Free convection and turbulent fluxes over complex terrain

FREE CONVECTION AND TURBULENT FLUXES OVER COMPLEX TERRAIN A dissertation submitted to the F ACULTY OF B IOLOGY , C HEMISTRY AND G EOSCIENCES AT THE U NIVERSITY OF B AYREUTH , G ERMANY to attain the academic degree of D R . RER . NAT . presented by R AFAEL E IGENMANN Dipl. Geoökol. born 29 April 1983 in Eichstätt, Germany Bayreuth, January 2013 I FREE CONVECTION AND TURBULENT FLUXES OVER COMPLEX TERRAIN Supervisor: Prof. Dr. Thomas Foken II Die vorliegende Arbeit wurde in der Zeit von August 2008 bis Januar 2013 in Bayreuth an der Abteilung Mikrometeorologie unter Betreuung von Herrn Prof. Dr. Thomas Foken angefertigt. Vollständiger Abdruck der von der Fakultät für Biologie, Chemie und Geowissenschaften der Universität Bayreuth genehmigten Dissertation zur Erlangung des akademischen Grades eines Doktors der Naturwissenschaften (Dr. rer. Nat.). Dissertation eingereicht am: 09.01.2013 Zulassung durch die Prüfungskommission: 16.01.2013 Wissenschaftliches Kolloquium: 24.04.2013 Amtierender Dekan: Prof. Dr. Beate Lohnert Prüfungsausschuss: Prof. Dr. Thomas Foken (Erstgutachter) Prof. Dr. Andreas Held (Zweitgutachter) Prof. Dr. John Tenhunen (Vorsitz) Prof. Dr. Bernd Huwe Prof. Dr. Michael Hauhs C ONTENTS III Contents C ONTENTS ........................................................................................................................III L IST OF MANUSCRIPTS .....................................................................................................IV A CKNOWLEDGEMENTS ..................................................................................................VIII S UMMARY .......................................................................................................................IX Z USAMMENFASSUNG .......................................................................................................XI 1 I NTRODUCTION .......................................................................................................... 1 2 E XPERIMENTAL DATA SETS AND THE LES MODEL ..................................................... 7 2.1 The COPS field campaign................................................................................ 7 2.1.1 Surface energy balance and turbulence network.......................................... 8 2.1.2 Post-processing and quality control of the turbulence data........................ 10 2.2 Data set of Nam Co station............................................................................. 10 2.3 LES model...................................................................................................... 10 3 R ESULTS .................................................................................................................. 11 3.1 Turbulent fluxes and energy balance closure over complex terrain............... 11 3.2 Near-ground free convection conditions (FCCs) over complex terrain......... 14 3.3 Convective structures in the boundary layer during FCCs............................. 17 4 C ONCLUSIONS ......................................................................................................... 22 R EFERENCES .................................................................................................................... 25 L IST OF APPENDICES ........................................................................................................ 36 A PPENDIX A: I NDIVIDUAL CONTRIBUTIONS TO THE JOINT PUBLICATIONS ........................ 37 A PPENDIX B: E IGENMANN ET AL . (2009)......................................................................... 41 A PPENDIX C: E IGENMANN ET AL . (2011)......................................................................... 55 A PPENDIX D: Z HOU ET AL . (2011)................................................................................... 68 A PPENDIX E: B RÖTZ ET AL . (2013).................................................................................. 80 E RKLÄRUNG .................................................................................................................. 100 IV L IST OF MANUSCRIPTS List of manuscripts The thesis is presented in cumulative form consisting of four manuscripts. Three manuscripts have been published in peer-reviewed journals. The fourth manuscript has been submitted for publication to a peer-reviewed journal. Published manuscripts: Eigenmann, R., Metzger, S., Foken, T., 2009. Generation of free convection due to changes of the local circulation system. Atmospheric Chemistry and Physics, 9: 8587–8600. Previously published as: Eigenmann, R., Metzger, S., Foken, T., 2009. Generation of free convection due to changes of the local circulation system. Atmospheric Chemistry and Physics Discussions, 9: 11367–11411. Eigenmann, R., Kalthoff, N., Foken, T., Dorninger, M., Kohler, M., Legain, D., Pigeon, G., Piguet, B., Schüttemeyer, D., Traulle, O., 2011. Surface energy balance and turbulence network during the Convective and Orographically-induced Precipitation Study (COPS). Quarterly Journal of the Royal Meteorological Society, 137: 57–69. Zhou, D., Eigenmann, R., Babel, W., Foken, T., Ma, Y., 2011. Study of near-ground free convection conditions at Nam Co station on the Tibetan Plateau. Theoretical and Applied Climatology, 105: 217–228. Submitted manuscript: Brötz, B., Eigenmann, R., Dörnbrack, A., Foken, T., Wirth, V., 2013. Early-morning flow transition in a valley in low-mountain terrain. Boundary-Layer Meteorology, submitted. L IST OF MANUSCRIPTS V Other publications not included in this thesis: In peer-reviewed journals: Kalthoff, N., Kohler, M., Barthlott, C., Adler, B., Mobbs, S.D., Corsmeier, U., Träumner, K., Foken, T., Eigenmann, R., Krauss, L., Khodayar, S., Di Girolamo, P., 2011. The dependence of convection-related parameters on surface and boundary-layer conditions over complex terrain. Quarterly Journal of the Royal Meteorological Society, 137: 70–80. Wulfmeyer, V., Behrendt, A., Kottmeier, C., Corsmeier, U., Barthlott, C., Craig, G.C., Hagen, M., Althausen, D., Aoshima, F., Arpagaus, M., Bauer, H.-S., Bennett, L., Blyth, A., Brandau, C., Champollion, C., Crewell, S., Dick, G., Di Girolamo, P., Dorninger, M., Dufournet, Y., Eigenmann, R., Engelmann, R., Flamant, C., Foken, T., Gorgas, T., Grzeschik, M., Handwerker, J., Hauck, C., Höller, H., Junkermann, W., Kalthoff, N., Kiemle, C., Klink, S., König, M., Krauss, L., Long, C.N., Madonna, F., Mobbs, S., Neininger, B., Pal, S., Peters, G., Pigeon, G., Richard, E., Rotach, M.W., Russchenberg, H., Schwitalla, T., Smith, V., Steinacker, R., Trentmann, J., Turner, D.D., van Baelen, J., Vogt, S., Volkert, H., Weckwerth, T., Wernli, H., Wieser, A., Wirth, M., 2011. The Convective and Orographically-induced Precipitation Study (COPS): the scientific strategy, the field phase, and research highlights. Quarterly Journal of the Royal Meteorological Society, 137: 3–30. In non-reviewed journals: Metzger, S., Foken, T., Eigenmann, R., Kurtz, W., Serafimovich, A., Siebicke, L., Olesch, J., Staudt, K., Lüers, J., 2007. COPS experiment. Convective and orographically induced precipitation study. 01 June 2007 - 31 August 2007, Documentation. Work Report, University of Bayreuth, Dept. Micrometeorology, ISSN 1614-8916, 34, 72 pp. VI L IST OF MANUSCRIPTS Data sets at the World Data Center for Climate (WDCC) in Hamburg, Germany: Dorninger, M., Eigenmann, R., Foken, T., 2012. cops_nebt_uv_flux: Surface layer scintillometer data from COPS energy balance network station run by University of Vienna during COPS 2007. World Data Center for Climate. DOI:10.1594/ WDCC/cops_nebt_uv_flux. http://dx.doi.org/10.1594/WDCC/cops_nebt_uv_flux Eigenmann, R., Foken, T., 2012. cops_nebt_ubt_flux: Eddy-covariance turbulence data from COPS energy balance and turbulence network stations run by University of Bayreuth during COPS 2007. World Data Center for Climate. DOI:10.1594/ WDCC/cops_nebt_ubt_flux. http://dx.doi.org/10.1594/WDCC/cops_nebt_ubt_flux Eigenmann, R., Foken, T., 2012. cops_nebt_ubt_mast: Profile data of wind speed, temperature and water vapor pressure at profile mast run by University of Bayreuth during COPS 2007. World Data Center for Climate. DOI:10.1594/ WDCC/cops_nebt_ubt_mast. http://dx.doi.org/10.1594/WDCC/cops_nebt_ubt_mast Eigenmann, R., Foken, T., 2012. cops_nebt_ubt_met: Radiation and soil measurement data from COPS energy network stations run by University of Bayreuth during COPS 2007. World Data Center for Climate. DOI:10.1594/WDCC/cops_nebt_ ubt_met. http://dx.doi.org/10.1594/WDCC/cops_nebt_ubt_met Eigenmann, R., Foken, T., 2012. cops_nsod_ubt: wind data from sodar-RASS run by University of Bayreuth during COPS 2007. World Data Center for Climate. DOI:10.1594/WDCC/cops_nsod_ubt. http://dx.doi.org/10.1594/WDCC/cops_nsod_ubt Kalthoff, N., Eigenmann, R., Foken, T., 2011. cops_nebt_imk_flux: Eddy-covariance turbulence data from COPS energy balance and turbulence network stations run by FZK/IMK-TRO during COPS 2007. World Data Center for Climate. DOI: 10.1594/WDCC/cops_nebt_imk_flux. http://dx.doi.org/10.1594/WDCC/cops_nebt_imk_flux Kalthoff, N., Eigenmann, R., Foken, T., 2011. cops_nebt_imk_met: meteorological data from COPS energy balance and turbulence network run by FZK/IMK-TRO L IST OF MANUSCRIPTS VII during COPS 2007. World Data Center for Climate. DOI:10.1594/WDCC/ cops_nebt_imk_met. http://dx.doi.org/10.1594/WDCC/cops_nebt_imk_met Pigeon, G., Eigenmann, R., Foken, T., Legain, D., Piguet, B., Traulle, O., 2012. cops_nebt_mf_flux: Eddy-covariance turbulence data from COPS energy balance and turbulence network stations run by Meteo-France/CNRM during COPS 2007. World Data Center for Climate. DOI:10.1594/WDCC/cops_ nebt_mf_flux. http://dx.doi.org/10.1594/WDCC/cops_nebt_mf_flux Schüttemeyer, D., Eigenmann, R., Foken, T., 2012. cops_nebt_ubn_flux: Eddycovariance turbulence data from COPS energy balance and turbulence network station run by University of Bonn during COPS 2007. World Data Center for Climate. DOI:10.1594/WDCC/cops_nebt_ubn_flux. http://dx.doi.org/10.1594/WDCC/cops_nebt_ubn_flux 2 I NTRODUCTION height (Mason, 1988). However, in the case of very low wind speeds (free convection regime), even patch lengths considerably smaller than the boundary layer height ( i z< λ , ≈ λ 250 m) are sufficient to affect the turbulence structure of the CBL (Shen and Leclerc, 1994; 1995). Free convection is a common phenomenon in the mixed layer of a CBL on clear sunny days with light wind speeds. It prevails when the part of turbulence kinetic energy (TKE) created by buoyancy dominates over that created by shear (e.g. Stull, 1988). Vertical motion is then primarily induced by density differences within the fluid. Coherent vertical structures (e.g. plumes or thermal updrafts) exist in the CBL during free convection leading to an effective vertical transport. The present thesis is focused on specific situations in which free convection is already observed over certain patches of land use very close to the ground (in the surface layer), where normally shear processes dominate. Air masses close to the surface show different characteristics than air masses further up in the boundary layer, e.g., are humid or have a characteristic chemical composition. These characteristics can then be transported upwards very effectively by free convection which starts already close to the ground. In the surface layer, free convection can be detected with the help of the stability parameter ζ for values of 1 − < ζ (e.g. Webb, 1962; Foken, 2008a). ζ is the quotient of the height z and the Obukhov length L (Obukhov, 1946): ( ) 3 * 0 '' u wgz L z v v θ θκ ζ −== (1) Here, * u is the friction velocity, g the acceleration due to gravity, v θ the mean virtual potential temperature, ( ) 0 '' v w θ the buoyancy flux at the surface and κ the Kármán constant (κ ≈ 0.4). According to Businger et al. (1971), ζ is related approximately linearly to the Richardson number for unstable conditions ( ζ < 0). The flux Richardson number f R is defined as the quotient of the buoyancy term B to the shear term S of the TKE budget equation (e.g. Stull, 1988; Garratt, 1992):         ∂ ∂ == z u wuw g SBR v v f '''' θ θ (2) Regarding a coordinate system aligned with the mean wind, the shear term S is the I NTRODUCTION 3 product of the momentum flux ''wu and the wind shear zu ∂∂ . With the knowledge of f R≈ ζ in an unstable surface layer, |L| can be interpreted as the height above the surface at which buoyant production first dominates over mechanical production of TKE (e.g. Stull, 1988). In other words, 1 − < ζ means that buoyancy-driven turbulence dominates over shear-generated turbulence for |L| < z. Thus, the requirement of 1 − < ζ is used in this thesis (Eigenmann et al., 2009, Appendix B; Eigenmann et al., 2011, Appendix C; Zhou et al., 2011, Appendix D; Brötz et al., 2013, Appendix E) to detect situations of near-ground free convection conditions (FCCs). Regarding Eq. (1), free convection is achieved by the conditions 0 * →u and ( ) 0'' 0 > v w θ . The surface turbulent fluxes, * u and ( ) 0 '' v w θ , can easily be measured by e.g. the eddy-covariance (EC) method (e.g. Swinbank, 1951; Foken et al., 2012). Recently, Mayer et al. (2008) observed the free convective coherent vertical transport of surface layer trace gases into upper regions of the boundary layer in the Alpine foreland during early-morning situations of low wind speed. The ceasing wind speeds could often be related to the onset of a mesoscale circulation system called Alpine pumping (see Lugauer and Winkler, 2005). During the period of changing wind direction, the horizontal wind speed vanished. The trace gases, transported upwards vertically by free convection, were then translocated advectively with the mean wind towards a mountain summit and altered the trace gas observations there significantly. Situations of nearground free convection can also be seen in the stability and data quality analysis of EC data obtained in an Alpine valley in Switzerland (see Hiller et al., 2008), although these authors did not address these events. Following these studies, the main objective of this thesis is to investigate in general the occurrence of FCCs over complex terrain and its impact on ABL characteristics. For this purpose, data of the three month field campaign of the Convective and Orographically induced Precipitation Study (COPS, e.g. Wulfmeyer et al., 2011) in summer 2007 were mainly used. The experiment took place over a heterogeneous low-mountain region typical for Central Europe, i.e. the Vosges Mountains and the Black Forest. These low-mountain ranges are well known to induce thermally driven orographic wind systems (e.g. Kossmann and Fiedler, 2000; Kalthoff et al., 2000; Barthlott et al., 2006; Meißner et al., 2007), which can be expected to trigger free convective situations similar to Mayer et al. (2008). Complex terrain also has an impact on turbulence flux measurements, e.g. on the EC 4 I NTRODUCTION measurements as used in this thesis. It is a well known phenomenon that, in micrometeorological field experiments conducted over complex terrain, the energy balance cannot be closed (e.g. Oncley et al., 2007; Foken, 2008b). Theoretically, the input of energy at the surface by net radiation (-Q S* ) must balance the sum of the sensible (Q H ), latent (Q E ), ground (Q G ) and storage heat flux (∆Q S ): SGEHS QQQQQ ∆+++=− * (3) The sum of Q H and Q E was found to only amount to about 70–90% of the available energy (-Q S* -Q G -∆Q S ) at the surface (e.g. Aubinet et al., 2000; Wilson et al., 2002; Mauder et al., 2006; Franssen et al., 2010). Contrary to that, in landscapes without heterogeneities, a closure of the energy balance was reported (e.g. Heusinkveld et al., 2004; Mauder et al., 2007; Foken, 2008b). For that reason, it is concluded that the landscape heterogeneity plays an important role for the non-closure of the energy balance. It is recognized that the missing flux components are transported within largescale eddies caused by the landscape heterogeneity (e.g. Foken, 2008b, Foken et al., 2010; 2011). These large-scale eddies cannot be captured by a single-point EC station, as they are organized above the EC tower, are quasi-stationary in space or have a longer wavelength than the usual 30 min averaging period of the EC fluxes. Indeed, strong indications exist that the residual over complex terrain is transported within advective and low-frequency flux contributions (e.g. Sakai et al., 2001; Finnigan, 2003; Malhi et al., 2004; Foken et al., 2006; Mauder and Foken, 2006) and within quasi-stationary circulations (e.g. Kanda et al., 2004; Inagaki et al., 2006; Steinfeld et al., 2007; Huang et al., 2008; Foken et al., 2010; Stoy et al., 2013). These secondary circulations are mainly buoyancy-driven due to differences in thermal heating of the land surface. Consequently, to partition the residual according to the buoyancy flux ratio, as proposed by Charuchittipan et al. (2013), seems to be an appropriate approach to close the energy balance. Based on the theoretical background introduced above, the overarching goal of this thesis is the investigation and description of FCCs detected at the height of EC measurements (≈ 2 m) over complex terrain. Major points of interest are the identification of the (micro-) meteorological and synoptical conditions which lead to these free convective events, the frequency and time of their occurrence and their I NTRODUCTION 5 impact on boundary layer properties. Moreover, this thesis aims to investigate the impact of complex terrain on turbulence flux measurements and on the energy balance closure. Regarding the manuscripts attached to this thesis, Eigenmann et al. (2009, Appendix B) introduces the method for the detection of FCCs with EC flux measurements and investigates their occurrence using data obtained in the Kinzig valley in the Black Forest during the COPS experiment. This study also addresses the impact of FCCs on ABL properties by analyzing ABL profiling measurements in the Kinzig valley and studying the spectral characteristics of the turbulence during FCCs. Within the COPS field campaign, an energy balance network was operated (see Sect. 2.1). This network consisted of sixteen individual EC stations distributed over the complex terrain of the COPS region. Different types of land use and different topographical features (valleys and mountaintops) were probed. Energy balance measurements in the COPS region were already carried out within the REKLIP project (e.g. Wenzel et al., 1997), although this project had a more climatological focus. These earlier investigations revealed a significant dependence of the energy balance components on the altitude and on the land-use types in the COPS area (Kalthoff et al., 1999; Wenzel and Kalthoff et al., 2000). Thus, the study of Eigenmann et al. (2011, Appendix C) aims at comparing the flux measurements obtained during COPS with the results of the REKLIP project. Moreover, this study calculates and discusses the energy balance closure at the COPS stations with regard to the surrounding heterogeneous terrain. Also, the occurrence of FCCs within the entire COPS region in relation to different land-use types and topographical features is investigated. In order to demonstrate that FCCs can also be detected within different settings of complex terrain besides the study area of COPS, a data set of Nam Co station on the Tibetan Plateau (see Sect. 2.2) is further investigated in this thesis (see Zhou et al., 2011, Appendix D). The dominating land-use type (Alpine steppe) at Nam Co station is rather homogeneous and the terrain in the immediate vicinity is flat. However, heterogeneity in this study is given by a land-lake surface and by frequently changing heating differences during cloud cover periods. These surface heating differences during cloud cover periods are enhanced on the Tibetan Plateau compared to lowland sites as diffuse radiation components are strongly reduced at high altitudes. A numerical simulation of the ABL flow characteristics at Nam Co station by Lu et al. (2008) recently indicated the occurrence of a thermally driven land-lake circulation system. 6 I NTRODUCTION A further main objective of this thesis is to adapt a LES model (see Sect. 2.3) to the heterogeneous terrain of a segment of the Kinzig valley using COPS data (Brötz et al., 2013, Appendix E). The aim of the model is to simulate the exchange processes in the ABL during the observed FCCs. Especially, the organisation of turbulence in the convective mixed layer, crucial for further vertical transport of the convectively released surface layer air masses, is studied. Therefore, simulations of the flow transition from the free covective low wind speed situation to the following period of along-valley wind are performed. An initial stable stratification in the early-morning hours is considered. The model provides the opportunity to estimate how far surface layer air masses are transported upwards within convective structures during the observed FCCs. E XPERIMENTAL DATA SETS AND THE LES MODEL 7 2 Experimental data sets and the LES model The results presented in the studies of Eigenmann et al. (2009, Appendix B; 2011, Appendix C) are based on data sets obtained within the framework of the COPS project (Sect. 2.1). The publication of Zhou et al. (2011, Appendix D) uses a seven month data set of measurements at Nam Co station on the Tibetan Plateau provided by the project partners involved in this specific study (Sect. 2.2). In the LES study of Brötz et al. (2013, Appendix E), data of the COPS project were used as boundary conditions of the model (Sect. 2.3). 2.1 The COPS field campaign The Convective and Orographically induced Precipitation Study (COPS) was an international field campaign (Wulfmeyer et al., 2008) embedded within the German 6year Priority Program SPP 1167 “Quantitative Precipitation Forecast PQP (Praecipitationis Quantitativae Predictio)” of the German Science Foundation (DFG). The field campaign was carried out from 1 June to 31 August 2007 in a low-mountain area in southwestern Germany and eastern France covering the Vosges Mountains, the Upper Rhine valley, the Black Forest and the Swabian Mountains. It was organized into different Intensive Observation Periods (IOPs), which observed specific convective situations with a synergy of meteorological instruments in order to identify the physical and chemical processes responsible for the deficiencies in Quantitative Precipitation Forecast (QPF) in low-mountain regions (Wulfmeyer et al., 2011). The investigation of the initiation of (moist) convection was a main part of the research efforts (e.g. Aoshima et al., 2008; Kottmeier et al., 2008; Kalthoff et al., 2009, 2011; Behrendt et al., 2011; Bennett et al., 2011, Corsmeier et al., 2011). During COPS, a large number of state-ofthe-art meteorological instrumentation, combining a synergy of in situ and remotesensing systems (e.g. radar and lidar systems), was operated (see Wulfmeyer et al., 2008, 2011). Measurements were obtained from networks (see Sect. 2.1.1; Eigenmann et al., 2011, Appendix C; Hauck et al., 2011), aircraft (e.g. Kiemle et al., 2011) and satellites and were intensified at specific sites (supersites). In the following, only the data sets of the COPS campaign used in this thesis will be introduced. 