Regional Management of Water Resources (Proceedings of a symposium held during the Sixth IAHS Scientific Assembly at Maastricht, The Netherlands, July 2001). IAHS Publ. no. 268, 2001. 73

Effects of damming the in semiarid northern —lessons learnt for future management

B. J. M. GOES* The Netherlands Institute of Applied Geoscience TNO (present workplace), PO Box 6012, 2600 JA Delft, The Netherlands e-mail: [email protected]

Abstract The Hadejia River is a part of the basin in northern Nigeria. Two major dams were built in the upstream part of the Hadejia River. The other main river connecting to the Yobe River is uncontrolled. By comparing discharge for the controlled and uncontrolled rivers it is shown that the first dam led to an average decrease of 33% in annual flow in the upstream part of the Hadejia River. The total annual flow in the Hadejia River further downstream, just above the Hadejia-Nguru Wetlands (HNW), was not significantly reduced as a result of the dams. This is related to a relatively small river flow reduction at lower flows in the upstream part of the Hadejia River. The major impact of the dams on the downstream part of the river is the change in regime from ephemeral to perennial. The introduced dry-season flows created favourable circumstances for the development of weed blockages in the HNW. Due to these blockages the Hadejia River stopped contributing to the Yobe River. Furthermore, after the completion of the dams the timing of the floods in the HNW became less predictable. Despite the decrease of total wet-season discharge relatively large flood extents are still experienced in wet years. This is explained by the fact that almost all the water entering the HNW through the Hadejia River now remains within the wetlands. The paper concludes with suggestions for improvement of the management of the dams

Key words dams; Nigeria; river basin management; Typha weeds; wetlands

THE STUDY AREA

The Hadejia-Jama'are-Yobe River basin is situated in semiarid northern Nigeria (Fig. 1). Most of the flow in the Hadejia River system (-80%) is controlled by (storage capacity: 1492 x 106 m3, completed in 1974) and Challawa Dam (972 x 106 m3, 1992). The Jama'are River is presently uncontrolled but plans exist to build a dam at Kafin Zaki. In the nearly flat middle and lower parts of the basin, the rivers spill into the flood plains during the wet season (June to October). The most extensive flood plain areas in the basin are the Hadejia-Nguru Wetlands (HNW). The Hadejia River splits into three channels in the HNW: the Marma Channel which flows into Nguru Lake, the Old Hadejia River which joins up with the Jama'are River to become the Yobe River and the relatively small Burum Gana River. The mean annual rainfall ranges from over 1000 mm in the upstream Basement Complex area to approximately 400 mm in the middle part of the basin and less than 300 mm near Lake

* Former work place: RJCN-Hadejia Nguru Wetlands Conservation Project. Fig. 1 The Yobe River basin. Effects of damming the Hadejia River in semiarid northern Nigeria 75

Chad. The reservoirs behind the dams supply two large, partly finished, formal irrigation schemes ( River Irrigation Project—KREP, and Hadejia Valley Irrigation Project—-HVIP), and contribute to Kano City Water Supply (KCWS). The traditional farming system in the basin is rain-fed. Flood farming and flood-recession farming provide an important food supplement in the wetlands. Furthermore, the wetlands are important for fishing, groundwater recharge, dry-season grazing and they are ecologically rich. This presentation is mainly a shortened version of Goes (in press.). The river flow data in this paper are taken from Diyam Consultants (1996), and Goes & Zabudum (1996, 1998, 1999).

CHANGES IN FLOW REGIME

Peak flow

Tiga Dam and droughts reduced the average peak flow in the upstream part of the Hadejia River at Wudil (Fig. 1) by 64% between 1964-1973 and 1979-1989 (Table 1), while in the uncontrolled Jama'are River at Bunga the average peak showed a reduction, as a result of droughts only, of 27% between these two periods. Thompson (1995) showed that the relative flow reduction between Wudil and Hadejia increases significantly when the 10-day discharge at Wudil is above 87-145 m3 s"1. The causes for the river flow reduction at relatively high flows require further study but are most likely attributed to flood plain inundation and the presence of at least four non-returning channels with an elevated riverbed where they connect to the main river. The decrease of peak flows at Wudil reduced the inflow in the non-returning channels and the floods. As a result shallow groundwater recharge probably decreased, although the lowering of the ground­ water table in the area might be limited (no reliable long-term groundwater level data are available) because of the relatively high rainfall in the upstream part of the basin. The average peak flow at Hadejia was reduced by 34% after the construction of Tiga Dam which is the same order of magnitude as the peak flow reduction at Bunga (27%) on the uncontrolled Jama'are River that represents the climatic influence (Table 1). In the 1990s the Hadejia peak flow had the same order of magnitude as in the 1960s. So the peak flow at Hadejia has not decreased (1979-1998) significantly as

