riparian countries that they shared a common space, a common though checkered, history, and hopefully a common future may explain their deep-seated commitment to proceed with seeking negotiated arrangements over the water resources of the Incomati. The new water sharing agreement, signed by the three water ministers during the World Summit on Sustainable Developmentin Johannesburg in August 2002, demonstrates this.

This case study attempts to find an answer to the following question: Why did open conflict not emerge between the riparian countries over the water resources of the Incomati, and why did cooperation prevail?

In order to answer this question, the case study presents, in separate sections, information about the natural characteristics of the basin, its political history, socioeconomic developments, and the existing legal and institutional framework. A subsequent section provides detailed empirical information about the negotiations that took place during the period 1967–2002. The concluding section attempts to answer the central question raised.

The Incomati case shows that water can indeed drive peoples and countries towards cooperation. Increased water use has led to intensified cooperation. However, as the Incomati basin approaches closure, water sharing will increasingly be a delicate balancing act between cooperation and competition.

Natural Characteristics

See Incomati/ Water Basin Profile: Physical and Hydrological Characteristics for detailed description

A Political-Historical Overview

Prior to 1948: The Colonial Experience

The Incomati basin is not only a system of watercourses that are interconnected; the basin also encompasses a space that is of significance in terms of culture, history, politics, and economy, comprising parts of southern Mozambique, northern Swaziland, and eastern South Africa, as well as Maputo Bay (formerly known as Delagoa Bay) in the Indian Ocean.

According to Newitt (1995) the bay has deeply influenced early developments in southeastern Africa, as people settled round its shores and along its rivers. Its people engaged in activities such as fishing, hunting whales, trading ivory, and maintaining an economy of cattle keeping and agriculture. The influence of the bay stretched across the Lebombo mountains to the interior, for instance through the trade of copper. The first Europeans to arrive were the Portuguese, who for some 200 years focused on the ivory trade. Most ivory had to be brought from considerable distances, and deepened and expanded the long-distance commercial networks to the hinterland (Newitt, 1995).

In 1721, a Dutch expedition build a fort in what was to become the city of Lourenço Marques, now the city of Maputo. Fifty-seven years later an Austrian settlement was created, and after 200 years of contact the Portuguese established their first permanent trading station there in 1781. Towards the end of the eighteenth century the bay was a hive of activity, with the British having established a post at the southern shores of the Bay. The city of Lourenço Marques was, however, founded only in 1790. In the nineteenth century, it became “the major trading port for the Swazi and Zulus, for the Tsonga chieftainships to the north and, after the Great Trek, for the Transvaal Boers as well” (Newitt, 1995, p. 293). The Boers first settled in the former Transvaal (now Mpumalanga) in the late 1830s.

By the 1850s, however, a hunting and slaving frontier community, barely controlled by the Republican government, had established itself in the Zoutpansberg and begun to form links with the Portuguese traders and ivory hunters from Inhambane and Lourenço Marques. . . . For two decades the region north-west of the Bay saw a mixed community of Boer and Afro- Portuguese traders and elephant hunters, cooperating to a greater or lesser extent with local chiefs . . . in developing a trade in ivory, slaves and cattle which made use of the Portuguese ports. (Newitt, 1995, p. 327)

Both Britain and the Transvaal Boers claimed Delagoa Bay (in 1861 and 1868, respectively). Portugal protested, and in 1869 the Boers and the Portuguese signed a treaty that recognized Portugal’s possession of the Bay, drew the eastern frontier of Transvaal along the ridge of the Lebombo Mountains, and agreed to build a road linking the highveld and the port. In 1875 Britain recognized Portuguese control over the Bay.

The discovery of important minerals in the Transvaal area during the second half of the nineteenth century reinforced the emerging regional economy. The enormous mine developments required an intensive support network in terms of transport, communication, agriculture, trade, and labor. Lourenço Marques with its harbor was strategically positioned. The construction of a railway connecting the city to Johannesburg, and passing through Ressano Garcia where the Komati and Crocodile rivers join to form the Incomati, was completed in 1895 (Libby, 1987).

Labor movement intensified after 1850, when migrant workers from southern Mozambique traveled via Johannesburg to Durban to work in the sugar industry. The discovery of diamonds in Kimberley in 1869 and gold in the Rand area (Gauteng) and near Barberton in 1886 further intensified this traffic of human labor (Nyathi, 1977).

From the middle of the nineteenth century the economy of southern Mozambique was thus dominated by its link with South Africa. This link was formalized by the signing of three treaties between Portugal and South Africa, including one on free trade (signed in 1875) and one on traffic with the Transvaal (1895), and legislation concerning the recruitment of mine workers from Mozambique in 1896 (Katzenellenbogen, 1982). In 1901, just after the South African War, the British High Commissioner renegotiated an agreement with the Portuguese to allow recruitment of mineworkers in exchange for direct payment in gold and preference for Delagoa Bay as a port. During 1903–7 southern Mozambique supplied some 50,000 laborers annually, or 60 percent of the total black labor force on the mines. By 1909, Lourenço Marques dominated 65 percent of all trade with the Transvaal, while Durban had 22 percent and the Cape ports only 13 percent (Beinart, 1994).

At the turn of the century the Sabi Game Reserve was proclaimed, encompassing an area between the Crocodile and Sabie rivers that would later form part of the much bigger Kruger National Park (proclaimed in 1926), extending north up to the Limpopo river. Despite the threat of malaria, pockets of European agriculture emerged along the river valleys of the Crocodile river near the towns of Barberton and Nelspruit. White settlement increased rapidly in the lowveld of the South African part of the Incomati basin after the First World War. The major crops cultivated were citrus, cotton, and tobacco (Packard, 2001).

During the same period, irrigation development started in the Incomati plains. A British-owned company established the sugarcane plantation and sugar mill at Xinavane in Mozambique around 1910. In 1914 it was linked by rail via Moamba to the Lourenço Marques–Rand railway. Portuguese industrialists created Maragra, which first cultivated bananas for export to Transvaal, but this venture collapsed because of export restrictions to South Africa.

The Period 1948–74: Economic Development

By 1952, some 250,000 Mozambican workers were staying in South Africa, of whom only 100,000 were there legally. In 1967, the earnings by these Mozambican migrant workers was estimated to be eight times the value of marketed agricultural produce of southern Mozambique (Coles and Cohen, 1977; Murteira, 2000).

The swift decolonization of the continent in the 1960s drove Portugal and South Africa closer together. This is most clearly demonstrated by the construction of the massive Cahora Bassa dam on the Zambezi. This project was a joint venture in which the Anglo American Corporation, as well as Portuguese capital, had a strong interest.

A generous concession was given to Eskom, the South African electricity company, which would buy the electricity generated by the dam (Minter, 1977). This electricity was seen as crucial for the further economic development of the Transvaal (Cliffe, 1976; Minter, 1977). Dam construction started towards the end of the 1960s, and coincided with increased guerrilla activities by Frelimo, the liberation movement that came into power at Mozambique’s independence in 1975. South Africa sent some 1,000 soldiers to protect the construction site (Coker, 1985; Cawthra, 1996).[3]

During this period agricultural development in the Incomati basin increased significantly. In the lowveld of the South African part of the basin, DDT was introduced in 1945 to control malaria (Packard, 2001). This led to the opening up of the area for commercial farming, and the forced removal of black farmers from areas designated for white farmers (Packard, 2001). The KaNgwane “homeland” along the lower Komati was created. Transvaal Suiker Beperk (TSB) started to develop irrigated sugarcane from 1965 onward along the Crocodile river, and constructed Malelane sugar mill. A paper mill was constructed in 1966 in the upper parts of the Crocodile river (Ngodwana). In 1960 Swaziland started to establish sugar plantations along the Umbeluzi river, adjacent to the Komati, with support from the Commonwealth Development Corporation (CDC), in which the King acquired a 50 percent shareholding after the country gained independence in 1968. In the Incomati plain the area cultivated with sugarcane increased both at Xinavane and at Maragra. New Portuguese settlers started to cultivate irrigated rice.

The Period 1975–94: Political Turmoil and Peace

Mozambique and Angola attained independence in 1975 as a result of the collapse of the Portuguese colonial regime in 1974. The Soviet Union and Cuba became directly involved in both countries, while the support of the African National Congress (ANC) within South Africa increased dramatically. These developments forced South Africa to revisit its regional strategy. Its military budget increased by 50 percent in the fiscal year 1974–5, and was further doubled by 1977–8. South Africa invaded Angola in 1975 and started to destabilize Mozambique by supporting Renamo, the resistance movement that competed for power with Frelimo. The killing of Steve Biko in 1977 caused a wave of international indignation, which resulted in the Security Council of the United Nations imposing an arms embargo on South Africa (Davies and O’Meara, 1985; Price, 1990).

