Do detection dogs respond differently to dried, frozen and live plant targets?

• Successful alternate method for training dogs on restricted invasive plant material. • Dried or frozen plant samples can be used as a training aid for dogs. • Dogs can detect untrained and live plant targets at high mean accuracies. • Dogs did not display a strong preference of the trained over the untrained target. Weeds can have detrimental impacts on agriculture and the environment, and effective detection of individual plants is a crucial component of weed management. An emerging detection tool is the use of dogs (Canis familiaris) trained to recognise a weed's scent. Scent-detection dogs are well established as a tool in other disciplines but deploying dogs within weed management poses novel challenges. It can be difficult (or even illegal) to source a large number of live target weeds for training and evaluation purposes, due to the risk that these targets will create further detrimental impacts. We investigate whether invasive weed samples could be processed into inert and therefore bio secure dried or frozen forms for use in training. There is currently limited understanding of whether training dogs on frozen or dried plant targets can transfer into successful detection of the live target. We trained four dogs to detect frozen plant samples and four dogs to detect dried plant samples and exposed all eight dogs to scent boards containing frozen, dried and/or live plant targets. All dogs detected the three target types with high precision (100 %) and sensitivity (>85 %). Only one dog showed a statistically significant preference for the frozen odour. Our findings support the use of frozen and dried plant matter as an effective training and evaluation tool for dogs, in cases where the live target cannot be safely used at a broad scale. [ABSTRACT FROM AUTHOR]