WADERS AND WATERFOWL IN THE FLOODPLAINS OF THE LOGONE, CAMEROUN, JANUARY, 1993 Jeroen CJ. van Wetten & Peter Spierenburg (eds.) Ecole de Faune (Cameroun) Foundation Working Group International Waterbird and Wetland Research (WIWO) WIWO-report nr. 67 1998
82
Embed
Waders and waterfowl in the floodplains of the Logone ...wiwo.org/wiworeport67cameroun1993.pdf · WADERS AND WATERFOWL IN THE FLOODPLAINS OF THE LOGONE, CAMEROUN, JANUARY, 1993 ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
WADERS AND WATERFOWL IN THE FLOODPLAINS OF THE LOGONE, CAMEROUN,
JANUARY, 1993
Jeroen CJ. van Wetten & Peter Spierenburg (eds.)
Ecole de Faune (Cameroun)
Foundation Working Group International Waterbird and Wetland Research (WIWO)
WIWO-report nr. 67 1998
This report can be ordered by paying Dfl. 20.-- plus 00. 15.-for administrationcost for each separate order from abroad, directly to postal giro account 2666009 or ABN-AMRO bankaccount 570216613 of Stichting WIWO, Fetha 23, 3633 CT Vreeland, Netherlands or by sending cash or an Eurocheque with 1he bankpassnurnber filled in at the back to that actress mentioning WIWO-report 67. The last two methods of payment are free of aillninistrationcost. Postage for surface mail is included in the price. Any information about WIWO-reports can also be obtained from that adress. For further instructions and the fulltis! of WIWO-reports see the last pages of this report.
WADERS AND WATERFOWL IN THE FLOODPLAINS OF THE LOGONE,
CAMEROUN, JANUARY, 1993
by Rigobert Azombo, Emmanuel Battakok, Arnold Bodekamp, Pim Edelaar,
Paul Scholten, Peter Spierenburg, Jeroen Cl van Wetten
edited by Jeroen C1. van Wetten & Peter Spierenburg
Ecole de Faune (Cameroun) P.O. Box 271, Garoua, Camroun
Foundation Working Group International Wader and Waterfowl Research (WIWO)
c/o Driebergseweg 16C 3708 JB l.eist
The Netherlands
WIWO-report Dr. 67 1998
Project participants Rigobert Azombo, Emmanuel Battakok, Arnold Bodekamp, Pim Edelaar, Paul Scholten, Peter Spierenburg, Jeroen CT van Wetten (team leader).
Financial support Stichting Fona, Vereniging tot Behoud van Natuurmonumenten, The Dutch Ministry of Agriculture, Nature Management and Fisheries, International Waterfowl Reserach Bureau (IWRB, now Wet land International)
Transport and Logistics Ecole de Faune, (Garoua, Cameroun), Centre d' Etude de I'Environnement et du Developpement au Cameroun, (Marou, Cameroun). IUCN/ CML, Waza Logone Project.
Acknowledgements. In the first place our gratitude goes to the agencies that have given financial support to the project. The director of the Ecole de Faune Mr. Jean N'gog Nje, Garoua Cameroun gave our work an immense input by providing a splendid Landrover, camping gear and some very luxurious fieldbeds, which gave us a firm backing during the nights. We also wish to thank Maliki Ali Bouba and Hanny Korthof of the CEDC (Centre d'Etude de l'Environnement et du Developpement au Cameroun) for the logistic support and the maintenance of our communication lines Without the skilful driving of J ames Che Pelote, the team members would never have reached and returned from the remote places we were now able to visit. The cheerful talks of James during the dark nights gave us all new spirits for the next morning.
Wetten, van lC]. & P Spierenburg (eds) 1998. Waders and waterfowl in the tloodplains of the Logone, Cameroun, January, 1993. WIWO-repport nr 67 Zeist.
CONTENTS
1 Introduction 1.1 Background of the project 1.2 Objectives
2 Area description 2.1 General 2.2 Field conditions in 1993
3 Methodology 3.1 Target species 3.2 Field methods 3.3 Data processing 3.4 Extrapolation
4 Results 4.1 Coverage and results per site 4.2 Total results for the study area 4.3 Comparison with previous studies 4.4 Discussion on results
5 Discussion of methods used 5.1 Evaluation of transect count method 5.2 Direct evaluation 5.3 Methodological comparison with previous studies 5.4 Conclusion on methodology
6 Conclusions and recommendations
7 References
APPENDICES
APPENDIX 1: APPENDIX 2: APPENDIX 3: APPENDIX 4:
Tables 4.1 to 4.16 Tables 5.1 to 5.4 Figure 3.1,3.2 and 3.3 Location of the study sites.
3 , ,)
5
7 7 7 9 12
15 15 18 19 20
24 24 27 28 29
30
31
APPENDIX 5: Calculation procedures for the King census method, the Two Belt method and the adapted Two Belt methods.
APPENDIX 6: List of observed species APPENDIX 7: Itinerary APPENDIX 8: List ofWIWO-reports
Waders and Walerblrds In the floodplams of the logone, Cameroun, 1993
1. INTRODUCTION
1.1. Background of the project
The Logone floodplains are part of the Lake Chad-basin, which is regarded as an area of great importance for wintering palearctic waders and other waterbirds (Perennou, 1991). However, (actual) ornithological data for the region are scarce. Before 1993, the most extensive recent surveys consisted of aerial counts between 1984 and 1987 (Perennou, 1991). These counts covered the whole Chad-basin but focussed mainly on anatidae. More recently Ruff and other waders were studied in parts of the floodplain area in Cameroun, mainly in the rice-fields near Lake Maga and Yagoua (OAG MUnster, 1991). Visits to the area by some team members in previous years already revealed that considerable numbers of waders and waterfowl were present. The Logone 1993 project represents the first detailed and extensive ground survey of the Logone floodplains north of Lake Maga and the Lake Chad area.
The survey can be considered a valuable contribution to the knowledge of winter distribution of waders and waterfowl in the Chad-basin, especially for Ruff and Black-tailed godwit and a number of Herons. The Chad-basin being a major wetland-system in the Sahel-zone, on a larger scale this survey contributes to the knowledge on wintering waders at inland localities in West-Africa. According to the Long Term Plan of WIWO (Engelmoer et al.., 1988) this can be regarded as a priority in data gathering on the wintering sites of migratory waders and waterbirds following the East Atlantic Flyway or the East Mediterranean Flyway.
Furthermore the Chad-basin is subject to major environmental threats, and the development of environmental management plans is urgently needed. In the central part of the Logone floodplain area north of Lake Maga initiatives are being taken in this direction, for example the Waza Logone Project (IUCN/Center of Environmental Science Leiden). The development of these plans need to be supported by reliable data on the ornithological significance of the area (Grimmett, 1987) while nature conservation value impact monitoring of the floodplain restoration attempts are useful.
1.2. Objectives
The objectives of the Logone 1993 project of WIWO and Ecole de Faune were:
Providing basic data on waterfowl and migratory wader and waterbirds of the Logone floodplains and the Lakc Chad area in North-Cameroun, in the context of the yearly IWRB (now Wetlands International) mid-winter counts of tropical and sub-Sahara wet land areas.
WIWO~report 67
Gathering relevant data on bird numbers and distribution for priority of conservation and protection measures in North-Cameroun.
Providing relevant data on the Zero-situation of wader and waterfowl numbers, before large scale floodplain restoration initiatives become implemented in the Waza Logone floodplain area.
In-field training of staff members of the Eeole the Faune, Garoua, into methods of bird counts and monitoring.
Initiating yearly counts of waders and waterbirds by staff and students of "Ecole de Faune", by organising yearly counts by students and staff of the "Eeole de Faunc", being an integrated part of the fieldwork in the wet/and course of this institute.
Forming an expertise-pool, especially at the "Ecole de Faune" to be used for the yearly IWRB (now Wet/and International) mid-winter counts in tropical and subSahara Africa in subsequent years.
This report concerns methods, results and conclusions ofthe first two objectives. It discusses the significance of the Logone floodplains and the Lake Chad area for wintering Palearctic waders and waterbirds, as well as Afro-tropical species.
2
Waders and Waterb!rds In the floodplalns of the logone. Cameroun.1993
2. AREA DESCRIPTION
2.1. General
The province 'Extreme Nord' of Cameroun is situated in the transition zone of the Sahel and Sudan climatic zones. Major rivers like the Benoue, Chari and Logone all have their catchment for a large part in the Sudan zone. Although the most important wetlands are located in the Sahel/Sudan transition zone, their hydrologkal characteristics are for a large part detelmined by the rains in the Sudan zone. Floods and peak discharges start in the JuneIJuly, with a maximum in AugusUSeptember/October. From mid October onwards rains cease to fall and the dry season extends up to the beginning of April. Mean rainfall figures vary from less then 500mm in the northern part (Lake Chad to Waza), 500-800 mm in the middle part (Waza-Maga) and 800-1000 mm in the southern part (Maga to Mayo KebilBenoue) (Drijver and Marchand, 1985).
The Logone floodplain and the adjacent Mayo Kebi floodplain cover an area up to 15.000 square kilometres. The maps of appendix 4 and 5 show the study areas. While the Mayo Kebi drains into the Benoue Basin, the Logone with its tributaries drains into Lake Chad. In the I 970'ties and early 1 980'ties the Extreme Nord has been severely touched by the Sahel droughts and reduced rainfall in the adjacent Sudan zone. Lake Chad shrank to one third of its original size and the Logone floodplains were deprived from any large scale inundation 's for subsequent years. In the second half of the 1980'ties and the last two years (1991, 1992) normal floods occurred. Although the area was severely degraded a quick recovery of vegetation, fauna populations and natural resource output could be observed soon after the return of the floods. The waterlevels in Lake Chad rose and in 1993 the actual size and level nearly matched the long term means (Burgis and Symoens, 1987; van Wetten and Spierenburg, 1993)
Interventions in the area mainly aimed at starting rice irrigation and reducing the dependency of the local popUlation on the unpredictable floods and rain patterns. For this, large irrigation schemes like SEMRY, I,II,III ,IV totalling some 30.000 hectares were constructed. An artificial lake at Maga blocks nowadays the natural inundation pattern of some 150.000 hectares Logone floodplains . In general one can say that less and shorter inundation's are the direct impact of this lake (Drijver and Marchancl 1985).
Drainage water from the irrigation schemes around the lake at Maga is since 1987 diverted by a small dam in the Logomatya into the depression of ZilimJZimado. This dam thus creates new and man made inundation's not only during the normal rain and flood season but also during the dry season. In total some 400.000 hectares are in this way provided with water and wetland conditions are greatly restored in this area (Haberland et van Oijen, 1991)
wrwO-report 67
A number of different ethnic groups have occupied the floodplains. Notably the KabaJayeGabri, Mesme and Moussey in the Sudan zone of Chad, and the Kotoko, Mousgoum in the Sahel zone of Cameroon and Chad, form the main groups which have settlements on the river banks. Nomadic pastoralists like the Fulbe or Shoa-Arabs are found in both the two zones.
Apart from floodplain agriculture, cattle raising and hunting, one of the main activities of people dwelling on this flood plain is fishery. Especially in those villages close to the Logone river, its side branches and its major low lying swamps, communities are found which depend for a great part of their livelihood on fishery, taking advantage of the yearly returning pulses of rain and river inundation's. Inundation's restore soil fertility, vegetation cover and floodplain productivity in general. Old river courses nowadays forming depressions and swamps are common in the floodplain. During river floods flood water flows firstly over the embankments to these depression and swamps. With even higher flood levels the surrounding floodplains are inundated. During the heights of the inundation the plains seem to be one great shallow lake with small settlements on isles in the middle of it. only when the flood water recedes features like marshes, small river branches and ponds become distinctive again (Drijver in prep., Haberland et van Oijen, 1991)
Fish fauna consists of two major groups of species. The first group survives the dry season in the downstream situated Lake Chad and migrates upstream shortly before the flooding period. The second group remains in the deeper river parts and in the deeper parts of the depressions and marshes. Both groups move into the low parts of the floodplain once the flood reaches the plains. Spawning and reproduction takes place for a number of species already in the river parts and marshes before the flood arrives, while others only start spawning and reproduction after the floodwater reach the depressions, marshes and plains. In general one can say that fish resources are abundant from the start of the flood inundation's onward and become scarcer towards the end of the dry season.
The dominant vegetation type of the flood plain is a steppe vegetation, dominated by perennial grass species. Shrubs, trees and annuals grasses and forbs are confined to the fringes of the flood plain basin, though the drought period has led to their proliferation towards the central parts of the flood plain. The lower plains and depressions are dominated by a floating vegetation of Echinochloa stagnina, important for cattle. Higher plains have a mixed species composition of tall-standing grasses like Hypharrhenia rufa, O,yza barthii, Echinochloa pyramidalis and Vetiveria nigritGlla. After retreat of the floods part of the area is burned by herdsmen in order to evoke a regeneration of fresh leaves from the tufts of the grasses. The perennial stream Logomatya and canals in the ZilimJZimado depression have well-vegetated, with reedbeds dominated by Vossia cllspidata (van der Zon, 1987; Drijver and Marchand 1985; Drijver in prep.)
The floodplain serves as an important foraging site for large mammals, especially Kob antelopes. Their presence is related to the existence of the Waza and Kalamaloue National parks, which incorporate part of the floodplain areas (van der Zon, 1987)
4
Waders and Waterblrds in the floodplalns of the Logone, Cameroun,1993
2.2. Field conditions in 1993
Although the rains in the Extreme North province were not so abundant as the previous year the discharges of the Logone and the Chari turned out to be excessive. As it turned out the Chari still transported large volumes of water north to Lake Chad and its water level seemed to be 2-3 meters above the normal level for mid January.