In order to address the political crisis, the newly appointed South African Prime Minister, P. W. Botha, forged a strong alliance between financial and military interests, adopting the “Total Strategy” (Davies and O’Meara, 1985; Hanlon, 1986). One offshoot of this strategy was the establishment of the Development Bank of Southern Africa (DBSA), in which the private sector participated. This bank was supposed to support infrastructure developments in allied states, which would serve as a buffer against the frontline states. This strategy failed with the landslide electoral victory in 1980 of Robert Mugabe in Zimbabwe. Zimbabwe became the last country to join the frontline states, which transformed into the SADCC the same year.[4] South Africa remained politically isolated (Davies and O’Meara, 1985; Hanlon, 1986; Price, 1990). However, in terms of trade and commerce, South Africa’s role in the region actually increased: during the 1970s its regional exports doubled. Despite the political rhetoric the frontline states did not cut their commercial ties with South Africa, since they were too dependent on it. Also the electricity supply from Cahora Bassa to South Africa continued, providing 10 percent of Eskom’s electricity requirements (Bienen, 1985; Libby, 1987).

In the meantime South Africa unilaterally decreased the number of Mozambican mineworkers allowed into South Africa from 127,000 in 1975 to 38,000 in 1978, and stopped the payments in gold to Mozambique that were part of the migrant mineworkers’ salaries (Davies and O’Meara, 1985; Green and Thompson, 1986). In accordance with its socialist policies, Mozambique for its part nationalized many industries, including the sugar industry at Xinavane and Maragra. During the 1980s, South Africa increased its destabilizing activities in neighboring countries, and its support to Renamo in Mozambique. The result was that by 1983 the countryside in southern Mozambique had become unsafe. Many rural people had to find refuge around Maputo. At the end of 1983, South Africa signed a secret non-aggression pact with Swaziland. In the midst of the insurgent activities, negotiations started between Mozambique and South Africa, which culminated in the signing of the Komati Agreement in March 1984.

Beinart (1994) explains why South Africa wanted to reach an agreement with Mozambique: it experienced severe economic problems after the gold price declined sharply in 1983; production from its manufacturing industry stagnated, and unemployment increased. This fuelled political turmoil. South Africa’s interest in entering into a new agreement may therefore be seen as a desperate effort to position itself as the key to regional peace, and in so doing regain some confidence of foreign investors.

Mozambique was faced by an even more dramatic collapse of its economy. It wanted to increase the volume of goods ferried through Maputo, to regularize the export of electricity from Cahora Bassa (and to stall the sabotage of electricity lines), and to increase the number of mineworkers allowed into South Africa. New agreements were reached on all three issues (Davies and O’Meara, 1985). Mozambique also sought openings to the West. In 1985 it became a member of the IMF and World Bank, and in 1987 it started constitutional reforms towards multi-party democracy (Hanlon, 1986; Libby, 1987; Minter, 1998; Cabaço, 1991; Zacarias, 1991).

The Komati agreement did not have the intended effect. A further increase in politically motivated violence within South Africa triggered a flight of capital and the rand lost half of its value (Beinart, 1994). Rather than decreasing, Renamo’s destabilizing activities increased and developed into a fully-fledged civil war. The death of Samora Machel in an unexplained plane crash in December 1986 shocked the world. It was only after Nelson Mandela was released from prison in February 1990 that the political atmosphere improved. In the same year the Mozambican government introduced a new constitution that provided for multi-party democracy, and started negotiations with Renamo. This resulted in the Peace Accord signed in October 1992. A UN peacekeeping force arrived in the country in 1993, and after some delays multi-party elections were held in November 1994. This development, together with the attainment of majority rule in South Africa after the elections in April 1994, put the lid on more than a decade of regional violence and heralded a new era of peace and regional cooperation.

During this period, irrigation areas in the South African parts of the Incomati basin increased rapidly, in the context of subsidies to white commercial farmers. Afrikaner capital (the Rupert family with their Rembrandt group of companies, as well as the Development Bank of Southern Africa – DBSA – and its chairman Simon Brand) was instrumental in further developing sugarcane cultivation along the Crocodile and Lomati rivers in South Africa. Swaziland commissioned a second sugar mill on the Umbeluzi. In Mozambique, new irrigation infrastructure was established – for instance, the Sabie–Incomati irrigation scheme (3,500 ha) – but the area actually irrigated dropped because of the civil war.

The Period 1995–2002: Liberalization and Regional Integration

New initiatives and developments in the Incomati basin indicate that political, commercial, and cultural ties across national borders have been intensifying since 1995. The most obvious political development was that South Africa was accepted into the fold of the Southern African Development Community (SADC) in that year. South Africa hosted the SADC summit in August 1995, and three months later also the SADC conference of water ministers.

Because of the new peace, investment of private South African capital in Swaziland and Mozambique rose dramatically. An example was the so-called “Maputo corridor,” which involved the construction of a new highway (toll) between Maputo and the border at Ressano Garcia, improving communications between Gauteng and Maputo. Another massive multinational investment was the construction of an aluminum smelter in the estuary of Maputo Bay, which involved many players, including South African mining interests and cheap energy from Cahora Bassa, supplied through Eskom.

South African sugar business took advantage of the new liberal policies of Mozambique, Tongaat-Hulett and Illovo obtaining equity in the two Mozambican sugar estates in the Incomati, Xinavane, and Maragra, respectively. As a result, the three sugar companies that dominate the South African market now all have interests in the Incomati water resources, the two largest ones in the Mozambican part of the basin. In 1997 the smallest of these three companies, TSB, which operates in South Africa, commissioned a second mill at Komatipoort. Indicative of the new South African political dispensation, TSB boasts that a significant proportion of sugar deliveries (40 percent) comes from small-scale producers.

The commitment to advance the plight of small-scale farmers runs through the recent policies of all three countries. Swaziland developed the Komati Downstream Development Project, which will irrigate 6,000 ha of sugarcane for smallholders from the new Maguga Nkomati Basin dam, a joint venture with South Africa (Mwendera et al., 2002). In the lower Komati and Lomati rivers in South Africa, the Nkomazi Irrigation Expansion Program involves the development of 6,500 ha of irrigated sugarcane for emergent black farmers, drawing water from the Maguga dam in Swaziland and Driekoppies dam in South Africa (Waalewijn, 2002). In Mozambique, the efforts are focused on rehabilitating existing irrigation infrastructure.

An icon of the new era of regional integration is the “peace park” concept, which involves the merging of three National Parks in three countries located in the Incomati and Limpopo river basins, namely Gaza (in Mozambique), Kruger (in South Africa), and Ghonarezhou (in Zimbabwe). The idea was mooted by Anton Rupert, the founder and chairman of the South African chapter of the World Wildlife Fund, who presented his initiative to the Mozambican President Chissano as early as 1991, just after Nelson Mandela was released and Mozambique had adopted its new constitution. Mr Rupert was well acquainted with the area, as he owned a private park adjacent to Kruger, as well as the TSB sugar company. By 2002, the Great Limpopo Transfrontier Park was a fact, the first elephants having been moved across the borders.

In sum, the contrast between the 1980s and the 1990s could hardly have been starker. Developments during the 1990s were characterized by cooperation and economic integration, and a new thrust of economic development. This rosy picture was temporarily disturbed by the floods of February 2000 that devastated southern Mozambique (see Box bellow). The floods triggered immediate assistance by South Africa and a watershed of relief support by the international community, and emphasized once more the need for further regional cooperation.

The floods of February 2000:
Heavy rains, which started in early February 2000, flooded parts of Mozambique’s southern provinces. The Save, Limpopo, Incomati, and Umbeluzi rivers, which have their head-waters in Zimbabwe, Botswana, South Africa, and Swaziland, reached their highest-ever recorded levels in early March, and many riparian communities were submerged for weeks. There were 699 deaths, ninety-five people disappeared, and a million required some form of emergency assistance.

Large sections of the major road connecting Maputo to the north were demolished. Bridges along the Limpopo flood plain and the railroad were damaged. About 20,000 cattle drowned and 140,000 hectares of crops were destroyed, with the largest irrigation scheme in the country (25,000 ha, along the Limpopo) seriously damaged. Health centers as well as water supply and sanitation infrastructure in many towns and villages suffered extensive damage, exposing a million people to water-borne diseases such as cholera, malaria, and diarrhea. The destruction caused by the floods is estimated at $600 million. Mozambique’s economic growth declined from 10 percent in 1999 to 2 percent in 2000.

(Source: Brito, 2002.)

Socioeconomic Developments and Water Use

See Incomati/ Water Basin Profile: Socio-Economic and Environmental Issues for detailed description

Legal and Institutional Framework

See Incomati/ Water Basin Profile: Transboundary Political and Institutional Setting for detailed description

Water-Related Conflict and Cooperation: 1967–2002

The Situation in the 1960s and 1970s: The First Dams and Formal Contacts

The Period up to 1972

The First Tripartite Technical Conference on Rivers of Common Interest between South Africa, Swaziland, and Mozambique (the latter two still under colonial rule) was held in Mbabane, April 24–25 1967, where it was agreed:

  • To adopt the principle of “best joint utilization.”
  • That Swaziland would accede to Part I (Rivers of Mutual Interest) of the 1964 Cunene Agreement (Bipartite Agreement between South Africa and Portugal in regard to Rivers of Mutual Interest and the Cunene River Scheme).
  • The full and free exchange of hydrological data.
  • Studies to be done leading to an integrated plan for each of the shared river basins.
  • Technical studies to be done for an international flood warning system.

