We present a method for estimating monthly burned area globally at 1° spatial resolution using Terra MODIS data and ancillary vegetation cover information. Using regression trees constructed for 14 different global regions, MODIS active fire observations were calibrated to burned area estimates derived from 500-m MODIS imagery based on the assumption that burned area is proportional to counts of fire pixels. Unlike earlier methods, we allow the constant of proportionality to vary as a function of tree and herbaceous vegetation cover, and the mean size of monthly cumulative fire-pixel clusters. In areas undergoing active deforestation, we implemented a subsequent correction based on tree cover information and a simple measure of fire persistence. Regions showing good agreement between predicted and observed burned area included Boreal Asia, Central Asia, Europe, and Temperate North America, where the estimates produced by the regression trees were relatively accurate and precise. Poorest agreement was found for southern-hemisphere South America, where predicted values of burned area are both inaccurate and imprecise; this is most likely a consequence of multiple factors that include extremely persistent cloud cover, and lower quality of the 500-m burned area maps used for calibration. Application of our approach to the nine remaining regions yielded comparatively accurate, but less precise, estimates of monthly burned area. We applied the regional regression trees to the entire archive of Terra MODIS fire data to produce a monthly global burned area data set spanning late 2000 through mid-2005. Annual totals derived from this approach showed good agreement with independent annual estimates available for nine Canadian provinces, the United States, and Russia. With our data set we estimate the global annual burned area for the years 2001-2004 to vary between 2.97 million and 3.74 million km², with the maximum occurring in 2001. These coarse-resolution burned area estimates may serve as a useful interim product until long-term burned area data sets from multiple sensors and retrieval approaches become available.

This study reports dimethyl sulfide (DMS) sea-to-air fluxes derived from a mass-balance/photochemical-modeling approach. The region investigated was the western North Pacific covering the latitude range of 0°-30°N. Two NASA airborne databases were used in this study: PEM-West A in September-October 1991 and PEM-West B in February-March 1994. A total of 35 boundary layer (BL) sampling runs were recorded between the two programs. However, after filtering these data for pollution impacts and DMS lifetime considerations, this total was reduced to 13. Input for each analysis consisted of atmospheric DMS measurements, the equivalent mixing depth (EMD) for DMS, and model estimated values for OH and NO3. The evaluation of the EMD took into account both DMS within the BL as well as that transported into the overlying atmospheric buffer layer (BuL). DMS fluxes ranged from 0.6 to 3.0 μmol m-2d-1 for PEM-West A (10 sample runs) and 1.4 to 1.9 μmol m-2d-1 for PEM-West B (3 sample runs). Sensitivity analyses showed that the photochemically evaluated DMS flux was most influenced by the DMS vertical profile and the diel profile for OH. A propagation of error analysis revealed that the uncertainty associated with individual flux determinations ranged from a factor of 1.3 to 1.5. Also assessed were potential systematic errors. The first of these relates to our noninclusion of large-scale mean vertical motion as it might appear in the form of atmospheric subsidence or as a convergence. Our estimates here would place this error in the range of O to 30%. By far the largest systematic error is that associated with stochastic events (e.g., those involving major changes in cloud coverage). In the latter case, sensitivity tests suggested that the error could be as high as a factor of 2. With improvements in such areas as BL sampling time, direct observations of OH, improved DMS vertical profiling, direct assessment of vertical velocity in the field, and preflight (24 hours) detailed meteorological data, it appears that the uncertainty in this approach could be reduced to ±25%. Copyright 1999 by the American Geophysical Union.

Biomass burning represents an important source of atmospheric aerosols and greenhouse gases, yet little is known about its interannual variability or the underlying mechanisms regulating this variability at continental to global scales. Here we investigated fire emissions during the 8 year period from 1997 to 2004 using satellite data and the CASA biogeochemical model. Burned area from 2001–2004 was derived using newly available active fire and 500 m. burned area datasets from MODIS following the approach described by Giglio et al. (2006). ATSR and VIRS satellite data were used to extend the burned area time series back in time through 1997. In our analysis we estimated fuel loads, including organic soil layer and peatland fuels, and the net flux from terrestrial ecosystems as the balance between net primary production (NPP), heterotrophic respiration (R_h), and biomass burning, using time varying inputs of precipitation (PPT), temperature, solar radiation, and satellite-derived fractional absorbed photosynthetically active radiation (fAPAR). For the 1997–2004 period, we found that on average approximately 58 Pg C year⁻¹ was fixed by plants as NPP, and approximately 95% of this was returned back to the atmosphere via R_h. Another 4%, or 2.5 Pg C year⁻¹ was emitted by biomass burning; the remainder consisted of losses from fuel wood collection and subsequent burning. At a global scale, burned area and total fire emissions were largely decoupled from year to year. Total carbon emissions tracked burning in forested areas (including deforestation fires in the tropics), whereas burned area was largely controlled by savanna fires that responded to different environmental and human factors. Biomass burning emissions showed large interannual variability with a range of more than 1 Pg C year⁻¹, with a maximum in 1998 (3.2 Pg C year⁻¹) and a minimum in 2000 (2.0 Pg C year⁻¹).

