Droughts are weather extremes that are hostile enough that plenty of sci-fi and post-apocalyptic stories use near-permanent droughts for apocalyptic backdrops (Waterworld notwithstanding). And for good reason—drought is part of the reality-based picture of modern climate change, as combined trends in rainfall and evaporation are bringing drier conditions to some regions. But understanding trends is a challenge: more rain is being delivered to other regions, drought conditions are naturally variable, and historical rainfall data is limited.
Researchers have typically turned to tree rings for archives of past droughts. By compiling records from many trees, historical maps called "drought atlases" have been built for a number of regions and can cover nearly a millennium. These can provide incredible historical information, including events like the megadroughts of the Western United States between 800 and 1300 CE. But each drought atlas is only one piece of the global picture.
A new study led by NASA's Kate Marvel pulls all these regional drought atlases together—along with recent data and climate-model simulations—to see what they can tell us about human impacts on drought since 1900.
A global atlas
The researchers' approach was similar to other studies that have examined patterns of climate change for a human fingerprint. Many climate-model simulations were averaged together to highlight the expected long-term drought trend driven by human-induced climate change. Real-world data was then analyzed to see if a similar trend has emerged from the typical range of natural variability. In this case, those tree-ring-based drought atlases provided the range of natural drought variability around the world.
This type of analysis breaks things down to a signal-to-noise ratio—the stronger the human-caused signal stands out from the natural, noisy variability, the clearer it is that humans are changing the climate. When changes are great enough, the trend leaves the range of past events entirely and enters what is essentially uncharted territory.
The model simulations with human greenhouse gas emissions reflect the general patterns seen in the data, with shifts toward more drought in North America and Europe but increasingly wet conditions around India and western China. But the way this plays out over the 20th century is interesting, leading the researchers to put brackets around three periods: 1900 to 1949, 1950 to 1975, and 1981 to the present.
A tale of three trends
Over the first half of the century, the signal of human-driven trends in drought became increasingly clear, crossing the 90-percent statistical confidence level. But through the 1950s, '60s, and early '70s, global temperature dipped, partly due to a rapid increase in sunlight-reflecting aerosol pollution. This brought the signal back toward the noise range.
Interpreting this as aerosol-driven, however, is complicated. Aerosol pollution affects precipitation in complex ways that are different from greenhouse gas-driven warming, and these interactions are still a weak point for climate-model simulations. So while the researchers see some hints that the global drought trends over this 25-year period resemble the effects of aerosol pollution, they can't say anything confidently.
The last period starts in 1981, when a couple of newer drought datasets come online. The data again shifts in the direction of a greenhouse gas-caused signal, but they don't quite leave the range of natural variability. Because it's a short amount of time, that range can more easily obscure a trend.
Since the model simulations continue in the future using a scenario of high greenhouse gas emissions, the researchers also looked at where these trends are headed. That shows that we can expect the human signal to emerge clearly from the noise in the near future. The global signal crosses the 99-percent statistical confidence level by the mid-2030s, though most individual regions take longer. Europe and parts of Asia would emerge by the 2040s, while North America and Mexico follow in the 2060s, owing to their greater natural variability.
So what can you take away from all this? The idea that there is a potentially detectable influence of aerosol pollution on drought patterns in the 1950s, '60s, and '70s is pretty noteworthy. But more broadly, the authors note, "Multiple observational datasets and reconstructions using data from tree rings confirm that human activities were probably affecting the worldwide risk of droughts as early as the beginning of the 20th century."
And that risk is only going to grow as the world keeps warming. "The human consequences of this," the researchers write, "particularly drying over large parts of North America and Eurasia, are likely to be severe."
Nature, 2019. DOI: 10.1038/s41586-019-1149-8 (About DOIs).
