Diversity and population densities of coraciiform birds in Zambezi riparian forest

Diversity and population densities of coraciiform birds in Zambezi riparian forest. A territory mapping method was used in 2015 to assess the population density of coraciiform species breeding in a riparian forest on the Zambezi River near Katima Mulilo, NE Namibia. The forest, c. 280 ha in surface area, was partly transformed by human settlement and arable grounds. A total of 13 species and 42 breeding pairs were recorded. Population densities were (pairs/100 ha) as follows: grey–headed kingfisher 4.3; giant, pied and woodland kingfisher, each one with 1.1; malachite kingfisher 0.7; striped kingfisher 0.4; white–fronted bee–eater 1.4; little bee–eater 1.1; lilac–breasted roller 0.7; broad–billed roller 1.1; African hoopoe 0.7; red–billed wood–hoopoe 1.1; and scimitar–billed wood–hoopoe 0.4. Data published through GBIF (Doi: 10.15470/s9rlud)


Introduction
Coraciiforms (kingfishers, bee-eaters, rollers and allies) belong to a highly attractive group of birds and may play an important ecological role in some habitats. They are especially well-represented in sub-Saharan Africa where they are grouped in three families: bee-eaters (Meropidae), rollers (Coracidae) and kingfishers (Alcedinidae). Some authorities also include the hoopoe (Upupidae) and wood-hoopoe (Phoeniculidae) families, as I also do in this study.
Many southern African species belonging to the order Coraciiformes are relatively wellstudied (e.g. Fry et al., 1992), especially in regard to their breeding biology and diet (Kopij, 2000(Kopij, , 2018a. However, little information on population density is yet available for most of them (Hockey et al., 2005). Since many coraciiform species breed in low densities, such studies are difficult because they have to be conducted on relatively large study plots.
This note reports on the distribution and population density of coraciiform species breeding in a riparian forest in NE Namibia. Since the riparian forest is under threat of degradation, it is important to create a basis for monitoring bird species that may be indicative of habitat naturalness.

Study area
The study area was located in the Zambezi Valley near Katima Mulilo in the Zambezi Region, NE Namibia. The study area comprised a forest stretching between the river and the international road from Zambia through Katima Mulilo to Botswana, lying on the left bank between the Wenela Bridge (border pass) and the Zambezi River lodge. This forest area is 7 km long and 200-700 m wide (mean = c. 400 m). The approximate surface area is therefore c. 280 ha.
The natural vegetation is classified as Riparian Zambezi Forest (Mendelsohn et al., 2009 The forest is interlaced with pans covered with grass and sedges flooded almost on a yearly basis. About one quarter of the land has been converted into arable grounds and built-up areas, but large trees usually remain even in these converted areas ( fig. 1).
The annual temperature for Katima Mulilo is 21 ºC. The average maximum temperature during the hottest month (September) is 35 ºC; the average minimum temperature during the coldest month (July) is 3 ºC. Humidity is 80-90 % in the most humid month (February) Fig. 1. Distribution of breeding pairs/occupied territories/cooperatives groups of the: kingfishers, bee-eaters, rollers, and hoopoes and woodhoopoes in a Zambezi Forest in NE Namibia: a, Zambezi riparian forest; b, grassland (flooded area); c, Colophospermum mopane forest; d, Kalahari woodland; e, arable grounds; f, urbanized built-up areas; g, rural area; h, Zambezi River; i, border of the study area. Fig. 1

Methods
A territory mapping method (Bibby et al., 2002) was used to assess the population densities of all coraciiform species nesting in the forest. The study area was divided into six sections. Birds were counted in each of the sections in one morning.Therefore, the whole study area was covered in six mornings. Such complete coverage was achieved four times in 2015, in August, September, October and November. During each count, all coraciiform birds seen or heard were plotted on the map 1:500. Special attention was paid to birds showing territorial behaviour or breeding display. Recordings of two or three simultaneously calling males were important in interpreting the results. A bird or pair of the same species recorded at the same site in at least two out of the four months was interpreted as residential/ breeding/territorial (following Bibby et al., 2002).

