Can Environmental Injustices be Addressed in Conservation? Settlement History and Conservation-Induced Displacement in the Case of Lyanshulu in the Zambezi Region, Namibia

The Zambezi Region of Namibia is known for its vast conservation landscape, which is part of the world’s largest transboundary conservation area – the Kavango Zambezi Transfrontier Conservation Area (KAZA TFCA). Less known are the settlement histories of the communities displaced from their villages during the planning and implementation of this conservation landscape. We reconstruct the settlement history of two families that lived on two islands in the Kwando River and the adjacent riverbanks in today’s Mudumu NP, and we describe the circumstances that led to the designation of a national park in this region and the subsequent conservation-induced displacement. The evicted families increasingly protest and legally contest estrangement from their ancestral land, which is now a protected area. We argue that current nature-conservation efforts can succeed only if issues of past and present environmental injustices are comprehensively addressed.


Introduction
The establishment of a conservation landscape in the Zambezi Region, at the heart of the world's largest transboundary conservation area, is emblematic of major challenges met with grand solutions. The Zambezi Region of Namibia is an important part of the Kavango Zambezi Transfrontier Conservation Area (KAZA TFCA), home to many tourism and hunting enterprises and famous for its abundant wildlife, especially its large elephant herds and magnificent savannah landscape. But the history of this conservation landscape -at least in some parts -includes the displacement of local residents due to conservation planning even before the declaration of Namibian independence in 1990, and has continued to have adverse effects on the livelihoods of those who were evicted from the area.
Our interest focuses on the displacement of the people from a place called Lyanshulu, located inside the Mudumu NP today -which has been largely disregarded in official accounts, except for a small note by Ashley (2000, p. 17) and the descriptions of Lenggenhager (2018, p. 165) and Bollig and Vehrs (2021, p. 17-18). Lyanshulu is both the name of a region and a village in the Zambezi Region of Namibia. In 1980, the village was located on the east bank of the Kwando River, which now lies within Mudumu National Park, whereas the new Lyanshulu (post-eviction) settlement is situated on the southern border of the park (see Fig. 1). 1 Along with the Bwabwata and Nkasa Rupara NPs, Mudumu NP is one of three national parks in the Zambezi Region. There are also a designated forest reserve, 15 community conservancies, and several community forests. In Namibia, more than 45 percent of the land is under some form of conservation regime (MEFT/NACSO, 2021).
Since further resettlements have taken place in the Zambezi region at different times, we refer to the movements of residents in different parts of the Zambezi region that were negotiated with the South African administration, as described by Bollig and Vehrs (2021), as relocations. In these cases, the decision to depart from the original settlement was generally not voluntary, and some affected people consider it illegitimate; however, these movements were negotiated with the local authorities, and agreements were reached. For the displacements we describe below, we assume that there was no prior communication and agreement on the relocation and that families were forcefully evicted. We further want to characterise this movement as a conservation-induced displacement of people motivated by strategies for nature and wildlife conservation. According to Terminski (2013, p. 11): "In the case of development-induced displacement or conservation-induced displacement, territory becomes an arena of specific conflicts between the interests of the public or private sector and the needs of people displaced or affected by particular development deci-sions.
[…] Displacement is primarily a phenomenon associated with the loss of land, which is a fundamental point of economic, social and cultural reference. In each of these cases the largest single cost of the conflict is paid by affected individuals and communities, who in many cases are not even the subject of the dispute leading to their displacement." Through using the notion of conservation-induced displacement in our case study, we want to draw our attention to different levels of environmental injustice. With brief reference to the history of environmental (in)justices and the roots of the movement in the United States of America here (Coolsaet, 2021;Murdock, 2021;Walker, 2012), we want to take account of Schnegg and Kiaka (2018) of environmental injustice in the context of Community Based Natural Resources Management (CBNRM) in Namibia as a starting point to both identify different layers of injustice and examine the role and responsibility of CBNRM-related institutions in addressing these issues. Kiaka (2018) gives a detailed overview of the transformation of the concept, as well as its application to the Namibian CBNRM context. He  Schlosberg (2003Schlosberg ( , 2007 and outlines three dimensions of environmental injustice -'distribution, recognition and participation' -that are essential to address potentially unjust situations in the context of CBNRM and conservation in Namibia. Together with Schnegg, he further discusses these three levels of analysis to 'explore (1) who gets what, (2) who has to live with what, and (3) whether people perceive this distribution to be equitable and fair' (Schnegg & Kiaka, 2018, p. 106).
Here, we address different topics that relate to (1) the settlement history of two Mayeyi families in the area that is today a national park; (2) the planning and implementation of conservation measures and the subsequent displacement of residents; (3) the perspectives of evicted residents and their claims to address social and environmental injustices of the past; and (4) the dilemma for the Namibian CBNRM programme of not being able to address these issues with the current conservation model.

