South Africa’s National Parks Rebuild Their Alien Species Lists With a Ten-Step Data Overhaul

South African National Parks have quietly completed one of the most ambitious data-cleaning exercises in invasion science, and the results are striking. A new study published in the journal NeoBiota describes how researchers consolidated three decades of scattered records into a single, living checklist of alien plants across all 21 national parks. The final tally reached 1,330 unique alien taxa, nearly double the 752 recorded in the previous compilation from 2017. That jump is not simply a story of more invaders arriving; it reflects a systematic effort to rescue data from old publications, herbarium specimens, government clearing contracts and citizen-science platforms, then reconcile them under one taxonomic standard.

The stakes are higher than a longer list might suggest. Protected areas sit at the centre of global conservation commitments, including the Kunming-Montreal Global Biodiversity Framework, whose targets demand expanded conservation areas, ecosystem restoration and action against alien species. Reliable species checklists underpin everything from early detection of new invasions to the allocation of scarce management budgets. Yet the study’s authors found that the raw material for such lists was fragmented, inconsistent and often unverifiable. Different sources used different names for the same plant, spelling errors abounded in management databases, and some species had been misidentified for years.

At the heart of the problem is a deceptively simple challenge: matching species names across datasets and across time. Alien species are especially troublesome because the scientific names used in their introduced ranges often lag behind those in their native ranges, where taxonomy moves faster. National checklists, which tend to focus on native flora, frequently omit aliens altogether or retain outdated names. The researchers resolved this by adopting the Global Biodiversity Information Facility (GBIF) taxonomic backbone, which follows the World Checklist of Vascular Plants and covers both native and alien species. Every record was assigned a unique taxon identifier, allowing names from wildly different sources to be reconciled automatically and updated as taxonomy evolves.

The team’s proposed workflow unfolds in ten steps, moving from defining the geographic and taxonomic scope of the list, through compiling and standardising data from diverse sources, to curating everything in a relational occurrence database from which up-to-date checklists can be generated on demand. Each occurrence record receives a globally unique identifier, and both the original verbatim name from the source and the cleaned, standardised name are retained. Records are annotated with information on identification certainty and spatial accuracy, so that uncertain observations can be filtered out or flagged rather than silently contaminating the list.

The data sources themselves tell a story about how biodiversity information is produced. Botanists working in the Garden Route National Park built their list from field surveys and a herbarium established in 1998, sending unfamiliar specimens to institutions such as the Compton Herbarium for expert identification. Kruger National Park’s herbarium, founded around 1950, now holds more than 15,000 specimens covering roughly 2,000 species. At the other end of the spectrum, the government’s Working for Water Information Management System, which tracks alien clearing contracts, contained numerous spelling errors, misidentifications and surrogate names, where one species was recorded in place of another with similar clearing requirements.

Citizen science emerged as a powerful but double-edged contributor. The researchers drew on more than 200,000 iNaturalist records from parks and their surrounding catchments, but included only research-grade observations, where multiple users have agreed on an identification. Even so, misidentifications slipped through, and the authors caution that research-grade status guarantees only that two users agree, saying nothing about their expertise. iNaturalist data added 124 species to the consolidated list, roughly 9.3 percent of the total, demonstrating the platform’s value for early detection while underscoring the need for verification standards. A dedicated ornamental plant survey in Kruger contributed the largest single batch of new records, 323 species.

Perhaps the most sobering finding is that about one-third of the alien species recorded in the parks were absent from South Africa’s national plant checklist. When matched against the national backbone via GBIF, 447 of 1,419 species-level taxa, or 31.5 percent, could not be found. The proportion was similar for species recorded only in catchments surrounding the parks, which the team compiled into watchlists for potential future incursions. This gap means that national reporting systems, on which international biodiversity indicators increasingly depend, may be systematically undercounting alien species, with consequences for how progress toward global targets is measured.

The exercise also revealed how easily numbers can mislead. Some park lists grew because new sources were incorporated or because the scope widened to include 206 extralimital taxa, species native to parts of South Africa but introduced elsewhere in the country, where they can hybridise with close relatives. Other lists shrank because misidentified or synonymous records were removed, with 77 taxa dropped from the final compilation and 236 unique taxa flagged as misidentified or incompletely identified in particular parks. An apparent increase might suggest control failure, while a decrease might look like successful eradication, when in reality both reflect changes in data handling. The authors argue that reporting frameworks need explicit guidance on classifying invasion status, treating extralimital aliens and verifying non-expert records.

Technical wrinkles abounded. Automated fuzzy matching occasionally produced absurd results, matching the common name Banana to Banara, a genus absent from the region, or steering a misspelled Cinnamomum camphora to an unrelated synonym. Special characters in hybrid names defeated the matching algorithm. Standard queries against GBIF took hours to days, so the team, with help from an AI coding assistant, rewrote their R scripts for parallel processing and to avoid re-checking names, cutting run times to minutes. Even then, 281 records representing 54 taxa required manual resolution, a reminder that no checklist workflow can be fully automated.

The payoff is a dynamic system rather than a static document. All cleaned records, complete with source details, confidence annotations and GBIF taxon keys, will feed into the SANParks Biodiversity Information Management System, where taxonomy is regularly re-aligned against the global backbone. Because the structure is compatible with Darwin Core standards and platforms such as GBIF and the Global Register of Introduced and Invasive Species, locally curated lists can interoperate with national and international reporting. The authors argue the approach is scalable well beyond South Africa, offering protected areas everywhere a practical route to checklists that stay alive, support prioritisation and, ultimately, make the fight against invasive species measurable.

Subject of Research: Streamlining alien plant species checklist compilation and prioritisation in South African National Parks

Article Title: Streamlining alien species listing processes to enable prioritisation and reporting in protected areas: The case of alien plants in South African National Parks

Article References: van Wilgen, N. J., Wright, E. J., Baard, J., Ahrends, B., Cole, N., Cheney, C., Foxcroft, L. C., & Spear, D. (2026). Streamlining alien species listing processes to enable prioritisation and reporting in protected areas: The case of alien plants in South African National Parks. NeoBiota, 105, 177-208. https://doi.org/10.3897/neobiota.105.159101

Image Credits: AI Generated

DOI: 10.3897/neobiota.105.159101

Keywords: invasive species, alien plants, protected areas, South African National Parks, biodiversity data, taxonomic standardisation, GBIF, iNaturalist, citizen science, species checklists, biodiversity monitoring, NeoBiota