Cameroon: Floodplain Rehabilitation through Prescribed Flooding of Waza-Logone

Project Stage:
Implementation

Start Date:
1992-10-30

End Date:
1992-10-30

Primary Causes of Degradation

Degradation Description

Since the 1950s, demand for irrigation water and electricity has resulted in the disruption of natural flooding regimes in river systems throughout Africa. In 1979, the Waza-Logone floodplain became yet another ill-fated system with the construction of an embankment along the Logone River and a dam forming Lake Maga. This dam, in combination with lower than average rainfall for the last two decades (Beauvilain 1995), has led to a significant reduction in the depth and extent of flooding over an area of 1500 km2, including Waza National Park. Consequently, productive perennial grasses have been supplanted by stands of annual grasses, limiting regrowth in the dry season and reducing the area’s carrying capacity for livestock, fisheries and wildlife (Scholte et al. 1996). Numbers of Kob antelope, for instance, dropped from 20,000 to 5,000 in the period 1979-1983, and declined further to 2,000 by the end of the 1985 drought. Ironically, the cultivation of irrigated rice, the main purpose for the Maga Dam, was largely a failure, as attested by the use of less than half of the irrigation scheme’s capacity and the continuing importation of rice.

Reference Ecosystem Description

Pre-dam vegetation was dominated by the perennial grasses Echinochloa pyramidalis (antelope grass), Vetiveria nigritana (black vetivergrass), Hyparrhenia rufa (jaraguagrass) and Oryza longistaminata (long-stamen rice) (Wesseling et al. 1994). Other common grasses included: Sorghum arundinaceum, Dinebra retroflexa, Pennisetum ramosum and Ischaemum afrum. According to Barth (1857) and Nachtigal (1889), the levees along water courses were far more densely wooded in the mid-nineteenth century than nowadays, most probably with the same dominant species (Celtis africana, Ficus spp, Tamarindus indica), although some Sudanian tree species, scarce nowadays, were mentioned as well (Vitellaria paradoxa).

Vegetation maps of the entire area and Waza National Park, based on aerial photographs from the mid 1960s (Gaston & Dulieu 1976 and Wit 1975, respectively), indicate that presently inundated floodplain communities used to extend 10-15 km further west. Other sources (Letouzey 1968) indicate that inundation depth was formerly 1 to 2 metres, whereas at the present time it is <1 metre.

Fauna of the Waza-Logone area includes: elephants, Kobs, giraffes, Roan antelope, Red-fronted gazelles, lions, jackals, hyenas, and more than 370 different species of resident and migratory birds spread across eight habitat types. While many of these avifauna species are important components of the ecosystem, the endangered Black-crowned crane (Balearica pavonina pavonina) is considered the area’s “flagship” species.

Project Goals

The project’s overarching goal is the hydrological and ecological rehabilitation of the floodplain in order to restore, to the extent possible, the physical conditions that existed before construction of the dam in 1979 (IUCN 1999).

The pilot reflooding experiment conducted in 1994 was intended to reinstate the natural flooding regime in an area of 600 km2 and to assess the rehabilitation capacity of vegetation disturbed as a result of the dam. Ultimately, these efforts are aimed at providing data for the evaluation of larger-scale reflooding options.

Monitoring

Monitoring Details:
Pre- and post- flooding monitoring were major components of the project (see the Project Planning and Design, Key Lessons Learned, and Long-term Management sections for details). Ongoing monitoring is conducted by the Government of Cameroon and IUCN

Description of Project Activities:
Studies conducted as part of the Waza-Logone Project in 1993 identified a number of reflooding options that were expected to have a major impact on the floodplain, including its human population. The pilot reflooding was initiated in 1994 by breaching an embankment that had blocked a former watercourse, the "˜Petit Goroma.' Through this newly opened watercourse, an additional flow of approximately 20 m3/s streamed into the floodplain. In 1997, another watercourse, the Areitekele, located just south of the Petit Goroma was opened, triggering a further 8-10 m3/s inflow into the floodplain. In order to evaluate the effect of this prescribed flood, extensive pre- and post-flood monitoring was conducted. Reference data had been collected along a 35-km transect in 1984 and 1985. The transect, consisting of six permanent plots each 20 x 50 m, started at the western edge of the floodplain, where Vetiveria nigritana (a perennial grass) used to be dominant but where annual vegetation had come to predominate (Wit 1975). The transect ended in the East in the well-inundated floodplain dominated by perennial grasses little changed since Gaston & Dulieu (1976) and Wit (1975). In the desiccated floodplain, 23 permanent 10 x 10 m plots were installed in a 1 x 0.5 km area around Tchikam. From 1993 onward, prior to the pilot reflooding, vegetation monitoring was re-activated in the same Tchikam plots as in 1985 and at a number of additional plots. In 1993, the number of permanent plots in the transect was increased from six to 40, and the new 6 x 6 m plots were located at intervals of 0.5 km. Floristic composition was monitored in a newly installed 20 x 16 km grid, situated in the centre of the pilot-release impact zone, upstream from the transect and Tchikam plots. The grid had 136 (6 x 6m) plots located at a distance of 0.5 or 1 km.