Methods: Cultural Mapping
As an interdisciplinary team trained in anthropology and geography, we have been conducting fieldwork in the Zambezi Region since 2018, for a total of 10 months to date. We have held several interviews, group meetings, and informal conversations in Lyanshulu. We conducted key interviews with the local authorities (induna) of the families and their representatives (ngambela). Further interviews were conducted with different stakeholders from the Ministry of Environment, Forestry and Tourism (MEFT), the national NGO Integrated Rural Development and Nature Conservation (IRDNC), the Balyerwa Conservancy management, and the Mayeyi traditional authority to understand their positions in the long-lasting conflict, in which the claims and complaints of the affected families are continuously overlooked. We returned the results of the initial mapping in 2020 to the evicted families, and they were intensively discussed in several group meetings to ensure a valid representation of their settlement history. In all cases, we worked with a local assistant to ensure that all persons involved in the research had the opportunity to participate in the discussions held in both Siyeyi and English.
In addition, we consulted archival material from the National Archive of Namibia to both document the conservation planning from a colonial governmental perspective and to triangulate the different kinds of data available (interviews, cultural mapping data, historical maps) with the records found in the archives.
Besides these standard ethnographic and geographical methods, we also applied a cultural mapping approach inside Mudumu NP with members of those families evicted from the park. The cultural mapping allowed us to respond to the request from the former inhabitants of Lyanshulu to address the events of the past. The cooperation between anthropology and geography also provided the opportunity to combine their respective disciplinary strengths: focusing on local perspectives while, at the same time, recording the site situation and connecting it to the spatial context of a map perspective, thus making it accessible to a broader audience (see also Dieckmann, 2021). In general, cultural mapping is seen as 'a systematic tool to involve communities in the identification and recording of local cultural assets' that allows a better understanding of local communities more generally 'in terms of cultural identity, vitality, sense of place, and quality of life' (Duxbury et al., 2015, p. 2).
We were accompanied by the local authorities of the two evicted Mayeyi families -Makombwe and Mufalali -to their ancestral land within Mudumu NP, including the islands of Qxkatwa and Kweka. On the islands, we visited the places where people used to settle, grow crops, or bury their relatives. In this river island environment, the canoe (mukoro) was the primary means of transport in the past. To ensure the safety of all participants, we were assigned a wildlife ranger by the Mudumu NP ranger station on behalf of MEFT. 2

Historical Accounts of Lyanshulu
In the Zambezi Region, relocations started in the 1930s when the region was under the South African administration. The number and scale of relocations increased until the 1970s when the East Caprivi had become a pseudo-independent homeland, and enforced settlement shifts were planned and carried out by an 'East Caprivian Administration.' Our archival research (and recent publications by Lenggenhager, 2015) highlight how all major decisions in the Bantustan administration were taken by White South African staff who liaised with traditional local authorities from time to time. The study area along the Kwando River in the western part of the Zambezi Region is inhabited mainly by Mayeyi, Mafwe, and Hambukushu people, who speak south-western Bantu languages that are closely related yet distinct enough to be recognized as separate languages. The Mayeyi ethnic 2 The assistance of a wildlife ranger was a required condition specified by the MEFT, particularly to ensure protection from larger wildlife, such as hippos, elephants, crocodiles, or predators. During the cultural mapping, we recognized these measures were absolutely necessary. Beyond that, we were by no means restricted by the MEFT or other institutions, and we were fully able to listen to the reports of our companions from the evicted community during the mapping and follow their instructions. Since the wildlife ranger also lives in that community (affiliated through marriage) and was supportive of the mapping, this configuration did not create any conflicts of interest (COI). group that resides in the southern part of the Zambezi came from the southern region around Lake Ngami. Bolaane (2013) describes how both Mayeyi and Hambukushu moved towards Lake Ngami from their former territory along the Zambezi River during the expansion of the Barotse kingdom in the mid-eighteenth century. In the early nineteenth century, in reaction to Tawana expansion towards Lake Ngami, some Mayeyi people moved further to the riverine areas between today's Namibian and Botswanan territories (Bolaane, 2013).
The earliest archival accounts of the settlement in today's Mudumu NP refer to the Makombwe family residing in the riverine area of the Kwando River in the late nineteenth century. A map of an exploratory trip in 1898 and 1899 published by Gibbons (1901), a British major who undertook a mapping of the Barotse kingdom at the end of the nineteenth century, refers to 'Nana Makomba' at the southern altitude of 18° 7' in the riverine area of the Kwando River. Some years later, Seiner (1909), an Austrian traveller, reporter, and cartographer in the service of the German colonial government in 1905/06 (Kostka, 2007), submitted the following report on 'Mona Makomkwe' and the place 'Lanschuru:' The river landscape of Muniambania [few kilometres north of Lyanshulu and still within today's Mudumu NP] is of strategic importance, because the river, which is 60 metres wide here, is without marshes and side channels when the water is low, as is the case at Mona Makomkwe [sic.], […]. The river's channel network south of Muniambania to Lanschuru [sic.] is 1 ½ to 3 km wide and elevated [...]. (Seiner, 1909, p. 72;own translation) Both accounts of 'Nana Makomba' and 'Mona Makomkwe' are references indicating the settlement of the Makombwe family in this location. Most probably, the terms 'Nana' and 'Mona' refer to the vernacular Siyeyi term 'mwana'-'son' or 'daughter' of Makombwe (Fig. 2). 3 The entries in Seiner's map and the corresponding report describe locations such as the village of 'Mona-Makomkwe' (at 18° 7′ 30'' south altitude) and the ford named 'Makomkwe' where European travellers used to cross the river -something that was possible in only a few places along the Kwando and Linyanti Rivers at that time. 4 For the early twentieth century, Seiner also describes more generally 13 Mayeyi villages located along the Kwando River with about 700 inhabitants, an average of more than 50 people per village.