Table 1 Mean annual river flow and peak discharge at six sites in the Hadejia-Jama'are-Yobe basin before and after the construction of Tiga Dam.

River Site Catch- Pre-Tiga Dam construction: Post-Tiga Dam construction: ment area Mean annual Mean Period Mean annual Mean Period flow: peak flow: peak discharge discharge (km2) (106m3) (mm) (mV) (106m3) (mm) (mV) Hadejia Wudil 16 380 1915 117 946 1964-73 1004 61 336 1979-89 Hadejia 30 430 718 24 99 1964-73 523 17 65 1979-89 Jama'are Bunga 7 980 2061 258 1227 1964-73 1431 179 890 1979-89* Katagum 15 000 1634 109 381 1979-89 Yobe Gashua 62 150 1397 22 182 1964-73 925 15 143 1979-89 Yau 84 138 423 5 37 1964-72f

* Excluding 1987. f Excluding 1968, 1969. 76 B. J. M. Goes

a result of the construction of the upstream dams. This is due to the relatively small river flow reduction between Wudil and Hadejia at lower flows at Wudil.

Annual flow

The annual runoff in the upstream part of the Hadejia River at Wudil decreased as a result of the increased formal water uses and evaporation losses from Tiga Reservoir (Table 1). Another cause of the reduced annual runoff was a number of droughts in the 1970s and mid-1980s. The following method was derived to calculate the effect of Tiga Dam on flows at Wudil. The cumulative mass curve of annual flows at Wudil and Bunga (an upstream site on the uncontrolled Jama'are River) shows a consistent gradient for the pre-Tiga Dam period (1964-1973, Fig. 2). This gradient remained the same through the drought period of 1972 and 1973. Thus, it is assumed that subsequent droughts would have resulted in a similar flow reduction at Bunga and Wudil in a situation with no dams. However, after the completion and filling-up of Tiga Reservoir (1979-1989) the gradient of the cumulative mass curve dropped from 0.95 to 0.64. Therefore, the reduction of annual river flow at Wudil as a result of Tiga Dam is calculated at 33% (i.e. gradient pre-Tiga minus gradient post-Tiga divided by gradient pre-Tiga). The gradient of the cumulative mass curve of annual flows at Bunga and Hadejia, downstream of Wudil just above the HNW, did not change significantly between the pre- and post-Tiga Dam periods (Goes, in press.). An upgrade of recent Bunga river flow data is needed before the impact of Challawa Dam (1992) on the annual river flow at Hadejia can be determined with a cumulative mass curve. Still, a major impact is unlikely since the recent (1994-1997) annual Hadejia discharge is of the same order of magnitude as in the 1960s. So the control structures did not reduce (1979-1997) the total annual flow at Hadejia. This is due to the relatively small river flow reduction between Wudil and Hadejia at lower flows at Wudil and the fact that the formal large upstream water users are not (yet) working at full capacity.

I _____ 1964-73 (pre-Tiga [Dam) : Wudil = 0.95 x Bunga (R2=99%)

1979-89 (post-Tiga Dam): Wudil = 0.64 x Bunga + 6.8 (R2=1 00%)

1974 78: filling up c f Tiga Reserv oir +^ ^1989

—1979 + 1973

0 5 10 15 20 25 30 35 40 45 Cumulative annual flow at Bunga [109 m3] Fig. 2 Cumulative annual river flow relation between Bunga (Jama'are River) and Wudil (Hadejia River): 1964-1989 (excluding 1987; Goes, in press). Effects of damming the Hadejia River in semiarid northern Nigeria 77

