‘Mega-masts’, climate change and the management of invasive mammals

Image – Rowan Buxton

It is well known that many New Zealand plants periodically have years with very high seed production (called masts) and that, for beech forests, populations of invasive rodents increase substantially in mast years. This, in turn, leads to a build-up of populations of stoats and increased predation on indigenous species by rodents and stoats. For example, research by DOC scientists has shown that outbreaks of ship rats result in severe impacts on orange-fronted parakeets and native bats, and that populations of yellowhead in the South Island are unlikely to persist through periodic irruptions of stoats.

Currently the management solution to mast-generated outbreaks of invasive species is pre-emptive control operations, usually with aerially-sown 1080 baits distributed over large areas of forest. Although using 1080 is an effective method of pest control, baiting large areas of forest is costly. For beech forest, planning begins as soon as there is heavy flowering by beech trees. Because flowering occurs a year in advance of seed production there is, in principle, ample time to check on the seed crop and the abundance of rodents. However, it is feasible to do this only at a few locations such as high-priority conservation areas with remnant populations of predation-sensitive indigenous species. Until recently there were no easy, affordable methods for estimating the spatial extent of a beech mast and therefore the appropriate size of pest control operations.

In 2013, Dave Kelly and colleagues, including several from Landcare Research, published a simple model for predicting masts using climate data. According to their ‘delta T’ (ΔT) model, the likelihood of masts by several species, including beech and tussocks, is positively correlated with the difference between average summer temperatures in successive years: a high positive value of ΔT (i.e. last summer warmer than the preceding summer) corresponds to a high likelihood of a mast in the coming year. In a follow-up paper, Pen Holland and colleagues have shown that ΔT can predict directly the probability of outbreaks of house mice after beech masts.

Roger Pech and his colleagues used the ΔT model to consider the following questions. (1) Are some areas in New Zealand more prone to masts? (2) Do masts always occur over very large areas? (3) How often have widespread ‘mega-masts’ (e.g. >50% of beech forest predicted to mast) happened in the past and (4) will they occur more frequently in the future? (5) How do mega-masts affect the cost of controlling invasive mammals?

Mandy Barron used temperature data from the National Institute of Water and Atmospheric Research’s (NIWA) Virtual Climate Station Network to calculate ΔT values for each 5 x 5 km grid cell across New Zealand for the last 40 years (Fig. 1).The model predicts that in some areas, for example on the west coast of the South Island, beech forest is likely to mast once every 3 years; this compares with an average of once every 5.4 years at the five sites with long-term datasets used by Dave Kelly and his colleagues. Previous modelling by Dan Tompkins suggested that very frequent beech masts ‘will lead to populations of stoats and ship rats becoming less irruptive and being maintained at appreciably higher abundances, while the ability of both current and in-development management approaches to suppress invasive mammals will be compromised.’

Mandy also calculated the percentage of the total area of beech forest predicted to have masted each year since 1974. Most masts are localised (e.g. since 1974 there have been 11 years when masts covered <10% of beech forest), but some were much more widespread (Fig. 2). Mega-masts were predicted on eight occasions: over the last four decades there has been a moving average of 5.2 mega-masts per 25 years. In cooperation with colleagues at NIWA, ΔT values to 2100 were predicted for four standard scenarios representing a range of possible future levels of greenhouse gas emissions: an optimistic scenario with peak CO₂ concentrations around 2025 followed by a substantial decline (labelled RCP 2.6); CO₂ increasing initially then declining after 2050 (RCP 4.5); CO₂ increasing until 2060 then declining slowly (RCP 6.0); and a pessimistic scenario with rapidly increasing greenhouse gas concentrations (RCP 8.5). Compared to the reasonably likely mid-range scenario RCP 4.5, the number of mega-masts with subsequent high-cost pest control is estimated to decrease for all other RCPs over the 21st century.

Mega-masts are particularly important for at least two reasons. Firstly, widespread outbreaks of rodents and their predators are likely to result in widespread impacts on indigenous species. This has potential to severely reduce connectivity between remnant populations of endangered species. Secondly, the cost of pest control following a mega-mast can be prohibitive. In its ‘Battle for our Birds’ programme, DOC has implemented a massive boost in pest control in response to the 2014 mega-mast (Fig. 1). Based on data for the period 2003 to 2013, it would cost DOC approximately $65 million to conduct aerial 1080 baiting over the entire 3,632,500 ha of beech forest predicted by the ΔT model to mast in 2014. Clearly, episodic pest control on this scale would require major financial and logistical planning.

This work was funded by the Ministry of Business, Innovation and Employment (C09X0909).

Roger Pech, Mandy Barron, Andrea Byrom, Dan Tompkins (Landcare Research)
Jenny Christie (DOC)
Andrew Tait (NIWA)