Update on the teams insect and plant monitoring
Written by Stephen Horn – Project Manager
Invertebrates are an important focus of the monitoring effort as mice have been massively destructive to the unique invertebrate community on the Antipodes. Moths are trapped using a funnel trap which utilises fluorescent lighting activated at night as an attractant. Pitfall traps were set up to capture ground dwelling invertebrates. Pitfall traps were setup 10 m apart and baited with a sweet smelling attractant in a plastic cup, which the insects fall into and are preserved. A plastic lid is positioned a few centimetres above the trap to protect it from rain. Moth light traps and insect pitfall traps were set up in the coastal, mid and upper region of Antipodes Island as well as Bollons and Leeward Island for comparison with sites where mouse impacts are not present. As with bird monitoring, invertebrate sampling has been taking place before and after bait applications so that any subsequent changes in the communities can be documented.
Preliminary results based on pitfall trap sampling, show significant differences in beetle and isopods diversity and abundance between the islands free of mice and Antipodes Island. The results complement and concur with previous studies comparing the fauna of off-shore islands particularly mouse-free Bollons Island and Antipodes Island where the latter has had local extinctions in the invertebrate communities attributed to predation by mice. It is expected that the removal of mice will allow the Antipodes Island ecosystem to recover and invertebrate communities to thrive, resulting in an increase in the diversity, abundance and distribution of the invertebrate fauna.
Pests can change vegetation composition as well. Prior to the eradication a satellite image with 2m resolution was acquired from Geo-Eye Imagery. A supervised classification was carried out remotely using expert botanical knowledge to assign regions of interest to a vegetation class. The classification assigns every pixel to the class that is most like the trained region of interests. The classes that were mapped were Soils (a combination of slips, basalt, bare rock), Coastal Poa littorosa, Lichen, Blechnum, Polystichum, Carex ternaria bog (sparse but dominant Carex ternaria with lots of other herby species amongst them), Inland tussock histiopteris (Poa litorosa and lots of low herbaceous vegetation including Histiopteris fern), Carex ternaria valley swamp (very dense tall Carex ternaria with nothing much else), Mixed vegetation and Low herbs. This classification has been groundtruthed by picking areas on the map that were assigned to one of these classes, locating these areas and recording what the true vegetation class was. To compensate for spatial error every sample unit was divided into 4 cells (5x5m) and assigned a value as correct or incorrect. This vegetation map will serve as a baseline for vegetation changes in the coming years due to the mouse eradication.