8 E XPERIMENTAL DATA SETS AND THE LES MODEL 2.1.1 Surface energy balance and turbulence network Under weak synoptic forcing, spatial heterogeneities of surface characteristics result in heterogeneities of turbulent fluxes of heat and moisture into the ABL and hence may determine if and where convection is initiated (e.g. Kottmeier et al., 2008; Kalthoff et al., 2011). Moreover, land surface exchange processes modulate thermally induced orographic flow systems, which in turn influence the initiation of convection (e.g. Barthlott et al., 2006; Meißner et al., 2007, Kalthoff et al., 2009). Therefore, a surface energy balance and turbulence network (Eigenmann et al., 2011, Appendix C) was set up during COPS. Data of high-quality surface flux measurements as well as standard surface micrometeorological measurements was stored at the World Data Center for Climate (WDCC) in Hamburg, Germany. This provided data is further used within the COPS community for the forcing and validation of the applied mesoscale models. The surface network consisted of sixteen stations operated by five collaborating institutes (Table 1). The EC measuring technique was applied in order to provide surface turbulent flux data of momentum, sensible and latent heat, expect at one site where scintillometer measurements were used. Latent heat flux was only measured at nine sites. At most of the sites, the remaining components of the surface energy balance were obtained by additional soil (Hauck et al., 2011) and radiation measurements. The instrumentation of the soil and radiation measurements is listed in Table V in Eigenmann et al. (2011, Appendix C). Heterogeneity in the COPS region exists due to orography and due to a patchy land-use structure. Therefore, measuring sites included locations in the valleys (Murg, Kinzig and Rench valley) and on mountaintops (Hornisgrinde, Igelsberg) of the Black Forest, locations in the Upper Rhine valley and locations on the eastern edge of the Black Forest. Moreover, the turbulent fluxes were measured over different types of land use, mainly grassland and agricultural fields. Table 1 summarizes all above-mentioned information. At some of these stations, Sodar/RASS instruments were also operated for profiling wind components and virtual temperature. In this thesis, only the Sodar/RASS system installed in the Kinzig valley is used (Eigenmann et al., 2009, Appendix B; Brötz et al., 2013, Appendix E). More details about the experimental setup can be looked up in Metzger et al. (2007) and Eigenmann et al. (2009, Appendix B; 2011, Appendix C). E XPERIMENTAL DATA SETS AND THE LES MODEL 9 Tabelle 1: Turbulence measuring sites of the COPS field campaign. The column ‘Code’ abbreviates the responsible institute: UBT (University of Bayreuth), IMK (Karlsruhe Institute of Technology), MF (Météo France), UV (University of Vienna), UBN (University of Bonn). Also given are the coordinates and the altitude above sea level (a.s.l.) of the sites as well as the applied instruments of the EC system (sonic anemometer, hygrometer - CSAT3: sonic anemometer by Campbell Scientific Inc., USA; USA-1: sonic anemometer by METEK GmbH, Germany; Solent R1012: sonic anemometer by Gill Instruments Ltd., UK; Young 81000: sonic anemometer by R. M. Young Company, USA; Solent HS: sonic anemometer by Gill Instruments Ltd., UK; KH20: krypton hygrometer by Campbell Scientific Inc., USA; LI-7500: open-path CO 2 /H 2 O gas analyzer by LICOR Biosciences, USA). The station UV1 used the Optical Energy Balance Measurement System (OEBMS1) with a Scintillometer SLS20 system by Scintec AG, Germany. The column ‘Land use’ indicates the target land-use type: grassland (G), maize (M), strawberry (S), fallow (F) and wheat (W). The column ‘Location’ sorts the stations by their location: valley (V) sites, mountaintop (T) sites, Upper Rhine (R) valley sites and sites in the lee (L) of the Black Forest. Table taken from Eigenmann et al. (2011, Appendix C), modified. Code Site Coordinates (lat., long.) Altitude Land use Sonic anemometer Hygrometer Location UBT1 Fußbach I * 48° 22′ 7.82′′ 8° 1′ 21.17′′ 178 M CSAT3 LI-7500 V UBT2 Fußbach II 48° 22′ 0.88′′, 8° 1′ 16.68′′ 180 F USA-1 - V UBT3 Fischerbach * 48° 16′ 57.40′′, 8° 7′ 56.28′′ 226 G CSAT3 KH20 V UBT4 Hagenbuch * 48° 16′ 54.59′′, 8° 12′ 16.81′′ 245 G CSAT3 LI-7500 V IMK1 Hornisgrinde * 48° 36′ 12.95′′, 8° 12′ 4.88′′ 1158 G Solent R1012 LI-7500 T IMK2 Baden Airpark * 48° 46′ 40.51′′, 8° 4′ 25.20′′ 120 G Solent R1012 LI-7500 R IMK3 Linkenheim * 49° 8′ 9.24′′, 8° 23′ 32.21′′ 96 W Young 81000 - R IMK4 Sasbach 48° 39′ 4.20′′, 8° 5′ 19.53′′ 133 G Solent R1012 - R IMK5 Oberkirch 48° 31′ 19.15′′, 8° 5′ 54.00′′ 203 S Solent R1012 - V IMK6 Bad Rotenfels 48° 49′ 29.35′′, 8° 17′ 30.55′′ 127 G Solent R1012 - V IMK7 Igelsberg 48° 31′ 40.85′′, 8° 25′ 50.35′′ 770 G Solent R1012 - T IMK8 Burnhaupt 47° 42′ 33.52′′, 7° 9′ 16.07′′ 299 M Solent R1012 - R MF1 Niederrott * 48° 26′ 34.40′′, 7° 32′ 38.36′′ 155 M Solent HS LI-7500 R MF2 Nordfeld * 48° 27′ 58.22′′, 7° 32′ 22.35′′ 156 M Solent HS LI-7500 R UV1 Deckenpfronn * 48° 38′ 21.12′′, 8° 49′ 7.86′′ 574 G - - L UBN1 Deckenpfronn * 48° 38′ 21.12′′, 8° 49′ 7.86′′ 574 G CSAT3 LI-7500 L * Additional measurement of radiation and soil components. 10 E XPERIMENTAL DATA SETS AND THE LES MODEL 2.1.2 Post-processing and quality control of the turbulence data In order to receive a data set of high-quality turbulent fluxes, a post-processing and quality control scheme was consistently applied to all turbulence stations. This allows for a high comparability of the surface fluxes between different sites. The scheme included the processing of the turbulence raw data with the software package TK2 (Mauder and Foken, 2004; Mauder et al., 2008) as well as a footprint analysis (Göckede et al., 2004, 2006) and a check for internal boundary layers (Raabe, 1983; Jegede and Foken, 1999) at each site. Detailed information about the individual post-processing and quality control steps and the corresponding references of the applied methods are given in Eigenmann et al. (2011, Appendix C). In this publication, also the application of the data quality control scheme for the selection of data can be looked up. The applied methods are in agreement with the recent recommendations by Foken et al. (2012). 2.2 Data set of Nam Co station The data set of Nam Co station on the Tibetan Plateau used in the study of Zhou et al. (2011, Appendix D) was provided by the project partners involved in this research and covers a time period of seven month from 21 March 2007 to 21 October 2007. Nam Co station was established by the Institute of Tibetan Plateau Research, Chinese Academy of Sciences in August 2005. Besides standard meteorological, radiation and soil measurements, flux measurements were carried out with an EC system. The turbulence data was processed in a similar way as the COPS data (see Sect. 2.1.2). More details are given in Zhou et al. (2011, Appendix D). 2.3 LES model In the framework of this thesis, a LES model was run by the DFG project partners to simulate turbulent exchange processes in a segment of the Kinzig valley. The used numerical model was the multiscale geophysical flow solver EULAG (Smolarkiewicz et al., 1997; Prusa et al., 2008). Among the broad range of application documented in literature, EULAG was also successfully applied to the ABL (see Smolarkiewicz et al., 2007; Piotrowski et al., 2009). A realistic topography was implemented into the model. The boundary conditions were chosen according to the COPS observations in the Kinzig valley at Fußbach (see Table 1). A detailed description of the model and the setup of the simulations can be looked up in Brötz et al. (2013, Appendix E). The objectives of the model application in respect of this thesis are formulated in Sect. 1. R ESULTS 11 3 Results 3.1 Turbulent fluxes and energy balance closure over complex terrain The COPS data set (see Sect. 2.1.1) allowed an investigation of the influence of land use and topographical location on the turbulent fluxes of sensible and latent heat. Fluxes were found to differ strongly between different types of land use (see Eigenmann et al., 2011, Appendix C). In the afternoon, the oasis effect (e.g. Stull, 1988) caused negative values of sensible heat over highly evapotranspirating land-use types (e.g. maize), while over others (e.g. grassland) this effect did not occur. However, flux differences could also be observed between sites with the same type of land use. This was because of varying surface characteristics even over the same type of land use due to different times of mowing of the grassland sites or due to different stages of the vegetation development in general. It became clear that on a specific day the flux values were strongly determined by different types of land use and surface characteristics, while the effect of altitude (mountaintop or valley) plays a minor role. Interpreting these findings within the scope of the COPS project, which deals with the initiation of convection over complex terrain, it can be concluded that on convective days with weak synoptic forcing, the spatial distribution of land-use characteristics may be decisive if and where convection is initiated. Hot spots of increased transport of heat or moisture into the ABL may form over certain patches of land use. However, it should be mentioned that thermally driven wind systems, which frequently occur in the Black Forest (e.g. Kossmann and Fiedler, 2000; Kalthoff et al., 2000; Barthlott et al., 2006; Meißner et al., 2007) may redistribute the surface-initiated heat and moisture distribution in the ABL and thus may modulate the local forcing of the initiation of convection by surface fluxes. To study the effect of altitude on flux differences, the fluxes over one type of land use (here: grassland, as most frequently measured) were averaged over the entire COPS period (see Eigenmann et al., 2011, Appendix C). The averaging minimized temporal flux differences due to mowing and vegetation development. Higher Bowen ratios were found at the top of the mountains and lower values in the valleys. This is in accordance with the former findings by Wenzel et al. (1997) and Kalthoff et al. (1999) in the framework of the REKLIP project. These authors also found an increase of the Bowen ratio from the Rhine valley to the top of the Black Forest within a one year data set. 18 R ESULTS Indeed, strong updrafts could be frequently observed in the valley boundary layer by the Sodar/RASS system during FCCs. The morning evolution of the vertical wind speed is exemplarily shown for COPS IOP 15b (13 August 2007) in Fig. 5 (lower panel). It can be seen that during FCCs – especially from 07:50 to 10:30 UTC – vertical wind speeds are strongly enhanced with values of locally more than 0.8 ms -1 (10 min average value). Such values can also be found within other studies using boundary layer profiling techniques (e.g. Barlow et al., 2011; Kiemle et al., 2011). Moreover, the corresponding profiles of virtual potential temperature in Fig. 5 (upper panel) give insight into the stratification of the valley boundary layer. As expected, at times of strong coherent updrafts (Fig. 5b and d), the stratification is approximately neutral or slightly unstable. However, in the short period of interruption of stronger updrafts at 09:10 UTC (Fig. 5c), stable stratification becomes evident with a profile similar to that observed before the period with FCCs (Fig. 5a). This indicates that outside of the updrafts the valley atmosphere is still stably stratified and that the observed updrafts deeply penetrate into the stably stratified valley boundary layer during FCCs. 50 100 150 200 250 300 350 400 19 21 23 25 height a.g.l. (m) (a) 50 100 150 200 250 300 350 400 19 21 23 25 (b) virtual potential temperature (°C) 50 100 150 200 250 300 350 400 19 21 23 25 (c) 50 100 150 200 250 300 350 400 19 21 23 25 (d) 60 140 220 300 380 460 540 620 700 height a.g.l. (m) time (UTC) (a) (b) (c) (d) Vertical wind speed (ms−1) >0.8 0.6 0.4 0.2 0.1 −0.1 −0.2 −0.4 −0.6 −0.8 <−0.8 05:00 06:00 07:00 08:00 09:00 10:00 11:00 12:00 13:00 Figure 5: Upper panel (a-d): profiles of the virtual potential temperature measured by the Sodar/RASS in the morning hours of COPS IOP 15b (13 August 2007) at Fußbach. The times of the profiles are marked in the lower panel. Lower panel: corresponding Sodargramm of the vertical wind speed in colour from 05:00-13:00 UTC. The black dashed vertical lines indicate the period with FCCs (06:30–10:30 UTC). From Brötz et al. (2013, Appendix E), modified. R ESULTS 19 To answer the question how these coherent updraft structures get organized during FCCs in the boundary layer of the Kinzig valley, the LES model was adapted to the conditions of the Kinzig valley (see Brötz et al., 2013, Appendix E). Flux differences between different types of land surfaces were found to be negligible in the observed early-morning periods with FCCs (see Eigenmann et al., 2011, Appendix C). Orography was assumed to mainly determine the convective structures in the valley in these periods. Indeed, the application of an ensemble and time mean analysis (see Brötz et al., 2013, Appendix E) of the vertical wind speed revealed that the complex orography imposes the valley flow coherent convective motions at specific locations during FCCs (see Fig. 6). Their persistence relative to the mountain ridges is in contrast to the changing locations of coherent structures in a CBL over flat homogeneous terrain. The quasi-stationary patterns of upand downdrafts in the valley are shown in Fig. 6 at about 130 m above the valley floor (300 m a.s.l.). This pattern resembles the typical spoke patterns known for CBLs with zero mean wind speed (e.g. Schmidt and Schumann, 1989). Regarding the location of the measurement site indicated in Fig. 6, it can be seen that it lies within an area of preferred coherent updraft during FCCs. Thus, the simulations support the observations of strong updrafts by the Sodar/RASS (see Fig. 5) during FCCs. An analysis of probability density functions of the simulated and measured vertical wind speed as described in Brötz et al. (2013, Appendix E) also showed that the site is located preferably in an updraft area during FCCs. The situation of turbulent transport in the valley during FCCs was further analysed by means of LES. The simulated TKE budget terms (see Fig. 8 in Brötz et al., 2013, Appendix E) showed that the boundary layer is buoyancy-driven during FCCs. A strong transport of TKE from the lower half of the valley boundary layer into upper regions can be seen in the simulations. However, the simulated following along-valley wind period with no FCCs showed that the quasi-stationary spoke patterns during FCCs are dissolved and irregular streak-like patterns become evident (see Fig. 5 in Brötz et al., 2013, Appendix E), typical for shear-driven boundary layers (e.g. Moeng and Sullivan, 1994; Weckwerth et al., 1997; Drobinski et al., 1998; Drobinski and Foster, 2003). The TKE transport term in the along-valley wind situation is much weaker compared to the period with FCCs. 20 R ESULTS 0 2 4 6 x (km) 0 2 4 6 y (km) -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 -0.05 0.05 0.1 0.2 0.3 0.4 0.5 0.6 Figure 6: Ensemble and time mean of the simulated vertical wind speed in ms -1 at 300 m a.s.l. in the Kinzig valley during FCCs (color-coded). Black solid lines mark the orography in steps of 50 m. Grey contours indicate the intersection with the orography. The red frame represents the section of the valley used for Fig. 7. The location of the measurement site is marked by the black circle. From Brötz et al. (2013, Appendix E). -0.003 0.000 0.003 gradient potential temperature (Km-1) 0.0 0.5 1.0 1.5 2.0 zzi-1 -0.003 0.000 0.003 gradient potential temperature (Km-1) 0.0 0.5 1.0 1.5 2.0 zzi-1 -0.4 0.0 1.0 QHQH,0 -1 0.0 0.5 1.0 1.5 2.0 zzi-1 -0.4 0.0 1.0 QHQH,0 -1 0.0 0.5 1.0 1.5 2.0 zzi-1 total up down subgrid total up down subgrid Figure 7: (left panel) Simulated vertical profiles of the gradient of the potential temperature and (right panel) of Q H (normalized) during FCCs. The solid line shows the values for all points in the valley (see red frame in Fig. 6), while the dotted (dashed) line shows the profiles for the updraft (downdraft) areas in the valley. The contribution to Q H from the sub-grid model is shown in the right panel as dash-dotted line. From Brötz et al. (2013, Appendix E), modified. R ESULTS 21 To further highlight the effective vertical transport during FCCs, Fig. 7 shows the simulated profiles of the flux of sensible heat and the corresponding vertical profiles of the gradient of the potential temperature for the valley area (see red frame in Fig. 6). The mean total profile and the mean profiles for upand downdraft areas are depicted separately. Regarding the total flux of sensible heat in the valley, it can be seen that the flux is counter to the temperature gradient in the center of the valley between about 0.4 z i and 0.8 z i . The total heat flux is mainly determined by the flux within the updraft areas. Regarding the updraft areas, the heat flux follows the (unstable to neutral) temperature gradient up to a height of 0.65 z i . A counter-gradient flux remains above this height up to about 0.9 z i . Counter-gradient fluxes are a common feature in turbulent flows (e.g. Schumann, 1987) and can also be found within forest canopies (see e.g. Denmead and Bradley, 1985). The described counter-gradient flux is a very effective transport mechanism for surface layer air mass properties to reach higher altitudes in the stably stratified valley. For illustration, Fig. 8 shows for the vertical wind and the temperature stratification an instantaneous vertical slice through the model domain during FCCs. Strong convective updraft structures can be seen within the valley penetrating through the stably stratified valley atmosphere up to approximately the height of the surrounding mountains. The simulated vertical counter-gradient transport in the valley agrees well with the Sodar/RASS observations shown in Fig. 5. 0 2 4 6 x (km) 0.2 0.4 0.6 0.8 1.0 1.2 1.4 z (km) 290 292 292 Figure 8: I nstantaneous situation of the simulated vertical wind speed in ms -1 (color-coded) shown in a vertical cross section perpendicular to the axis of the valley during FCCs. Black lines are isolines of potential temperature in steps of 0.2 K. Intersection with the orography is shaded in grey. From Brötz et al. (2013, Appendix E), modified. 22 C ONCLUSIONS 4 Conclusions The following conclusions related to free convection and turbulent fluxes over complex terrain can be drawn: (i) High effort was put into a uniform processing and quality control of the surface turbulent flux measurements conducted at sixteen stations during COPS (Eigenmann et al., 2011, Appendix C). This allowed for a high comparability of the measurements as methodical differences could be excluded. Differences in flux values on days with weak synoptic forcing were found to be strongly determined by varying land surface characteristics. The role of altitude plays a minor role, although a slight increase of the Bowen ratio from the valleys to the mountaintops – similar to former experiments in this area (Wenzel et al., 1997; Kalthoff et al., 1999) – could be found. From this it can be concluded that on a single day with weak synoptic forcing, the initiation of convection can be forced locally by the spatial distribution of land surface characteristics. (ii) Values of the non-closure of the energy balance typical for agricultural sites were found at the sites of the COPS field campaign regarding average conditions. The landscape heterogeneity causing not measurable advective and low-frequency flux components is assumed to be responsible for the observed residual. This hypothesis is reinforced by the fact that the residual was found to increase with the onset of the oasis effect. Moreover, in the observed low wind speed situations with free convection conditions (FCCs) – characterized by a more vertically oriented turbulent exchange regime – the residual was found to vanish on average. Due to the more vertical orientation of the exchange regime during FCCs, the landscape heterogeneity at a single measurement site is of minor importance and missing advective flux components are strongly reduced. With the onset of the wind after the periods with FCCs, the heterogeneity of the surrounding landscape becomes again more important and advective flux components and other heterogeneity effects lead again to an unclosed energy balance. (iii) It was shown that the relative contributions of sensible and latent heat missing in the considered periods with no FCCs compared to the periods with FCCs have exactly the proportions of the buoyancy flux ratio. This C ONCLUSIONS 23 finding supports the theory (e.g. Foken et al. 2008b) that the missing flux components are transported within buoyancy-driven secondary circulations which develop over heterogeneous terrain. Moreover, this finding also supports the recent suggestion of Charuchittipan et al. (2013) to use the buoyancy flux ratio approach for the correction of the energy balance. The usually used approach by Twine et al. (2000) to distribute the residual according to the Bowen ratio seems to be not appropriate. (iv) FCCs could be frequently observed in situations of weak synoptic forcing during COPS. This finding confirmed the initial question of this thesis based on the work of Mayer et al. (2008) whether FCCs occur in general over complex terrain. As the occurrence of FCCs is bound to vanishing wind speeds ( 0 * →u ) and simulataneously enhanced positive values of the buoyancy flux ( ( ) > 0 '' v w θ 20 Wm -2 ), FCCs were observed in all probed valleys of the Black Forest during the reversal of the valley wind system from downto up-valley winds in the morning hours. Thus, the occurrence of FCCs is mainly controlled by the adaption of thermally driven orographic or local wind systems to heating differences over complex terrain. However, also the surface characteristics with their impact on the resulting turbulent fluxes may control the occurrence of FCCs. For instance, it was shown that the negative values of sensible heat during the oasis effect inhibit FCCs. (v) With the data set of Nam Co station on the Tibetan Plateau, the conclusions drawn in (iv) could be confirmed. However, FCCs were more frequently observed in the afternoon in combination with cloud cover periods at this site. As diffuse radiation components are strongly reduced at the altitude of the Tibetan Plateau compared to lowland sites, the cloud cover periods caused a strong cooling of the land surface within a short time. Thus, the adaption of the land-lake breeze at Nam Co station to the changing situation of heating differences led to more periods of vanishing wind speeds and thus to more FCCs in combination with cloud cover periods compared to the COPS region. (vi) A spectral analysis of the turbulent scales and an analysis of the vertical wind speeds measured by the Sodar/RASS revealed the existence of coherent updraft structures in the Kinzig valley during FCCs. The Sodar/RASS 24 C ONCLUSIONS measurements further showed that these updrafts penetrate into the stably stratified valley boundary layer indicating that surface layer air mass characteristics can be effectively injected into higher regions of the valley boundary layer during FCCs. It can be assumed that these fluxes in the earlymorning hours are crucial for the preconditioning of the CBL for a further development of moist convection in the course of the day. However, the experimental design of COPS – contrary to the trace gas measurements in the study of Mayer et al. (2008) which tracked a pulse of surface layer trace gases up to a mountain summit – was not able to answer such questions. An experiment similar to that of COPS, but focused on a certain segment of e.g. the Kinzig valley, should be considered in future plans of boundary layer experiments. (vii) The application of the LES model confirmed that the measurement site in the Kinzig valley is located preferably in an updraft region during FCCs. This was demonstrated by an ensemble and time mean analysis of the vertical wind speed in the valley, which showed that the complex orography forces the valley flow to form coherent convective structures at spatially fixed locations relative to the surrounding mountain ridges. This is in contrast to the random locations of convective structures in case of a CBL over flat homogeneous terrain. (viii) Simulated profiles of the sensible heat flux showed that the total flux out of the valley is mainly determined by the flux within these coherent updraft structures during FCCs. The total heat flux was found to be a flux counter to the temperature gradient in these early-morning situations. This finding underlines the observations summarized in (vi) that surface layer air mass properties can be effectively transported into higher regions of the stably stratified valley boundary layer during FCCs. According to the simulation, these updrafts reach up to approximately the height of the surrounding mountains. From this, it can be concluded that the vertically transported surface layer air mass properties during FCCs can be horizontally translocated with the above-valley or large-scale flow and can also alter ABL characteristics elsewhere. R EFERENCES 25 References Aoshima, F., Behrendt, A., Bauer, H.