Eddy covariance and scintillation measurements of atmospheric exchange processes over different types of vegetation

Introduction and objectivesGood comprehension of the energy and mass cycles and their effect on climate dynamics is crucial to understanding, predicting and anticipating ecological changes due to possible future climate perturbations. Here direct and long-term flux density measurements of greenhouse gases from various ecosystems provide means to supply such fundamental knowledge. For the global water vapour and carbon cycles, however, the interactions between different spatial scales become important, where extrapolating from canopy flux density measurements to global budgets lead to practical and theoretical problems. This thesis focuses on the direct and long-term measurement of surface flux densities and interaction processes at the canopy (< 1 km scale within the framework of the Surface Layer Integration Measurements and Modelling (SLIMM) project. Furthermore, some characteristics and limitations of the scintillation technique are studied in two field experiments in New Zealand.As indicated in Chapter 1, the first objective of this project was the direct and continuous long-term measurement of the surface flux densities of radiation, momentum, heat, water vapour and carbon dioxide (CO 2 ) to study the effect of biological and climatic processes that regulate carbon dioxide exchange of this ecosystem at the canopy scale. At the same time these data were used to study the effect of plant related and environmental conditions on the interaction of carbon dioxide and water vapour exchange, to satisfy the second objective of the thesis. The third objective focussed on the prospect of obtaining both the spatial averaged sensible heat flux density and momentum flux density from scintillation measurements.Generally, a compromising point measurement of the mean horizontal wind speed or friction velocity is used to calculate the sensible heat flux density from the temperature structure parameter. By using two scintillometers at two heights, point measurements to obtain the atmospheric stability can be omitted. The fourth objective of this thesis was to study the influence of absorption fluctuations on the average sensible heat flux density derived from the scintillation technique.Carbon dioxide exchange and the effect of biological and climatic processesCarbon dioxide exchange was measured, using the eddy covariance technique, during a one and a half-year period in 1994 and 1995. The measurements took place over a former true raised bog, characterised by a shallow peat layer and tussock vegetation dominated by Molinia caerulea . Peat soils in the Northern Hemisphere's wetlands contain about one third of the worlds carbon pool. Many regions in the arctic tundra, however, have changed from sinks to sources for CO 2 over the past decade but this can not simply be generalised.The growing season extended from May until late October, with a maximum LAI in August of 1.7. The carbon balance showed a net release of 97 g CO 2 m -2y -1(265 kg C ha -1y -1) from the peat bog ecosystem to the atmosphere. During June, July and August there was net consumption of CO 2 , while during the rest of the year there was net production of CO 2 . The maximum daytime net exchange rates were about -0.5 mg CO 2 m -2s -1(-11.3μmol CO 2 m -2s -1) with an average peak exchange rate of -0.2 mg CO 2 m -2s -1(-4.5μmol CO 2 m -2s -1), in a period where the LAI ranged between 1 and 1.7. A high vapour pressure deficit (>15 hPa) corresponding with high temperature was found to reduce the net CO 2 exchange rate by on average 50%.Apart from these factors, LAI and the soil temperature co-determined the net exchange of CO 2 . The total nocturnal respiration during the growing season was within the same order as the average daytime net photosynthetic rate. Temperature was found to be the main factor controlling soil respiration, with a Q 10 of 4.8.The effect of plant related and environmental conditions on the interaction of CO 2 and H 2 O exchangeThe tussock grassland, dominated by Molinea caerulea , was covered with a dense layer of dead organic material from the previous growing seasons. During the summer months, the daytime carbon dioxide uptake often showed a single early morning maximum and a decline in uptake during the rest of the day. Surprisingly, maximum water vapour flux densities were not greatly reduced. The surface cover and the small value of the leaf area index were the main reasons for this phenomenon.The layer of dead organic material acted as an insulating blanket to the transport of water vapour from the soil to the atmosphere. Furthermore, the canopy was far from closed with a peak leaf area index of 1.7 in early August. For both low vapour pressure deficit (< 15 hPa) and high vapour pressure deficit (> 20 hPa) at high surface temperatures, the vegetation showed similar behaviour resulting in a clear reduction of the daytime CO 2 uptake. Temperature was therefore inferred to be main the reason for a reduction in CO 2 exchange. The response of the stomata to atmospheric humidity was deduced to be small possibly due to the abundant availability of soil water. Instead transpiration increased with increasing vapour pressure deficit. The latter was stimulated by the surface temperature, which often exceeded the optimum temperature for photosynthesis and led to an increase in the atmospheric evaporative demand.The scintillation techniqueAn optical or electromagnetic wave propagating through a turbulent atmosphere exhibits fluctuations in intensity known as 'scintillations'. In atmospheric turbulence, fluctuations in temperature, humidity and pressure cause density fluctuation