By Daniel Collins
Kaitaia’s main source of drinking water is the Awanui River. At the township, the mean annual river flow is about 6000 litres per second, most of which comes in winter and spring. February and March are typically the driest months, with a mean annual low flow (MALF) of 557 litres per second (PDF, Table 10.12). But during the 2010 drought, the river’s flow dropped precipitously low, in March reaching the lowest on record of below 320 L/s. The town had a consent to take up to 5 million litres of water per day, but normally only when the river flow was above 460 L/s. With conditions so dire, water restrictions had long been in place, and an emergency directive to continue taking water was granted. But while this drought illustrates the possible year-to-year variations about the average, how might the average flow conditions change as climate changes?
This was a question I was posed last year, as part of a project to examine the potential impacts of climate change on Maori communities. The challenge was to answer the question quickly. This essentially ruled out simulation modelling, so I turned to statistics.
I knew I had a long record of flow for the river, as well as of rainfall and temperature in the area. I also knew that projections of rainfall and temperature change had already been documented for the area on a seasonal basis [see MFE]. Lastly, as the area never experiences snowfall, it’s fair to say that the nature of the hydrological processes will not substantially change. Thus I refined the question to become: What were seasonal Awanui River flows like in the past during conditions that resemble climate change projections of the future?
Answering the new question meant that I had to construct a statistical model of mean seasonal flow as a function of mean seasonal temperature and rainfall. I kept the model as simple as possible, to avoid giving us a false sense of certainty, but not so simple as to violate established hydrological principles. I also made sure that the forecasts I was making, which were of average conditions, were within the historical range of variability, and so I was not stretching the model beyond breaking point.
In the end, the mid-range projections of the seasonal flows pointed to reductions throughout the year, but mostly in winter and spring (see the coloured bars in the figure below). The error bars for the two forecast periods indicate the range of uncertainty from the climate modelling. This uncertainty means that projections for the all-important summer flows ranged from a substantial decrease (-45% by 2090) to a moderate increase (+25% by 2090). As for winter and spring, reductions were projected across the board.
This has important implications for water resources planning for the future. With a reduction in mean flows during spring, conditions would be increasingly primed for a summer hydrological drought. And while mean summer flow could increase, the chances are higher that it will decrease, perhaps by a lot. This points to more low flows in the Awanui River, and probably more extreme low flows at that. How Kaitaia should adapt depends a lot on the capacity and aspirations of the community, but there are two good candidates. One is the ramping up and institutionalisation of water conservation within the town (e.g., dual flush toilets, less water-intensive gardens, household rainwater storage). The other is a greater reliance on reservoirs to store winter flow for the specific purpose of alleviating summer droughts. A risk with the second option, however, is that during times of plenty it could be used to grow demand, which would defeat the purpose of drought resilience.