Impact of time and spatial averages on the energy balance closure
- Secondary circulations are large and relatively stationary eddies, which are caused by the surface heterogeneity and normally reside away from the ground. They are believed to be the cause of the energy balance closure problem at the earth's surface, because their contribution to the turbulent fluxes is missed by a fixed eddy-covariance tower measurement that has a typical averaging time of 30 minutes. In this thesis, data from the LITFASS-2003 experiment was used to investigate the impact of time and spatial averages on the energy balance closure. This data consisted of many observations over a large heterogeneous landscape that could generate secondary circulations; some of which might be still near the earth's surface.
For the time average analysis, the averaging time was extended to increase the possibility that secondary circulations were picked up by the sensor. Two approaches, which were the modified ogive analysis and the block ensemble average, were applied to analyze the data from ground-based measurements. The modified ogive analysis requiring a steady state condition, could extend the averaging time up to a few hours and suggested that an averaging time of 30 minutes was still overall sufficient for the eddy-covariance measurement over low vegetation. The block ensemble average, on the contrary, did not require a steady state condition, but could extend the averaging time to several days. However, this approach could only improve the energy balance closure for some sites during specific periods, when secondary circulations existed in the vicinity of the sensor. Based on this approach, it was found that the near-surface secondary circulations mainly transported sensible heat, which led to an alternative energy balance correction by the buoyancy flux ratio approach, in which the attribution of the residual depended on the relative contribution of the sensible heat flux to the buoyancy flux. The fraction of the residual attributed to the sensible heat flux by this energy balance correction was larger than in the energy balance correction that preserved the Bowen ratio.
In the spatial average analysis, two energy balance correction approaches, the buoyancy flux ratio and the Bowen ratio approaches, were applied to the area-averaged fluxes (composite fluxes) in order to include contribution from secondary circulations. These composite fluxes were aggregated from multiple ground-based measurements. The energy balance corrected fluxes were validated against the spatial average fluxes, which were measured by an aircraft and a large aperture scintillometer (LAS). In this validation, the backward Lagrangian footprint model was used to estimate the source area of the measurement. It was found that both energy balance correction approaches did improve the agreement between time and spatial averages fluxes. This suggested that the contribution from secondary circulations could be properly accounted by the energy balance correction.
All findings in this thesis, therefore, depict that secondary circulations significantly transport energy in the atmospheric surface layer. The energy balance correction, accomplished by using either the Bowen ratio approach or the buoyancy flux ratio approach, is necessary to estimate the actual vertical transport of energy at the earth's surface.
Documentation of the EVENT-HMMS Experiment 2012 – Microclimatological effects of rain-out shelters within EVENT II
- no abstract
Free convection and turbulent fluxes over complex terrain
- The impact of complex terrain on the land-atmosphere exchange is investigated in this thesis. Here, free convection, a very effective vertical transport mechanism as turbulence is predominantly driven by buoyant forces, is explicitly addressed. Recently, it was shown for certain situations over complex terrain that free convective injections of surface layer air masses into the atmospheric boundary layer (ABL) can alter the ABL properties significantly. This study aims at the general identification and description of such situations of near-ground free convection conditions (FCCs) over complex terrain. For this purpose, data obtained during the COPS (Convective and Orographically induced Precipitation Study) field campaign in summer 2007 were used. Within this project, several surface flux measurement stations were installed, mainly in valleys and on mountaintops of the Black Forest, southwestern Germany. Turbulent fluxes were calculated with the eddy-covariance (EC) method and were used to detect FCCs with the help of a stability parameter. The flux measurements were further combined with ABL profiling measurements (Sodar/RASS) and a large-eddy simulation (LES) model in order to investigate the impact of FCCs on ABL properties. The effect of complex terrain on the energy balance closure and on spatial and temporal flux differences was also studied with these flux data.
FCCs were detected on about 25% of the days during the three month COPS experiment. In situations of weak synoptic forcing, thermally driven orographic (e.g. valley winds) or local wind systems developed over the complex terrain due to heating differences. During the adaption of these wind systems to changing heating differences (e.g. during the reversal of the valley wind from down- to up-valley winds in the morning), the horizontal wind vanished. If, at the same time, the buoyancy flux was positive and enhanced, buoyant forces exceeded the usually prevailing shear forces in the surface layer and FCCs were detected. Moreover, it was demonstrated that FCCs are not restricted to the COPS region. Also, a data set of Nam Co station on the Tibetan Plateau showed FCCs during the reversal of a thermally driven land-lake breeze. However, at this high-altitude site, FCCs were more often detected in the afternoon compared to the COPS region due to the frequent change of heating differences during cloud cover periods.