The Discovery of Lyanshulu and the Planning of the Mudumu National Park
Apart from the numerous cartographic records from the early twentieth century, few written records are available for Lyanshulu in particular, and further detailed descriptions did not emerge until after the South African administration had taken over. The records of the administration officer Kruger from the 1970s are especially important: We were now entering some of the most charming country of river, forest, glades and plains where Nature seems to have bestowed special favours -Lianshulu 5 [sic.].
[…] There was Game abounding, too, Impala in the Mopane glades frequently adding a touch of graceful and colourful life.
[…] Again Nature had been over-kind in providing magnificent trees on a high bank with a deep waterway passing by and a most picturesque outlook over this large island named Nxakatwa [sic.]. (1984c, p. 20) But Kruger is also concerned about the decline of wildlife numbers, especially through colonial hunting, that he perceives between his first term in office in the 1940s and his observations in his second term in the 1970s. From his perspective, a crucial turning point for the future of the region and its residents was reached in the 1970s. Conservation and tourism were, in Kruger's opinion, clearly focussed on presenting 'unspoiled Africa' to 'wealthy clients,' although he is also aware that establishing conservation areas is a 'delicate' matter: Entering the field about 1977 was an enterprise to attract wealthy clients from overseas and southern Africa who, as one understood it, would be flown from international airway stops directly to a camp on the Mashi River [Kwando River] up by a few miles from Lianshulu. […]. This scheme [between Lyanshulu and the islands of Nkasa and Lupala in the south] was estimated to bring in a lot of money, not only for the management but also for the Caprivi government. Whatever the achievements may in the meantime have been, there can be no doubt that such a scheme in the context of the Caprivi would require both imaginative and delicate application, well-thought-out and clear rules and unquestionable integrity, not to mention firstclass service in all directions.
[…] Towards conservation, therefore, one would start with the Forest Reserve already defined and not to be tampered with. Add to this the sparsely occupied (1976) and charming Lianshulu -Lizauli frontage on the Mashi River [Kwando] of about 15 miles 6 and its broadening interior heading back north-eastwards to within reach of the said Forest Reserve […] With the achievement of something of that kind the bringing in of 'clients' prepared to pay for the privilege of a week or two in unspoiled Africa would hold out very much better prospects, Nkasa Island to be part of the package. (1984d, p. 15-16) The plans to establish conservation areas along the Kwando River soon became more concrete and were reported to officials in South Africa.
The first map with proposed conservation areas was created in 1976 by a South African conservation officer (van der Waal, 1976). This hand-drawn map proposed a much larger hunting reserve in the southern parts of the region (see Fig. 3.2.2), and, remarkably, all existing settlements were excluded from the conservation planning. Indeed, the settlements were often even located directly adjacent to the boundaries of the protected areas. A few years later, van der Vegte et al. (1983) produced a more detailed report, including a map (see Fig. 3  scheme), 7 van der Vegte et al. also outlined conservation areas, a game farm, and the forest reserve, which was already established according to the plans of the South African forest expert Breitenbach (1968). 8 In the 1970s and 1980s, interest in the idea of turning the Caprivi into a conservation landscape picked up, although van der Vegte et al. indicate this was initially also combined with the implementation of agricultural development zones: Something that cannot be disregarded in development of the Caprivi is the role nature conservation can play as a revenue earner and the role that it must play … to ensure rational, long-term utilization of the region's natural resources and, through conservation of soils, ensure sustained yields in agriculture. With its rivers, swamps, floodplains and Lake the region has consid-  (4) erable tourist potential. At present the Nature Conservation authorities are in a tenuous position. The once plentiful and varied game has largely been shot out, particularly over the last 10 years, and it will take some time, and legal protection, if the numbers are to be built up again. There are no conservation areas or reserves as such although the Division of Forestry has control over the Forestry Reserve in the north-west. The hunting concession and safari camp in the Lianshulu area have been abandoned and there are no other facilities for tourists, except limited accommodation in Katima. 9 (1983, p. 6) Concerning the role of settlements in the planning, an important difference from the older van der Waal draft (1976) is that the Lyanshulu settlements now lie within the planned conservation area boundaries. Van der Vegte et al. (1983, p. 154-155) also proposed three different parks that could potentially be designated, one being the Mudumu Reserve, as they name it ( Fig. 3 Two elements were of particular importance: the creation of a conservation landscape along the Kwando River, and the establishment of a wildlife migration corridor between Botswana and Namibia. However, it is difficult to trace the exact reconstruction of the decisions made and the people involved in the process that led to the proclamation. Lenggenhager (2018) assesses the accounts of the time between 1980 and 1990 but also states that although many in the Zambezi Region know about the forced eviction, hardly any records can be found about it. 10 While he recounts interviews with people directly involved in the eviction, he also indicates no certainty about when it occurred because his informants refer to the years 1982 and 1985. However, our Lyanshulu informants report that they were evicted on 15 June 1980. Lenggenhager (2018) concludes that the eviction took place not later than 1984, and a nature-conservation law-enforcement unit evicted the Lyanshulu people and thus established the basis for introducing the national park. He also reviews the complex political situation before the proclamation and recalls the difficulties that arose between the people affected -who are of Mayeyi ethnic affiliation -and their political representation through Chief Mamili, who belongs to the Mafwe ethnic group. 11 Though the Mamili family was in power and had ruled over a so-called confederacy of Mafew, Mayei, Matotela, Hambukushu, and Khwe people since the early twentieth century (see also Kangumu, 2011), a dispute arose because decisions were being made by people who were neither affected nor in any legitimate position to represent those being affected in the forced eviction (Lenggenhager, 2018, p. 167).