The background of this meeting remains unclear, but interest in the water resources of the Incomati was rising sharply in all three countries and unilateral development plans were formulated.[5] Two follow-up conferences were held (in 1969 in Lourenço Marques, and in 1972 in Mbabane), but no progress on the preparatory studies was made. During the 1969 meeting, the three countries agreed that projections for water requirements in the year 2000 had to be made. However, at the meeting of 1972 no data on actual water use were shared. It was therefore agreed that the countries would make available such data by March 1973. During the same meeting, South Africa committed itself not to proceed with further developments in the Incomati basin without first consulting the other two riparian states.[6]

During these initial meetings on the Incomati, Swaziland’s main concern was a guaranteed minimum flow during the dry season in the Komati River. As a trade off it was prepared to forego the floods for storage in upstream South Africa (that is, in Vygeboom dam). Swaziland accepted the recommendations of a technical committee, which set low flow levels to be guaranteed.[7] The meeting of March 1973 never took place. The next tripartite meeting was only held in 1982 (exactly ten years after their last meeting), as the Tripartite Permanent Technical Committee (TPTC). It is likely that the three countries found it difficult to meet during 1972–81 because of developments in Mozambique.

The Period 1972–81

After Mozambique attained independence in 1975, its relation with South Africa quickly deteriorated. At the political level relations between Mozambique and South Africa remained strained, and only improved substantially from 1990 onwards. It is therefore important to acknowledge that the three countries remained on speaking terms with regard to the water resources of the Incomati.

In contrast, the relation between Swaziland and South Africa grew closer, though not always without tension.[8] Between 1978 and 1981, Swaziland and South Africa delegations formally met at least eleven times, negotiating the uses of the Komati River. Of these eleven meetings, seven were of the Joint Permanent Technical Committee (JPTC), the bilateral counterpart of the TPTC. In order to dispel Swaziland’s fear of being pushed into an agreement, South Africa agreed to pay half the cost of an independent consultant to act on Swaziland’s behalf.[9]

During this flurry of meetings and negotiations the first ideas of a joint Komati Basin Development Plan were conceived. In 1992, these ideas would culminate in a treaty on the development and utilization of the water resources of the Komati river basin, the establishment of the Komati Basin Water Authority (KOBWA), and the construction of the Maguga dam in Swaziland and Driekoppies dam in South Africa. But before this treaty was signed, both countries would still meet bilaterally many times (thirty-three times, including those at ministerial level, between 1982–91).

During the period 1976–80, Mozambique and Swaziland held talks on the Umbeluzi waters, on which the city of Maputo depends. This resulted in the signing in 1976 of the “Umbeluzi Agreement,” in which Swaziland committed itself to leaving 40 percent of the flow in the White and Black Umbeluzi in the river for use by Mozambique.[10] When in May 1979 the lowest flow ever in the Umbeluzi at the Mozambican side was recorded because of the filling of the new Mnjoli dam (capacity 153 Mm3) in Swaziland, and Maputo experienced severe shortages, both countries held successful negotiations on respecting this bilateral agreement. After the talks the river flow significantly increased.

Given the political context of the time, it is remarkable that South Africa and Mozambique met twice in 1978 to discuss issues related to shared waters. The first meeting was held on May 8, and was discretely organized under the auspices of the CEO of the South African Railways. A month later, an official delegation of the South African Department of Water Affairs visited Maputo and met a delegation of the Direcção Nacional de Aguas (DNA).

By 1980, Swaziland and South Africa acknowledged that Mozambique was an interested party in the Incomati, and that its interests had to be taken into account. In February 1981, floods occurred in the Save, Limpopo, Incomati, and Umbeluzi rivers. In the southern rivers floods were modest, but in the Limpopo and Save extensive flooding occurred. In a subsequent meeting held in 1981 between representatives of Mozambique and South Africa on Inhaca island (off the coast of Maputo), issues related to flood management were discussed. The local newspaper in Mozambique excitedly spoke of “water diplomats” from South Africa and Mozambique meeting.[11] This must have been quite an event, given that official contacts between both countries were extremely limited.

The few contacts that existed between Mozambique and the other two riparian countries centered on the exchange of hydrological information and plans for future developments. Although hydrological information was shared more or less freely (which proved important during the floods of 1981 and 1984), there was great reluctance to inform each other frankly on actual water use and plans for future development. In addition there was a considerable language gap. Mozambique was Portuguese-speaking and not many nationals had mastered the English language. Expatriates who spoke on behalf of the Mozambican delegation often facilitated negotiations. The Swaziland delegation also made use of expatriates, mainly to bring in expert knowledge.

Another characteristic of the negotiations on the Incomati can be discerned from the above, and would be repeated in the years to come: Swaziland and South Africa held many more bilateral meetings than there were tripartite meetings. This must have made Mozambique feel uneasy, as it was clear that it was negotiating with two partners that, at least during trilateral negotiations, formed a closed front, and talked a common language, literally and figuratively speaking. Swaziland, however, views its role of having facilitated the coming on board of Mozambique in the tripartite talks as crucial.[12]

The Situation in the 1980s: Tensions in the Basin

The Period 1982–4

The year 1982 was a drought year. Many people in southern Mozambique died of starvation. Water levels in the Incomati basin were low. The sugar companies on the Incomati suffered severe crop losses. For the first time it was realized by many that the water resources of the river were limited, and that the steadily increasing upstream uses (Kwena dam on the Crocodile river (155 Mm3) had just been commissioned) directly impacted on downstream users.

When, after a ten-year lull, the fourth tripartite meeting between Mozambique, Swaziland, and South Africa was held in Mbabane on July 6 1982, the Incomati River, as measured at the border between South Africa and Mozambique, only carried a trickle (40 l/s) where it used to carry at least 6 m3/s. It ran completely dry two months later in September 1982, the first time since recording started in 1953. During this meeting the following was agreed:

  • Mozambique recognized the 1964–7 Rivers Agreement, accepting the agreement made by its former colonial power, Portugal.
  • The Tripartite Permanent Technical Committee (TPTC) was formally established (this meeting would be known as the first TPTC meeting, although the formal agreement between the three governments establishing the TPTC was only signed on February 17 1983 during the third TPTC meeting in Pretoria; see Annex 1).
  • In the near future a report would be prepared outlining the water developments in each country during the period 1972 to 1982.
  • In view of present developments, talks should start on the Incomati (rather than on the Maputo river, the only other river of common interest to the three countries).

Less than two months later the second TPTC meeting was held, this time in Maputo (the venue would rotate among the three countries). By now, the drought was a fact. Mozambique expressed its concern about the lack of water in the Incomati during the dry season at the border. South Africa explained that because of upstream developments it was no longer possible to release water during the dry season. Both countries agreed that there was an urgent need to coordinate further plans of water development, and that they would exchange information on studies already undertaken. Reference was made to the Komati river basin development plan being formulated jointly by South Africa and Swaziland.

Four subsequent TPTC meetings were held in 1983 and 1984.[13] The topics discussed during these meetings all centered on the exchange of information. Mozambique asked South Africa to provide discharge data on the Komati and Sabie rivers near its border. South Africa obliged. Mozambique informed the other riparians that construction of the Corumana dam had started.

Knowing what South Africa and Swaziland’s development plans were in the upstream part of the Incomati basin was an important issue for Mozambique. South Africa and Swaziland first submitted short abstracts on their joint Komati basin development plan, but Mozambique wanted to see the full report. This was made available; the draft development plan, as formulated by the JPTC, envisaged the construction of Driekoppies (South Africa) and Maguga (Swaziland) dams. The dams would “stabilize existing water use and allow for modest irrigation expansion.” Mozambique expressed its satisfaction with this draft interim report.

For its part, Mozambique presented, during the fifth TPTC meeting (Maputo February 27–8 1984), its “framework” report on the Incomati, inspired by Article 5 of the Helsinki Rules (which provides a list of all factors to be taken into account when establishing “a reasonable and equitable share” of the water resources of an international basin). South Africa agreed with Mozambique’s suggestion to formulate a single “framework” for the entire Incomati basin. In the next meeting (sixth TPTC meeting, Berg en Dal, August 30 1984), Swaziland and South Africa accepted Mozambique’s Incomati report. But when, during the same meeting, Swaziland and South Africa submitted their joint report on the Komati, Mozambique observed that the format used differed from its “framework” document. Mozambique further observed that similar documents would have to be made for the Sabie and Crocodile rivers. South Africa promised to make available a preliminary report on the Sabie River to Mozambique.

In the corridors of this meeting, the head of the Mozambican delegation mentioned that his country eventually aimed at reaching an agreement on the Incomati basin that would have to address two key issues, namely:

  • No further developments in the South African part of the Sabie River, in order not to harm the Corumana dam (under construction) on the Mozambican side of the Sabie.
  • A guaranteed minimum flow in the Incomati at Komatipoort/Ressano Garcia. [14]

More than six years later, both issues would be addressed in the Piggs Peak agreement.

