Emission-dominated gas exchange of elemental mercury vapor over natural surfaces in China

Xun Wang,Xinbin Feng,Wei Zhu,Jonas Sommar,Che-Jen Lin,Wei Yuan

doi:10.5194/acp-16-11125-2016

Abstract

Abstract. Mercury (Hg) emission from natural surfaces plays an important role in global Hg cycling. The present estimate of global natural emission has large uncertainty and remains unverified against field data, particularly for terrestrial surfaces. In this study, a mechanistic model is developed for estimating the emission of elemental mercury vapor (Hg0) from natural surfaces in China. The development implements recent advancements in the understanding of air–soil and air–foliage exchange of Hg0 and redox chemistry in soil and on surfaces, incorporates the effects of soil characteristics and land use changes by agricultural activities, and is examined through a systematic set of sensitivity simulations. Using the model, the net exchange of Hg0 between the atmosphere and natural surfaces of mainland China is estimated to be 465.1 Mg yr−1, including 565.5 Mg yr−1 from soil surfaces, 9.0 Mg yr−1 from water bodies, and −100.4 Mg yr−1 from vegetation. The air–surface exchange is strongly dependent on the land use and meteorology, with 9 % of net emission from forest ecosystems; 50 % from shrubland, savanna, and grassland; 33 % from cropland; and 8 % from other land uses. Given the large agricultural land area in China, farming activities play an important role on the air–surface exchange over farmland. Particularly, rice field shift from a net sink (3.3 Mg uptake) during April–October (rice planting) to a net source when the farmland is not flooded (November–March). Summing up the emission from each land use, more than half of the total emission occurs in summer (51 %), followed by spring (28 %), autumn (13 %), and winter (8 %). Model verification is accomplished using observational data of air–soil/air–water fluxes and Hg deposition through litterfall for forest ecosystems in China and Monte Carlo simulations. In contrast to the earlier estimate by Shetty et al. (2008) that reported large emission from vegetative surfaces using an evapotranspiration approach, the estimate in this study shows natural emissions are primarily from grassland and dry cropland. Such an emission pattern may alter the current understanding of Hg emission outflow from China as reported by Lin et al. (2010b) because a substantial natural Hg emission occurs in West China.

Highlights

Accurate inventories of mercury (Hg) emission are the foundation for assessing Hg global biogeochemical cycling (Selin, 2009; Streets et al, 2009, 2011)
2014; Bash, 2010; Scholtz et al, 2003; Zhang et al, 2012b), this model (1) builds a new scheme for estimating the air– soil flux based on the reduction pathways of reactive Hg in soil identified in the literature, (2) develops a scheme for the Hg flux exchange over rice paddy, which is an important land use feature in China, and (3) updates the scheme for the air– snow interface and chemical parameters for air–foliage flux (Table 1)
Using a mechanistic model incorporating the present state of understanding in Hg transformation in soils and on foliage surface with up-to-date datasets of soil characteristics and land use changes, the natural emission of Hg0 vapor in China is estimated to be 465.1 Mg yr−1, including 565.5 Mg yr−1 of emission from soils, 9.0 Mg yr−1 of emission from water bodies, and −100.4 Mg yr−1 deposition by vegetation

Summary

Introduction

Accurate inventories of mercury (Hg) emission are the foundation for assessing Hg global biogeochemical cycling (Selin, 2009; Streets et al, 2009, 2011). Estimates of natural Hg emission are poorly constrained and have large uncertainties (±2000 Mg yr−1), limiting the understanding of global and regional Hg cycling budgets In light of the control of anthropogenic Hg emission by the legally binding Minamata Convention (Pacyna et al, 2016), a better quantification of natural Hg emission is critical in evaluating the effectiveness of policy actions (Selin, 2009; Pirrone et al, 2010; Song et al, 2015)

Methods

Results

Conclusion