Australian River Assessment System: AusRivAS Physical Assessment Protocol

Melissa Parsons, Martin Thoms and Richard Norris
Cooperative Research Centre for Freshwater Ecology
University of Canberra
Monitoring River Health Initiative Technical Report Number 22
Environment Australia, 2002
ISSN 1447-1280
ISBN 0 642 54888 9

3 Data Collection

• 3.1 Sampling site dimensions
• 3.2 Overview of the variables included in the protocol
• 3.3 Control and response variables
• 3.4 Field data collection
  • 3.4.1 General overview
  • 3.4.2 Instructions for the measurement of field variables
  • 3.4.3 Sampling times
• 3.5 Office data collection
  • 3.5.1 Instructions for the measurement of office variables

The sampling design for the physical assessment protocol consists of two aspects. First, reference sites are stratified across the landscape according to broad climatic regions and geomorphological zones (see Part 2). Then, physical, chemical and habitat information is collected locally from each reference site, and in future, each test site. Any site at which data are collected is called a sampling site, and will be referred to by this name throughout this document.

3.1 Sampling site dimensions

The length of a sampling site is a function of stream size (Table 3.1), and is defined as 10 times the channel bankfull width. Upon arrival at each sampling site, bankfull width of the channel should be measured or estimated (see Part 5) and the length of the sampling site calculated. Use a tape measure to quantify the sampling site length, until distances can be estimated accurately by eye.

To facilitate ease of movement along the length of the sampling site, the protocol has been designed in a manner that minimises the transportation of heavy or cumbersome sampling equipment over long distances (see cross-section variables section in Part 5). More information about field sampling is provided in Section 3.4.1 and a list of recommended field sampling equipment is provided in Appendix 2.

3.2 Overview of the variables included in the PHYSICAL ASSESSMENT protocol

Variables for inclusion in the protocol were selected using a three-step process. Firstly, a comprehensive list of the physical and chemical variables collected in the Index of Stream Condition (Ladson and White, 1999), the River Habitat Audit Procedure (Anderson, 1993a), the River Habitat Survey (Raven et al., 1998), AusRivAS, River Styles (Brierley et al., 1996) and Habitat Predictive Modelling (Davies et al., 2000) was drawn up. The variables suggested at the Habitat Assessment Workshop (see Section 1.2.2) were also included. Then, each variable was examined in light of what it indicates about river condition, or how it relates to geomorphological process. Lastly, the list was trimmed of duplicated, highly variable, hard to measure and redundant variables, to form a final set for inclusion in the protocol.

Over 90 field and office based variables are included in the protocol (Table 3.2). The variables are divided into control and response types (see Section 3.3) and are grouped according to broad categories (Table 3.2). These broad categories represent the main physical components of river systems, and also incorporate factors that are important for ecological function. Site observations include factors that are collected in AusRivAS to indicate the general condition of a sampling site.

Additionally, there is a small amount of repetition in the choice of some variables. The repetition has been deliberately incorporated into the protocol and is analogous to the social survey practice of asking the same question in several differently worded versions. Repetition of some variables will ensure that a large set of high quality data, that covers all the important physical components, is available to construct the predictive models (see Section 3.4.1).

3.3 Control and response variables

The variables included in the protocol are divided into control and response types and have very different functions in the construction of a predictive model.

Control variables – are large-scale environmental factors that control the expression of local-scale habitat features. Control variables are used as predictor variables in a predictive model and are analogous to the physical, chemical and habitat information collected in AusRivAS (see Section 1.3.2). Control variables are generally measured in the office (see Table 3.2 for exceptions). Also, control variables are usually large scale variables that are measured within the catchment area upstream of a site, or within a stream segment that is 1000 times the bankfull channel width. Exceptions are alkalinity, valley shape, local landuse, latitude and longitude, which are measured locally at the sampling site (Table 3.2).

Response variables – are local-scale environmental features. Response variables are used to form groups with similar physical features and are analogous to the macroinvertebrate information collected in AusRivAS (see Section 1.3.2). Response variables are all collected in the field and thus, are measured on a local scale. The exception is planform channel pattern, which should be verified using maps and aerial photographs.

3.4 Field Data Collection

3.4.1 General overview

Field data collection occurs in a similar manner as AusRivAS. Upon arrival at a sampling site, determine the bankfull channel width and calculate the length of the sampling site. Locate the sampling site so as to be 'representative' of the major bedform types present in the area. Then, follow the instructions given in Part 5 for the measurement of each variable. At larger sites, sampling may need to be conducted and recorded in sections, then combined. If this occurs, combination of data from different sections should be done while still at the sampling site, and overall observations of the site are still fresh in the memory!