Due to only mediocre rainfall the southern part of the floodplains turned out to be dry already, with only scattered places with remaining water, while on the contrary in the north the lake borders provided extended zones of inundated areas and marshes river due to excessive river discharges. Appendix 5 comprises detailed maps of the study sites.
As irrigation had only just started with the planting and growing of rice in the nurseries, the large irrigation schemes of SEMRY II,III and IV were generally dry. Only at Maga some 450 hectares were under first irrigation. The expected high numbers of birds as seen during previous counts were not observed (Perennou, J 991 ; OAG Munster, 1991). Only the shores of Lake Maga and the adjacent ponds near the dyke contained interesting numbers of waterfowl and waders. As Lake Maga was filled up to its brinks in order to await the coming demand of water for irrigation, the marshy lake fringes were very narrow. This would probably account for the small numbers of both ducks and waders observed in this wetland.
The floodplains north ofMaga up to ZimadolHinale have recovered considerably during the last two rain and flood seasons (1991-1992 and 1992-1993). The late rains in 1992 prevented early burning of the grass vegetation in these floodplains. Compared with other years, the team members had the impression that in January 1993 only a small part of the grass vegetation was burned by fires, while in previous years almost 100% was burned down at this stage of the dry season. The high vegetation of grass (sometimes up to 2.5 meter) seemed utterly unsuitable for Ruffs and other waders occurring in dry habitat. In the rare plots where burning, herdsmen or cattle had already removed the high vegetation, substantial numbers of Ruffs were observed.
In the previous months (NovemberlDecember 1992) drainage water from the SEMRY IT Scheme was converted into the depressions of Zilim, Mandabe, Hinale and Zimado. These depressions still contained considerable amounts of water. The lush vegetation of Oryza barthii, Echinochloa spp", Hyparrhenia rufa, and Vetiveria nigritana covered most of the mudbanks and ponds that were observed in other years. Herons, ibises and storks were now, instead oflarge numbers ofPalaearctic ducks and waders, very abundant in these depressions. Especially the Black heron, Intermediate heron and Black-headed heron as well as the African Spoonbill were found to be numerous.
Before the construction of an earthen dam in the Logomatya in 1987, drainage water from SEMRY 1I flowed into the Logone. Nowadays this water is first diverted in to the depressions ofZilim, Manadabe, Hinale e.a. and partly reaches the El Beid watershed. Intensive Corn and Mouskuari (dry season Sorghum) cultivation on the banks of the El Beid (profiting from the steady flow of drainage water) created mud banks where all vegetation was removed. As it was observed at the moment of our visits these banks were important sites for a number of smaller
5
WIWQ-report 67
waders like the Little stint. Once cultivation has started again on the banks of the El Beid the river will gradually lose its importance for these waders as a suitable foraging area. The downstream part of the El Beid north ofFotokol is currently occupied by Nigerian forces and not open to visitors. The extensive mudflats and shallow waters in the estuary of the El Beid still remains an interesting area for an ornithological reconnaissance.
The water level of Lake Chad in 1991-1992 had reached a high level not experienced in the last 15 years. In January 1993 the borders of the lake did not exceed the line drawn on the maps from 1952, but people still regarded the size of the lake and the waterheights as exceptional compared with the last two decennia. Due to the immense inundations and the very lush vegetation bird counts proved to be a hard task. Only with an equipment which would allow for an extended stay and visit of the lake shores an adequate census could have been made. In spite of the logistic shortcomings of our team a large number ofbirds was counted on a few limited transects. The bird numbers observed indicate that the Lake Chad has again become one of the major wintering sites for certain Palaearctic waders and duck species. Ruffs, Blacktailed god wits, Garganey and Pintails as well as Little stints were observed on all transects. The team members had the impression that the low numbers of the above mentioned species in the southern part of the Logone plains might partly be caused by the immense zones of inundation on the borders of Lake Chad in the winter of 1992-1993 which proved to be more suitable for these species.
6
Waders and Waterblfds In the floodplalns of the Logone, Cameroun.1993
3. METHODOLOGY
3.1. Target species
Two groups of target species were considered, for which different types of data were gathered :
- Ruff and Black-tailed godwit were the main target species. Data collection concentrated on the distribution of these species over the study area, related to their feeding ecology. The idea was to compare the importance of dry floodplain (with seeds as major food) with that of moist areas like depressions and rice-fields (with evcrtebrates as food). In practice, the virtual absence of Ruffs and Black-tailed godwits at a number of sites where we had expected them (dry floodplain, depressions), necessitated a verification of a larger number of alternative sites. Due to the relatively low numbers encountered little attention could be given to additional data collection (for example sex ratio, presence of marked individuals, direct observations of feeding behaviour, roosts)
- The rest of the species of interest to the IWRB midwinter waterfowl census in West Africa formed the second group. The low numbers of Ruff and Black-tailed godwit often gave rise to a shift in priority to this "secondary" group of species. In order to arrive at an estimate as accurate as possible, these species were recorded in all counts. Sites with possible high concentrations were checked.
Besides the two target groups raptar species and some warbler species (for example Sedge warbler) were given attention too, as these could easily be integrated into certain counts (line transects) . Finally all identified species were recorded, for the water bird sites a list as complete as possible was pursued.
3.2. Field methods
The study area was the Logone floodplains and related water bodies at the Cameroonian side of the Logone, between Lake Maga in the south, Lake Chad in the north, the Logone in the east and the El Beid in the west. Evidently this area of approximately 10000 sq. km is too vast for a total count by ground survey. Furthermore a variety of water bird habitats is present. As hydrology and ecology of the area are not yet fully understood, it is hard to predict the condition of different potentially interesting sites at the time of visiting.
For these above mentioned reasons a classification of water bird sites was made in the field , supported by field observations, maps and Lalldsat and SPOT satellite images. Different ccnsus mcthods wcre used for different sites, thus responding to site characleristics like accessiblity, surveyability and extensiveness. Like bird numbers, these site characteristics proved to be quile unpredictable, in spite of prior visits to the area in February/March 1990,
7
WIWO-report 67
1991 and April and October 1992. The variety of methods used was beneficial to the training aspect of the expedition.
Total counts were undertaken at a minority of (well-defined) sites, notably Lake Maga, the Kalamaloue National Park, part ofWaza National Park and a number of smaller water bodies in the floodplain . In most cases sample counts were regarded as more appropriate. The sample counts can be classified in several categories:
A. River bank census A river bank census is actually a total count of a certain stretch of river or lake-side. The distance covered should be large enough for reasons of extrapolation. This method was used for the Logone river, the smaller rivers Logomatya and El Beid, some larger depressions and the lake-side of Lake Chad. With the exception of the latter two localities these are all sites which were easy to survey, so practically no birds will be missed. Double counts were ruled out by only taking account of birds that are passed by (or birds that flee opposite to the walking direction). Effort was made to cover a good percentage of the total distance, and to divide the counts over different stretches. For reasons of reduced surveyability the counts at Lake Chad and the large depressions only yield relative numbers of the birds present at the lake- and depression-sites.
B. Line transect with recording of sighting distance (Bibby et al .. 1992; Jarvinen and Vaisanen, 1975; van Lavieren) The line transect with recording of sighting dislance was used in extensive, homogeneous habitats, especially where accessibility and surveyability were low. The transects were done by foot, by car and by canoe. The method was lIsed in dry floodplain (by foot and by car), depressions (by foot) and swamps and waters with reedbeds (by canoe). As for each observation the distance perpendicular to the transect is recorded, an analysis of detectability and sighting distance is possible and bird densities can be estimated. A sufficient number of sightings is essential to this purpose. This method is not applicable at sites where the distribution of birds is related to a linear habitat type (like a lake-side or a river).
The transect counts were carried out in a way that data could be processed following different line transect methods. The perpendicular distance to the transect of each sighting was recorded (full-distance recording), in intervals of 10 metres at short range and 100 metres at long range. Special attention was given to birds exactly on the transect, as detectability at distance zero was supposed to be 100 % (Bibby et al. 1992). Distances were estimated by eye, with regular checking for in-between observer variability. Initial calibration was done at the irrigation schemes of Maga, where the standard dimensions of the "casiers" provided a good training environment for estimating distances in the field. Counting units were birds that displayed a clear relationship with the terrain, so herons and storks on the ground, low hunting raptors or ruffs that were flushed. The counts were carried out by car (road counts), by [oat and by canoe. The counts by foot and by car concerned the same area and some transects were counted more than once. Observation conditions appeared to be optimal in the morning. Around midday the activity of birds was low and in the evening there seemed to be some interference with movements to roosting sites. With the exception of the transects in the Tikele depression, all transects have been counted in the morning.
8
Waders and Waterblrds In the floodptalns of the Logone. Cameroun. 1993
C. Line transect with a fixed observation belt The line transect with a fixed observation width is more suited to sites which are well surveyable and contain large densities. Contrary to the previous method, this method presumes that within a certain observation width all birds are recorded. This method was used in the rice-fields ofMaga.
D. Counts of migration to and from roosts Counts of migration to and from roosts were executed at several sites, around sunrise and for a longer period around sunset. This method was to give an impression of the abundance of certain species (herons, Ruff, Black-tailed godwit) at inaccessible and extensive sites. For some species it even augmented the actual recorded number of a site, (Mandabe surroundings).
3.3. Data processing
In line with the specific character of each of the field methods used, different data processing methods have been used :
Total counts No data processing has been undertaken concerning the total counts. The absolute numbers recorded have been taken as the minimum estimate for the site, even if the site was not fully covered. Confidence intervals could not be calculated. The counts of migrations to and from roosts are treated as total counts, even though often not all migration directions could be covered.
River bank counts River bank counts have a sample character. The samples consisted of a number of transects of 1 up to 7 km of river bank, depression or lake side per site. For each transect, the mean number of birds per km was calculated. Analysis of the Lake Chad data suggest that this value has a lognormal distribution, judging from the linear relationship between the mean and the standard deviation after log-transformation (Slob, 1987). This relationship is shown in Figure 3.1. So a transformation of Ln(x+ 1) was applied in order to approach a normal distribution. The factor I is introduced to avoid zero values in the data. With the transformed data the mean and the 90 % confidence interval were calculated. The back-transformed mean is the maximum likelihood estimator for the population median (Slob, 1987).
In the following we will refer to the back-transformed mean as the flmedian" . Calculation:
0= EXP(Ln(D»
CL90 = EXP(Ln(CL90)
9
WIWO-report 67
Si=1 n Ln(Di+ I) Ln(D) = ---------------
N
--- 2 _ Si=ln(Ln(Di+ J) - Ln (D»
Ln(CL90) = D ± (SQRT --------------------------) I (SQRT(n) n-J
D ~ back-transfonned mean corresponding to the "median" referred to in the text CL90 = back-transfonned 90% confidence interval Di = density or number per km for transect i n = number oftransects
For extrapolation over the entire stretch of the river, lake-side or depression-side, the median and the back-transfonned lower 95% confidence limit (corresponding to the lower limit of the two-sided 90% confidence interval) were considered most useful. Data are presented as Cl) the actual observed number on the sampled stretch, (2) an estimate over the entire stretch based on the lower 95% confidence limit and (3) an estimate over the entire stretch based on the median value.
Line transect counts In the line transect counts, densities (number of birds per ha) per transect were calculated according to the methods that are described below. Eventually only the data obtained through the King census method were retained, as an evaluation of the different methods (Ch. 5) indicated that this method was the most suited in our study design. Similarly to the river bank counts median densities and back-transfonned 95% lower confidence limits were calculated and used for extrapolation over the entire area of thc site. As differences in the results of the roadcounts and counts by foot were relatively small (Ch. 5), in the calculations the individual transects of both methods were combined. The values for the transects counted more than once were averaged to yield a single value for the transect.
The calculation procedures follow the King census method, the Two Belt method and two adapted variants of the Two Belt method.
A King census method In the King ccnsus method the perpcndicular distance to the transect of each individual has to be estimatcd. Flocks are dealt with as being composed of a number of individuals. The density of individuals is calculatcd as the mean in an area with the transect length and a width of double the sighting distance, using thc formula:
10
D=N/2*L*S
D = density of individuals in the area N = total number recorded L = transect length
Waders and Waterbirds in the floodplams of the Logone, Cameroun, '993
S = mean sighting distance of individuals, perpendicular to the transect
If there are no recordi~~ D is s~osed to a.JJQfoach zero.
Bl Two Belt method In the Two Belt method a fixed transect width is used, resulting in two series of observations: inside and outside the inner observation belt. The quotient of the number of observations inside the inner belt and the total number recorded in both belts is used for estimating densities, given a certain detectability curve.
To determine the most appropriate detectability curve and band width, in Figure 3.1. the total number of sightings is plotted against perpendicular sighting distance. The car, foot and canoe transects are each treated separately. Figure 3.2. further refines these detection curves for four species- groups, each composed of species of supposed comparable detectability.
The detectability curves have a predominantly negative exponential shape. However, for some species-groups or count types there is a tendency that detectability declines slowly at short distances, and more steeply beyond a certain distance from the transect. In these cases a normal model might better fit the detection curve. The distribution tends most strongly to the negative exponential distribution in the road counts and less so in the canoe counts. Furthermore species with high detect ability (large size, often in flight or easy to flush) give rise to a more pronounced negative exponential distribution. Detectability diminishes from herons/storks to raptors to ruffs and finally other species (mainly ducks, geese and small waders). The negative exponential model is chosen as the basis for data analysis and forms the basis for calculated densities.
The 50%-limit is situated at a larger distance when species are more conspicuous, especially regarding their flight behaviour. The distance is the largest for raptors and ruffs and diminishes to herons/storks an other species. Good sighting conditions also result in larger distances: the largest for the counts by foot and diminishing to the road counts and canoe counts. The 50%limit corresponds to an optimal observation belt width (Bibby et al. 1992). In the analysis a belt width of 100 metres is used for the counts by foot, 50 metres for the road counts and 25 metres for the canoe counts. Data are derived from the full-distance recording data of the King census method.