-S., Wulfmeyer, V., 2008. Statistics of convection initiation by use of Meteosat rapid scan data during the Convection and Orographically-induced Precipitation Study (COPS). Meteorol. Z., 17: 921–930. Aubinet, M., Grelle, A., Ibrom, A., Rannik, Ü., Moncrieff, J., Foken, T., Kowalski, A. S., Martin, P.H., Berbigier, P., Bernhofer, Ch., Clement, R., Elbers, J., Granier, A., Grünwald, T., Morgenstern, K., Pilegaard, K., Rebmann, C., Snijders, W., Valentini, R., Vesala, T., 2000. Estimates of the annual net carbon and water exchange of forests: the EUROFLUX methodology. Adv. Ecol. Res., 30: 113–175. Avissar, R., Schmidt, T., 1998. An evaluation of the scale at which ground-surface heat flux patchiness affects the convective boundary layer using large-eddy simulations. J. Atmos. Sci., 55: 2666–2689. Baidya Roy, S., Avissar, R., 2000. Scales of response of the convective boundary layer to land surface heterogeneity. Geophys. Res. Lett., 27, 533–536. Bange, J., Spieß, T., Herold, M., Beyrich, F., Hennemuth, B., 2006. Turbulent fluxes from Helipod flights above quasi-homogeneous patches within the LITFASS area. Bound.-Lay. Meteorol., 121: 127–151. Barlow, J.F., Dunbar, T.M., Nemitz, E.G. Wood, C.R., Gallagher, M.W., Davies, F., O'Connor, E., Harrison, R.M., 2011. Boundary layer dynamics over London, UK, as observed using Doppler lidar during REPARTEE-II. Atmos. Chem. Phys., 11: 2111– 2125. Barthlott, C., Corsmeier, U., Meißner, C., Braun, F., Kottmeier, C., 2006. The influence of mesoscale circulation systems on triggering convective cells over complex terrain. Atmos. Res., 81: 150–175. Barthlott, C., Schipper, J.W., Kalthoff, N., Adler, B., Kottmeier, C., Blyth, A., Mobbs, S., 2010. Model representation of boundary-layer convergence triggering deep convection over complex terrain: A case study from COPS. Atmos. Res., 95: 172– 185. Behrendt, A., Pal, S., Aoshima, F., Bender, M., Blyth, A., Corsmeier, U., Cuesta, J., Dick, G., Dorninger, M., Flamant, C., Di Girolamo, P., Gorgas, T., Huang, Y., Kalthoff, N., Khodayar, S., Mannstein, H., Träumner, K., Wieser, A., Wulfmeyer, V., 2011. Observation of convection initiation processes with a suite of state-of-theart research instruments during COPS IOP 8b. Q. J. R. Meteorol. Soc., 137: 81–100. 26 R EFERENCES Bennett, L.J., Blyth, A.M., Burton, R.R., Gadian, A.M., Weckwerth, T.M., Behrendt, A., Di Girolamo, P., Dorninger, M., Lock, S.-J., Smith, V.H., Mobbs, S.D., 2011. Initiation of convection over the Black Forest mountains during COPS IOP15a. Q. J. R. Meteorol. Soc., 137: 176–189. Betts, A.K., Ball, J.H., Beljaars, A.C.M., Miller, M.J., Viterbo, P.A., 1996. The land surface-atmosphere interaction: A review based on observational and global modeling perspectives. J. Geophys. Res., 101: 7209–7225. Beyrich, F., Mengelkamp, H.-T., 2006. Evaporation over a heterogeneous land surface: EVA_GRIPS and the LITFASS-2003 experiment - an overview. Bound.-Lay. Meteorol., 121:5–32. Brötz, B., Eigenmann, R., Dörnbrack, A., Foken, T., Wirth, V., 2013. Early-morning flow transition in a valley in low-mountain terrain. Bound.-Lay. Meteorol., submitted. Brunsell, N.A., Mechem, D.B., Anderson, M.C., 2011. Surface heterogeneity impacts on boundary layer dynamics via energy balance partitioning. Atmos. Chem. Phys., 11: 3403–3416. Businger, J.A., Wyngaard, J.C., Izumi, Y., Bradley, E.F., 1971. Flux-profile relationships in the atmospheric surface layer. J. Atmos. Sci., 28: 181–189. Chandra, A.S., Kollias, P., Giangrande, S.E., Klein, S.A., 2010. Long-term observations of the convective boundary layer using insect radar returns at the SGP ARM climate research facility. J. Clim., 23: 5699–5714. Charuchittipan, D., Babel, W., Mauder, M., Beyrich, F., Leps, J.P., Foken, T., 2013. Extension of the averaging time of the eddy-covariance measurement and its effect on the energy balance closure. Bound.-Lay. Meteorol., to be submitted. Corsmeier, U., Kalthoff, N., Barthlott, C., Aoshima, F., Behrendt, A., Di Girolamo, P., Dorninger, M., Handwerker, J., Kottmeier, C., Mahlke, H., Mobbs, S.D., Norton, E.G. Wickert, J., Wulfmeyer, V., 2011. Processes driving deep convection over complex terrain: A multi-scale analysis of observations from COPS-IOP 9c. Q. J. R. Meteorol. Soc., 137: 137–155. Courault, D., Drobinski, P., Brunet, Y., Lacarrere, P., Talbot, C., 2007. Impact of surface heterogeneity on a buoyancy-driven convective boundary layer in light winds, Bound.-Lay. Meteorol., 124: 383–403. Defant, F., 1949. Zur Theorie der Hangwinde, nebst Bemerkungen zur Theorie der Bergund Talwinde. Arch. Meteorol. Geophys. Bioklim., A1: 421–450. R EFERENCES 27 Denmead, D.T., Bradley, E.F., 1985. Flux-gradient relationships in a forest canopy. In: B.A. Hutchison, Hicks B.B. (Editors), The forest-atmosphere interaction, D. Reidel Publ. Comp., Dordrecht, Boston, London, pp 421–442. Drobinski, P., Brown, R.A., Flamant, P.H., Pelon, J., 1998. Evidence of organized large eddies by ground-based doppler lidar, sonic anemometer and sodar. Bound.-Lay. Meteorol., 88: 343–361. Drobinski, P., Foster, R.C., 2003. On the origin of near-surface streaks in the neutrallystratified planetary boundary layer. Bound.-Lay. Meteorol., 108: 247–256. Eigenmann, R., Metzger, S., Foken, T., 2009. Generation of free convection due to changes of the local circulation system. Atmos. Chem. Phys., 9: 8587–8600. Eigenmann, R., Kalthoff, N., Foken, T., Dorninger, M., Kohler, M., Legain, D., Pigeon, G., Piguet, B., Schüttemeyer, D., Traulle, O., 2011. Surface energy balance and turbulence network during the Convective and Orographically-induced Precipitation Study (COPS). Q. J. R. Meteorol. Soc., 137: 57–69. Finnigan, J.J., Clement, R., Malhi, Y., Leuning, R., Cleugh, H.A., 2003. A re-evaluation of long-term flux measurement techniques. Part I: Averaging and coordinate rotation. Bound.-Lay. Meteorol., 107: 1–48. Foken, T., Wimmer, F., Mauder, M., Thomas, C., Liebethal, C., 2006. Some aspects of the energy balance closure problem. Atmos. Chem. Phys., 6: 4395–4402. Foken, T., 2008a. Micrometeorology. Springer, Berlin, Heidelberg, 308 pp. Foken, T., 2008b. The energy balance closure problem: An overview. Ecol. Appl., 18: 1351–1367. Foken, T., Mauder, M., Liebethal, C., Wimmer, F., Beyrich, F., Leps, J.P., Raasch, S., DeBruin, H.A.R., Meijninger, W.M.L., Bange, J., 2010. Energy balance closure for the LITFASS-2003 experiment. Theor. Appl. Climatol., 101: 149–160. Foken, T., Aubinet, M., Finnigan, J.J., Leclerc, M.Y., Mauder, M., K.T., Paw U, 2011. Results of a panel discussion about the energy balance closure correction for trace gases. B. Am. Meteorol. Soc., 92(4): ES13–ES18. Foken, T., Leuning, R., Oncley, S.P., Mauder, M., Aubinet, M., 2012. Corrections and data quality. In: M. Aubinet et al. (Editors), Eddy covariance: A practical guide to measurement and data analysis, Springer, Dordrecht, Heidelberg, London, New York, pp. 85–131. Franssen, H.J.H., Stöckli, R., Lehner, I., Rotenberg, E., Seneviratne, S.I., 2010. Energy balance closure of eddy-covariance data: A multisite analysis for European 34 R EFERENCES Weckwerth, T.M., Wilson, J.W., Wakimoto, R.M., Crook, N.A., 1997. Horizontal convective rolls: determining the environmental conditions supporting their existence and characteristics. Mon. Weather Rev., 125: 505–526. Weigel, A.P., Chow, F.K., Rotach, M.W., 2007b. The effect of mountainous topography on moisture exchange between the “surface” and the free atmosphere. Bound.-Lay. Meteorol., 125: 227–244. Wenzel, A., Kalthoff, N., Fiedler, F., 1997. On the variation of the energy-balance components with orography in the Upper Rhine Valley. Theor. .Appl. Climatol., 57: 1–9. Wenzel, A., Kalthoff, N., 2000. Method for calculating the whole-area distribution of sensible and latent heat fluxes based on climatological observations. Theor. Appl. Climatol., 66: 139–160. Whiteman, C.D., 1990. Observations of thermally developed wind systems in mountainous terrain. In: W. Blumen (Editor), Atmospheric processes over complex terrain. Meteorological Monographs, 23 (45), American Meteorological Society, Boston, USA, pp. 5–42. Whiteman, C.D., 2000. MountainMeteorology: Fundamentals and Applications. Oxford University Press, USA, 355 pp. Wilson, K., Goldstein, A., Falge, E., Aubinet, M., Baldocchi, D., Berbigier, P., Bernhofer, Ch., Ceulemans, R., Dolman, H., Field, C., Grelle, A., Ibrom, A., Law, B.E., Kowalski, A., Meyers, T., Moncrieff, J., Monson, R., Oechel, W., Tenhunen, J., Valentini, R., Verma, S., 2002. Energy balance closure at FLUXNET sites. Agr. Forest Meteorol., 113: 223–243. Wulfmeyer, V., Behrendt, A., Bauer, H.S., Kottmeier, C., Corsmeier, U., Blyth, A., Craig G., Schumann, U., Hagen, M., Crewell, S., Di Girolamo, P., Flamant, C., Miller, M., Montani, A., Mobbs, S., Richard, E., Rotach, M.W., Arpagaus, M., Russchenberg, H., Schlüssel, P., König, M., Gärtner, V., Steinacker, R., Dorninger, M., Turner, D.D., Weckwerth, T., Hense, A., Simmer, C., 2008. The Convective and Orographically induced Precipitation Study: A research and development project of the World Weather Research Program for improving quantitative precipitation forecasting in low-mountain regions. B. Am. Meteorol. Soc. 89: 1477–1486. Wulfmeyer, V., Behrendt, A., Kottmeier, C., Corsmeier, U., Barthlott, C., Craig, G.C., Hagen, M., Althausen, D., Aoshima, F., Arpagaus, M., Bauer, H.-S., Bennett, L., Blyth, A., Brandau, C., Champollion, C., Crewell, S., Dick, G., Di Girolamo, P., R EFERENCES 35 Dorninger, M., Dufournet, Y., Eigenmann, R., Engelmann, R., Flamant, C., Foken, T., Gorgas, T., Grzeschik, M., Handwerker, J., Hauck, C., Höller, H., Junkermann, W., Kalthoff, N., Kiemle, C., Klink, S., König, M., Krauss, L., Long, C.N., Madonna, F., Mobbs, S., Neininger, B., Pal, S., Peters, G., Pigeon, G., Richard, E., Rotach, M.W., Russchenberg, H., Schwitalla, T., Smith, V., Steinacker, R., Trentmann, J., Turner, D.D., van Baelen, J., Vogt, S., Volkert, H., Weckwerth, T., Wernli, H., Wieser, A., Wirth, M., 2011. The Convective and Orographicallyinduced Precipitation Study (COPS): the scientific strategy, the field phase, and research highlights. Q. J. R. Meteorol. Soc., 137: 3–30. Zardi, D., Whiteman, C.D., 2013. Diurnal mountain wind systems. In: F.K. Chow, S.F. De Wekker, B.J. Snyder (Editors), Mountain Weather Research and Forecasting, Springer, Dordrecht, Heidelberg, New York, London, pp. 35–119. Zhou, D., Eigenmann, R., Babel, W., Foken, T., Ma, Y., 2011. Study of near-ground free convection conditions at Nam Co station on the Tibetan Plateau. Theor. Appl. Climatol., 105: 217–228. 36 L IST OF APPENDICES List of appendices A PPENDIX A: I NDIVIDUAL CONTRIBUTIONS TO THE JOINT PUBLICATIONS A PPENDIX B: E IGENMANN ET AL . (2009) A PPENDIX C: E IGENMANN ET AL . (2011) A PPENDIX D: Z HOU ET AL . (2011) A PPENDIX E: B RÖTZ ET AL . (2013) A PPENDIX A 37 Appendix A: Individual contributions to the joint publications The results presented in this cumulative thesis were obtained in collaboration with other scientists. Thus, many authors contributed to the publications listed in the appendices B to E in different ways. This section is to specify my own contributions to the individual manuscripts. Appendix B Eigenmann, R., Metzger, S., Foken, T., 2009. Generation of free convection due to changes of the local circulation system. Atmospheric Chemistry and Physics, 9: 8587–8600. • The data used in this publication were obtained during the COPS field campaign and include the measurements of one energy balance and turbulence station and a nearby Sodar/RASS system. The set-up and maintenance of the instruments and the routinely performed data quality control during the experiment involved many people of the Department of Micrometeorology (A. Serafimovich, L. Siebicke, K. Staudt, J. Lüers, J. Olesch). S. Metzger was mainly responsible for the on-site data collection and continuous operation during the three month field campaign. I also supported the field work of S. Metzger during a two weeks stay. • I alone was responsible for the post-processing of the eddy-covariance, lowfrequency soil and radiation measurements and the Sodar/RASS data. The innovative data analysis procedure for the detection of near-ground free convection conditions (FCCs), the analysis of spectral characteristics of the turbulence during these situations and the analysis of the Sodar/RASS data was performed by myself. W. Babel supported me with his knowledge about spectral analysis methods. I alone wrote the complete text of the manuscript. • B. Brötz contributed with many fruitful discussions and comments, especially in the review process of the publication. • T. Foken supervised this work and contributed with many helpful ideas and discussions. 38 A PPENDIX A Appendix C Eigenmann, R., Kalthoff, N., Foken, T., Dorninger, M., Kohler, M., Legain, D., Pigeon, G., Piguet, B., Schüttemeyer, D., Traulle, O., 2011. Surface energy balance and turbulence network during the Convective and Orographically-induced Precipitation Study (COPS). Quarterly Journal of the Royal Meteorological Society, 137: 57–69. • This publication is an overview over the post-processing and quality control of the turbulence data of all sixteen stations of the surface energy balance and turbulence network during COPS. The network consisted of five different institutions (Department of Micrometeorology, University of Bayreuth; Institute for Meteorology and Climate Research, Karlsruhe Institute of Technology; Meteorological Institute, University of Bonn; Department of Meteorology and Geophysics, University of Vienna; CNRM-GAME, Météo France). Each institution conducted the field work during the experiment independently. Responsibilities for the field work of the Department of Micrometeorology were already mentioned with the publication above. Most institutions provided the results of the processing of the turbulence raw data with the software package TK2. For the University of Bonn and for our department, the processing of the data with TK2 was done by myself. • My contribution was the development of a consistent turbulence data postprocessing scheme applied to all sixteen measuring sites. I alone did the quality control of the turbulence data, including a footprint analysis and a check for internal boundary layers. For this purpose, each institution provided land use data for the corresponding sites. • I myself performed the data analysis about some typical features of turbulent flux data and the occurrence of FCCs within the entire COPS region. I alone wrote the text of the manuscript including some comments of the coauthors. • My supervisor T. Foken contributed to this publication through many discussions and initiated the scientific exchange between the different institutions involved. A PPENDIX A 39 Appendix D Zhou, D., Eigenmann, R., Babel, W., Foken, T., Ma, Y., 2011. Study of near-ground free convection conditions at Nam Co station on the Tibetan Plateau. Theoretical and Applied Climatology, 105: 217–228. • The data used in this publication were provided by the Institute of Tibetan Plateau Research, Chinese Academy of Sciences, within the framework of the scientific collaboration of this study originated by T. Foken and Y. Ma. This study aimed at demonstrating the applicability of the method for the detection of FCCs also in another investigation area with a different setting of heterogeneity (land-lake surface). • I myself, W. Babel and T. Foken introduced D. Zhou into the post-processing of the turbulence data, including the usage of TK2 and footprint analysis tools, and into the investigation and relevance of FCCs at Nam Co station. • D. Zhou performed the processing of the data under close instruction of myself and W. Babel. D. Zhou also wrote a first version of the manuscript. • I myself intensively revised and rephrased the initial manuscript version of D. Zhou before submission. • T. Foken, W. Babel and Y. Ma contributed to the progress of the manuscript in several discussions. Appendix E Brötz, B., Eigenmann, R., Dörnbrack, A., Foken, T., Wirth, V., 2013. Early-morning flow transition in a valley in low-mountain terrain. Boundary-Layer Meteorology, submitted. • B. Brötz was responsible for the performance of the applied large-eddy simulations (LES) under the guidance of V. Wirth and A. Dörnbrack. • The observational data from COPS, necessary for the adaption of the model to the complex terrain of a segment of the Kinzig valley, were provided by myself and T. Foken. 40 A PPENDIX A • The scientific content of this study was intensively discussed in many fruitful meetings of the coauthors. Each of them contributed with ideas on data analyses in manifold ways. The study aimed at investigating the convective structures in the valley during the observed FCCs responsible for the transport of surface layer air masses into higher regions of the boundary layer. Simulated data was compared with the observations. A great number of telephone conferences between myself and B. Brötz refined the content of the manuscript and led to its final version. • The text was written in close cooperation with myself and B. Brötz considering many helpful comments of the supervisors. I myself mainly wrote the introduction of the observational data (Section 2.1), the interpretation of the modification of the energy balance closure during FCCs (Section 3.1) and the interpretation of the observations and the simulations with respect to the vertical transport situation during FCCs (Section 3.3). A. Dörnbrack strongly supported the writing in the part of the description of the model (Section 2.2). A PPENDIX B – E IGENMANN ET AL . (2009) 41 Appendix B: Eigenmann et al. (2009) Atmos. Chem. Phys., 9, 8587–8600,2009 www.atmos-chem-phys.net/9/8587/2009/ ©Author(s) 2009. This work is distributed under the CreativeCommons Attribution 3.0 License. Atmospheric Chemistry and Physics Generation of free convection due to changes of the local circulation system R. Eigenmann1,S. Metzger1,*,and T.Foken1 1Department of Micrometeorology,University of Bayreuth, Bayreuth, Germany *nowat: Institute for Meteorology and Climate Research –Atmospheric EnvironmentalResearch (IMK-IFU), Karlsruhe Institut of Technology,Garmisch-Partenkirchen, Germany Received: 31 March 2009 – Published in Atmos. Chem. Phys. Discuss.: 7May 2009 Revised: 29 October 2009 – Accepted: 5November 2009 – Published: 12 November 2009 Abstract. Eddy-covariance and Sodar/RASS experimental measurement data of the COPS (Convectiveand Orographically-induced Precipitation Study) field campaign 2007 are used to investigate the generation of near-ground freeconvectionconditions (FCCs)inthe Kinzig valley,Black Forest, Southwest Germany.The measured high-quality turbulent flux data revealed that FCCs are initiated near the ground in situations where moderate to high buoyancyfluxes and asimultaneously occurring drop of the wind speedwere present. The minimum in wind speed –observable by the Sodar measurements through the whole vertical extension of the valleyatmosphere –is the consequence of athermallyinduced valleywind system, which changes its wind direction from down to up-valleywinds in the morning hours. Buoyancythen dominates overshear within the production of turbulence kinetic energy near the ground. These situations are detected by the stability parameter (ratio of the measurement height to the Obukhov length) when the level of free convection, which starts abovethe Obukhov length, drops belowthat of the sonic anemometer.An analysis of the scales of turbulent motions during FCCs using wavelet transform shows the occurrence of large-scale turbulence structures. Regarding the entire COPS measurement period, FCCs in the morning hours occur on about 50% of all days.Enhanced surface fluxes of latent and sensible heat are found on these days. 1Introduction The COPS (Convectiveand Orographically-induced Precipitation Study) field campaign was undertaken from 1June to 31 August 2007 within the lowmountain range of the Correspondence to: R. Eigenmann ([email protected]) Black Forest, the Vosges Mountains and the Swabian Jura with the Rhine rift valleyas apronounced topographic lowland plain in between (Wulfmeyer et al.,2008). Rainfall in the COPS area is characterized by subgrid-scale convection initiation (CI) processes, e.g. orographically or thermallyinduced local circulation systems,triggered by the complex terrain, thus complicating the exact modeling and forecasting of precipitation events (Meißner et al.,2007;Barthlott et al., 2006). The problems of modeling convectiveclouds and precipitation –its time, amount and location –are the premature initiation of convection, the simulation of convective precipitation events as being too spatially widespread and the overestimation of precipitation on the windward compared to the lee side over low-mountainranges, e.g. the Black Forest (Schwitalla et al.,2008). The interaction of the land surface with the overlyingatmosphere crucially affects the energy and water cycle over manytemporal and spatial scales (Betts et al.,1996). The spatial distribution and patchiness of individual land use elements can have a strong impact on theatmospheric boundary layer (ABL) evolution and its thermodynamic structure, as changes of the surface energy budget directly influence the surface turbulent fluxes of moisture, momentum and heat, which act as the link between the atmosphere and the underlying soil-vegetation system (Pielke,2001). Dynamical phenomena in the ABL can be related to changing surface characteristics, since gradients in sensible heat flux produced by evapotranspiration, albedo and soil property discontinuities induce local or secondary circulation systems (Segal and Arritt,1992). Together with diurnal mountain winds developing over mountainous terrain (Whiteman,1990), land surface-atmosphere interactions and the related physical processes within complexterrain are the keyto the local occurrence and timing of the initiation of convection, cloud formation and precipitation (Hanesiak et al.,2004;Pielke,2001; Chen and Avissar,1994;Rabin et al.,1990;Banta,1990, 1984;Raymond and Wilkening,1980). Besides land-surface Published by Copernicus Publications on behalf of the European Geosciences Union. 