Climate change and the hydrological cycle
The prospects of significant and damaging changes in the hydrological cycle due to the increase in atmospheric greenhouse gas concentrations were raised in earlier IPCC reports and restated more strongly in the most recent, 2007 Fourth Assessment Report (AR4). Now, the U.S. Climate Change Science Program (CCSP) issued its final Synthesis and Assessment Report on Abrupt Climate Change with an entire chapter dedicated to the subject of “Hydrological Variability and Change” addressing this potential climate hazard. Overall, the chapter confirms the IPCC AR4 conclusions giving an extensive survey of the history of U.S. droughts and the physical mechanism behind their occurrence. Here we highlight a few points regarding the major conclusions stated in the CCSP Report.
1. Two types of drought:
As described in the report, historical instrumental observations and pre-instrumental reconstructions based on proxy-data (and going back, in large details, two millennia) show that the U.S. West has been repeatedly subjected to multiyear, damaging droughts. These and other data also show that the multiyear droughts are linked with changes in ocean surface temperatures in the tropical Pacific and Atlantic Oceans. In particular, the droughts tend to occur when the eastern tropical Pacific is colder than normal (a protracted La Niña state) and are more severe if, at the same time, the tropical North Atlantic is warmer than normal. Climate models show that these ocean temperature variations cause changes in the patterns of the atmospheric circulation and its moisture transport, which lead to the hydrological changes. These sea surface temperature changes arise, it is thought, from natural variability of the tropical atmosphere-ocean system.
Model projections of 21st century climate response to increased atmospheric concentration of greenhouse gases, depict an entirely different type of drought. In the U.S., the difference is not so much in the location of the drying regions; here too the West will be inflicted. The difference is in the global extent of the drying regions and the causal mechanism. Global warming will lead to a worldwide reduction in rainfall in the subtropical latitude belts. The belts that circle the globe between about 15 and 35 degree latitude on both sides of the equator. Some locations in this latitude belt, such as the Sahara Desert are currently already bone dry but other, such as the U.S. Southwest, northern Mexico and the Mediterranean countries do receive adequate rainfall for human needs.
2. How does global warming cause subtropical drying?
To understand the mechanism of greenhouse warming related droughts we note that regionally, the hydrological cycle is a balance between two large terms: the atmospheric moisture influx into the region and the local difference between precipitation and evaporation (two smaller terms: storage of moisture in the upper ocean or in the ground and runoff of surface water in streams and rivers, complete the balance). Much of the subtropics lie over the world ocean areas where climate models indicate that when greenhouse gas concentrations increase and the surface warms “dry” regions, from which the atmosphere extracts moisture, will experience increased drying while “wet” regions, where atmospheric moisture inflow allows precipitation to exceed evaporation, will get wetter. This is a direct result of the warmer atmosphere being able to hold more moisture and thus existing patterns of atmospheric moisture transport intensify. Since currently the atmosphere exports moisture out of the subtropics and evaporation in these areas exceeds precipitation, this imbalance will increase in the future, drying these belts even further.
Over land the situation is somewhat different. Here the surface can only provide for evaporation if it receives precipitation. It must therefore be the atmosphere that supplies land regions with moisture. According to the above argument, land areas should therefore receive more moisture in the future. However, in the subtropics this atmospheric moisture transport is quite weak and climate models indicate that global warming will weaken it further. In the future, the models suggest, atmospheric circulation will change slightly such that the subtropical dry zones will expand towards the poles. The expansion will cause the precipitation-bearing, mid-latitude storm to tracks further poleward than at present. This will decrease the already weak atmospheric moisture transport into subtropical land areas and increase their aridity. This seems to be the faith of the U.S. Southwest and northern Mexico, the countries surrounding the Mediterranean Sea, South Africa, Argentina, and southern Australia.
3. Future droughts and “natural” climate variability
One implication of the existence of two types of drought is that the two can coexist, sometimes adding and at other times offsetting one another. As anthropogenic warming progresses, the subtropical drying will gradually intensify but could also temporally be exacerbated if protracted cooling occurs in the tropical Pacific (due to the occurrence of a La Niña). In contrast, if the tropical Pacific experiences a protracted warming (El Niño) the anthropogenic drought might be alleviated. In all, it is important to remember that natural climate variability has a considerable impact on the hydrological cycle in the U.S. West (and elsewhere) and that therefore we will continue to experience cycles of dry and wet intervals as the baseline gradually shifts towards a drier climate. In this case, the CCSP and IPCC 21st century projections do not address the impact of natural climate variability. They rather focus on the change in the climate base line, and have to be taken in that context.
4. How certain are these projections?
The CCSP report refers to subtropical drying as likely (implying that there is a larger than 65% chance that it will occur). To understand the reason for this claim we note that climate models represent a state of the art compromise between physical understanding of these interactions, the ability to quantify them numerically, and computing power. Both precipitation and evaporation are governed by processes and interactions that are small-scale, hard to quantify, and computationally expensive to model. Observations of the changing climate, which are used to test model projections and assess their reliability, are in the case of hydrological variables notoriously suspect. This is because their large variability in time and space and because their ground observation network is sparse (extremely so in the case of evaporation). All this leads to less reliable projection than, say, in the case of temperature. The climate science community has appreciated these difficulties and has invested large efforts into improving the knowledge of hydrological processes and in obtaining better and higher-resolution global hydrological data. Improved scientific understanding enables a qualitative assessment of the overall feasibility and realism of model projections, while the use of multiple model projections help address individual model deficiencies and biases in simulating the global and regional change. Model consensus on the turn to drier conditions in the U.S. West is high although not unanimous. The brief physical arguments for subtropical drying above (point 2) support the numerical projections of the models. The scientists participating in the synthesis report combined all this information in arriving at their uncertainty estimate.
Interested readers can find additional discussions of the past and future of U.S. Droughts on the LDEO Drought Research web pages.
— Yochanan Kushnir and Richard Seager