Following the negative exponential distribution, the density of individuals is calculated as:
JI
WIWO-report 67
D = 5'" N '" L '" «-Ln(1-Nl/N»/w)
D = density of individuals N = total number of individuals, in both belts L = transect length NI = number of individuals within the inner belt w = width of the inner belt
If sightings are limited to a single belt, or the total number of sightings is zero, the density cannot be calculated.
B2 Two Belt method with recording of flocks In order to improve the method's resistance to outliers (large flocks at large distances), the Two Belt method tries to calculate density offlocks/sightings. With median flock size or number of individuals per sighting, the density of individuals can be calculated
B3 Two Belt method with confidence intervals based on variance in flock size In order to improve precision, variance in flock size has been used to calculate confidence intervals. Flock density data from all transect per site were averaged and multiplied with the median flock size and the back-transformed confidence limits. Median flock size and the backtransformed 90% confidence interval were obtained after Ln(x+ I) transformation. The underlying assumption is that variance in bird density is determined by variance in flock size rather than variance in flock density.
Line transect counts with a fixed observation belt In the second type of line transect counts, the density of birds is calculated by dividing the number per transect by the surface covered, that is the length of the transect x the belt width. Extrapolation of the field data was done in the same way as for the river bank counts.
Finally, estimates per site have been made, if necessary combining the results of different counting n-methods. In general the minimum estimate is based on the lower 95%-confidence limit. The most likely estimate is based on the median. However, if these estimates are not available (total counts), or turn out lower than the actual observed number, the latter is taken as the minimum estimate for the site.
3.4. Extrapolation
For extrapolation over the sites, satellite imagery and geographical maps were used. The extrapolation procedure passed through three stages:
Classification of sites In the field waterbird sites were classified in a number of categories. Counts were organised in a way that absolute or relative data for each category were obtained. The categories that were
12
Waders. and Waterbirds m the f100dplams of the logone, Cameroun,1993
distinguished are summarised in Table 3.1.
Mapping of land units Satellite images and topographic maps were used to identify land units corresponding to the categories of water bird sites that were distinguished. The land units were mapped visually, using colour and texture characteristics of the satellite images and landforms indicated on the topographic maps. Depending on the material available, only the larger scale sites could be retained. Use was made of:
- a print on a 1 :275 000 scale of a Landsat-MSS satellite image of 8 January 1987 - topographic maps 1:200 000 and 1: 1 500 000; 1950-51, revisions resp. 1975 and 1983.
The first was used for mapping the southern part of the study area, the latter for the El Beid stream, Lake Chad and the northern part of the Logone river. In general, the information that could be taken from the satellite image print and the maps proved to be sufficient to our purpose. The amount of detail was good, only for the ZilimlZimado depression additional information was taken from a print of a SPOT -satellite image 1: 100 000, in order to refine the mapping of the flooded depressions.
Furthermore, little difficulty was encountered in relating the information of the satellite image print to field observations, as the image was recorded in the month of January. The fact that the year of recording was drier than the study year, as shown by the extensive burnt patches in the southern part of the flood plain, had no severe consequences for the identification and mapping of land units. An exception to this are the areas of bare soil that were being inundated at the time of our study. These were also insufficiently covered in the field study and so the extension of the inundated area could not be assessed. For Lake Chad the information from the topographic maps was checked with a Quick-look of a Landsat-TM satellite image of23 January 1993. This facilitated a rough establishment of the current extension of the lake. Detailed mapping is only possible if the satellite image itself is available.
Determination of areas and stretch length The area or stretch length of each land unit mapped were measured manually, with the help of a distance measurer. The total extension of each category and the overall coverage per category is summarised in Table 3.1 on the next page. Chapter 4 gives more detailed information on the extension of specific sites and the coverage.
13
WIWO-fepor167
Table 3.1. Categories of waterbird sites; total area (in square kilometre) or stretch (in kilometre) of each habitat category and sample coverage.
CATEGORY TOTAL COVERAGE AREA/STRETCH
Floodplain -flooded depression with marshy vegetation 50 sq. km. 15 % -dry depression (dominated by Echinochloa stagnina 65 sq. km. 20 % "Bourgou") -dry flood plain (dominated by tall standing grasses 1100sq. Km 2 % Hypharroenia rufa, Oryza barlhii, Echinichloa pyramidalis, Vetiveria nigritana), with burnt patches -rice irrigation scheme -recently inundated muddy areas 130 sq. Km 10 %
max. 50 sq .km max. 4 %
Open water - lake 130 sq.km 75 % - perrenial river 300 km 5 % - temporary stream with muddy banks 200 km 12 % Marshy habitat - perrenial stream with marsh vegetation 190 km 12 % - lake side with marsh vegetation and muddy banks 150 km 6 %
14
Waders and Waterblrds In the floodpla!ns of the Logone, Cameroun, 1993
4. RESULTS
4.1 Coverage and results per site
The results per site are presented in Tables 4.1 to 4.7 in Appendix 1. Each table gives the results of the different field methods used at the site and total estimates. Tables 4.8. to 4.10. give the totals for all sites. Following Chapter 3.3., the estimates are presented as:
the actual observed bird numbers, sununated over samples/transects (obs.) the minimum estimate of the bird nwnbers, corresponding to an extrapolation of the back-transformed lower 95% confidence limit (min.) the median value estimate of the bird numbers, corresponding to an extrapolation of the back-transformed mean (med.).
In the text each site is described briefly and reference is given to methods used and coverage.
Kysmatary, Garoua Kysmatary is a small lake in the Benoue-valley near Garoua. It has extensive reedbeds, scrub vegetation and some open water. It covers an area of approximately 30 ha (12/1). The lake was counted from a viewpoint, from which 2 ha of open water could be seen. Data are treated as a total count, though the low surveyability of the site has certainly led to an underestimate. The results are presented in Table 4.1.
Lake Maga and SEMRY irrigation schemes Lake Maga covers an area of 13 000 ha. With a canoe a total count was made of 75 % of the lake. In addition the small pools and swamps north of the Maga barrage were covered completely. No floodplain areas were observed at the southern lake-side (16/1).
The Maga and Pouss irrigation schemes (SEMR Y II) cover a total area of 6200 ha. At the time of visiting irrigation had started in an area of approximately 450 ha. A verification of the nonirrigated areas showed that no waterbirds were present. Only some raptors were found (Marsh harrier, Tawny eagle). In the irrigated area 5 km oftransects with fixed belt width of 100 m (one-sided) were done by foot. The area thus covered was 50 ha, so 11 % of the 450 ha of area under actual irrigation. Beyond 100 m sight was obstructed by the irrigation earthworks, but inside the 100 m wide irrigation compartments, with the soil being bare, all birds were likely to be recorded. Observations of birds beyond 100 m are included as additional observations into the total number observed (Table 4.2.). Counts of evening migration to roosts revealed a roost of thousands of Yellow wagtails, but no roosts of Ruffs or Black-tailed godwits (1 5/1 and 16/1). At the irrigation schemes of Kousseri (SEMRY ITJ) and Zimado (SEMRY IV) irrigation had not yet started. The irrigation schemes ofGuirvidig (part ofSEMRY IT) and Yagoua (SEMRY I) were not visited. The results arc presented in Table 4.2.
15
WIWO·report 67
Depression Tikeh~ A large depression with an Echinochloa stagnina-vegetation is situated north of Tikele, just downstream of the SEMR Y irrigation schemes It covers an area of approximately 6500 ha. At the time of visiting it had mostly dried up. In this area transects counts were executed by car (8 km) and by foot (6 km), covering approximately 20 % of the area (17/1). Additional observations concern amongst others a roost of J 0 000 Ruffs. The results are presented in Table 4.3.
Dry flood plain, pools and depressions in the surroundings ofZina. At the time of visiting the biggest part (more than 90 %) of the flood plain area between Tikele up to Kousseri had already dried up, surface water being restricted to scattered pools and minor depressions. This dried-up floodplain covers an area of approximately 800 sq. km. In this area transect counts were executed, mainly focusing on the centre of the floodplain around Zina. Transect counts were carried out by foot for a total of2l.5 km. Transect counts by car amounted to 37 km. Additionally, one stretch of 20 km was counted at three days (17 to 2011, 5/2). This will correspond to approximately 3 % of the total area. Total counts were held of pools and depressions encountered along the road form Zina to Ivye. These sites of pools and depressions are mainly concentrated between Zina and Zilim. Probably these represent a major part of all pools and depressions in the area, though the smallest are likely to be overlooked (1711 to 2011)
The Logomatya and Mayo Vrick canaVstream stretches over 90 km. During a river bank count between Zina and Ngodeni 7 km were covered = approximately 8 % of the total stretch. The large roost of Night herons in Zina village is included in this count (1711). For the minimum and median estimates, the Logomatya data were combined with those from the canals in the ZilimlZimado depression. The results are presented in Table 4.4.
Depressions between ZiJim, Mandabe and Zimado Under influence of a dam in the Logomatya at Zilim, depressions in this floodplain section were still inundated at the time of visiting . The total extension of the inundations could not be assessed in detail (due to limited accessibility). The combination of field observations and satellite imagery allows a rough estimate for the major depressions.
Three depressions were visited. The Zilim depression was partly counted from two viewpoints (termite hills) . More detailed counts were carried out along 3 km of the depression-side ("river bank count"), covering approximately 40 % of the total stretch of the depression-side (18/1 and 19/1)
16
Waders and Waterbirds in the floodplClom s of the Logone, Cameroun,1993
The Mandabe depression was counted from a single viewpoint (19/1) In the Hinale depressions a count was done along 2 km of the depression-side covering approximately 10 % of the total stretch (4/2) .
Outside the actual floodplain area a swampy depression was encountered along the road from Waza to Kousseri near Zigague. A total count from a viewpoint was held, supplemented by a more intensive "river bank" count along 1.5 km the depression-side. The latter count corresponds to approximately 30 % of the total stretch of the depression side (2411 and 2511) . Downstream the depression continued in a small canal, a tributary of El Beid. Along this canal 1.5 km of river bank counts were held. Upstream of this depression inundated areas were still present at the north side ofWaza National Park. This area was not visited.
The data gathered proved to be insufficient for an extrapolation over the total depression area. They are therefore presented as total counts of the depressions visited. This Jack of representative counts was due to the complex hydrology of the area and its consequences for foraging opportunities for birds. At the north-east side of the area (Hinale, Ngawani and probably Khalkoussam) water levels were still high, creating only moderate conditions for fishing birds as well as waders. At the south side (Zilim) water levels were already dropping, creating good conditions for storks, ibises and herons. With water flowing in a north-west direction, at the west side (Mandabe, Kalgoulou, west side of the Hinale depression) fish accumulated, giving rise to large concentrations of mainly herons (and fishermen). At the outer north-west edges, water was still rising, (re-)inundating bare mud flats and thus creating excellent conditions for waders as well as ducks. The latter situation was observed at the small depression of Zigague. Probably, the same situation occurred north ofKhalkoussam as the satellite image of January 1987 shows green vegetation next to extensive bare areas. An elephant research team that visited the area during our study period, reported inundation's in this section, and observed a couple of thousand of White-faced tree-ducks. At Zimado this species was observed on evening migration from pools near the Logone (on the Chad side) in the direction of the northern part of this floodplain section. An accurate estimate of the area inundated in this northern part would only be possible on the basis of recent satellite imagery (first months 1993).
Canals connecting the depressions stretch over approximately 100 km (running partly through the centres of the depressions). Along canals at Mandabe and Hinale a total of 7 km of "river bank" counts were carried out, as well as a total count at the 2 ha Lake Mandabe (18/1, 1911, 20/1). Furthermore, transect counts were done by canoe in the Hinale depressions, along another 8 km (3/2) . So approximately IS % of the total canal stretch was counted.
In between the depressions an area of approximately 40 000 ha of dry floodplain was present. Here 10.2 km of transect counts with a fixed belt width of 200 m were executed. So approximately I % of the total area was covered (17/1, 19/1).
Finally, migration to and from roosts was recorded successfully at Mandab6 (19/1, 20/1). The results of all counts are presented in Table 4.5.
17
WIWO-report 67
Logone/Chari river Between Maga and Lake Chad the Logone and Chari river extend over 300 km. River bank counts were held at three localities along this stretch, covering a total of 15 km = 5 % of the total river stretch (! 611, 2011, 2111, Pouss, Zimado and Kalamaoue). The Logone river south of Lake Maga was not visited. The results are presented in Table 4.6.
El Beid stream The El Beid stream stretches over 204 km. River bank counts were held at two localities, covering a total of 26 km = approximately 12 % of the total stretch (2111 and 2211). The Serbewel stream, a branch of the Logone/Chari north of Ko us se ri, was not visited. It is unknown whether this stream presents interesting waterbird habitats. The results are presented in Table 4.6.
Kalamaoue National Park In the 6000 ha Kalamaoue National Park a number of ancient river bnlnches (oxbows) are found. A total count was held covering all surface waters in the park (2111). Results are presented in Table 4.6.
Waza National Park Of the 1700 sq. km Waza National Park only the forest section could be visited (approximately 200 sq. km). In this area a total count of all ponds (pools) (6) was held (2511) . Ponds in the floodplain section of the park were not visited (approximately 20 ponds). The results are presented in Table 4.6.