and with it fluctuations in the refractive index ( n ). These refractive index fluctuations cause random refraction and absorption of electromagnetic (EM) radiation passing through the turbulent atmosphere, changing the characteristics of the wave. Scintillation of light is related to these phenomena and is experienced at a receiver as fluctuations in the light intensity caused by interference of refracted light and absorption of the light. Scintillometers measure the turbulent intensity of the refractive index fluctuations of the air from the intensity fluctuations of a received signal expressed in the refractive index structure parameter, C n2.The measured C n2value is related to the structure parameters of temperature C T2, humidity C Q2and a covariant term C TQ , respectively. To calculate the sensible heat flux density from C n2compromising point measurements of the Bowen ratio,β, and friction velocity, u * , are necessary. Generally, the deficiency in the available spatial measurement of u * is overcome by using a point measure of the average wind speed, u and surface roughness, z 0 , but the necessity forβoften remains unresolved.By using two scintillometers at different heights above the surface, a spatial measurement of the Obukhov length, L o , and u * can be derived without incorporating compromising point measurements of the friction velocity or alternatively the average wind speed combined with a measure of the roughness length. The presumption that such measurements are representative of the entire transect usually holds for homogeneous surface cover but may not be valid for patchwork terrain. The two-scintillometer technique is referred to as the C T2-profile method.Refraction is the result of normal and anomalous dispersion. If, however, the frequency of the emitted EM wave is close to a resonance frequency (absorption lines) of atmospheric constituents, like water vapour and carbon dioxide, absorption becomes important. To quantitatively describe the combined effect of refraction and absorption, a complex refractive index structure parameter, C n2, is introduced. Here the phenomenon of absorption is represented by the imaginary part of the refractive index and is solely determined by single absorption lines and their corresponding absorption coefficients (β i ), resulting in a total absorption coefficient for a band of lines (Hill et al. , 1980). The absorption line strength is temperature dependent, while the absorption line width is temperature, humidity and pressure dependent.The contribution of absorption fluctuations to C n2is generally neglected, that means to have a real component only. In reality C n2includes both a real part, C nR2, due to refraction and an imaginary part, C nI2, attributable to the absorption mechanism. Any additional source of scintillation such as a contribution from absorption fluctuations could conceivably corrupt the estimation of the sensible heat flux.Measuring sensible heat flux density over pasture using the C T2- profile methodTwo large aperture scintillometers were positioned at heights ( z ) of 10 and 1.5 m with beams propagating horizontally over pasture for distances of 3.1 km and 141 m respectively. From each scintillometer a half-hourly average value of the path-averaged, temperature structure parameter ( C T2) was obtained in unstable atmospheric conditions. The result suggested C T2to scale with height as z -2/3. Using the C T2- profile method, a path averaged measure of the Obukhov length ( L o ) was calculated for each half hour period. L o was used to determine the friction velocity and the surface layer temperature scaling parameter, T * . The scintillometer sensible heat flux density, H sc , was then calculated from H sc = -ρC p u *T * . A time series of half-hourly averaged H sc compared to H ec obtained by the eddy covariance method agreed to within 10%, with R 2= 0.67, for a range of unstable conditions (-0.2≤( z/L o )≤-0.01).Using a Large Aperture Scintillometer to measure absorption and refractive index fluctuationsThe contribution of refraction and absorption fluctuations to the measured scintillation were observed for a near-infrared absorption region using a NOAA designed large aperture scintillometer. The logarithm amplitude spectra were shown to decay with a frequency as f-8/3for both the absorption and scattering mechanism. For the absorption mechanism this is in line with similar observations made at microwave and infrared frequencies. However, for finite transmitting and receiving apertures, theory predicts a stronger decay of the scattering mechanism due to aperture averaging. The spectral shape is characterised by a region of low frequency absorption, higher frequency refraction separated by a flattish transition zone. The upper observed corner frequency ( f C2 ), compared well with the calculated values using the measured transverse wind speed ( v ) for a known aperture radius. The lower corner frequency ( f C1 ) position was shown to be sensitive to the ratio of the real and imaginary part of the refractive index structure parameter, ( C nR2/C nI2) 3/8and v . The part of the spectrum associated with the absorption scintillations was observed to be much less than that due to refraction until the evening when decreasing C nR2caused C nR2/C nI2to decrease and absorption to become significant. If absorption is ignored, this may have consequences for calculating nocturnal surface heat flux densities. During unstable, daytime conditions the large aperture scintillometer is most sensitive to refractive scintillations despite having an infrared source transmitting in a lossy atmosphere. But also under these conditions, the low frequency absorption part of the spectrum is observable.