Weed Risk Assessment for non-botanists Peter A.Williams Landcare Research Nelson, New Zealand [email protected].

Weed Risk Assessment for non-botanists

Peter A.Williams

Landcare ResearchNelson, New Zealand

[email protected]

Definitions

• Weed – a plant species controlled ( or potential to be so) because of its impact

• Impact – interference with human health, economy, or environment etc.

• Weed risk – probability x intensity of weed impact

• Invasiveness – spatial spread

Invasion is crossing barriers

• Human responses: love it or hate it

• Spatial: continents to farm-scale

• Biological: pollinators, dispersers etc.

• Environmental: temperature, nutrients etc.

Four Elements of Risk

• Probability - likelihood, chance

• Uncertainty - the unknowable

• Consequences - impact in space and time

• Manageability - resources, timing

Constraints to weed prediction

• Uniqueness of species x environment matches

• Non-reoccurring events

• Not measurable directly prior to invasion

• Continuous change through time

Science, sociology,economics

People Pathways

Weed-riskassessment

PlantsHabitats

< 1 spp.per year p.a.

16 spp.per year

WRA opportunities

Risk component

Pre entry

Post entry

Probability Low High

Uncertainty High Low

Consequences

Estimate Measurable

Manageability

Estimate Measurable

Elements of a Pre Border WRA

• Origin of the plant and climate match

• History of the species

• Pathways of introduction

• Establishment and spread likelihood

• Impacts features

WRA outcome thresholds and examples

Score range

Outcome

Example species

Actualscore

< 1 Accept Camelia -14

1-6 Evaluate Cineraria +1

1-6 Reject Gorse +22

Table 1. Accuracy rates for weeds and non-weeds using the WRA system from various studies outside Australia (with locally-adapted questions but without the decision tree of Daehler et al., 2004). Sensitivity is the percentage of known weed species which had a ‘reject’ outcome. Specificity is the percentage of known non-weed species which had an ‘accept’ outcome.

Study

Sensitivity(% true weeds

predicted)

Specificity(% true non-

weeds predicted)

Pheloung et al. 1999 69% 92%

Daehler and Carino 2000

90% 54%

Daehler et al. 2004 ?% 66%

Jefferson et al. 2004 100% 40%

Krivanek and Pysek 2006

73% 71%

Mean ?? 65%

Elements of a Post Border WRA

• Evaluate weed risksInvasiveness criteria

Impacts criteriaPotential distribution

• Feasibility of controlCurrent distributionCosts and duration

• Weed management prioritiesWeed risks versus feasibility of control

Mathematics and Post Border WRA

• Evaluate weed risksmodelling traits

modelling spread environment matching

• Feasibility of controloptimum search strategiesdetermining real costs

• Weed management prioritiescost benefit analysis