We knew there must have been a connection between climate warming and damaging droughts in the world’s largest rainforest, as the two globally hottest years on record (2005 and 2010) coincide with the two record droughts in the Amazon – which led to large-scale forest degradation and the release of several gigatonnes of extra CO2 into the atmosphere.

But why exactly? Shouldn’t global warming lead to a simple increase in the general circulation – so even more rain in the wet tropics? As always, practice is more complicated.

Newly collected paleoclimatic data from between 49,500 to 16,000 years ago including warmer and colder episodes during the last ice age help to better explain the emerging rainforest droughts. The Amazon it seems has been a bit unfortunately placed – just south of the equator.

Earth scientists from the University of Massachusetts and the University of Minnesota have uncovered there may be a link across thousands of kilometres of atmosphere between high latitude northern hemisphere warming and precipitation decrease in the southern hemisphere tropics, which includes the Amazon basin, they write in a new publication in Science.

Warm up north, dry in southern tropics

Measurements of stalagmites and other cave sediments from South America linked to the Greenland ice cores temperature dataset show that rainfall increased in the southern hemisphere tropics during northern hemisphere cold spikes – and decreased during warmer intervals.

More specifically the South American Summer Monsoon is weakened during warm intervals within Dansgaard/Oeschger and Heinrich climate fluctuations. Such fluctuations occur on a millennium timescale.

According to the authors this suggests an “anti-phase relationship to Northern Hemisphere Monsoon” – which would likely increase under northern hemisphere warming.

[This monsoon increase could be noticeable across the southern tip of the Indian subcontinent, where indeed monsoon rains have increased – but not over the Ganges basin – where monsoon rains are blocked by the Asian Brown Cloud. Following the same logic one would be inclined to suggest perhaps droughts in the Horn of Africa would decrease – and Queensland monsoon rains too – although both phenomena are much stronger related to ENSO variability.]

Other climate links to Amazon drought

To suggest a climatic mechanism between northern hemisphere warming and Amazon droughts – another recent Science publication is also of interest. In November a research team led by the University of California in Irvine concluded Amazon wildfires (a result of droughts) could be correlated to sea surface temperature anomalies, both from the El Niño/La Niña Southern Oscillation (ENSO) in the Pacific – but also, and this is a stronger correlation, to the Atlantic Multidecadal Oscillation index.

A temperature increase of one degree in the Pacific (as during an El Niño phase) would lead to substantially higher forest fire risk in the following months in the Amazon. More surprisingly though this same wildfire increase occurs as a result of a positive temperature anomaly of just 0.25 degrees in the North Atlantic.

These researchers have even devised a climate forecasting model in order to predict the Amazon wildfire chance – which according to them performed well predicting the exceptional drought of 2010.

Decreased South American monsoon seasonality

The reason – to our basic understanding of meteorology – is that higher water temperatures lead to stronger convection over the oceans, so the formation of depressions, and would therefore lure away or diffuse part of the Intertropical Convergence Zone (ITCZ) of the South American Summer Monsoon.

Connecting the two Science studies suggests this process not only occurs on a year-to-year timescale – but can lead to net precipitation decreases across the Amazon basin over periods of centuries to millennia as well.

[Image: South American Summer Monsoon 1979-1995 average air pressure, wind and precipitation distribution, courtesy: V. Kousky and M. Halpert, NCEP/NCAR archive.]

© Rolf Schuttenhelm | www.bitsofscience.org