Researchers at NASA and the University of California, Irvine, using a variety of satellite data, expect Amazon forests this year to see a below-average wildfire season. The fire season in this region of South America typically begins in May, peaks in September and ends in January.

"2014 looks to be a quiet year for fires in the Amazon region, as sea surface temperatures remained neutral in the Pacific and cool in the Atlantic through April, late in the wet season," said Doug Morton a co-investigator on the project at NASA's Goddard Space Flight Center in Greenbelt, Maryland.

The forecast is based on a model of fire season severity developed by Yang Chen and Jim Randerson at the University of California, Irvine, in collaboration with scientists at NASA. The model, published 2011 in Science, considers historical fire data from NASA's Terra satellite, along with sea surface temperature data from National Oceanic and Atmospheric Administration buoys.

"This season may be the calm before the storm," Randerson said.

"Although the conditions were more favorable for higher levels of rainfall and soil moisture build-up over the last six to eight months, conditions are now changing. With the probability of El Nino building this fall, we have to be prepared for higher fire risk and damages in 2015."

A previous study by the same researchers established that sea surface temperatures in the tropical Pacific and Atlantic oceans can be used to forecast fire risk during the Amazon fire season three to six months prior to peak fire activity.

Warmer sea surface temperatures toward the end of the wet season lead to reduced precipitation and drier soils at the start of the dry season. A drier start to the fire season sets the stage for higher fire activity across the region.

The team also uses data from the joint NASA/German Aerospace Center (DLR) Gravity Recovery and Climate Experiment (GRACE) mission to follow changes in soil water during the dry season. GRACE measurements represent the "memory" of past rainfall events and the level of water stress for Amazon trees.

This year, conditions look similar to those in 2009, a year with some of the lowest fire activity during the Terra record, which goes back to the year 2000. Average and below-average temperatures in the Pacific and Atlantic oceans are consistent with conditions averaged over the long-term that are shown to produce a quiet fire season.

"We can be confident in the seasonal fire forecast because the climate system - sea surface temperatures, large-scale patterns of rainfall, and atmospheric circulation - typically don't change too abruptly," Morton said.

The 2013 season was an exception. Warm sea surface temperatures in the Atlantic in the early and middle of the Amazon wet season prompted the model to forecast an average to above-average fire season. Then, sea surface temperatures in the tropical Atlantic returned to neutral conditions in the late wet season, between February and June, allowing normal passage of tropical rainfall to move south into the Amazon in a way the model did not anticipate.

"Our forecast of fire season severity was too high in 2013, yet we were in good company," Morton said. "Hurricane forecasts also projected an above-average year for tropical storms in the Atlantic. Sea surface temperatures in the tropical north Atlantic affect both hurricanes and fires. We're continuing to investigate the climate connections between tropical storms and fires to better understand the Earth system."

Much of the fire activity across the Amazon region can be explained by climate variability, yet intentional fires for land management or deforestation are also important. In regions like Brazil, where deforestation rates have declined sharply since 2005, year-to-year variability in fire detections from NASA satellites is closely tied to climate.

However, in regions like Peru, agricultural management and deforestation fires have increased in recent years, so a climate-only forecast is likely to underrepresent the risk of fire activity.