This project developed flexible and statistically robust systems to calibrate and improve sampling strategies where required for the detection of post-harvest grain storage pests, and for demonstrating the absence of EPPs that could impact on market access

Research outcomes:

This project has delivered the first review of bulk grain sampling methodologies and statistical techniques since the late 1970’s. The review illustrated that a need existed to develop a new sampling methodology that considered how insect species behaved in grain bulks. A new statistical approach was then developed and was shown to outperform existing approaches (detect insects) by up to 400 percent. The new approach performed well in all conditions however its benefits were most evident where insects where highly restricted to certain portion of the grain mass. This is an important finding since these types of infestation are common in storages, particularly where infestations are a result of localised factors within storages.

Our research has illustrated that, unlike current sampling methodologies based on grain bulk size, sampling programmes are more efficient if based on a fixed numbers of samples. Furthermore, we illustrated that increasing sample number was more important than increasing sample size, particularly where infestation contained in small areas of the grain bulk.

This project has developed a statistical methodology, which for the first time allows grain producers and bulk grain handlers to determine at some level certainty how effective their sampling programmes are. Grain producers can use this approach to determine the optimal sample number to maximise detection of pests. Extensions to the approach also allow sampling programmes to be developed for Integrated pest management programmes. This technique, when utilised, will minimise bulk rejections ensuring grain going to storage/port meets the desired level of pest freedom.

Research implications:

Sampling methods in Australia typically are not based on a sound statistical or biological basis.

Intake and outturn from bulk handling facilities are the elements of the supply chain most ‘at risk’. Sampling methodologies currently vary significantly among regions and may not provide the level of detection that is required.

The new sampling approach as tested against existing statistical detection approaches was shown to provide the highest detection rate of all models examined. This method was also shown to outperform current statistical methods used in Australia.

The new sampling method provides a simple method to determine the number of samples required to maximise detection in grain bulks. The model was designed so that it can be adapted to account for varying conditions due to seasonal and geographical variation given appropriate data.

The model has been extended such that it can be used for IPM where detection at a zero tolerance threshold is not required. Insect detectability has also been incorporated into the model to improve detection estimates.

Acknowledgements:

The project is grateful for the cooperation of Cooperative Bulk Handling (CBH), Viterra and Graincorp. We would also like to thank the Peterson family (Killarney, Qld) for use of on farm storages.