Overview of the 2015 fire season severity prediction

This webpage presents a prediction of fire risk for the 2015 dry season in high biomass burning regions of South America. The following figure presents fire season severity indices (FSSI, ranging from 0-100) for 6 states in Brazil (Acre, Amazonas, Maranhão, Mato Grosso, Pará, and Rondônia), 3 departments in Bolivia (El Beni, Pando, and Santa Cruz), and one country (Peru) using sea surface temperature information through the end of May. Green indicates below average predictions of fire activity whereas orange and red indicate above average activity. Because of warm sea surface temperatures (SSTs) in the tropical Pacific Ocean and near-normal SSTs in the tropical Atlantic during boreal winter and spring, the predicted 2015 fire season shows a distinctive east-west gradient across the Amazon. In eastern and southeastern regions of the Amazon, including the states of Maranhão, Pará, Mato Grosso and Rondônia, projected FSS is close to the 65th percentile relative to the long term mean FSS. In contrast, FSS in southwestern and central Amazonia are projected to be below the long term mean. Terrestrial water storage at the end of the wet season, shown below for April 2015, provides additional support for differences in fire risk between the eastern and western Amazon. A detailed description of the prediction method is given here.

Check predictions for other years: 2012 | 2013 | 2014 | 2015

Fire observations and predictions

FSS predictions compared to observations

This figure compares the observed and modeled FSS in South America fire regions. The black solid lines are observations for past years, and the black dashed horizontal lines represent the all year mean observed values. The orange lines are FSS derived from the empirical model. The orange shades indicate the range of predicted FSS in 2015 for each region. The numbers in the parentheses represent the coefficient of determination (r2). Information on sea surface temperatures through Apr of 2015 were evaluated for these predictions. Click the boxes below the figure to see previous predicitons using earlier data.

Month of prediction:

Satellite observation of active fires

FSS is the sum of active fire counts (FC) during the fire season. Active fires are the thermal radiation anomalies created by fires that are detected by satellites. Our empirical model was based on active fire observations from Moderate Resolution Imaging Spectroradiometer (MODIS) on board the NASA Terra satellite.

  • Maps of monthly active fires as observed by MODIS in South America. Click "<" or ">" for active fire data in previous or next year/month. Click "Enable/Disable smoothing" to see the location of each observed fire.

  • A map of FC density (in numbers per million hectares per year) averaged over 2001-2014. The location of 10 geographical regions considered in this study is marked. Click on the map for a larger version.

  • Fraction of active fire counts in each region (among the 10 regions considered in this study). Move to pointer over each wedge to see the mean annual active fire counts in each region.

  • Time series of total FC (in numbers per month) in the 10 regions of South America since 2000.

Fire-climate conditions

Ocean climate indices

To predict FSS in South America, we used two climate indices that represent the sea surface temperature anomalies in Pacific and Atlantic: ONI (Ocean Nino Index) and AMO (Atlantic Multidecadel Oscillation index). The following figure shows time series of ONI and AMO since 2000.

Relationship between ocean climate indices and FSS

  • Linear regression analyses show that the annual FSS had largest correlation with ONI and AMO during months between October previous year and April. Here we show the monthly ONI and AMO values during this time period for eath year. The colors from dark blue to dark red represent the order of lowest to highest fire years.

  • The following figure compares interannual variations of sea surface temperatures and fire season severity in South America. The upper panel shows the time series of mean ONI and AMO anomalies averaged over the period from October (previous year) and April (current year). The anomalies are relatave to the 2001-2010 mean. The bottom panel shows the time series of annual FSS.

Terrestrial water storage

NASA's Gravity Recovery and Climate Experiment (GRACE) satellite shows below average terrestrial water storage (TWS) over most of eastern Amazonia in April 2015, indicating less soil moisture recharging during the wet season than in previous years. Without access to sufficient soil water, even deep-rooted Amazon trees may experience drought effects, reducing evapotranspiration and lowering atmospheric moisture levels during the dry season. Reduced atmospheric moisture, in turn, dries surface fuels and increase fire risk. Below-average TWS in the eastern Amazon is consistent with our SST-based forecast for above average fire activity in these regions in 2015. Positive TWS anomalies in the western Amazon confirm projections of low fire season severity in these regions.


The method of annual FSS prediction is based on Chen et al. (2011) with some modifications. We developed our empirical model of FSS using fire counts detected by MODIS onboard NASA's Terra satellite along with Oceanic Nino Index (ONI) and Atlantic Multidecadal Oscillation index (AMO) SST anomaly time series. Sea surface temperatures prior to the onset of the fire season have the strongest relationship with the number of satellite observed fires during the fire season in many areas of South America. The lead times enable us to make a prediction for the upcoming fire season.


  • Active fire counts
  • We used MODIS collection 5 global monthly fire location product (MCD14ML). We sampled the geographic coordinates of individual fire pixels (at a 1×1 km spatial resolution) that had a confidence level greater than 30%, and calculated the monthly FC within each 0.5° pixel after applying a cloud fraction correction. Persistent hot spots from MODIS observations and gas flare pixels in NOAA Global Gas Flare Estimates were excluded because the burning in these pixels is primarily associated with petroleum production rather than landscape fires. We then calculated the monthly FC for each region (6 states in Brazil (Acre, Amazonas, Maranhao, Mato Grosso, Para, Rondonia), 3 departments in Bolivia (El Beni, Pando, Santa Cruz), and one country (Peru)). The sum of FC during the fire season (defined as the 9-month period centered at the peak fire month) was recorded as the annual FSS for each region.

