Natural and Anthropogenic Mercury Sources
27
fluxes between 3 and 20 ng m -2 h -, from the lakes to the overlying atmosphere.
On one occasion, a net deposition was observed, but all of these values must be
questioned because those measurements were hampered by blank problems
(Kim and Lindberg 1995). It was also found that fluxes during daytime were
larger than at night, indicating that sunlight, biological activity, and temperature
might play an important role in the Hg volatilization process.The authors were
also able to demonstrate pronounced seasonal differences in the Hg volatilization
rate, with fluxes in May and June being much larger than in November. This
phenomenon could be due either to temperature alone or to other seasonal cycles
associated with it. More recently, methods have been developed which allow Hg
fluxes to be measured directly over water surfaces with the micro meteorological
modified Bowen ratio gradient approach (Lindberg et al. 1996). During June,
1994, the first measurements of Hg vapor fluxes over a boreal forest lake were
made at Lake Gardsjon, Sweden. Using highly accurate methods with multiple
replicate samplers, the authors measured concentration gradients of Hg vapor,
CO2, and H20 over the lake surface. Mercury was found to be readily emitted
from the lake surface, and there was no evidence of Hg(o) dry deposition to the
lake surface. Emission rates over the lake averaged 8.5 ng m- 2 h-', and appeared
to be weakly influenced by water temperature and solar radiation. Overall, the
fluxes ranged from -2 to 18 ng m -2 h -', comparable to fluxes previously
measured using surface chambers as discussed above. Overall, the surface water
of the lake appears to be a more active zone for Hg exchange than the
surrounding soils based on two independent studies (Xiao et al. 1991; Lindberg
et al. 1998).
All other studies on Hg emission from lakes use calculations and models to
estimate Hg fluxes from measured concentrations of atmospheric mercury and
DGM in the water phase. It is then assumed that all gaseous Hg in both the
waterphase and the atmosphere is Hg( 0) - which is a reasonable initial
assumption because other volatile Hg species in the ambient atmosphere or in
water occur only at very low concentrations (e.g. Stratton and Lindberg 1995;
Bloom et al. 1996 a,b) - and that this Hg(o) partitions between water and
atmosphere according to Henry's law. With regard to this, most surface lake
waters seem to be supersaturated in Hg(o), which is assumed to give rise to an
evasional flux. The difference between the concentrations of Hg(o) measured
in the lake water and back calculated from the atmospheric concentration is fit
into a thin film model and combined with wind speed and other meteorological parameters to estimate the Hg flux from the lake to the atmosphere. It
is evident that this calculation is based on pure phase thermodynamic
properties and therefore unlikely to reflect reality in all natural ecosystems.
Although, it is undetermined for many systems to what extent theoretical
calculations and experimentally determined fluxes disagree, one study in a
contaminated lake experimentally measured both DGM and fluxes simultaneously (Lindberg et al. 1997). This study found that measured fluxes exceeded
those modeled from DGM using reported values of the exchange coefficients in
other lakes.
The majority of studies on mercury exchange between lakes and the
atmosphere have been conducted using modeling approaches in a set of seven
27
fluxes between 3 and 20 ng m -2 h -, from the lakes to the overlying atmosphere.
On one occasion, a net deposition was observed, but all of these values must be
questioned because those measurements were hampered by blank problems
(Kim and Lindberg 1995). It was also found that fluxes during daytime were
larger than at night, indicating that sunlight, biological activity, and temperature
might play an important role in the Hg volatilization process.The authors were
also able to demonstrate pronounced seasonal differences in the Hg volatilization
rate, with fluxes in May and June being much larger than in November. This
phenomenon could be due either to temperature alone or to other seasonal cycles
associated with it. More recently, methods have been developed which allow Hg
fluxes to be measured directly over water surfaces with the micro meteorological
modified Bowen ratio gradient approach (Lindberg et al. 1996). During June,
1994, the first measurements of Hg vapor fluxes over a boreal forest lake were
made at Lake Gardsjon, Sweden. Using highly accurate methods with multiple
replicate samplers, the authors measured concentration gradients of Hg vapor,
CO2, and H20 over the lake surface. Mercury was found to be readily emitted
from the lake surface, and there was no evidence of Hg(o) dry deposition to the
lake surface. Emission rates over the lake averaged 8.5 ng m- 2 h-', and appeared
to be weakly influenced by water temperature and solar radiation. Overall, the
fluxes ranged from -2 to 18 ng m -2 h -', comparable to fluxes previously
measured using surface chambers as discussed above. Overall, the surface water
of the lake appears to be a more active zone for Hg exchange than the
surrounding soils based on two independent studies (Xiao et al. 1991; Lindberg
et al. 1998).
All other studies on Hg emission from lakes use calculations and models to
estimate Hg fluxes from measured concentrations of atmospheric mercury and
DGM in the water phase. It is then assumed that all gaseous Hg in both the
waterphase and the atmosphere is Hg( 0) - which is a reasonable initial
assumption because other volatile Hg species in the ambient atmosphere or in
water occur only at very low concentrations (e.g. Stratton and Lindberg 1995;
Bloom et al. 1996 a,b) - and that this Hg(o) partitions between water and
atmosphere according to Henry's law. With regard to this, most surface lake
waters seem to be supersaturated in Hg(o), which is assumed to give rise to an
evasional flux. The difference between the concentrations of Hg(o) measured
in the lake water and back calculated from the atmospheric concentration is fit
into a thin film model and combined with wind speed and other meteorological parameters to estimate the Hg flux from the lake to the atmosphere. It
is evident that this calculation is based on pure phase thermodynamic
properties and therefore unlikely to reflect reality in all natural ecosystems.
Although, it is undetermined for many systems to what extent theoretical
calculations and experimentally determined fluxes disagree, one study in a
contaminated lake experimentally measured both DGM and fluxes simultaneously (Lindberg et al. 1997). This study found that measured fluxes exceeded
those modeled from DGM using reported values of the exchange coefficients in
other lakes.
The majority of studies on mercury exchange between lakes and the
atmosphere have been conducted using modeling approaches in a set of seven