The Sodar/RASS as well as the LES model showed the presence of coherent updraft
structures in the developing early-morning convective boundary layer (CBL) in the Kinzig valley (Black Forest) during FCCs. Spectral analysis of the EC data in these situations indicated the existence of large-eddy turbulent scales – typical for thermal updrafts in the CBL – already close to the ground. An ensemble and time mean analysis of the simulated flow field in the valley further confirmed that the Sodar/RASS was located preferably in an updraft region during FCCs. In a CBL over flat homogeneous terrain, the locations of convective structures would occur randomly. However, over the complex orography of the Kinzig valley, the updraft structures were found to develop in
quasi-stationary patterns at specific locations relative to the surrounding mountain ridges. The model further showed that the flux through the valley boundary layer is mainly determined by the flux within these coherent updrafts. In combination with the Sodar/RASS observations, the model also showed that these updrafts deeply penetrated into the stably stratified valley boundary layer up to approximately the height of the surrounding mountains leading to an effective upward counter-gradient transport of surface layer air mass properties during FCCs.
The analysis of the turbulent fluxes at the different COPS sites showed that the flux values were strongly determined by varying land surface characteristics. Also an increase of the Bowen ratio with increasing altitude could be detected. These findings are in accordance with former studies in this area. As expected, the energy balance was found to be unclosed on average during the entire COPS period, with values of the residual typical for heterogeneous landscapes. However, regarding only the periods with FCCs, no residual occurred on average. This is due to the fact that the landscape
heterogeneity is of minor importance in case of the more vertical oriented exchange regime during FCCs, so that missing advective flux components became strongly reduced in these situations. Moreover, it was found that in comparable periods with no FCCs, flux components were missing with exactly the proportions of the buoyancy flux ratio, thus suggesting a correction of the energy balance according to the buoyancy flux ratio approach. These results support recent publications on the energy balance closure
ExchanGE processes in mountainous Regions (EGER)- Documentation of the Intensive Observation Period (IOP3) June, 13th to July, 26th 2011
Henrique F. Duarte
Advection at a forest site – an updated approach
- The exchange of carbon dioxide (CO2) across the vegetation-atmosphere interface of a spruce forest was investigated. Horizontal and vertical advection are recognized as important terms of the Net Ecosystem Exchange (NEE) budget in addition to the commonly measured turbulent flux and storage flux. Direct advection measurements are challenging because of the instrumental accuracy required to observe small concentration gradients and small wind velocities and because of the spatio-temporal measurement resolution required to observe complex 3-D flow phenomena. This work presents an experimental multi-analyzer setup for the observation of horizontal CO2 concentration gradients with high temporal resolution and good spatial resolution with no tradeoff between the two. A statistical approach was developed to correct for inter-instrument bias by applying a conditional time dependent bias correction. This approach relies on properties of probability density distributions of concentration differences between one sample point and the spatial average of the sample point field. Sub-canopy CO2 concentration gradients observed with the above presented system showed a high spatial variability which was dependent on vegetation structure. Local concentration perturbations correlated with statistical properties of coherent structures and were explained by vertical exchange between CO2 enriched sub-canopy air and low concentration above-canopy air. The small-scale variability of CO2 concentration gradients brings into question the representativity of horizontal advection measurements for the control volume if observed with a low spatial resolution. Vertical advection estimates rely on accurate measurements of vertical wind velocity (w). Different procedures were applied during coordinate rotation to align the coordinate system of the sonic anemometer with the long-term stream lines. Spatial variability of the wind field was addressed by a sector-wise coordinate rotation. An investigation of temporal aspects of vertical wind velocity showed significant contributions from low frequencies in the spectrum of w. The impact of the data set length used for coordinate rotation on w and on vertical advection was investigated and observed to be large. A sequential coordinate rotation with controlled window length was proposed. Advection contributed significantly to NEE during the night and during transition periods at the Waldstein-Weidenbrunnen (DE-Bay) FLUXNET site. Daily NEE budgets were more realistic, compared to NEE from turbulent flux and storage change alone, if direct advection measurements from continuous and bias corrected gradient sampling were included, reducing the estimated daily carbon sequestration of the forest by almost 50 %.
