Recent drought events underscore the vulnerability of Amazon forests to understorey fires. The long-term impact of fires on biodiversity and forest carbon stocks depends on the frequency of fire damages and deforestation rates of burned forests. Here, we characterized the spatial and temporal dynamics of understorey fires (1999–2010) and deforestation (2001–2010) in southern Amazonia using new satellite-based estimates of annual fire activity (greater than 50 ha) and deforestation (greater than 10 ha). Understorey forest fires burned more than 85 500 km 2 between 1999 and 2010 (2.8% of all forests). Forests that burned more than once accounted for 16 per cent of all understorey fires. Repeated fire activity was concentrated in Mato Grosso and eastern Pará, whereas single fires were widespread across the arc of deforestation. Routine fire activity in Mato Grosso coincided with annual periods of low night-time relative humidity, suggesting a strong climate control on both single and repeated fires. Understorey fires occurred in regions with active deforestation, yet the interannual variability of fire and deforestation were uncorrelated, and only 2.6 per cent of forests that burned between 1999 and 2008 were deforested for agricultural use by 2010. Evidence from the past decade suggests that future projections of frontier landscapes in Amazonia should separately consider economic drivers to project future deforestation and climate to project fire risk.

Aim In any region affected, fires exhibit a strong seasonal cycle driven by the dynamic of fuel moisture and ignition sources throughout the year. In this paper we investigate the global patterns of fire seasonality, which we relate to climatic, anthropogenic, land-cover and land-use variables.Location Global, with detailed analyses from single 1°¥ 1°grid cells. MethodsWe use a fire risk index, the Chandler burning index (CBI), as an indicator of the 'natural' , eco-climatic fire seasonality, across all types of ecosystems. A simple metric, the middle of the fire season, is computed from both gridded CBI data and satellite-derived fire detections. We then interpret the difference between the eco-climatic and observed metrics as an indicator of the human footprint on fire seasonality. ResultsDeforestation, shifting cultivation, cropland production or tropical savanna fires are associated with specific timings due to land-use practices, sometimes largely decoupled from the CBI dynamics. Detailed time series from relevant locations provide comprehensive information about these practices and how they are adapted to eco-climatic conditions. Main conclusionsWe find a great influence of anthropogenic activities on global patterns of fire seasonality. The specificity of the main fire practices and their easy identification from global observation is a potential tool to support land-use monitoring efforts. Our results should also prove valuable in the development of a methodological approach for improving the representation of anthropogenic fire practices in dynamic global vegetation models.

Abstract. Tropical forests have been a permanent feature of the Amazon basin for at least 55 million years, yet climate change and land use threaten the forest's future over the next century. Understory forest fires, which are common under the current climate in frontier forests, may accelerate Amazon forest losses from climatedriven dieback and deforestation. Far from land use frontiers, scarce fire ignitions and high moisture levels preclude significant burning, yet projected climate and land use changes may increase fire activity in these remote regions. Here, we used a fire model specifically parameterized for Amazon understory fires to examine the interactions between anthropogenic activities and climate under current and projected conditions. In a scenario of low mitigation efforts with substantial land use expansion and climate change -Representative Concentration Pathway (RCP) 8.5 -projected understory fires increase in frequency and duration, burning 4-28 times more forest in 2080-2100 than during 1990-2010. In contrast, active climate mitigation and land use contraction in RCP4.5 constrain the projected increase in fire activity to 0.9-5.4 times contemporary burned area. Importantly, if climate mitigation is not successful, land use contraction alone is very effective under low to moderate climate change, but does little to reduce fire activity under the most severe climate projections. These results underscore the potential for a fire-driven transformation of Amazon forests if recent regional policies for forest conservation are not paired with global efforts to mitigate climate change.

Abstract. Vegetation fires have been acknowledged as an environmental process of global scale, which affects the chemical composition of the troposphere, and has profound ecological and climatic impacts. However, considerable uncertainty remains, especially concerning intra and inter-annual variability of fire incidence. The main goals of our global-scale study were to characterise spatial-temporal patterns of fire activity, to identify broad geographical areas with similar vegetation fire dynamics, and to analyse the relationship between fire activity and the El Niño-Southern Oscillation. This study relies on 10 years (mid 1996–mid 2006) of screened European Space Agency World Fire Atlas (WFA) data, obtained from Along Track Scanning Radiometer (ATSR) and Advanced ATSR (AATSR) imagery. Empirical Orthogonal Function analysis was used to reduce the dimensionality of the dataset. Regions of homogeneous fire dynamics were identified with cluster analysis, and interpreted based on their eco-climatic characteristics. The impact of 1997–1998 El Niño is clearly dominant over the study period, causing increased fire activity in a variety of regions and ecosystems, with variable timing. Overall, this study provides the first global decadal assessment of spatial-temporal fire variability and confirms the usefulness of the screened WFA for global fire ecoclimatology research.

Quantification of biogenic carbon fluxes from agricultural lands is needed to generate comprehensive bottom-up estimates of net carbon exchange for global and regional carbon monitoring. We estimated global agricultural carbon fluxes associated with annual crop net primary production (NPP), harvested biomass, and consumption of biomass by humans and livestock. These estimates were combined for a single estimate of net carbon exchange and spatially distributed to 0.05°resolution using Moderate Resolution Imaging Spectroradiometer satellite land cover data. Global crop NPP in 2011 was estimated at 5.25 ± 0.46 Pg C yr À1, of which 2.05 ± 0.05 Pg C yr À1 was harvested and 0.54 Pg C yr À1 was collected from crop residues for livestock fodder. Total livestock feed intake in 2011 was 2.42 ± 0.21 Pg C yr À1, of which 2.31 ± 0.21 Pg C yr À1 was emitted as CO 2 , 0.07 ± 0.01 Pg C yr À1 was emitted as CH 4 , and 0.04 Pg C yr À1 was contained within milk and egg production. Livestock grazed an estimated 1.27 Pg C yr À1 in 2011, which constituted 52.4% of total feed intake.Global human food intake was 0.57 ± 0.03 Pg C yr À1 in 2011, the majority of which was respired as CO 2 .Completed global cropland carbon budgets accounted for the ultimate use of approximately 80% of harvested biomass. The spatial distribution of these fluxes may be used for global carbon monitoring, estimation of regional uncertainty, and for use as input to Earth system models.