Do past fires explain current carbon dynamics of Amazonian forests?

Amazon forests process and store large quantities of carbon in vegetation and soils. These forests, assumed to be mostly 'old-growth' and fire-free are exhibiting a remarkable feature-they are gaining the equivalent mass of a small car every year in aboveground biomass per hectare of forest (0.89 Mg/ha/yr). These gains are attributed to increasing atmospheric CO2, which has a fertilization effect on tree growth. However, fire in these 'old-growth' forests may be more recent than expected (in the last few centuries), and regrowth from fire together with soil charcoal-which has a fertility effect-may be contributing together with CO2 fertilisation to the observed increases in biomass in these apparently mature forests.

Current understanding of drivers of these increases is limited by, (i) an unknown fire history of plots used to monitor change, and (ii) lack of information of how resource change affects these forests. The effects of pre-modern fire on forest regrowth and the gain have not been evaluated.

Our pilot analysis of radiocarbon dated fire from soil charcoal indicates that even the wettest Amazon rainforest has burned, with forests considered to be 'old-growth' having burned within the last few centuries, and 70% of plots (n=70) containing visible soil charcoal fragments. Periodic drier climate and fire use by Native Americans before their populations collapsed ~450 years ago following Europeans colonisation may have resulted in a higher fire frequency than currently observed. Forest regrowth from these and more recent fires may still be occurring in forests considered to be 'undisturbed', e.g., some trees may grow to be 980 years old in central Amazonia, so that forest considered 'old-growth' may still be approaching equilibrium as long-lived trees mature following fire.

Fire also produces soil pyrogenic carbon (PyC) as charcoal that is found in archaeological sites in terra preta soils and in upland soils across Amazonia far from evidence of human settlement. Soil PyC increases soil fertility on otherwise nutrient poor soils, and being resistant to decomposition, may have increased soil fertility across the Amazon. Pre-modern fire and soil PyC are therefore two important ingredients in understanding how Amazonian forests currently function. We will determine whether regrowth following past fire and soil PyC fertility effects in 'old-growth' permanent forest plots across Amazonia contributes to the observed carbon sink.

We have developed a basin-wide network of on-the-ground sample plots, and because methods of measuring the forest with these are fully standardised even across nations they represent an excellent opportunity to measure the response of Amazon forest to historic fire and soil PyC. In permanent forest plots across the Amazon Basin our Objectives are: 1) determine spatial patterns in 'time since last fire'; 2) determine soil PyC stocks, and how these are affected by climatic, edaphic conditions, and fire intensity; 3) using results from (1) and (2), determine whether spatial patterns of productivity and carbon gains in aboveground tropical forest trees in Amazonia are consistent with regrowth from historical fire disturbance and by soil PyC acting as a soil fertility enhancer.

Our research will improve understanding of fire patterns across the Amazon for long-term forest plots (the same plots used to estimate the current carbon sink). We will provide a first quantification of: soil PyC stocks, basin-wide environmental drivers of soil PyC stocks, and whether soil fertility is greater where soil PyC levels are higher. This will be a first large-scale test of whether forest productivity, structure, and increases in carbon can be attributed to regrowth from historic fire and soil PyC fertility effects. The results are vital for conservation planning, to estimate the longevity of the carbon sink, and for policy such as Reducing Emissions from Deforestation and Degradation (REDD+).