Lake Chad In 1992-1993 Lake Chad reached an extremely high level compared to that of the last two decades of drought. The observed borders of the lake waters were almost matching with the 1950-1951 (revision 1975) topographic maps (1200 000) demarcations. A rough estimate was made of the extension of the inundated southern shores of the lake. The total stretch ofinundated lake-side amounted to about 150 km, in a zone that was on the average 1.5 km wide. At two sites, 9.2 km of "river bank" counts were completed. Furthermore a transect count of3.5 km was carried out by canoe (23/1 and 2411) . In the extrapolations it was assumed that the inundation zone was covered partly by the river bank counts (strip of 500 m wide) and partly by the canoe counts (strip of 1 km wide). So, the coverage of the area was about 6 % for the river bank counts, and 1 % for the canoe counts. Results are presented in Table 4.7.
4.2. Total results for the study area
In table 4.8.,4.9. and 4.10. the total results of the survey are presented . For explanation and discussion on the extrapolation see chapter 5.
18
Waders and Waterblrds In the fl00dptains of the Logone, Cameroun,' 993
4.3. Comparison with previous studies
The most extensive study carried out in the Lake Chad basin is that ofRoux et Jarry between 1984 and 1987. This study concerned large scale aerial surveys in midwinter (January), covering amongst others our entire study area. Mid-winter data (mostly qualitative) on parts of our study area are available from a number of smaller studies (in: OAG Munster, 1991; Haberland en van Oijen, 1991; van der Zon, 1987). Elsewhere in the Chad basin, the studies of RSPB in the Hadedji-Nguru floodplains in northern Nigeria are worth mentioning (perennou, 1991).
I I
4.3.1. Qualitative comparison
Table 4.11. and 4.12. show the species that were not observed in our study, but of which records exist, resp. in the study area or the Chad Basin and Benoue Valley as a whole.
The tables suggest that the Chad Basin is of importance to a number of duck species and Recurvirostra avosetta, that were not observed in our study. The Benoue Valley harbours a number of characteristic species (plover and heron species), that have not been reported from the Chad Basin.
4.3.2. Quantitative comparison
Data from previous years, concerning the entire Chad Basin are presented in Table 4.13.
Aerial surveys 1984, 1986, 1987 A comparison with the results of the extensive aerial surveys between 1984 and 1987 (Roux et Jarry in: Perennou, 1991) leads to some interesting observations. The aerial surveys allow a better coverage oflarge concentrations and inaccessible sites. Ducks and geese are obviously better covered by the aerial counts. For these species, the ground counts cannot effectively be extrapolated, because of the occurrence of these species in large concentrations. Furthermore important sites for ducks and geese are mainly situated at some central parts of Lake Chad (Perennou, 1991). The relatively low numbers at Maga in 1993 are probably real (compare OAG Munster, 1991) and possibly due to the high water level of the lake.
Survey OAG Munster 1991 A comparison with the results from counts in the surroundings ofMaga in 1991 and some additional counts in parts of the study area (OAG Munster, 1991), shows important differences for a number of species. In 1991 in the rice irrigation schemes large numbers of herons, ibises and waders were found. However, for Ruff and White-faced tree-duck numbers were comparable. Over the whole study area, the total numbers observed in 1991 exceed those of 1993 for a number of species, for example most waders, including Ruff and Hadada ibis. The differences originate from the favourable conditions at the rice irrigation schemes in 1991, which where mostly dry in 1993. Extrapolation of the 1993 data generally results in figures of the same order of magnitude as 1991 (median estimate). Even though in 1991 important
19
wrwo.report 67
numbers of Ruffs and Black-tailed godwits were recorded outside the rice irrigation schemes (Haberland et van Oijen, 1991), it can be assumed that in 1991 waders were more concentrated at the casily accessible irrigation schemes, In 1993 waders were presumably dispersed over a larger area, and with more emphasis on the northern part of the study area,
4.4. Discussion on results
This study has produced the first data from ground counts in most of the study area, Only from the surroundings of Lake Maga and from Waza comparable data exist from previous years (OAG Munster, 1991). The aerial surveys carried out between 1984 and 1987 constitute the most important reference point, as they covered our entire study area, With the available data the discussion of results concerns:
- a comparison of data from 5 years from Lake Maga and its surroundings - an evaluation of the distribution of species over the areas in 4 years, by comparing our data to
the aerial survey data, - a discussion of the distribution of species over the areas throughout the seasons - an evaluation of the numbers encountered in view of present estimates of the entire West-
African bird populations,
4,4,1, Five years of data from Lake Maga
With the aerial surveys and the ground counts combined, now a 5 year set of data exists for Lake Maga and its surroundings, Table 4,14, summarises the data, including the 1993 data,
Table 4,14, shows that important fluctuations in numbers occur, The years 1987 and 1991 stand out for the abundance of respectively ducks, geese, Crowned crane and herons, Ruff, other waders, For 1987 this situation is probably due to favourable conditions at the fringes of the lake: after a period of drought these were inundated again, Unfortunately, the excessive numbers of ducks and geese were not confirmed in recent years, probably because of redistribution, though methodological causes can not be excluded (Ch, 5.3,), For 1991 the large area under irrigation was very favourable for Ruffs, other waders and herons, However, 1991 being a year of extreme drought, unsuitability of other sites could be also of importance, The 1993 data represent a situation in which irrigation had not yet started in its full extent and in which Lake Maga itself was filled up to its brim, with practically no marshy fringes, Furthermore, favourable conditions at other sites especially the ZilimlZimado depression and Lake Chad, will have resulted in a redistribution of the species mentioned, In this view the large numbers of Ruff and Spur-winged goose encountered in the Tikele depression, just north of the irrigation schemes are very significant.
4.4.2, Distribution of the waterbird populations over the study area in different years,
In general it can be expected that waterbirds have a more dispersed distribution in years with
20
Waders and Waterbirds in the tloodplalns ot the Logone, Cameroun,1993
favourable climatological conditions. Drier years necessitate concentration. Based on previous studies in drought years like 1984, 1986, 1991 (Roux et Jarry; OAG MOnster, 1991; Haberland en van Oijen, 1991) concentration sites in our study area are probably situated in the surroundings of Lake Maga, at the Logone river, the ZilimlZimado depression and the more central parts of Lake Chad. More southern refuge areas on Camerounian territory are likely to be situated in the Benoue valley (van der Zon, 1987). Our data present a climatologically more favourable year and show that an important proportion of some species of waterbirds in the area occur dispersed over the extensive f1oodplains. However, as data from previous favourable years exist only from 1987 and due to the limited coverage of most studies, it is not sure whether this distribution is typical for a favourable year. In any case, our study shows that even in a favourable year important concentrations can occur, for example in the ZilimlZimado depression and at the inundated fringes of Lake Chad.
Human activities have important impacts on some of the concentration tendencies, both in drier and more favourable years,. Lake Maga is an obvious large scale example: in dry years the Maga dyke has a severe negative impact on the suitability of the f100dplain for waterbirds, in some years eliminating any inundation. On the other hand, the perennial lake and the associated irrigation schemes, always provide important areas of alternative habitats, evoking concentration of birds. The case of the ZilimlZimado depression shows that even smaller scale waterworks can create very significant areas of suitable habitat to many species in a period when other sites have already dried up. Besides the waterworks man also influences the suitability of sites through numerous management practices. This year, at ZilimlZimado fisheries management created abundance offish in the downstream section of the depression, thus leading to a huge concentration of fish-eating bird species. Furthermore grazing and burning practices of herdsmen showed to be a key factor for the occurrence of Ruffs and other species in the dry f1oodplains.
4.4.3. Distribution of the waterbird populations over the study area in different seasons and throughout the season.
The concept of midwinter counts are a very valuable tool for regional and supra-regional comparisons. However, the highly dynamic and variable hydrological and antrophogenic conditions of the Sahelian wetlands severely hampers the interpretation of data. In November 1992 some 12,000 Black-tailed godwits were observed at a roosting site just north ofWaza. As this site had dried up by January 1993 and conditions all over the floodplain of the Logone did become unfavourable hardly any Black-tailed godwits were seen during this survey. During occasional visits in previous years large flocks of Ruffs were observed on burned f100dplain vegetation's. Without burning the 1 to 2 meter high grasses and herbs limit the accessibility for birds to the seed and ground insect resources. Delayed burning in 1993, due to late rains, was the reason that the largest part of the Logone f100dplain was still covered with high grasses. Only in the grazed areas and the rare burned areas a number of Ruffs were observed.
The extended inundation's along the borders of Lake Chad may have been an excellent foraging and roosting condition for waterfowl. Initial surveys with students of the Ecole de
21
WlWo-report 61
Faune in November 1992 revealed that at the peak of the inundation's of the Logone floodplains large numbers of waterfowl are found all over the southern parts. Although additional data are required for a detailed analyses it seems that the southern part of the Logone floodplain and the Mayo Kebi is of greater importance in the period OctoberlNovember. As soon as the Lake Chad inundation's start in the north a gradual shift of bird concentrations through the middle part of the Logone floodplain up to Lake Chad occurs.
4.4.4. Importance of the study area in the West-African context.
In order to evaluate the importance of the study area as a waterbird site, the results can be compared to estimates for the entire waterbird population of West Africa. These populations concern resident as well as wintering species. Table 4.15. shows for the relevant species the presumed percentages of the total West-African population. A general remark to be made is that little is known yet about central African bird populations (for example those in the Sudd area in Sudan and the Upper reaches of the Chari in Chad and the Central African Republic) and possible exchanges with the bird populations of the Chad basin.
The results of extrapolation show some exorbitantly high figures for a number of species. For some these even exceed current population estimates (Squacco heron, Crowned crane, Pratincole). This could be a reason to question the validity of some extrapolations. However in the case ofSquacco heron, Pratincole, but also Glossy ibis, Black-headed heron and Marabou stork, structural undercounting in previous studies is a likely cause as well (Perennou, 1991). This is due to either the often inconspicuous behaviour (former three species) or a preference for drier habitats (latter two species). For Crowned crane and predominantly east-african species like Hadada ibis, Sacred ibis, Yellow-billed egret, an additional explanation can be sought in the lack of knowledge of the size of central African bird populations (south-eastern Chad, Sudan, Central African Republic).
Table 4.15. shows the extreme importance of the ZilimJZimado depression for amongst others Squacco heron, Night heron, Great white egret, Little egret, Yellow-billed egret, Sacred ibis, White stork, Woodstork, African spoonbill, Crowned crane, White pelican. These figures show the unique character of this man-induced habitat, resulting from a relatively small hydrological intervention. The Logone floodplains evidently harbour more birds than was suspected up till recently. The extrapolated figures suggest that even in the extensive dry floodplain (January) heron species, storks, ibises, Ruff, Pratincole some duck and geese species occur in internationally significant numbers. For the Gr~y heron the floodplain as a whole is probably one of the few important sub-sahara wintering sites. The area's supposed importance for Egyptian goose and was not confirmed in our study (compare Perennou, 1991).
Furthermore the extrapolated values for Lake Chad stand out, especially for some Palearctic species like Glossy ibis, Purple heron, Squacco heron, Ruff, Black-tailed godwit, Black-winged ~ilt. This indicates that the area has regained its importance as a major wintering site for these Palearctic species, all important from a conservation viewpoint. The area's importance for ducks and geese could only be confirmed for Fulvous treeduek, due to methodological constraints as described in Chapter 5.3. However the observation of some large flocks of
22
Waders and Walerblrds In the floodplains 01 the logone, Cameroun, 1993
palearctic duck species (Garganey, Pintail) could indicate that the current restoration of the Lake's water levels has been beneficial to these species as well .
The importance of Lake Maga for White-faced treeduck, and Pratincole was confirmed (P6rennou, 1991). Other duck species and Ruff were not observed in exceptional numbers (Chapter 6.1.).
Most other sites visited are characterised by one or more species occurring in internationally significant numbers, for example Crowned crane at Waza, Gull-billed tern at the Logone river, Black-winged stilt at the El Beid river. For a number of species (for example Black heron, Black-headed heron, Marabou stork, Openbill stork) current popUlation estimates do not exist and these species are not included in Table 4.15 . Their inclusion would certainly stress once more the importance of most sites. With the exception of Black heron, in West-Africa the species mentioned have not yet been recorded in comparable numbers (p6rennou, 1991).
The main reproductive period of the afro-tropical species in the area is supposed to be the inundation period (August-November) . At least for herons and ibises, breeding was reported in this period by the local population. A very important heronry is expected to exist at the extensive river banks between Zimado and Kousseri. Records for reproduction in January are limited to Marabou stork with hundreds of nests at Zina village and (possibly) tens at Guirvidig; Spur-winged plover, of which a chick was observed at a waterhole in Waza and Grey-headed gull, of which 60 pairs in a post-breeding stage were aggregated at a colony in the inundated area near Hil6 Halifa at Lake Chad.
23
WIWO-report 67
5. DISCUSSION OF METHODS USED
The results allow for an evaluation of methods used on different levels:
- comparison of the results of (countinglcalculating) methods that were used simultaneously on the same sites. This comparison permits a choice between the transect count methods tested in this study;
- direct evaluation of the methods used, judging from field experiences; - comparison with results from previous studies. This comparison can give insight into
methodological aspects as well as the extent to which the results are compatible with the results from other years.
5.1. Evaluation of transect count methods
In the transect counts detailed data were gathered, allowing an evaluation of different line transect methods. These line transect methods vary in the way distance is recorded (full distance recording vs. two belt recording). Furthermore there is a choice in recording flocks or individual birds. Confidence intervals can be based either on variance of flock or individual density per transect or variance of flock size. In the latter case the variance of the density of flocks per transect is supposed to be negligible.
The countingldata processing methods evaluated are: A the King census method, with full distance recording of individuals B I the Two Belt method, with two belt distance recording of indi viduals B2 the Two Belt method, with two belt distance recording of flocks B3 the Two Belt method, with two belt distance recording of flocks and confidence
intervals based on variance in flock size.
Calculation procedures for each method are described in Chapter 3.