Dry-season flow

As a result of the dry-season releases from Tiga Dam, which are made to supply water to the large formal users upstream of Hadejia, an alteration of the Hadejia River regime from ephemeral to perennial occurred. Between 1963 and 1973 on average only 2% of the annual runoff passed Wudil during the dry season (November to May). After Tiga Dam was built (1979-1989) this percentage increased to 21%. The dry-season river flows at Hadejia increased from 4% of the annual flow (1964-1973) in the uncontrolled river to 16% (1977-1991) after the completion of Tiga Dam, and 32% (1993-1997) after the completion of the Challawa Dam (Fig. 3).

EFFECTS OF CHANGES IN FLOW REGIME

Weed development and flow reduction

Due to weed (mainly Typha domingenis) and silt blockages in the old Hadejia River (Fig. 1) virtually no water (<1%) from the Hadejia River has been contributing to the Yobe River since at least the early 1990s (Thompson, 1995; Goes & Zabudum, 1996, 1998). Typha started to grow in the Hadejia Barrage shortly after its completion in 1992. It was observed that the Typha grows in zones of the barrage with shallow (less

100

01-Apr 29-Apr 27-May 24-Jun 22-Jul 19-Aug 16-Sep 14-Oct 11-Nov 09-Dec 06-Jan 03-Feb 03-Mar 31-Mar Fig. 3 Mean daily hydrographs at Hadejia before and after the construction of dams. 78 B. J. M. Goes than 1-1.5 m deep) fluctuating water (Eng. Kazaure, personal communication). Typha reeds cannot survive in deeper water because their roots anchor them. Reeds did not invade the uncontrolled Jama'are River system because of the lack of water during the dry season and the high wet-season peak flows that flush the main channels clean. This means that the diversion of wet-season river flows to dry-season releases from the dams created favourable circumstances for the reeds to develop in the Hadejia River system.

Less reliable timing of floods

Despite a decrease in wet-season river flows at Hadejia, relatively large flood extents are currently, in the post-dams period, still possible in wet years (flood maps from IUCN-HNWCP). This is explained by the fact that the Hadejia River stopped contributing to the Yobe River due to the blockages. So almost all the water entering the HNW through the Hadejia River remains within the wetlands. Rice fanners in the HNW prefer the floods to arrive 1-4 weeks after the onset of the rains. This is needed for young plants to grow strong enough to withstand the flooding and grazing by fish (Hadejia et al., 1994). Contrary to what would be expected, after the completion of the dams the timing of the floods in the HNW became less predictable (Fig. 4). Even dry-season flooding occurred in the HNW. The dry-season floods of 1997-1998 destroyed 30-60% of the flood recession farms along the Marma Channel (Goes & Zabudum, 1998).

Water wastage during the dry season

Dry-season flows in the HNW during the second half of the dry season (January-May) are wasteful because there are not many demands during this period. Small-scale irrigation in the HNW is mainly practised during the first part of the dry season and not during the hot later part of the dry season. One of the main reasons for the fact that the discharge in the Hadejia River is maintained at higher than optimal levels during the second half of the dry season is to ensure that the pits of KCWS downstream of Tiga Dam are adequately filled (Diyam Consultants, 1996). The excessive dry-season releases are likely to increase further because the new (1999) water intake to KCWS downstream of Challawa Dam (Fig. 1), which was constructed under the World Bank assisted National Water Rehabilitation Project, has the same engineering mistake.

WATER MANAGEMENT IN THE BASIN

Most governmental and nongovernmental organizations that have an interest in the management of water resources in the basin are inward-looking. For example, there are two River Basin Development Authorities within the basin; one for upstream and one for downstream (IUCN-HNWCP, 1999). The potential water requirements in the Hadejia and Jama'are river systems are, respectively, 2.6 and 1.8 times larger than the mean available surface water resources. This can be largely attributed to proposed Effects of damming the Hadejia River in semiarid northern Nigeria 79

Hydrological Year

63 65 67 69 71 73 75 77 79 81 83 85 87 91 93 95 97 99 01-Apr

Tiga Darn completed Challawa Dam completed 01-May

31-May

01-Jul

31-Jul

30-Aug

30-Sep

30-Oct

29-Nov

1964-73: mean start date of long wet-season flood 28 June, standard deviation 12.8 days

1974-98: mean start date of long wet-season flood 4 July, standard deviation 20.1 days

Fig. 4 Periods during which the Hadejia discharge exceeded 19m3s"' (threshold value above which over bank flow starts, Goes, in press).

