Modelling the decline in Sporobolus anglicus detections toward functional eradication: a case study in the Marlborough Sounds, New Zealand

Abstract

SummarySporobolus anglicus (C.E. Hubb) P.M. Peterson & Saarela (synonym Spartina anglica) is a highly invasive coastal weed that forms dense monocultures in intertidal mudflats and estuaries, displacing native reedbeds and associated fauna. Introduced to Aotearoa New Zealand from Britain in the early 1900s to aid coastal land reclamation, it became a conservation concern by the 1960s, prompting control efforts in the South Island from the 1970s. This study presents a detailed S. anglicus eradication case study from the Marlborough Sounds, where detections are nearing zero. We model changes in detection probability over time in Te Hoiere / Pelorus Sound, a complex estuarine catchment. We aim to inform control efforts and assess the effectiveness of ongoing management, by evaluating the probability of non-detection as a proxy for functional eradication. Using data from the Department of Conservation (DOC), we analysed detections across 79 search blocks between 2013 and 2024. A generalised linear mixed model was used to generate predicted detections through to 2040, using a modified dataset with pseudo-zero values for probable absences. The data were modelled as presence/absence with a binomial distribution, to identify the first year with a <0.01 probability of a positive detection (upper 99% confidence limit). Results suggest that by 2032, the likelihood of further detections under current management practices is remote, and that functional eradication may have occurred. We interpret this decline in detection probability as indicative of management success. Model outputs can support decision-making as to when active surveillance might reasonably be ceased. To accelerate the tail-end of eradication efforts, we recommend intensifying search effort and widening delimitations within the catchment over the next five years, to ensure removal of any remaining individuals. We also propose the use of environmental DNA as a cost-effective backstop for after operational wind-down. Implications for ManagersO_LIOur modelling predicts that S. anglicus detections in Te Hoiere / Pelorus Sound should decline to levels consistent with functional eradication by 2032, and may render continued manual surveillance uneconomical after this date. C_LIO_LIFull, regular and repeated surveillance of suitable habitats is needed within a five-year intensive monitoring period at the tail-end phase of eradication, as detections approach zero and managers consider withdrawal. C_LIO_LIWhile statistical analyses support eradication management decisions, absolute certainty of absence is unattainable. Decisions must balance technical feasibility with practical risk tolerance. C_LIO_LIEnvironmental DNA could provide an effective post-withdrawal monitoring tool, to allay the risk of re-invasion. C_LI