Intra- and inter-annual changes in isoprene emission from central Amazonia

Abstract. Isoprene is a chemical compound emitted naturally by soil, microorganisms, plants, and animals into the atmosphere. But plants are the largest emission source, and the amount of emission depends on plant species, weather conditions, and environmental conditions, including exposure to environmental stresses such as heat and drought. Isoprene is very reactive in the atmosphere and is involved in atmospheric physicochemical processes that can impact atmospheric chemistry, air quality, and regional climate. Quantification and understanding of the atmospheric processes influenced by isoprene result from a combination of observational experiments and estimates obtained from computational models. However, only a few long-term observational experiments have been conducted in the largest source of isoprene to the global atmosphere – the Amazon rainforest, and there are still uncertainties in the model estimates. Recent experiments have also shown that the models have greater uncertainty when estimating intra- and inter-annual variations in isoprene. This study aimed to improve our understanding of isoprene emission from a central Amazonian site by considering biological and environmental factors influencing emission on intra- and interannual time scales. By combining observational datasets, we adapted a widely used computational model of isoprene emission to observed conditions in the field. Our observations indicated that isoprene emission was not as high as the model estimated when the forest experienced environmental stress, such as heat and drought, in the 2015 El-niño year. In addition, observations revealed that the model performed well when diurnal variations were analyzed but not when long-term variations occurred. The performance for estimating intra- and inter-annual isoprene emission improved when the model was modified on two biological factors – (i) the amount of different leaf ages throughout the year and (ii) the emission rates of these different leaf ages. This shows that isoprene emission estimates can be improved when biological processes are mechanistically incorporated into the model.