Sensitivity of mountain precipitation distribution to temperature: a modelling case study for the Southern Alps, New Zealand

Temperature affects precipitation through changing its magnitude and its distribution. Magnitude changes are generally attributed to the change in the atmosphere's moisture holding capacity. In mountain regions precipitation distribution changes are associated with the different fall trajectories and accretion efficiencies of snow and ice with respect to rain and the change in elevation of cloud formation. These effects do not all work in concert, preventing simple estimations of new precipitation distributions under different temperature scenarios. If precipitation distribution changes cross a hydrological divide, the catchment scale precipitation totals may be affected differently to that predicted by changes in atmospheric moisture alone. In New Zealand the electricity supply is heavily weighted to hydro-electric generation, with the majority of the hydro storage fed by lee-side mountain precipitation in the central Southern Alps. Large scale studies have indicated the hydro-electricity water supply is secure under a warmer climate through the increased water vapour holding capacity of the warmer atmosphere, but these studies are incapable of identifying subtle changes in precipitation distribution at the catchment level. As a case study to investigate possible precipitation distribution changes in the Southern Alps under different temperatures, a series of Weather and Research Forecasting (WRF) model runs was carried out for a single observed storm event. A control run was first carried out using reanalysis input data, and the modelled total precipitation within the major catchments was found. Scenario model temperature sensitivity runs were then executed with the reanalysis temperature incremented in degree steps from -3 to + 3 compared to the control run. The total catchment precipitation from each scenario was compared to the control totals. The selected storm event occurred in autumn and was in the 80th percentile of storm rainfall totals at one of the long-term rain gauge locations in the region. It was a reasonably common storm type with a moist flow of warm air crossing the mountains followed by a cold front. For the positive temperature scenario model runs, all catchments had an increase in total precipitation, though the amounts varied from catchment to catchment (e.g. for the plus 2 degree scenario the catchment precipitation increases ranged from 2 to 18 % with respect to the control run). For the negative temperature scenarios the results were mixed, particularly for negative three degrees when the windward catchments had a reduction in precipitation (< -10 %) whereas the leeward (hydro) catchments had a major (> 20 %) increase in precipitation. The results indicate that, with all else being the same, it is possible for the temperature controls on precipitation distribution to have a significant impact on catchment-scale storm totals. These changes can be enough to offset, and in some cases overcome, precipitation magnitude changes expected from increased atmospheric water vapour.