42 A PPENDIX B – E IGENMANN ET AL . (2009) 8588 R. Eigenmann et al.: Generation of free convection dueto changes of the local circulation system interactions and orography,mesoscale and synoptic scale features are important for convectiveprocesses (Wulfmeyer et al.,2007). Kottmeier et al. (2008)discusses several mechanisms relevant for CI in the COPS region. The present study aims at the detection and the description of near-ground free convection conditions (FCCs) by using eddy-covariance (EC) measurements. FCCs can be detected at the height of the EC measurement (see Sect. 2.1)close to the ground with the help of the stability parameter (see Sect. 3.2)and occur if the buoyancyterm dominates over the shear term within the turbulence kinetic energy equation. In the case of detection of FCCs near the ground, convectiveelements are closely related to ground sources and can more effectively transport quantities of moisture, heat and trace gases enhanced in near-ground regions intothe ABL. Moreover,following Shen and Leclerc (1995), the dimensions of our targeted land use type (corn field) of the EC measurements (see Sect. 2.1)are large enough (>250m) that the surface fluxes are able to exert amajor influence on ABL thermodynamics and turbulence structure. Recently,Mayer et al. (2008)found that FCCs detected close to the bottom of avalleyresult in astrong and sudden ozone decrease at amountain summit (Hohenpeissenberg) in the morning hours. In their study,the FCCs are triggered by asimultaneously occurring wind speed minimum, which reduces shear and leads to adominance of buoyancy.On about half of the days these wind speed minima could be attributed to the onset of Alpine Pumping in the alpine foreland (Lugauer and Winkler,2005)associated with achangeof wind direction causing adrop of the horizontal wind speed. Hence, other mesoscale or local circulation systems initiated by complexterrain –e.g. the well known slope and valleywinds in the Black Forest (Kossmann and Fiedler,2000; Kalthoffet al.,2000)–are expected to trigger FCCs. For that reason, the focus of the present study is to demonstrate the applicability of the EC method to the detection of FCCs in experimental data obtained during the COPS field campaign in the Kinzig valley(Black Forest), which was found in earlier studies (e.g., Meißner et al.,2007)to establish apronounced diurnal thermally-induced valleywind system. Moreover,the processed turbulent flux data passing through adetailed quality assurance and control effort are used to select and describe these FCCs in detail. 2Materials and methods 2.1 Site description and experiment set-up During the COPS field campaign, anetwork of 17 energy balance and turbulence stations was set up over the entire COPS region (Wulfmeyer et al.,2007). The energy balance and turbulence station under investigation in the present study is located in the Kinzig valleynear Fußbach (48◦22′7.8′′ N, 8◦1′21.2′′ E, 178ma.s.l) on the western edge W−E distance from tower [m] N−S distance from tower [m] deciduous tree conifer jerusalem artichoke corn meadow field/fallow garden slope street building stream power pole small power pole large x↑↑ N −500 −250 0 250 500 −500 −250 0 250 500 Fig.1. Land use map (1×1km2)at the Fußbach site with the location of the eddy-covariance station indicated as ablack cross in the middle of the map abovethe target land use type (corn). The additional red cross marks the position of the Sodar/RASS system. Fetch distances of the eddy-covariance system depending on wind sector can be obtained from Table 1. of the Black Forest, Southwest Germany.The Kinzig valley at Fußbach is oriented in aN-S direction, thus specifying the main wind direction, and has avalleywidth of about 1.3 km. Mountain crests in the immediate vicinity of the Fußbach site reach maximal values of about 450ma.s.l. The target land use type was acorn field (length: 260 m, width: 140m) located within the patchy,agricultural land use of the Kinzig valley(see Fig. 1). In this study,dataof an EC tower and anearby Sodar/RASS system are used. The EC system(measurement height: 2.29 m, sampling rate:20Hz) measured turbulent fluxes of momentum, sensible and latent heat as well as carbon dioxide (CO2)abovethe corn field using aCSAT3 (Campbell Scientific, Inc.) sonic anemometer for recording the wind vector and the sonic temperature TSand aLI-7500 (LI-COR Biosciences) open-path gas analyser for water vapor (H2O) and CO2concentrations. The Sodar/RASS system consisted of aphase array Doppler Sodar DSDPA.90-64 with a1290 MHz RASS extension by Metek GmbH and provided vertical profiles of wind velocity components, wind direction and acoustic temperature with avertical resolution of 20m and atemporal resolution of 10min. More detailed information about the measuring set-up and background data can be obtained from Metzger et al. (2007). 2.2 Quality control effort The EC flux data measured at the Fußbach site was processed and quality controlled applying the latest micrometeorological post-field data processingstandards (e.g., Mauder et al., 2006)in order to obtain adata set of the desired quality and accuracy,which can be utilized for further fundamental Atmos. Chem. Phys., 9, 8587–8600,2009 www.atmos-chem-phys.net/9/8587/2009/ A PPENDIX B – E IGENMANN ET AL . (2009) 43 R. Eigenmann et al.: Generation of free convection due to changes of the local circulationsystem 8589 research. Accordingly,the turbulent flux rawdata recorded with the EC method was post-processed with the comprehensivesoftware package TK2 developed at the University of Bayreuth (Mauder and Foken,2004), which comprises state of the art flux corrections and post-field quality control including tests on thefulfillment of integral turbulence characteristics and stationarity (Foken and Wichura,1996;Foken et al.,2004). Theoretical assumptions actually restrict the EC method to homogeneous terrain, but the increasing requirement for continuous monitoring of flux data (Aubinet et al.,2000;Baldocchi et al.,2001)forced the application of the EC method within highly structured terrain such as that in the COPS region. Thisstep is supported by thedevelopmentofasite evaluation and characterisation approach (G¨ ockede et al.,2004, 2006), which combines the flux data quality approach (Foken et al.,2004)withaforward Lagrangian footprint model (Rannik et al.,2000,2003). The approach is able to identify site-specific spatial quality structures and the spatial representivity of the measured flux data in the context of the underlying land use distribution and has been recently employed on sites of the CarboEurope network by Rebmann et al. (2005)and G¨ ockede et al. (2008). In this study,it is used –together with an internal boundary layer evaluation procedure –in order to obtain target land use typerepresentativeturbulent flux data sets of the required high quality usable for further analyses. The check for possible internal boundary layers, which form as aconsequence of changes ofthe underlying surface characteristics, is implemented by using the following fetch-height relation to roughly estimate the height of the newequilibrium layer δdepending on fetch x(Raabe,1983;Jegede and Foken, 1999)neglecting weak stability effects (Savelyevand Taylor,2005): za≤δ=0.3√x(1) The aerodynamic measurement height zashould be lower than δin order to guarantee that the EC measurement takes place within the newequilibrium layer establishing over the target land use type. 2.3 Spectral analysis Spectral analysis methods are used in this paper to study the temporal scales of the turbulence duringthe period of FCCs (see Sect. 3.2). Methods used are the continuous wavelet transform (CWT) and the computation of power spectra, both applied to the time series of EC rawdata of the vertical wind speed, the horizontal wind speed components andthe sonic temperature from 05:00 to 13:00UTC. Toprepare data for the CWT,the time series were block averaged from the original 20Hz rawdata to asampling frequencyof 0.5Hz in order to drastically reduce computational time of the CWT without altering the results significantly (e.g., Thomas and Foken,2005)as the decisive scales range in the order of several seconds to afewminutes. Subsequently,all block averaged time series apart from the vertical wind speed (no trend removal necessary) have been detrended using polynomial regression. The resulting high-pass filtered time series is obtained by substracting the fitted polynomial from the block averaged time series. The residual time series is used for the calculation of the CWT using the Morlet wavelet. The CWT and plotting routine of the normalized wavelet power spectra was done by using the software package sowas (software for wavelet spectral analysis and synthesis: Maraun and Kurths,2004;Maraun et al., 2007)implemented in the statistical computing software R. Results of apointwise significance test (significance level: 0.95) performed by Monte Carlo simulations (1000 realizations) are indicated by black solid lines in the plots of the normalized wavelet power spectra. The calculation of the power spectra of detrended (by polynomial regression) and tapered 20Hz time series of the vertical wind speed, the longitudinal wind speed and sonic temperature before, during and after the period of FCCs within the time range from 05:00 to 13:00UTC was realized by applying afast Fouriertransform (FFT) to the computed autocorrelation function. Smoothing of the rawperiodogram was performed using amodified Daniell smoother window technique. 3Results and discussion 3.1 Quality control of the turbulent flux measurements This section presents some results of the detailed quality control effort adapted to the eddy-covariance flux data as described in Sect. 2.2. Processed turbulent fluxes of sensible (QH)and latent heat (QE), friction velocity u∗as well as the CO2net ecosystem exchange NEE are depicted in Fig. 2as Hovmøller-type plots where the color baron the right side represents the calculated values with white areas indicating data failure (9.9% for each flux). Furthermore, footprint analysis (see Sect. 2.2)were performed according to the site evaluation and characterisation approach of G¨ ockede et al. (2004,2006). To gain insight into the average flux contribution over the entire measurement period of the target land use type “corn” as afunction ofdifferent wind sectors and stability classes, appropriate sorted data havebeen individually processed within the footprint analysis procedure. The results are listed in Table 1and reveal good average flux contributions of more than 92% for all wind sectors during unstable or neutral cases. However, for stable stratification easterly and westerly sectors haveto be considered critically,as flux contributions below80% can be found, with the minimum (67%) in the 240◦wind sector.Admittedly,the latter finding has to be regarded with the knowledge that the densityof available data is lowin the www.atmos-chem-phys.net/9/8587/2009/ Atmos. Chem. Phys., 9, 8587–8600,2009 50 A PPENDIX B – E IGENMANN ET AL . (2009) 8596 R. Eigenmann et al.: Generation of free convection dueto changes of the local circulation system Table 2. Mean onset and cessation times [UTC] of the up-valleywind direction and the mean times of FCCs of those days classified as “event days”, with standard deviation (SD), and number (n) for the individual months (June, July,August) and for the whole COPS measurement period at Fußbach. Onset of up-valleywind direction Cessation of up-valleywind direction FCCs Period Mean SD nPeriod Mean SD nPeriod Mean SD n June 08:01 01:14 8June 17:36 00:55 7June 07:33 01:08 8 July 08:03 00:37 5July 17:23 01:35 5July 08:14 00:51 5 August 09:06 00:46 10 August 17:56 01:10 10 August 08:48 00:42 10 whole 08:29 01:02 23 whole 17:42 01:10 22 whole 08:14 01:01 23 01/06/2007 24/06/2007 17/07/2007 09/08/2007 01/09/2007 00:00 06:00 12:00 18:00 00:00 Time [UTC] onset cessation FCCs sunrise sunset Fig.7. Onset and cessation times of the up-valleywind direction and the corresponding periods of FCCs in the morning hours at days classified as “event days” regarding the entire COPS measurement period at Fußbach. Also depicted are the times of sunrise and sunset. classified as “event days” coinciding with the paradigm of the generation of FCCs at COPS IOP8b outlined in Sect. 3.2. Furthermore, 19 days (21%) can be denoted as “intermittent days”, as they do not exhibit a clear diurnal, persistent valley wind circulation. FCCs occur but can only be attributed to brief duration –several minutes up to afewhours –changes from down-valleytoup-valleywinds during the day,these sometimes not even reaching afull wind rotation of180◦. Despite the fact that most of the FCCs of these “intermittent days” seem to be triggered by short changes of wind direction of varying duration, it was decided to separate the “intermittent” from the “event days” inorder to havesimilar flow patterns initiating the FCCs and thus aclearly structured data set. The reason for the intermittence of the valleywinds is a decrease of the solar energy input, e.g. due to cloud shading. Finally,37 days (40%) at Fußbach can be characterized as “non-event days”, as FCCs do not appear.Thirteen days (14%) cannot be evaluated due to data failure. Figure7showsall days classifiedas“eventdays”(23) with the onset and cessation times of the up-valleywind direction, the periods in which FCCs occurred and the times of sunrise and sunsetduring the entire COPS campaign. The mean onset and cessation times of the up-valleywind direction andthe mean times of FCCs of those days classified as “event days”, with standard deviation and number,are listed in Table 2for the individual months and the whole measurement period. Remarkable is the adjustment of the onset of the up-valleywind direction and of the FCCs to the seasonal change of sunrise evident in themean valuesof the individual months. The mean duration of the periods in which FCCs occur –depicted in Fig. 7 – is 84 min with astandard deviation of 57min. Toclarify the difference of “event” and “intermittent days” as related to the diurnal valleywind system,persistence values Pwere calculated from the west-east and south-north components (uand v,respectively) and the horizontal wind speed vhof the EC sonic anemometer following Lugauer and Winkler (2005): P(t)=qu(t)2+v(t)2 vh(t)(6) where the temporal vector mean of the horizontal wind speed is divided by the arithmetic mean of the horizontal wind speed at every time of the day.Pcan adopt values between 0and 1, where 1means that every day at that time the wind blewfrom the same direction. Figure 8adepicts the calculated persistence Pfor all days, “event days”, “intermittent days” and “non-event days”. Twopronounced eye-catching minima can be found the “event days”, i.e. one at the times the up-valleywinds start (at about 08:00UTC) and another at the times the up-valleywinds rotate back to down-valley winds (at about 18:00UTC), indicating highly variable wind directions during these times. The Pvalues above0.75 for the rest of the time point to aquasi-identical flowpattern on the 23 selected “event days” at Fußbach. The small persistence values of the “intermittent days” between 06:00 and 18:00UTCcan be attributed to the non-persistence of the up-valleywinds, as briefly lasting up-valleywindsare interrupted by frequent rotations back to the original wind directiondueto adecrease of thesolarenergyinput (through cloud shading) which otherwise usually drives constantly blowing up-valleywinds. Atmos. Chem. Phys., 9, 8587–8600,2009 www.atmos-chem-phys.net/9/8587/2009/ A PPENDIX B – E IGENMANN ET AL . (2009) 51 R. Eigenmann et al.: Generation of free convection due to changes of the local circulationsystem 8597 0 0.25 0.5 0.75 1 0:00 6:00 12:00 18:00 0:00 P [-] all days (n=92) “event days” (n=23) (a) -40 0 40 80 120 0:00 6:00 12:00 18:00 0:00 QH [W m-2] “intermittent days” (n=19)“non-event days” (n=37) (d) -2 -1 0 1 2 3 0:00 6:00 12:00 18:00 0:00 ζ [-] (b) 0 300 600 900 0:00 6:00 12:00 18:00 0:00 Global radiation [W m-2] (e) 0 0.1 0.2 0.3 0.4 0:00 6:00 12:00 18:00 0:00 Time [UTC] u* [ms-1] (c) 0 100 200 300 400 0:00 6:00 12:00 18:00 0:00 Time [UTC] QE [W m-2] (f) Fig.8. Persistence P[−](a) and mean diurnal trends of the stability parameter ζ[−](b),friction velocity u∗hms−1i(c),sensible heat flux QHhWm−2i(d),global radiation hWm−2i(e) and latent heat flux QEhWm−2i(f) for all days of the COPS measurement period (92), days classified as “event days” (23), “intermittent days” (19) and “non-event days” (37) at Fußbach. Thirteen days cannot be classified due to data failure. Mean diurnal trends of ζ(Fig. 8b), u∗(Fig. 8c), QH (Fig. 8d), global radiation (Fig. 8e) and QE(Fig. 8f) of the individually classified days can be applied to characterize the FCCs. The mean values of ζindicate the occurrence of FCCs in the morning hours at about 08:00UTC on both “intermittent” and “event days”. After this distinct minima of ζat about 08:00UTC (until about 12:00UTC), the “intermittent days” showslightly lower values of ζ,indicating that the intermittence of the valleywind system causes more frequent triggering of FCCs. Aclear difference in the magnitude of u∗comparing “event” and “non-event days” (see Fig. 8c) at about 08:00UTC underlines that the drop of u∗caused by the valleywind reversal in the morning hours triggers FCCs. The generally lower values of u∗on the “intermittent days” can be explained by the non-persistence of the valleywinds resulting in acontinuous drop of the horizontal wind speed. Above-average values of QHand global radiation (Figs.8d and 8e) can be found on the “event days”, thus indicating the preferred occurrence of FCCs inclear,undisturbed weather situations with high solar radiation driving the valleywinds, which act as the trigger mechanism for FCCs by providing the necessary minimum of u∗in the early morning transition period. Moreover,about 100Wm−2higher QEvalues (Fig. 8f) can be observed on the “event days” compared to the “non-event days”, which also contribute, due to density effects, to the destabilization of near-ground air masses. www.atmos-chem-phys.net/9/8587/2009/ Atmos. Chem. Phys., 9, 8587–8600,2009 52 A PPENDIX B – E IGENMANN ET AL . (2009) 8598 R. Eigenmann et al.: Generation of free convection dueto changes of the local circulation system 4Conclusions Acomprehensivequality assurance and control effort, including footprint analysis and acheck for internal boundary layers, was adapted to the EC turbulence measurements at Fußbach in order to obtain high-quality surface flux data usable for the detection (ζ<−1) and description of FCCs in the Kinzig valley.FCCs were found to be triggered by achange of the local valleycirculation system from down to up-valley winds in the morning hours and occurred on about half of the total 92 COPS days summing up days classified as “event” and “intermittent days”. This frequent occurrence of FCCs confirms the assumption of Mayer et al. (2008)that other regions showing complexterrain –besides the alpine foreland investigated in their study –might face trigger mechanisms, such as local or mesoscale circulation systems leading to the convectiverelease of near-ground air masses into the ABL. FCCs initiated by achange of the valleywinds were also found by Hiller et al. (2008)in an alpine valleyin Switzerland following their stability and data quality analysis. Unfortunately,these authors did not address these events. The large-eddy scale character of the turbulence near the ground during periods of FCCs could be confirmed by applying spectral analysis methods (see Figs. 5and 6), thus suggesting that plume-likecoherent structures of rising air masses emerge from the detected FCCs. These large-scale structures can more effectively transport quantities of heat and moisture enhanced in near-ground regions intothe ABL. As enhanced surfacefluxes of latent and sensible heat were found on the “event days” compared to “non-eventdays” (see Fig. 8)as well as increased upward vertical wind speeds in the Sodar measurements during FCCs (see Figs. 3eand 3f), aclear effect of the detected FCCs on the thermodynamic structure of the ABL is obvious. Tosum up, FCCs arelikely –in addition to other orographic or landscape effects –to have a non-negligible impact on ABL temperature and moisture profiles and to play arole for CI processes. In order to directly simulate the impact of FCCs on ABL thermodynamics and on possiblysubsequent cloud formation, alarge-eddy simulation (LES) model will be applied for further investigations. Acknowledgements. The project was funded within the Priority Program 1167 “QuantitativePrecipitation Forecast PQP (Praecipitationis Quantitativae Predictio)” by the German Science Foundation (DFG), second and third (Fo226/19-1) period. The authors wish to acknowledge the support and data provision by the participants of the COPS experiment and the COPS Operations Center as well as Bj¨ orn Br¨ otz for manyfruitful discussions and comments. Last but not least, the authors want to thank all people which took part in thefield work, especially Andrei Serafimovich, Lukas Siebicke, Katharina Staudt, Johannes L¨ uers and Johannes Olesch. Edited by: G. Vaughan References Aubinet, M., Grelle, A., Ibrom, A., Rannik, U., Moncrieff, J., Foken, T., Kowalski, A. S., Martin,P.H., Berbiger,P., Bernhofer, C., Clement, R., Elbers, J. A., Granier,A., Gr¨ unwald, T., Morgenstern, K., Pilegaard, K., Rebmann, C., Snijders, W., Valentini, R., and Vesala, T.: Estimates of the Annual Net Carbon and Water Exchange of Forests The EUROFLUX Methodology,Adv. Ecol. Res., 30, 113–176, 2000. Baldocchi, D., Falge, E., Gu, L. H., Olson, R., Hollinger,D., Running, S., Anthoni, P., Bernhofer,C., Davis, K., Evans, R., Fuentes, J., Goldstein, A., Katul, G., Law,B., Lee, X. H., Malhi, Y., Meyers, T., Munger,W., Oechel, W., U, K. T.P., Pilegaard, K., Schmid, H. P., Valentini, R., Verma, S., Vesala, T., Wilson, K., and Wofsy,S.: FLUXNET: Anewtool to study the temporal and spatial variability of ecosystem-scale carbondioxide, water vapor,and energy flux densities, B. Am. Meteorol. Soc., 82, 2415–2434, 2001. Banta, R. M.: Daytime Boundary-Layer Evolution over Mountainous Terrain. Part 1: Observationsof the Dry Circulations, Mon. Weather.Rev., 112, 340–356, 1984. Banta, R. M.: The role of mountain flows in making clouds, in: Atmospheric processes over complexterrain, edited by Blumen, W., vol. 23 (45) of Meteorological monographs,pp. 229–283, American Meteorological Society,1990. Barthlott, C., Corsmeier,U., Meißner,C., Braun, F., and Kottmeier, C.: The influence of mesoscale circulation systems on triggering convectivecells overcomplexterrain, Atmos. Res., 81, 150–175, 2006. Bergstr¨ om, H. and H¨ ogstr¨ om, U.:Turbulent exchange above a pine forest. II.Organized structures., Bound.-Lay.Meteorol., 49, 231– 263, 1989. Betts, A. K., Ball, J. H., Beljaars, A. C. M., Miller,M. J., and Viterbo, P.A.: The land surface-atmosphere interaction: Areviewbased on observational and global modeling perspectives, J. Geophys. Res.-Atmos., 101, 7209–7225, 1996. Beyrich, F.: Mixing height estimation from sodar data –Acritical discussion, Atmos. Environ., 31, 3941–3953, 1997. Chen, F.and Avissar,R.: Impact of Land-Surface Moisture Variability on Local ShallowConvectiveCumulus and Precipitation in Large-Scale Models, J. Appl. Meteorol., 33, 1382–1401, 1994. Deardorff, J. W.: ConvectiveVelocity and Temperature Scales for the Unstable Planetary Boundary Layer and for Rayleigh Convection, J. Atmos. Sci., 27, 1211–1213, 1970a. Deardorff, J. W.: Preliminary Results from Numerical Integrations of the Unstable Planetary Boundary Layer,J. Atmos. Sci., 27, 1209–1211, 1970b. Eigenmann, R.: Investigation of conditions initiating free convection using energy exchangemeasurements. COPS-experiment, Black Forest, 2007, Master’sthesis, University of Bayreuth, Germany,2008. Foken, T.: The energy balance closure problem: An overview,Eco. Appl., 18, 1351–1367, 2008a. Foken, T.: Micrometeorology,Springer,NewYork, 1st edn., 2008b. Foken, T.and Wichura, B.: Tools for quality assessment of surfacebased flux measurements, Agr.Forest. Meteorol., 78, 83–107, 1996. Foken, T., G¨ ockede, M., Mauder,M., Mahrt, L., Amiro, B. D., and Munger J. W.: Post-field data quality control, in: Handbook of Micrometeorology: AGuide for Surface Flux Measurement and Atmos. Chem. Phys., 9, 8587–8600,2009 www.atmos-chem-phys.net/9/8587/2009/ A PPENDIX B – E IGENMANN ET AL . (2009) 53 R. Eigenmann et al.: Generation of free convection due to changes of the local circulationsystem 8599 Analysis, edited by Lee, X., Massman, W., and Law,B., 181– 208, Kluwer,Dordrecht, 2004. Foken, T., Wimmer,F., Mauder,M., Thomas, C., and Liebethal, C.: Some aspects of the energy balance closure problem, Atmos. Chem. Phys., 6, 4395–4402, 2006. Gao, W., Shaw,R. H., and PawU, K. T.: Observation of organized structures in turbulent flowwithin and above a forest canopy, Bound.