Both national parks -Nkasa Rupara (inaugurated as Mamili NP; in van de Vegte et al.: Nkasa Lupala) and Mudumu -were proclaimed on 1 March 1990, shortly before the declaration of Namibia's independence. Before proceeding to the period following the proclamation of the national parks, we present the settlement history from the perspective of the Lyanshulu residents, accompanied by archival accounts referring to the Makombwe family.

Resettlement History of the Makombwe and Mufalali Families
We focus on reconstructing the settlement history of two families -Makombwe and Mufalali -on the Qxkatwa and Kweka islands and the riverbanks in Balyerwa and 9 The abandonment of these structures in the 1980s was used as an argument that the existing facilities were in a state of decay and it was imperative that conservation be implemented in the region. This was accomplished in the following years by establishing the Mudumu NP and Nkasa Rupara NP, while community conservation schemes followed at the end of the 1990s. These constitute the basis for the tourism and trophy-hunting exploitation of the Zambezi Region's conservation landscape (Bollig & Vehrs, 2021). At the time of van der Vegte et al. 's report (1983), only the Forest Reserve in the north of the region and the Caprivi Game Park in the west existed, both designated in the 1960s. 10 As Lenggenhager (2018, p. 162) points out, few publications cover the Caprivian history of the 1980s and many 'records relating to the Administration for Caprivians seem to have failed to have found their way into the archives.' One can assume that there is some material on this case in the archives of the South African Defence Force (SADF) in Pretoria. 11 The two Mayeyi households of Makombwe and Mufalali were also under the authority of the Mamili chieftaincy during that time.
Lyanshulu until they were evicted in the early 1980s. Generally, a family's entity is organized around a village head (induna) (often the oldest representative). The core of the family (parents and their children) often comprises one household (lapa), and the married siblings and children of the induna form separate households. This kind of village organization produces many small households, all of which are affiliated with the village head. Thus, while we discuss only two evicted families, this encompasses many households and a much larger number of people than might seem apparent (Figs. 1, 4, 5, and 6). 12 Figures 1, 4, 5, and 6 illustrate the places we visited during the cultural mapping. Figure 1 shows an overview of the study area and the settlements, whereas Fig. 4 displays the Qxkatwa and Kweka islands in detail, where the Makombwe family initially lived and where the two families settled together after 1960. Figure 5 covers the early settlement history of the Mufalali family before they moved to Qxkatwa and Kewka islands to join the Makombwe family. Finally, Fig. 6 shows the late settlement history of the 1970s on the eastern riverbank before the forced eviction took place. The detailed Figs. 4, 5, and 6 also display selected gravesites of important community members, the tree where the people were vaccinated by Kruger in the 1940s, 13 and the harbours (mukoro landing sites) 14 that are This resettlement history refers exclusively to those people who were evicted in the 1980s; we do not include the settlement histories of other families (of different ethnic backgrounds) that lived in the region at other times and that also migrated out before the eviction. 13 The elders refer to the vaccination campaign as being for sleeping sickness. However, tsetse fly infestations were addressed by the South African administration through ground and air campaigns in the 1940s, and again in the 1960s and 1970s (Bollig & Vehrs, 2021). However, as Kreike (2010) points out, smallpox outbreaks were recorded for Ovamboland and the Caprivi in 1929, and a vaccination campaign was executed in the Oukwanyama district in 1945, three years after another smallpox outbreak. It is possible that Lyanshulu people are referring to a smallpox vaccination campaign at Lidimwe. 14 Harbours are central places in this fluvial landscape that were constantly used with the canoes and are therefore still well remembered. Even though the term 'harbour' might be often associated with infrastructure not present in this case, we believe this term is appropriate here. particularly important for mobility in this fluvial environment. Furthermore, former animal pens, and places where dry-season gardens and wet-season fields were established, are also mapped.

Settlement of the Makombwe Family on Qxkatwa Island
The first settlement recorded is the village called Mwanahakuwu, with about 30 lapa, 15 established in the 1880s, followed a few years later by the village of Kawayire in the 1890s and Lidimwe in the mid-1890s. Mwanahakuwu was the first village on the eastern side of the Kwando River (Fig. 1). Some agricultural fields were established around the villages; however, gathering and hunting were the most important Mayeyi livelihood strategies, and farming and also goat and cattle husbandry had only limited relevance in the late nineteenth century. When the settlement in Lidimwe was abandoned, the Makombwe family split in two.
One segment moved out of today's Mudumu NP territory towards Lizauli (some kilometres north), while the other remained in the region and moved to the island of Qxkatwa.
The first settlement on Qxkatwa (also spelled Nakatwa) was established at the end of the nineteenth century in the south-east of the island by Shita Mushakwa from the Makombwe family, who was buried in close vicinity to his first homestead in the early twentieth century (Fig. 4). Shita Makombwe, his son, became the next headman and lived on Qxkatwa Island until his death in Lidimwe in the 1940s (Fig. 6).