Results and discussion
In total 13 coraciiform species were recorded as breeding in 2016: six kingfishers, two beeeaters, two rollers, two wood-hoopoes, and one hoopoe species (see dataset published in GBIF, Doi: 10.15470/s9rlud). The coraciiform community in the riparian forest was dominated by kingfishers, which comprised more than 50 % of all breeding coraciiform species. By far, the most numerous   (table 1). The bee-eaters were represented by the white-fronted bee-eater Merops bullockoides breeding in a small colony on the bank of the river, and the little bee-eater Merops pisillus breeding in a built-up area within the forest. Overall density of bee-eaters was 2.5 pairs/ 100 ha (table 1).
Territories of three pairs of the broad-billed roller Eurystomus glaucurus and two territories of the lilac-breasted roller Coracias caudatus were recorded in the area. Rollers nested, therefore, in a density of 1.3 pairs/100 ha (table 1). N There were also two pairs of the African hoopoe Upupa africana, three breeding units of the red-billed wood-hoopoe Phoeniculus purpureus and one pair of the scimitar-billed wood-hoopoe Rhinopomastus cyanomelas. The overall density of all hoopoes and woodhoopoes was 2.1 pairs/100 ha.
Out of 42 breeding pairs of coraciiform birds, 15 territories have been established in a pure forest (35.7 %), 9 territories in built-up areas (21.4 %), 11 on the border of built-up and forested areas (26.2 %), 6 (14.3 %) on the border of built-up and arable areas and only three territories (7.1 %) located mainly in the grassy pans ( fig. 3).
The number of coraciiform species recorded in the riparian forest near Katima Mulilo appears to be very high (N = 13), probably one of the highest in southern Africa. It should also be mentioned that in close proximity to the study area, a large colony (c. 500 nests in 2016) of the Southern carmine bee-eater Merops nubicoides, c. 50 pairs of the white-fronted bee-eater and a few pairs of the swallow-tailed bee-eater Merops hirundineus have been recorded (Kopij, 2018a). In the neighbouring town of Katima Mulilo, well-endowed with indigenous and exotic trees (highly modified riparian forest), the number of coraciiform species was nine (Kopij, 2016); in the Kalahari woodland (12 km long transect) in the same Zambezi Region (also 9; Kopij, 2017), in Kasane on Chobe River (5; Kopij, 2018c). Outside the Zambezi Region the number of coraciiform species was much lower, e.g. in urbanized areas in the north central Namibia it was four (Kopij, 2014); in Bloemfontein it was also four (Kopij, 2001a(Kopij, , 2015, in Roma Valley, Lesotho (3; Kopij, 2001b), in Swakopmund on Atlantic coast (none; Kopij, 2018b).
To date, few data have been published on the population densities of coraciiform birds in Africa (Hockey et al., 2005). For the lilac-breasted roller the highest density was recorded in Mababe Depression in the Chobe National Park, Botswana (Herremans et al., 1997). However, the density varied considerably even within the same year. For example, in N and E Botswana, it was highest in August/October (10 birds/100 km), and lowest in November/December (2.5 birds/100 km) (Herremans and Herremans-Tonnoeyr, 1994). The little bee-eater nested at a density of two pairs/100 ha in broad-leaved woodland in the Limpopo Province, South Africa (Tarbotton et al., 1987). Kopij (1999Kopij ( , 2001aKopij ( , 2015 recorded 32 African hoopoe breeding pairs (i.e. 0.6 pairs/100 ha) and three scimitar-billed wood-hoopoe breeding pairs (i.e. 0.1 pairs/100 ha) in 5,000 ha of urbanized, well-timbered habitats of Bloemfontein, South Africa (Kopij, 2001a(Kopij, , 2015.
High species diversity and population densities of coraciiform species recorded in riparian forest indicate that this vegetation, even if partly modified by human activities, may play a very important role in the protection of these species, which can be regarded as indicators of a high level of biodiversity. Protection of any remaining stretches of this rich and unique forest vegetation along the Zambezi River should be among the highest conservation priorities in Namibia.