Wildfires threaten the Amazon rainforest and may affect the local climate, in turn making future fires more likely. Suzanne Bevan and Peter North describe how satellite data is changing our view of what fire means for this vulnerable but crucially important ecosystem.
Research into the effects of burning biomass – wood, foliage and other plant matter – on rainfall over the Amazon may not seem an obvious topic to choose after my PhD work on using satellite-borne radar to measure how fast glaciers in the Arctic were flowing. But both research topics involve processing large amounts of data collected and transmitted back to Earth by satellites flying thousands of miles a day across the globe.
Over the Amazon, as in the Arctic, local feedback processes can amplify the effects of global climate change on the region. And just like the Arctic, the Amazon is an area so vast and inaccessible that remote sensing from space is the best way to monitor it.
The Amazon is the largest remaining rainforest on Earth and plays a major role in regulating the planet’s climate. But tens of thousands of square kilometres of Amazon rainforest are destroyed each year by slash-and-burn practices, which local people use to clear land for farming. As well as this deliberate deforestation, climate modellers fear that climate change will make the Amazon warmer and drier, causing the forest to die back.
These scientists use climate models that include a dynamic vegetation component – they represent the ability of vegetation to grow or die back in response to climate change. The models do not, however, consider the possible feedback effects between smoke from the burning forests and the region’s rainfall patterns, partly because we don’t understand these interactions well enough.
There were more fires and more smoke in dry years. And the more smoke there was, the later the dry season ended.
Aerosols are tiny solid or liquid particles, less than a thousandth of a millimeter across, that float suspended in the atmosphere. Around the world there are many sources of aerosols, such as desert dust, sea spray and industrial pollution, and their concentrations in the atmosphere vary throughout the year and from place to place.
While they remain in the atmosphere, aerosols scatter and absorb sunlight, producing a cooling at the Earth’s surface which in some areas may be up to three times greater than the warming caused by increasing greenhouse gases. Aerosols also act as cloud condensation nuclei, providing a seed onto which water vapour can condense to form cloud droplets. This means adding aerosols to the atmosphere can change the properties of clouds, changing their reflectivity, their lifetimes and also their ability to produce precipitation (see p9).
During the rainy season, the atmosphere over the Amazon rainforest is so clean it has been referred to as a green ocean. In contrast, during the dry season in September and October biomass burning pollutes the atmosphere with smoke aerosols. Forest fires in the Amazon do not occur naturally. They happen because people deliberately start them, and in a dry year it’s much more likely these fires will ‘leak’ beyond the area they were intended to burn.
Fire in the forest
The key question is, do these fires intensify or extend the drought by suppressing rainfall? If so, droughts could cause more fires, which would then make further drought more likely – a vicious circle. Answering this question means disentangling cause and effect, and requires repeated observations of rainfall, fires and smoke aerosols over as long a period as possible. Recently, satellite remote sensing has begun to provide the data we need.
Comparative map of effects of biomass burning. In July 2005 the atmosphere over the Amazon region was clean and aerosol optical depths were low. By September 2005, when biomass burning occurred, aerosol optical depths were higher than they would have been over an extremely polluted city.
Aerosol Optical Depth (AOD) is a measure of the total amount of sunlight scattered or absorbed by aerosols throughout the depth of the atmosphere. This lets us estimate how many aerosol particles are in the atmosphere at a particular moment. But measuring it from space is difficult because satellites, looking down from above the top of the atmosphere, see light from the sun scattered back by both the surface and the atmosphere.
At Swansea we have developed an algorithm to use data from radiometer instruments designed in the UK, which have been flying onboard European Space Agency satellites since 1995. These instruments view the Earth from two different angles and use four different wavelengths (colours), letting us separate light scattered by the Earth’s surface from light scattered by atmospheric aerosols. The rainfall observations we used are based on a combination of data from rain gauges and observations from satellite microwave and infrared instruments. The data on fires come from the dates and locations of night-time hot pixels identified in satellite thermal images, showing areas of land that are much warmer than their surroundings.
Using these three collections of data, we found that from 1995 to the 2008 dry-season, AOD was strongly correlated with the number of fires and inversely correlated with the amount of precipitation. In other words, as expected, there were more fires and more smoke in dry years. We also found that the more smoke there was, the later the dry season ended. These results were very interesting but they were compatible with the possibility of a local climate feedback effect rather than proving it. Were the fires suppressing rainfall, or were they just spreading in response to dry conditions?
We know that Amazon forest fires are not a natural phenomenon and, therefore, that the amount of burning is influenced by economics and legislation as well as by climate. For example, when prices of agricultural crops rise, farmers have more incentive to cut down and burn forest to clear land to plant them. With our unique 13-year time series of dry-season AODs we were able to identify trends in biomass burning over several years, and to relate these to external factors.
We found that when land values were increasing, due to worldwide demand for soybeans and beef, concentrations of aerosols in the atmosphere rose. More recently, from 2004 onwards, falling soybean prices, a strong Brazilian currency and active government intervention reduced the demand for land and AODs declined as less forest was burned. 2005 was an exception – aerosols were extremely high, probably because of an unusually severe drought that year.
Our research shows that biomass burning in the Amazon region, driven by local and global economics, can affect the local climate in ways which amplify the regional consequences of global climate change. Now that global demand is rising again for soybeans and beef, and for ethanol to replace fossil fuels, there is a risk that deforestation and burning will begin to rise again and exacerbate the effects that the drying and die back of the Amazon rainforest are predicted to have on the climate.
Dr Suzanne Bevan and Dr Peter North work on remote sensing and environmental change in the School of the Environment & Society at Swansea University. Email: firstname.lastname@example.org or email@example.com. The research described in this article was funded by the NERC National Centre for Earth Observation.
S. L. Bevan, P. R. J. North, W. M. F. Grey, S. O. Los, and S. E. Plummer (2009). Impact of atmospheric aerosol from biomass burning on Amazon dry-season drought. Journal of Geophysical Research – Atmospheres 114.