Why does the southern hemisphere have smaller annual temperature variations than northern hemisphere

First published in IGBP's Global Change  magazine Issue 76, January 2011

Most studies that reconstruct the climatic conditions of the past centuries to millennia tend to focus on the northern hemisphere. But now an intriguing multicentennial record of temperature and precipitation in southern South America is available. Raphael Neukom and Jürg Luterbacher elaborate on its significance.

Raphael Neukom is a Postdoctoral Research Fellow at the Oeschger Centre for  Climate Change Research (OCCR), University of Bern, Switzerland and at the School of Earth Sciences, University of Melbourne, Australia. E-mail: .

Jürg Luterbacher is Professor for Physical Geography in the Department of Geography, Justus Liebig University of Giessen, Germany. E-mail: .

We need to know about past climate because it is important both for understanding how the climate system works and for improving the accuracy of projections of future climate change. But the instrumental climate record covers just the past 150 years or so. To overcome this limitation, reconstructions of the climate further back in time rely on what are termed as “proxies” – historical documents or natural archives that allow indirect estimates of variables such as temperature and rainfall. Yet, much of this research has focused on the northern hemisphere; the southern hemisphere mostly remains terra incognita. Now, results of an international research initiative mean we can begin to quantify continental-scale climate variations in southern South America before the instrumental record began. These results, synthesised by Neukom et al. (2010), provide a high-resolution record of temperature and precipitation in South America going back several hundred years.  

South America is an especially important landmass in the southern hemisphere, for it spans a range of climates that are influenced by multiple drivers such as the El Niño Southern Oscillation, Antarctic Climate and the high Andes, for example. Although proxies with good age control are sparse in tropical South America, the southern half of the continent provides many more proxies that we can use to infer past climatic fluctuations. Ice cores from Andean glaciers and ice fields, drilled at up to 6100 metres above sea level, record snow accumulation and the chemical composition of the atmosphere over time. Trees living up to 3500 years and growing at altitudes up to 5000 metres above sea level respond to variations in temperatures and/or drought, recording this information in the width of their annual growth rings. We derived additional information from lake and marine sediments. Furthermore, we used historical documents from the time of the Spanish colonisation, now stored in many different archives in Europe and the Americas, that report on agricultural yields and the climatic (and non-climatic) causes of yield fluctuations.

It is one thing to have proxy records, though, and another to pool together available information for specific locations and then infer the palaeoclimatic history of half a continent. In the area of investigation, proxies were not distributed evenly in space and time. Some proxies were more suited to estimating summer temperatures than winter ones. And some were actually from areas outside of southern South America but were known to have the same set of controls on their climate. All of this meant that we needed an elaborate statistical methodology to reconstruct the annual history of summer and winter temperatures, and of precipitation (rain and snow) of the region. We now have summer temperatures stretching back more than 1000 years. But winter temperatures could be reconstructed only for the past 300 years or so due to the more limited number of proxy data that resolve winter temperature conditions. We constructed summer and winter precipitation to the late 15th and late 16th centuries respectively.

Comparisons and contrasts
The new records now allow us to compare the climate evolution of both hemispheres, leading to some interesting observations. Consider the comparison between the summer and winter temperature trends for southern South America and Europe, for which seasonally resolved temperature reconstructions are available (Figure 1). During some periods, the summer temperatures in the two regions seem to have fluctuated quite synchronously, for example in the 17th and 20th centuries. This co-variation could arise from global controls such as changes in solar irradiation, large volcanic eruptions or decadal-scale changes in the behaviour of globally relevant climate phenomena such as the El Niño-Southern Oscillation. It could just as well be a chance phenomenon. Other periods do not show a synchronicity for summer temperatures, and the winter temperatures generally do not seem to vary in consort. It is likely that the effects of the global forcing mechanisms were superimposed with and perturbed by strong regional to hemispheric-scale influences during these periods. To pinpoint the causes of these variations, we will require reconstructions from other regions and climate-model simulations.

Figure 1. Decadal-scale summer (top) and winter (bottom) temperature fluctuations in southern South America (Neukom et al. 2010) and Europe (Luterbacher et al. 2007) over the past 500 years, expressed as anomalies with respect to the 1901-1995 reference period. Shaded areas represent the reconstruction uncertainties expressed as two standard errors. The black lines extend the temperature curves to the present using instrumental data.  

