Pollination is an essential ecological service that has been studied intensively in the last decade because of its importance in agriculture and conservation, and because of the threat of global change. In South Africa, the important Cape Floristic Region and unique vegetation types (e.g. Succulent Karoo and Nama Karoo), where temperature and water balance will change markedly, are at great risk. Due to their importance as pollinators, holdings of insect specimens have grown rapidly in natural history collections, and so have the formats, inconsistencies and inaccuracies in biodiversity information. In South Africa a shortage of the right skills has hampered the development of biodiversity information management and the adoption of available standards. In this context a pilot project entitled Museum Data Migration had been initiated in April 2010 to:

• Clean specimen records from museums
• Migrate data to Specify 6
• Train end-users
• Develop capacity to continue this work

The project was divided into two parts:

• Part 1 – Data cleaning and migration: They extended the Museum Data Migration project. Their philosophy was to improve data quality in the museum – too often specimen records became separated from the voucher specimens. This included development of the Catalogue of Afrotropical Bees.
• Part 2: They conducted research into biodiversity ontology in the domain of bee specimens and pollination information. They developed a semantic mediator to integrate expert behavioral ecology information, on flower-visiting and pollination from specimen-records in distributed, heterogeneous specimen databases. They also trained a MSc. Student in molecular systematics of the bee genus Scrapter.

• A web interface to the implemented system for semantic interoperability among heterogeneous, distributed databases of flower-visiting records.
• The project plans to improve the quality and accessibility of enriched flower-visiting records for use in research on African pollinators.

• Coetzer, W; O Gon; M Hamer; F Parker-Allie (2012) A new era for specimen databases and biodiversity information management in South Africa. Biodiversity Informatics. 8: 1–11. (link)
• Coetzer, W; D Moodley; A Gerber (2013) A Case-Study of Ontology-Driven Semantic Mediation of Flower-Visiting Data from Heterogeneous Data-Stores in Three South African Natural History Collections. In The Semantic Web: ESWC 2013 Satellite Events. (P Cimiano; M Fernandez; V Lopez; S Schlobach; J Volker; Eds.) Lecture Notes in Computer Science. No. 7955. p. 87-100. (link)
• Coetzer; W; D Moodley; A Gerber (2014) A knowledge-based system for discovering ecological interactions in biodiversity data-stores of heterogeneous specimen-records: A case-study of flower-visiting ecology. Ecological Informatics. 24: 47-59. doi:10.1016/j.ecoinf.2014.06.008. (link)
• Coetzer, W; Moodley, D; Gerber, A (2016) Eliciting and Representing High-Level Knowledge Requirements to Discover Ecological Knowledge in Flower-Visiting Data. PLoS One. 11: e0166559. (link)
• Coetzer, W; Moodley, D; Gerber, A (2017) A knowledge-based system for generating interaction networks from ecological data. Data and Knowledge Engineering. 112: 55–78. (link)

Specimen Digitization and Biodiversity Information Improvement

• A total of about 48 000 specimen records of pollinators were digitized at Iziko Museums, Naturalis and the Agricultural Research Council. 16000 specimen labels were photographed in the ARC bee collection, which not only allowed the project to capture 6275 undigitized specimens, but also gave them new ecological information and allowed them to correct hundreds of errors and inaccuracies (also among existing digital records).
• The Catalogue of Afrotropical Bees contains 2449 accepted species that match COL LSIDs and 352 that need LSIDs; Accepted subspecies that match / need LSID: 9 / 75; Synonyms that match / need LSID: 82 / 4500. There are 7483 citations (that have a nomenclatural status) among 702 references. Country distribution is included (standardized abbreviations).
• A MSc. thesis on the molecular phylogeny of the bee genus Scrapter was produced by Qiniso Mthethwa.

Semantic Interoperability

• This project designed and implemented a system architecture for semantic interoperability using an ontology of flower-visiting behavioural ecology. The improved biodiversity information that was digitized was used to design the ontology and system architecture, and to design the system implementation.

This project learned two primary lessons:

• On international collaboration: get the contract to collaborate signed before your funded project kicks off
• About students: your interview should be more rigorous than a job interview, and get references from more than one person