The underlying science behind climate models has greatly improved in the past couple of decades, partly due to a concerted research effort known as the Coupled Model Intercomparison Project. Through that work, scientists have improved the accuracy of climate models and their representation of Earth's physical processes.
Another type of model used to project climate impacts, however, still has significant uncertainty, researchers say. These models, called impact models, take the climate model outputs that show changes in the atmosphere and ocean systems and translate them into effects on things like agriculture, flooding, drought and even human health.
Such results would be incredibly useful for planners and policymakers, but impact models are still in their early stages, with somewhat unreliable projections.
That's a major reason why scientists engaged at the Potsdam Institute for Climate Impact Research in Germany have been working to improve them. That effort, known as the Inter-Sectoral Impact Model Intercomparison Project, or ISI-MIP, was launched in early 2012.
Earlier this week, a group of more than 30 research teams spanning 12 countries all involved in the ISI-MIP project published their first results in a special issue of the journal Proceedings of the National Academy of Sciences
Impact models are "the second step after the climate models, and it's a relevant step for what climate change means for humans," said Katja Frieler, a climate impacts researcher at the Potsdam Institute who is coordinating the project.
While the results from many of the model runs have considerable uncertainty, this first effort is part of a comparative, iterative process similar to the one that early climate models went through. Eventually, the effort will make impact models more robust, Frieler said.
"Even if there are differences between the models, it's extremely helpful to analyze where they come from and where actually our uncertainties lie," she said.
The papers from the project were also submitted to the Working Group 2 of the Intergovernmental Panel on Climate Change, which will be able to use their results in its assessment of the impacts of climate change.
Can plants adapt to frequent droughts?
One of the project's studies, for example, explored climate change's impact on hydrological drought, which is a type of drought associated with decreased runoff, leading to water shortages in rivers, aquifers, reservoirs and other parts of a watershed.
Led by Christel Prudhomme, hydrologist with the Centre for Ecology & Hydrology in Wallingford, England, researchers concluded that under all but one of seven global impact models used in their study, drought is expected to increase in both global extent and frequency by the end of the century.
Although this study uses a different metric for drought (earlier work focused on soil moisture), this finding is not qualitatively different from the existing body of research on climate change and drought, said Martin Hoerling, a meteorologist with the National Oceanic and Atmospheric Administration's Earth System Research Laboratory.
But an inconsistent result in Prudhomme's work raises questions about how a warming planet might interact with the biosphere, potentially altering future drought projections.
One impact model used in the study, which takes into account how plants might adjust to an increase in carbon dioxide, did not show a significant increase in drought. Called the JULES model, it assumes that plants will "breathe" less in a CO2-enriched atmosphere and therefore will not lose as much water in the process.
"If the plants don't dry up the soil as much, then you have more water available for runoff," Prudhomme said.
Prudhomme believes the results of the JULES model should cause researchers to re-examine their earlier conclusions on hydrological drought under climate change.
"There is a lot of uncertainty because the transformation from rainfall to river flow is complex," she said. "All of these models, they try to solve this complexity slightly differently."
She added, "It's time now to seriously look at hydrological models and understand their deficiencies, understand the processes we can't capture well yet ... and we need to understand better why that is the case."
Hoerling agreed, saying, "These guys have illustrated another layer of complexity that we really need to work out," namely, how plants respond to increasing greenhouse gases and how that affects soil moisture and runoff.
But he also stressed that although the JULES model demonstrates that the global ecosystem could mitigate climate change's impacts on surface runoff and hydrology, all signs still point to the conclusion that droughts will worsen by the end of the century.
"What I would be reluctant to draw from this is somehow that these authors have raised a new red flag that maybe drought isn't going to increase." he said. "The overall body of evidence [states] that the occurrences of drought will increase."
A plunge into the complexities of flooding
Another paper was the first to use both climate models and river flow simulations from a variety of hydrologic models to address changes in global flood risks. Despite the fact that scientists expect flooding impacts to increase with climate change, there are few published papers in this research area.
The results, based on a worst-case warming scenario known as RCP 8.5, suggest that in half the river areas examined, 30-year floods will become more frequent by the end of the century, said Rutger Dankers, a scientist at the United Kingdom's Met Office and lead author of the paper.
In areas where flood risk decreased, Dankers suggested it might be due to reductions in snowfall linked to climate change.
Yet the impact models used in the flood study also had some disagreements, which are a source of uncertainty, he noted. For example, some models might predict flooding increases for an area, while others predict decreases.
That information about uncertainty can be given back to the impact modelers, who can then use that to improve the models and learn why they have contradictory outcomes.
"What are the reasons behind that and what are the processes we aren't sure of?" Dankers asked.
Other results from the impact model studies published in the special issue looked at impacts on water availability, agriculture and health (ClimateWire, Dec. 17).
One study quantified the amount that water scarcity increases due to climate change in addition to scarcity from simple population growth (ClimateWire, Dec. 18).
The research also identified hot spots of climate change impacts in the Amazon, the Mediterranean and East Africa, where different stresses, like drought and agriculture, can interact.
Frieler, of the Potsdam Institute, said her group planned to continue working on improving impact modeling studies. It is also making the results of its model runs publicly available to any researcher who would like to use them.
The goal, she said, is to do for impact models what the Coupled Model Intercomparison Project did for climate models: concentrate scientific focus on specific areas for improvement and eventually end up with a robust, reliable set of models.
"We think that it would be great to establish something like that for the impacts in the same way," she said.