The Makombwe family also had a settlement in Limbo Lisangani in the 1930s and went to Lidimwe in the early 1940s (where they resided next to a Mbukushu village and where Kruger vaccinated all residents) (Fig. 4). The settlement is described as quite extensive, although people frequently moved between Qxkatwa and Lidimwe during these years, and residential construction reflected this transient lifestyle since they consisted mainly of meshed wooden structures not covered with clay mixtures like more permanent structures in other locations. Dry-season gardens were also established next to the Rushika channel. When the seasonal watercourse dried up, gardens on the channel bed were established where maize, sorghum, millet, pumpkin, and melons were cultivated along with other crops. Furthermore, more extensive rain-fed agriculture was now (indicated as 'fields' in the maps) established in the rainy season. There was no animal husbandry.
In the late 1940s, the Makombwe family moved back to Limbo Lisangani, where they stayed throughout the mid-1950s when they moved to a place known as Matako, closer to the island of Qxkatwa, where they remained for a few years. In 1958, significant flooding occurred, forcing them to move to the main island of Qxkatwa. However, they continued their seasonal migration to Lidimwe for farming in the dry season 16 while practicing rain-fed agriculture on Qxkatwa (Morton, 1996, p. 55): Floodplain fields were far more productive and reliable than those in dryland areas, and the Bayei [sic.] in particular preferred this form of cultivation. They cleared fields as the dry season began and the floods from the north began to arrive. Flood waters would then inundate the fields for several months between June and September, depending on location. When the waters subsided, crops would be immediately planted and relied on rainfall for additional moisture. Such fields could produce harvests even in drought years, […]. Such fields in the modern era produce twice as much grain as dryland fields, and very likely did so in precolonial times. When the Makombwe family moved back to the main island of Qxkatwa in the early 1960s, they were accompanied by the Mufalali family and lived on both islands at different times.

Settlement of the Mufalali Family in Balyerwa in the Twentieth Century
The settlement history of the Mufalali family in Balyerwa started in the 1940s when they settled close to the Botswanan side of the Kwando River in Wapa (the name of a water pool). In the early 1940s, the village comprised around 30 lapa., but due to a tsetse fly (Glossina morsitans) infestation of their settlement and pastures, followed by an order from the South African administrative officer Kruger (see Gumbo, 2010;Morton, 1996), 17 they moved to the Namibian side with a large herd of cattle and started a new settlement in Muzungu I (Fig. 5. The area is named Li#ana). 18 The Mufalali family stayed here for one year before moving on to Muzungu II village, where they lived from 1942 to 1947, before establishing another village nearby, Muzungu III, where they stayed until 1961 when they joined the Makombwe family on the islands of Qxkatwa and Kweka where they believed it would be easier to harvest specific resources (such as water lilies, which were then a staple food). Besides the two Mayeyi families, a Mbukushu family also lived in the region for a short period, probably in the late 1930s and early 1940s. The Hambukushu of Balyerwa lived at Liwe (Fig. 1) for about five years before moving to Lidimwe (where they met with the Makombwe family). The settlement in Liwe was considerably smaller than the Mayeyi villages in Li#ana and consisted of about 10-15 lapa (with about 50 to 80 people). The Hambukushu family moved out of the region before the eviction of the 1980s.
Millet, sorghum, melons, and vegetables were cultivated in the surrounding area during the wet season (Fig. 5). As well as wet-and dry-season farming, considerable resources such as game meat, fish, water lilies, wild fruits, and honey were acquired through hunting and gathering. Cattle -often used for ploughing fields -were also kept in the vicinity of the villages. However, due to the tsetse fly prevalence in the riverine area, cattle suffered severe losses from time to time; and in 1947, the Mayeyi people in Balyerwa lost all their cattle in one outbreak. 19 This was when the first goats were acquired from other Mayeyi families in the Nsheshe and Mbambazi areas. By that time, 'white men' (mukua) were buying the skins of wild animals, which also made acquiring goats easier. 20 Another important place, Shuno, is a large water pool on the riverbank that marks the limits of the fields at the time of the settlements of Muzungu I, II, and III) and is also where wildlife used to gather. In the 1940s and 1950s, Shuno was also considered the boundary between the Makombwe and Mufalali families used to delineate hunting rights.

Common Settlement of Makombwe and Mufalali Families in the Riverine Area in the Mid-Twentieth Century
The common settlement of the two families on the islands in the Kwando River lasted from the early 1960s to 1970. Over time, the families inhabited different villages on both Qxkatwa and Kweka islands (Fig. 4). Mufalali and Makombwe families at times occupied more than one village on Qxkatwa in the 1960s, and both families had a village and a cultivation site on Kweka Island. It is, therefore, important to note that the social organization in terms of families does not always correspond to the settlement structures. There were also times when family members split up or reunited.
On Kweka Island, villages were quite similar in size, the Mufalali village having about 15-20 lapa and the Makombwe village 10-12 lapa. In 1970, Kruger relocated the settlements on the islands in the Kwando River, offering villagers food in exchange for relocating to the Lyanshulu area on the riverbank.