  • ONI
  • The Oceanic Niño Index (ONI) is a 3-month mean SST anomaly in the Niño 3.4 region (5°N-5°S, 120°-170°W) of the Pacific. We obtained the ONI time series from the NOAA National Weather Service Climate Prediction Center.

  • AMO
  • The Atlantic Multidecadal Oscillation index (AMO) represents a similar 3-month mean for the North Atlantic (0°-70°N). We obtained the AMO index time series from the NOAA Earth System Research Laboratory website.


We defined our empirical predictive model as a linear combination of the two climate indices sampled during the months of maximum correlation:


FSSpredicted is the predicted FSS in region x and year t. The parameter τc indicates the lead time (number of months before the peak fire month) when the prediction was made. a and b are spatial varying coefficients that represent the sensitivities of FSS in each region to ONI and AMO, individually, and c is a constant. ONI and AMO were sampled each year during months with lead times τONI and τAMO relative to the peak fire month (m) in each region. Given a target τc, the optimal τONI and τAMO values were derived from a series of linear regressions using ONI and AMO values at different months (with a cutoff(minimum) lead time of τc).


Based on the data (ONI and AMO) availability and the peak fire month, we derived the τc for each region. We then applied the predictive model with corresponding coefficients (a, b, and c) and optimal lead times (τONI and τAMO) to derive the FSS in the target fire year. The range of the prediction was calculated using the 1-sigma uncertainty estimates for the parameters of the predictive model. Therefore, we have a set of predictions derived from different months (though with different confidence).



  • Chen, Y., J. T. Randerson, D. C. Morton, R. S. DeFries, G. J. Collatz, P. S. Kasibhatla, L. Giglio, Y. Jin, M. E. Marlier, Forecasting fire season severity in South America using sea surface temperature anomalies, Science, 334, 787-791, 2011. [link]
  • Chen Y., I. Velicogna, J. S. Famiglietti, and J. T. Randerson, Satellite observations of terrestrial water storage provide early warning information about drought and fire season severity in the Amazon, J. Geophys. Res. - Biogeosciences, 118, 1-10, 2013. [link]
  • Chen, Y., J. T. Randerson, D. C. Morton, Y. Jin, G. J. Collatz, P. S. Kasibhatla, G. R. van der Werf, R. S. DeFries, Long-term trends and interannual variability of forest, savanna and agricultural fires in South America, Carbon Management. 4(6), 617-638. [link]
  • de Linage, C., J. S. Famiglietti, and J. T. Randerson (2014), Statistical prediction of terrestrial water storage changes in the Amazon Basin using tropical Pacific and North Atlantic sea surface temperature anomalies, Hydrol Earth Syst Sc, 18(6), 2089-2102.[link]
  • Chen, Y., J. T. Randerson, D. C. Morton (2015), Tropical North Atlantic ocean-atmosphere interactions synchronize forest carbon losses from hurricanes and Amazon fires, Geophysical Research Letters, in press.


  • FC : Active fire counts, defined as the number of fire/hotspots observed by satellite.
  • AMO: Atlantic Multi-decadal Ossilation index, representing sea surface temperature anomaly in North Atlantic. We used 3-month mean of Kalplan SST anomalies in North Atlantic (0-70N). Data available at NOAA Earth System Research Laboratory website.
  • Fire season: The fire season is defined here as the period from 4 months before the peak fire month to 4 months after the peak fire month.
  • FSS: Fire season severity, defined as the sum of FC during the fire season (4 months before the peak fire month to 4 months after the peak fire month) for each year.
  • FSSI: FSS index, a measure of FSS based on historical mean values and standard deviation in the same region. FSSI = 50*(1+ERF((FSS-FSSoavg)/sqrt(2)/FSSostd), where FSSoavg and FSSostd are mean and standard deviation of observed FSS during 2001-2010. ERF is the error function.
  • MODIS: Moderate Resolution Imaging Spectroradiometer, an earth observation remote sensing instrument on board the Terra satellite and Aqua satellite.
  • ONI: Ocean Nino Index, representing sea surface temperature anomaly in Eastern tropical Pacific. We used 3-month mean of ERSST.v3b SST anomalies in the Niño 3.4 region (5N-5S, 120-170W). Data available at NOAA Climate Prediction Center website.
  • Terra: A NASA research satellite in a sun-synchronous orbit around the earth. It carries a payload of five remote sensors including MODIS.


This work is funded by the Gordon and Betty Moore Foundation through Grant GBMF3269 and the US Agency for International Development (USAID).

This work is the result of a collaboration between University of California, Irvine (Yang Chen and Jim Randerson), NASA Goddard Space Flight Center (Doug Morton and James Collatz), Columbia Univeristy (Ruth DeFries and Miriam Marlier), University of Maryland (Louis Giglio), and Duke University (Prasad Kasibhatla).

NASA provided the satellite observations of fires and NOAA provided the sea surface temperature time series used in our analysis. The interactive figures and maps were generated using Google's Chart API , Maps API, and Fusion table API .

Doug Morton at NASA Goddard Space Flight Center and Pineda Llopart Serrano translated this forecast website to Porturguese and Spanish.


The model predictions contained on this website are highly experimental. They cannot be used to predict the occurrence of individual fires. Use of this information for planning purposes should also draw upon other independent and reliable climate information sources. The Regents of The University of California will not be liable for any consequences that may occur if you rely on this information.


We are interested in hearing your thoughts on our Web site. You may contact us by email us, or click here to leave your comments.
blog comments powered by Disqus