Long term, high quality estimates of burned area are needed for improving both prognostic and diagnostic fire emissions models and for assessing feedbacks between fire and the climate system. We developed global, monthly burned area estimates aggregated to 0.5° spatial resolution for the time period July 1996 through mid-2009 using four satellite data sets. From 2001–2009, our primary data source was 500-m burned area maps produced using Moderate Resolution Imaging Spectroradiometer (MODIS) surface reflectance imagery; more than 90% of the global area burned during this time period was mapped in this fashion. During times when the 500-m MODIS data were not available, we used a combination of local regression and regional regression trees developed over periods when burned area and Terra MODIS active fire data were available to indirectly estimate burned area. Cross-calibration with fire observations from the Tropical Rainfall Measuring Mission (TRMM) Visible and Infrared Scanner (VIRS) and the Along-Track Scanning Radiometer (ATSR) allowed the data set to be extended prior to the MODIS era. With our data set we estimated that the global annual area burned for the years 1997–2008 varied between 330 and 431 Mha, with the maximum occurring in 1998. We compared our data set to the recent GFED2, L3JRC, GLOBCARBON, and MODIS MCD45A1 global burned area products and found substantial differences in many regions. Lastly, we assessed the interannual variability and long-term trends in global burned area over the past 13 years. This burned area time series serves as the basis for the third version of the Global Fire Emissions Database (GFED3) estimates of trace gas and aerosol emissions.

Tropical deforestation contributes to the build-up of atmospheric carbon dioxide in the atmosphere. Within the deforestation process, fire is frequently used to eliminate biomass in preparation for agricultural use. Quantifying these deforestation-induced fire emissions represents a challenge, and current estimates are only available at coarse spatial resolution with large uncertainty. Here we developed a biogeochemical model using remote sensing observations of plant productivity, fire activity, and deforestation rates to estimate emissions for the Brazilian state of Mato Grosso during 2001–2005. Our model of DEforestation CArbon Fluxes (DECAF) runs at 250-m spatial resolution with a monthly time step to capture spatial and temporal heterogeneity in fire dynamics in our study area within the ''arc of deforestation'', the southern and eastern fringe of the Amazon tropical forest where agricultural expansion is most concentrated. Fire emissions estimates from our modelling framework were on average 90 Tg C year⁻¹, mostly stemming from fires associated with deforestation (74%) with smaller contributions from fires from conversions of Cerrado or pastures to cropland (19%) and pasture fires (7%). In terms of carbon dynamics, about 80% of the aboveground living biomass and litter was combusted when forests were converted to pasture, and 89% when converted to cropland because of the highly mechanized nature of the deforestation process in Mato Grosso. The trajectory of land use change from forest to other land uses often takes more than one year, and part of the biomass that was not burned in the dry season following deforestation burned in consecutive years. This led to a partial decoupling of annual deforestation rates and fire emissions, and lowered interannual variability in fire emissions. Interannual variability in the region was somewhat dampened as well because annual emissions from fires following deforestation and from maintenance fires did not covary, although the effect was small due to the minor contribution of maintenance fires. Our results demonstrate how the DECAF model can be used to model deforestation fire emissions at relatively high spatial and temporal resolutions. Detailed model output is suitable for policy applications concerned with annual emissions estimates distributed among post-clearing land uses and science applications in combination with atmospheric emissions modelling to provide constrained global deforestation fire emissions estimates. DECAF currently estimates emissions from fire; future efforts can incorporate other aspects of net carbon emissions from deforestation including soil respiration and regrowth.

Various land-use transitions in the tropics contribute to atmospheric carbon emissions, including forest conversion for small-scale farming, cattle ranching, and production of commodities such as soya and palm oil. These transitions involve fire as an effective and inexpensive means for clearing. We applied the DECAF (DEforestation CArbon Fluxes) model to Mato Grosso, Brazil to estimate fire emissions from various land-use transitions during 2001–2005. Fires associated with deforestation contributed 67 Tg C/yr (17 and 50 Tg C/yr from conversion to cropland and pasture, respectively), while conversion of savannas and existing cattle pasture to cropland contributed 17 Tg C/yr and pasture maintenance fires 6 Tg C/yr. Large clearings (>100 ha/yr) contributed 67% of emissions but comprised only 10% of deforestation events. From a policy perspective, results imply that intensification of agricultural production on already-cleared land and policies to discourage large clearings would reduce the major sources of emissions from fires in this region.

Drainage of peatlands and deforestation have led to large-scale fires in equatorial Asia, affecting regional air quality and global concentrations of greenhouse gases. Here we used several sources of satellite data with biogeochemical and atmospheric modeling to better understand and constrain fire emissions from Indonesia, Malaysia, and Papua New Guinea during 2000–2006. We found that average fire emissions from this region [128 ± 51 (1σ) Tg carbon (C) year⁻¹, T = 10¹²] were comparable to fossil fuel emissions. In Borneo, carbon emissions from fires were highly variable, fluxes during the moderate 2006 El Niño more than 30 times greater than those during the 2000 La Niña (and with a 2000–2006 mean of 74 ± 33 Tg C yr⁻¹). Higher rates of forest loss and larger areas of peatland becoming vulnerable to fire in drought years caused a strong nonlinear relation between drought and fire emissions in southern Borneo. Fire emissions from Sumatra showed a positive linear trend, increasing at a rate of 8 Tg C year⁻² (approximately doubling during 2000–2006). These results highlight the importance of including deforestation in future climate agreements. They also imply that land manager responses to expected shifts in tropical precipitation may critically determine the strength of climate–carbon cycle feedbacks during the 21st century.

