Improving the Temperature Dependence of Modeled Isoprene Emissions in MEGAN with Remote Sensing Constraints

Isoprene is the most abundant biogenic volatile organic compound in Earth's atmosphere. It is emitted naturally by many species of plants, primarily in tropical regions. The seasonal cycle of isoprene emissions is largely controlled by variations in meteorology, particularly surface temperature and sunlight, as well as variations in leaf area and leaf age throughout the year. The oxidation of atmospheric isoprene by the hydroxyl (OH) radical has a large impact on many chemical species that are relevant for air quality and climate, including nitrogen oxides, ozone, carbon monoxide, and secondary organic aerosols. Accurate isoprene emissions estimates are therefore required in order to improve air quality and climate models in the tropics. However, major discrepancies exist between modeled and observationally constrained isoprene emissions estimates, due in part to uncertainties in isoprene emission models. Such models parameterize the relationships between isoprene emissions and different driving variables, such as air temperature, based on experimental measurements. The accuracy of these parameterizations in a particular ecosystem depends on the availability of representative measurements, which introduces large uncertainties in remote tropical regions where surface measurements are scarce. In this study, we use satellite-derived isoprene emission estimates to improve the temperature dependence of the isoprene emission parameterization in the widely used Model of Emissions of Gases and Aerosols from Nature (MEGAN) in several tropical regions. This is done using a Metropolis-Hastings Markov Chain Monte Carlo data assimilation framework. This approach offers a means of improving the seasonal cycle of modeled isoprene emissions in regions where there are limited ground-based observations to sufficiently constrain the emission estimates in the model.