News & Views | Published:

Climate change

Human influence on rainfall

Nature volume 470, pages 344345 (17 February 2011) | Download Citation

Rising concentrations of anthropogenic greenhouse gases in the atmosphere may already be influencing the intensity of rainfall and increasing the risk of substantial damage from the associated flooding. See Letters p.378 & p.382

The varying distribution of fresh water across the globe, involving complex patterns of rainfall in space and time, crucially affects the ecosystems and infrastructure on which human societies depend. The recent severe floods in Australia, Sri Lanka and Brazil, which were partly associated with an episodic cooling in the equatorial Pacific Ocean (La Niña), highlight the effect of natural fluctuations in atmospheric circulation systems on rainfall distributions. However, global warming resulting from anthropogenic emissions of greenhouse gases may have compounded the effects of such fluctuations, a possibility that is considered in two papers in this issue1,2.

Min et al. (page 378)1 provide evidence that human-induced increases in greenhouse-gas concentrations led to the intensification of heavy precipitation events over large swathes of land in the Northern Hemisphere during the latter half of the twentieth century. They combined a rigorous 'detection and attribution' framework with extreme-value theory (a statistical technique designed for analysing rare events) to place daily rain-gauge data and climate-model simulations on a common scale. A tentative but intriguing finding by these authors is that climate models may underestimate the effects of anthropogenic global warming on rainfall intensification, a possibility that has implications for projections of future climate.

Climate models typically do not have sufficient resolution to satisfactorily represent processes at the level of cloud formation, and there is considerable variation in simulations of the relationship between rainfall extremes and warming in the tropics3,4. Pall et al. (page 382)2 get around this issue by considering an event relating to a large-scale weather pattern that can be represented by high-resolution versions of such models. They link human influence on global-warming patterns with an increased risk of severe flooding that is associated with a displacement in the North Atlantic jet stream (a fast, eastward-moving ribbon of air at around 8–12 kilometres altitude). The consequence of jet-stream displacement was recurrent, unusually prolonged periods of heavy rain over England and Wales in autumn (September–November) 2000 (Fig. 1).

Figure 1: The floods of autumn 2000.
Figure 1

Geese take to the inundated streets of York, northern England. Image: A. DENNIS/AFP/GETTY

The authors' approach involved conducting multiple climate simulations of this event, with realistic control runs and further scenarios in which the anthropogenic warming patterns were artificially removed. Building up convincing statistics required thousands of model simulations, which were made possible through the use of idle time on personal computers volunteered by members of the public.

Feeding the precipitation simulations into river-flow models allowed Pall et al. to gauge the influence of anthropogenic greenhouse-gas emissions on the risk of damaging floods occurring in England and Wales during autumn 2000. They conclude that flooding risk was indeed “substantially increased” by these emissions, although the exact scale of the effect is difficult to estimate. However, their results vary with the range of anthropogenic warming patterns estimated separately from four coupled ocean–atmosphere climate models. The set of different models used to simulate the relationships between temperature, precipitation and flooding also requires further evaluation. Nevertheless, applying their technique to further severe flooding events may prove valuable for informing policies aimed at improving adaptation to climate change.

More generally, in considering changes in the pattern and intensity of precipitation, the underlying physical and geographical context provides essential background. The root cause of the reported changes in rainfall with increasing atmospheric temperature is centred on atmospheric water vapour5. The convergence of moisture-laden air masses leads to air uplift, cloud formation and eventual precipitation (rain, snow or hail). A warming atmosphere can carry greater quantities of gaseous water (approximately 6–7% more is carried per °C of warming near Earth's surface, as determined by the Clausius–Clapeyron equation). It has long been thought that the intensity of the heaviest rainfall is modulated through this simple relationship, again increasing at around 6–7% per °C. However, water vapour also controls precipitation by influencing the radiative cooling of the atmosphere (through the absorption and emission of infrared radiation). The temperature dependence of radiatively induced increases in total global precipitation (around 2–3% per °C of warming5,6) is substantially lower than the 6–7% per °C increase in intense rainfall implied from changes in atmospheric moisture.

Intense rainfall is inherently local, but is fuelled by a supply of atmospheric moisture from farther afield that may otherwise have contributed to more moderate rainfall elsewhere. So rapid increases in precipitation intensity in one region imply a decrease in intensity, duration and/or frequency in other regions7. Many climate models show this projection through an increasing intensity of rainfall in wet regions (above the 2–3% per °C rate of increase in global precipitation due to radiative cooling), together with a tendency towards diminished rainfall in the already dry subtropics5,6. Because of the implications for future flooding and drought, it is vital to establish the physical basis for these changes5, and to verify theory with further observational evidence1,3,8.

The detailed physical mechanisms determining changes in the distribution of extreme rainfall also require further investigation. Additional latent energy released within storms can invigorate vertical motion, increasing rainfall intensity to above 6–7% per °C (ref. 9). Alternatively, limitations on sources of moisture and counteracting dynamical feedbacks could explain increases in heavy-rain intensity of less than 6–7% per °C (ref. 8). One other possibility to explore is the role of atmospheric aerosols (especially those that absorb sunlight), which may influence global and regional precipitation by modulating the energy balance between the atmosphere and Earth's surface10,11.

Predicting regional changes in the water cycle presents a considerable challenge, but that endeavour is essential in formulating strategies for adapting to and mitigating such changes. Subtle shifts in large-scale atmospheric circulation may affect local rainfall to a much greater extent than the thermodynamic processes relating to atmospheric water vapour content. Understanding the regional responses of rainfall patterns to global warming is therefore crucial. In the meantime, as these two papers1,2 demonstrate, robust physics, combined with carefully constructed observing systems and detailed modelling, indicate that the frequency of intense rainfall events is likely to increase with anthropogenic greenhouse-gas-induced warming.


  1. 1.

    , , & Nature 470, 378–381 (2011).

  2. 2.

    et al. Nature 370, 382–385 (2011).

  3. 3.

    , , , & Environ. Res. Lett. 5, 025205 (2010).

  4. 4.

    & Proc. Natl Acad. Sci. USA 106, 14773–14777 (2009).

  5. 5.

    & J. Clim. 19, 5686–5699 (2006).

  6. 6.

    & Nature 419, 224–232 (2002).

  7. 7.

    , , & Bull. Am. Meteorol. Soc. 84, 1205–1217 (2003).

  8. 8.

    , & Geophys. Res. Lett. 115, D17102, doi:10.1029/2009JD013384 (2010).

  9. 9.

    & Environ. Res. Lett. 5, 025208 (2010).

  10. 10.

    , & Geophys. Res. Lett. 35, L17706, doi:10.1029/2008GL034842 (2008).

  11. 11.

    et al. Geophys. Res. Lett. 37, L1470, doi:10.1029/2010GL043991 (2010).

Download references

Author information


  1. Richard P. Allan is in the Department of Meteorology, University of Reading, Reading RG6 6BB, UK.

    • Richard P. Allan


  1. Search for Richard P. Allan in:

Corresponding author

Correspondence to Richard P. Allan.

About this article

Publication history



Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Newsletter Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing