SAMPLING PLANS FOR GREYBACK CANEGRUB DERMOLEPIDA ALBOHIRTUMTO AID MANAGEMENT DECISIONS AT FARM AND DISTRICT LEVEL

NUMBERS of greyback canegrubs were monitored in canefields in central and northern Queensland from 2003–2010. Canegrubs were counted annually from March–May under 20 cane stools in each field. Two different sampling schemes were used in different fields, with stools dug in each of four transects the length of each field, five per transect, or from the four corners and the centre of each field, four per position. Relationships between the mean and variance of grub counts did not differ significantly between the two schemes when calculated over a similar range of means. Mean grub counts did not differ significantly between the inner and outer transects or the central and corner positions in either scheme. Taylor’s power law was used to describe the mean-variance relationships for individual fields and for whole districts, and these relationships were used to derive optimal sampling plans. To estimate grub densities in individual fields with a standard error equal to 0.25 of the mean, a sample of 20 stools per field will only be adequate at high grub densities of more than 2 per stool. However, to estimate district-wide grub densities with the same precision, it will usually be better to sample fewer than 20 stools per field and sample more fields. For a given grub density, the optimal number of fields and samples per field depend on the time it takes to sample stools within fields and the cost of travel between fields. The latter will vary greatly depending on the layout of the district and the way a survey is structured— whether a grub survey is a stand-alone activity or done in conjunction with other activities, and the size of the sampling team. More sampling is needed to achieve higher precision of density estimates.