Entrainment of native fish through irrigation systems is an environmental impact of irrigation activities. Entrained fish are almost always permanently lost to the river system and fish can also suffer significant mortalities as they pass through offtakes. In this study, entrainment rates of fish through different irrigation intake systems were evaluated. Various riverine pumps on the Comet, Nogoa and Mackenzie rivers and the irrigation diversion channels originating from Fairbairn Dam were compared. Irrigation outlets were sampled during natural and allocated flow events with specialised nets to capture all fish entrained over a 100-minute period. Entrainment rates were calculated as fish per unit time and as fish per megalitre (ML) extracted. Larval nets with flow meters were also set to calculate the number of fish larvae entrained per ML.

Nearby river and impoundment reference sites were sampled during the same flow events. Catch rates from the reference sites were used as a covariate in generalised linear models to harmonise comparisons between irrigation systems. Reference site catch rates, when compared with entrainment rates also helped identify species and size classes of fish that were more or less susceptible to entrainment.

The results from the diversion channels originating from Fairbairn Dam suggest that gravity fed diversions entrain significantly more fish per ML and per unit time than pumped diversions. The water extraction rate was far less important than whether the channel was gravity fed or pump fed.

Several factors were considered when comparing riverine pumps, including pump rate (ML/day), intake location and depth (intake configuration), and flow type pumped (allocated flow, natural within-bank flow or overbank flow).  For most species there was a general trend for increasing entrainment rates as pump rates increased, although the fish entrained per ML increased at a lesser rate than fish entrained per unit time as pump rate increased. Pump intake position and depth significantly impacted entrainment rates, with shallow bankside intakes generally entraining far fewer fish than bankside deep, mid-river channel or side-channel pump intakes. Some very large pumps with shallow bankside intakes entrained far fewer fish than some smaller pumps nearby with different intake configurations. There was some variation between species and size classes on which intake locations and depth had the greatest impact, but for most species and size classes, entrainment through bankside shallow intakes was consistently low.

Pumping from overbank flows (flows where the river covers the bench) entrained far fewer fish than pumping from both natural within-bank flows and allocated flows. However, it is highly unlikely that any irrigator pumps solely from overbank flows. There was no statistically significant difference between allocated flows and natural within bank flows in terms of total numbers of fish entrained. Allocated flows tended to entrain more small fish, whereas fish >100 mm length appeared to be more susceptible to entrainment on natural within-bank flows. The pelagic larvae of golden perch were only entrained on natural within bank flows. Pumping natural within bank flows probably has a marginally higher biological impact on fish than pumping from allocated flows. Further replication would help determine these differences more conclusively.

Based on individual pump licenses and operations it is possible to predict the likely severity of impact of a pump. This can be done by considering pumping rate, pump intake position and depth (intake configuration), flow type(s) pumped and annual licensed allocation (total volume licensed to pump of any flow type). By cross multiplying score metrics for these different categories, a score can be derived for different pumps in a river system. The scores can help prioritise pumps for mitigation actions such as screening. The highest scoring pumps will be those predicted to be in the greatest need of mitigation action. However, for mitigation actions, feasibility and cost, based on the site characteristics also need to be considered. Any group, agency or peak body wishing to invest in pump screening, in some cases may achieve better outcomes for fish per unit cost by screening several slightly lower ranked pumps, rather than expending a large amount of money on a single highly ranked pump that is logistically difficult and thus expensive to screen.

The following recommendations have been derived from this research project.

1. Gravity fed diversions should be considered a high priority for mitigation of impacts to fish. Further investigations into impacts of riverine gravity fed diversions are recommended.
2. Pumped diversions can be prioritised using a four-part scoring system that considers flow type being pumped, intake location and depth (intake configuration), pump rate and total volume pumped per annum. Consideration also needs to be given to feasibility of screening a site (including cost) as part of the prioritisation process.
3. Future pumped irrigation developments should consider factoring in screening at the design and construction phase when it will be cheaper to install screens, compared to retrofitting them later.
4. Further replication of sampling will provide more confidence in the metrics for flow type being pumped, intake location and depth, and pump rate.
5. Further research needs to be conducted into the cost benefits of screening to provide irrigators confidence that pump screening will not significantly impact on their financial position.