An unrelated natural event reinforced the need for speedy information exchange: the Demoina flood that hit the lower Incomati in early 1984. During the fifth TPTC meeting (Maputo, February 27–8 1984) the countries noted the difficulty in communicating flood warnings speedily because conventional communication systems failed. Swaziland and South Africa promised to explore satellite communication options. During the sixth TPTC Mozambique presented a report on the 1984 floods, and South Africa committed that in future occasions the South African Broadcasting Corporation (SABC) would transmit messages via TV and radio in case telephone links were destroyed, as well as by satellite communication via Portugal.

The Period 1985–91

From 1985 to 1988 the TPTC did not meet. In the same period, Swaziland and South Africa met at least nineteen times, fourteen times as a JPTC, and five times at ministerial level. Swaziland and South Africa were making progress with formulating their joint Komati basin plan. A small hitch in the negotiations was that Swaziland did not accept financing through the Development Bank of Southern Africa, DBSA.c[15]

South Africa and Mozambique only met at the highest level: first at a meeting at Komatipoort between P. W. Botha, the premier of South Africa, and Samora Machel, the president of Mozambique (March 16 1984), when they signed the Komati Agreement. This agreement focused entirely on immediate security issues, and not on water. In 1985, two meetings were held between the ministers of water (April 17 at Komatipoort, and May 9 in Maputo). From 1985–8, as the security situation in southern Mozambique further deteriorated, the possibilities of Mozambique and South Africa negotiating water issues at technical and political levels were further curtailed.

Despite Mozambique’s political and security problems, and after it had completed the massive Corumana Dam, the TPTC met twice in 1989. The need for such a meeting was possibly reinforced by South Africa and Swaziland’s urgent wish to proceed with the building of two new dams on the Komati River, in accordance with their joint Komati basin plan. The hydrological context was that again the Incomati River fell dry in September that year.

The seventh TPTC meeting (Mbabane, February 21 1989) was characterized by a hardening of positions between the three countries. The eighth TPTC meeting, held one month later, was one of tough negotiations. This was probably when the clashes of interests between South Africa and Swaziland on one hand, and Mozambique on the other, were made explicit for the first time.

During the seventh TPTC meeting, Mozambique acknowledged receipt of the First Phase Komati Basin Development Plan (dated February 1987), but stated that the plan did not take into account its needs. South Africa announced that it intended to construct Driekoppies Dam as soon as possible, and Swaziland stated that it would make a final decision on the construction of the Maguga dam by the end of 1989. South Africa stated that the purpose of Driekoppies and Maguga dams was to utilize floodwater not used by Mozambique. Mozambique retorted that it already experienced severe shortages, and that the JPTC plan itself stated that “serious shortages of water are expected to occur along the Incomati River in Mozambique.” Mozambique further accused South Africa and Swaziland of transferring Komati water out of the basin. Mozambique further denied ever having agreed to the construction of the two dams.

In an effort to resolve the deadlock, South Africa invited Mozambique to define its immediate needs but only along the Incomati upstream of the confluence with the Sabie, since the remaining needs could be met from the recently completed Corumana dam. Mozambique did not accept this suggestion, arguing that the Komati could not be separated from the rest of the basin, and that it required sufficient flow for present and future needs. The meeting finally agreed that South Africa would draft a position paper answering Mozambique’s questions concerning Driekoppies dam, and that Mozambique would make its water needs known.

The eighth TPTCmet only four weeks later in Maputo (March 21 1989). South Africa submitted its position paper, presented as a joint paper with Swaziland under the aegis of the JPTC. Mozambique, for its part, stated that its actual, short-term, medium-term, and potential water requirements were given in the 1984 “Framework” report: requirements that were not considered by the South Africa/Swaziland study. Mozambique further stated that it had never received any formal proposal by South Africa/Swaziland on the construction of the Driekoppies and Maguga dams. It noted that, whereas South Africa alleged that Mozambique had agreed to proposals in the consultant’s interim report, no such agreement was recorded in the minutes of the TPTC meetings. South Africa responded by maintaining that the dams would not substantially reduce the availability of utilizable water in Mozambique. Mozambique emphasized that the dry season flow coming to Mozambique at Ressano Garcia had decreased substantially over the last twenty years. With the proposed plans, upstream water use in South Africa and Swaziland would further increase from about 700 to 900 Mm3/a after implementing Phase One.

South Africa stated that it could not wait for the implementation of Phase One, and it reiterated that it wanted to know Mozambique’s needs between the border and the confluence with the Sabie, as the immediate water requirements on the Lower Incomati could be met from Corumana dam. Mozambique replied that Corumana should not be used to correct a situation caused by increased water use by upstream countries, but to expand its own irrigated area.

Mozambique finally stated that it wished to reach agreement on a division of the Sabie water first, and that it then would like to reach a general agreement on the entire Incomati basin. Mozambique had suggested this earlier during the fifth TPTC meeting. With Corumana dam on the Sabie River in place, it was in Mozambique’s immediate interest to secure an uninterrupted supply. Mozambique further agreed to prepare a paper detailing its water requirements in order to allow the other countries to accommodate its needs.

After the meeting, South Africa formally requested Mozambique’s “no objection” for the construction of Driekoppies dam. Mozambique responded in an official note signed by the Minister of Cooperation, stating some conditions, the most important being a water sharing agreement on the Sabie, and a guaranteed minimum flow in the Incomati at Ressano Garcia.

After the tough experience of the seventh and eighth TPTC meetings, Mozambique requested the World Bank to provide technical advice. Dutch experts from Euroconsult reviewed the existing data on water availability and water use in three rivers in Maputo Province, with emphasis on the Incomati (DNA, 1989). The study was commissioned in September 1989, and proved that water flows at Ressano Garcia had indeed diminished significantly due to upstream abstractions over the period 1952–86. The report contained a secret annex with advice to the Mozambican government with respect to the upcoming negotiations that would lead to the Piggs Peak agreement.

The ninth TPTC was held on November 16 1989 in the Kruger National Park in South Africa but little is known of what transpired. Developments during 1990 indicate that Swaziland and South Africa would not proceed with constructing Driekoppies and Maguga dams without the consent of Mozambique, and that both countries were willing to establish a water sharing arrangement with Mozambique. This willingness came at the time when Swaziland negotiated with the World Bank about financing its contribution towards the construction of Maguga dam. The World Bank, adhering to a long-established practice, demanded a declaration of “no objection” from Mozambique. This forced South Africa and Swaziland to reach agreement with Mozambique.

A water use arrangement was the central issue of the tenth TPTC meeting, held at Piggs Peak in Swaziland on February 14 1991. A day later, the ministers responsible for water met in the Tripartite Ministerial Meeting. The ministers agreed to approve the first phase of the Komati Basin Development Plan (that is, the construction of Driekoppies and Maguga dams), but also, pending agreements resulting from the Joint Inkomati Basin Study, that as interim measures the base flow at Ressano Garcia should be maintained at no less than 2 m3/s, and that South Africa would consult the TPTC prior to constructing any water work larger than 250,000 m3, or any water abstraction larger than 110 l/s on the Sabie.

Also part of the agreement was the decision to conduct a joint study of the water resources, demands, and development potential of the entire Incomati basin. The idea for such a study came from the positive experience of both Swaziland and South Africa when they formulated their joint development plan on the Komati. Conducting this joint study proved a powerful tool towards cooperation, as it broke down suspicions and created alliances at the technical level that influenced the process all the way up to the political level (see also Van Niekerk, 1989). It was with this philosophy in mind that the Joint Inkomati Basin Study (JIBS) was proposed.[16]

During the same TPTC meeting, South Africa made available to Mozambique and Swaziland a report on water resources planning of the Sabie river catchment, which contained a recommendation to construct Injaka dam. This dam was meant to ensure sufficient domestic water for towns and communities in this area.

The Situation in the 1990s: Peace, Protocols, and Bilateral Projects

The Period 1991–7

From 1991 to 1997 no major new dam was commissioned on the Incomati basin, but water use continued to increase. This, combined with the great drought of 1992, resulted in the Incomati drying up again at the border, violating the agreed minimum flow at Ressano Garcia of 2 m3/s. Mozambique complained, but South Africa argued that it was caused by the extreme drought. In the meantime, South Africa did not prevent sugarcane farmers building a weir immediately upstream of the border at Komatipoort, further affecting river flow.[17] In the South African part of the Sabie, communities suffered from severe water scarcity, which underscored the need for Injaka dam.[18]

Two major political developments with a positive impact on the Incomati basin were that South Africa attained majority rule in 1994, and that Mozambique held multi-party elections for the first time. Both developments consolidated the new era of peace and stability that emerged around 1991.

After the important Piggs Peak meeting of February 1991, the TPTC met six more times during the period 1991–7. The most important issue on the agenda during these meetings was progress on the Joint Incomati Basin Study (JIBS), which commenced in 1992. During these meetings, South Africa also regularly informed Mozambique about developments concerning the Injaka dam. The fourteenth TPTC meeting held in July 1995 also had the Umbeluzi basin on the agenda. The Maputo basin only started featuring during the seventeenth TPTC meeting held in May 1998. In 1995, the Joint Incomati Basin Study was completed in curtailed form, because of difficulties with gathering relevant data in Mozambique and the apparent reluctance of Mozambique to actively cooperate.

