Review of Quality Assurance and Quality Control in the SRA Macroinvertebrate Theme

"December 2007".

Project Number: Review of Quality Assurance and Quality Control in the SRA Macroinvertebrate and Fish Themes Phase I - M/BUS/114 XE199.

MDFRC item.

1 of 2 reports assoicated with project see (A Review of Quality Assurance and Quality Control Procedures in the Sustainable Rivers Audit Fish Theme).

A review of Quality Assurance/Quality Control (QA/QC) procedures for the macroinvertebrate theme of the Murray–Darling Basin Commission’s Sustainable Rivers Audit was undertaken by the Murray-Darling Freshwater Research Centre between March 2006 and July 2007. The review of macroinvertebrate procedures involved visiting the four different state survey teams in the field and in their laboratories to review procedures. The field visits involved visiting sites in each of the states to review differences between and within states. The operators from each agency were evaluated against a standard questionnaire. All states conducted their field sampling, laboratory identifications and data enumerations in a very professional manner, generally to a very high standard. The organisation of the field component, compliance with Occupational Health and Safety requirements, allocation of resources to the project, laboratory facilities and staffing for the project were excellent for all states. Most operators were familiar with, and followed the methods described in their state AUSRIVAS manuals. There appeared to be no blatant breaches of the procedures, as the state agencies genuinely tried to follow the SRA methods or a modified version of their own methods. However, these methods were often subjected to variations and interpretations by individuals. These should be addressed by introducing standard QA/QC procedures. The methods in the AUSRIVAS manuals vary between states and this should be addressed by having a single SRA manual. The development and use of a single SRA set of procedures was strongly supported by the field operators. There were five major areas where differences were identified between and within states agencies. [1] Live Pick and Laboratory Sorting It is important to mention the use of the two different methods, live picking and laboratory sorting, as the issue remains a major inconsistency between the states. However, the respective methods are documented in the respective state manuals and the states comply with these methods. This is an age old question, discussed in depth in many forums and will not be discussed further here. However, the question is important for the SRA and if consistency in picking / sorting is to be achieved, the matter should be addressed by the SRA. [2] Sample collection The methods undertaken in the field occasionally varied considerably between operators, within and between agencies. The variation was due to a difference in protocols, differences in interpretation of the protocols, or in some cases deviations from the protocols. The data collected from these methods is qualitative and the only way to maintain Quality Assurance over these data is to be consistent throughout all phases of the procedures undertaken. The problem areas were: (a) In determining the habitat to be sampled some operators sampled a section of each of the available microhabitats, e.g. snags, vegetation, clear bank strictly in proportion to the abundance of the habitat in the reach. In contrast, other operators concentrated on sampling habitats known to possess large abundances of animals. Both the sampling techniques fit within the prescribed methods, but potentially collect a very different number of taxa. To alleviate the problem it is recommended that the method be standardised. The recommended method is for collecting effort to be in proportion to the available habitat. (b) The agencies manuals stipulate a sweep should cover a distance of 10 metres. Checking of the operators in the field revealed an extremely large variation in the area sampled and a corresponding variation in the quantity of material collected. Initial checks showed that the areas sampled ranged from 7-12 m and the number of sweeps taken varied from 33 to 107. These large variations in the sweep netting method cause a high degree of inconsistency, which can readily be avoided by tightening the method. A logical solution is to set a standard number of sweeps that must be taken or a recommended time period to undertake sampling. (c) There are slight variations in the “Live Pick” methods used by each agency, which will cause some variation in the number of taxa picked. This is a minor problem, but to ensure consistency it is recommended that a standard protocol for live picking is introduced. This should specify a minimum time of 30 minutes and a maximum time of 60 minutes and a set number of animals to be collected. (d) Due to the following variations from the AUSRIVAS methods it is recommended that the SRA documentation should list (1) When a riffle is absent then an, second edge sample should be taken and (2) The minimum dimensions in which two edge samples can be performed. (e) To enhance quality assurance it is suggested that the specific dates for sampling periods for SA and Vic should be documented in either the state AUSRIVAS manuals or as part of SRA Protocols. (f) To alleviate variations in the collection of water samples it is recommended that a standard method should be drawn up, based on the NATA recommended method. The method should include more information on taking a water sample, where to collect the sample from and at what depth, and possible pit falls to avoid. It is recommended that the collection of water samples be included in the annual workshops. [3] Equipment To date each state agency has used readily available commercial equipment for their field gear and this has led to a range of net shapes and sizes, and a range of sorting trays. To standardize the equipment it is recommended that: (a) A standard (size and shape) net frame be introduced, which would comply with ISO 7828- 1985, International Standard. The ISO is reproduced in Appendix 1. The MDBC has agreed to pay for a standard net for all states that are using the small triangular net. (b) A range of tray sizes have been used for live picking, with 38-43 cm x 27-30 cm (inside dimensions) tray being the most practical and usable. For consistency it is recommended that a 38-43 cm x 27-30 cm (inside dimensions) tray should be used. [4] Taxonomy All laboratories had an extensive collection of taxonomic literature, which included the most current identification keys. However, laboratories had no standard list of the appropriate keys to use and the choice was left to individuals. This becomes an issue when several versions of a key exist or if a new key is published, resulting in different identifiers using different keys to identify the same animal. There is also much variation between laboratories as to which taxa should be included. Some agencies include the micro-crustaceans, where others don’t. Likewise many include semi aquatic bugs, other don’t. These discrepancies in the taxa recognised are another source of inconsistency. To standardize the keys and taxa it is recommended that: [1] A list is drawn up of the most appropriate keys to be used, which would be used by all agencies. [2] A standard list of the taxa to be identified is produced, with a defined taxonomic level to which each taxon is to be identified. Much of the discrepancy is from several agencies including Copepoda, Cladocera, Ostracoda and the others not including them. Other discrepancies emanate from some agencies combining species from two families as one group, others disregarding some families as they consider them semi-aquatic, and others including some of the primitive families. If the aim of the project is basin wide, then these groups should be included as they occur over the whole basin, not just the drier areas where they are more abundant. A list of the taxa involved is presented below. [a] Groups inconsistently included: Copepoda Cladocera Ostracoda [b] Families which are inconsistently combined as one group: Physidae and Planorbidae Corbiculidae and Sphaeriidae [c] Families which are inconsistently considered semi-aquatic: Chrysomelidae Curculionidae Brentidae Heteroceridae Staphylinidae Carabidae [d] Primitive families which are inconsistently included: Hydridae Clavidae Spongilldae It is recommended that the above taxa be identified, along with mussels and snails to family level, and that all are included in the models. [5] QA/QC The agencies have, to varying degrees, documented procedures for Quality Assurance. Quality Control is only conducted routinely on training and identifications by one laboratory, and at another only when requested. Both of these agencies document the cross checking of new operators and routinely checked a percentage of the samples processed. The quality of all laboratories maybe satisfactory, but without documented proof of checking, it cannot be assumed to have been undertaken. This also applies to field collecting. To address these QA /QC issues the following recommendations are proposed: [a] Establish a QA/QC procedure for checking of identifications and enumerations. This should include cross checking 5% of all samples by a senior operator and the long term implementation of an inter-laboratory check. [b] It is proposed to address some of the QA/QC issues relating to the field component by conducting a regular field workshop with operators from each state present. The workshop would involve a training component followed by a forum to discuss the relevant techniques with the aim of establishing a uniform approach. It is strongly recommended that this workshop be undertaken, as all operators interviewed in the field believed that it would be beneficial, and would help in eliminating many of the inconsistencies within and between the states. [c] Another major issue emanating from this project is that many diffe