Assessment Protocol
Initial measurements of a weed infestation will be made before agents are released to provide baseline information.
An important component of all biocontrol projects is assessing how successful they have been.
Proof of impact is needed to back up anecdotal evidence that agents are doing a good job and provide justification for continued investment in biocontrol. It is equally important to also identify where agents are not up to scratch so additional species can be sought to strengthen the attack or alternative control methods developed. However, the cost of undertaking assessment studies has to date proven to be a major obstacle both in New Zealand and worldwide. When a project has clearly been highly successful it is unappealing to channel further resources into a former problem when so many others still require attention. Likewise if a project appears to have failed there is little incentive to spend precious resources documenting this in more detail, although there is a danger here that changes are actually happening but are too subtle to be easily noticed. “However, the National Biocontrol Collective (NBC), the major funder of the development and release of new weed biocontrol agents in New Zealand, has recently strengthened its commitment to ensuring that some assessment is undertaken,” confirmed Lynley Hayes. They asked Landcare Research to develop a suitable protocol outlining minimum standards plus further options where additional resources are available.
Scientists at Landcare Research have thought long and hard about how assessment could be tackled in a simple and cost-effective manner yet still yield useful data. “Our suggested approach is to get NBC field staff to collect simple data from many sites around the country but only when it is essential to do so,” explained Lynley. Members of the NBC have indicated that they are interested in obtaining data they can share with colleagues, managers, councillors and ratepayers, rather than the loftier goal of obtaining data that could convince scientists. Data that demonstrate a correlation will therefore generally suffice, rather than conclusive proof of cause and effect (which is much more difficult and expensive to obtain). Where possible, information that could be published in peer-reviewed journals will be collected. “Landcare Research will continue to use government funding to undertake more complex cause-and-effect population and ecosystem-level studies for a few flagship projects, and to assess the impact of agents that target reproductive structures only,” said Lynley. Because seed-, fruit- and flower-feeders do not impact on existing weed populations, but rather the future replacement of them, this type of assessment is more challenging and so has been excluded from the national protocol, at least for the time being.
The NBC met twice in 2014 to discuss the proposed protocol and suggest refinements. “There are five possible steps, and the number of steps an organisation undertakes depends on results achieved, resources available and the level of proof required,” explained Lynley. A lead organisation will be nominated to take overall responsibility for ensuring adequate follow-up occurs for each biocontrol agent. The lead organisation will generally be the one that acted as the applicant to release that agent, and will act as a project champion, involving other organisations as necessary. NBC members will collect data for the weeds that are considered significant in their region. The lead organisation will compile and share the information with those with an interest in it.
Step One: Agree Desired Outcomes & Collect Baseline Data
A key first step is to clarify what successful weed biocontrol would look like. The NBC agreed to a single desired end outcome: that where weeds are widespread, their harmful impacts will be reduced; and where weeds are less widespread (or absent), they will be prevented from becoming a problem. Intermediate outcomes to show if progress is being made towards the end outcome include demonstrating a reduction in weed abundance, density or vigour.
Due to the limitations of current data it will largely be necessary to start collecting baseline data for current weed targets from scratch rather than build on existing datasets. Nationwide, 10–20 good potential, or actual, release sites will be selected taking into account significant regional and national variation. Photos of the sites will be taken. Some sites/weeds will lend themselves to photos that can be analysed using digital software, but if not, ordinary before-and-after shots will be taken to provide a visual record. If the site is already a release site, the abundance of agents present, or their damage levels, will be assessed and recorded. The weed infestation at the site will be defined as either major (as far as eye can see), moderate (>100 m2), or minor (<100 m2). At the densest accessible point the percentage cover of the weed (and height for some species) will be estimated. These photos and measurements will be repeated every 2–3 years.
Step Two: Check for Establishment
NBC members will visit at least 75% of release sites at least once to check whether agents they have released have established. The release point will be checked for 5 minutes, following by wider searching of the site for 10 minutes. If agents fail to establish then no further monitoring will be undertaken. If establishment is uncertain, a watching brief will be kept for at least 10 years.
Step Three: Assess Agent Population Build-up/Damage Levels
If the agent has clearly established then its population or damage levels will be evaluated. Where it is easy to collect/count individual insects, population size will be estimated using broad categories (tens/hundreds/thousands). The method used will depend on the agent: e.g. beat bushes a set number of times, or estimate the number seen over a set time period, etc. Damage will be estimated for pathogens and also where insect population assessments will be difficult for cryptic species. The following categories will be used: occasional (signs of damage present but not common), patchy (signs of damage are present but are variable throughout the site, e.g. some plants may have no damage, and others may have heavy damage but this would be rare), heavy (the majority of plants are showing signs of damage and at least some plants are beginning to show signs of severe defoliation/damage or stress), and heavy (severe damage is obvious and widespread).
If agent population/damage levels are low, a watching brief will be kept for at least 10 years. If they remain low, more research may be required to understand why. If agent populations are high or damage appears significant then the impact on the weed population will be measured. Some members of the NBC may also check for non-target damage to support Landcare Research’s efforts to follow up on this for all species released, and study how widely dispersed agents are to inform redistribution efforts.
Step Four: Assess Impact on Weed Population
The impact of an agent on the weed population will be assessed by continuing to take photos and measure weed abundance and agent population/damage levels every 2–5 years. Some organisations may choose at this point to set up plots that can be manipulated experimentally to prove cause and effect. If no impact is measured then again further research may be needed to understand why. However, if an impact on the weed population is measured through either method then the ecosystem consequences of this can be evaluated and an economic analysis of the project undertaken.
Step Five: Assess Ecosystem Consequences & Undertake Economic Evaluation
Ecosystem consequences will be evaluated through continuing to take photos and measure weed abundance and agent population/damage levels every 5–10 years. Some organisations may again choose at this point to set up experimental plots from which more detailed measurements of change can be made, such as what species are replacing the weed. Organisations may also opt to undertake an economic evaluation of the costs and benefits provided by the project.
“Assessment activities based on the above protocol will get underway in 2015. Detailed instructions and forms have been prepared for the projects that are ready to begin now: broom (Cytisus scoparius), Japanese honeysuckle (Lonicera japonica), lantana (Lantana camara), privet (Ligustrum spp.), tradescantia (Tradescantia flumeninsis), and woolly nightshade (Solanum mauritianum). These are available from the Landcare Research website (see Changes to Pages). “In 2016 we will review how well the protocol is working and if necessary make some tweaks,” concluded Lynley.
The National Biocontrol Collective consists of 12 regional councils/unitary authorities nationwide plus the Department of Conservation.