In recent years summer flows in the lower Selwyn River/Waikirikiri have dropped to very low levels and this has led to questions being raised about what is causing these low flows. When the Selwyn River at Coes Ford dried in February 2017, this was the first time that the lower Selwyn had not maintained permanent flow at this location since records began at the site in 1984. This dry period generated wide public interest in the river flows and the impacts of both climate and abstraction.

As directly measuring the independent effects of climate and abstraction on flow was not possible in the Selwyn Catchment, a desktop study using existing data was undertaken. This study utilised conceptual understandings, analytical methods, and numeric modelling to better understand the drivers of low flows. As summer flows in the lower Selwyn River are spring- fed, the interactions between surface water and groundwater were a key component which needed to be captured in this study. To do this, a MODFLOW groundwater model was developed to simulate changes in groundwater levels and surface water flows. While model development and testing were the main area of focus, this was complemented by trend analysis carried out for rainfall and flow measurement sites within the study area.

The GMS interface was used to implement the MODFLOW code, using a five-layer model and a uniform 1 km grid. Model parameters were adjusted to calibrate observed steady state (average) groundwater levels and lowland stream flows. Once calibrated, the model was used to test changes in recharge and abstraction. This indicated that both could influence groundwater levels, river flows and the extent of the dry reaches in the Selwyn and Irwell Rivers. By converting from a steady state to transient model, the daily changes in longitudinal flows were able to be simulated.

Trend analysis was caried out using the Mann-Kendall test, which indicated that flows in the summer months (December- April) in the lower Selwyn River are showing significant declining

trends. These declines are not observed in the upper Selwyn River, the other surrounding rivers, or in rainfall within the study area. While the trend analysis alone does not fully explain the causes of the declining flows, it does rule out rainfall within the catchment and inflows from the hills as being the drivers of the decreasing flows.

The converging lines of evidence from the trend analysis and numeric modelling suggest that the changes in recharge drive the year-to-year variability but that the effects of abstraction are the likely cause of the longer-term declining trends. This means that the year-to-year variability is overlaid on a long-term declining trend. In more recent dry seasons, the flows in the lower Selwyn River are lower than the flows which would have occurred in historic dry seasons with lower levels of abstraction. Scenario testing also indicated that the decline in flows and extent of dry reaches could worsen if abstractors used a greater portion of their authorised volumes.