The main assumption underlying analysis of lin~ transect counts is that detectabil ity of birds diminishes with distance from the transect. Most detailed analysis is achieved through calculating detection curves from full distance recording data (Bibby et al. 1992). However, large samples are necessary for this. Even if species are grouped together in species-flocks (for example herons, storks, waders), realising a minimum of 100 observations per transect takes a large effort, given the relatively low bird densities of the flood plain habitat (scc Ch. 5.1.1.). The above mentioned methods are selected because they are supposed to producc higher efficiency in analysing field data. The evaluation concerns ifthese methods combine efficiency with prccision and accuracy (small confidence intervals, resistance to oUlliers). The road counts, canoe counts and counts by fool ar~ compared as well.
24
Waders and Walerbirds In the floodplalns of the Logone, Cameroun.1993
5.1.1. Efficiency
In the field the King method is the most time c.onsuming, as each observation distance has to be recorded as accurate as possible. The Two Belt method is more efficient in this respect.
Another aspect is the efficiency in reaching an acceptable sample size. Bibby et al. (1992) propose 40 recordings as a minimum for calculating precise densities. From Table 5.1. it can be derived that this overall sample size was not achieved for either the road, foot or canoe transects, when flock recordings were to be used for analysis. A sample size of 40 individual birds was more easily reached, but only for the selection of numerous species presented in the Table. Treating individual birds as independent data, whether in a flock or not, obviously allows for a smaller sample size. So methods A (King) and Two Belt-Bl depend less on extensive data than Two Belt-B2 and B3. Furthermore, the King method offers better opportunities than the Two Belt-B 1 method for calculating confidence intervals based on variance between sub-transects. For the Two Belt method data need to be available from both belts, so generally a larger number of recordings is needed. The King method is more efficient regarding the sample size.
Concerning the number of recordings per km I, the counts by foot can be considered more efficient regarding the number of recordings per km than the road counts. The same applies to the number of recordings per hour, though here differences are less marked. For some species the road counts are even somewhat more efficient regarding counting time: Crowned crane, Spur-winged plover. The canoe counts, which were carried out in a different area, produce an efficiency in time and per km comparable to the counts by foot. Table 5.1. shows some marked differences between the two methods. but these will relate partly to differences in species composition.
When examining the mean walking, riding, boating speed of the different methods, it appears that the counts by foot and the canoe counts cannot be carried out in much less time. So the time saving Two Belt method could only be of interest for the improvement of the road counts. From an efficiency point of view the best methods seem to be the counts by foot and the canoe counts, with full-distance recording following the King census method. The combination road count/Two Belt method could be interesting as well.
5.1.2. Accuracy
In comparing the King and Two Belt methods, there are two methodological considerations concerning accuracy of the methods. In the first place, from Figures 3.1. to 3.3. it shows that there is some effect of preferential, rounded distances around 50 and 100 metres in
I An underlying assumption to the evaluation of number or recordings per km or per hour is that hirds are evenly distributed over the study area. for the extensive dry floodplain habitat, and the species considered this appears a rather save assumption. Moreover. if this condition is severely violated, the utility of transect counts is low altogether. The number of transects counted is insufficient for a detailed analysis of the required sample size (for example by constructing fl diagram showing the cumulative number of birds for transect I to i)
25
WIWO-report 67
respectively road counts and counts by foot. This mainly affects the accuracy of the King method. Application in the field of distance estimates in two belts, according to the Two Belt method, will eliminate this problem.
Secondly, the Two Belt-BI method is more susceptible to outliers than the King method: large flocks at large distance have an important impact on the results , as these do not fit in with the exponential model used. Analysis showed that flock size and recording distance do not interfere for species that occur predominantly in small or medium size flocks (up to 20 birds). For a species like Ruff, occurring in very small to very large flocks, the Two Belt-B I method is unsuitable. In order to solve this problem in the Two Belt-B2 and B3 methods, densities are calculated using the densities of flocks/recordings and median flock size (after logtransformation).
A comparison of the results for a selection of species (Table 5.2.) shows that the median densities calculated with the King method are generally lower than those for the Two Belt methods. Especially the Two Belt B3 method produces relatively high absolute values. Outliers change this general pattern significantly. For species that occur in very small as well as large flocks (for example Ruff, Spur-winged goose, Pratincole here included under waders), the Two Belt methods produce a figure that is lower than King or in the same order of magnitude. For the canoe counts the difference in order of magnitude between King and Two Belt is very pronounced. Possibly this is due to the lower applicability of the exponential model in this case.
When comparing the values of the counts by foot with the road counts it appears that for both the King method and the Two Belt the correspondence is reasonable, if the transects are long enough. Two Belt B3 gives a somewhat better fit than King.
5.1.3 . Precision
Table 5.3. gives two-sided 90 % confidence intervals for the Zina data of roadcounts and counts by foot. It shows that the Two Belt B I and B2 methods have the lowest precision, judging from the width ofthe 90 % confidence interval. The width of the confidence interval is larger for King than for Two Belt B3. The confidence interval for Two Belt B3 is based on variance in flock size, whereas for King it is on variance between densities per transect. The Two Belt B3 method thus has the methodological disadvantage of not incorporating the variance in flock density per transect.
5.1.4. Conclusion on methodology of transcct counts
Data are obtained in the field the most eftkientJy (time, distance covered) by counts by foot and canoe counts, with full-distance recording following the King census method. The extra effort nceded for full-distance recording compared to two-belt recording does not lead to a severe loss of time in the field. A major advantage of the King census method is that it generally allows for smaller sample sizes, and so allows for a division in more sub-transects.
2(,
Waders and Waterb.,ds In the ftoodplains of the Logone. Cameroun. 1993
However, the Two Belt method is interesting for use in the field, in view of improving the efficiency of the road counts.
The choice of a data processing method has CClnsiderable consequences for the outcome. However, it is hard to tell which of the methods produces the most accurate results. The applicability of the Two Belt method is hampered by the effect of outliers, but the B3 adaptation reduces this. This method also gives the best fit between the road counts and the counts by foot. It's narrow confidence intervals seem advantageous too, but these result from not taking into account variance in flock density per transect. The Two Belt B3 method seems to be a reasonable option for the processing oft~e transect data, though the lower values resulting from the King method could be more on the safe side. In this study this last reason takes us to adopt the King method for thc calculation of minimum bird densities in the area.
5.2. Direct evaluation
In general the stratified census method has proved successful, both for carrying out in the field as for obtaining realistic results. Extrapolation from the line transect and the riverbank counts produced satisfactory figures for most species, both for the cxtrapolated medians and the lower confidence limits. However, the upper confidence limits appeared to give a quite unrealistic picture.
A major challenge in a stratified census is to balance the use of sample and total counts. Problems occurred at sites that were too large or had too low surveyability for a total count, but on the other hand were too small or too heterogeneous for sample counts. The ZilimJZimado depression is an example of a very heterogeneous site. Extrapolation from the sample counts was difficult and so the final estimate turned out too low. At a number of sites the total sample size amounted to more than ten percent of the total stretch or area. In this cases extrapolation often produced figures that hardly exceeded the actual observed number, or were even below it. So a coverage of 5 to 10 % seems to be the most effective for sample counts. Each stretch or transect counted must then have a minimum size, determined by the presence in each sample of the a number of characteristic species . For the rich fringes of Lake Chad, the minimum sample size appeared to be around 1 km, for the road counts it amounted to about 10 kilometres. So minimum sample size, a sufficient number of samples (depending on the heterogeneity of the distribution) and the area of the site determine whether a sample count is effective or if a total count should be pursued. As total counts are quite time-consuming, compromises are inevitable for sites like the ZilimJZimado depression, and results may never cover all birds presents.
Observation of migration to and from roosts is a possibility for a corn:etion of estimates from sites with low surveyability. But though migration counts were tried at most places (Lake Chad, Maga, Zigaguc, Mandabe, Zimado, Zina), only in the latter two cases interesting numbers wcre observed. Observation of roost migration of Ruffs and mack-tailed godwits failcd altogether. Finding a good observation spot will take more time than available in a stratified census.
27
WIWO-report 67
The use of satellite imagery proved generally satisfactory. The combination of field observations and the Landsat January 1987 image facilitated estimates of the areas of sites with a precision of up to 100 ha for the smallest sites. Major problems occur at sites with large year to year fluctuations in hydrology, especially the fringes of hydrological units like the ZilimlZimado depression and Lake Chad. These sites can be dry in one year and provide important areas of suitable habitats in the other. Possible solutions to this problem are discussed in Chapter 7.
A more fundamental way to use satellite imagery is to investigate correlation's between bird abundance and reflectance (compare Avery and Haines-Young, 1990). For the moment, hydrology and ecology of African wetlands and their interaction with the presence of birds, are not yet sufficiently understood for analysis along this line. For midwinter censuses best opportunities may exist for species like ducks and geese that feed mainly on plants. Correlation's between reflectance and plant-biomass are certainly possible. For herons and waders correlation's could be more complicated, as their distribution is determined to respectively abundance of fish or evertebrates and the occurrence of (mostly small scale) muddy fringes and shallow waters at lakes, depression etc .. These are elements that are less likely to be correlated to detectable features of available satellite images available. For Ruff possibilities may exist for a correlation with burned vegetation, though the relationship with the timing of the bushfires and the presence of other food resources is not yet sufficiently understood. Finally opportunities may exist for a wider range of species in other seasons, in our study area for example in October or November, when the floodplains are inundated
5.3. Methodological comparison with previous studies
5.3.1. Qualitative comparison
A qualitative comparison with previous studies mentioned in Chapter 4 could show if certain species have been missed in the current study, due to methodological constraints.
Tables 5.4. and 5.5. reveal that the qualitative coverage of our study was generally good. Of the species with regular occurrence in the area, only some Palearctic duck species have been missed. For further discussion see next paragraph. Furthermore, rallidae species offer severe methodological problems, and probably have not been covered adequately by any study yet. This persistent knowledge gap requires attention through the development of specific census methods (possibly tape recording/playback methods, netting)
53.2. Quantitative comparison
Aerial surveys 1984, 1986, 1987 A comparison with the results of the extensive aerial surveys between 1984 and 1987 (Roux et Jarry) leads to some interesting observations. The aerial surveys allow a better coverage oflarge concentrations and inaccessible sites. Ducks and geese are obviously better covered by the aerial counts. For these species, the ground counts cannot effectively be extrapolated, because of the occurrence of these species in large concentrations (clumped distribution)
28
Waders and Wat~rblrds in the floodplains of the Logone, Cameroun,1993
Furthennore, absolute numbers of herons, storks, ibises and some waders and gulls (Ruff, Black-tailled godwit, Black-winged stilt, Grey-headed gull) recorded by aerial survey are higher than as observed in ground sample counts. After extrapolation (median value) figures are of the same order of magnitude. Because of the smaller area concerned in the ground counts in 1993, it can be stated that aerial survey is less effective than ground survey for these species. For the waders mentioned the advantage is less pronounced because these species sometimes occur in large concentrations.
For a number of species ground counts appear superior even regarding absolute numbers: for example small waders, Pratincole (soaring behaviour), White stork (soaring behaviour), Night heron (day-time roosts), Black-headed heron (dispersed, mainly in drier habitats).
Survey OAG Munster 1991 The OAG Munster survey was characterised by a prolonged stay at a single site (Lake Maga and surroundings). The results from other sites (Waza, Lake Chad) have a more qualitative character. The Maga data are summarised in Table 4.15. The prolonged stay at Maga has resulted in a very good coverage, and the results can be regarded as the maximum number present at the time. In January 1991, under severe drought conditions, the Maga area was probably a concentration area for most species. The total observed numbers for a number of species (waders, including Ruff, herons, Hadada ibis) are thus higher than those observed in our studies. In sample counts the actual observed number is generally low. In January 1993 situation this was accentuated by the more dispersed distribution of the birds, than that of 1991.
The 1991 example shows that a stratified census requires visits to a large number of potential sites, as overlooking of one concentration easily leads to severe underestimates (the actual observed number in sample counts is generally low). Furthennore a longer effort on a site evidently results in higher numbers. Possibly local movements could be an explanation for this, making a comparison between stratified census data and multiple day counts on one site hazardous.
5.4 Conclusions on methodology
A stratified census, focussed on sampling, bears the risk of overlooking important concentrations, or being unable to extrapolate from a limited number of large groups observed. This risk is most severe for species with a clumped distribution, especially geese and ducks and to a lesser extent some waders like Ruff and Black-tailed godwit. For the latter species the effect is more or less compensated by the fact that ground surveys record these species more effectively than aerial surveys, even when dispersed. The stratified census is the most effective for species with a dispersed distribution. It allows covcrage of a wide species range, though not including rallidae. With a comparably lower effort larger numbers are recorded than by aerial survey. Table 5.4. summarises the applicability of different methods for, different species-groups in the context of water bird surveys in Sahelian wetlands.
29
WIWO-report 67
6. CONCLUSIONS AND RECOMMENDATIONS
The present study has established the importance of the Logone floodplain as a major site for wintering and resident waders and waterbirds in the Chad-basin. Estimates as well as the actual observed numbers of birds exceed those of previous studies. Furthermore it appeared that Lake Chad is recovering as a major wintering site for palearctic waders and waterbirds. However both for the Logone floodplain as for Lake Chad, these one year data give not yet reason for a definite review of the ornithological importance of these areas.
The relatively high numbers recorded in this study compared to previous ones, probably relate to a large extent to the methodology used. A stratified census, based partly on sample counts and extrapolation with the help of satellite imagery, appears a promising way to dcal with wader and waterbird populations dispersed over extensive wetland areas. Compared to aerial surveys and single site total counts, strong points of a stratified census are extensive coverage and good detail. The method can be considered as a valuable contribution to the development of methodologies for inventory and monitoring of (wintering) waterbird populations in tropical wetlands.
