Vertical profile and flux measurements of ammonia in a deciduous forest in Japan towards improvement of bi-directional exchange model

Ammonia (NH3) is a major component of reduced nitrogen. A bi-directional exchange of NH3 takes place be-tween the atmosphere and surface. Although the study of reduced nitrogen is gaining prominence in recent years, quantification of NH3 deposition is technically challenging due to its complex properties; consequently, this generates significant uncertainty regarding the estimation of total nitrogen deposition. Based on past studies, it has been reported that Asia has numerous high-risk nitrogen deposition sites; however, only a few studies have investigated the NH3 exchange in this region. To clarify the exchange process and achieve improved accuracy regarding the estimation of nitrogen deposition in Asia, we conducted measurements of vertical profile and flux of NH3 in a forest in Japan during summer (July 21-August 1, 2015), winter (February 23-February 29, 2016), and autumn (September 27-October 11, 2016). We measured daytime and nighttime NH3 concentrations at four or five heights of an observation tower in the forest by using filter-pack holders, and determined NH3 fluxes by using the aerodynamic gradient method. During daytime, NH3 showed emission above the canopy in summer, and near the forest floor in winter. There was no clear emission trend in autumn. During nighttime, NH3 showed deposition from the top of the canopy to the forest floor for all observed periods. NH3 exchange in the forest possibly had seasonal and diurnal variations. The flux measurements showed large NH3 emissions at daytime in summer and small depositions at nighttime in summer and autumn. Mean flux during the observation periods was 0.026 mu gm(-2) s(-1). The daytime flux during leafy periods was found to have a strong correlation with solar radiation. At daytime in summer, flux also had strong correlations with relative humidity and horizontal wind speed. We also estimated NH3 fluxes in summer by employing a bi-directional exchange model to verify the applicability of the model at the study site. Inferred fluxes obtained via the model showed large emissions during daytime and agreed with the observation results; however, the model could not reproduce the nighttime deposition. This discrepancy was improved by applying more suitable cuticular resistances.