New Zealand Journal of Ecology (2020) 44(2): 3416

Does habitat manipulation enhance native woody seedling recruitment in a dryland river floodplain?

Research Article
Debra M Wotton 1,2*
Philip B. Grove 3
Dave Kelly 2
  1. Moa’s Ark Research, PO Box 11270, Wellington 6142, New Zealand
  2. Biological Sciences, University of Canterbury, Christchurch, New Zealand
  3. Environment Canterbury, Christchurch, New Zealand
*  Corresponding author

Recruitment failure of native plants is common in dryland ecosystems in New Zealand. We investigated whether herbicide control of invasive grasses or restoration of either shrub cover or the natural disturbance regime (river gravel deposition after flooding) can promote seedling establishment of the critically endangered shrub Olearia adenocarpa and two common species also found in river floodplain ecosystems (Carmichaelia australis and Sophora microphylla). Olearia adenocarpa seedlings have been observed only after invasive grasses were controlled with herbicide and when browsing mammals were also excluded. We used a field experiment to compare seedling establishment in four treatments: (1) shade cloth shelters, a proxy for native shrub cover; (2) river gravel addition, to simulate gravel deposition during flooding; (3) herbicide application, to kill invasive grasses; (4) untreated controls. For C. australis, provision of shade and shelter (hereafter referred to as shade) and herbicide application both promoted seedling emergence compared to controls, but few seedlings survived after two years. For O. adenocarpa, we found weak evidence that shade, herbicide and gravel increased seedling emergence, but the effects were small and few seeds germinated. Establishment of S. microphylla seedlings was unaffected by shade, but gravel and herbicide treatments reduced seedling emergence compared to controls. Seed germination was also low for S. microphylla. During a drought in the second spring, soil moisture was higher in all treatments than in untreated controls. Soil moisture was lower in gravel plots than in untreated controls at all other times. Shade also increased soil moisture in the first spring. The effects of herbicide on soil moisture were variable. Our findings indicate that the loss of native shrublands and invasion by non-native grasses limits the establishment of C. australis seedlings in dryland river floodplains, and that the effects of these changes are not easily overcome. While O. adenocarpa seedlings cannot compete with invasive grasses, other factors also limited recruitment of this species, and S. microphylla, in this study.