(expansions of) large- and small-scale irrigation projects and the proposed (IUCN-HNWCP, 1999). The installation of the proposed (Diyam Consultants, 1996) flow proportioning structure at Likori in the Hadejia-Nguru Wetlands is conceived as the best option for conveying water from the Hadejia River to the Yobe River because it can be controlled in such a way that there are no disastrous impacts on uses in the wetlands. Furthermore, the intake works for KCWS need to be improved by, for example, lowering the intake to the riverbed in combination with a weir.

CONCLUSIONS AND RECOMMENDATIONS

The average annual flow in the upstream part of the Hadejia River decreased by 33% after the completion of Tiga Dam (1979-1989). Furthermore, the peak flow was reduced and the river regime changed from ephemeral to perennial. The impact of these reductions is probably limited in the upstream section of the river due to the relatively high and more reliable rainfall in this area. The annual runoff and peak flows further downstream at Hadejia, just upstream of the HNW, did not decrease (1979-1997) as a result of the construction of the control structures. This is due to the relatively low river flow reductions upstream at low flows at Wudil and the fact that the formal large upstream water users are not (yet) working at full capacity. 80 B. J. M. Goes

The dry-season river flows at Hadejia increased from 4% of the annual flow in the uncontrolled river to 32% after the completion of the two dams. These dry-season flows: - create favourable circumstances for the development of weed and silt blockages in the HNW which prevent the Hadejia River from contributing to the Yobe River, - lead to dry-season floods which are disadvantageous for farmers and herders, and - waste water. Two improvements are recommended for the management of the control structures: (a) During the second half of the dry-season releases from the dams should be limited only to the three large formal water uses. That water can be saved for, for example, the restoration of a part of the historic contribution from the Hadejia to the Yobe River. (b) The timing of the floods should become more predictable so farmers can plan the planting of seeds accordingly. The potential water requirements in the basin exceed the water availability in an average year. Therefore, a proper basinwide demand control policy is essential. There is also an urgent need for an institutional organization that manages the water resources on a basinwide scale.

Acknowledgements Core funding (1995-1998) of the IUCN Hadejia-Nguru Wetlands Conservation Project was provided by the European Union. Bridging funds have been provided by: the Department for International Development, the Royal Society for the Protection of Birds and the Government. Michael van der Valk and an anonymous reviewer made numerous valuable comments on an earlier version of this paper. Michael also drew the map.

REFERENCES

Diyam Consultants (1996) Yobe Basin Water Resources Study Report on Stage la. Federal Environmental Protection Agency, Nigeria. Goes, B. J. M. (in press.) Effects of river regulation on aquatic macrophyte growth and floods in the Hadejia-Nguru Wetlands and flow in the Yobe River, northern Nigeria; implications for future water management. Paper accepted for Regulated Rivers Research and Management. Wiley, Chichester, UK. Goes, B. J. M. & Zabudum, A. N. (1996, 1998 and 1999) Hydrology of the Hadejia-Jama 'are-Yobe River Basin: 1992- 95, 1996-97, 1998-9. IUCN-Hadejia-Nguru Wetlands Conservation Project, Nigeria. Hadejia, I. A., Shu'aibu, M. & Polet, G. (1994) Farmer's Opinion on Timing of Artificial Flood Releases in the Hadejia- Nguru Wetlands. Hadejia-Nguru Wetlands Conservation Project, Nigeria. IUCN-HNWCP (1999) Water Management Options for the Hadejia-Jama 'are-Yobe River Basin. IUCN-Hadejia-Nguru Wetlands Conservation Project, Nigeria. Thompson, J. R. (1995) Hydrology, Water Management and Wetlands of the Hadejia Jama'are Basin, Northern Nigeria. PhD Thesis, University of London, London, UK