• About this site|
• Site index|
• Contact us|
• Jobs

Changes in drought risk with climate change

Executive Summary

1. Introduction

2. Drought Risk under Current Climate

3. Drought Risk under Scenarios of Climate Change

4. Other Issues

5. References

6. Technical Appendix

4. Other Issues

4.1 Assumptions and Limitations

A number of assumptions and simplifications had to be made in the analysis presented in this report.

• Future changes in potential evapotranspiration (PET) have been estimated from historical regression relationships involving temperature, wind, and precipitation. The use of precipitation as a proxy for solar radiation (and possibly also for humidity influences) is not ideal, but necessitated by what global climate model data were available.
• In calculating future PET, we have assumed that carbon dioxide increase has no net effect: that is, increases in stomatal resistance (reducing transpiration) more or less compensate for increasing leaf area in terms of the effect on evapotranspiration. This assumption is consistent with the best information available to date, but because the CO₂ effect has a large potential impact on drought, we have carried out additional sensitivity tests (discussed in appendix section 6.11)
• The water balance calculation used to derive potential evapotranspiration deficit (PED) is straightforward, but also involves a number of assumptions about soil depth and pasture response to water stress. Different formulae have been used in the literature. Nevertheless, we have used what we consider to be the most appropriate assumptions for New Zealand pastoral regions. Section 6.4 in the technical appendix demonstrates that while the absolute PED level is quite sensitive to assumed soil depth, the change in PED and return periods is fairly robust.
• The future scenarios are driven by what changes the global models project for the New Zealand region, and these can vary considerably from model to model. This report has tried to cover a reasonable range of possibilities by considering two models with differing patterns of local rainfall and temperature change, and two scalings of global temperature change. As expected from its large rainfall decreases in eastern NZ, the Hadley model shows the more extreme drying. The CSIRO model also suggests a similar but weaker tendency in spite of small rainfall increases in the east over the annual cycle. The agreement in PED tendency between models is likely a result of the CSIRO model having a weak gradient (wetter in the west and drier in the east) in rainfall change over the critical summer months when PET is highest. This agreement in seasonality at a critical time of year could be fortuitous, so we cannot rule out the possibility that other models might project more benign conditions in the 2080s than this report suggests.
• The New Zealand climate change scenarios used in this report span the central portion but not the full range of IPCC projections of possible global temperature changes (1.4 to 5.8°C by 2100). Thus changes in drought risk which are smaller than those projected under our "low-medium" scenario are possible, particularly if substantial international action is taken to reduce greenhouse gas emissions. Similarly, changes greater than our "medium-high" scenario are also possible.
• Estimates of future PED are derived by applying offsets to the current climatology. Only changes in means of the underlying climate elements have been considered. The calculations still lead to changes in extremes (ie, droughts), but the results could be modified by future changes in daily and interannual variability, which we have not considered since climate models at present do not provide consistent projections for changes in variability.
• The projected changes in PED are relative to an historical baseline (1972-2003), a period probably somewhat drier in the east than for the 20th century overall because of natural decadal variations in the climate (appendix, section 6.7). Long-period natural variations will continue to influence drought risk from decade to decade, in addition to the changes expected from increased greenhouse gases.
• The PED calculations, and comments on drought frequency, are for unirrigated pasture. Irrigation can in principle offset increases in drought risk where sufficient water for irrigation is available. The parameters underlying the PET calculations are also specific to pasture. Change maps are shown for the whole country, and obviously a lot of it is not in pasture. However, we believe the changes we calculate would give at least qualitative guidance for land not currently in pasture, such as forests.
• This report focuses on drought risk, and does not explore possible implications of climate change for heavy rainfall and flooding. The report indicates that many parts of New Zealand are likely to become drier on average, but this is in terms of the moisture availability for pasture growth. It does not necessarily mean the frequency of very heavy rainfall events and floods will decrease. Other work (Wratt et al., 2003) suggests the frequency of very heavy rainfall events may in fact increase in many parts of New Zealand, even in those areas where the annual rainfall decreases on average.

4.2 Issues for Further Research

Further research could assist in either reducing the uncertainty of projections made in this study, or in assessing the implications of climate change more generally on New Zealand water resources.

• Only two model patterns of local climate change were considered in this study. Obviously, increasing the number of scenarios would give a better indication of the range of possible future changes. A number of international modelling institutions plan to make their latest simulations publicly available as part of the IPCC Fourth Assessment, including experiments specifically targeting low and high emissions pathways, as well as stabilization of CO₂-equivalent concentrations at various levels. However, we do not expect the uncertainty range for New Zealand drought estimates to be substantially reduced in the near future.
• High resolution regional modelling of New Zealand climate change, currently under development by NIWA, would be preferable to relying on statistical downscaling of rainfall and potential evapotranspiration. This may provide better insights to potential changes in weather patterns and local climate changes for New Zealand, but it would not reduce uncertainties due to differences in global temperature change and broad regional climate change patterns as projected by different global climate models.
• A better understanding of the CO₂ fertilisation effect on evapotranspiration is crucial to reducing uncertainty about future drought frequency. However, the contradictory effects of increasing stomatal resistance (which decreases water loss) versus increasing leaf area and plant cover (which increases overall evapotranspiration) have been known about for some time, and a resolution does not appear to be near. Some attention should perhaps be focussed on grass species with inherently more efficient water use.
• In order to evaluate other consequences for water resources, future projections of river flow and water availability for irrigation should be considered to complement the current study on soil moisture.
• Future changes in climate variability are also possible, both natural and anthropogenic. Studies of changes in daily temperature and rainfall variability would lead to a better understanding of future water variability.