The study confirms that the interpretation of results from any study at inland wetland localities in West-Africa is often problematic due to the annual variation in suitability of different habitats and in the distribution of waterbirds. Even relatively small changes in hydrological or ecological conditions can have drastic effects (Smit en Piersma, 1989). In our study this was shown for example through the virtual absence of Ruffs from dry floodplain that had not yet been burned by herdsmen, or the relatively unfavourable conditions at Lake Maga and the irrigation schemes, as irrigation had not yet started at its full extent. Evidently the combined effect of climatological and human factors influence the suitability of the complex of wetland sites in the Chad basin in a rather unpredictable way. Smaller scale interventions can have unexpected outcomes as well, as is shown for the case of the ZilimlZimado depression, being essentially managed for fishery and cattle herding purposes.
The low predictability of the abundance ofa species at a certain site as well as the distribution pattern over the area, has important consequences for the design of monitoring and inventory studies. Field methods can be expected to respond to different distribution patterns in different ways. An aerial survey produces a better coverage of species occurring in large compact concentrations, like anatidae; our stratified census responds best to dispersed distributions. If it is assumed that birds in the Chad-basin display more concentrated distributions in drought years, the implications of the choice of field methods are evident. In this light improvement of the coverage of large concentrations in a stratified census is a point of attention in further development of methodologies.
In the field the stratified census as we applied it, can be improved by more attention to thc balance betwecn total and samplc counts, the sample size and especially the spreading of the samples over the different sites. As the fieldwork gains in efficiency and spreading over the area, a better coverage oC (often unexpected) concentrations of birds will be enhanced. Of course the limitations of further upgrading of fieldwork lie in mobility and time constraints.
30
Waders and Waterblfds in the floodplalns of the Logone Cameroun.1993
The use of satellite imagery in monitoring and inventory studies has an intrinsic large potential for dealing with the problem of annual fluctuations in field conditions. Major limitations however lie in the availability of good quality images for the intended period, the term of delivery and the costs involved. Field conditions in West-African inland wetlands relevant for the abundance of birds, generally change within the term of weeks. For optimal use this necessitates the use of satellite images corresponding as close as possible to the study period. This implicates that most analysis has to been done after the field work has been carried out, with the risk of overlooking important sites and of not been able to purchase appropriate images. In an inventory study, the effectiveness of the use of satellite images can be enhanced by:
- using an image of the same period from a previous year for orientation and verification in the field,
- make positioning in the field as precise as possible, for example by using GPS (Global Positioning System)
- initial aerial survey to determine and identify the major characteristics of the area and hydrological situation.
For a monitoring study can be added: - setting up a database of satellite images of previous years, so the situation of the present year can be roughly assessed by correspondence with previous years with comparable hydrological conditions.
The costs involved in the use of satellite imagery for wader and waterfowl studies impl ies elaborate co-operation with other research institutions and intervening agencies.
Finally an important element to be developed in wader and waterfowl studies in our study area is to use local informants for keeping up with recent developments in the area concerning land use practices and its consequences for opportunities for waders and waterfowl. Besides the local informants it was found that a detailed knowledge of the team members about the geographical, infrastructural and hydrological situation, is of great importance.
31
WIWO-repOrl 67
7. REFERENCES Avery, M.I. and R.H. Haines-Young (1990). Population estimates derived tTom remotely-sensed imagery for
Calidris a/pina in the Flow Country of Caithness and Sutherland. Nature 344, 860-862. Bibby, c.J., N.D. Burgess, D.A. Hill (1992). Bird Census Techniques. Academic Press, London. 257 pp. Burgis, M.J. and J.J Symoens (eds.) (1987). AtTican wet lands and shallow water bodies. Directory. ORSTOM,
Paris. p. 233-279. Drijver C.A. (in prep.). Management of AtTican wet lands. CML, Leiden. Drijver C.A. and M. Marchand (1985). Taming the Floods. CML, Leiden. 87 pp. Haberland, B. et J. van Oijen (1991). Une dique en terre. Effets socio-ecologiques de la dique de Zilim. These
de doctorat, Centre des Etudes de l'Environnement, Ul1iversite a Leyde. 181 pp. Engelmoer, M. (ed) et al.. (1988). WIWO Long Term Plan 1989-1993. Zeist. 41 pp. Grimmett, R. (1987). Review of the problems affecting palearctic migratory birds in AtTica: the findings and
recommendations resulting tTom the ICBP migratory birds questionnaire. ICBP, Cambridge. 240 pp. Jiirvinen, O. and R.A. Vaisiinen (1975). Estimating relative densities of breeding birds by the line transect
method. Oikos 26, 316-322. Lavieren, B. van (1982). Wildlife management in the Tl"Opics. A guidebook to the warden. Part I, Introduction,
Taking fieldnotes and Wildlife census methods. School of environmental conservation mangement, ClAW I, Bogor, Indonesia.
Louette, M. (1981), The birds of Cameroun. An annotated check-list. Verhandelingen van de Koninklijke Academie van Wetenschappen. Bruxelles. 296 pp.
Moreau, R.E. (1972). The Paleartic-AtTican Bird Migration Systems,. London Academic Press. 384 pp. OAG MOnster (1991). Report of the ornithological expedition to northern Cameroun. January/February 1991.
OAG MOnster. 40 pp. Perennou, c. (1991). Les nlcensements internationaux d'oiseaux d'eau en AtTique tropicale. IWRB, Slimbridge.
137 pp. Sokal, R.R. and F.J. Rohlf (1981). Biometry: the principles and practice of statistics in biological research. 2nd
ed. San Francisco: Freeman, 1981. p. 147-152,419-421. Slob, W. (1987). Strategies in applying statistics in ecological research. Vrije Universiteit Amsterdam, 1987. 112
pp. Smit C. J. and T. Piersma (1989). Numbers, midwinter distribution and migration of wader populations using
the East Atlantic flyway. In: Boyd, H. and J.Y. Pirot (1989); Flyways and reserve networks for waterbirds. IWRB Spec. Pub!. no. 9, Slimbridge. p. 24-63.
Wetten, J.c.J. and P.J. Spierenburg (1993). The Camerounian-Dutch midwinter wader and waterfowl expedition to the floodplains of the Logone. January 1993. Short report._WIWO. 14 pp.
Zon, T. van der (1987). Seminaire sur les zones humide,. Ecole de Faune Garoua, Cameroun. 275 pp.
32
APPENDICES
APPENDIX 1: Tables 4.1 to 4.16 APPENDIX 2: Tables 5.1 to 5.4 APPENDIX 3: Figure 3.1, 3.2 and 3.3 APPENDIX 4: Location of the study sites APPENDIX 5: Detailed maps of three study sites APPENDIX 6: List of observed species APPENDIX 7: Itinerary APPENDIX 8: WIWO-reports published
additionaL observations, not included in the standardized counts. transect countl by canoe transect count. with fixed beL t width transect counts by foot counts of migrations to and from roosting sites river bank coUllts totaL counts transect counts by car
The nl.d>er preceding the count type indicates the number of transects or river bank counts. The total transect Length is given in brackets after the count type.
ii
Appendici
Table 4.2.: Bird ntm>ers at Maga (lake Haga and SEHRY irrigation schemes).
lake Dike SEMRY Total area 13000 ha 6200 ha 19200 ha lC TC 5BC(5km) obs. obs. obs. min. med. obs. m;n~ med.
obs. ntm>er observed Aa = additional observations, not included in the standardi zed counts. min. = minillUll estimate CC transect COlllts by canoe
(95 " confidence limit) BC transect COlllts with fixed bel t width med. = median value estimate FC transect COlllts by foot
MC counts of mi grat ions to and from roosting sites RB river bank counts le total counts RC transect counts by car
The nunber preceding the count type indicates the "unber of transects or river bank counts. The total transect length is given in bracket. after the count type.
lii
WIWO-report 67
Table 4.3. : Bird numbers in the depression of Tikele.
dry floodplain Total area 6500 ha AO 2FC(6km) lRC(8km) obs. obs. obs. min. med. obs. min. med.
AO = additional observations, not included in the standardized counts. CC = transect counts by canoe
(95" confidence limit) meet. = median value estimate
BC = transect counts with fixed belt width FC = transect counts by foot HC = counts of migrations to and from roosting sites RS = river bank counts TC = tot a l counts RC = transect counts by car
The nunber preceding the count type indicates the nunber of transects or river bank counts. The total transect length Is given in brackets after the count type.
IV
Appendlcl
Table 4.4. : Bird numbers in floodplain habitats in the surroundings of Zina.
pools Logomatya dry floodplain 10 ha 90 km 80000 ha TC 4RB(7km) 3FC(21.Skm) 3RC(37km) AO obs. obs. min. med. obs . obs . min. med . obs. obs. min . med.
obs. = nlEber observed AD = additional observations, not included in the standardized counts . min. = m;n;rrun estimate CC = transect counts by canoe
(95 X confidence limit) BC = transect counts with fixed belt width med. = median value estimate Fe = transect counts by foot
MC = counts of migrations to and from roosting sites RB = river bank counts Te = tote l count& RC = transect counts by car
The nlEber precedi ng the count type i ndi cates the nlEber of transects or river bank counts . The total transect length i9 given in brackets after the co...,t type.
v
WIWO-report 67
Table 4.5: Bird numbers in the Zilim/ zimado depression
dry floodplain canals (south) canals (north) 40000 ha 40 km 60 km 2BC(8.2km) 4RB(7.5km) 3CC(8km) obs. min. med. obs. min. med. obs. min. med.
abs. nurrber observed AD = additional observations, not included in the standardized counu. min. = mininun estimate CC transect counts by canoe
(95 X confidence limit) BC transect counts with fixed belt width med. = median value estimate FC transect counts by foot
MC counts of mi gr8t ions to and from roosting sites RB river bank counts Te total counts RC transect counts by car
The nurrber preceding the count type i ndi cates the nl.flber of transects Or river bank counts. The tota l trensect length is given in brackets after the count type.
VI
Appendlcl
Table 4.5. (continued)
Mubl i Zit im L.Man.Mand. Zigague direction Total area 200ha 500ha 2ha 300ha 200ha south north 40000ha RB lC TC TC TC MC MC Aa obs. obs. obs. obs. obs . obs. obs. obs. ebs. min. med.
obs. ~ nuOOer observed AD = additional Observations, not included in the standardized counts. min. ::I mininun estimate CC = transect counts by canoe
(95 X confidence I imitl BC = transect counts with fixed bel t width med. ~ median value estimate FC = transect co\~nts by foot
MC = counts of migrations to and from roosting sites RB 1111 river bank c:ounts rc = total counU RC = transect counts by car
The nunber preceding the count type indicates the nuOOer of transec:ts or river bank counts. Th .. total transect length is given in brackets after the count type.
vii
WIWO-report 67
Table 4.6. Bird numbers at the Logone river, the El Beid river and Waza and Kalamaloue National Parks.
Pelecanus onocrotslus Phalscrocorsx africanus Nycticorax nycticorax Ardeola ralloides Bubulcus i bi s Egretta ardes i aca Egretta alba Egretta intermedia Egretta garzetta Ardea cinerea Ardea melanocephala Anastoous lsmell igerus leptopti Los cruneniferus Ibis ibis Threskiornis aethiopica Bostrychi a hagedash Plegadis falcinellus Dendrocygna vi duata Plectropterus galTbensis Sarkidiornis melanota C; reus macraurus Ci rcus aerugi nasus Bslearics psvonina Actophi lornis sfricana Vanellus spinosus Vanellus senegallus Charadrius hiaticula Charadrius dubius Charadrius pecuarius Charadrius alexandrinus limosa l imosa Tringa nebularia Tringa stagnati l is Tringa glareola Tringa ochropus Tringa hypoleucos Tringa erythropus Gallinago gall inago Cal idris minuta Ph i lomachus pugnax Himantopus himantopus Pluvianus aegyptius Glsreola prstincola larus ridibundus larus ci rrhocephalus larus fuscus Sterna nilotica Chl idonias leucoptera totsl
obs. n~r observed min. = mininun est;mate
logone 300 km 5RB (12km) obs. min. med.