Modeling the exchange of energy and matter within and above a spruce forest with the higher-order closure model ACASA
- Multilayer SVAT-models that contain an advanced turbulence scheme are necessary for the detailed simulation of all relevant exchange processes above and within a forest canopy. The Advanced Canopy-Atmosphere-Soil Algorithm (ACASA) model incorporates such an advanced turbulence scheme, the third-order turbulence closure. This study presents the application of the ACASA model for a spruce forest at the Waldstein-Weidenbrunnen site in the Fichtelgebirge (Germany). The comprehensive micrometeorological and plant physiological measurements performed during the EGER project (ExchanGE processes in mountainous Regions) provided the necessary data base for this purpose, particularly eddy-covariance and sap flux measurements at several heights within the canopy. Thorough model tests were a main focus of this study and led to an improvement of the investigated model. This included both the exploration of the sensitivity and predictive uncertainty of the modeled fluxes and the analysis and correction of model errors that were encountered while working with the model. Furthermore, the ability of the ACASA model to reproduce measured quantities within and above the forest canopy was assessed, with an emphasis on the vertical structure of evapotranspiration and its components. To study the sensitivity and predictive uncertainty of the ACASA model, the Generalized Likelihood Uncertainty Estimation (GLUE) methodology was employed for two five day fair weather periods. Here, the sensitivity of the sensible heat flux, the latent heat flux and the net ecosystem exchange above the forest canopy was assessed. This analysis allowed the identification of influential parameters for the three fluxes. The fluxes were strongly sensitive to only a few parameters while the problem of equifinality was revealed for many parameters. Equifinality is a common problem for complex process-based SVAT-models. The calculated uncertainty bounds showed the ability of the ACASA model to well reproduce the fluxes for two periods with different meteorological conditions. Furthermore, the results of the GLUE analysis indicated weaknesses in the model structure concerning the soil respiration calculations. The latest ACASA version includes multiple improvements in comparison to older model versions which were introduced after a comparison of modeled within- and above-canopy fluxes and turbulence statistics with measurements. The former version of the ACASA model did not explicitly close the energy balance. Rather, an error was included in the model output. This modeled error, however, did not agree with the measured residual at our site and was shown to reach substantial magnitudes depending on the value of the leaf area index. Thus, a method to ensure a closed energy balance for all layers in ACASA was introduced. Measured third-order velocity statistics were largely underestimated by the former ACASA version, which required correcting the calculation algorithms for the third-order moments in the latest ACASA version. Comparisons of third- and second-order velocity statistics showed that simulations of the latest ACASA version were improved but only partly reproduced measurements. Sap flux and eddy-covariance measurements at several heights within the profile provided estimates of all components of evapotranspiration of the forest and its vertical distribution. Canopy transpiration of the stand measured with the eddy-covariance technique delivered larger estimates than measured with the sap flux technique. Possible reasons for this mismatch are discussed, such as a contribution of evaporation from intercepted water that was still present at the beginning of the study period and differences between the eddy-covariance footprint and the area to scale up sap flux measurements. The modeled evapotranspiration components by ACASA compared well to these measurements when taking the uncertainties of these measurements into account. Also, modeled in-canopy profiles of canopy (evapo-) transpiration agreed well with measurements, with a better agreement of mean profiles for daytime, a partly and completely coupled canopy than for nighttime and a decoupled canopy. Largest contributions to canopy (evapo-) transpiration stem from the upper half of the canopy at daytime, whereas during nighttime, the contribution shifted towards lower parts of the canopy. Additionally, model simulations of the 3D model STANDFLUX were included in this study. This study revealed that the ACASA model is a powerful tool to simulate in detail a large range of the relevant exchange processes within and above a spruce forest site. At the same time existing weaknesses in the model code were identified that should be improved in future ACASA versions.