Ecological Outcomes Achieved

Eliminate existing threats to the ecosystem:
The vegetation on the newly-inundated floodplain is typical of African seasonally-flooded grasslands, with its species-poor perennial stands (Denny 1993). The 1993-2000 monitoring, supported by field observations in 2002, indicates that in 2002 the perennial rhizomatous grasses Oryza longistaminata and Echinochoa pyramidalis occupied almost the totality of the reflooded area once again. Indeed, cover of perennial grasses in the reflooded zone, most notably these two species, increased from 41% in 1994 to 62% in 1997 and 75% in 1999 (a 7% annual average). The limited number of observations in the grid above the 1994 floodline did not allow a clear distinction in the perennial species' increase, although it seems to be initially only Panicum anabaptistum and Ischaemum afrum. At the same time perennial-grass cover was increasing in the reflooded area, a decrease in the relative cover of annual grasses and herbs was also observed. Between 1994 and 1999, relative cover of these species decreased from 58% to 23%. Sorghum arundinaceum, an annual grass dominant since the mid-1980s, experienced an overall decrease in relative cover in the reflooded zone from 26% to 16% between 1994 and 1997, despite higher inter-annual fluctuations. In areas above the 1994 floodline, S. arundinaceum still fluctuated around 35%, but from 1998 onward, coinciding with the 1997 extension of the floodline, relative cover above the 1994 line dropped dramatically as well. If the conversion of annual to perennial grassland proceeds at the same rate as that observed in the period 1994-1999 (i.e. approximately 7%), the recovery of a 100% perennial-grassland state would be achieved after the 2003 flooding season. In the period 1992-2000, the number of waterbirds observed during the dry season increased from 60,000 to 105,000. In this same period, 19 waterbird species were present in numbers surpassing 1% of their known population size (Wetlands International 2002). Furthermore, the number of species meeting this criterion in yearly observations doubled between 1992 and 2000. A few species surpassed the 1% mark in all years of the study period: White-faced whistling duck (Dendrocygna viduata), numbering up to 6% of its known population size; Black-crowned crane (Balearica pavonina pavonina) (up to 16% of its population); and Collared pratincole (Glareola pratincola) (up to 39% of its population). The Black-headed heron (Ardea melanocephala) colony, already of exceptional size before the start of the reflooding, tripled during the study period from 750 to a total of 2500 nests. This colony is now a multiple of the next largest known colonies. Because the increase in Ciconiiformes in Waza-Logone was not paralleled by similar trends in other West African floodplains, it is concluded that floodplain rehabilitation played a significant role. That being said, large piscivorous Ciconiiformes--Leptoptilos crumeniferus, Mycteria ibis and Pelecanus rufescens--enjoyed only limited increases, and this is attributable to the repeated destruction of their breeding colonies, resulting in a general population decline, as observed elsewhere in West Africa.

Factors limiting recovery of the ecosystem:
Vetiveria nigritana was dominant on the floodplain prior to construction of the Maga Dam, but it failed to regenerate after reflooding. A tussock grass, V. nigritana is dependent upon seed dispersal as its means of propagation. Whereas the tussock grasses Ischaemum afrum and Panicum anabaptistum, also dependent upon generative propagation, appeared as quickly as the dominant rhizomatous species, V. nigritana did not regenerate at all. The 1996, 1999 and 2002 monitoring showed V. nigritana tussocks to be in generally good condition, producing seeds in abundance in an area where vegetation cover is low, potentially permitting gap phase replacement (Van der Valk 1992). In the eastern part of the transect, where V. nigritana lost its dominant position to Oryza longistaminata and Echinochloa pyramidalis between 1985 and 1993 and despite the continuing flooding, it is assumed that the strong competitive nature of these species prevented its regeneration. However, further west in the transect, where V. nigritana used to be dominant and is still present in depressions, the lack of regeneration might be attributed to changes in the hydrological regime--namely, the late arrival of floodwater and/or the low sediment level of floodwater. In spite of this project's efforts to restore the pre-dam hydrology through prescribed flooding, major changes in the river system's hydrological regime have persisted. Whereas floodwater from both the Logone River in the east and smaller watercourses draining the Mandara Mountains in the southwest once inundated the Waza-Logone floodplain, these smaller channels are now trapped by Lake Maga. Consequently, the flooding season is now shorter, and floodwaters from the Logone River arrive relatively late--particularly in the western part of the floodplain. In the Tchikam plots and the western portion of the transect, for example, floodwater does not arrive until late October, and by this time, the area has already dried up following the rainy season. Moreover, water traveling to this western area must flow over 40 km, through dense vegetation that reduces its speed and leads to a loss of sediment load, resulting in less fertile soil conditions and a lack of sediment downstream (Breen et al. 1988). It is thought that these changes in hydrology might explain the failure of Vetiveria nigritana to regenerate in some parts of the study area. The recovery of human-supported resources, such as grasslands for cattle grazing and heron nests protected by villagers, exceeds the increase in resources with which humans compete--i.e. Kob antelope and large piscivorous waterbirds. Human populations in the region have been quick to respond to the improved conditions and are out-competing wildlife in many cases. The isolated Waza-Logone population of 4,000 Kob antelope had a limited capacity to increase before the reflooding--at most by an annual rate of 50% (i.e. 2,000 antelope). Thus, the "˜reservoir' of 4.7 million cattle in the Conventional Lake Chad Basin (Martin et al. 1996), of which 200,000 visit the floodplain, represents a significant source of competition. A desire to redirect migration routes and take advantage of new resources has brought a growing number of villagers to the floodplain and threatens to increase grazing intensity to a point beyond which wildlife (in this case the Kob) cannot compete. Some species of water birds have also suffered as a result of competition with the local human population. Among the water birds that showed a trend less than the general two-fold increase were the large, predominantly-piscivorous resident Marabou and Yellow-billed Storks. These species have come under increasing persecution from local fishermen, and their colonies have been repeatedly destroyed in many cases. Although this is a relatively recent development in Waza-Logone, it has been common elsewhere in West Africa for many years.