10 368
o 6
19 4 6 2
50 13 9
20 76 16 5 o 2
83 o
17 o 2 1 o 8 9 o 5 o o 3 4 1 o o 4 o o 1
130 3
21 1300
1 74 22 27 80
2392
o 0 865 3462
o 0 20 100 26 231 o 52 o 69 o 42
185 558 42 187 56 193 76 309 o 15 o 58 o 61 o 0 o 16
10 678 o 0 o 131 o 0 o 28 o 15 o 0 o 84 o 87 o 0 o 79 o 0 o 0 o 41
14 82 o 0 o 15 o 0 o n o 0 o 0 o 16
15 806 o 60
68 292 8 796 o 16
400 1174 o 193
114 460 89 727
1980 11214
El Beid 204 km 6RB (26km) ob.. min. med.
o o
50 7
26 o 2 o
16 24 7
45 o o 6 1
16 o
26 o 2 1 o o
79 o 6
208 38 2 o
38 4
38 53 47
8 1
844 5
178 o o o o o o o
lna
o 0 o 0 o 162 o 26 o 151 o 0 o 21 o 0
65 151 150 211
11 55 o 141 o 0 o 0 o 42 o 10 o 103 o 0 o 3 o 135 o 0 o 18 o 5 o 0
P. africanus N. nycticorax A. ralloides B. ibis E. alba E. intermedia E. garzetta A. cinerea A. melanocephala A. purpurea C. ciconia C. epi scopus A. lamell igerus l. cruneni ferus I. ibis T. aethiopica B. hagedash P. alba D. viduata A. aegypt i aca P. galTbens i s S. melanota c. rnacrourus C. aeruginosus H. vocifer P. hal i aetus B. pallonina A. africana V. spinosus V. tectus C. hiaticula C. dubius C. pecuarius T. nebularis T. stagnatil is T. glsreola T. ochropus T. hypoleucos T. erythropus G. gall inago C. minuta P. pugnax H. himantopus P. aegypt ius l. cirrhocephslus S. nilotica C. hybridus
total
1 3 o o 8 3
14 25 6 1 1 4 4 3
11 15 5
12 1650
4 o
78 1 5 3 2 o 1
13 o 7 7 2
39 1
26 14 13 1 5
34 68 22 5 4 5 2
2128
o o
13 1 6 o 1 4
139 o 6 8 o
10 1 1 o 1
1311 o 6 8 2 o 1 o
590 1
23 3 o o o 2 o 6 9 3 3 o 1 o
16 o o o o
2176
additional observations, not included in the standardized counts. transect counts by canoe
(95" confidence limit) med. = median value estimate
Aa = CC BC FC MC RB
transect counts with fixed belt width transect counts by foot counts of migrat;ons to and from roosting sites river bank counts
TC tota l counts RC transect counts by car
The nurber preceding the count type indicates the n~r of transects or riv&r bank counts. The total transl!Ct length is given in brackets after the count type.
viii
Appendlci
Table 4.8. : Bird numbers in the Lake Chad Basin (total observed)
Kys- Maga Tikel.! Zina 2il im Lo- El Kala lIaza Lake Total matary gone Beid Maloue Chad obs. obs. obs. obs. obs. obs. obs. obs. obs. ob •. ob ••
obs. = number observed AO = additional observations, not included in the standardized counts. min. = mininun estimate CC = transect counts by canoe
(95 X confidence limit) BC = transect counts with fixed belt width med. = median value estimate FC = transect counts by foot
HC = counts of migrations to and from roosting sites RB = river bank counts TC = tota I counts RC = transect counts by car
The number precedi ng the count type i ndi cates the nuJber of transects or river bank counts. The total transect length is given in brackets after the count type.
xi
WIWO-report 67
Table 4.9. : Bird numbers in the Lake Chad Basin (estimate based on minimum values)
Kys- Maga likele Zina Zi I im Lo- El Kala lIaza Lake lotal matary gone Beid Maloue Chad obs. min. min. min. min. min. min. obs. obs. min. min.
Table 4.11.: Midwinter records in the study area of species that were not observed in our study (1980-1992)
years
Anhlnga rufa 1984
max. nu!i)er
site source
(Roux et Jarry) Clconi. nlQra 1987
1987,1991 1984,86, 87, 91 1984,86,87 1984
1 1 1
Maga (Roux et Jarry) Scopus ~retta Anas clypeata Aythya nyroca
5000 300
1 2
700
Maga, lake Chad (v.d. Zon, 1987; OAG MUnster, 1991) lake Chad, lIaza (Roux et Jarry; OAG MUnster, 1991) lake Chad, Maga (Roux et Jarry; v.d. Z~n, 1987)
Aythya fer i ne Thala.sornis leuconotus Recurvl rostra .vosetta Burh I nus senegal ens I s Burhlnus capensls Gallinula chloropus Fullca atra Porphyr I 0 porphyr i 0
(Roux et Jury) (Roux et Jarry) (Roux et Jarry) (OAG HUnster; 1991) (OAG MUnster; 1991) (Roux et Jarry; v.d. Z~n, 1987) (Roux et Jarry) (Roux et Jarry) (v.d. Zon, 1987; Haberland et van 01 jen, 1991)
Table 4.12.: Additional species list for the Lake Chad basin and the Benoue valley (midwinter). Records before 1980 are marked with * Species already mentioned in Table 4.11. are omitted.
max. nr.
Podiceps nlgrlcollis 3 *Phalaerocorax carbo [xobrychus s turml I [xobrychus ml nutus Butorldes strl atus Cieonia abdimii Platalea leucorodla 41 Phoeniconaias minor 11 Anser alblfrons 1 Anas pene l ope 71 Anas strepera 8 Anas crecca 176 Anas hottentota 70 Ana. eapen. I • *Marmaronetta angustlrostris Aythya fuligula 26 Grus grus Anthropoides vi rgo Porlana parva Porlana porlana ·Porzans marginal is Crex egregia Gall inula angulata Porphyrio porphyrlo Podi ca senega l ens I s Vanellu. albiceps Vanellus vanellu. Pluvial is squatarola Charadrius trlcollaris *Charedr;U8 leschenaul t; i Charadrlus forbesl I Charedri us marg i natus *LilOOsa lapponica Arenari8 interpres
2 1 1 2
1 4 6
Gall inago media 1 Calidris alplna 2 wCal idr;s alba GlareoLa cinerea
remarks
Benoue Benoue Benoue Benoue irreguLar Hadej i a-Nguru
1984-1987 Aerial survey Lake Maga, Logone floodplains, Lake Fitri, Lake Chad (Perennou, 1991) 1988·1990 Ground survey Hadedji'Nguru, Nigeria (Peremou, 1991) 1991 Ground survey Lake Maga and surroundings (DAG Munster, 1991) 1993 data current study
A. 1- 10 birds C. 101- 500 birds B. 11- 100 birds D. 501-1000 bi rds
:\'\'111
Appendlcl
Table 4.14. : Waterbird numbers counted at Lake Maga 1984-1993
1984-1987 Aerial survey (Roux et Jarry. in: V8n der Zon, 1987) 1991 Ground survey (OAG MUnster. 1991) 1993 current study
A. 1- 10 bi rds C. 101- 500 birds 8. 11- 100 birds D. 501-1000 bi rds
Appendlcl
Table 4.15.: Percentages of the estimated West-african (wintering) population per species per site. Estimates from Perennou, 1991. The post-breeding population is estimated as three times the number of pairs. Only sites with more than 1 % of a certain species are mentioned.
X of the West-African population obs. min. med.
Phalacrocorax africanus (75,000) Logone river 0.5 1.2 4.6 Lake Chad 0.1 1.9 3.8
Ardeola ralloides (12,000) Maga 0.5 0.5 1.3 depression likel~ 0.4 0.4 5.5 dry floodplain Zina 0.7 1.7 7.3 depression Zi I im/Zimade 6.4 44.7 101.0 Lake Chad 0.7 61.6 65.5
Bubulcus ibis (300,000) Maga 0.4 0.5 3.3 dry floodplain Zina 0.1 0.1 5.8 Lake Chad 0.1 1.1 2.0
Egretta alba (30,000) dry floodplain Zi"" 0.0 0.9 4.6 depression Zi I im/Zimado 3.3 4.1 7.8 Lake Chad 0.2 6.2 7.0
Egretts intermedi. (9,000) dry floodplain Zi"" 0.1 0.6 9.8 depression Zi I im/Zimado 5.8 5.8 16.7
Egretts garzetta (60,000) dry floodplain Zina 0.1 0.4 5.2 depress i on Z i l i m/Z i made 1.6 1.6 9.9 Lake Chad 0.1 0.8 1.4
Ardea cinerea (120,000) depress i on Z il i m/ IZ i made 0.5 0.5 2.5 Ardea purpurea (20,000) Lake Chad 0.2 6.9 16.5 Ciconia ciconia (60,000) dry floodplain Zi"" 0.4 0.4 4.6
depression Zi l im/Zimado 2.0 2.0 2.5 Ibis ibis (15,000) dry floodplain ZinB 0.2 0.2 4.9
Sarkidiornis melanotus (50,000) dry floodplain Zl na 0.1 0.1 7.6 Vanellus spinosus (100,000) depress ion li kele 0.1 0.1 1.1
Lake Chad 0.2 0.5 1.4 Ph i lomachus pugnax (1,000,000) Haga 0.5 0.5 1.1
depression Tikele 1.6 1.6 1.6 dry floodplain Zlna 0.5 0.5 6.1 Lake Chad 0.1 0.8 1.4
Himantopus himantopus (25,000) dry floodplain Zins 0.1 0.1 3.8 El Beid 0.7 1.0 3.0 Lake Chad 1.0 15.2 13.0
Glareola pratincola (10,000) Haga 6.8 6.8 6.8 depression likel~ 3.3 3.3 8.3 depression Zi I im/Zimado 6.8 6.8 6.8 logone r; ver 13.0 13.0 13.0 Lake Chad 0.3 0.3 1.7
Sterna nilotic8 (10,000) Logone river 0.3 1.1 4.6 ChI idonias leucoptera (20,000) Haga 2.0 2.0 2.0
XXI
wrwO-repoct 67
The observations at lake Chad concern quadrat 87 (P6rennou, 1991). The extrapolations coneern 150 km of lake-side at the southern fringe of lake Chad. The (provisional) estimate of the aile of the West-African population according to Perennou (1991) is presented between brackets after each spec! e •• obs. = nunber observed min. = mininun estimate med. = median estimate (see ch . 4)
xxii
Appe ndicl
APPENDIX 2: Tables 5.1 to 5.4
Table 5.1.: Mean number of individual birds or flocks recorded per km transect for the most numerous species, using all data. The foot and the road counts took place in the same area, the canoe counts in a different one.
Ind ividual bl rds flocks foot road canoe foot road
total km counted 30.5 100.0 13.5 total hours counted 11 23 4.5 mean speed (km/h) 2. 8 4.4 3.0
Table 5.2.: Bird densities (birds per ha) for the Tikele depression and the surroundings of Zina, resulting from the data of the counts by foot and the road counts. Calculation methods: King A, Jarvinen B1, B2 and B3 (see text).
Tikel!! Tikele foot (6.5 km) road (8 km) A Bl B2 83 A Bl 82/83
Table 5.3.: Two-sided 90 X confidence Intervals for the 2ina road count and count by foot data. Calcula' tion methods: King A, Jarvinen Bl, B2 and B3 (see text).
Table 5.4. hpplicability of three waterbird survey methods in Sahelian wet lands for different species-groups.
AH i • L survey
peL icans cormorants, herons, ibisses storks, spoonbi L Ls, cranes ducks, geese Ruff. BLack-tailed godwi t other wader., guL Is, terns raptors raiLs, crakes
+ (+)
incLuding PratincoLe
good appL i cabiL i ty lIIOderate appL i cabi L i ty
+
(+) +
(+)
xxv
Ground survey, Ground survey, focus. i ng s"""L ing on concentrations
(+) (+) +
+ (+) (+)
(+) + +
(+) +
+
WIWO-report 67
APPENDIX 3: Figure 3.1, 3.2 and 3.3
Figure 3.1: Mean numbers of birds per km at the lake Chad-side, plotted against the standard deviation on a log-normal distribution. each species is represented by + (cross) in the figure
Figure 3.2: Distribution of birds over the observation distances (in meters) in the road counts, counts by foot and canoe counts. Based on all data.
Figure 3.3: Distribution of birds over the observation distances in the counts by foot for respectively herons/storks/ibisses/crane, raptors, ruff and other species. Based on all data.
xxvi
Appendlcl
l oq ,---------~--------------------------------_,----------------~--,
-I ..
-I lOQ
Figure 3.1: Mean numbers of birds per km at the lake Chad-side , plotted against the standard deviation on a log-normal distribution. each species is represented by + (cross) in the figure
xwii
WIWO-,eport 6 7
~ 60
Herons, storks, ibisses and cranes --. -- - ··1
60 + ..
40
30
20 - .~
10 - ~-~- ... -, ~ o _ L ___ -L._L "-'f= t ~--~-;--. ---.,jv- I *"-=-* • ok
median dial. cla&& + median dist. class * median disL class
Figure 3.2: Distribution of birds over the observation distances (in meters) in the road counts, counts by foot and canoe counts. Based on all data.
XXIX
.. ~o :,
~, .~. 30 -
20 ;-
-+- raplors ~ rllff ~()lhers
Figure 3.3: Distribution of birds over the observation distances in the counts by foot for respectively herons/storks/ibisses/crane, raptors, ruff and other species, Based on all data,
Due to the outbreak of new firearm clashes in the south-west of Chad the initial idea of an integral count of both the Chadian and Camerounian parts of the Logone floodplain had to be abandoned. Instead all our attention became focused on the west bank of the lower part of the Logonc river and the El Beid region
10/1/93 arrival in Garoua 11/1/93 preparations at Ecole de Faune, Garoua 12/1/93 preparations at Ecole de Faune, count at Kysmatary, Garoua 13/1/93 preparations at CEDC, Maroua, count at Lagdo 14/1/93 preparations at CEDC, Maroua 15/1/93 travel from Garoua to Maga, site reconnaissance, canoe rental 16/1/93 counts in the surroundings of Maga, travel to Zina 17/1/93 counts in the surroundings ofZina 18/1193 counts at the depression of Zilim 19/1/93 travel to Mandabe, counts in the surroundings ofMandabe 20/1/93 counts at Mandabe and Zimado, travel to Kousseri, courtesy visit to the Prefecture 2111193 counts at Kalamaoue National Park, travel to and counts at El Beid river 22/1/93 counts at El Beid river, travel to Hali Alifa (Lake Chad) 23/1/93 counts at Lake Chad 24/1/93 counts at Lake Chad and travel to Waza, counts of the floodplains north ofWaza 25/1/93 counts at Waza National Park and surroundings 26/1/93 travel to Garoua 27/1/93 presentation of results at Ecole de Faune, Garoua 2/2/93 travel to Zimado 3/2/93 counts at Zimado 4/2/93 counts at Zimado 5/2/93 travel to Maroua
Appendlcl
APPENDIX 8: WIWO-reports published
The WIWO-report series intends to be a regular medium by which occasional reports of ornithological expeditions or journeys can be published. In this way regular distribution and availability are guaranteed. Quality differences are accepted,
since the authors remain responsible.