Relocation of Makombwe and Mufalali Families to the Riverbank in the 1970s
Two main phases mark the resettlement to Lyanshulu between 1970Lyanshulu between and 1980Lyanshulu between : until 1973, the two families lived in separate villages close to the river and not far from the 17 The resettlements in the 1940s and the 1970s were in response to tsetse fly infestations, and our informants emphasized the peaceful character of these relocations,who reporting that these were negotiated agreements. However, not all relocations in the 1940s were peaceful, as Kruger (1984c, p. 22) recounts: 'Many of the Mayeyi people had been living on the large islands of Lupala and Nkasa but owing to the Tsetse Fly there and the dangers of Sleeping Sickness Brittz had moved them out, rather against their will I gathered.' Superintendent Brittz (the first administrator of the South African administration in Schuckmannsburg from 1929 to 1939) also 'warned the Bechuanaland tribesmen on the island that, as there had been a change in control, they could no longer continue to occupy the island' (Kruger, 1984a, p. 38). harbour; after 1973, they established a large, common settlement known as Lyanshulu. At the time of their move, none of the families had cattle, 21 but they owned considerable herds of goats (an estimated total of more than 100 goats) kept on a small island opposite the harbour (Fig. 6), where they were not affected by the tsetse fly, were protected from predators, and were unable to reach the croplands. During this time, people either had to bring in cattle to plough their fields or use manual labour, was also noted by Kruger (1984c, pp. 21-22): The people of Lianshulu [sic.] lived at that time on the island, the village head being a good man named Makombwe.
[…] the Tsetse Fly still (1976) prevents the keeping of cattle by Makombwe and a few other nearby villages due, it would seem, to re-infestation from the other side and possibly inaccessibility of some places for spraying. Makombwe and his neighbours have to be content with keeping goats, the movements of which are closer to the village. They also have to use hoes for cultivating their fields […].
The common Lyanshulu settlement existed from 1973 until the early 1980s. It consisted mainly of the same structures as standing today (a mopane-wood construction with clay-reinforced wall and a grass-thatched roof). However, a few houses already had iron sheet roofs.
In the early 1980s, the Makombwe and Mufalali families in Lyanshulu were abruptly evicted by a nature-conservation law-enforcement unit that burned the village. Memories of this day remain vivid since not only were the houses burnt, but people were not allowed to gather their belongings and had to leave their goats in the kraals on the island (an estimated total of 150 goats).

Forced Eviction of Makombwe and Mufalali Families from Lyanshulu
The only sources of information on the day of the eviction are the descriptions of those who were evicted, who describe being informed by a local government representative 22 that they had to leave all their belongings and immediately board the transport vehicles that had been brought in to go to Libburakurube, which lies south of the later designated national park boundaries. There was no prior consultation or warning of the relocation plans, nor had any negotiations taken place at any time. 23 Any signs of resistance were met with threats of armed physical violence: We were chased from our land with no warning or anything. In the morning around 6 am, we just heard cars outside and they [the law enforcement unit] started telling us to move and we were very surprised as no questions were allowed. 24 (Elderly person from Lyanshulu, anonymous, August 28, 2019) 25 We never saw anyone from the government going to my father [the headman of the Lyanshulu village] telling that they want our land, so that an agreement can be signed. (Elderly person from Lyanshulu, anonymous, August 28, 2019) The people of Lyanshulu insist that they had no opportunity to negotiate the eviction or to identify a new settlement place. 26 The government removed them from their residential area, designated the Mudumu NP some years later. Thus, they not only lost their homes, livestock, structures, and belongings, but also access to the graves of their family members and their historical and social institutions (khuta). 27 The South West African administration made promises about constructing a school and a clinic, building houses, opportunities for employment in the future tourism industry around the Mudumu NP, and establishing fields for 22 We were unable to obtain more detailed information about the people in charge of the eviction and destruction of property. However, the Caprivi government had started their planning of conservation areas in the 1970s, in close cooperation with officials from Windhoek and Pretoria, and both traditional authorities of the Zambezi region were part of the Caprivi government at that time. 23 We were able to find the minutes of the 'Cabinet Meetings 1979-1980. Department of Justice and Interior' in the Namibian Archive in which 'Practical problems involved?' were discussed in the context of conservation planning, and the 'Removal of people at Lyanshulu' was on the agenda (Caprivi Government, 1979, p. 3). No, decisions were made according to this, but the two traditional authorities R.M. Mamili (Chairman: Chief Minister) and J.M. Moraliswani (Minister: Department of Justice and Health) were present. However, during our research we could not ascertain if or why the communication was interrupted at this point, and found no archival records that could shed further light on the issue. 24 Lenggenhager (2018, p. 167) also reports on the forced nature of the eviction by officials from the Directorate of Nature Conservation, who came with lorries, threatened to burn the village, and made extravagant promises about the new settlement scheme. 25 All interview partners are rendered anonymous. 26 They also refer to earlier resettlements by the colonial administration (e.g. through Kruger) that they accepted after consultations between the colonial officials and local authorities. The previous resettlements were predominantly based on political concerns. In the early twentieth century, residents moved across the border into the South West African territory to escape a planned taxation in the Betchuanaland Protectorate, and again in the 1960s and 1970s, people were moved from the riverine areas due to large-scale spraying campaigns against the tsetse fly. 27 We gathered a list of 52 people buried on the islands and in Lyanshulu and Balyerwa, but even this is incomplete (see Fig. 1, 4, 5, and 6). 21 However, a smaller cattle kraal was established around a termite hill (Fig. 4), indicating people owned cattle for at least a short time, though this situation might not have continued for long due to the tsetse fly. The construction around the termite hill ensured that the rainwater drained from the kraal during the wet season. cultivation; according to the evicted families, none of these promises has materialized.