Transport Processes of Reactive Trace Gases in the Atmospheric Boundary Layer
- Transport of trace gases within the atmospheric boundary layer plays a key role in feedback processes between the earth’s surface and the atmosphere and consequently in ecosystem budgets of carbon and nitrogen (among many more). For a correct quantification of the exchange between surface and atmosphere, it is crucial to understand the transport processes involved and to determine limitations of the presently available measurement techniques in order to apply the right technique with respect to the currently active transport processes. This dissertation focuses on three topics: (a) The analysis of effects of vertical transport mechanisms on surface measurements of trace gases, (b) the appropriate choice of an experimental setup to assess specific measurement errors of moving measurement systems and (c) the application of a series of measurement techniques for surface fluxes of reactive trace gases to determine their degree of agreement and to assess potential source of deviations. To study the impact of vertical transport mechanisms on surface trace gas measurements, this thesis presents a comprehensive set of measurements at the surface and within the atmospheric boundary layer (by tethered balloon). It enables the attribution of a recurrent negative excursion of ozone mixing ratios in the morning hours at a mountain summit to a very efficient vertical transport by free convection. It has been shown that, due to the rapid vertical transport, a layer of approximately 20 m thickness developed at the equilibrium height of the free convection, being located within the residual layer. It had a chemical composition similar to the air close to the ground while being surrounded by residual layer air masses. Hence, very strong gradients of the chemical composition were found within the residual layer. Evidence was found, that such a transport occurs rather frequently at this location, affecting at least 18 % of the days between April and September. To assess measurement errors introduced by the application of scanning methods as compared to gradient approaches, a higher temporal resolution of the vertical profiles was needed. Because of limitation inherent to a tethered balloon, an elevator based profiling system was installed, providing a temporal resolution of 10 minutes with a maximum ceiling of 100 m. Prior to the investigation of transport processes, the proper functioning of correction algorithms for the so-called dynamical error was investigated under real atmospheric conditions. This dynamical error is inherent to all moving measurement systems and arises from the non-zero response time of the deployed sensors. It has been shown that existing algorithms as well as one developed by the authors reliably balance the dynamical error. Furthermore it has been demonstrated, that the elevator data correlate with reference data at fixed levels with coefficients of determination being always greater than 0.992 at every level (10, 20, 40, 60, 80, 98 m). To evaluate the applicability of different flux measurement techniques for the determination of surface fluxes of reactive trace gases, three different approaches were compared. In order to determine surface fluxes of trace gases, a new modification of the modified Bowen ratio method was used. In this modification, the measurements of sensible heat flux and of the gradients were horizontally separated. This allowed the simultaneous measurement of the fluxes of various trace gases without creating errors due to flow distortion by bulky inlet systems. Surface emission fluxes of nitric oxide were found to be in the range 0.02 – 0.15 nmol m-2 s-1 (night/day), nitrogen dioxide fluxes varied around 0.1 nmol m-2 s-1 (deposition) with slightly positive values in the early afternoon, indicating emission. Ozone deposition fluxes ranged from close to zero to about 6 nmol m 2 s-1. A laboratory parameterization of biogenic soil emission fluxes of nitric oxide from incubated soil samples yielded values from 0.025 nmol m-2 s-1 to 0.12 nmol m-2 s-1 for environmental conditions encountered during the field campaign. This was in excellent agreement with the fluxes from field observations. Besides the comparison of field fluxes with laboratory data, a case study (1 night) comparison of carbon dioxide and ozone fluxes between two field methods was done. Results from the modified Bowen ratio method have been compared to fluxes derived from the integral boundary layer budget method. Both methods yielded similar mean carbon dioxide fluxes during the night. In contrast, ozone fluxes deviated between both methods. This deviation was attributed to chemical in-situ loss of ozone during night time within the profile being integrated by the budget method.