Notably, in South America the 20th-century warming was less pronounced in the context of the preceding centuries than for instance in Europe, especially during the summer (Figure 1). For example, we found that the average surface temperature in southern South America rose 0.44°C in the 20th century. European records representing northern hemisphere conditions show a more pronounced rise of 0.79°C. The lower rate of warming in the southern hemisphere, which is also evident in climate models, can be explained by the role played by the waters in this hemisphere. The Southern Ocean has a large capacity to take up heat and store it in its waters, which leads to a smaller warming at the surface. The southern, water-dominated hemisphere thus reacts with a certain lag to global warming as compared to the northern, land-dominated hemisphere. This could be why the temperatures in southern South America are not yet extraordinary compared with previous centuries. In contrast, comparable northern-hemisphere reconstructions do show 20th-century temperatures that are markedly different from past centuries. In fact, changes in atmospheric dynamics have even led to cooling over a small area in coastal Chile in recent decades (Falvey and Garreaud 2009).

In contrast to the more muted warming, we find that in recent decades precipitation has changed substantially in some areas of southern South America. For example, the patterns of annual rainfall during the past four centuries in the catchment of the Laguna Mar Chiquita, a large lake in northern Argentina, are rather different from those in central Chile (Figure 2). In the former region, a large jump in rainfall amounts (an increase of more than 100 millimetres per year on average) occurred in the 1970s, signalling a shift from a relatively dry regime to the presently wet one. In contrast, central Chile currently suffers from a prolonged drying; modern conditions are probably drier than at any time over the last four centuries. In general, our data and analysis suggest that summers in many parts of southern South America have become progressively wetter, whereas winters have become drier.

Figure 2. 400 years of annual precipitation history at Laguna Mar Chiquita (northern Argentina; upper panel) and Central Chile (lower panel). Precipitation was reconstructed using a probabilistic “ensemble reconstruction” approach. The median represents the most probable values. The dotted line represents 0-millimetre anomalies (reference period 1901-1995). Modified after Figure 3 from Neukom et al. (2010).

Societal implications
We know that key historical events in this region may have coincided with changes in climate. Earlier work has shown, for example, that periods of warfare and migration to fortified sites in the Andean Altiplano in the 14th and 15th centuries coincided with and may have resulted from severe drought. Climate reconstructions of the sort discussed in this study can help provide a far more detailed context for these events. They also have implications for contemporary societies. The economies on both sides of the Andes are highly dependent on the fresh water that stems from the mountains for irrigation and hydropower generation. Recent changes in the total amount and the seasonal distribution of precipitation have significant consequences for these sectors. Our reconstructions show that such changes have occurred in the past and should be incorporated into future economic planning and adaptation strategies.

Climate-model simulations for the 21st century project up to 50 percent reduction in precipitation relative to the present day conditions, mainly in the central Chile area. In combination with future melting of Andean glaciers, this may lead to critical reductions in water availability, which may strongly affect agriculture, freshwater supply and hydropower generation in some areas. In the northern and southern parts of the study area, models project rather wetter conditions, which may benefit the agricultural sector in the highly populated area between Buenos Aires and Rio de Janeiro.

The southern hemisphere has not received the kind of attention from climate researchers that it deserves. Our research is the first to reconstruct the regional climate in any part of this hemisphere at a high temporal resolution. Although the results raise more questions than they answer at this stage, we hope they will provide a foundation for further regional studies in South America in particular and the southern hemisphere in general. More importantly, they are expected to refine our understanding of how and why climate changes at local and regional scales, and thereby guide our responses to future change.

This research was conducted under the auspices of the LOTRED-SA project, an initiative of IGBP’s Past Global Changes (PAGES) project and the recently launced PAGES 2K  initiative. This initiative aims to reconstruct the climatic variation of all continents at high temporal and spatial resolution over the past two millennia.
http://www.pages.unibe.ch/science/last2millennia.html

References
Falvey M and Garreaud R (2009) Journal of Geophysical Research DOI:10.1029/2008JD010519.

Luterbacher J et al. (2007) Geophysical Research Letters, DOI: 10.1029/2007gl029951.

Neukom R et al. (2010) Climate Dynamics, DOI: 10.1007/s00382-010-0793-3.

Neukom R et al. (2010) Geophysical Research Letters, DOI: 10.1029/2010glo43680.