Probabilistic estimation of changes in BVOC emissions due to projected land-use change: atmospheric and policy implications

Humans have altered Earth's natural land surface for thousands of years, e.g., agriculture, deforestation, biomass burning. Scientists have studied extensively how these activities alter the biogeochemical cycles (e.g., water and carbon), and how they affect surface albedo and consequently the radiative balance of the atmosphere. Vegetation emit a number of biogenic hydrocarbons (isoprene, C5H8, represents the largest emission) with chemical lifetime of days or less that can lead to the production of ozone, a surface air pollutant and a greenhouse gas in the upper troposphere. We know from in situ measurements that different vegetation emit these reactive hydrocarbons at different rates so that altering the spatial patterns of vegetation will alter budgets of biogenic hydrocarbons with serious implications for surface ozone. We know that tropical ecosystems emit ~70% of the global isoprene budget. A good understanding of how human activity will alter Earth's natural land cover underpins climate projections of the next few decades. The main scientific objective of this proposal is to reduce uncertainty in land-use change projections related to human activity and subsequent changes in the magnitude and distribution of biogenic hydrocarbon emissions and surface ozone. Projecting human activity and its role in land-use change is difficult, involving many different socio-economic issues. However, it is an integral part of the Earth system and if we are to improve climate projections we must consider this human element. Up until now, the human contribution to land-use change has been considered independent of the rest of the climate system. Past work has focused primarily on developing complex models that consider a whole range of socio-economic phenomena but this complexity effectively limits the exploration of alternative socio-economic futures because of computational constraints. What we propose is to develop a simpler model, with only a few describing parameters, that captures many of the broad scale features of the more complex models. Such a model enables a more comprehensive exploration (and uncertainty analysis) of alternative futures. This probabilistic approach allows us to determine the most statistically likely future. In our work we will use the developed model to explore the most likely human-induced land-use change over the next 40 years. An excellent example of a current socio-economic driver of land-use change is biofuel production. Biofuels represent an arguably cleaner alternative to burning traditional fossil fuels and is increasingly being used as an ingredient in car diesel fuel and in power stations. Palm oil is the most prominent biofuel, which is grown largely in Southeast Asia. Production of palm oil, and associated rates of tropical deforestation, is increasing rapidly in response to demand, representing significant changes in tropical land-cover. In many cases hydrocarbon emissions from palm oil are larger than the indigenous crops so that increased biofuel production effectively increases the global biogenic hydrocarbon budget with (yet unquantified) implications for surface air pollution. Future biofuel production will be considered in the land-use model. We will estimate emissions of biogenic hydrocarbons using inventories based in situ measurements so that for projection of land-use we will have an associated distribution of biogenic hydrocarbon emissions. To quantify the impact of these emissions on surface air pollution we will use a start-of-the art computer model of atmospheric chemistry and transport. Running such a model is computationally intensive for a large ensemble of biogenic hydrocarbon emission inventories so we will limit our calculation to the minimum, maximum and most likely global biogenic hydrocarbon emissions. Our calculations will significantly reduce uncertainty on the human element on land-use change and implications for surface air pollution.