WlWO-reports published:
1. Altenburg W., M. Engelmoer, R. Mes & T Piersma 1982. Wintering waders at the Banc d'Arguin, Mauritania. Comm. 6 Wadden Sea Working Group. Leiden, The Netherlands. Sold out.
2. Kersten M., T Piersma, C. Smit & P Zegers 1983. Wader migration along the Atlantic coast of Morocco, March 1981. Texel, The Netherlands. 1 25,-
3. Philippona J. 1985. Waterbirds at some wetlaflds in Turkey and Greece. Zeist, The Netherlands. f 4,-4. Bijlsma R.G. & F.E. de Roder 1985. Waders along the coast of Thailand during November and
December 1984. Zeist, The Netherlands. f 10,-5. De Roder F.E. 1985. Waterbirds on some Turkish wetlands, October/November 1983. Zeist, The
Netherlands. f 8,-6. Van den Berk V., R.G. Bijlsma & F.E. de Rodcr 1985. The importance of some wetlands in Turkey for
transient and wintering birds in Turkey. Zeist, The Netherlands. f 10,-7. Piersma T. 1985. Wader studies and waterbirds in the Nakdong Estuary, South Korea, in September
1984. Zeist, The Netherlands. f 12,50 8. Chalabi B., J. Harrison & G. van DiJk 1985. Les zones humides du Nord-Est Algerien en 1984. Zeist,
The Netherlands. f 10,-9. Ens B. 1985. Tussen Sahara en Siberie. Ewijk. The Netherlands. f 5,-9. Ens B. 1985. Entre Sahara et Siberie. Ewijk, The Netherlands. Sold out. 10. Van den Berk V., D. van Dorp, O. van Hoorn & R. Vas 1986. Cranes and waterfowl counts of some
Turkish wetlands. Zeist, The Netherlands. f 7,50 11. Van Dijk AJ., K. van Dijk, L. Dijksen, T van Spanje & E. Wymenga 1986. Wintering waders and
waterfowl in the Gulf of Gabes, Tunisia, January-March 1984. Zeist, The Netherlands. f 30,-12. Bijlsma R.G. & F.E. de Roder 1986. Notes on the birds of some wetlands in Turkey. Zeist, The
Netherlands. f 7,50 13. Dijksen L.J. & F.J. Koning 1986. Mid-winter waterfowl census, Turkey 1986. Zeist, The Netherlands. f
10,-14. Schilperoord L. & M. SChilperoord-Huisman 1986. Observations of waterbirds in some wetlands in
Turkey, July/August 1986. Zeist, The Netherlands. f 7,50 15. Altenburg W 1987. Waterfowl in West African coastal wetlands: a summary of current knowledge.
Zeist, The Netherlands. f 10,-16. Bennet CA, R.G. Bijlsma & R. Stouthamer 1982. Survey of waterbirds on Egyptian wetlands,
autumn 1981. Zeist, The Netherlands. Sold out. 17. Schekkerman H. & J.C.J. van Wetten 1987. An ornithological winter survey of Lake Turkana, Kenya.
Zeist, The Netherlands. f 7,50 18. Dijksen L.J. & P van der Wolf 1987 Mid-Winter waterfowl census Turkey January 1987 Zeist, The
Netherlands. f 7,50 19. Lensink R. 1987. Notes on the birds of some wetlands in North-East Greece and Turkey. Zeist, The
Netherlands. f 7,50 20. Meininger PL. (ed.) 1990. Birds of the wdlands in North-East Greece, spring 1987 Zeist, The
Netherlands. f 30,-21. Dijksen L.J. & A-M.C. Blomert 1988. Mid-Winter waterfowl census Turkey, January 1988. Zeist, The
Netherlands. Sold out. 22. Van der Have T., V M. van den Berk, J.P. Cronau & M.J. Langeveld (eds.) 1988. South Turkey
Project. A survey of waders and waterfowl in the Cukurova deltas, spring 1987 Z~ist, The Netherlands. Sold out.
23. Altenburg W & J. van der Kamp 1989. Etude ornithologique preliminaire de la zone cotiere du nordouest de la Guinee. Joint report of WIWO and ICBP. WIWO-report 23 and ICBP study report 3D, Zeist, The Netherlands and Cambridge, England. Sold out .
. 1
wrWO·repart 67
24. Chalabi B. & G. van Dijk 1988. Les zones humides dans la region de Anaba et El Kala en Mai 1987. Zeist, The Netherlands. f 7,50
25. Ens B.J., T Piersma, W.J. Wolff, L. Zwarts (eds.) 1989. Report of the Dutch-Mauritanian project Banc d'Arguin 85/86. Joint report of WIWO and RIN nr. 25, Zeist, The Netherlands. f 30,-
26. Altenburg W., E. Wymenga & L. Zwarts 1992. Ornithological importance of the coastal wetlands of Guinea-Bissau. Zeist, The Netherlands. Sold out.
27. Ens B.J., T. Piersma, W.J. Wolff & L. Zwarts 1990. Homeward bound: Problems waders face when migrating from the Banc d'Arguin, Mauritania, to their northern breeding grounds in spring. WIWO/RIN, ZeistlTexel. f 55,-
28. Van Winden A, K. Mostert, P. Ruiters, M. S_k_ & H. de Waard 1989. Waders and waterfowl in spring 1988 at Eber Galu. Zeist, The Netherlands. f 10,-
29. Van den Berg AB. 1988. Moroccan Slender,.billed Curlew survey, winter 1987-88. Joint report of WIWO and ICBP nr. 29, Zeist, The Netherlands and Cambridge, England. Sold out.
30. Keijl G.O., M.W.J. van Roomen, P.S. Ruiters & A Wijker 1992. Migration of waders and other waterbirds along the Mediterranean coast of Israel, spring 1989. Zeist, The Netherlands. f 20,-
31. Dijksen L.J. & A-M. Blomert 1989. Mid-winter waterfowl census Turkey January 1989. Zeist, The Netherlands. f 10,-
32. Schekkerman H. & M.W.J. van Roomen (eds.) 1993. Migration of waterbirds through wetlands in Central Anatolia, spring 1988. Zeist, The Netherlands. f 15,-
33. Spiekman H. & N. Groen 1993. Survey (If breeding waders and relations between breeding performance and lemming densities, North-East Taimyr, Siberia, 1992. Zeist, The Netherlands. f 15,-
34. Van den Berk V.M., J.P. Cronau & T.M. van der Have 1993. Waterbirds in the Van Province, eastern Turkey, May 1989. Zeist, The Netherlands. f 15,-
35. Altenburg W. & J. van der Kamp 1991. Ornithological importance of coastal wetlands in Guinea. Joint report of WIWO and ICBP. WIWO-report 35 and ICBP study report nr. 47, Zeist, The Netherlands and Cambridge, England. f 10,-
36. Van der Have TM., S. van der Sant, Y. Verkwil & J. van der Winden (eds.) 1994. Waterbirds in the Sivash, Ukraine, spring 1992. Zeist, The Netherlands. f 20,-
37. Dijksen L.J. & A-M. Blomert 1993. Mid-wintE!r waterfowl census Turkey January 1990. Zeist, The Netherlands. f 15,-
38. Spiekman H.W., G.O. Keijl & P.S. Ruiters (eds.) 1993. Waterbirds in the Kneiss area and other wetlands, Tunisia. Eastern Mediterranean Wader Project, spring 1990. Zeist, The Netherlands. f 25,-
39. Wolff W.J. (ed.) 1998. Waders in Guinea-Bissau, October 1992-May 1993. The end of the East Atlantic Flyway. f 25,-
40. Meininger P.L & GAM. Atta (eds.) 1994. Ornithological studies in Egyptian wetlands 1989/90. FORE-report 94.01, Zeist, The Netherlands. f 35,-
41. Schepers F.J. & E.C.L. Marteijn (eds.) 1993. Coastal waterbirds in Gabon. Zeist, The Netherlands. f 30,-
42. Hirschfeld E., SA Moharned & T Stawarczyk 1992. Bahrain wader study 1991. Zeist, The Netherlands. f 10,-
43. Brehme S., T. Muller & J. Redlich 1992. Bird observations in the Danube delta and in the Dobrodgea (Romania). Zeist, The Netherlands. J 10,-
44. Spiekman H. 1992. Wader ringing in Tunisia 1962-1986. Zeist, The Netherlands. f 10,-45. Hustings F. (ed.) 1994. Bird census in the K_z_l_rmak delta, Turkey, in 1992. Zeist, The Netherlands.
f 25,-46. Meininger P.L., PA Wolf, DA Hadoud & M. FA Essghaier 1994. Ornithological survey of the coast
of Libya, July 1993. Zeist, The Netherlands. f 15,-47. Groen NM. & P.J. Zomerdijk (eds.) 1994. Waders and waterbirds along part of the Atlantic coast of
Morocco, autumn 1991 - spring 1992. Zeist, ne Netherlands. f 20,-48. Kivit HA, H. Nijmeijer & A Ova a (eds.) 1994. Wader and waterfowl migration in the Cukurova
deltas, South Turkey, spring 1990. Zeist, The Netherlands. f 25,-49 HagemelJer W.J.M. (ed.) 1994. Wintering waterbirds in the coastal wetlands of Albania, 1993. Zeist,
The Netherlands f 20,-50. Dijksen L.J. & M C M Klemann 1994. Wintering geese in Turkey, January 1992. Zeist, The
Netherlands. f 15,-51. Gemtsen G.J. & N.M. Groen 1995. Icelandic Black-tailed Godwit project 1993. Zeist, The
Netherlands. J 15,-53. De Nobel W.T (ed.) 1995 Birds of the Messo1onghi wetlands. Eastern Mediterranean Wader Project,
11
Appendlcl
Taimyr, Siberia, in 1991. Zeist, The Netherlands. f 20,-56. Van den Brink B, R.G. Bijlsma & TM. van der Have (eds.) 1998. European Swallows Hirundo rus/ica in
Botswana. Zeist, The Netherlands. f 20,-57. Tulp I., L.W. Bruinzeel, J. Jukema & O. Stepanova 1997 Breeding waders at Medusa Bay, western
Taimyr, in 1996. Zeist, The Netherlands. f 20,-58. Van den Brink B., R.G. Bijlsma & T.M. van der Have (eds.) 1998. European songbirds and Barn Swallows
Hirundo rus/ica in Ghana: a quest for Constant Effort Sites and swallow roosts in December/January 1996/97. Zeist, The Netherlands. f 15,-
59. Veerman 1. & H. Wessels 1998. Cranes Grus grus in NW-Russia, autumn 1996. Zeist, The Netherlands.
f 10,-60. Willems F.J. & E. de Vries 1998. Ecological aspects of Pygmy Cormorants Phalacrocorax pygmeus in
Prespa, Greece, May-August 1996. Zeist, The Netherlands. f 25,-61. Tulp I., H. Schekkerman, T. Piersma, J. Jukema, P. de Goeij & J. van de Kam 1998. Breeding waders at
Cape Sterlegova, northem Taimyr, in 1994. Zeist, The Netherlands. f 25,-62. Keijl G.O., P.S. Ruiters, TM. van der Have, A bij de Vaate, E.C.L. Marteijn & R. Noordhuis 1998. Waders
and other waterbirds in the United Arab Emirates, autumn 1994 and spring 1995. Zeist, The Netherlands
f 25,-63. Schepers F.J., G.O. Keijl, P.L. Meininger & J.B. Rigoulot 1998. Oiseaux d'eau dans le Delta du Sine
Saloum et la Petite Cote, Senegal, janvier 1997. WIWO, Zeist, The Netherlands & Direction Pares Nationaux du Senegal, Dakar, Senegal. f 30,-
67. J.C.J van Wetten & P. Spierenburg 1998. Waders and waterbirds in the floodplains of the Logone, Cameroun, January 1993. f 20,-
These reports can be ordered by mentioning the numbers and paying the mentioned amounts plus 00. IS.-for administrationcost for each separdte order from abroad, directly to postal giro account 2666009 or ABN-AMRO bankaccount 570216613 of Stichting WIWO, Fetha 23, 3633 eT Vreeland, Netherlands or by sending cash or an Eurocheque with the bankpassnumber filled in at the back to that adress . The last two methods of payment are free of administrationcost. Postage for surface mail is included in the prices. Any information about WIWO-reports can also be obtained from that adress.
iii
WIWO Foundation Working Group International Waterbird
and Wetland Research
WIWO, a foundation according to Dutch law, was established in 1983 to form an initiative-supporting centre for Dutch ornithologists who organised ornithological expeditions to West Africa, and to encourage that kind of future activities. The aims of WIWO are to initiate, stimulate and organise scientific research on waterbirds and wetlands and to publish the gathered knowledge for conservation of the populations involved and their habitats. WIWO is an organisation working only with unpaid volunteers which have a varying scientific background. The research data collected in the field during WIWO-projects require often laborious efforts at home to evaluate, analyse and report them. Up to 1995 more than 50 projects have been carried out in 18 different countries.
During the expeditions important data have been collected on waterbirds in a large number of wetlands along the entire East Atlantic Flyway, Mediterranean Flyway and in the Arctic. Furthermore, information has been collected on the function, importance and extent of protection and threats to the studied wetlands. WIWO's main role is to gather knowledge on migration, wintering and breeding of waterbirds and their habitats, to provide a sound basis for the conservation of the populations and areas concerned. The results are made available to all organisations and individuals active in nature conservation through a series of WIWO-reports as well as scientific and popular papers. Up to 1994 about 50 reports and more than 120 articles have been published. WIWO does not act as co-ordinating organisations such as Birdlife International and the International Union for the Conservation of Nature (lUCN). The main strength of WIWO is its capability to mobilise volunteers with a professional attitude for ornithological research expeditions.
All correspondence should be sent to WIWO, P.O. Box 925, NL-3700 AX Zeist,