The land allocation for residences and cultivation where the village residents were removed continues to be a contentious matter. The Lyanshulu people are being forced onto the least favourable land, and many have moved directly to the southern border of the national park, designated the 'New Lyanshulu gate set' (Fig. 1). No official land rights were designated to them on the land they now occupy, and the legal board that mediates land-use conflicts (the local khuta) was riven with problems for some time, as three different khutas were established during political frictions that arose in the new area of residence. Furthermore, the political conflict had reached such an intensity that some families have changed their affiliation between the two traditional authorities (from the Mayeyi TA that would typically be in charge to the Mafwe TA under the Mamili family that was in charge before the Mayeyi TA was inaugurated).
Those evicted from Lyanshulu retain a strong sense of belonging to their ancestral lands and a strong desire to return to them. They have been refused permission to return to their former settlement (New Era, 2015;NBC, 2016). They are also concerned that the present government is turning a blind eye to the violations committed by the colonial administration.
Every time we report it [our situation] to the government and the traditional authority, they tell us to go back home with no positive response. Would they please just let us go back to our old Lyanshulu and stop calling it a park, because that is our home. (Elderly person from Lyanshulu, anonymous, August 28, 2019) Since they were evicted under the colonial administration, the people of Lyanshulu had hoped that this injustice would be rectified after Namibia's declaration of independence. But their demands and repeated attempts to raise their concerns with government officials continue to be ignored by the current authorities, conservation institutions, and traditional authorities. A small hope arose with the recent report of the Ancestral Land Commission, discussed below (Republic of Namibia, 2020). Nevertheless, the Mayeyi Traditional Authority's recent decision overshadowed these hopes. It excluded one of the two families from the Balyerwa Conservancy after political frictions emerged and the Makombwe family disaffiliated from the Mayeyi TA. This dispute may also result in a second displacement, as suggested by the Mayeyi TA in an official letter to the MEFT (Mayeyi Traditional Authority, 2022), this time from a community-based conservation area. As the Lyanshulu people claim, there are no prospects for them where they currently live due to the lack of any security in terms of tenure and land use, as well as the risks attached to living so close to the park, such as human-wildlife conflicts (New Era, 2017).
Their aspirations are evident: they want to return to Lyanshulu. But all their efforts to gain recognition of their case from local authorities and government officials have failed. 28 In Namibia, there are examples of other groups living or having lived (until their evictions) in regions that are today national parks, such as the Hai//om in Etosha NP (Dieckmann, 2007) or the Khwe in Bwabwata NP (Koot et al., 2016). While Lyanshulu people criticise government authorities to ignore the Lyanshulu case, the difficulties of litigation with the Namibian state and the unforeseen ramifications involved, as observable in the Hai//om case (Koot & Hitchcock, 2019) and in Bwabwata (Taylor, 2012;Hitchcock, 2012), are not much acknowledged by the Lyanshulu families.

The Fight for Recognition and the Ancestral Land Rights Report
The two families who reside south of Mudumu NP struggle with the consequences of displacement. Their eviction in the 1980s, along with the subsequent events and failure to recognise the inequities of the situation, have led to challenging living conditions for the Lyanshulu people today. The failure of the South West African administration to offer compensation for their eviction before the declaration of independence, and the continued absence of an adequate response by the subsequent Namibian government and traditional authorities in charge, have led to severe resentment and general distrust towards these institutions.
In 2020, the 'Commission of Inquiry into Claims of Ancestral Land Rights and Restitution' submitted its nearly 800-page report to the Namibian President's office. It identified that 'most testimonies from Zambezi Region (88%) make claims of ancestral land loss due to colonial dispossession by means that included proclamations of national parks and forced removal' (Republic of Namibia, 2020, p. 206), and it lists the demands of those evicted and suggests viable compensation measures. In the Lyanshulu case, the Mayeyi Traditional Authority's contribution to the report claims: The colonial policy on forced removal was exercised on this community and forced them to move their residence to another area occupied by another community. This generated animosity between the two communities forcing the other community to resort back [sic.] and live closer to the Mudumu National Park. (Republic of Namibia, 2020, p. 176) As laid out in the government report, restitution can take several forms. These are listed as (i) restoration in full or in part of areas claimed as ancestral land, (ii) compensation for areas claimed as ancestral land, (iii) alternative land, (vi) renaming of places, (v) payouts from businesses and from the exploitation of natural resources on areas claimed as ancestral land; (vi) recognition of heroes and heroines, and erection of statues in their honour; (vii) community land to be regarded as ancestral home; (viii) preferential treatment in resettlement programmes; (ix) recognition in land reform and resettlement policies; (x) erection of monuments in areas of historical, educational and spiritual value; (xi) access to grave sites of ancestors; and (xii) reburial of victims of genocide and wars of resistance (Republic of Namibia, 2020, p. 30). Based on the above claims, the report recommended the following measures for the case of Lyanshulu: (i) The community of Lyanshuru [sic.] be allowed to revert to their original area of sojourn and other areas so determined by the Commission… [sic.] (ii) Ancestral land rights and restitution be done to the satisfaction of the entire community without favour or discrimination; (iii) Resolutions of the Commission should be shared to all participants in the region; (iv) The Commission restores any part of the region where ancestral land rights are infringed upon by others; (v) A mechanism be put in place to allow participants to appeal against unsatisfactory resolutions of the Commission.