The Bilateral Komati Development Project

During this period, arguably the most important development on the Incomati basin was the establishment of the Komati Basin Water Authority (KOBWA) between Swaziland and South Africa, based on two treaties signed by both countries on March 13 1992. One treaty established the Joint Water Commission JWC, replacing the JPTC, which would “act as technical adviser to the Parties on all matters relating to the development and utilization of water resources of common interest to the Parties.” With the other treaty both countries committed themselves to building the Maguga in the Swaziland part of the Komati River, and the Driekoppies in the South African part of the Lomati River. This treaty specified how the costs would be shared (South Africa would fund Driekoppies dam and 60 percent of Maguga), as well as how the water would be shared in the Komati/Lomati (32.5 percent for Swaziland, 67.5 percent for South Africa). The treaty also established the Komati Basin Water Authority, which would be the bilateral agency operating the dams on the Komati/Lomati.

Driekoppies Dam (251 Mm3) on the Lomati River was completed by South Africa in 1998. The official commissioning ceremony was held on September 17 1998 by the then Minister of Water Affairs and Forestry of South Africa, Dr Kader Asmal, in the presence of the then Minister for Natural Resources and Energy of Swaziland, Prince Sobandla. At the occasion, the dam was renamed to Lake Matsamo. Around the same time the sod-turning ceremony for Maguga dam in Swaziland (which would be commissioned in 2002) was performed jointly by King Mswati III of Swaziland and former South African President, Nelson Mandela. Both ceremonies clearly underlined that the dams were a joint venture between the two countries.

Developments on the Incomati, 1996–8

In July 1996, South Africa and Mozambique agreed to establish a Joint Water Commission on rivers of mutual interest, in due consideration of the interests of the other riparians of these rivers (that is, Swaziland on the Maputo and Incomati, and Botswana and Zimbabwe on the Limpopo). In the same year, South Africa announced that it would start with the construction of Injaka dam (120 Mm3) on the Sabie River, as it could not wait any longer for an agreement because of domestic and environmental needs. While South Africa had during earlier meetings tabled its intention at the TPTC, the Mozambican authorities considered it a surprise and a violation of the Piggs Peak agreement of 1991.

In 1997, Mozambique and Swaziland started to hold meetings concerning the establishment of a Joint Water Commission, along similar lines to the JWC between Swaziland and South Africa, and that between Mozambique and South Africa. The most important topic discussed during the two meetings held in 1998 was the intention of Swaziland to increase its irrigated area on the Usutu river (part of the Maputo basin) with 11,500 ha under the Lower Usuthu Smallholder Irrigation Project (LUSIP). In April 1999 both countries reached an agreement at the technical level on LUSIP, and in July 1999 the JWC was formally established.

The SADC–EU Conference: May 1997

In May 1997, the SADC Water Sector organized a conference on international river basins in Maseru, Lesotho. In the run-up to that conference, some tension came to the surface between South Africa and Mozambique. South Africa was unhappy about some formulations in the draft text of a paper that would be presented during that conference. Strongly worded references to alleged violations by South Africa of its earlier agreements on the Incomati basin irked South Africa. Mozambique obliged and rephrased sensitive parts of the paper (Box 5; the final paper was published as Carmo Vaz and Lopes Pereira, 2000).

Possibly facilitated by the willingness of Mozambique to take South Africa’s concerns seriously, during the conference the South African water minister committed his country to honor earlier agreements, to right wrongs of the past, and to do everything possible to ensure the agreed minimum flow in the Incomati at its border with Mozambique. This did not materialize in 1998 (see Box bellow) but in 1999 South Africa did manage to deliver the agreed flow at the border.[19]

News clip: “South Africa accused of keeping water” :
Moamba, November 16, 1998

Residents of the district of Moamba, in the southern province of Maputo, are accusing the South African authorities of violating the agreement concerning the use of the water from the Incomati River, that flows through both countries, reports the daily paper Noticias on November 3.

Moamba administrator Romao Mutisse stated that the South Africans take more than their share of Incomati water, which deprives the Mozambican side of water for irrigation and jeopardizes the expected good harvest in the present agricultural season.

“If it rains, as the forecasts say it will, the 1999 harvest will be good,” said Mutisse. “If it does not rain, but the South Africans release water as they should, we can still have a satisfactory harvest, mainly along the banks of the Incomati. But if it does not rain, and the South Africans continue violating their obligations concerning the use of international waters, then we will have serious problems in the Incomati valley.”

Over the last few months farmers in Moamba have been complaining of shortages of water for irrigation, which they blame on the South African unilateral decision to retain the water on their side.

(Source: AIM Mozambique News Agency Report No.147, November 16 1998.)

The Situation since 1999: Towards a New Agreement

The Year 1999: a Turning Point?

Whereas by 1998 relations between Swaziland and South Africa were at their best, relations between Mozambique and South Africa concerning the Incomati were not. Around 1999 a number of new initiatives were started, which would create the conditions for this relationship to improve, and would lead to the three riparian countries signing a new agreement on the Incomati and Maputo basins in 2002. These initiatives included the following:

  • The TPTC established in 1998 the “Incomati System Operation Task Group” (ISOTG), which would advise on how all of the major dams on the Incomati, including both KOBWA dams (Driekoppies and Maguga), should be operated in order to achieve equitable water distribution. The major development here was that the operation of both dams was made subservient to the interests of all three riparian countries (and not solely of Swaziland and South Africa).
  • The commitment of Mozambique to proceed, in 1999, with the second phase of the Joint Incomati Basin Study (JIBS), with funding from Danida. The original study of 1995 would be complemented with data and information on the Mozambique part of the basin that were missing. Where necessary the earlier study would also be updated with new data.
  • The initiative to conduct a Joint Maputo Basin Study. This effectively broadened the agenda of the TPTC, which had so far mainly focused on the Incomati. The study was intended to provide information that would lead to a water sharing agreement between the three riparian countries on the Maputo. The importance of conducting this study increased when Swaziland indicated its intention to undertake a new smallholder irrigation project on the Usuthu River in Swaziland, known as LUSIP. The agencies willing to fund LUSIP demanded a water sharing agreement on the Maputo basin between the three countries. The Joint Maputo Basin Study started in 2000 and was completed in 2001.
  • The TPTC subsequently decided that if a water sharing agreement on the Maputo basin was required, then it would be better to reach a similar agreement on the Incomati as well, and to incorporate both into one encompassing interim agreement for both river basins. The Inco-Maputo Task Group was established in May 1999 to prepare drafts for the TPTC, and met many times (twenty meetings between May 1999 and February 2002).

The Floods of February 2000

See Box above

Nearly 700 people died and 95 more disappeared, many others were dislocated and much infrastructure severely damaged during the floods of the Limpopo and Incomati in February 2000. South Africa assisted Mozambique with rescue operations. For many Mozambicans, this was probably the first time to view the South African military as “brothers,” and created a tremendous measure of goodwill. Just as with previous floods, such as Demoina in 1984, this event once more emphasized the need for basin-wide coordination across borders, and called for effective means of “real-time” information exchange in order to mitigate as much as possible potential hazards of future floods.

The Joint Incomati Basin Study (JIBS), 2001

During 2001 and 2002, two major dams were commissioned in the basin: Injaka (120 Mm3) on the Sabie River in South Africa, and Maguga (332 Mm3) on the Komati River in Swaziland. This brought the total storage capacity in the basin to 2,060 Mm3. His Majesty King Mswati III of Swaziland and South Africa’s Deputy President Jacob Zuma conducted the official opening ceremony of Maguga dam on April 5 2002. Significantly, Mozambique’s Minister of Public Works and Housing, Hon. Roberto White, was present during the ceremony. The dam was renamed as the Maguga Nkomati basin dam.

The second phase of the Joint Incomati Basin Study (JIBS) was finalized and submitted to the TPTC in April 2001. Although the study has not yet been formally approved by the TPTC, some conclusions of the study are worth mentioning:

  • There is insufficient water in the Incomati to fulfill all the plans of the three riparian states. Mozambique’s ambitious plans for irrigation development should be scaled down significantly and be limited to 36,000 ha over and above its current 22,000 ha.
  • In order to cope with the high pressure on the water resources, Mozambique will have to increase storage capacity, first through raising the existing dam wall of Corumana (additional storage: 495 Mm3), and second by building the Moamba Major dam (700 Mm3). South Africa will have to construct Mountain View dam in the Sand river (245 Mm3). With these dams in place, storage capacity would be just above the average annual water generation in the basin at the confluence of the Sabie and the Incomati (see Figure 7 in original publication); that is, with an average residence time of one year.
  • A sophisticated suite of computer models (Water Resources Yield Model and Water Resources Planning Model) was developed by consulting engineers BKS (Pty) Ltd (South Africa) and Acres International (Canada) for and in collaboration with the South African Department of Water Affairs and Forestry. The Water Resources Yield Model (WRYM) was used to analyze the water availability and water supply to the various water users in the basin. It allowed JIBS to measure the impact of various development scenarios. It appears that all three countries trust the model and the outcomes of the various development scenarios considered.
  • JIBS emphasized the importance of experts from all three riparian countries having access to the WRYM computer model, as well as to new models that would assist with operational issues. For a considerable time, it remained unclear to Swaziland and Mozambique whether or not their experts would have access to the WRYM model. This was later clarified by South Africa. In the mean time, however, experts from Swaziland and Mozambique did not use the opportunity to run development scenarios on their own. This might have assisted them in the trilateral negotiations.