-Lay.Meteorol., 47, 349–377, 1989. G¨ ockede, M., Rebmann, C., and Foken, T.: Acombination of quality assessment tools for eddy covariance measurements with footprint modelling for the characterisation of complexsites, Agr.Forest. Meteorol., 127, 175–188,2004. G¨ ockede, M., Markkanen, T., Hasager,C. B., and Foken, T.: Update of afootprint-based approach for the characterisation of complex measurement sites, Bound.-Lay.Meteorol., 118, 635–655, 2006. G¨ ockede, M., Foken,T., Aubinet, M., Aurela, M.,Banza, J., Bernhofer,C., Bonnefond, J. M., Brunet, Y., Carrara, A., Clement, R., Dellwik, E., Elbers, J., Eugster,W., Fuhrer,J., Granier,A., Grunwald, T., Heinesch, B., Janssens, I. A., Knohl, A., Koeble, R., Laurila, T., Longdoz, B., Manca, G., Marek, M., Markkanen, T., Mateus, J., Matteucci, G., Mauder,M., Migliavacca, M., Minerbi, S., Moncrieff, J., Montagnani, L., Moors, E., Ourcival, J. M., Papale, D., Pereira, J., Pilegaard, K., Pita, G., Rambal, S., Rebmann, C., Rodrigues, A., Rotenberg, E., Sanz, M. J., Sedlak, P., Seufert, G., Siebicke, L., Soussana, J. F., Valentini, R., Vesala, T., Verbeeck, H., and Yakir,D.: Quality control of CarboEurope flux data -Part 1: Coupling footprint analyses with flux data quality assessment to evaluate sites in forest ecosystems, Biogeosciences., 5, 433–450, 2008. Gurvich, A. S.: Experimental investigation of frequencyspectra of the vertical wind velocity in the atmospheric surface layer, Dokl. Akad. Nauk SSSR, 132, 806–809, 1960. Gurvich, A. S.: Spectra of the vertical wind-velocity fluctuations and their relation to micrometeorological conditions, Atmospheric Turbulence, (Proc. In-ta Fiziki Atmosf. Akad. Nauk SSSR, No. 4), 101–136, 1962. Hanesiak, J. M., Raddatz, R. L., and Lobban, S.: Local initiation of deep convection on the Canadian prairie provinces, Bound.-Lay. Meteorol., 110, 455–470, 2004. Hiller,R., Zeeman, M. J., and Eugster,W.: Eddy-covariance flux measurements in the complexterrain of an Alpine valleyin Switzerland, Bound.-Lay.Meteorol., 127, 449–467, 2008. Inagaki, A., Letzel, M. O., Raasch, S., and Kanda, M.: Impact of surface heterogeneity on energy imbalance: Astudy using LES, J. Meteorol. Soc. Jpn., 84, 187–198, 2006. Jegede, O. O. and Foken, T.: Astudy of the internal boundary layer due to aroughness change in neutral conditions observed during the LINEX field campaigns, Theor.Appl. Climatol., 62,31–41, 1999. Kaimal, J. C., Wyngaard, J. C., Izumi, Y., and Cote, O. R.: Spectral characteristics of surface-layer turbulence, Q. J. Roy.Meteor. Soc., 98, 563–589, 1972. Kalthoff, N., Binder,H. J., Kossmann, M., Vogtlin, R., Corsmeier, U., Fiedler,F., and Schlager,H.: Temporal evolution and spatial variation of the boundary layer over complexterrain, Atmos. Environ., 32, 1179–1194, 1998. Kalthoff, N., Horlacher,V., Corsmeier,U., Volz Thomas, A., Kolahgar,B., Geiss, H., Mollmann Coers, M., and Knaps, A.: Influence of valleywinds on transport and dispersion of airborne pollutants in the Freiburg-Schauinsland area, J. Geophys. Res.- Atmos., 105, 1585–1597, 2000. Kanda, M., Inagaki, A., Letzel, M. O., Raasch, S., and Watanabe, T.: LES study of the energy imbalance problem with Eddy covariance fluxes, Bound.-Lay.Meteorol., 110, 381–404, 2004. Katul, G. G. and Chu, C. R.: Atheoretical and experimental investigation ofenergy-containing scales in the dynamic sublayer of boundary-layerflows, Bound.-Lay.Meteorol., 98, 279–312, 1998. Katul, G. G., Chu, C.R., Parlange, M. B., Albertson, J. D., and Ortenburger,T.A.: Low-wavenumber spectral characteristics of velocity and temperature in the atmospheric surface layer, J. Geophys. Res., 100, 14243–14255, 1995. Kolmogorov,A. N.: Rassejanie energii pri lokolno-isotropoiturbulentnosti, Dokl. AN. SSSR., 32, 22–24, 1941. Kossmann, M. and Fiedler,F.: Diurnal momentum budget analysis of thermally induced slope winds, Meteorol. Atmos. Phys., 75, 195–215, 2000. Kossmann, M.,Vogtlin, R., Corsmeier,U., Vogel, B., Fiedler,F., Binder,H. J., Kalthoff, N., and Beyrich, F.: Aspects of the convectiveboundary layer structure over complexterrain, Atmos. Environ., 32, 1323–1348, 1998. Kottmeier,C., Kalthoff, N., Corsmeier,U., Barthlott, C., van Baelen, J., Behrendt A., Behrendt, R., Blyth, A., Coulter,R., Crewell, S.,Dorninger,M., Foken, T., Hagen, M., Hauck, C., H¨ oller,H., Konow,H., Kunz, M., Mahlke, H., Mobbs, S., Richard, E., Steinacker,R., Weckwerth, T., and Wulfmeyer,V.: Mechanisms initiating convection during the COPS experiment, Meteorol. Z., 17, 931–948, 2008. Liebethal, C. and Foken, T.: Evaluation of six parameterization approaches for the ground heat flux, Theor.Appl. Climatol., 88, 43–56, 2007. Lugauer,M. and Winkler,P.: Thermal circulation in South Bavaria - climatology and synoptic aspects, Meteorol. Z., 14, 15–30, 2005. Maraun, D. and Kurths, J.: Cross Wavelet Analysis. Significance Testing and Pitfalls, Nonlinear.Proc. Geoph., 11, 505–514, 2004. Maraun, D., Kurths, J., and Holschneider,M.: Nonstationary Gaussian Processes in Wavelet Domain: Synthesis, Estimation and Significance Testing, Phys. Rev.E., 75, 016707, doi:10.1103/PhysRevE.75.016707, 2007. Mauder,M. and Foken, T.: Documentation and instruction manual of the Eddy covariance software package TK2, Work Report University of Bayreuth, Department of Micrometeorology,26, ISSN: 1614-8916, 42 pp., 2004. Mauder,M., Liebethal, C., G¨ ockede, M., Leps, J. P., Beyrich,F., and Foken, T.: Processing and quality control of flux data during LITFASS-2003, Bound.-Lay.Meteorol., 121, 67–88, 2006. Mayer,J. C., Staudt, K., Gilge, S., Meixner,F.X., and Foken, T.: The impact of free convection on late morning ozone decreases on an Alpine foreland mountain summit, Atmos. Chem. Phys., 8, 5941–5956, 2008. Meißner,C., Kalthoff, N., Kunz,M., and Adrian, G.: Initiation of shallowconvection in the Black Forest mountains, Atmos. Res., 86, 42–60, 2007. Metzger,S., Foken, T., Eigenmann, R., Kurtz, W., Serafimovich, A., Siebicke, L., Olesch,J., Staudt, K., and L¨ uers, J.: COPS experiment, Convectiveand orographically induced precipitation study, 01 June 2007 – 31 August 2007, Documentation, Work Report University of Bayreuth, Department of Micrometeorology,34, www.atmos-chem-phys.net/9/8587/2009/ Atmos. Chem. Phys., 9, 8587–8600,2009 54 A PPENDIX B – E IGENMANN ET AL . (2009) 8600 R. Eigenmann et al.: Generation of free convection dueto changes of the local circulation system ISSN 1614-8916, 72 pp., 2007. Monin, A. S. and Yaglom, A. M.: Statistical Fluid Mechanics, Volume II: Mechanics of Turbulence, MIT Press, Cambridge, Massachusetts, USA, and London, England,UK, 1975. Pielke, R. A.: Influence of the spatial distribution of vegetation and soils on the prediction of cumulus convectiverainfall, Rev.Geophys., 39, 151–177, 2001. Raabe, A.: On the relation betweenthe drag coefficient and fetch abovethe sea in the case of off-shore wind in the near-shore zone, Z. Meteorol., 33, 363–367, 1983. Rabin, R. M., Stadler,S., Wetzel, P.J., Stensrud, D. J., and Gregory, M.: Observed Effects of Landscape Variability on Convective Clouds, B. Am. Meteorol. Soc., 71, 272–280, 1990. Rannik, U., Aubinet, M., Kurbanmuradov,O., Sabelfeld, K. K., Markkanen, T., and Vesala, T.: Footprint analysis for measurements over aheterogeneous forest, Bound.-Lay.Meteorol., 97, 137–166, 2000. Rannik, U., Markkanen, T., Raittila, J., Hari, P., and Vesala, T.: Turbulence statistics inside and over forest: Influence on footprint prediction, Bound.-Lay.Meteorol., 109, 163–189, 2003. Raymond, D. and Wilkening, M.: Mountain-Induced Convection under Fair Weather Conditions, J. Atmos. Sci., 37, 2693–2706, 1980. Rebmann, C., G¨ ockede, M., Foken, T., Aubinet, M., Aurela, M., Berbigier,P., Bernhofer,C., Buchmann, N., Carrara, A., Cescatti, A., Ceulemans, R., Clement, R., Elbers, J. A., Granier,A., Grunwald, T., Guyon, D., Havrankova, K., Heinesch, B., Knohl, A., Laurila, T., Longdoz, B., Marcolla, B., Markkanen,T., Miglietta, F., Moncrieff,J., Montagnani, L., Moors, E., Nardino, M., Ourcival, J. M., Rambal, S., Rannik, U., Rotenberg, E., Sedlak, P., Unterhuber,G., Vesala, T., and Yakir,D.: Quality analysis applied on eddy covariance measurements at complexforest sites using footprintmodelling, Theor.Appl. Climatol., 80, 121–141, 2005. Savelyev,S. A. and Taylor,P.A.: Internal boundary layers: I. Height formulae for neutral and diabatic flows, Bound.-Lay.Meteorol., 115, 1–25, 2005. Schwitalla, T., Bauer,H.-S., Wulfmeyer,V., and Zaengl, G.: Systematic errors of QPF in low-mountain regions as revealed by MM5 simulations, Meteorol. Z., 17, 903–919, 2008. Segal, M. and Arritt, R. W.: Nonclassical mesoscale circulations caused by surface sensible heat-flux, B. Am. Meteorol. Soc., 73, 1593–1604, 1992. Shen, S. H. and Leclerc, M. Y.: Howlarge must surface inhomogeneities be before theyinfluence the convectiveboundary layer structure? Acase study,Q. J. Roy.Meteor.Soc., 121, 1209– 1228, 1995. Staudt, K.: Determination of the atmospheric boundary layer height in complexterrain duringSALSA 2005, Master’sthesis, University of Bayreuth, Germany,2006. Steinfeld, G., Letzel, M. O., Raasch, S., Kanda, M., and Inagaki, A.: Spatial representativeness of single tower measurements and the imbalance problem with eddy-covariance fluxes: results of a large-eddy simulation study,Bound.-Lay.Meteorol., 123, 77–98, 2007. Stull, R. B.: An introduction to boundary layer meteorology,Atmospheric sciences library,Kluwer Academic Publishers, Dordrecht, The Netherlands, 1988. Stull, R. B.: Meteorology for scientists and engineers, Brooks/Cole, Pacific Grove, California, 2. edn., 2000. Thomas, C. and Foken, T.: Detection of long-term coherent exchange over spruce forest using wavelet analysis, Theor.Appl. Climatol., 80, 91–104, 2005. Whiteman, C. D.: Observations of thermally developed wind systems in mountainous terrain, in: Atmospheric processes over complexterrain, edited by Blumen, W., vol. 23 (45) of Meteorological monographs,5–42, American Meteorological Society, Boston, Massachusetts, USA, 1990. Wulfmeyer,V., Behrendt, A., Adrian, G., Althausen, D., Aoshima, F., van Baelen, J., Barthlott, C., Bauer,H. S., Blyth, A., Brandau, C., Corsmeier,U., Craig, G., Crewell, S., Dick, G., Dorninger, M., Dufournet, Y., Ehret, G., Engelmann, R., Flamant, C., Foken, T., Hauck, C., Girolamo, P.D., Graßl, H., Grzeschik, M., Handwerker,J., Hagen, M., Hardesty,R. M., Junkermann,W., Kalthoff, N., Kiemle, C., Kottmeier,C., Krauss, L., Long, C., Lelieveld, J., Madonna, F., Miller,M., Mobbs, S., Neininger, B., Pal, S., Peters, G., Radlach, M., Richard, E., Rotach, M., Russchenberg, H., Schl¨ ussel, P., Schumann, U., Simmer,C., Steinacker,R., Turner,D., Vogt, S., Volkert, H., Weckwerth, T., Wernli, H., Wieser,A., and Wunraum, C.: Convective and Orographically-induced Precipitation Study.COPS Field Report 2.1., https://www.uni-hohenheim.de/spp-iop/documents/ COPSFieldReport2.pdf,2007. Wulfmeyer,V., Behrendt, A., Bauer,H.-S., Kottmeier,C., Corsmeier,U., Blyth, A., Craig, G., Schumann, U., Hagen, M., Crewell, S., Di Girolamo, P., Flamant, C., Miller,M., Montani, A., Mobbs, S., Richard, E., Rotach, M. W., Arpagaus, M., Russchenberg, H., Schluessel, P., Koenig, M., Gaertner,V., Steinacker,R., Dorninger,M., Turner,D. D., Weckwerth, T., Hense, A., and Simmer,C.: The Convectiveand Orographicallyinduced Precipitation Study: AResearch and Development Project of the World Weather Research Program for Improving QuantitativePrecipitation Forecasting in Low-Mountain Regions, B. Am. Meteorol. Soc., 89, 1477–1486, 2008. Zubkovskii, S. L.: Frequencyspectra of the horizontal windvelocity fluctuations in the atmospheric surface layer, Izv.Akad. Nauk SSSR, Ser.Geofiz, No. 10, 1425–1433, 1962. Atmos. Chem. Phys., 9, 8587–8600,2009 www.atmos-chem-phys.net/9/8587/2009/ A PPENDIX C – E IGENMANN ET AL . (2011) 55 Appendix C: Eigenmann et al. (2011) Surface energy balance and turbulence network during the Convective and Orographically-induced Precipitation Study (COPS) R. Eigenmann,a*N.Kalthoff, bT. Foken,aM. Dorninger,cM. Kohler,bD. Legain,dG. Pigeon,d B. Piguet,dD. Sch¨ uttemeyereand O.Traulled aDepartment of Micrometeorology, UniversityofBayreuth, Bayreuth,Germany bInstitute for Meteorology and Climate Research, Karlsruhe Institute of Technology,Karlsruhe, Germany cDepartment of Meteorology and Geophysics, University ofVienna, Vienna, Austria dCentre National de Recherches M´ et´ eorologiques –Groupe d’´ etude del’Atmosph` ere M´ et´ eorologique (CNRM-GAME), M´ et´ eo France –CNRS, Toulouse, France eMeteorological Institute, University of Bonn, Bonn, Germany *Correspondence to: R. Eigenmann, Department of Micrometeorology, University of Bayreuth, Universit¨ atsstrasse 30, 95440 Bayreuth, Germany. E-mail: r[email protected] Experimental data of the energy balance and the turbulence network installed duringtheConvectiveandOrographically-induced Precipitation Study (COPS) field campaign of2007 are presented in this study. The network aims at providing continuous surface flux and other surface micrometeorological data of the required high accuracy and quality for further fundamental research. An overview of the turbulence data processing and data quality control, including footprint analysis and acheck for internal boundary layers, isgiven. The consistently applied approach allows for ahigh comparability of the turbulent flux data ofsensible and latent heat. The reaction of surface fluxes during the observed frontal passage during the Intensive Observation Period (IOP) 9c(20 July 2007) is presented. As surface fluxes were measured over different land-use types and at different locations within the COPS area, the effect of land use and orography onturbulent fluxes isdiscussed with the help ofIOP 8b (15 July 2007). The flux differences between individual sites due tovarying surface characteristics are often larger than the flux differences with changing altitude. The oasis effect observed for the highly evapotranspirating maize fields isfound to increase the residuum of thesurface energy balance. At all sites and during both IOPs the occurrence ofnear-ground free convection conditions (FCCs) is investigated. During the oasis effect, FCCs do not occur. Copyright c 2011 Royal Meteorological Society Key Words: eddy-covariance; turbulent fluxes; energy balance closure; residuum; oasis effect; near-ground free convection conditions Received 12 February 2010; Revised 27August 2010; Accepted 31 August 2010; Published online inWiley Online Library 2February 2011 Citation: Eigenmann R,KalthoffN,Foken T, Dorninger M,Kohler M,Legain D,Pigeon G,Piguet B, Sch¨ uttemeyer D,Traulle O. 2011. Surface energy balance and turbulence network during the Convective and Orographically-induced Precipitation Study (COPS). Q.J. R. Meteorol. Soc. 137:57–69. DOI:10.1002/qj.704 1. Introduction Anetwork of surface energy balance and turbulent flux measurement stations was set upduring the comprehensive Convective and Orographically-induced Precipitation Study (COPS) field campaign (Wulfmeyer et al.,2008). The campaign took place insouthwestern Germany and eastern France from 1June to 31 August 2007. It was Copyright c 2011 Royal Meteorological Society 56 A PPENDIX C – E IGENMANN ET AL . (2011) organized into Intensive Observation Periods (IOPs), which observed specific convective situations with asynergy of meteorological instruments (Wulfmeyer et al.,2011). Theaim of theCOPS energy balance and turbulence network was to provide information about the temporal and spatial heterogeneity ofhigh-quality turbulent flux values of sensible and latent heat for individual IOPs. This is important, since under weak synoptic forcing spatial inhomogeneities of surface characteristics result in inhomogeneities ofturbulent fluxes ofheat and moisture into the atmospheric boundary layer (ABL) and hence may determine ifand where convection isinitiated (e.g. Aoshima et al.,2008; Kottmeier et al.,2008, Behrendt et al.,2011; Bennett et al.,2011; Weckwerth et al.,2011). Ananalysis of flux measurements with amore climatological focus was already performed inthe COPS area within previous studies (Wenzel et al.,1997). These investigations revealed asignificant dependence ofthe energy balance components on the altitude and onthe land-use types inthe area of the UpperRhine valleyandthe BlackForest (Kalthoffetal.,1999, Wenzel and Kalthoffet al.,2000). The turbulence network of the COPS campaign iscomparable tothat ofthe LITFASS2003 experiment(Mengelkamp et al.,2006), where turbulent fluxes were measured within anarea of20 ×20 km2over all relevant land-use types and were area-averaged bymeans of atile approach (Avissar, 1991). The aggregated surface fluxes showed good agreement with area-integrated fluxes from long-range scintillometer and airborne measurements (Beyrich et al.,2006). However, not all existing land-use types were covered with turbulence measurement stations in the case ofthe COPS network. Therefore, an areaaveraging offluxes based onmeasurements isimpossible for the COPS period. More important for the COPS set-up are flux differences between the Upper Rhine valley and the mountaintops and valleys of the Black Forest and flux differences between differentland-use types. Flux differences between different land-use types mainly occur due to altering surface characteristics such as albedo, emissivity, leaf area index (LAI), canopy heightand structure (e.g. Munn, 1966; Oke, 1987). Different stages of vegetation development and mowing ofgrassland sites also lead to temporal variations of surface characteristics at asingle location. The turbulent fluxes atindividual locations are also influenced bydifferences in altitude. (Wenzel et al., 1997) give theoretical considerations for the dependence of the Bowen ratio (Bo)onaltitude. They showed that an increase or decrease of Bo with altitude depends on theairtemperatureandtherelationofthetemperature and humidity gradient. Observations of both an increase or decrease of Bo with altitude were made (e.g. Kessler, 1985). Accordingly, the first objective of this study is the investigation offlux differences due to different locations and land-use characteristics and thepresentation of thedata processing and quality control ofthe flux data. The influence of the surface fluxes onthe ABL conditions and on the preconvective environment isinvestigated indetail inthe study of Kalthoffet al.,(2010). Using some stations of the COPS turbulence network, the authors focused onthe relationship between soil moisture, surface fluxes, ABL conditions and convective indices. It was found that the convective indices depend on ABL conditions, which inturn are influenced by the energy transformation at the Earth’s surface. However, due to aweak correlation between the surface fluxes and the ABL conditions, especially over the Black Forest, the authors Figure 1. Topographic map of the turbulence measuring sites ofthe COPS field campaign (see also Table I). concluded that advective processes also determine the ABL characteristics over the complex terrain of the COPS area. The second objective of the present study is the investigation ofthe non-closure ofthe surface energy balance. Aresiduum of 10–30% is often found in micrometeorological field experiments (e.g. Foken et al., 2010, Oncley et al.,2007) and atFLUXNET sites (Wilson et al.,2002). Anoverview ofthe energy balance problem is given inCulf et al. (2004) and Foken (2008b). The main reason for the residuum is assumed tobethe landscape heterogeneity causing advective and low-frequency flux components (Foken et al., 2006) and secondary circulations (Inagaki et al.,2006), which are not caughtbythestandard eddy-covariance measurements and which transport thesurplus ofenergy. The imbalance has important consequences for the use of surface turbulent flux data for ground truth and model validation. As the energy balance is, bydefinition, closed in most of the applied models, the residuum has tobeconsidered when comparing modelled and measured flux values (e.g. Kracheret al.,2009). Asa first guess itisacommon procedure todistribute the residuum according tothe Bowen ratio (Twine et al., 2000). Moreover, the present study extends the investigation of the occurrences of near-ground free convection conditions (FCCs) ofEigenmann et al. (2009) in the Kinzig valley of the Black Forest to all sites in the COPS region. FCCs occur during situations ofhigh buoyancy fluxes which coincide with very low wind speeds. Buoyancy-driven turbulence then dominates over shear-driven turbulence near the ground, which results in an effective vertical transport mechanism of heat and moisture, enhanced in near-ground regions, into upper parts of the ABL. Asthe dimensions of the target land-use types of most of the COPS sites are large enough (>250 m) that the surface fluxes Copyright c 2011 Royal Meteorological Society Q. J. R. Meteorol. Soc. 137:57–69 (2011) A PPENDIX C – E IGENMANN ET AL . (2011) 57 Table I.Turbulence measuring sites ofthe COPS field campaign.The column ‘Code’ abbreviates the responsible institute (UBT: University ofBayreuth; IMK: Karlsruhe Institute of Technology; MF: M´ et´ eo France; UV: University of Vienna; UBN: University of Bonn) followed by arunning number. Also given are the coordinates (latitude, longitude), the altitude (metres a.s.l.) and the EC set-up (sonic anemometer, hygrometer –CSAT3: sonic anemometer byCampbell Scientific Inc., USA; USA-1: sonic anemometer byMETEK GmbH, Germany; Solent R1012: sonic anemometer byGill Instruments Ltd., UK; Young 81000: sonic anemometer byR.M.Young Company, USA; Solent HS: sonic anemometer by Gill Instruments Ltd., UK; KH20: krypton hygrometer byCampbell Scientific Inc., USA; LI-7500: open-path CO2/H2Ogas analyzer by LI-COR Biosciences, USA). The station UV1 used the Optical Energy Balance Measurement System (OEBMS1) with a Scintillometer SLS20 system byScintec AG, Germany insteadofthe EC measuring technique. The column ‘Land use’ indicates the target land-use type: grassland (G), maize (M), strawberry (S), fallow (F) and wheat (W). The column ‘Location’ sorts the stations by their location: valley (V) sites, mountaintop (T) sites, Upper Rhine (R) valley sites and sites in the lee (L) ofthe Black Forest. Code Site Coordinates Alt. Land use Sonic Hygrometer Location (lat., long.) anemometer UBT1 Fußbach Ia48◦22′7.82′′,8 ◦1′21.17′′ 178 MCSAT3 LI-7500 V UBT2 Fußbach II 48◦22′0.88′′,8 ◦1′16.68′′ 180 FUSA-1 -V UBT3 Fischerbach a48◦16′57.40′′,8 ◦7′56.28′′ 226 GCSAT3 KH20 V UBT4 Hagenbuch a48◦16′54.59′′,8 ◦12′16.81′′ 245 GCSAT3 LI-7500 V IMK1 Hornisgrinde a48◦36′12.95′′,8 ◦12′4.88′′ 1158 GSolent R1012 LI-7500 T IMK2 Baden Airpark a48◦46′40.51′′,8 ◦4′25.20′′ 120 GSolent R1012 LI-7500 R IMK3 Linkenheim a49◦8′9.24′′,8 ◦23′32.21′′ 96 WYoung 81000 -R IMK4 Sasbach 48◦39′4.20′′,8 ◦5′19.53′′ 133 GSolent R1012 -R IMK5 Oberkirch 48◦31′19.15′′,8 ◦5′54.00′′ 203 S Solent R1012 -V IMK6 Bad Rotenfels 48◦49′29.35′′,8 ◦17′30.55′′ 127 GSolent R1012 -V IMK7 Igelsberg 48◦31′40.85′′,8 ◦25′50.35′′ 770 GSolent R1012 -T IMK8 Burnhaupt 47◦42′33.52′′,7 ◦9′16.07′′ 299 MSolent R1012 -R MF1 Niederrott a48◦26′34.40′′,7 ◦32′38.36′′ 155 MSolent HS LI-7500 R MF2 Nordfeld a48◦27′58.22′′,7 ◦32′22.35′′ 156 MSolent HS LI-7500 R UV1 Deckenpfronn a,b 48◦38′21.12′′,8 ◦49′7.86′′ 574 G- - L UBN1 Deckenpfronn a48◦38′21.12′′,8 ◦49′7.86′′ 574 GCSAT3 LI-7500 L aAdditional measurement of radiation and soil components. bNot included inthe data analyses of section 4for comparability reasons. areabletoinfluencethestructureoftheABL(Shenand Leclerc, 1995), the detected FCCs may have astrong impact on ABL characteristics and hence on the pre-convective environment intheCOPS region. Animpact ofFCCs on vertical wind speeds (Eigenmann et al.,2009) and onozone concentrations (Mayer et al.,2008) in the ABL was recently demonstrated. Two IOPs are selected in the framework of this study: IOP 8b (15 July 2007) and IOP 9c (20 July 2007). These IOPs are chosen because the processes ofconvection initiation (CI) are intensively investigated and well understood (e.g. Kottmeier et al.,2008; Kalthoffet al.,2009; Barthlott et al., 2010). The mechanisms leading toCIare very different on both days. According toKottmeier et al. (2008), convection during IOP 8b is locally initiated, which means that surface fluxes and valley winds play an important role for CI, while convection during IOP 9c occurs near prefrontal convergence zones. The article is organized asfollows. Section 2outlines the experimental set-up. Section 3describes the processing and quality control ofthe turbulence data as well as the determination of the energy balance closure and the detection ofFCCs. Section 4shows and discusses the effect of land use and location onflux measurements, the energy balanceclosureandtheoccurrenceofFCCs.Insection5a summary and conclusions are given. 