The heat generated by the atmospheric radiative forcing of the surface is dissipated by the turbulent flux and thermal radiation. The partitioning between the two mechanisms is dependent on the surface temperature and the static stability of the near-surface air. The turbulent flux itself is partitioned between sensible (dry) and latent (moist) heat flux; the relative partitioning is strongly controlled by vegetation and soil moisture. Sensible heat flux is relatively less efficient than latent heat flux in dissipating heat; when there are (soil moisture) controls on evapotranpiration, the greater partitioning to sensible heat flux results in a rise of ground temperature. This in turn influences the dissipation of heat by radiative cooling at the surface. Soil capillary action controls the rate at which atmospheric evaporative demand and vegetation roots withdraw water from soil storage. Soil moisture also affects the thermal inertia and shortwave albedo of the surface.
In two-way land-atmosphere interaction, meteorological phenomena both act as
the forcing and react to the forcing by the surface heat and moisture state.
Land surface modification of the lower atmospheric environment and the
atmospheric forcing of these land surface conditions form feedback loops
which are significant factors in modulating the variability of the weather
and climatic systems. Brubaker and Entekhabi [1994] quantify the
magnitude of the individual feedback processes associated with the two-way
interaction between the surface and the atmospheric boundary layer. They
develop a simple analytical model of two-way land-atmosphere interaction and
then perform linear stability analysis on the coupled system in order to
measure the relative strengths of the feedback mechanisms. Brubaker
and Entekhabi [1994] also find that the dynamics of surface heat and
moisture fluxes at the land surface are governed by components with diverse
time scales. Variability in both weather and climate are therefore
influenced by the surface conditions. The predictability and analysis of
fluctuations in the atmospheric environment require the accurate
representation of these feedback mechanisms and two-way land-atmosphere
interaction. For example Yang et al. [1994] test the forecast errors
of five and ten day integrations of a numerical weather prediction model
with various initializations of the initial soil moisture state. They find
that for the five to ten day lead, most of the error is confined to the
lower atmosphere. The air temperature forecast error is reduced from 2.9
C to 1.1
C for the five day and down to 1.3
C for the ten day lead-time forecast when the interactive soil moisture
state is correctly initialized. The forecast errors for the near surface air
relative humidity are reduced from 15% to 7.6% for the five day lead-time
and down to 8.0% for the ten day lead-time forecasts. Anthes et al.
[1989] also conclude that sensible and latent heat fluxes have statistically
significant effects on the forecast skill in regional numerical weather
prediction.