Development of a maize forage sampling system to accurately determine dry matter percentage

Field sampling, sub-sampling and laboratory methodology for maize forage DM % determination, from both trucks and finished stacks, were investigated in the 2001-02 and 2002-03 seasons. A number of sampling and sub-sampling techniques were tested. A target 95 percentile confidence interval of ±1.0 % from the true mean DM % appears to be a realistic target for maize forage assessment when samples are collected from completed stacks and sub-sampled in the field. For stack sampling, testing showed that to be 95 % confident of estimating the true DM % to within 1.0 % for 95 % of stacks sampled, seven individual auger-cores (samples) need to be taken equidistantly along the formed stack, with each sample submitted for DM %testing individually. To reduce the number of samples submitted to the laboratory, the use of a riffle-box proved the most accurate and unbiased sub-sampling technique. To get the same level of accuracy, 10 auger-core samples must be collected and combined through the riffle-box, from which 1-2 sub-samples then need laboratory testing. Determining forage DM % by sampling from trucks proved more difficult. A minimum of 52 trucks per paddock (source) need to have one composite sample (sample made up of at least four hand-scoop samples combined per truck) tested individually to be 95 % confident of estimating the true DM % to within 1.0 % for 95 % of paddocks sampled. Sub-sampling of truck samples was not conducted.