2. Experimental set-up The surface energy balance and turbulence network consisted ofsixteen stations set upwithin the heterogeneous landscape ofthe COPS area (see Figure 1and Table I). The eddy-covariance (EC) measuring technique was applied in order toprovide high-quality and continuous surface turbulent flux data of momentum and sensible and latent heat. Additionally, soil (Hauck et al.,2011) and radiation measurements as well as standard surface meteorological data were recorded at most of the sites. Heterogeneity in the COPS region exists due to orography and due to apatchy land-use structure. Therefore, measuring sites included locations in the valleys (V) and on mountaintops (T)ofthe BlackForest, locations inthe Upper Rhine (R) valley and locations in the lee (L) ofthe Black Forest (see Table Iand Figure 1). Conversely, the turbulent fluxes were also measured over different land-use types (see Table I): grassland (G), maize (M), strawberry (S), fallow (F) and wheat (W). Mainly grassland sites (nine stations) and maize fields (four stations) were probed. Thetypical EC measuring set-up consisted ofasonic anemometer and afast-response hygrometer according toTable I. The measurement height of all stations was in therange of 1.8–10 m, and thesampling frequency ofthe EC raw data amounted to10, 20 or 32 Hz. The time stamp used was UTC (Universal Time Copyright c 2011 Royal Meteorological Society Q. J. R. Meteorol. Soc. 137:57–69 (2011) 58 A PPENDIX C – E IGENMANN ET AL . (2011) Table II. Flux contributions from the target land-use type (grassland), in%,dependent on wind direction sector and stability class for the station Hagenbuch (UBT4, see Table I).Average flux contributions over the entire COPS period are shown. Moreover, the internal boundary layer evaluation procedure for average conditions over the entire COPS period is presented for Hagenbuch. The height ofthe new equilibrium layer, δ0,iscalculated with the fetch, x,according to Eq. (1). Both variables are listed for different wind direction sectors. The flagging scheme of Table III is applied with an aerodynamic measurement height of za=2.0m. Wind sector: 30◦60◦90◦120◦150◦180◦210◦240◦270◦300◦330◦360◦ Flux contributions, in %, from the target land-use type (grassland): Stable 78 85 89 94 93 91 83 83 96 97 97 91 Neutral 9497959899989695999910096 Unstable 100 100 99 100 100 100 100 97 100 100 100 100 Internal boundary layer evaluation: x(m) 95 105 97 101 84 115 44 48 95 247 239 101 δ0(m) 2.9 3.1 3.0 3.0 2.8 3.2 2.0 2.1 2.9 4.7 4.6 3.0 Flag000000100000 Coordinated). All data were transformed into acommon data format and transferred tothe COPS data base operated bythe World DataCenter for Climate (WDCC)in Hamburg, Germany. 3. Turbulence data processing 3.1. TK2 processing The ECflux data of the COPS turbulence network were processed and quality-controlled with the software package TK2 developed by the Department ofMicrometeorology, University of Bayreuth (Mauder and Foken, 2004; Mauder et al.,2008). Following Mauder et al.(2006), the processing steps and flux corrections listed below are applied tothe EC raw data while running TK2. •Calculation ofaverages, variances and covariances for anaveraging interval of 30 min and taking into consideration the time delays between different sensors and excluding physically invalid values and spikes (Vickers and Mahrt, 1997). •Cross-wind correction ofthe sonic temperature if necessary (depending on sonic anemometer type). •Correction ofoxygen cross-sensitivity for krypton hygrometers (Tanner et al.,1993, van Dijk et al., 2003). •Planar fit coordinate rotation (Wilczak et al.,2001). •Correction ofspectral loss due to path-length averaging, spatial separation ofthesensors and the frequency dynamic effect of signals (Moore, 1986). •Conversion of buoyancy into sensible heat flux (Schotanus et al.,1983, Liu et al.,2001) •Correction for density fluctuations (WPL correction) to determine fluxes of the scalar quantities H2Oand CO2(Webbet al.,1980; Fuehrer and Friehe, 2002; Liebethal and Foken, 2003, 2004). The impact ofthese processing and flux-correction steps on flux estimates and energy balance closure isdiscussed in Mauder and Foken (2006). Quality tests implemented inTK2 consist ofastationarity test and atest on the fulfilment ofintegral turbulence characteristics (ITC) for each turbulent flux (Foken and Wichura, 1996; Foken et al.,2004). According toFoken et al. (2004), the final quality flag (1–9) isassigned to aspecific half-hourly turbulent flux value by combining the quality flags for stationarity and ITC. Classes 1–3 can beused for fundamental research, classes 4–6 for general use such as continuously running systems and classes 7–8 for arough orientation. Turbulent flux values marked with aquality flag ofclass 9should berejected. This information isalso provided in the COPS data base at the WDCC (see section 2) as well as the information on the footprint modelling results (section 3.2) and theinternal boundary evaluation (section 3.3), asdescribed below. 3.2. Footprint modelling In order toevaluate thespatial representativity ofthe measured turbulentflux data in the context of the underlying land-use distribution, afootprint model was applied toall EC flux sites of the COPS turbulence network. Following the approach of G¨ ockede et al. (2004, 2006), the flux data-quality flagging scheme of Foken et al. (2004) is combined with the Thomson (1987) three-dimensional Lagrangian stochastic trajectory model ofLangevin type (Wilson and Sawford, 1996). The parametrization followsthe flow statistics and the effect of stability onthe profiles used inRannik et al. (2003). The approach of G¨ ockede et al. (2004, 2006) allows the determination ofthe footprint climatology in relation to the land use (spatial representativity)and the spatial distribution of quality flags (spatial quality structure) as well as the calculation ofthe flux contribution ofthe target land-use type to the total flux measured for each half-hourly turbulent flux value. Recently, this flux data-quality evaluation approach was adapted to the sites of the CarboEurope network byRebmann et al. (2005) and G¨ ockede et al. (2008). As an example, TableII shows theflux contribution from thetarget land-use type at thestation Hagenbuch (UBT4, see Table I) for different wind direction sectors and Copyright c 2011 Royal Meteorological Society Q. J. R. Meteorol. Soc. 137:57–69 (2011) A PPENDIX C – E IGENMANN ET AL . (2011) 59 stability classes. The entire COPS measurement period is considered. Asexpected, flux contributions from the target surface increase from stabletowards unstable conditions. The main wind directions (120, 150, 270 and 300◦)withthe highest data density (not shown) show flux contributions over 93% even under stable conditions. The lowest flux contributions are found insectors with low data density, e.g. 78% in the 30◦sector during stable conditions. Overall, themeasured turbulent flux data atHagenbuch are highly representative of the target land-use type. 3.3. Internal boundary layers All measurements ateach ECstation of the COPS turbulence network were checked for the existence ofinternal boundary layers (IBL), which form asaresult of changes ingroundsurface characteristics adjacent to the tower. The exchange of energy and matter over the target land use isaffected by neighbouring areas when measuring above the IBL. The followingequation wasusedto estimate roughlythe heightof the new equilibrium layer, δ0,depending on the fetch, x,for each 30◦wind sector (Raabe, 1983; Jegede and Foken, 1999): za<δ 0=0.3√x.(1) The aerodynamic measurement height, za,isdefinedas the geometric height minus the zero-plane displacement. The latter can be determined with two-thirds of the canopy height (e.g. Foken, 2008a). In order toguarantee that the measurement takes place within the new equilibrium layer that establishes over the target land-use type, zashould be lower than δ0.Aboveδ1,the measurement cannot berelated to the target land-use type: za>δ 1=0.5√x.(2) Between δ0and δ1,atransition area is assumed. Aweak effect of stability onδ0,1 is neglected (Savelyev and Taylor, 2005). For the outcome ofthe IBL evaluation procedure, a flagging scheme was defined depending on the relation ofza to δ0and δ1(see Table III). This scheme was applied toeach half-hourly turbulent flux value. Flux values flagged with 1 should be considered with care and flux values flagged with 2should berejected inany case, asthe measured flux cannot beascribed to thetarget land-use type. As an example, Table II shows the results of the IBL evaluation procedure for average conditions over the entire COPS measurement period for the station Hagenbuch (UBT4, see Table I).The height of the new equilibrium layer, δ0(see Eq. (1)), and thefetch,x,arebothgiven for different wind direction sectors. Itcan be seen that in the 210◦sector the aerodynamic measurement height, za=2.0m,equalsthe height of the new equilibrium layer, δ0=2.0. This indicates that the flux measurement takes place at the lower edge of the transition area ofthe IBL and thus is flagged with 1(see Table III). Table III. Flagging scheme for the internal boundary layers. Condition Flag Description za<δ 00Measurement below δ0 δ0≤za≤δ11transition area za>δ 12Measurementaboveδ1 3.4. Data selection For the results shown in section 4,the following data selection criteria were applied; these result from the dataquality tests described insections 3.1, 3.2 and 3.3. Only turbulent flux data with •TK2 quality flags ≤6(see section 3.1), •flux contributions from the target land-use type >70% (see section 3.2) and •flags of the IBL evaluation procedure ≤1(seesection 3.3) are considered further for data analysis. Due tothe high landscape heterogeneity at the COPS measurement sites, we decided toallow turbulent flux data with flux contributions from the target land-use type >70% instead of>80% as was recommended byG¨ ockede et al. (2008) as areasonable value for representative measurements in his study above tall vegetation. Theresulting data availability after theapplication of these data selection criteria is shown inTable IV. Overall, most of thesites reach anadequate data availability ofmore than 50%. However, six stations also show very low data availability below 30%. This can be mainly attributed to data rejection due to low flux contributions from the target landuse type, either identified bythe applied footprint model or by the IBL evaluation procedure. Unfavourable fetch conditions were present at these sites, as they were closely situated at the boundaries of the target land-use types, which results inanincrease ofthe influence of neighbouring areas. At the station Linkenheim (IMK3) the large measurement height of 10 misresponsible for the failure ofthe IBL criteria and thusfor data rejection. 3.5. Energy balance closure Apossible residuum, res,ofthesurface energy balance can be evaluated at all measuring sites oftheturbulence network where additional radiation and soil measurements were carried out and both turbulent fluxes of sensible heat, QH,and latent heat, QE,were measured (see Table I). At the surface, the net radiation, Q∗ S,istransformed into QHand QEand into the ground heat flux, QG: −Q∗ S=QH+QE+QG+res.(3) Thestorage of heatin the upper soillayer is included inthis study within the value of QGand was calculated according to the‘simple measurement’ method after Liebethal and Foken (2007). Other storage terms (plants, air, etc.) and photosynthesis can beneglected as they are usually very small (Foken, 2008b). 3.6. Near-ground free convection conditions Near-ground free convection conditions can bedetected with the ECmeasurements bycalculating the stability parameter, ζ,which is the quotient ofthe measurement height, z,and the Obukhov length, L: ζ=z L=−zκgw′θ′ v0 θvu3 ∗ .(4) Copyright c 2011 Royal Meteorological Society Q. J. R. Meteorol. Soc. 137:57–69 (2011) 66 A PPENDIX C – E IGENMANN ET AL . (2011) Eigenmann R,Metzger S, Foken T.2009. Generation of free convection due tochanges of the local circulation system. Atmos. Chem. Phys. 9: 8587–8600. Foken T.2008a.. Micrometeorology.Springer: Heidelberg. Foken T.2008b.. The energy balance closure problem: Anoverview. Ecol. Appl. 18:1351–1367. Foken T,Wichura B. 1996. Tools for quality assessment ofsurface-based flux measurements. Agric. Forest Meteorol. 78:83–105. Foken T,G¨ ockede M, Mauder M, Mahrt L,Amiro BD, Munger JW. 2004. ‘Post-field data quality control’. InHandbook of Micrometeorology:A Guidefor Surface FluxMeasurement and Analysis,Lee X, Massman W, Law B(eds). Kluwer: Dordrecht, pp 181–208. Foken T,Wimmer F,Mauder M,Thomas C,Liebethal C. 2006. Some aspectsoftheenergy balanceclosure problem. Atmos. Chem. Phys. 6: 4395–4402. FokenT,MauderM,LiebethalC,WimmerF,BeyrichF,LepsJP,RaaschS, DeBruin HAR, Meijninger WML, Bange J. 2010. Energy balance closure for the LITFASS-2003 experiment. Theor. Appl. Climatol. 101: 149–160. Fuehrer PL, Friehe CA. 2002. Flux corrections revisited. Boundary-Layer Meteorol. 102:415–457. G¨ ockede M, Rebmann C,Foken T. 2004. Acombination of quality assessment tools for eddy covariancemeasurements with footprint modelling for the characterisation of complex sites. Agric. Forest Meteorol. 127:175–188. G¨ ockede M, Markkanen T,Hasager CB, Foken T. 2006. Update of afootprint-based approachfor the characterisation of complex measurement sites. Boundary-Layer Meteorol. 118:635–655. G¨ ockede M, FokenT,Aubinet M, Aurela M, Banza J,Bernhofer C, Bonnefond JM, Brunet Y,Carrara A, Clement R,Dellwik E, Elbers J, EugsterW,FuhrerJ,GranierA,Gr¨ unwald T, Heinesch B, Janssens IA, Knohl A,Koeble R,Laurila T,Longdoz B,Manca G, Marek M, Markkanen T,Mateus J,Matteucci G, Mauder M, Migliavacca M, Minerbi S,MoncrieffJ,Montagnani L, Moors E,Ourcival JM, Papale D,Pereira J,Pilegaard K,Pita G,Rambal S,Rebmann C, Rodrigues A,Rotenberg E, Sanz MJ, Sedlak P,Seufert G, Siebicke L, Soussana JF, Valentini R, Vesala T, Verbeeck H, Yakir D.2008. Quality control ofCarboEurope fluxdata –Part 1: Coupling footprint analyses with flux data quality assessment toevaluate sites in forest ecosystems. Biogeosciences 5:433–450. Hauck C,Barthlott C, Krauß L,KalthoffN,Sch¨ adler G,Kottmeier C. 2011. Soil moisture variability and its influence onconvective precipitation over complex terrain. Q. J. R. Meteorol. Soc. 137(S1): 42–56, DOI: 10.1002/qj.766. Inagaki A,Letzel MO, Raasch S, Kanda M.2006. Impact of surface heterogeneityonenergyimbalance:AstudyusingLES.J. Meteorol. Soc. Jpn 84:187–198. Jegede OO, Foken T.1999. Astudy of the internal boundary layer due to aroughness change in neutral conditions observed during the LINEX field campaigns. Theor. Appl. Climatol. 62:31–41. KalthoffN, Fiedler F,Kohler M,Kolle O, Mayer H,Wenzel A.1999. Analysis of energy balance components asafunction of orography and land use and comparison ofresults with thedistribution of variables influencing local climate. Theor. Appl. Climatol. 62:65–84. Kalthoff N, Adler B,Barthlott C, Corsmeier U,Mobbs S, Crewell S, Tr¨ aumner K, Kottmeier C, Wieser A, Smith V,DiGirolamo P.2009. Theimpact ofconvergence zones onthe initiation of deep convection: AcasestudyfromCOPS.Atmos. Res. 93:680–694. KalthoffN, Kohler M, BarthlottC,Adler B, Mobbs SD,Corsmeier U, Tr¨ aumner K, Foken T,Eigenmann R, Krauß L,Khodayar S,Girolamo PD. 2010. The dependence ofconvection-related parameters on surface and boundary layer conditions over complex terrain. Q. J. R. Meteorol. Soc.,DOI: 10.1002/qj.686. Kessler A. 1985. Heat balance climatology.Elsevier: Amsterdam. Kottmeier C, KalthoffN, Barthlott C,Corsmeier U, van Baelen J, Behrendt A, Behrendt R, Blyth A,Coulter R, Crewell S,DiGirolamo P, DorningerM,FlamantC,FokenT,HagenM,HauckC,H ¨ oller H, Konow H, Kunz M, MahlkeH, Mobbs S, Richard E,SteinackerR, Weckwerth T,Wieser A,Wulfmeyer V. 2008. Mechanisms initiating convection during the COPS experiment. Meteorol. Z.17:931–948. Kracher D,Mengelkamp HT, Foken T.2009. The residual ofthe energy balanceclosure and its influenceon theresults of three SVAT models. Meteorol. Z.18:647–661. Liebethal C,Foken T. 2003. On the significance of the Webb correction to fluxes. Boundary-Layer Meteorol. 109:99–106. Liebethal C,Foken T. 2004. On the significance of the Webb correction to fluxes. Corrigendum.Boundary-Layer Meteorol. 113:301. Liebethal C,Foken T. 2007. Evaluation of six parameterization approaches for the ground heat flux. Theor. Appl. Climatol. 88: 43–56. Liu HP, Peters G, Foken T.2001. New equations for sonic temperature variance and buoyancy heat flux with an omnidirectional sonic anemometer. Boundary-Layer Meteorol. 100:459–468. Mauder M, Foken T.2004. ‘Documentation and instruction manual of the eddycovariance software packageTK2’, Work Report University of Bayreuth, Dept. of Micrometeorology, 26, 42 pp. University of Bayreuth: Bayreuth; ISSN 1614-8916. Mauder M, Foken T.2006. Impact of post-field data processing on eddy covariance flux estimates andenergy balance closure. Meteorol. Z.15: 597–609. Mauder M, Liebethal C,G¨ ockede M, Leps JP, Beyrich F, Foken T.2006. Processing and quality control offlux data during LITFASS-2003. Boundary-Layer Meteorol. 121:67–88. Mauder M, Foken T,Clement R, Elbers JA, Eugster W, Gr¨ unwaldT, Heusinkveld B,Kolle O. 2008. Quality control ofCarboEurope fluxdata –Part 2: Inter-comparison of eddy-covariance software. Biogeosciences 5:451–462. Mayer JC, Staudt K, Gilge S,Meixner FX, Foken T. 2008. The impact of free convection on late morning ozone decreases onanAlpine foreland mountain summit. Atmos. Chem. Phys. 8:5941–5956. Mengelkamp HT, Beyrich F, Heinemann G,Ament F, Bange J,Berger F, B¨ osenberg J, FokenT,Hennemuth B, Heret C,HunekeS,Johnsen KP, Kerschgens M, Kohsiek W,Leps JP, Liebethal C,Lohse H, Mauder M, Meijninger W, Raasch S, Simmer C,Spiess T,Tittebrand A, Uhlenbrock J, Zittel R.2006. Evaporaton over aheterogeneous land surface–TheEVA-GRIPS project. Bull. Am. Meteorol. Soc. 87: 775–786. Moore CJ. 1986. Frequency response corrections for eddy correlation systems. Boundary-Layer Meteorol. 37:17–35. Munn RE. 1966. Descriptive micrometeorology.Academic Press: New York. Oke TR. 1987. Boundary layer climate.Methuen: London. Oncley SP, Foken T, Vogt R, Kohsiek W,DeBruin HAR, Bernhofer C, Christen A, van Gorsel E,Grantz D,Feigenwinter C,Lehner I, Liebethal C,Liu H, Mauder M, Pitacco A, Ribeiro L,Weidinger T. 2007. The Energy Balance Experiment EBEX-2000. Part I: Overview and energy balance. Boundary-Layer Meteorol. 123:1–28. Raabe A.1983. On the relation between the drag coefficient and fetch above the sea inthe case of off-shore wind inthe near-shore zone. Z. Meteorol. 33:363–367. Rannik ¨ U, Markkanen T,Raittila J, Hari P, Vesala T. 2003. Turbulence statisticsinside and over forest: Influence on footprint prediction. Boundary-Layer Meteorol. 109:163–189. Rebmann C,G¨ ockede M, Foken T,Aubinet M, Aurela M, Berbigier P, Bernhofer C,Buchmann N,Carrara A, Cescatti A, Ceulemans R, Clement R,Elbers JA, Granier A,Gr¨ unwaldT,GuyonD, Havr´ ankov´ aK,Heinesch B,Knohl A,Laurila T,Longdoz B, Marcolla B, Markkanen T,Miglietta F, MoncrieffJ,Montagnani L, Moors E,Nardino M, Ourcival JM, Rambal S,Rannik ¨ U, RotenbergE, Sedlak P, Unterhuber G, Vesala T, Yakir D.2005. Quality analysis applied oneddy covariance measurements at complex forest sites using footprint modelling. Theor. Appl. Climatol. 80:121–141. Savelyev SA, Taylor PA. 2005. Internal boundary layers: I.Height formulae for neutral and diabatic flows. Boundary-Layer Meteorol. 115:1–25. Schotanus P,Nieuwstadt FTM, DeBruin HAR. 1983. Temperature measurement with asonic anemometer and its application toheat and moisturefluxes. Boundary-Layer Meteorol. 26:81–93. Shen S, Leclerc MY. 1995. How large must surface inhomogeneities be before they influencetheconvective boundary layer structure? Acase study. Q. J. R. Meteorol. Soc. 121:1209–1228. Stull RB. 1988. An Introduction to Boundary Layer Meteorology,Atmospheric Sciences Library. Kluwer Academic Publishers: Dordrecht. Tanner BD, Swiatek E,Greene JP. 1993. ‘Density fluctuations and useofthekrypton hygrometer insurfacefluxmeasurements’. In Management of Irrigation and Drainage Systems: Integrated Perspectives,Allen RG(ed). American Society ofCivil Engineers: New York, USA, pp 945–952. Thomson DJ. 1987. Criteria for the selection ofstochastic models of particle trajectories in turbulent flows. J. Fluid Mech. 180:529–556. Twine TE, Kustas WP, Norman JM, Cook DR, Houser PR, Meyers TP, Prueger JH, Starks PJ, Wesely ML. 2000. Correcting eddy-covariance flux underestimates over agrassland. Agric. Forest Meteorol. 103: 279–300. Copyright c 2011 Royal Meteorological Society Q. J. R. Meteorol. Soc. 137:57–69 (2011) A PPENDIX C – E IGENMANN ET AL . (2011) 67 van Dijk A,Kohsiek W, DeBruin HAR. 2003. Oxygen sensitivity of krypton and Lyman-alpha hygrometers. J. Atmos. Ocean. Tech. 20: 143–151. Vickers D,Mahrt L. 1997. Quality control and flux sampling problems for tower and aircraft data. J. Atmos. Ocean. Tech. 14:512–526. Webb EK, Pearman GI, Leuning R. 1980. Correction offlux measurements for density effects due toheat and water-vapor transfer. Q. J. R. Meteorol. Soc. 106:85–100. Weckwerth TM, Wilson JW, Hagen M, Emerson TJ, Grebe L.2011. Radar climatology of theCOPS region. Q. J. R. Meteorol. Soc. 137(S1): 31–41. Wenzel A, Kalthoff N. 2000. Method for calculating the whole-area distribution of sensible and latent heat fluxes based onclimatological observations. Theor. Appl. Climatol. 66:139–160. Wenzel A, Kalthoff N, Fiedler F.1997. On the variation of the energy balance components with orography in the Upper Rhine Valley. Theor. Appl. Climatol. 57:1–9. Wilczak JM, Oncley SP, Stage SA. 2001. Sonic anemometer tilt correction algorithms. Boundary-Layer Meteorol. 99:127–150. Wilson JD, Sawford BL. 1996. Review of Langrangian stochastic models fortrajectories intheturbulent atmosphere.Boundary-LayerMeteorol. 78:191–210. WilsonK,GoldsteinA,FalgeE,AubinetM,BaldocchiD,BerbigierP, Bernhofer C,Ceulemans R, Dolman H, Field C,Grelle A, Ibrom A, Law BE, KowalskiA, Meyers T,MoncrieffJ,Monson R, Oechel W, Tenhunen J, Valentini R,Verma S. 2002. Energy balance closure at FLUXNET sites. Agric. Forest Meteorol. 