Detecting signals of climatic shifts and land use change from precipitation and river discharge variations: The Whanganui and Waikato catchments

The Whanganui and Waikato river catchments have somewhat different degrees of exposure to the westerly wind systems. It is of interest to determine whether the two regions have similar times of occurrence of any concurrent shifts in river discharge and rainfall, with particular reference to mean value changes. Concurrent rainfall and runoff shifts are indicative of climatic variation but catchment land use changes (which will influence only discharge change) have also been occurring in both catchments, particularly with respect to forest planting or forest clearance. This thesis gives a summary of both climatic and land use change effects within the two catchments. If it happens that both catchments have similar climatic change-points then the data can be combined to provide a more robust framework for future water right specifications in both regions. Also, any similar responses to land use change may enable some degree of anticipation as to how future land use changes might lead to similar discharge responses. Change-points in river and rainfall time series flows were determined by an objective approach to detect breaks of slope in cumulative mass plots. Using repeated least squares fitting of piecewise linear segments, time points of maximum difference are determined as measured by the minimum least-squares in 2-segment fitting. Randomisation of time ordering of the original data was then employed to check that changes in the cumulative plots were statistically significant. Many significant but minor shifts were detected but a number of the shifts shown evidently in the rainfall and runoff cumulative mass plots. A set of change-points due to land management impacts were identified as discharge changes in the absence of concurrent rainfall changes. Rainfall-runoff linear relationship changes associate with changes in discharge time series. Change-points in rainfall and runoff times were detected at 44 flow gauges and 59 rainfall sites. There is some indication of a degree of natural geographic grouping with spatial correlation of times of discharge change. The times of the detected changes tend to cluster, with similar times for the same sign of change toward either greater or lower values of rainfall and discharge. The alternation of positive and negative signs is interesting as it was found 1981 and 1998 were times of negative shifts, while 1988 and 1994/1995 were times of positive shifts. Almost over the whole Waikato and Whanganui region, the changes in rainfall and runoff appears to relate to El Nino and La Nino events, which is of practical interest for water right considerations. The driver of the shifts in rainfall pattern was found to be the changes in high rainfall events, which can change the rainfall-runoff linear relationship in some areas. The land use component of some of the shifts was evaluated also and found only in 10 of the 44 flow gauges. The type of the land-use can be categorized into three groups: hydropower diversion, flood control system and afforestation. Within the study catchments, the impact of hydropower diversion is more significant than the other two types. With regards to the Whanganui catchments, the operation of the Tongariro power scheme from 1973 decreased low-flow by 89% in the Wahkapapa River and around 42% and 26% of the flow in the Whanganui River at Pariaka and Te Maire respectively. In the Whanganui catchments, the impact of farmland and native forest on river discharge was compared and farmland in the Ongarue catchment reduced flow much more than the native forest in the upper Whanganui catchment. The relationship of the high flow (Q90) in the two catchments is quite close to the ratio relationship of the catchment area, however, the relation of the low flows in the two catchments is fairly different and exceeds much more than the 0.75 ratio relationship of the catchment area