(Republic of Namibia, 2020, p. 176-177) From the perspectives of the Lyanshulu people, two strategies could be applied: either relocation to and rebuilding of their former settlement in Lyanshulu, including land-use rights in the Mudumu NP, or the implementation of compensations and the configuration and legalization of land rights in the area south of the park. How the displacement can finally be addressed, and whether solutions from other countries, such as the South African contractual national park model, could serve as a role model for the direct involvement and participation of displaced communities (as described by Reid, 2001 for the case of the Makuleke community) remain open questions.
The theoretical implications of possible compensations or even a return to their ancestral land are highly speculative but address important questions; for instance, whether the human-wildlife conflict (HWC) is already more intense in the direct vicinity of the park. It remains debatable whether a return to 'Old Lyanshulu' is at all possible with hundreds of cattle and the need to build new infrastructures such as schools and roads and provide water and electricity access, while at the same time, population numbers have greatly increased compared to the 1980s.
More generally, conservation planning and the conservationist movement that hopes to increase the extent of protected areas (i.e., 30% in 2030, or the half Earth Initiative) must consider the importance of addressing historical injustices regarding conservation planning and implementation to make biodiversity conservation successful in the long run. In this context, new, large-scale projects such as the KAZA TFCA, with their recent history of transboundary conservation, must also take on the histories of the smaller conservation units that they have absorbed to avoid being blamed for perpetuating colonial structures and injustices.
Moreover, the Lyanshulu case described here is similar to other conservation-induced displacements (see Dieckmann, 2020, on the Hai//om in Etosha, and Chiweshe, 2022, for Nambya people in the Hwange N.P. in Zimbabwe). Of course, all cases are unique and must be understood within their specific contexts. However, conservation strategies must prove that they can address environmental injustice and the displacement of people, as Larkins' (2021) critique of conservation also highlights, but also practical issues arising for residents are directly affected by conservation measures. The adoption of CBNRM strategies is expected to address these issues, as it emphasizes a direct link between conservation and residents.

Outlook: CBNRM and Environmental Justice
The question remains to what extent current conservation institutions can address issues of environmental justice that are fundamental not only for the future of the Namibian conservation path but also issues of environmental equity and environmental rights in general that go beyond human rights. Two questions face conservation planners: conservation for whom, and at whose expense? Kiaka (2018, p. 26) addresses the underlying assumptions of the CBNRM scheme in Namibia in which 'equity is presumably already inoculated into its design by devolving power and authority over resource management from central governments to the communities.' However, the Namibian CBNRM design does not meet the high expectations of participation, devolution of power, and a fair distribution of benefits, but rather 'emerges as an arena of asymmetrical power struggles between hierarchies of knowledge, land use practices, domination and subordination' (ibid: 27).
If the CBNRM programme wants to address the environmental injustices of the past, it must attend to different dimensions of justice within the environmental justice framework. To achieve restorative justice, which aims to restore the capabilities of local communities (see Zhu and Lo (2021), three different goals must be achieved: recognitional justice, distributional justice, and procedural justice. In the Lyanshulu case, this would mean that the displacement is officially recognized (beyond the deficient and technical reference in the report of the ancestral land commission) and publicly addressed. Beyond the communication with the displaced families and an official statement of the Namibian government, this would also include that the settlement history is officially displayed as part of the park's history -something that currently does not exist. Furthermore, such recognition would allow debate about compensations for those families evicted from their ancestral lands. This task is complex since the eviction took place 40 years ago, and many conflicts arose among those residents now living in the region south of Mudumu NP, entailing not only the identification of monetary and infrastructural assets but also the question of land rights and ownership. Lastly, restorative justice refers to enabling displaced families to have secure access to land and the capacity to build a viable future, as well as a role in all decisions about the park and a share of revenues derived, i.e., from tourism enterprises, inside Mudumu NP.
So far, the opportunities for participation of the Lyanshulu people in national park structures identified in the planning process have been minimal, and apart from a few jobs in the ranger station and the lodge, represent only options to offer cheap wage labour, sell homemade handicrafts along the roadside to passing tourists, or perform at times bizarre shows for visitors in the 'cultural villages' in order to convey an image of 'traditional' ways of life. In our understanding, the Namibian government needs to respond to these injustices and address rehabilitation in the face of past events, and the identification of ways forward. It is vital to address the history of colonial injustices in Namibia and build a locally grounded legitimacy for the protected areas being created.
However, the recent communication between the Mayeyi Traditional Authority and MEFT indicates that they shall no longer be a part of the Balyerwa Conservancy and shall relocate from their current settlement area. Dealing with this new situation, which is essentially a political one, will show to what extent CBC schemes can address conservationrelated issues and community-based conservation's social and political challenges. If the Makombwe family is displaced again, it will show that CBC cannot deal with the wicked history of conservation.