CO2 and Isotope Flux Measurements above a Spruce Forest
- The measurement of the turbulent carbon dioxide (CO2) exchange by the eddy covariance (EC) method has become a fundamental tool for the quantitative determination of the atmospheric CO2 net ecosystem exchange (NEE) and the investigation of the carbon mass balances of ecosystems. Such measurements require a high degree of quality control in order to prevent systematic errors. The determination of the annual sum of NEE and filling of data gaps is complicated by characteristic diurnal and seasonal variation in the governing gross flux components of assimilation, i.e. photosynthetic uptake of CO2, and respiration. In this dissertation, a set of criteria is suggested for the identification of high quality NEE data. They are applied to data obtained above a spruce forest in the Fichtelgebirge Mountains in Germany. The application of the quality criteria resulted in less systematic distribution of data gaps compared to a commonly applied criterion based on the friction velocity u-star measured above the canopy. The suggested method is therefore able to reduce the risk of double accounting of nighttime respiration fluxes and systematic error in the annual sum of NEE. The isotopic flux partitioning method can be applied to quantify the assimilation and respiration flux components. Especially above forest ecosystems, it requires isotope flux measurements with high analytical precision in order to resolve small gradients in the isotopic signature of the turbulent exchange. A conditional sampling instrument was developed and tested in laboratory and field experiments. By combining the hyperbolic relaxed eddy accumulation method (HREA), whole-air sampling and high precision isotope ratio mass spectrometry (IRMS), 13CO2 and CO18O isotopic flux densities (isofluxes) could be measured with an estimated uncertainty of 10-20% during a three day intensive measuring campaign of the field experiment WALDATEM-2003 (Wavelet Detection and Atmospheric Turbulent Exchange Measurements 2003). Thorough quality control was applied at all stages of the experiment, including the data evaluation. The sampling process and the assumption of similarity in the turbulent exchange characteristics of different scalars (scalar similarity) were assessed by simulation of HREA sampling based on high temporal resolution data of the turbulent energy and gas exchange. Above three different vegetation types, distinct diurnal changes of scalar similarity were observed and attributed to events on time scales longer than 60 s, which most likely represent changes in the source/sink strength or convective or advective processes. Poor scalar-scalar correlations indicate the risk of systematic underestimation of fluxes measured by HREA. There is some evidence for good scalar similarity and a generally linear relation between bulk CO2 mixing ratios and its isotopic signatures in the turbulent exchange. However, the slope of that relation was observed to change temporarily so that especially for the EC/flask method temporal and spatial scales represented in flask samples must carefully be considered. HREA isoflux measurements have a footprint similar to the footprint of EC measurements and are therefore able to integrate small-scale heterogeneity in ecosystems. CO2 mixing ratios and delta-13C and delta-18O isotopic signatures measured in updraft and downdraft whole-air samples allowed determining ecosystem integrated and truly flux weighted isotopic signatures of the atmospheric ecosystem gas exchange and ecosystem isotope discrimination Delta-e and Delta-E on half-hourly timescales. The observed diurnal variability demonstrates the need for their repeated high precision measurement at ecosystem scale for the evaluation of isotopic mass balances. For the isotopic flux partitioning method, additional data on the integrated canopy isotope discrimination Delta´-canopy from independent measurements or validated models is indispensable. An observed fast equilibration of isotopic disequilibria D13C and D18O between the assimilation and respiration fluxes may indicate that the successful application of the isotopic flux partitioning method is limited to short periods after significant environmental changes on the scale of few days.
Quality Assurance for Eddy Covariance Measurements of Turbulent Fluxes and its Influence on the Energy Balance Closure Problem
- The eddy covariance method enables direct measurements of turbulent fluxes at the earth’s surface. Such measurements are required to study the surface energy balance and the exchange of gaseous air constituents. Due to improvements in the construction of adequate sensors and the progress in computer technology during the last decades this method is now well-established. However, a general failure to close the energy balance equation has been reported for many micrometeorological field experiments. This unresolved problem motivated this dissertation, which aims at the quality assurance for eddy covariance measurements. The presented concept for quality assurance comprises investigations on the accuracy of the deployed sensors and on the impact of the data analysis for such measurements. A specific focus is set on possible implications for the determination of reliable CO2 flux estimates, since this issue gained importance during the last years for studies on the global carbon cycle related to global warming due to the green house effect. Data from several field experiments in Germany, California and Nigeria form the experimental basis for these investigations. A software package was developed to perform the necessary post processing for all eddy covariance measurements presented in this thesis. The results of sensor intercomparison experiments show a typical random error of eddy covariance measurements of 5% for the sensible heat flux and 10% for the latent heat flux, if sensors are well-calibrated and maintained and the assumptions for this method are fulfilled. The applicability of an objective quality assessment scheme of flux data was demonstrated for large datasets from a field campaign comprising 14 measuring systems. The energy balance closure problem has been studied at two exemplary sites. The energy balance could not be closed for measurements over an agricultural area in Germany. For this experiment the sum of turbulent heat fluxes was 30% smaller than the available energy at the surface. In contrast, no systematic bias of the energy balance could be found for measurements over fallow bush-land in Nigeria, although the measures of quality assurance were similar. Neither differences in instrumentation nor in the post-field data processing between both experiments can explain these findings. A further analysis of the dataset from the agricultural area in Germany showed that additional flux contributions can be found when extending the averaging time of covariances beyond the conventional 30 minute interval. The energy balance can even be closed for this site when applying an averaging time of 24 hours. Longwave flux contributions seem to be generated here by the much stronger heterogeneity of the surrounding terrain compared to the more or less homogeneous environment of the Nigerian site. The filtering of heterogeneity induced flux contributions from very low frequency covariances through the commonly used averaging times of less than 30 minutes is identified as major reason of the energy balance closure problem. To improve the understanding of the processes leading to low frequency flux contributions a more detailed analysis of further experiments in combination with large eddy simulation modelling are required. The impact of post-field data processing was not only evaluated for energy flux estimates but also for fluxes of CO2, which showed similar additional flux contributions for extended averaging times. Finally, it was demonstrated that the quality assessment scheme presented in this thesis provides a fundamental and robust rejection criterion for a successful gap-filling strategy to determine annual sums of CO2 net ecosystem exchange.
