Comparisons of the habitat preferences of different African kingfisher species have in the past mainly concerned the influence of single habitat variables on kingfisher presence [for example Monadjem et al. (1994) compared perch height preferences between Giant kingfisher (Ceryle maxima), Half-collared kingfisher (Alcedo semitorquata) and Pied kingfisher (Ceryle rudis)].
In this study, by regarding multiple habitat variables we compare the habitat preferences of four African species of kingfisher: Giant, Half-collared, Pied and Malachite (Alcedo cristata), along the Kilombero River, southern Tanzania.
Riverine habitat of the Kilombero River is cleared for purposes such as agriculture and livestock grazing and this may negatively affect the resident bird community (Baker & Baker, 2002). We aim to highlight those habitat variables that influence kingfisher habitat use, in an area experiencing moderate levels of anthropogenic pressure.
Materials and methods
The banks of the Kilombero River are heavily vegetated with shrubs, trees and vines and this habitat holds important bird populations including African Finfoot (Podica senegalensis) and Pel’s Fishing Owl (Scotopelia peli; Baker & Baker, 2002). Much of the watercourse is shaded by extensive tree canopy.
Bird surveys were conducted between August 2005 and February 2006 (ten surveys per month from 07.30 hours to 09.30 hours and from 13.00 hours to 15.00 hours) along a 3-km stretch of the Mafinji River, a branch of the Kilombero River, in the Kilombero Valley (8°34′S. 8° 34′E). The Kilombero River flows south-westerly between the Selous Game Reserve and the Kilombero Game Controlled Area (Fig. 1).
Six experienced observers walked along the river (spanning across the whole river) and opportunistically searched for kingfisher species. Surveys were carried out on sunny, calm days, as kingfishers are the most active when visibility is good (Douthwaite, 1976).
During the survey, the following were recorded: (i) kingfisher species, (ii) perch height, measured using a tape measure (m), (iii) an estimate of overhanging branches <3 m in height (as a percentage of the river bank 20 m either side of the perched bird), (iv) an estimate of overhanging branches >3 m in height (as a percentage of the river bank 20 m either side of the perched bird), (v) river width (m), measured using a tape measure, (vi) river depth (m), measured using a plastic tape measure in the central river channel and 1 m from the north and south bank (=average taken of the three), (vii) river speed, a ‘poo-stick’ method used, in which the time it took a stick to travel 10 m in the water was recorded using a stop clock in the central river channel and 1 m from the north and south banks (=average taken of the three).
One-way analysis of variance was suitable for determining which habitat variables influenced kingfisher habitat preference.
The most frequently observed kingfisher species during this survey were Pied and Malachite kingfishers, with Giant kingfisher being the least observed species (Table 1).
Table 1. The number of individuals of each of the surveyed kingfisher species, the percentage of individuals from each species across all species and the average number of individuals of each species observed per day
% of all individuals observed
Average number of individuals observed per day (four survey hours)
The values presented here should not be taken as precise species richness values because the same individual may have been recorded more than once. These values must therefore only be used as frequency estimates.
Giant kingfisher was observed significantly more often in wider river stretches and the Malachite kingfisher showed a significant preference for narrower river stretches (F2.40 = 3.91, P <0.01; Fig. 2a). The Giant and Pied kingfishers utilized higher perches (>3 m) and the Half-collared and Malachite kingfishers preferred lower (<1 m) perch heights (F2.40 = 89.83, P <0.01) (Fig. 2b).
Malachite and Half-collared kingfishers were found in significantly higher numbers in areas of the river with a high proportion of branches (<3 m within 20 m of observed individuals), compared with the Pied and Giant kingfishers (F2.40 = 2.96, P <0.05) (Fig. 2c). Giant and Pied kingfishers favoured a greater depth of water than the Half-collared and Malachite kingfishers (F2.40 = 4.55, P <0.01) (Fig. 2d).
The proportion of branches (>3 m) and river speed did not influence kingfisher habitat preference (P >0.05).
Different habitat preferences between kingfisher species based on their foraging strategies have been documented previously (Fry, Keith & Urban, 1988; Monadjem, 1996). Monadjem et al. (1994) found Giant and Pied kingfishers to favour perch-sites 2–4 m high, whilst Half-collared kingfishers favoured perches <2 m in height. Our study supports this earlier research and has found that Malachite kingfishers also prefer lower perch-sites <1 m high.
Half-collared and Malachite kingfishers preferred shallower sections of the river. This may be because Half-collared and Malachite kingfishers perch at lower heights from the water, and therefore the water depth required for full submersion (thus improving the chance of catching prey) is shallower than it would be if they were plunging from a greater height; where as the Giant kingfisher may forage in deeper waters to reduce the chance of injury that could be caused if this large bird plunged into shallow waters.
Although Pied kingfishers have alternative hovering foraging strategies (Douthwaite, 1976), this active searcher strategy was only recorded on two occasions during this survey (C. Bonnington, unpublished observations). The width and depth of this watercourse probably discourage this hovering foraging behaviour. Therefore, we can assume that foraging from a stationary perch is the favoured strategy for the kingfishers in our study.
The destruction of vegetation near the river’s edge may cause reduction in suitable perches for Giant and Pied kingfishers (removal of taller flora) and Half-collared and Malachite kingfishers (removal of smaller flora). In addition, species such as the Half-collared kingfisher are particularly sensitive to human disturbance and prefer well-wooded streams (Fry, Fry & Harris, 1992). Therefore, any human activity that opens up the river habitat, would negatively impact this species.
This study has reported habitat niche segregation between (i) Giant and Pied kingfishers and (ii) Half-collared and Malachite kingfishers, with Giant and Pied kingfishers favouring foraging areas with higher perch-sites and deeper and wider river stretches, and Half-collared and Malachite kingfishers preferring lower perch-sites near shallower, narrower river stretches.
This study was produced as part of the Frontier-Tanzania Savanna Research Programme, which is a collaboration between the University of Dar es Salaam (UDSM) and the Society for Environmental Exploration (SEE). I would like to thank Robert Cooper (Frontier-Tanzania), a Research Assistant during July–September 2005, for developing the methodology used during this study. Neil Baker (Tanzania Bird Atlas), Darcy Ogada, Elias Mungaya (Wildlife Conservation Society of Tanzania) and Matthew Grainger provided comments on an earlier edition. I would also like to thank the personnel at Ulanga District Department of Natural Resources and Kilombero Valley Teak Company (KVTC), for all their help and assistance. Particular thanks to Peter Msangameno, Abdul Mwangalile and Christopher Nchimbi (game scouts) and all staff and Research Assistants (Frontier-Tanzania). I would like to dedicate this work to our long-serving game scout Israel Mwansele, who passed away recently. It was a pleasure working with him, and he will be sadly missed.