Measurements of atmospheric trace gases provide evidence that fire emissions increased during the 1997/1998 El Niño event and these emissions contributed substantially to global CO₂, CO, CH₄, and δ¹³CO₂ anomalies. Interpretation and effective use of these atmospheric observations to assess changes in the global carbon cycle requires an understanding of the amount of biomass consumed during fires, the molar ratios of emitted trace gases, and the carbon isotope ratio of emissions. Here we used satellite data of burned area, a map of C₄ canopy cover, and a global biogeochemical model to quantitatively estimate contributions of C₃ and C₄ vegetation to fire emissions during 1997–2001. We found that although C₄ grasses contributed to 31% of global mean emissions over this period, they accounted for only 24% of the interannual emissions anomalies. Much of the drought and increase in fire emissions during the 1997/1998 El Niño occurred in tropical regions dominated by C₃ vegetation. As a result, the δ¹³CO₂ of the global fire emissions anomaly was depleted (−23.9‰), and explained approximately 27% of the observed atmospheric decrease in δ¹³CO₂ between mid-1997 and the end of 1998 (and 61% of the observed variance in δ¹³CO₂ during 1997–2001). Using fire emissions that were optimized in an atmospheric CO inversion, fires explained approximately 57% of the observed atmospheric δ¹³CO₂ decrease between mid-1997 and the end of 1998 (and 72% of the variance in δ¹³CO₂ during 1997–2001). The severe drought in tropical forests during the 1997/1998 El Niño appeared to allow humans to ignite fires in forested areas that were normally too moist to burn. Adjacent C₄ grasses (in woodlands and moist savannas) also burned, but emissions were limited, in part, by aboveground biomass levels that were 2 orders of magnitude smaller than C₃ biomass levels. Reduced fuel availability in some C4 ecosystems may have led to a negative feedback on emissions.

New burned area datasets and top-down constraints from atmospheric concentration measurements of pyrogenic gases have decreased the large uncertainty in fire emissions estimates. However, significant gaps remain in our understanding of the contribution of deforestation, savanna, forest, agricultural waste, and peat fires to total global fire emissions. Here we used a revised version of the Carnegie-Ames-Stanford-Approach (CASA) biogeochemical model and improved satellite-derived estimates of area burned, fire activity, and plant productivity to calculate fire emissions for the 1997–2009 period on a 0.5° spatial resolution with a monthly time step. For November 2000 onwards, estimates were based on burned area, active fire detections, and plant productivity from the MODerate resolution Imaging Spectroradiometer (MODIS) sensor. For the partitioning we focused on the MODIS era. We used maps of burned area derived from the Tropical Rainfall Measuring Mission (TRMM) Visible and Infrared Scanner (VIRS) and Along-Track Scanning Radiometer (ATSR) active fire data prior to MODIS (1997–2000) and estimates of plant productivity derived from Advanced Very High Resolution Radiometer (AVHRR) observations during the same period. Average global fire carbon emissions according to this version 3 of the Global Fire Emissions Database (GFED3) were 2.0 Pg C year⁻¹ with significant interannual variability during 1997–2001 (2.8 Pg C year⁻¹ in 1998 and 1.6 Pg C year⁻¹ in 2001). Globally, emissions during 2002–2007 were relatively constant (around 2.1 Pg C year⁻¹) before declining in 2008 (1.7 Pg C year⁻¹) and 2009 (1.5 Pg C year⁻¹) partly due to lower deforestation fire emissions in South America and tropical Asia. On a regional basis, emissions were highly variable during 2002–2007 (e.g., boreal Asia, South America, and Indonesia), but these regional differences canceled out at a global level. During the MODIS era (2001–2009), most carbon emissions were from fires in grasslands and savannas (44%) with smaller contributions from tropical deforestation and degradation fires (20%), woodland fires (mostly confined to the tropics, 16%), forest fires (mostly in the extratropics, 15%), agricultural waste burning (3%), and tropical peat fires (3%). The contribution from agricultural waste fires was likely a lower bound because our approach for measuring burned area could not detect all of these relatively small fires. Total carbon emissions were on average 13% lower than in our previous (GFED2) work. For reduced trace gases such as CO and CH₄, deforestation, degradation, and peat fires were more important contributors because of higher emissions of reduced trace gases per unit carbon combusted compared to savanna fires. Carbon emissions from tropical deforestation, degradation, and peatland fires were on average 0.5 Pg C year⁻¹. The carbon emissions from these fires may not be balanced by regrowth following fire. Our results provide the first global assessment of the contribution of different sources to total global fire emissions for the past decade, and supply the community with an improved 13-year fire emissions time series.