113:223–243. WulfmeyerV,BehrendtA,BauerHS,KottmeierC,CorsmeierU,BlythA, CraigG,Schumann U,HagenM,Crewell S,DiGirolamoP,FlamantC, MillerM,MontaniA,MobbsS,RichardE,RotachMW,ArpagausM, Russchenberg H,Schl¨ ussel P,K¨ onig M,G¨ artner V, Steinacker R, Dorninger M,Turner DD, Weckwerth T,Hense A, Simmer C.2008. The convective and orographically-induced precipitation study: A Research and Development Project ofthe World Weather Research Program for improving quantitative precipitation forecasting inlowmountain regions. Bull. Am. Meteorol. Soc. 89:1477–1486. Wulfmeyer V,Behrendt A,Kottmeier C, Corsmeier U,Barthlott C, Craig G,Hagen M, Althausen D,Aoshima F, Bauer HS, Bennett L, Blyth A,Brandau C, Crewell S,Dick G,DiGirolamo P,Dorninger M, Dufournet Y,Ehret G, Engelmann R,Flamant C, FokenT,Gorgas T, GrzeschikM,HandwerkerJ,HauckC,H ¨ oller H,Junkermann W, Kalthoff N, Kiemle C, KlinkS,K¨ onig M,Krauss L,Long CN, Madonna F, Mobbs S, Neininger B,Pal S, Peters G, Pigeon G, RadlachM,Richard E, RotachMW, Russchenberg H,Schwitalla T, Smith V,SteinackerR,Trentmann J, Turner DD, van Baelen J,Vogt S, Volkert H,Weckwerth T, Wernli H, Wieser A, Wirth M.2011. The Convective and Orographically-Induced Precipitation Study(COPS): thescientificstrategy, thefield phase, and research highlights. Q. J. R. Meteorol. Soc. 137(S1): 3–30, DOI: 10.1002/qj.752. Copyright c 2011 Royal Meteorological Society Q. J. R. Meteorol. Soc. 137:57–69 (2011) 68 A PPENDIX D – Z HOU ET AL . (2011) Appendix D: Zhou et al. (2011) ORIGINAL PAPER Theor Appl Climatol (2011) 105:217–228 DOI 10.1007/s00704-010-0393-5 The study of near-ground free convection conditions at Nam Co station on the Tibetan Plateau Degang Zhou &Rafael Eigenmann &Wolfgang Babel& Thomas Foken&Yaoming Ma Received: 18 May 2010 /Accepted: 6December 2010 /Published online: 1January 2011 #Springer-Verlag 2010 Abstract This study investigates the near-ground free convection conditions(FCCs)based on eddy covariance (EC) measurements at Nam Co station near the Nam Co Lake on the Tibetan Plateau (TP). The spatial and temporal structure of EC measurements at this station is evaluated by using the comprehensive softwarepackage TK2 together with afootprint model. The obtained high-quality turbulent flux data are used to study the occurrence of FCCs, which can be detected with the EC system by calculating the stability parameter.Two types of generation of FCCs can be identified. (1) During the wind direction change ofadiurnal thermally forced land-lake circulation system inthe morning, strongly reduced wind speeds and simultaneously high buoyancy fluxeslead to aperiod ofdominance of buoyancy over shear,and hence, to the occurrence of FCCs. (2) On days with the appearance of clouds,the land-lake circulation is weakened or reversed, dependenton the temperature gradients between the land and the Nam Co Lake. During the period ofadaptation of the land-lake breeze to the alternating situation of heating differences, wind speeds decrease and buoyancy again dominates over shear near the ground. These are the situations where FCCs are also detected during the entire day at Nam Co station. The investigation of FCCs regarding the wholemeasurement period shows that FCCs can be mainly attributed to case (1) duringthe non-monsoon period, while FCCs are generatedbybothmechanisms(1and2)duringthe monsoon season.Animpact of the FCCs on the nearground profiles of air temperature and humidityisdemonstrated. The FCCs are assumed to play an importantrole for the land surface-atmosphere exchange processes and the atmospheric boundary layer (ABL) conditionson the TP by providing an effective transport mechanism of near-ground air mass characteristics into upper parts of the ABL. 1Introduction The Tibetan Plateau (TP) is the largest and highest plateau in the world. Because of the topographic feature of the TP, the underlying surface absorbsalarge amountofsolar radiation and undergoes dramatic diurnal and seasonal changes of surface heating andwater fluxes. Such energy and water cycles on the TP play an important role in the Eastern Asian Monsoon circulation (e.g., Yeand Gao 1979). Several large field experimentssuch as GAMETibet and CAMP-Tibet were carried out to investigate the land surface-atmosphere exchange processes on the TP, whose findings are reported in many studies (e.g., Yang et al. 2002,2004;Ma etal. 2002,2005;Zuo et al. 2005). The terrain on the TP consists of alot of different land use types such as lake, glacier,alpine steppe, bare soil, and wetland with different scales. As aconsequence, there are complex interactions between different types of land use and the atmosphere, which affect the development of thermally inducedcirculationsand waterand energy cycles (Whiteman D. Zhou (*) Centre for Monsoon System Research, Institute of Atmospheric Physics, Chinese Academy of Sciences, P.O. Box 2718, No. 6, Bei’ertiao, Haidian district, Beijing 100190, China e-mail: [email protected] R. Eigenmann :W.Babel :T.Foken Department of Micrometeorology,University of Bayreuth, Universitätsstrasse 30, 95440, Bayreuth, Germany Y. Ma Institute of Tibetan Plateau Research, Chinese Academy of Sciences, P.O. Box 2871, Beijing 100085, China A PPENDIX D – Z HOU ET AL . (2011) 69 1990;SegalandArritt 1992;Lietal.2006;Luetal.2008), andtheoccurrence oflargethermal convectivecells (Chen etal.2002). In general, land surface-atmosphere interactionsarehighly decisivefor the localoccurrence and timing ofthe initiation ofconvection, cloud formation, and precipitation (e.g., Hanesiak etal. 2004;Chenand Avissar1994;Rabin etal. 1990;Banta1990;Raymond and Wilkening 1980). Following the study of Eigenmann et al. (2009), the presentstudy aimsatthe investigation of near-ground free convection conditions(FCCs) in experimentaldata at Nam Co Station for Multisphere Observation and Research, ChineseAcademy of Sciences (hereinafterNam Co station).FCCs observed at ground level may result in an effective, vertical transport of near-ground air mass characteristics into upper parts of the atmospheric boundary layer (ABL). An impact of the FCCs on the vertical wind speeds in theABL measured with aSodar and on ozone concentrations at amountain summit was recently demonstrated by Eigenmann et al. (2009)and Mayer et al. (2008), respectively.Both authorsdetected FCCs with the help of an eddy covariance (EC) system by calculating the stability parameter from directly measured turbulent flux data. In the present study,the same approachwill be appliedtodetect FCCs at Nam Co station. As the EC data are used for the detection of FCCs, this study will also give ashort introduction in the EC data processing and EC data quality including footprint analysis at Nam Co station. A more detailed quality assessmentofECdata at Nam Co station can be looked up in Metzger etal. (2006). The preconditions for the occurrence of FCCs are periods of high buoyancyfluxeswhich coincide with astrong decreaseofthe wind speed. Buoyancy then dominates over shear within turbulence production near the ground. In the case of Eigenmann et al. (2009), the drop of the wind speed was initiated by achange of the diurnal valley circulation system in the Kinzig valley,Black Forest,southwestern Germany.The nighttime down-valley winds cease and upvalley winds slowly build up in the morning, leading to the low wind speeds duringthe period ofwind direction change in the Kinzig valley.Similar findings werereported by Mayer et al. (2008). These authors found that ozone drop events on amountain summit in the Alpine foreland of southern Germany can be attributed to the occurrence of FCCs, which effectively transport ozone-free air masses from the nearby valleytowards the measuring site at the mountain top. In their study,the low wind speed periodin the morning, which triggers the occurrence of FCCs, is caused by achange of the wind direction of amesoscale circulation system knownasAlpine Pumping (Lugauer and Winkler 2005). Similar mechanismsleading to FCCs are assumed to occur at Nam Co station on the TP.Anumerical simulation of the ABL flow characteristics in the region of the Nam Co station by Lu et al. (2008)recently indicated that local circulation systems near Nam Co station include a wind circulation between anearby mountain range and its forelandaswellasasmaller scale circulation induced by the inhomogeneity ofthermal forcing between the large Nam Co Lakeand the grasslands.This land-lake circulation system is expected to trigger FCCs at Nam Co station. In general,the FCCs are assumed to play an importantrole for the land surface-atmosphere exchange processes on the TP similar to the findings in other study areas as described above. The article is organized as follows. Section 2introduces the fieldmeasurements at Nam Co station. Section 3 outlines some typicalfeatures ofEC data quality by applying afootprint model, and Section4presents the findings of the investigation of FCCs at Nam Co station. Conclusions are given in Section 5. 2Field measurements at Nam Co station NamCostation (30.773 N,90.963 E,4,745 ma.s.l.)was established by theInstituteofTibetan PlateauResearch, ChineseAcademyof Sciences in August 2005.Thestationis locatedat1km distance southeastofNamCo Lake,whichis thesecond largest saline lake on theTP(Guanet al.1984). Moreover,thestationis situated 15 kmnorth-northwestof a mountain range(NyainqentanglhaMountain)oriented from west-southwest to east-northeast with an altitude of up to 5,700 ma.s.l. Figure 1showstheposition of theNamCo stationand thedistribution oftheland use. Asmall inner lakeorientedfromwest-southwest to east-northeast is located 300 mnorthwestern,anda52-m meteorological tower40 mnortheastern of theECflux site.Furthermore,a solarpanel,thestationmain building,andanotherbuilding arelocated 70 m, 90 m, and 105 mabout10°westby south of theflux site,respectively.Theterrainis slightlyinclined with the altitudedropping by about1mon a250-mdistance along thedirectionfromtheflux site to theinnerlake.The dominating land use class is alpine steppe,referredtoas grass(−)inFig.1,witha canopy height notexceeding 5 cm. Asmall portion ofgrass is higher and denser,whichis named grass(+).Tosumup,the canopy atNamCo station canbeconsideredasratherhomogenous. The 52-m meteorological tower was set up to observe the lower boundarylayer with wind speed, air temperature, and relative humidity measurements in five levels (1 m, 2m,4m, 10 m, and 20 m), and wind direction measurements in three levels (1 m, 2m,and 20 m). Additionally, pressure andprecipitation were measured as wellasthe soil temperature and the soil moisture content insix levels (0 cm, 10 cm, 20 cm, 40 cm, 80 cm, and 160 cm). The averaging interval of all these measurements was 10 min. 218 D. Zhou et al. 70 A PPENDIX D – Z HOU ET AL . (2011) The EC flux site was equipped with aCSAT3 (Campbell Scientific, Inc.) sonic anemometer and aLI-7500 (LI-COR Biosciences) open-path H 2 O/CO 2 gas analyzer at aheight of 3m.The samplingfrequency was 10 Hz. Additionally, upward and downward components of long wave (Kipp & Zonen CG3) and shortwave radiation (Kipp &Zonen CM3) were measured with an averaging interval of 30 min.All measurements are available for about 7months from 21 March 2007 to 21 October 2007. However,data failure occurred from 20 April to 24 May due to instrument malfunction. 3Data processing and quality control of turbulence measurements The turbulent flux measurements at Nam Co station were carried out with the EC method. Thismethod is one of the best measurement techniques currently available for the determination ofmaterial and energy fluxes between the atmosphereand the underlying surface (Moncrieff 2004), although some general problems are still present, e.g., the widespread shortfall of the sum of sensible and latent heat flux measured by this methodascompared to the available energy (e.g., Aubinet et al. 2000;Foken 2008a). The EC methodhas becomeby now awidely accepted tool for the determination ofmass and energy fluxes applied by several flux networks such as Ameriflux, CARBOEUROFLUX and AsiaFlux, overall coordinated within FLUXNET (Baldocchi etal. 2001).Moreover,the focus of flux observationshas shiftedprogressivelyfrom idealand homogeneous sites to sites within complex and heterogeneous terrain (e.g., Schmid 2002). In general, the adoption of the EC method is based on the assumption that certain statisticaland meteorological requirements arefulfilled(e.g., Massmanand Lee2002;Foken andWichura 1996)and that the equipmentis workingreliably.Withthedevelopmentof thequality controland site characterization procedure includingfootprintinvestigations (e.g., Fokenetal.2004; Mauder et al.2006;Göckedeet al.2004,2006,2008), sitespecific spatialqualitystructures andthespatialrepresentativeness of themeasuredfluxescanbe identified,giventhat the underlying land use distribution is available. This approach improves ourunderstanding of theflux measurements and survey in complexterrainandisalsoappliedto theECflux data atNamCo stationwithinthepresentstudy. The turbulent flux data were post-processed with the comprehensive softwarepackage TK2 developed at the Department of Micrometeorology,University of Bayreuth (Mauderand Foken2004;Mauderetal.2008). The softwarecomprises all stateofthe art flux corrections and post-fieldquality control including tests for steadystateand integral turbulencecharacteristics(Foken and Wichura 1996;Foken et al. 2004). Planar fit rotation according to Wilczak et al.(2001)was done for six periods, P1 to P6, due to changes in vegetation structure and large-scale meteorological conditions. The start timeand end timeof these periods, and the momentum roughness length ofgrass (−)derived from Monin–Obukhov similarity theoryduring these periodsare reported in Table 1.Here, P1, P2, and P6 are periods during the non-monsoon period, P4 and P5 periods during the inner-monsoon period, and P3, aperiod duringthe transition of both. Figure 2shows the diurnal and seasonal variations of the sensible heat flux, the latent heat flux, the friction velocity, u * ,aswell as the wind direction for the entire measurement periodatNam Co station. The results indicate that the sensible heat flux dominates the energy transfer from the Earth’s surface into the atmosphere before monsoon. However,with the start of the monsoon period in July,the values of the sensible heat flux decreaseand latent heat dominates the turbulent energy exchange at the surface until theretreatofmonsoon in October.Thefriction velocity is usually smallduring the night and increases in 31.0N 30.3N 90.0E Nam Co Lake Nam Co Station for Multisphere Observation and Research x (km) y (km) -2 -1 012 -2 -1 0 1 2 91.5E Nam Co Lake Nam Co Station for Multisphere Observation and Research Nam Co Lake Nam Co Station for Multisphere Observation and Research x (km) -2 -1 012 -1 0 1 2 g rass(-) g rass(+)water settlement g rass(-) g rass(+)water settlement Fig. 1Position and land use map of Nam Co station. The land uses at Nam Co station include short grass, referred to as grass(−), denser grass, referred to as grass(+), water,and settlement The study of near-ground free convection conditions at Nam Co 219 A PPENDIX D – Z HOU ET AL . (2011) 71 the daytime. Adiurnal circulation systempresentatNam Co station becomes visible regarding the wind direction in Fig. 2d.Winds coming from northwest to southwestprevail in the daytime, while easterly to southerly winds dominate at night. During non-monsoon, the diurnal circulation is superimposed by asouthwesterly flow,while during monsoon, an increased southeasterly component is obvious. Footprint analyses according to Göckede et al. (2004, 2006)incombination with the data quality flagging scheme of Foken et al. (2004)wereperformed with aforward Lagrangian footprint model (Rannik et al. 2003;Göckede et al. 2005). Detailed terrain data wereused for the land use distribution. Metzger et al. (2006)alreadyinvestigated the temporal and spatial quality distribution of turbulent flux data at Nam Co station. Their results indicated that the general poor data quality of the sensible heat flux might be attributed to the occurrence of organized structuresor mesoscale flow patterns in the boundarylayer at the site. In this study,only ashort overview of the results of the quality control effort applied on the turbulent flux data shall be given in the following. Footprint calculations were performed in all periods, P1 to P6. The momentumroughness lengths of the lake and of the settlement were set constant to 0.002 mand 0.5 m, respectively,and the momentum roughness length of grass (+) was set to double the value of grass(−)ineach period as listed in Table1.The resolution of the land use map used for the footprint calculations amounts to 20 m(see Fig. 1). In the following, only the footprint analysis results of the periodP5are presented. The results in the other periods are similar and not discussed withinthis study.Moreover,the quality flagging scheme according to Rebmann et al. (2005) is used for all following figures, with flags ranging from 1 (good data quality) to 5(poor data quality). Figure 3shows the spatial distribution of the quality flags of the friction velocity duringthe period P5 as related to the footprint climatology.Itcan be seen that duringunstable and neutral stratifications, quality ratings within the 5% effect level ring mostly consist of classes 1and 2, which means that the data are suitable for fundamental research. However,during stable stratification, quality flags sometimes degrade to class 3 or4.Inthe case of the footprint analyses of the sensible heat flux (not shown), most of the quality flags are rated as 1 or2during unstable stratification, while during neutral and stable stratifications, quality flags degrade even to 5. The reason for the poor quality ratings(flag 5) of the friction velocity during the day and duringthe night is obviously different.Eighty-one percent of the bad values rated as 5among the nighttime data can be related to bad integral turbulence characteristics. This means that there is abig deviation of the friction velocity from statistical Table 1The start time and end time of the planar fit periods P1 to P6 distinguished for the processing of turbulent flux data at Nam Co station in 2007. The momentum roughness length, z 0m ,ofgrass(−)during each period is also given Period P1 P2 P3 P4 P5 P6 Start time 22 Mar 25 May 21 Jun 29 Jul 20 Aug 18 Sept End time 20 Apr 19 Jun 26 Jul 17 Aug 17 Sept 21 Oct z 0m (mm) 3.9 3.8 5.3 7.3 6.0 4.0 0 6 12 18 0 0 6 12 18 Local time (a) Apr Jun Aug Oct 0 n.V. −250 −50 0 50 200 600 (b) Apr Jun Aug Oct 0 6 12 18 0 n.V. −300 −50 0 50 200 600 Local time (c) Apr Jun Au g Oct n.V. 0 0.05 0.1 0.2 0.4 2 (d) Apr Jun Au g Oct 0 6 12 18 0 n.V. N E S W Fig. 2Thediurnalvariations of athesensibleheatflux(Wm −2 ), bthelatentheatflux(Wm −2 ), c thefriction velocity,u * (m s −1 ), anddthewind direction over the entiremeasurement period at NamCostation.Thevaluesof thevariablesarecolor-coded (valuesofthevariablesaregraycodedfortheprintedversion) 220 D. Zhou et al. 72 A PPENDIX D – Z HOU ET AL . (2011) similarity characteristics under strong stablestratification, which might be related to the occurrence of intermittency or gravity waves. On the contrary,68% of the bad values rated as 5among the daytime data can be attributedto stationarity problems. Regarding the spatial distribution ofquality flags, the bad-rated measurements in the western and northeastern upwind sectorcan be explained by the influence of the station buildingand the meteorological toweron the measurements. As the effect of aslightly inclined terrain is eliminated by the planar fit rotation and the bad quality ratings indeed do not concentrate on acertain sector,the poor quality flags within other sectors than the western and northeastern cannotbeattributed to the local topography. Consequently,this implies that instationarity is the reason for the bad quality ratings in the other sectors in the daytime, which is caused by changing mesoscale or local flow patterns. Moreover,the surface energy imbalance was investigated at Nam Co station. For example, the average non-closure duringthe periodP5reaches 31%. This value is similar to those found in previous studieson the TP (Tanaka et al. 2001;Yang et al. 2004). The soil heat flux in the present study is calculated by using six layers of soil temperature and soil moisture content and applying the thermal diffusion equation and correction method according to Yang and Wang (2008). In the present literature of the energy balanceclosure discussion (e.g., Foken et al. 2010; Foken 2008a;Culf et al. 2004), the landscape heterogeneity,which induces advective and low-frequencyflux components (e.g., Fokenetal. 2006), and secondary circulations (e.g., Inagaki et al. 2006), which may transport the surplus of energy not caught by the standard EC measurements, are considered to mainly cause the residuum of the surface energy balance. In summary,the terrain features at Nam Co station are predestinated to induce thermallydriven circulations, thus, influencing the nonclosure of the surface energy balance. 4Near-ground free convection conditions 4.1 The detection of FCCs FCCs occurintheatmosphere if the buoyancyterm dominatesoverthe sheartermwithinthe turbulence kinetic energy equation (e.g., Stull 1988). Thebuoyancy term (B), B¼g qv w0q0 v  0;ð1Þ consists oftheproductof thebuoyancy parameter, gqv 1,and thebuoyancyfluxat thesurface,w0q0 v.The buoyancy parameter isthe quotientofthe acceleration due to gravity,g,and the mean virtualpotential temperature, qv.The shearterm(S), S¼ u0w0@u @z;ð2Þ is theproduct of the momentumflux, u0w0,and the wind shear,@[email protected] quotient of BandScan also be expressedasthe flux Richardson numberR f (e.g., Stull y (m) −400 −200 0 200 400 20 5 Overall 20 5 Unstable x (m) y (m) −400 −200 0200 400 −400 −200 0 200 400 20 5 Neutral x (m) −400 −200 0200 400 20 5 Stable n.V. 1 2 3 4 5 Fig. 3Spatial distribution of quality flags of the friction velocity as related to footprint climatology depending on atmospheric stratification at Nam Co station. The flux site is indicated by the central, black (white in the printed version) crosshair,whereas quality flags (1–5) after Rebmann et al. (2005)are distinguished according to the color bar.Black (white in the printed version) contour lines display the 5% (dotted) and 20% (solid)effect level rings of the measurements The study of near-ground free convection conditions at Nam Co 221 A PPENDIX D – Z HOU ET AL . (2011) 73 1988). During unstable stratification (R f <0),R f equals the stability parameter (e.g., Ayra 2001), z¼z L¼ zkgw0q0 v  0 qvu3 » ;ð3Þ where zdenotes the measurement height, LtheObukhov length, κthevon-Kármán’sconstant,andu * thefriction velocity.Consequently,by calculating the stability parameter,ζ,fromdirectly measured ECturbulentflux data, free convection conditionscan bedetectedforζ<−1(e.g., Foken 2008b). Regarding Eq. 3,the fulfillment ofζ<−1 for theoccurrence of FCCs requires that high buoyancy fluxes coincide with small valuesofu * at the same time. In the following sections, the averaginginterval of the EC flux data is reduced from 30 min to 5min in order to get abetter insight into the temporal structure of the FCCs. Therefore, it should bekept in mindthat the depicted turbulent fluxes do not include spectral energy from eddies of larger scales. Moreover,the investigation of FCCs is restricted to days with no precipitationduringthe day and to timeperiods when the sensible heat flux exceeds the thresholdof20Wm −2 .This threshold seems to be reasonable for the criteria ofhigh buoyancy fluxes required for the occurrence of FCCs according to Eq. 3.Low values of u * occur at Nam Co station duringthe wind direction change of aland-lake wind circulation system between the Nam Co Lakeand the grassland, which dominates the local flow structure at the measuring site (see e.g. Lu et al. 2008). Hence, the change of the thermally inducedcirculation is able to causetheoccurrence of FCCsat NamCostation.Two typesofprocessesleadingtothelow wind speedperiodsin combinationwithawind directionchangeofthecirculation system,and consequently to theoccurrence ofFCCs,can be distinguishedand arepresentedin thenextsection. 