Birds of Zambia – a comprehensive review and development of a novel digital identification tool

Background

Ok, you can forget about all the fancy books concerning birds of Southern Africa. Yes, including Roberts Birds of Southern Africa, which while it claims to be the ‘most comprehensive single volume handbook on the African continent’ is…well…a little too comprehensive and doesn’t adequately explore the birds of Zambia.

Aspinwall and Beel have admirably dedicated a book solely to Zambian birds in ‘A Field Guide to Zambian Birds not found in southern Africa’. However, their illustrations leave much to the imagination and don’t provide readers with quick, accessible tool to identify birds.

For years ornithologists have therefore been desperately calling out for not only a definitive text on Zambian birds but for a modern method to rapidly identify birds of Zambia. In this prospective, observational study, we set out to characterise birds in the landlocked African country of Zambia and develop a robust digital pipeline for rapid and accurate bird identification.

Methods

Over the last 4 months, a dedicated, international, intergenerational team of birders painstakingly observed and documented birds from at least 5, maybe 6 different locations in Zambia.

Authors (AN, AJ, FN) camped out in hides (sat on the grass in their backyard) and using a Nikon Prostaff 3s or two toilet rolls tapped together observed birds in their natural habitat (the authors habitat that is). On scrap paper from the ZADP office AN and FN carefully reproduced the birds before colouring each individual bird (generally within the lines) with a 24 component Fabriano pencil set (fig 1). Distinctive bird behaviours were duly described.

Figure 1: Author FN and her Fabriano pencil set

The reproduction was immediately digitalised (photo taken with phone) and sent in a de-identified, but re-identifiable format (whatsapp) to expert birders in Tasmania, Australia (authors MB and PJ). Given the detailed information and lifelike drawings it was unsurprising that a proposed identification (ID) was rapidly achieved in the majority of cases. Using an established database of Zambian birds (google images) authors AN and FN correlated the proposed bird ID with the realtime bird.

Results

Out of a potential 779 Zambian bird species (according to wiki) a total of 23 birds were observed with sufficient detail to form a reproduction (figures 2-24).

17/23 birds were identifiable by authors PJ and MB with a median response time of 5.3 +/- 0.2 minutes. 17/17 proposed identifications were deemed correct after correlation with the Zambian bird database (ie gold standard).

5 birds could not be identified: 2 represented new species identified from Kalomo, South-Eastern Zambia, observed solely by author FN and subsequently named the Orange Eyed Beaker and Pleuder Plover. 1 unidentified birds was actually another male paradise flycatcher, despite one of the authors already doing a very good drawing of it. The remaining 2 unidentified birds had initial descriptions, behaviours and drawings by authors AN which were clearly inadequate. Following the provision of additional descriptive information (size, colouring etc) a correct identification was ultimately made in both cases.

The red-faced mousebird was renamed a superhero bird.

Discussion

Here, we have provided a definite and comprehensive analysis of Zambian birds. This novel, proof of concept study describes a unique digital technique for bird identification in Zambia. This is cheap, practical, and easily translational approach for identifying birds. We have also identified two previously undescribed species, further reflecting the richness of bird-life in Zambia.

Future studies will involve a continent-wide roll-out of this system and the development of an interactive app allowing realtime chat with birders PJ and MB.

We propose that this novel system be now accepted as the gold-standard for bird identification in Zambia.

Citation:

Neal F, Neal A, Jarman A, Brock, M, Jarman P. Birds of Zambia – a comprehensive review and development of a novel digital identification tool. Blogger 2017.

Figures (2-24):