Previous modelling studies suggest that SRM could stabilize average global temperatures, but in so doing would also reduce average global rainfall. But the trouble with global averages is that they don't tell you about the detail. What will a SRM world feel like in Pakistan, Canada, Australia and Spain, for example?

To try and answer this question Katharine Ricke, from Carnegie Mellon University, US, and her colleagues used a large-ensemble coupled climate model to examine the temperature and rainfall anomalies for 23 different land regions, under 54 different SRM scenarios. Each model was initiated in the model-year 2005 and simulated global climate for an 80-year period. Compared to the control scenario, where no SRM was applied and average global temperatures increased by 2.5 °C, all model runs were able to stabilize global temperatures back to the 2005 average. But what was striking was the regional variation. "We were able to demonstrate that different levels of SRM would be preferable for different regions," Ricke told environmentalresearchweb.

Essentially the researchers showed that it is impossible to use SRM to return the climate of all regions back to some historical baseline, such as the climate of the late 20th century. Instead there will always be winners and losers, with some regions benefiting more from one particular SRM scenario than others.

For example when India and eastern China are compared in the year 2070, the model suggests that high levels of SRM would bring eastern China's baseline temperature back to within 0.5 °C of its 1990s baseline, but it would also make the region significantly drier. Meanwhile, under the same conditions, India would become much cooler – 1.5 °C less than the 1990s baseline – and a little wetter than it is now.

However, in pretty much all cases, carrying out some form of SRM was better than doing nothing. "Our results confirm that solar-radiation management would generally lead to less extreme temperature and precipitation anomalies, compared with unmitigated greenhouse-gas emissions," says Ricke.

Nonetheless, Ricke and colleagues believe that the regional differences in climate response to different SRM scenarios are going to make it very difficult to reach an agreement about which course of action to take. For example, in one area the main priority might be to protect a snowpack and sustain water resources, while another region may prefer a different SRM scenario that assures the continued operation of a summer monsoon system. And even within one region different groups of people may have differing interests. "For example, indigenous peoples may want to preserve the summer Arctic sea ice that is critical to their way of life, whereas maritime industries may prefer summers with an ice-free Arctic Ocean," Ricke and her colleagues write in their paper.

Detailed impact studies to assess the benefits and drawbacks of different forms of SRM still need to be carried out. Meanwhile governments, businesses and other interested parties might want to start practising their negotiation skills.

The researchers reported their work in Nature Geoscience.