4. Assumptions and uncertainties
Several necessary assumptions have been made in using this approach to assess quantitatively the impacts on flooding of climate change. One of those is that the frequency of the events being studied will not change. That is, that while the intensity of these events will increase, the chances of the storms "turning up" has not changed. In fact, the expectation from climate change studies is that these significant storms are likely to be more frequent, particularly in the west. This means that, over the long term future, the actual number of inundation events may be under-estimated.
In changing the temperature of the air and sea, we have made an assumption that the relative humidity will not change, which implies that the air will take up extra moisture. In the absence of any evidence or compelling reason to the contrary, this assumption seems reasonable. The weather modelling approach appears to be realistic, given the future storms simulated by the model operate in much the way as today's storms.
Another significant assumption is that the storms chosen for modelling are representative of the storms likely to be encountered in the future. Global climate change models are not predicting a significant change in the typical types of storms. Thus the storms, if carefully chosen, are likely to be representative of those that will occur in the future.
The conversion of rainfall into flows assumes that the land use does not change. Given that the climate results in an increased temperature, two changes may occur:
- The natural vegetation may change over time. Precisely what these changes might be is not easily predicted. The consequences for floods may well be less or more interception of rainfall by vegetation, decreased or increased absorption of moisture by the ground, and changes to riverside vegetation resulting in less or more resistance to river flow during times of flooding.
- Changes in the way people use the land, such as more or less urbanisation, and/or changes in agricultural production. Increased land clearance for agriculture may well lead to increased sediment in the river and a gradual build up of the downstream river bed with its attendant increase in flood risk.
- 3. Higher average temperatures are likely to change evaporation rates and freezing levels, which in turn could affect soil permeability and the speed with which rainfall runs off. These factors can be modelled in principle, but usually would be very complex.
It is beyond the scope of this report to investigate the complex implications of changes in land use under climate change for flood risk. It requires combining the climate change with the implications for vegetation change, which in its turn depends on the soil types and terrain. Where people are involved, regional, national, and even international economics may influence social decisions on the type and extent of changes in land use by humans.
Once a flood reaches the flood plain, the flood inundation calculations assume that:
- Flooding will occur from over-topping of the riverbank and/or stop banks, e.g., slumping failure of the stop banks by undermining is not considered.
- The riverbed will scour in a predictable manner. This is a large source of error in inundation modelling because the degree and location of river scour is difficult to determine and can be greatly influenced by random factors such as a large tree falling into the river.
- The topography of the flood plain is up to date, and, in particular, that the crest heights of stop banks are accurately known.
The importance of these factors will vary from case to case, but need to be considered in assessing and mitigating flood risk.
In addition to the assumptions outlined above, there are a number of uncertainties associated with the models and calculations recommended in this report. The main uncertainties with relevance for the resulting change in flood risk include:
- Future climate change will depend on future greenhouse gas emissions. The scenarios used in this report assume that no global efforts are made to reduce greenhouse gas emissions. An effective implementation of global agreements such as the Kyoto Protocol and subsequent next steps could reduce the amount of global climate change and its consequences on flood risk.
- Climate models used to predict temperature changes for New Zealand have uncertainties, and there are differences in regional patterns between models. While there is relatively good agreement of the general temperature trend between models, it cannot be ruled out that future climate change could, for example, change ocean circulation patterns in such as way that the temperature change for New Zealand could be even higher, or lower, than the range indicated in the guidance material currently available.
- The climate models are currently unable to predict whether there will be a systematic change in weather patterns, and natural climate patterns such as the El Niño/La Niña pattern. A change in those patterns could result in a change of the frequency with which typical storm systems reach New Zealand.
