Exchange Processes between a Coniferous Forest and the Atmosphere

This thesis deals with the research question: which processes are relevant in controlling the exchange fluxes between the forest and the atmosphere and how can this control be quantified? Answering this question is relevant for research in the fields of air pollution, weather and climate and remote sensing. To answer this question a measurement program has been performed over and in a dense Douglas fir forest (Speulderbos), near the village of Garderen, the Netherlands. Variables were monitored related to, the state of the atmosphere, the state of soil water and the state of the trees. Forest response was determined by measuring outgoing radiation fluxes and the surface fluxes of momentum, sensible and latent heat. Well-known concepts from micro-meteorology and hydrology were used for the interpretation of the measurement. At times, these concepts had to be adapted for the specific forest situation. One concept is surface layer similarity theory, which enables to categorise observations in a convenient way. A second concept, related to surface layer similarity, is on resistances relating spatial differences in variables to their corresponding fluxes. A third concept is the Penman-Monteith equation, which enables the discrimination between atmospheric control and plant control on transpiration.

Turbulent exchange within the roughness sublayer is investigated. The importance of terrain inhomogeneity is studied with footprint analysis and with an inhomogeneous surface layer model. A windspeed dependence of roughness length for momentum and displacement height is found. Surface layer similarity theory is extended to describe the roughness-sublayer influence. This involves the introduction of an additional length scale related to the geometry of the forest. It is found that well- defined flux profile relations exist for momentum and sensible heat in the roughness layer of the current forest. In the roughness layer the exchange of temperature is more efficient than the exchange of momentum. This is in contrast to results for the surface layer, but in concord with previous findings for dense forests.

The scalar excess resistance, which describes the difference between momentum transport and scalar transport at the surface/atmosphere interface, is investigated by using measurements of infrared surface temperature. Surface radiation temperature and aerodynamic surface temperature, obtained by extrapolating the air-temperature profile to the surface, are not necessarily equal. By assuming equality between the two, it is shown that a consistent description of the relation between sensible heat flux and temperature difference between the surface and the atmosphere is obtained, at least for daytime cases. The excess resistance for the current forest is much smaller than values found for low vegetation. It is shown that the enhanced exchange efficiency of heat, relative to momentum, in the roughness sublayer attributes to this low value. An alternative analysis is presented to separate this roughness layer effect from the transfer resistance at the forest/atmosphere interface. The value found for this alternative excess resistance, is more in line with low vegetation values. For neutral cases the two methods give the same results for temperature differences between the surface and the surface layer. It is shown that stability effects give rise to a discrepancy between the two methods. The observations show some evidence in favour of the alternative method. The difference between forest interior air temperature and air temperature at canopy height is related to storage heat flux and sensible heat flux by applying the concepts of gust penetration and surface renewal. The analysis suggests that the renewal of interior air caused by gust penetration is slow due to the presence of a very dense crown layer.

For night-time cases, the equality between aerodynamic and radiation surface temperature breaks down when wind speeds are low and longwave cooling is high. The analysis shows that forest air becomes decoupled from the air aloft. Longwave cooling at the crown layer triggers canopy convection which transports cooled air from the crown layer to the forest interior. The existence of a convective surface temperature in the crown layer is deduced from the measurements. A two-layer radiation/energy balance model is constructed. The model explains the difference between radiation- and aerodynamic surface temperature in terms of the distribution of storage heat and sensible heat over the two model layers.

Transpiration for dry conditions is investigated by using the Penman-Monteith equation with a Jarvis type of formulation for the surface resistance. First the closure of the surface energy balance is checked. Overall closure is within the range of estimated measurement error. However, at times deviations occur which can be attributed to wind direction. With respect to transpiration it is found that surface resistance reacts strongly to water vapour deficit changes. This is related to the good aerodynamic coupling of the rough forest to the atmosphere. In spring, a clear increase in transpiration is observed after shoot growth. Soil water response is clearly present before mid summer, after that the forest seems less susceptible to draught. Probably the root system adapts to the dry situation. An analysis of residuals between observed and modelled transpiration shows that deviations occurred at the same wind direction where the energy balance closure broke down. The variance in the residuals appears to be two times larger than estimated from atmospheric statistics. Important contributions to this variance are correlated over periods of one day. This suggests that standard statistical techniques lead to an underestimation of the confidence intervals of estimated model parameters. Two other models are evaluated. A new formulation suggested by Monteith, where stomatal response to moisture deficit is replaced by a response to transpiration itself, is investigated. This formulation appears to be equivalent to the Jarvis formulation with an atmospheric moisture deficit response. The Priestley-Taylor formula is adapted to include soil water response. It performs reasonably well given the simple nature of the formulation.

Interception measurements and xylem sapflow measurements are exploited to investigate the interaction between evaporation and transpiration in a partially wet forest. The Penman-Monteith equation is generalised to describe this interaction. Explicit expressions are obtained for evaporation and transpiration. After optimisation the model is capable of describing both evaporation and transpiration reduction fairly well. Due to parameter interdependency, error bounds on individual parameter estimates, related to evaporation, are large. Independent estimates of the parameters, although crude, are shown to be within the confidence region of the optimisation results. It is shown that evaporation rates are smaller than the frequently used formula: wet fraction times potential evaporation. Most of the transpiration reduction comes from energy consumption by the process of evaporation and the impact of the humidity conditions close to the needles. Only a small amount of stomatal blocking due to intercepted water is needed to explain the remaining reduction. This is in concord with the observation that stomata of the Douglas fir are at the lower side of the needle, which is only partially wetted during rain.

The response of the forest to external forcings can be described by a number of parameters related to model descriptions of various processes. The results for the current dense Douglas-fir forest are compared with other forest studies. A good agreement is found for the relation between the geometric parameters, canopy height, displacement height and roughness length for momentum. The exchange coefficients in the roughness layer for momentum and heat agree qualitatively with typical values found for other dense forests. Quantitatively significant and as yet unexplained differences remain. The very small scalar excess resistance found for the current forest is in agreement with the only other comparable dense forest study, which appeared recently in the literature. Transpiration rates as a function of external conditions are broadly in line with results found at other forest sites in the temperate climate.

New in this thesis are the results on; the estimation of displacement height; the difference between aerodynamic surface temperature and radiation surface temperature at night time; night time convection; changing transpiration response to soil water stress during the season; and the interaction between evaporation and transpiration reduction during wet conditions.