Detection and Analysis of Coherent Structures within and above Tall-vegetated Canopies
- Coherent structures are an inherent phenomenon of the atmospheric turbulent flow in the proximity to tall-vegetated canopies. Although coherent structures have called increasing attention of the turbulence community during the past decades, the basic mechanisms of their emergence, their contributions to exchange processes and their importance for conventional flux determination methods remain poorly understood or even unknown. In particular, no studies have been published yet dealing with long-term observations rather than exemplary case studies using short-term data. This dissertation aims at the enhanced understanding of the driving mechanisms and statistical properties of coherent structures within and above tall-vegetated canopies through an extensive analysis using long-term observations. Thereby, it places a specific emphasis on the implications of coherent structures for exchange processes and assesses their impact on conventional flux determination methods such as the eddy covariance and relaxed eddy accumulation techniques. Data were mainly obtained using tower-based single-point turbulence measurements and acoustic remote sensing technique (Sodar-Rass) during extensive field campaigns conducted in summer 2003. The developed wavelet software tool for detection and analysis of coherent structures is verified to extract coherent structures objectively under varying environmental conditions and thus allows determining their statistics in long-term datasets. In the proximity to the plant canopy, the temporal scales of coherent structures typically range between 20 s and 35 s. The temporal scales of coherent structures in the horizontal wind velocity, the sonic temperature and the concentration of carbon dioxide and water vapour exceed those of the vertical wind velocity. Within the canopy, the temporal scales of all vector and scalar variables collapse at approx. 24 s to 28 s resulting in an enhanced symmetry. Besides this, coherent structures with temporal scales up to 220 s are evidenced well above the canopy with the aid of acoustic remote sensing. The application of the canopy mixing-layer analogy to the data partially yields departures of the ratio m between the streamwise structures spacing of coherent structures and the canopy shear scale from the prediction m = 7...10. The departures are due to the influence of the terrain affecting the shape of the canopy wind profile and therefore the vertical wind shear. An agreement is found for flows which are forced to reorganise downstream of flow obstacles. The vertical wind shear is identified as the main driving force from which coherent structures emerge close to the canopy. In the layer well above the canopy diabatic processes facilitate the generation of coherent structures of large temporal scales. Clearcuts in a fairly homogeneous canopy cause additional structures in the turbulent flow with large temporal scales. Coherent structures contribute about 16 % to total the momentum transfer and about 26 % to the total fluxes of buoyancy, carbon dioxide and latent heat. A scheme for the qualitative classification of exchange regimes between the atmosphere and the canopy is developed analysing the ejection and sweep phases of coherent structures along the vertical profile in the canopy. The presence of coherent structures causes flux errors in the eddy covariance method below 4 %. The effect of this flux error for long-term observations is negligible as individual flux errors average out. Coherent structures of large temporal scales significantly influence the scalar similarity required for the relaxed eddy accumulation technique. These flow structures are responsible for the diurnal changes of the scalar similarity observed in the traces of acoustic temperature and concentration of carbon dioxide and water vapour.