Socio-Economic & Community Outcomes Achieved

Economic vitality and local livelihoods:
The Waza Logone project has not only improved the livelihoods of local communities by bringing back water, it has been active in numerous areas to create a movement for sustainable development. For example, conflicts between herders and agriculturists, once rife, have disappeared by negotiating new rules for transhumance and compensation for damages. The provision of clean drinking water has reduced water-borne diseases by seventy percent. Finally, the project helped set new rules for burning of grasslands to improve their nutritional value for cattle and wildlife, and remove parasites.

Key Lessons Learned

Floodplain rehabilitation should be conducted in a probing approach with simple and long-term (“˜continuous’) monitoring that allows interventions and regular redefinitions of targets. Because of expected imbalances in Conservation versus Development, appropriate mitigating measures should be prepared beforehand, e.g. by the designation of distinct areas for conservation and other competing land uses. In addition, one may consider zoning with exclusive rights for present local communities that form a social buffer against the installation of newcomers. Formal agreements on the exclusive use of resources may facilitate such measures. To enhance the efficiency of conservation, it will be important to increase the intervention capacity of protected area personnel well before major changes, such as those induced by floodplain rehabilitation, are undertaken. This often implies that institutional arrangements are taken up early. In the long term, one aspires toward a governmental commitment for conservation and local development in the way they have been mobilised in Cameroon for security in the late 1990s. In addition, one should strive toward more coherence between local, national and international conservation policies and interventions.

The following are five key lessons learned throughout the implementation of this project:
1. The importance of considering different response times of ecosystem components, and their implications for rehabilitation results
2. The difficulty of setting aside resources for conservation due to the tendency of plants and animals to distribute themselves evenly according to resource availability (i.e. the Ideal Free Pre-emptive Distribution Model)
3. The risk of human population build-up at the border of protected areas, a process accelerated by the integration of conservation and development
4. The need to secure the commitment of all parties in consensual protected-area management
5. The need to have adequate training for protected-area personnel and to ensure that all parties have effective capacities for intervention

Long-Term Management

The large impact of the pilot reflooding stressed the need to strengthen the management of Waza National Park and its direct surroundings in order to cope with the rapidly changing situation, especially the arrival of pastoralists and fishermen to exploit the newly reflooded areas. A management plan was formulated in collaboration with local communities (sedentary as well as nomadic), protected area authorities, traditional and administrative authorities and the Ministry of Environment and Forestry. Experiences with the implementation of this management plan highlighted the incapacity of park personnel to deal with new forms of collaboration with local communities. This prompted the implementation of a training program at the Garoua Wildlife College for guards of the national parks in north Cameroon. Courses were developed to enhance the knowledge and skills of protected-area personnel and encourage dialogue with local communities to improve conservation of Waza-Logone and other protected areas in Cameroon and West-Central Africa.

One community-level approach tested in Waza-Logone was social fencing, where local communities exclude “˜outsiders’ from exploiting the surroundings of a protected area, an approach that is no longer taboo (Anonymous 2004). This was part of management agreements between protected-area authorities and local communities, where community management of a sensitive area was rewarded by exclusive exploitation rights, as well as other privileges like small-scale development assistance.

Sources and Amounts of Funding

20 million SFr The budget for the Waza-Logone floodplain rehabilitation totals 20 million SFr. Funding is provided principally by the Netherlands Ministry for Development Cooperation (DGIS), but additional grants and in-kind support have also been obtained through the World Wide Fund for Nature (WWF), the European Commission and the Government of Cameroon.

Other Resources

Scholte, P., S. de Kort, and M. van Weerd. 2000. Floodplain rehabilitation in far northern Cameroon: expected impact on bird life. Ostrich 71(1&2): 112-117.

Scholte, P., 2005. Doctoral thesis, Leiden University.

Primary Contact

Organizational Contact