3.1 Energy Balance Closure
The eddy covariance data is averaged to daily time steps then the
closure error percentage is calculated per day as:
\(\%\ closure\ error=\frac{\left(Rn-G\right)-(LE+H)}{\text{Rn}}*100\)Eq. 2
This daily closure error was then averaged to monthly closure error
(Figure 4) for seasonal analysis. The closure error is generally smaller
in spring and summer months than in winter months. Months with negative
closure mean the combined latent, ground, and sensible heat fluxes are
greater than the radiation possibly resulting from the saturated nature
of the area resulting in large latent heat fluxes despite lower
radiation. The month with the least closure error is March 2018 followed
by August 2017, with closures of 0.163% and 3.39%, respectively. In
March 2018, net radiation is low, but also, the heat fluxes are small
contributing to a small closure percentage. The winter of 2018 was also
a dry year with minimal snow, which could also contribute to this
smaller closure error. When the ground is saturated it is possible that
some of the energy penetrating the ground surface is used to heat water
that runs off into the stream leaving some energy fluxes unaccounted for
by the tower. Once the ground is no longer saturated and the snowpack is
gone, incoming energy can contribute more directly to the heat flux
terms in Eq. 2.
These results also show that summer closure is better, on average, than
fall and winter closure. In winter, the site is covered in snow and a
comparatively stable planetary boundary layer (PBL) persists, which
creates smaller eddies and laminar flow which makes it difficult for the
sensor to measure heat fluxes (Baldocchi, Hinks, & Meyers, 1988;
Eshonkulov et al., 2019). A more stable PBL and more laminar flow regime
may also be resulting in a much greater flux measurement sampling area
or “footprint” which may come from hillslope or other non-riparian
areas. Prevailing winds are also much stronger during the winter time,
particularly during storm events, which may also contribute to
enlargement of the flux tower footprint. Lastly, we are not explicitly
accounting for the thermal energy stored in the snowpack, nor within
standing water under the tower, which would be absorbing (or releasing)
energy and is therefore unaccounted for by the tower measurements.
Combined, the closure error results show that the energy flux estimates
are less reliable in the winter as the closure error is much larger and
all of the energy in the system is not accounted for by the heat fluxes
or there is more energy in the form of heat fluxes than is accounted for
by incoming net radiation. Therefore, from here on, our analyses focus
predominantly on the warm season flux characteristics.
To better understand the daily cycle of energy components during the
time of good energy balance closure, we averaged the net radiation,
latent heat, sensible heat, ground heat, and closure error percentage at
each hour of the day from June 1 to August 31 for all years. Solar noon
is also shown (Figure 5). Net radiation is greatest at 14:00 MT (local
time) with sensible heat reaching its maximum at the same time. Latent
heat reaches its maximum slightly later due to the delay in energy
transfer as this system seems to be more energy-limited. Ground heat
takes the longest to reach its peak as the heat transfer from the
atmosphere to the subsurface takes the most time. The diurnal data also
shows the decrease of closure error throughout the day with the best
closure occurring at 18:00 around the same time as the ground heat flux
maximum. Closure error is larger in the evening and early morning when
the atmosphere tends to be more stable and eddies are smaller resulting
in more difficult measurements of sensible and latent heat fluxes. The
eddy covariance tower behavior is reasonable for this riparian area,
thus giving us confidence in the flux tower data.