The way we humans are affecting the climate is changing rainfall patterns over land and sea, scientists at Lawrence Livermore National Laboratory (LLNL) in California have found. Kate Marvel and Céline Bonfils compared precipitation ‘fingerprints’ in satellite data against what climate models showed would result from actions like adding greenhouse gases to the atmosphere. “Everyone knows that temperatures are rising, but figuring out how that affects other aspects of the climate is tricky,” Kate told me. “We’ve shown that global precipitation is changing in the way climate scientists expect it to. The odds of the observed trends being due to natural climate variability are very low.”
Changes to rain, snow and all the other forms of falling wetness collectively known as precipitation are undeniably important, given their power to bring floods and droughts. Scientists have already shown that, over land, wet areas are getting wetter and dry areas are getting drier. These studies rely on data measured directly on land, reaching back almost a century. The long record gives scientists a lot of data to test, making it easier to tell human influences from the many natural rainfall patterns. Yet Kate and Céline wanted to use satellite data instead. Though these have only been recorded since 1979, each measurement is more reliable, and the satellites also cover the oceans.
“With such a short record, it’s often difficult to identify the ‘signal’ of climate change against the background of completely natural variability,” Kate explained. For example, the wet-gets-wetter, dry-gets-dryer strengthening of the Earth’s water cycle happens because warmer air can hold more water vapour. But that can be caused by the El Niño climate pattern, as well as by increasing greenhouse gases. Our activities can also change how air circulates above the planet, pushing dry regions and storm tracks toward the poles – but so can the La Niña pattern.
Climate change rides tandem
“It’s very easy to get either of these mechanisms from natural variability, but it’s very unlikely to see both occurring at the same time,” Kate stressed. “The satellite record shows that these two effects are happening increasingly in tandem. That’s incompatible with our best understanding of natural fluctuations in the climate system, but completely in line with how we think human activities affect the system.”
Kate and Céline have developed a new way to exploit the fact that it’s unusual for both effects to happen at the same time. “We used multiple climate models to identify the characteristic pattern of human induced climate change – the fingerprint,” she said. That fingerprint method makes it easier to compare against what might be expected from natural variability, and identify changes in the climate system caused by other factors.
In a Proceedings of the National Academy of Sciences of the USA paper published on Monday, the scientists followed mathematically-smoothed satellite precipitation data from pole to pole each year. They focussed on where precipitation peaked and bottomed out, and the points halfway in between, looking at how those points move, and how much rainfall changes in each place. They then used the same method on data from models cross-checked by the Coupled Model Intercomparison Project that LLNL hosts.
Knowing what we face
Averaging over several models to help make it easier to separate the human fingerprint from natural variability, Kate and Céline created two scenarios. In the first, they kept heat energy trapped by greenhouse gases in the atmosphere at levels last found before the world started industrialising. In this case, the extreme wet and dry spots moved further towards the equator as they got wetter and drier respectively, the response they expected. In the second scenarios, the heat energy trapped increases rapidly, driving the extreme wet and dry spots further towards the poles, as they get wetter and drier respectively. Overlaying the satellite fingerprint onto models using the two scenarios, the LLNL scientists worked out the odds that the trends in the satellite data were due to natural climate variability. They found they were much lower than the 1-in-20 chance level scientists normally call ‘statistically significant’.
Kate says this method could help detect whether changes in other climate data are caused by humans, but that it’s not suitable for giving detailed information for specific places. “We focused on very large-scale effects,” she added. “To provide useful regional information, we need good small-scale observations and multiple reliable high-resolution climate models, neither of which is available. But this does have real-world implications, particularly for those who live near dry zones or in the tropics.”
And in more general terms, she adds that these results further builds the case that humans are driving climate change. “It’s important to understand a) what climate change looks like and b) whether we can see the tell-tale signs of anthropogenic [human-caused] climate change in the observations,” she says. “We have very good evidence that human activity is affecting lots of variables, from temperature to circulation patterns to water vapour. This study adds to the evidence that climate change is indeed happening and is probably down to us.”
Kate Marvel and Céline Bonfils (2013). Identifying external influences on global precipitation Proceedings of the National Academy of Sciences of the USA DOI: 10.1073/pnas.1314382110