Towards an Interim Agreement: 2000–2

The discussions initiated in 1999 on an interim agreement for the Incomati and Maputo basins were tedious. A major breakdown occurred in mid-2000 when the Inco- Maputo Task Group, charged with drafting the agreement, was finalizing the sixth draft. The major issue was whether two annexes that were also being drafted, namely on the exchange of information and on transboundary impact, should be part of the current interim agreement or should become part of the final agreement, envisaged to be reached before the year 2010. The task group could not solve this stalemate. Given its limited mandate, the task group had to refer the matter back to the TPTC for resolution. The stalemate was finally resolved at that level, but significantly delayed the drafting process.

By the end of 2001 the negotiations had yielded little, and some observers close to the negotiations were pessimistic whether an interim water sharing agreement was within reach. However, during the first months of 2002 substantial progress was made. The three water ministers, who met on May 7 2002 in Ezulwini, Swaziland, declared in an official press release, that they:

  • Had discussed the text of the draft “Tripartite Interim Agreement on the Protection and Sustainable Utilization of the Water Resources of the and Maputo Watercourses” (TIA).
  • Confirmed that the TPTC’s “Resolution on the Exchange of Information and Water Quality Standards” would be passed prior to the formal signing of TIA.
  • Confirmed their commitment to signing the TIA during the World Summit on Sustainable Development in Johannesburg, South Africa, in August 2002.

As indicated in Incomati/ Water Basin Profile: Transboundary Political and Institutional Setting, the Tripartite Interim Agreement was indeed signed during the World Summit on Sustainable Development in Johannesburg, on August 29 2002. Three weeks earlier, the TPTC had signed the resolution on the exchange of information and water quality.

A Preliminary Analysis of the Interim Agreement of 2002

The Tripartite Interim Agreement is a very comprehensive document, setting out the laudable objectives of protecting the water resources of the Incomati and Maputo basins and utilizing these in a sustainable manner (see Annex 3 in original publication)). Moreover, the agreement is bold in that it specifies, in an annex concerning “flow regimes,” the water withdrawals in the three riparian countries that are allowed under this agreement (Table 11; compare with Table 5). The quantitative data in this annex were derived from the JIBS study, although some have been modified (notably environmental water requirements).[20]

The most striking feature of the agreement is that it allows a significant increase (nearly 30 percent) in the consumptive water uses of the water resources of the Incomati by all three countries, allowing the commitment level to increase from 51 percent in 2002 to 65 percent in the near future. This includes water reserved for the future needs of the city of Maputo (an interbasin transfer of 88 Mm3/a). The agreement is thus based on the premise that more (secure) water can be created by:

  • increasing the capacity of existing dams (Corumana dam in Mozambique and Vygeboom dam in South Africa)
  • constructing new dams (Moamba Major dam in Mozambique; Tonga, Mountain View, and New Forest dams in South Africa; Silingane and Ngonini dams in Swaziland).

This premise has been corroborated by the Inco-Maputo Task Group, which carefully analyzed the proposed allocations, and conducted a systems analysis by feeding these allocations into the Water Resources Yield Model (WRYM) developed under JIBS. The outcomes were debated at length by the task group, and eventually all delegations were satisfied that the allocations can be provided at reasonable levels of assurance.[21] However, it is clear that such levels of assurance can only be achieved at a high cost, not only in terms of finance but also in terms of increased evaporation losses from dams. Further additional allocations (as specified in a separate annex on “Reference projects” in the agreement) will therefore have to be considered very carefully.

Even with increased storage capacity, the significant increases in water withdrawals will most likely lead to more frequent shortages. These shortages will have to be resolved by relying on Article 1 and Clauses 5 and 6 of Article 4 of the flow regime annex of TIA, which define procedures for managing water use during droughts. Article 1 assigns priority to water for domestic, livestock, and industrial use,as well as to ecological water requirements. Implicitly, runoff reduction due to afforestation also takes priority, since this type of water use cannot be altered overnight. This means that in case of water shortage, the irrigation sector (with 51 percent of withdrawals by far the largest water user in the basin) will have to decrease its abstractions. Consequently, the irrigation sector will experience more frequent shortages in future. Enforcing this priority rule will be a challenge in terms of water management in all three countries, and more so if the political leverage of some large irrigators is considered (think, for example, of the sugarcane industry in all three countries).

The TIA’s premise that more secure water can be made available resulted in the negotiating parties not having to critically evaluate current water uses. The large consumptive water use by exotic tree plantations as well as by some large interbasin transfers was not questioned. Afforested areas in all three countries are allowed to increase significantly, and existing interbasin transfers may continue.

The manner in which TIA defines first priority uses and other uses may not be entirely consistent with the national water laws of the three countries. An example is the National Water Act (1998) of South Africa, which, in Part 3, prioritizes only the Reserve, that is, water to satisfy basic human needs as well as to protect aquatic ecosystems. TIA, however, assigns the status of first priority use to the water transferred out of the Incomati basin from Nooitgedacht and Vygeboom dams for the use of cooling thermal power plants in the adjacent Olifants catchment, part of the Limpopo basin. The same priority is given to the water reserved for the future needs of the city of Maputo. This position may be questioned.[22]

Concerning water requirements of the ecosystems, the interim agreement defines target in stream flows for the Sabie, Crocodile, Komati, and Incomati rivers. At Ressano Garcia, the minimum flow target is 2.6 m3/s, which is higher than agreed in 1991 at Piggs Peak (2 m3/s). At the estuary (Marracuene) the minimum flow target is set at 3 m3/s, which is less than recommended by JIBS (5 m3/s), but double the minimum flow recommended by Mozambique in 1984. Further detailed research on the water requirements of the Incomati estuary is clearly required, and the costs and benefits of various minimum flows estimated. This is not only relevant because many poor households derive important benefits from this estuary, but also because the integrity of the estuary and the whole of the Incomati River within Mozambique is important for the aquatic ecosystems in South Africa and Swaziland.[23]

Despite these critical observations, the Tripartite Interim Agreement is an important positive achievement, and a landmark in the sharing of international waters. The three countries have not only accepted the equity principle in utilizing the Incomati and Maputo water resources, but have also been able to translate this into concrete, measurable, and thus enforceable, commitments. They obviously take this agreement very seriously, as they invested three long years in negotiating it. However, negotiations on the mooted Comprehensive Agreement, to be concluded by 2006, will be tough and complex, as the Incomati basin is approaching “closure.”

From Competition to Cooperation: a Conclusion

The Incomati river basin has a dynamic and checkered history of water sharing between Mozambique, South Africa, and Swaziland. The basin has seen a change from colonial rule to more or less totalitarian regimes and a subsequent transition to democratic rule. These changes were accompanied by violent interventions. At a regional scale, SADC was transformed from a governmental anti-apartheid organization of the frontline states into a regional organization for political and economic integration. During these turbulent times, the region has seen years of severe drought, hunger, and floods. There were thus ample reasons and triggers for serious conflicts over water to emerge. Yet the tensions over water that indeed existed never escalated into open conflict. To phrase it more poignantly, during a large part of the 1980s South Africa and Mozambique were practically at war and literally fighting, except over water. This is remarkable, and the following question begs an answer: Why didn’t open conflict emerge between the riparian countries over the water resources of the Incomati, and why did cooperation prevail?

It is impossible to single out one clear-cut answer to this question. The answer must take into account the larger socio-political context that shaped the negotiations, which, as was shown, was fairly complex, multi-faceted, and sometimes contradictory. The answer must also consider the specifics of water availability in the three countries, and the manner in which the geography and hydrology of the Incomati basin tie them together and draw them into specific dependency relations, as well as their water needs and future development ambitions. Finally, the interests of individual water users (the majority of whom have remained voiceless) must be taken into account.

Socio-Politics at the Level of the Nation-States

The inequality of the three riparian countries in terms of geographic position, economy, and technological know-how is striking. In the Incomati the upstream country is the most powerful, whereas the downstream country is comparatively weak. In such a situation, the upstream country may be tempted to ignore the interests of other riparians, and the potential transboundary impact and damage of its development activities on downstream countries. This would provide fertile ground for open conflict to emerge.

