Incorporating climate change estimates into flow estimation can reveal various issues pertinent to engineering design. Some of these issues are discussed here.
7.1 Appropriate use of historical records
Gradual shifts in climate and flood risk have important implications for engineering design. An essential element of a design flood study is the prediction of the future risk of extreme floods. As the climate changes, however, historical observations will be less and less indicative of future events. In other words, future flood statistics will diverge from the past’s. Statistical flood data analysis methods, and their applications, will need to change to reflect this (see Milly et al, 2008).
This is not to say that historical data is useless. It will remain invaluable as a means to calibrate hydrological models and to observe how flooding is indeed changing under climate change, as well as being useful in certain screening and advanced techniques discussed in this manual. Furthermore, because flood risk will change as climate changes, it will be necessary to identify the future time period for which a design is required and then determine the flood risk for that period.
For example, living with climate change means it is no longer meaningful to simply define a single value for the 50-year average recurrence interval (ARI) flood peak for a river. Instead, there will be several 50-year ARI flood peaks defined for time windows of practical interest (eg, for the ‘current’ climate, for a ‘short’ lifetime 2010–2040, and for a ‘long’ lifetime 2010–2090). Design studies will need to consider the lifetime of the structure being planned and select the appropriate flood peak that goes with that design lifetime. It is worth noting here that many assets still in use today have far outlived their design lifetimes.
For the quantitative estimation of flow, information is available for changes in seasonal rainfall, rainfall intensity, temperature, sea level and evapotranspiration. At the time of producing this manual, there are only qualitative assessments of the impacts of climate change on storminess, including impacts on storm surge and waves. This is an active area of research, and more quantitative information is expected to be released over the next few years.
7.2 Reporting the estimates
It is important when giving rainfall, flow and inundation estimates to comment clearly on what has been considered and what was beyond the scope of the project. This includes the scenarios chosen, the assumptions made, and the basis for the choices of parameters chosen in the modelling. This is particularly true when providing estimates of freeboard. The main impacts of climate change (eg, through rainfall changes) should be incorporated in the main flow estimation. Freeboard should then allow for some of the uncertainties in these climate change estimates, especially where they are likely to be greater (eg, with higher sea-level rise).
7.3 Dealing with uncertainty in estimates
There may be significant uncertainties in the estimates of rainfall, flow and inundation. These arise through uncertainties in rainfall inputs, parameter choices in modelling, errors in modelling, and assumptions about antecedent conditions. The errors from these uncertainties could be as large as the expected climate change impacts. Where possible, the error bounds of the calculations should be estimated. However, because climate change is likely to have a significant impact on flow, and much of that impact is able to be estimated, these broader uncertainties should not prevent efforts to include climate change in flow estimation. More confidence in the estimates can be given through replication. In other words, uncertainty can be reduced if the results from two screening methods and a more complex physical modelling method for the same catchment all point in the same direction.
7.4 Professional judgement
Professional judgement will often form an important part of the process of flow estimation. This judgement could be applied to the emission scenario choice, the choice of modelling parameters, the interpretation of past data, and in estimating confidence in the final results. Indeed, judgement may be most important when considering issues that have yet to be quantified, such as the effect of climate change on antecedent conditions, snow, aggradation, erosion and coastal outflow conditions.
7.5 Scenario choice
The estimates of rainfall, flow and inundation developed by the procedures outlined in this manual are likely to be used as the primary input into the risk assessment of future flooding. To help in this risk assessment process it will be necessary to choose a number of climate change scenarios to span the future possibilities. The Climate Change Effects manual suggests choosing a mid-low and a mid-high scenario. Given the current global emissions, and the likely emission paths, it may be more appropriate to choose a ‘middle of the pack’ emissions scenario such as A1B to represent a mid-low estimate, and a higher scenario such as A1FI to represent a mid-high estimate. These scenarios correspond to temperature changes for New Zealand of around 2ºC and 3ºC by 2100 respectively.
To help in your risk assessment you need to choose a number of climate change scenarios to span the possible future possibilities. For example, you might examine the consequences of a base level of temperature rise of 2ºC by 2100 but also consider the consequences of at least 3ºC rise.
Significant climate change research is being undertaken at the time of writing, targeted at providing information that will aid flow estimation and engineering design. This includes detailed information on changes to:
- extremes in temperature, wind and rainfall
- offshore waves and storm surge
- storm paths and intensity
- snowfall and accumulation.
Much of this research is due to provide the first results within the next few years (2010–2013). Planners and engineers will need to be alert to the arrival of this information and the implications it has for their work. It is also clear that the science will continue to provide new information. Decisions will need to be made now, on the best information available, but where possible these decisions should not lock in options that minimise the ability to adapt at a later date.