Sampling should only be conducted under baseflow or low flow conditions. It is important not to sample under high flow conditions, because visibility of channel features will be reduced and the watermark will be obscured at cross-sections. In addition, health and safety issues should be considered at all times, but are of particular concern under high flow conditions.

Variables measured in the field have been selected to maximise information about stream character, but are also designed to minimise the amount of sampling equipment required (see Appendix 2). This facilitates ease of movement along the entire length of the sampling site and it is vitally important that the whole length of the sampling site is included in the assessment. Many local variables are assessed over the area of the sampling site (see Part 5) and thus, it is important to observe the overall status of each of these variables within the entire sampling site. This will involve walking greater distances than is generally encountered with AusRivAS sampling.

IMPORTANT

It is critical that all local scale variables are collected at every sampling site. In the physical assessment protocol, the physical, chemical and habitat variables are not used in the same way as in AusRivAS. The local scale variables are used to form groups of sites with similar features. Subsequently, the features present at a test site are compared against those present at a reference site and form the basis for derivation of O/E scores (see Section 1.3.2). Failure to measure a local physical, chemical or habitat variable at any reference site is analogous to losing taxa out of a macroinvertebrate kicknet sample collected for AusRivAS, and will ultimately detract from the robustness of physical predictive models

3.4.2 Instructions for the measurement of field variables

Standardised and detailed instructions on the measurement and interpretation of each field-based variable are given in Part 5. It is important that sampling teams familiarise themselves with these methods prior to the commencement of field work (see Appendix 1). This manual should also be available in the field for reference and cross checking if necessary.

The suggested sequence of work at a typical sampling site is given in Figure 3.1. This sequence of work can be adjusted to suit the needs of different sampling teams, although any sequence of work must ensure that all parts of the stream are observed and that all variables are measured. The sequence of work may also need to be adjusted for large rivers that require boat or canoe access.

Sampling sequence Take water quality and GPS readings at the starting point Walk to one end of the sampling site. Closely observe the channel along the way. Assess variables such as local impacts, local landuse, valley shape etc. Perform cross-section 1 in an appropriate bedform type. Travel back towards the starting point, closely observing the channel along the way. Stop at a point that is about mid-way along the sampling site. Assess additional variables from previous observations of the channel. Perform cross-section 2 in an appropriate bedform type. Travel the remainder of the way back to the starting point, again closely observing the channel along the way. Assess remaining variables. Perform cross-section 3 in an appropriate bedform type. Check that all variables have been recorded.

Figure 3.1 Suggested sequence of work at a wadeable sampling site with three cross-sections.

3.4.3 Sampling times

The physical assessment protocol is a rapid, semi-quantitative assessment method (see Section 1.3.1.4). When functional predictive models are fully implemented, this method will provide an assessment of physical stream condition that can be 'turned out' approximately 3-5 days after test site sampling. This turn out rate can be achieved because the majority of data collection occurs in the field. Laboratory processing of samples is not required, and is limited to the collection of office based predictor variables.

Further, the rapid aspect of the method is also applicable to field data collection, where sampling times have been substantially reduced in comparison to traditional geomorphological survey techniques. The approximate time required at different types of sampling sites is given in Table 3.3. However, sampling times may vary considerably depending on factors such as experience of the sampling team, site access, flow and weather conditions, ease of movement along the river, depth of the river, substrate type and periphyton cover, location of cross-sections and number of cross-sections. Thus, these times should be used as a guide only.

3.5 Office data collection

3.5.1 Instructions for the measurement of office variables

Standardised and detailed instructions on the measurement and interpretation of each office-based variable are given in Part 5. Many of the office-based variables, such as landuse and catchment characteristics can be measured using a GIS, while others will need to be measured directly off topographic maps. While not as critical as the collection of local scale variables, it is important to make an effort to measure all of the large-scale variables (i.e. those generally collected in the office). These variables are used as predictor variables and as such, have been included to cover the range of hierarchical links that may exist between local-scale and large-scale factors.

It should also be noted that for each office-based variable measured within a catchment (see Part 5), the term catchment always refers to the catchment area upstream of a site. This definition of a catchment standardises on the premise that regardless of catchment size, it is the large scale physical and geomorphological processes that occur upstream of a site, rather than downstream of a site, that determine the local scale features that will be found there.