LWRRDC Milestone Report 2. Development and implementation of QA/QC protocols for sampling and sorting components of the MRHI agency bioassessment program
Internal Report 247
Humphrey C, Storey A & Thurtell L
Supervising Scientist Division
About the report
A comprehensive report on quality assurance/ quality control in rapid bioassessment projects with preliminary guidelines for implementation in the Australian MRHI' has been completed and submitted as an attachment. Further revisions of this document will be provided to LWRRDC in ensuing Milestone reports. Preliminary results on an 'Assessment of the efficiency of four types of device for subsampling of aquatic macro invertebrate samples' has also been submitted as an attachment to this report. This subsampling R&D is progressing satisfactorily.
Of most potential concern to the MRHI is the apparent poor representativeness of invertebrate community structure arising in agency live-sorted samples. This is the preliminary assessment of work arising from external QA/QC of agency sorting procedures (Attachment 4). It is not possible at this stage to assess the significance of these results as they might affect the accuracy of developing agency models until associated R&D to refine QA/QC acceptance criteria has been undertaken (described in Attachment 3). This R&D includes (i) classification and model outcome (o/e) results after error simulations performed on existing agency data sets, and (ii) classification and o/e results following incorporation of ‘corrected’ agency data that is derived from external QA/QC.
Most members of the MRHI Technical Advisory Committee have been made aware of preliminary results arising from external QA/QC and the R&D required to assess the significance of the results. It has been generally agreed by TAC members that the R&D approach that would best and most definitively address this issue is that arising from (ii) above, i.e. assessing the effect upon existing agency classifications and model outcomes when 'corrected' data (derived from external QA/QC) are seeded in agency data sets and modelling conducted using these altered data sets. In order to expedite this R&D, it will be necessary to provide additional funds towards the conduct of external QA/QC; as reported above, existing staff resources are available only to process relatively few of the available agency residues and these resources must also cover QA/QC of preserved subsamples. For this evaluation, there is the possibility of processing live-sort residues from all 50 samples gathered by one WA sub-agency (Murdoch University) in round 1, as well as the additional residues that are available from other agencies. The advice of the NRHP coordinator will be sought on this matter and a number of scenarios provided to him on additional samples available, and funds required, for this work.