The previous sections have shown that up to 1991 South Africa developed its water resources to a large extent without considering the needs of Mozambique, while taking a more careful approach with regard to Swaziland. The more considerate approach towards Swaziland may be explained by several factors. First, South Africa is situated both upstream and downstream of Swaziland. Denying water to Swaziland would directly affect water availability in South Africa’s most heavily committed part of the basin, namely immediately downstream of Swaziland along the Komati. Second, Swaziland was one of the few independent countries in Africa that maintained good political relations with the apartheid regime, which South Africa did not want to put in jeopardy. Third, South African companies had economic interests in various waterrelated ventures in Swaziland. Thus, South Africa could not ignore Swaziland’s interests in the Incomati. Yet at various crucial moments in the history of water sharing, South Africa also took Mozambique’s interests, which it could have ignored, into account.

This was the case with the Piggs Peak Agreement reached in 1991, whereby South Africa and Swaziland acknowledged that Mozambique was an interested party in their bilateral Komati development project. Although the agreement was signed under pressure from the World Bank (it would only finance Swaziland’s Maguga dam if Mozambique accepted it), and coincided with the dawn of a new era of peace, it was rooted in the bilateral and tripartite negotiations that preceded it.

The Piggs Peak agreement is an example of a triple-win. Swaziland got what it wanted and could pride itself in having brokered negotiations between two antagonistic states. South Africa could show that it played its new role of responsible neighbor. Mozambique could also show its goodwill (as it allowed the Komati Development Plan to go ahead), and got critical undertakings by South Africa concerning the Sabie River and a guaranteed minimum flow at Ressano Garcia.

Subsequent developments show that South Africa remains committed to considering downstream interests. This is explicitly stated in its National Water Act of 1998, and was also demonstrated by the minister of water affairs in 1997 who personally guaranteed the minimum flow at the border with Mozambique at Ressano Garcia. This commitment has been reconfirmed with the signing of the Tripartite Interim Agreement in 2002. The role of SADC in this development was also significant, as the new South Africa was eager to show its new face to the region. This was most convincingly displayed when South Africa rendered speedy and life-saving support to Mozambique during the floods of February 2000.

The record of the tripartite negotiations contains one stain. The construction of Injaka dam on the Sabie catchment led to frictions between South Africa and Mozambique. South Africa maintained that it met all requirements before constructing the dam. Mozambique disagreed. [24]

Can More Water Be Squeezed out of the Incomati?

Whereas the riparian countries are committed to take each other’s interests into account and to find negotiated solutions, there is a tendency, and a need, in all three countries to satisfy all users (existing and prospective) as much as possible. South Africa faces the challenge of righting the wrongs of the past by making more water available to emergent black farmers, but without questioning existing water uses by the historically advantaged elite (sugar, paper, thermal power, and so on). In this mode of compromise it cannot consider the Incomati water resources as finite and “closed,” but prefers to project further increases of water withdrawals. Similarly, Swaziland wishes to increase access to water by small-scale producers, without affecting the large established sugar companies. Mozambique, for its part, faces an enormous challenge of economic development after more than a decade of civil war and economic stagnation. It therefore plans for a significant irrigation expansion and may require additional water for the city of Maputo from the Incomati.

The Tripartite Interim Agreement (TIA) honors the wishes and development plans of all three countries and thus suggests a “win–win” outcome. As was observed in the previous chapter, water scarcity is likely to occur more frequently. If in such situations the arrangements as spelled out in TIA are not fully complied with, downstream Mozambique will bear the brunt.

With water abstractions on the increase, and hence the frequency of water scarcity, new tensions may build up in the basin, and new negotiated solutions will be required. These negotiations will, of necessity, be more difficult than those of the past. As the Incomati basin is approaching “closure,” they will be “zero-sum.” In the not too distant future, some consumptive uses will have to decrease if other uses are allowed to increase.

The Role of Interest Groups and Stakeholder Participation

Among the most surprising aspects of the negotiations on the Incomati are the apparent voids: those interactions that did not occur, or did not come out in the open. The tripartite negotiations seem to have been largely confined to a handful of officials from the three countries. It remains unclear to the authors whether and how water users and user representatives were involved in them. And yet those are the ones with the real stakes.

Within each country large socioeconomic interests are associated with access to Incomati water. In Mozambique the most important interest groups include smallscale farmers, the two large sugar estates, those deriving benefits from the estuary, and the needs of the city of Maputo to secure additional water in the near future. In Swaziland, the two major interest groups include the established sugar estates and the small-scale farmers who also wish to grow sugarcane. In South Africa, the situation is more varied, with established commercial irrigators, newly emergent black farmers, the TSB sugar mills, the Sappi paper company, and the Kruger National Park all having strong interests in accessing Incomati waters.

There is no doubt that many of these interests have been able to articulate their wishes to country negotiators. It will require more detailed research to establish how this was accomplished, revealing the pressures under which the country delegations sat around the negotiation table. But it is known, for instance, that Anton Rupert, the founder of the Rembrandt group that owns the TSB sugar company, has direct access to the highest political circles, not only in South Africa but also in Mozambique. Other companies may have remained sidelined in the negotiation process, but have found other ways of accessing Incomati water. Such is the case with the two biggest sugar companies in South Africa, Illovo and Tongaat-Hulett, who recently acquired significant stakes in the two sugar estates in the Mozambican part of the Incomati basin. In addition, Tongaat-Hulett has interests in sugar production in the Umbeluzi basin irrigated with Incomati water. South African companies are therefore competing for the Incomati water resources in all three riparian countries. If the water cannot go their way, these companies go where the water is!

The fact remains that no formal and transparent procedure for water users to participate in the trilateral negotiations existed. The danger, then, is that water user groups participate through other means, such as “old boy networks.” This creates an uneven playing field whereby the majority of water users may remain without voice. In future this will have to change. Stakeholder representatives must get formally involved in the sharing agreements at the basin scale. This will allow the negotiations to “deepen,” and the agreements to be more effectively implemented and enforced. Water users have to “buy” into the agreements, and widen their often limited perspective to the entire basin.

The water reforms in all three countries (Swaziland’s new Water Act is expected to come into force soon) provide a unique opportunity. The newly established Catchment Management Agencies in South Africa and Regional Water Authorities in Mozambique, and the proposed River Basin Authorities in Swaziland, all have a strong component of stakeholder participation. Such water management bodies, including water user boards, should be stimulated to establish contacts across borders. This will enhance mutual understanding, and will allow them to give practical form to the bilateral and trilateral agreements reached between states.

Negotiating a Water Sharing Agreement: Some Practical Lessons

Widening the Scope of Negotiations to the Entire Basin

The first negotiations started mainly between Swaziland and South Africa about the river that was most heavily committed at that moment: the Komati. At that time it was not opportune to take a basin-wide perspective. Only gradually did the entire basin emerge as the only appropriate unit on which to base agreements.

The process of negotiating water in the Incomati has therefore been iterative, with an initial focus on the smaller spatial scales. With the steadily increasing pressure on water resources, the interconnectedness between the various parts of the basin became apparent (exemplified by the drying up of the Incomati at the border between Mozambique and South Africa). This led to widening the scope of the water allocation process also to the largest spatial scale, the watercourse as defined by the UN Convention and the revised SADC Protocol.

Broadening the Scope of Negotiations to Include Other Issues

When by 1998 a stalemate was reached on the Incomati, a relatively minor development created new opportunities for an agreement to be reached. The Lower Usutu Smallholder Irrigation Project (LUSIP), on a river basin that was less contested than the Incomati (the Maputo basin), brought to the negotiating table by the least controversial riparian country (Swaziland), caused the negotiation base to be broadened. With the inclusion of the Maputo basin, new combinations of negotiation positions were suddenly possible. One country might want to score “more” on the one basin and be somewhat “flexible” on the other, depending on that particular country’s needs. This broadening of the scope of negotiations therefore offered more options, enhancing the chances of a positive outcome.

Another instance of broadening the scope of the negotiations was when the JIBS study added the environment to the list of water using sectors. The recognition by the upstream countries that indeed the environment required water gave a fresh argument in favour of an old issue that had been the subject of negotiations ever since the Incomati river fell dry for the first time in 1982, namely the establishment of minimum flows. When in 1998 South Africa adopted its new Water Act, it had to reserve water for environmental needs within its territory. Since environmental water requirements are largely non-consumptive, this means that more water will flow into the downstream country, satisfying an important issue for Mozambique.

Broadening the scope of the negotiations therefore offered more give and take options, which enhanced the chances of a positive outcome.

Structuring the Negotiations: A Phased Approach

At most times during the Incomati negotiations, the three riparian countries did honor each other’s requests for information. The biggest opportunity to go beyond sharing information, and achieve a shared vision on the facts and future of the basin was created with the agreement at Piggs Peak to conduct the Joint Inkomati Basin Study. This opportunity, however, was not exploited to the full. This created a serious drawback, in that a new water sharing agreement took much longer than initially envisaged: three years of tedious negotiations. A similar chance was missed when access to the WRYM computer model developed under JIBS remained unclear. As a tool to analyze the impacts of development scenarios, WRYM can potentially create convergence among the negotiating parties.

It is apparent that it is imperative to conduct the negotiations in a structured and systematic way, from information exchange, via shared understanding of facts, to reaching agreement about a shared vision. Only thereafter will it be feasible to reach a formalized and comprehensive agreement. The Incomati shows that shortcuts are costly in terms of time and opportunities lost.

