Flood modelling and predicting the effects of land use change on the flood hydrology of mountainous catchments in New Zealand using TopNet
Thesis DisciplineCivil Engineering
Degree GrantorUniversity of Canterbury
Degree NameMaster of Engineering
The management of New Zealand’s freshwater resources has come under increasing pressure from different industrial and environmental stakeholders. Land use change and the pressure it can put on water resources has been a significant issue regarding resource management in New Zealand. A significant mechanism driving land use change has been the growth of forestry, dairy farming, and other agricultural industries. Improvements in agricultural and forestry science and irrigation techniques have allowed new, previously less arable areas of New Zealand to be subject to land use change, such as the conversion of tussock grassland to pasture in steep, mountainous regions in the South Island. Studies regarding the effects of land use change in such catchments, especially with focus on flood hydrology, appear to be limited, despite the importance of managing catchment headwaters to minimise flood risk downstream.
The TopNet model was used in this research project to evaluate the potential effects of land use change on flood hydrology in mountain catchments. It is a semi-distributed continuous rainfall-runoff model developed by the National Institute of Water and Atmospheric Research (NIWA). It has been widely used in New Zealand, and applications have included modelling water yield and the effect of climate change in catchment networks. However, it was not developed specifically for predicting flood flows. Hence, testing the model for flood peak prediction in mountainous catchments was also performed, and may show that TopNet can be a useful tool in resource management in New Zealand.
The Ahuriri and Pelorus River catchments were used in this investigation. Both are steep catchments located in the South Island. The Ahuriri River catchment, in the Waitaki Basin on the eastern side of the Southern Alps, is a semi-arid catchment dominated by tussock grassland. The surrounding catchments are heavily influenced by infrastructure for hydroelectric power (HEP) generation and more recently irrigation for dairy farming. The Pelorus River catchment is located at the northern end of the South Island. It is primarily covered in native forest, but adjacent catchments are subject to agricultural and forestry development.
The ability of the TopNet model for each catchment to predict flood flows were tested using a selection of historical flood events. Rainfall input to the model was at a daily timestep from the virtual climate station network (VCSN), and the method of disaggregating the daily estimate into an hourly rainfall series to be used by the model was found to have a significant influence on flood prediction. Where an accurate historical rainfall record was provided from a rainfall gauge station within the catchments, the disaggregation of the daily rainfall estimate based on the station data produced a significantly more accurate flood prediction when compared to predictions made using a stochastic disaggregation of the daily rainfall estimate. The TopNet models were modified to reflect land use change scenarios: the conversion of tussock grassland to pasture and the afforestation of tussock in the Ahuriri River catchment, and the conversion of forested land to pasture and the harvest of plantation forestry in the Pelorus River catchment. Following a past study into modelling the effects of land use change using TopNet, three key model parameters were modified to reflect each land use scenario: saturated hydraulic conductivity KS, canopy storage capacity, and the canopy enhancement factor. Past studies suggested a wide range of suitable values for KS, although also acknowledged that KS depends heavily on the specific catchment characteristics. A sensitivity analysis showed that KS had a significant influence on flood peak prediction in TopNet. It is recommended that further investigation be conducted into suitable values for KS.
TopNet appeared to predict the effect of land use change on flood magnitude in mountainous catchments conservatively. Past studies of land use change suggested that the effect on flood flows should be significant, whereas TopNet generally predicted small changes in flood peaks for the scenarios in each catchment. However, this may suggest that the topography, geology, and soil properties of steep catchments are more important to flood hydrology than land cover. Further investigation into the effect of such catchment characteristics is recommended. Nevertheless, TopNet was shown to have the potential to be a useful tool for evaluating and managing the effects of land use change on the flood hydrology of mountainous catchments in New Zealand.