Dispersal potential rather than risk assessment scores predict the spread rate of non‐native pines across New Zealand

Abstract

Listen

Translate article

Abstract Pine (genus Pinus) species are planted extensively for forestry purposes in areas where they are non‐native, with the result that biological invasions by many of these species are of considerable concern in many regions. Owing to the economic importance of these species, management approaches must focus on reducing the risk that they will spread from plantings, which requires access to accurate and rigorous risk analysis protocols. However, while a variety of tools are available that accurately predict which species are likely to naturalise, their performance in forecasting the subsequent rates of spread of species once established remain unproven. We compared 10 measures of risk that could be used to forecast the likelihood of pine spread, including established risk assessment tools as well as dispersal‐related plant traits: expert assessment of spreading vigour, the Z‐score, the Australian weed risk assessment protocol (WRA), seed terminal velocity, tree height, modelled dispersal potential, and planting effort. We examined the extent to which these measures were consistent in their estimate of risk and assessed their ability to accurately predict the spread rates of 10 non‐native naturalised pine species within New Zealand. Of the measures of risk assessed here, only the models of dispersal potential were able to significantly predict the species' spread rates within New Zealand. In addition, seed terminal velocity was significantly correlated with expert assessment of the species spreading vigour, indicating these two measures capture similar aspects of risk. Modelled long‐distance dispersal from forests had the greatest explanatory power for the species spread rates in New Zealand. Synthesis and applications. Common weed risk assessment tools and expert knowledge were unable to predict the spread rates of non‐native naturalised pine species in New Zealand, and should not be used to provide an index of spread risk. Instead, we recommend a move towards the use of dispersal models when assessing the spread risk of these species, even at national scales. Current practices relying on expert assessment are likely to underestimate the spread rate of species currently considered ‘low risk’, suggesting that these tools may be inadequate for predicting spread of these species.