The Unique Role of the Downstream Country

The downstream country was the party that had most to lose, and thus had to work hardest to claim its share. At times when contacts were less intense, upstream parties developed without impediment. The only option available to the downstream country was to insist on keeping the lines of communication open, even in times of war, and to show a willingness to reach negotiated solutions. The case of the Incomati shows that most of the time the other riparian countries obliged. This also implies, however, that the downstream country is often put in a position of responding and reacting to upstream faits accomplis, and finds it difficult to influence the agenda.

Technical Negotiations may Continue when Politics Fail

Discussions at the technical level can be very efficient, especially when official contacts are not possible at the political level. Mozambique has been able to secure a share of its waters thanks to the technical meetings that continued during the period 1974–91, when the official political relations were almost entirely hostile.

At times it was necessary for the weaker countries to engage external technical support (Swaziland received technical advice from the US Army Corps of Engineers, Mozambique from Euroconsult). In the case of Mozambique, this one-off type of technical support came at the right time, and led to the country becoming a more credible negotiator, taking on a firm and clear position. It is important to acknowledge that this did not weaken the other parties, but led to a credible, equitable, and legitimate sharing arrangement (Piggs Peak in 1991) that was largely upheld by the new Tripartite Interim Agreement (2002).

The Need for Capacity Building

There is a marked difference between the three riparian countries in terms of relevant expertise, despite significant capacity building efforts in Swaziland and Mozambique in the recent past. This may pose a threat to the successful implementation of existing agreements, and the conclusion of future negotiations. In devoting a separate article to capacity building in the Tripartite Interim Agreement, all three countries acknowledge the importance of this issue. To level the playing field there is need to continue investing in human resources.

From Conflict to Cooperation

So why was there not open conflict between the riparian countries over the water resources of the Incomati, and why did cooperation prevail?

First, because riparian countries are made up of people who share a common space and a common history, and therefore also a common future. There is an inherent pressure for nation-states to behave as good neighbors, even when political ideologies diverge. Furthermore, there are outside pressures on nation-states to act responsibly, and to honor existing regional and international conventions. The strong political ties between states in Southern Africa are an important and positive resource. Moreover, multilateral banks intent on funding large water projects normally demand declarations of no objections by other riparian states.

The second reason is the particular political developments in both Mozambique and South Africa. Just when the need for an agreement was highest, the cold between both countries started to melt, allowing the Piggs Peak agreement to be reached in 1991.

Third, there was a third riparian country whose role as broker was accepted by the other two riparian countries, because of its particular political and hydrological position vis-à-vis the other two countries.

Finally, and perhaps most importantly, potential conflicts were evaded by allowing more water to be abstracted and more dams to be built. So far the negotiations can therefore be considered non-zero sum games. However, as the Incomati basin fast approaches closure this situation is bound to change. The comprehensive agreement, to be concluded by 2006, will have to address this fact. Water sharing will increasingly be a delicate balancing act between cooperation and competition.

The hypothesis that water drives peoples and countries towards cooperation is supported by the developments in the Incomati basin. Increased water use has indeed led to rising cooperation. When the next drought comes and Mozambique, South Africa, and Swaziland enforce their recently concluded agreement, and voluntarily decrease those water uses deemed less essential, then the hypothesis has to be accepted.


  1. The word "Incomati" used to be spelled in South Africa and Swaziland as "Inkomati". Both spellings refer to the entire river basin. We follow the convention and use Incomati when referring to the entire basin, except when quoting official documents that do not follow this convention. Please note that "Nkomati" (Swaziland) or "Komati" (South Africa) refers to the Komati river, which is one of the tributaries in the Incomati basin. In Mozambique "Incomati river" is used to refer to the main stem of the river inside that country. (The JIBS, however, used "Inkomati" to refer to the entire river basin, and in quoting them we follow their convention.)
  2. The Umbeluzi river basin covers a total area of 5,460 Mm2, of which 41 percent is located in Mozambique, 58 percent in Swaziland, and only 1 percent in South Africa. The Maputo river basin covers a total area of 30,000 Mm2, of which approximately 57 percent is located in South Africa, 37 percent in Swaziland, and 6 percent in Mozambique.
  3. Another project was financed with South African capital, namely the construction of a dam on the Cunene river, which forms the border between Namibia and Angola. In 1964 Portugal and South Africa signed the so-called “Cunene Agreement,” which covered all waters of mutual interest to Portugal and South Africa. This agreement therefore included the waters of the Incomati basin (Bipartite Agreement between South Africa and Portugal in regard to Rivers of Mutual Interest and the Cunene River Scheme).
  4. The Southern Africa Development Coordination Conference, which was transformed into SADC (Southern Africa Development Community) in 1992.
  5. Dr Theo van Robbroeck recalls that because of the Cunene hydroelectric project on the border between Angola and the then South West Africa (Namibia), there was a lot of interaction between the South African authorities, then responsible for that project, and the Portuguese. It can be surmised that there was a desire to extend such discussions to other water resources of common interest between these two authorities, but which, of necessity, Swaziland was to be a part. The Portuguese may already have been worried about upstream water use in the Incomati due to the commissioning of the Nooitgedacht Dam. At that time it was also known that Eskom envisaged further power stations in the same area (personal communication, Dr Theo van Robbroeck, July 25 2002).
  6. The meeting of the Standing Technical Committee on Parameters (Mbabane, November 21 1972) agreed that works in excess of 10 Mm3 storage and 10 Mm3/a abstraction (or 0.32 m3/s) would be recorded by the TPTC
  7. Personal communication Eng. Gebre Libsekal, April 2002.
  8. Around 1976, a conflict emerged between Swaziland and South Africa on the Komati River. In the eyes of Swaziland, South Africa had reneged on the earlier agreement of maintaining a certain minimum flow in the Komati River at the border with Swaziland (personal communication Eng. Gebre Libsekal, April 2002). In September 1979, the drought of that year triggered the request by Swaziland to South Africa to increase its water releases from the Vygeboom Dam (upstream in the Komati) to 68 cusec (1.9 m3/s). The drought of 1982 caused a water shortage in Swaziland and in July 1982, Swaziland requested South Africa to release more water in the Komati River. It is not known whether the 1979 and 1982 requests were granted.
  9. As a result, the United States Corps of Engineers were hired at a cost of $56,000 (personal communications, Mr Peter van Niekerk, January 2002, and Dr Theo van Robbroeck, February 2002).
  10. This agreement, one and a half pages long, does not regulate issues related to exchange of information, joint control of measuring stations, and water quality, nor does it impose regular meetings between the two countries for verification of its implementation (Carmo Vaz and Lopes Pereira, 2000). A major weakness of the agreement is that it requires Swaziland to release to Mozambique 40 percent of the combined measured flows (not naturalized) past GS3 and 10. Further developments upstream of these gauges can significantly reduce the flow without violating the agreement (personal communication, Mr Sidney Dhlamini, July 2002).
  11. Personal communication, Mr Peter Van Niekerk, January 2002.
  12. Personal communication, Eng. Gebre Libsekal, April 2002.
  13. The third and fourth TPTC meetings were held in 1983, and the fifth and sixth in 1984. The seventh TPTC meeting was initially scheduled to be held in 1985 but did not take place until 1989. Whereas the TPTC met six times during 1982–4, the JPTC (Swaziland and South Africa) met eleven times in the same period.
  14. Personal communication, Mr Hubert Savenije, November 2001.
  15. Personal communication, Mr Peter Van Niekerk, January 2002.
  16. Personal communication, Mr Peter Van Niekerk, October 21 2002.
  17. Interestingly, these same farmers would also fight a battle upstream, namely with the South African electricity company Eskom, which diverts large amounts of water out of the Upper Komati River (Waalewijn, 2002).
  18. Personal communication, Mr Mike Muller, October 24 2002.
  19. The internal arrangement in South Africa was that 1.1 m3/s should come from the Komati River, and 0.9 m3/s from the Crocodile. At the border with Mozambique, both rivers join (Waalewijn, 2002, p. 79).
  20. It is important to note that TIA does not explicitly consider evaporation losses from dams, as these have not been included in the allocations presented in the flow regime annex of the agreement.
  21. Personal communication, Mr Niel Van Wyk, October 24 2002.
  22. Compare this also with the consent by Mozambique to the interbasin transfer from the Pungwe river basin for the primary needs of the city of Mutare in Zimbabwe, and its refusal to allow Zimbabwe to transfer water from the same river out of the basin for irrigation purposes, a position which was accepted by Zimbabwe. Both countries reached this agreement on September 27 1995, after eighteen months of negotiations.
  23. Personal communication, Mr Niel Van Wyk, October 31 2002.
  24. One observer is of the opinion that had the JIBS study from the start been conducted in the manner originally envisaged, namely all three countries actively taking part in it simultaneously, it would have been less likely that disagreement would have emerged on the Injaka dam. (Personal communication, Mr Peter Van Niekerk, October 21 2002.)

See also

External Resources

8804 Rating: 2.3/5 (63 votes cast)