4.2 The types of generation ofFCCs 4.2.1 Generation of FCCs due to achange of the diurnal circulation system The firsttype of generation of FCCs is similar to the case detected by Eigenmann et al. (2009)inthe Kinzig valley, Black Forest, southwestern Germany,whereFCCs appear in the morning hours during the change of the diurnal valley circulation system from down to up-valley winds. Typical for this type of generation of FCCs is that the drop of the horizontal wind speed, which is together with high buoyancy fluxes aprecondition for the occurrence of FCCs (see Eq. 3), is caused by the wind direction change of a diurnal circulation system in the morning after sunrise. This diurnal circulation system is forced by solar heating on fair weather days. The low wind speeds occur duringthe wind direction change when the previously prevailed wind ceases and winds from another direction start to build up according to the thermal forcing. At Nam Co station, adiurnal circulation system can be observed on many fair weather days as shown in Fig. 4. Regarding the wind statistics for the entire measurement period, it can be confirmed that athermallyinduced, landlake wind circulation between Nam Co Lakeand the grassland near the station is dominating the local wind regime at Nam Co station (see also Fig. 1). During the daytime (Fig. 4a), the frequent occurrence of alake breeze is obviousfrom the high frequency of winds coming from northwestern directions (7.3%). Higher frequencies can also be found in the west-southwestern sector.On the contrary, duringthe night (Fig. 4b), aland breeze with southsoutheastern wind directions dominates the flow regime. A diurnal land-lake wind circulation at Nam Co site was also recently confirmed by the numerical simulations of Lu et al. (2008). Figure 5gives an example for the occurrence ofFCCs on 29 June 2007. Each 5-minvalue, wherethe stability parameter,ζ,islower than −1, is marked by gray dotted lines in Fig. 5a.It isobviousthat FCCs are presentduring the time from 0625 to 0740 hours local time (in this study, local time corresponds to localsolartime=Beijing time −2h) in themorning.At thistime,aftersunrise,thesensible and latent heat flux (Fig.5b)arealreadylarge enough,butthe friction velocity,u * (0.11ms −1 ), andthehorizontal wind speedarestill verysmall (Fig.5c), so that FCCs are triggeredaccording to Eq.3.Thesmallvaluesofu * andthe horizontal wind speedcanbe relatedtothechange of the wind direction(Fig.5e) fromsoutheastto northwestduring theonsetof thelakebreeze.Thedownwardsolarradiationin Fig. 5d indicatesfairweatherconditions on this day,which forcethediurnalcirculation system.Finally,thedata quality flagsofu * ,ranging from 1(good data quality) to 5(lowdata quality) accordingtoRebmann et al.(2005), aredepictedin Fig. 5f forhalf-hourintervals. Note that thequalityflagsof u * during thedaytimearegood (flag1).However,during the FCCs from 0630 to 0730 hours, theflagsdegradeto 3 dueto stationarity problems. 4.2.2 Generation of FCCs due to the adaption of the land-lake wind circulation system to surface heatingdifferences duringcloud cover The second type ofgeneration of FCCs is closelyrelatedto the appearanceofcloudsduringtheentire daytime.Clouds frequently occur duringthemonsoon period atNamCosite and decreasethedirectsolar radiation which reachesthe surface.Asa consequence,thelakebreeze betweenthe Nam Co Lakeandthe grassland atNam Co station,which normally builds up in the daytime on undisturbedradiation days(see 222 D. Zhou et al. 74 A PPENDIX D – Z HOU ET AL . (2011) Fig. 5d,e), is weakened or even areversalcan beobserved. The cloud shading leadstoastrong and fast drop ofthe surface temperature overthe grassland,while thelake surface temperatureonlyexperiencesasmall decreaseduetothehigh heatcapacity ofwater.During theperiod ofweakening or reversalofthe circulation system,horizontalwind speeds and sheararereduced, and buoyancy dominatestheproduction of turbulence,thusleading totheoccurrence ofFCCs.If the reversal of the circulationsystem(land breeze)occursduring cloud shading,FCCscan also betriggeredat thetimethe land breeze shiftsback againtoalake breeze shortlyafterthe disappearanceofclouds.At thesetimes,temperature gradients and wind speeds are loweredagain, buoyancy dominatesovershearand FCCs occur. Figure 6depictsatypicalcaseon 23 August2007 with severalperiodsof FCCs marked by thegray dottedlines in Fig. 6a.The latent heat flux generallyexceeds thesensible heat flux on that day(Fig.6b). Theperiodsof FCCscoincide with periodsofsmall values of thefriction velocity(Fig.6c). Thedailycourse of thedownwardsolarradiation(Fig.6d) showstheappearance of clouds in thedaytimewith a minimumof174 Wm −2 at 1200 hours. At 1200 hours, a reversal of the circulation system is obviousfromFig.6e. From 1100 to 1140 hours, thesurface skin temperature and the airtemperature at 2mheight at NamCostation decreasesby 3.5Kand 0.8K,respectively,whiletheair temperature abovethelake canbe assumedtoshowonly some smalland slow changes. As a consequence,thelake breeze circulationisweakened.WindscomefromtheNam Co Lake (western directions)until 1205 hours, shift toaland breeze (eastern,southeastern directions)until 1425 hours, androtatebacktowind directionscoming fromthelake (north,northwestern directions)afterwards.Thesurface skin temperature and the airtemperature againincrease by 11.1K 9% 6% 3% WEST EAST SOUTH NORTH 0 − 2 2 − 4 4 − 6 6 − 16 N=4372 m s−1 (a) 12% 8% 4% WEST EAST SOUTH NORTH 0 − 2 2 − 4 4 − 6 6 − 14 N=3836 m s−1 (b) Fig. 4Wind statistics at Nam Co station during the entire measurement period aduring the daytime and bduring the night -2 -1 0 1 (-) (a) 0 200 400 Heat flux (W m-2) (b) 0 0.2 0.4 0.6 0.8 u* (m s-1) (c) 0 2 4 6 8 U (m s-1) 0 500 1000 Dsr (W m-2) (d) 00:00 06:00 12:00 18:00 00:00 0 90 180 270 360 Wind direction ( ) Local time (e) 00:00 06:00 12:00 18:00 00:00 1 3 5 Quality flag u* (-) Local time (f) H E ζ ° λ Fig. 5Daily courses of athe stability parameter,bthe sensible and latent heat flux (W m −2 ), cthe friction velocity and the horizontal wind (m s −1 ), dthe downward solar radiation (W m −2 ), ethe wind direction, and fthe quality flags of the friction velocity on 29 June 2007. The periods of FCCs are indicated by the gray dotted lines in each plot The study of near-ground free convection conditions at Nam Co 223 A PPENDIX D – Z HOU ET AL . (2011) 75 and by 3.8Kfrom1300 to 1440 hours,respectively.During thereversalofthelake breeze to windscoming fromtheland (1205 hours)and during therepeated onsetof thelakebreeze (1400–1500 hours),FCCsaredetected(see Fig.6a). These FCCs aretriggered by theperiodsof low wind speedsand u * (Fig.6c), whichcome along withthe cessation andtheonset of thelakebreeze.Furthermore, thetwoperiodsofFCCsat 0750 and 1015 hourscanalso berelatedto cloud cover periods(see Fig.6d)and toaweakening oftheprevailing circulationsystem.Thequalityflagsofu * areshownin Fig. 6f.During theFCCsat 0750 hoursand noon,data qualityispoor (flag 3 and 5, respectively). At 0750 hours, the poor data quality can beattributed to stationarity problems,whileat noon,stationarity problemstogetherwith badintegral turbulence characteristics arefound.Duringthe otherperiodsof FCCs on this day,data qualityisrathergood (mainlyflags1and 2). The second type of generation of FCCs is quitedifferent to the first type, as the appearance of clouds determines the occurrence of FCCs duringthe day.Onsome days, FCCs can be attributed to both types of generation.The distribution of FCCs during the entire measurement period is addressed in Section 4.4. 4.3 Impact of FCCs on near-ground boundarylayer conditions The impactofthe FCCs on the boundarylayer conditions near thesurface isinvestigated with the help ofthe measurements of the meteorological tower.The results are exemplarily shownfor 29 June 2007, already described in Section 4.2.Figure 7showsthe horizontal wind speed, the air temperature, and the specific humidity from 0530 to 0830 hours in five levelsofthe meteorological tower as well as the averaged profiles of these quantities before, during, and after the period ofFCCs. During the period of FCCs, astrong collapse of the horizontal wind speed (Fig. 7a)can be observed, and also, the vertical gradient of the horizontal wind speed decreases compared to the time periods before and after FCCs (Fig. 7b). The surface skin temperature showsanincreasing trend (Fig. 7c). It is lower than the air temperature until 0650 hours, but exceeds it from 0700 hours onwards. From 0650 to 0700 hours, the air temperature is consistent with the surface skin temperature. Togetherwiththe vertical gradient of the air temperature which is nearlyzero duringFCCs (Fig. 7d),thisfinding points to astrong vertical mixing ofthe airtemperature duringthistime. The specific humidity shows an increasing trend during the FCCs (Fig.7e). Again, the vertical gradient of specifichumidity (Fig.7f)issmall during the FCCs, which implies aneffective transport and mixing of specific humidity into upper partsof the ABL. In summary,it is obviousthatFCCshave animpacton near-ground ABLmoisture andtemperatureprofiles.Heated airmassesareanimated to ascend in convective pulse-like motions, whichresult in vertically uniformvalues of the air temperature, humidity,and thewind speed.Consequently, near-ground airmass characteristicsareeffectivelytrans- -2 -1 0 1 (-) (a) 0 100 200 300 Heat flux (W m-2) (b) 0 0.5 1 u* (m s-1) (c) 0 5 10 U (m s-1) 0 500 1000 Dsr (W m-2) (d) Cloud 00:00 06:00 12:00 18:00 00:00 0 90 180 270 360 Wind direction ( ) Local time (e) 00:00 06:00 12:00 18:00 00:00 1 3 5 Quality flag u* (-) Local time (f) H E ζ ° λ Fig. 6Daily courses of athe stability parameter,bthe sensible and latent heat flux (W m −2 ), cthe friction velocity and the horizontal wind (m s −1 ), dthe downward solar radiation (W m −2 ), ethe wind direction, and fthe quality flags of the friction velocity on 23 August 2007. The periods of FCCs are indicated by the gray dotted lines in each plot 224 D. Zhou et al. 82 A PPENDIX E – B RÖTZ ET AL . (2013) Early-morning flowtransition inavalleyinlow-mountainterrain 3 campaign isto understand theinfluence oftheorography ofalow-mountainrangeon precipitation.Severalsurface flux measurementstationswereinstalledthroughout therespective region during the campaign (Eigenmann etal., 2011;Kalthoff etal., 2011).Inaddition, ground-basedSodar/RASS instrumentswereinstalled on someofthesestations.Ourstudy usesdatafromoneofthesesiteswhere a full energy balance station and aSodar/RASS operatedsimultaneously.ThesiteFußbachislocatedintheKinzig valley,whichisatypical low-mountain valley oftheregion withapronounced valleywind systemdeveloping on fair weatherdaysinsummer. Inmicrometeorologicalfieldexperimentsthe energy balance isoften notclosed(e.g. Oncleyetal., 2007;Fokenetal., 2010).Although many uncertaintiesare connectedwith thedetermination ofthe componentsofthe energy balance (Mahrt,2010),strong indications exist that theresidualoccursduetotransportby large-scale eddiesorsecondarycirculations whicharenotcaptured by the eddy-covariance method (e.g.Mauderand Foken,2006;Foken,2008;Fokenetal., 2010,2011;Stoy etal., 2013).Asthesesecondarycirculationsare mainlyassociatedwiththebuoyancyflux,thepartitioning oftheresidualaccording tothe buoyancyflux ratioapproach,proposed by Charuchittipanetal.(2013),appearsto be appropriatetoclosethe energy balance.Thebuoyancyflux ratioapproach partitionstheresidual according tothebuoyancyflux ratioinstead oftheusually usedBowenratio(Twine etal., 2000).Thus,alarger fraction oftheresidualwould be attributedtothesensibleheatflux withthebuoyancyflux ratioapproach.AlsotheCOPS energy balance siteFußbach ofthis study showsanaverageresidualof21%during the entirefieldcampaign (see Eigenmann etal., 2011). Theremainderofthisarticleisorganizedasfollows.Section2 describesthedata and themethodsapplied,in particularthenumericalmodeland its set-ups.Section3 presents theresultsand thediscussion ofthem.The conclusionsaregiveninSect.4. 2Methodsand data 2.1Observationaldata Thedatausedinthis study wereobtained by observationsconducted during theCOPS experiment inthelow-mountainterrain oftheKinzig valley.Turbulence dataweremeasuredat aheightof2mabovethevalleysurface and friction velocityu∗,sensibleheatQHand latent heatQEwere calculatedwiththe eddy-covariance method (EC) (Fokenetal., 2012).An averaging timeof30minwasusedfortheEC.Contributionstothefluxeswithatimescale exceeding these30mincannotbe captured.Thegeographical location ofthesiteFußbach was48◦22′7.8′′ N,8◦1′21.2′′ E,178ma.s.l.(theposition ismarkedinFig.3).Thelocal timeisUTC+1 hour.Detailsabout thesite and themeasurementset-up can befound in Metzgeretal.(2007)and Eigenmann etal.(2009,2011).Closetotheturbulence station,a Sodar/RASS system measured verticalprofilesofwind componentsand virtual temperature. Moreover,theremaining componentsofthe energy balance,netradiation Q∗ S,and soil heat flux QG,weremeasuredat theECsite.An overviewofthedataprocessing,qualitycontrol, and flux characteristicsoftheturbulentdata aswell asthe calculation ofthe energy balance isgiveninEigenmann etal.(2011). Thestudy ofEigenmann etal.(2009)identified 23 daysduring thethree-monthCOPS campaign withfree convective conditionsbased on theECmeasurementsinthe earlymorning hours.Free convective conditionswereidentified by thestability parameter ζ = zL−1for ζ <−1,whereListheObukhov length.Theseperiodswere characterized by low A PPENDIX E – B RÖTZ ET AL . (2013) 83 4Bj¨ ornBr¨ otz etal. horizontalwind duetothereversalofthevalleywind systemfromdown-valleywindsto up-valleywinds.Themeantimeoftheoccurrence ofthefree convectivesituationson these 23 daysis0815 UTCwithastandard deviation of1 hour. Theverticalwind speedwderivedfromtheSodarobservationswasanalysedforthe identifiedlow wind periods.Intheremainderofthe articlethisperiod oflow wind will bereferredtoasp1and thesubsequentperiod ofup-valleywind will bereferredtoasp2. In ordertomaketheindividualdayscomparable,eachSodarsampleofwwasnormalized withtheDeardorff convectivevelocityw∗(Deardorff,1970).Themeasurementheightz wasnormalizedwiththeheightoftheboundarylayerzi.Surface buoyancyfluxesforthe calculation ofw∗werederivedfromtheECmeasurementsand valuesofziweredetermined by asecondarymaximuminthereflectivity profilesoftheSodarmeasurementsasdescribed inEigenmann etal.(2009)and suggested by Beyrich(1997).Afterthat,histogramsofww−1 ∗ were calculatedforthree characteristicheightsofzz−1 i=0.25,0.50 and 0.75. 2.2Simulations 2.2.1NumericalModel Thenumericalsimulationswere conducted by meansofthemultiscalegeophysicalflow solverEULAG (Smolarkiewicz etal., 1997;Prusa etal., 2008).EULAG solvesthenonhydrostatic,anelastic equationsofmotion,herewritteninanextended perturbationalform (Smolarkiewicz and Margolin,1997): ∇·( ρ bv)=0,(1) Dv Dt=−∇ π ′−g Θ ′ Θ b+M+D+F− α v′,(2) D Θ ′ Dt=−v·∇ Θ e+H− βΘ ′.(3) De Dt=S(4) Thesetofanelastic equations(1)-(4)describesthe anelasticmass continuityequation (1), thethree componentsofthemomentumequation (2),and thethermodynamic equation (3), respectively.The equation (4) forthesubgrid-scale(SGS)turbulentkinetic energy (TKE) ecompletesthesystemofequations.In(1)-(4),theoperators∇and ∇·symbolize gradient and divergence,whileD/Dt= ∂ / ∂ t+v·∇isthematerialderivative,and visthephysical velocity vector.Thevector representing thegravitationalacceleration g=(0,0,−g)Toccurs inthebuoyancytermofEqu.(2).Thequantities ρ b(z)and Θ b(z)refertothebasicstates, prescribed hydrostaticreference profilesusuallyemployedinthe anelastic approximated equations(Clarkand Farley,1984). Inaddition tothehorizontally homogeneousbasicstate,amoregeneralambient(environmental)stateisdenoted by thesubscripte.The corresponding variablesmay varyinthe horizontaldirectionsand they havetosatisfyEqu.(1)-(3);see Prusa etal.(2008) foradiscussion ofambientstate and itsbenefits.Theprimed variablesv′and Θ ′appearing inEqu. (2)-(3)correspond to deviationsfromthe environmentalvariablesveand Θ e.Thequantity π ′inthelinearized pressuregradient terminEqu.(2)denotesadensity normalized pressure deviation. Thetermsproportional to α and β denotewave absorbing devicesusedat theupper boundary ofthe computationaldomain.Thesource termsDand Hnotexplicitlystated 84 A PPENDIX E – B RÖTZ ET AL . (2013) Early-morning flowtransition inavalleyinlow-mountainterrain 5 inEqu.(2)and (3)symbolize theviscousdissipation ofmomentumand thediffusion of heat,respectively.Fsymbolizesanadditionalforcing forspecifiedsimulations,see below. Theformulation oftheTKEproduction and dissipation termhiddeninSand the applied parametersfollowthedescription ofSorbjan(1996). ThequantityMdenotesmetricforcesduetothe curvilinearity oftheunderlying physicalsystem.Inthepresentwork,anon-orthogonal terrain-following systemofcoordinates (x,y,z)=(x,y,H(z−h)/(H−h)) isusedwhichassumesamodeldepthHand anirregularlowerboundaryh(x,y)(Gal-Chenand Somerville,1975;Smolarkiewicz and Margolin, 1993;Wediand Smolarkiewicz,2004).The explicit formulation ofthetransformedsystem ofequationscan befound inPrusa and Smolarkiewicz (2003)or,morerecently,inK¨ uhnlein etal.(2012).Insymbolicform,theresulting systemofmotion fortheprognosticvariables Ψ =u,v,w, Θ ,ecan bewrittenasaflux-formEulerianconservation law ∂ ∂ t( ρ ∗ Ψ )+∇·(v ρ ∗ Ψ )= ρ ∗F Ψ (5) where ρ ∗= ρ bG,withGastheJacobian ofthetransformation.Afinitedifference approximation ofEqu.(5)is Ψ n+1=MPDATA Ψ n+0.5 ∆ tF Ψ  n,vn+1 2, ρ ∗+0.5 ∆ t ρ ∗F Ψ  n+1(6) whereMPDATA1standsforthenon-oscillatoryforward-in-time(NFT)advection transport schemedescribedinSmolarkiewicz and Margolin(1998).The elliptic equation forpressure is solvediterativelywithaKrylov-sub space solver,see Thomasetal.(2003).Bothelements areintegralpartoftheEULAG and arefundamentalforthestability ofthe code and the reliability oftheresults. 2.2.2 Simulation strategy Among thebroadrangeofapplicationsdocumentedinliterature,EULAG was successfully appliedtoatmosphericboundary-layerflows(see Smolarkiewicz etal., 2007;Piotrowski etal., 2009).Forthequestionsinvestigatedinthispaper,theset-up waschoseninthefollowing way. Thenumericalsimulationsare conductedinadomain of(Lx,Ly,H)=(7680m,7680m, 2430m)witharegulargridsize of ∆ x= ∆ y= ∆ z=30m.Forasimulation of2.5h physical time,30000 timestepswith ∆ t=0.3sarenecessary.Theheighth(x,y)ofthelower boundaryistakenfromtheASTERdigital topographicdataset(NASALand Processes DistributedActiveArchiveCenterNASALPDAAC,2001)ina30m×30mregular resolution.Inall simulations shown herethe computationaldomainisperiodicinthehorizontal directions.Toenablethisperiodicitythetopography was smoothlyrelaxedwithinaframe around the actualregion ofinterest.Duetothe complex orography and thelowinversion layerheight thewidth oftheframe could be chosento be300m. Forall simulationsananelasticbasicstatewithabackground stratification N=0.01s−1 isusedaccording toClarkand Farley(1984),resulting inexponentially decreasing ρ band increasing Θ b-profiles. Toinvestigatetheguiding questionsofthis study,two differentsimulation set-upswere designed.First,idealizedsimulationsofanevolving convectiveboundarylayer (CBL)over flat terrainh(x,y)=0were conductedand theyaredenoted by S,and,secondly,theCBL 1MPDATAstandsforMultidimensionalPositive-DefiniteAdvection TransportAlgorithm A PPENDIX E – B RÖTZ ET AL . (2013) 85 6Bj¨ ornBr¨ otz etal. Table1Set-upsforthesimulationsinthis study name terrainstratification zi,(m) wind forcing dx,(m) belowzi S1flatneutral800 off 20 S2flat– 0 off 20 R1complex neutral800 off 30 R2complex – 0 on (during p2)30 was simulated over realistictopography h(x,y)and thesesimulationsaredenoted by R,see Table1. All simulationswereinitializedwitharesting fluid,and two differentambientpotential temperatureprofiles Θ e(z)were appliedto distinguish betweenanearlymixedlayerwitha capping inversion layeratzi=800m(SimulationsS1and R1)and astablystratifiedambient state covering thewholedepth ofthe computationaldomain(simulationsS2and R2): SimulationsS1and R1: Θ e=( Θ 0forh≤z<zi Θ 01+N2 g(z−zi)forzi≤z≤H(7) SimulationsS2and R2: Θ e= Θ 01+N2 gzforh≤z≤H(8) Whitenoisewithanamplitudeof0.001ms−1wasaddedtotheinitialverticalwind fieldin ordertoinitiate convectivemotions.Forthe ensemblerunsanalysedinSection 3.2,eight independentrealizationsweresimulated using theset-up R1.Forthispurpose,therandom generatorwas seeded differentlyat theinitial timeforeachrealization.Becausethethree wind componentsare zero before adding thenoise,therandomdisturbance is100%ofthe absolutevalueofthewind vector.Thisensuresthat the eightrealizationsarestatisticallyindependent.Thenumericalsimulationswere conductedforadryatmosphere.At thesurface asensibleheatflux QH=0.05Kms−1was specifiedinall runs.Thehomogeneousheating can bejustified becauseflux differencesbetween different typesofland surfacesturned out to benegligibleintheobservedearly-morning situations(see Eigenmann etal., 2011).The effectoforographicshading isnot takenintoaccount.Orographically-inducedflows(valley winds,upslopeflows,etc.)are expectedtomainly dominatethepropertiesoftheCBLinthe valleyat thistimeoftheday. During thenight-day-transition,the along-valleywindsarepartofamountain plaincirculation betweenamountainmassive and anadjacentplane,in ourcasetheBlackForestand theUpperRhineValley,respectively.Duetothesmall computationaldomain,the effectof thismeso-scale circulation on theflowinthevalleyismodelled by anadditionaldynamical forcing F=(0,−v0(z) τ −1,0)forthemeridionalwind componentv,where τ =tend −tbegis theperiod whentheforcing isapplied.Thereference profileforthehorizontalwind speed v0(z)wasderivedfromtheSodarobservations.Inthesimulation R2the additionalforcing Fisapplied.Theperiod fromt=0totbegrepresentstheobservedlow wind period p1.The period fromtbegtothe end ofthesimulation representstheup-valleywind period p2.Inthe remainderofthe articletwosimulation timesreferredtoast1foratimeinp1and t2fora timeinp2are chosenin ordertocomparedifferencesinthesimulated periodsp1and p2. 86 A PPENDIX E – B RÖTZ ET AL . (2013) Early-morning flowtransition inavalleyinlow-mountainterrain 7 p1p2 QH (−Qs* − QG) −1 0.0 0.2 0.4 0.6 0.8 1.0 (a) p1p2 QE (−Qs* − QG) −1 0.0 0.2 0.4 0.6 0.8 1.0 (b) p1p2 (QH+QE) (−Qs* − QG) −1 0.0 0.2 0.4 0.6 0.8 1.0 (c) Fig.1Barplotsofthemean ofthefluxesofQH(a),QE(b)and thesumofboth(c)normalizedwiththe available energy −Q∗ S−QGduring thelow wind speed period p1and thefirst2 hoursofp2.Averagevalues forthe23 selected daysat theFußbachsite aregivenforboth periods.Alsoshownarethe95%confidence intervalswhichindicatesignificantdifferencesinthemean valuesfor (a)and (c). 3Resultsand Discussion 3.1Modification ofthe energy balance by thevalleywind Toanalysethe effectofthevalleywind on the energy balance oftheECflux measurements, theselected periodsp1and p2are analysedseparately.Figure1showsthemean ofthefluxes QHand QEand thesumofboth normalizedwiththe available energy −Q∗ S−QGduring the low wind speed period p1and thefirst2 hoursofp2.Thisintervalwaschosenin orderto makethedatabasisofboth periodscomparable.Aclosedenergy balance meansthat the ratio ofthesumoftheturbulentfluxesQH+QEand the available energy −Q∗ S−QGisequal to one.Altogether,the energy balance isclosedinp1,whileinp2aresidualof16%occurs on average(see Fig.1c).Thelattervalueisclosetothe averageresidualof21%found during the entireCOPS campaign at this site(see Eigenmann etal., 2011). Regarding Fig.1a and b,therelativeflux contributionsmissing in period p2compared to period p1have exactlytheproportionsofthebuoyancyflux ratio.Thebuoyancyflux ratiowould distribute about85%oftheresidual toQHand 15%toQEforatypicalBowen ratio ofabout0.45 intheobservedearly-morning situations.As such,Fig.1supportsthe application ofthebuoyancyflux ratioapproach(see Charuchittipanetal., 2013) forthe correction ofthe energy balance.Themissing flux componentsin period p2are assumedto be transportedwithin buoyancy-drivensecondarycirculationsnotcaptured by theECmeasurements(e.g.Foken,2008).Thetransferofthemissing energy intothesecondarycirculation mainly happensatsignificantsurface heterogeneitieswhichcan befound overcomplexterrain.Advection-dominated processes(also notcaptured by theEC)probablyleadtothe transportofthemissing energy totheseheterogeneities.Aswind speedsvanishin period p1,no energy istransferredintosecondarycirculationsand the energy balance isclosed. However,the along-valleywind in period p2leadstomissing advectiveflux components and thustotheobservedresidual inthisperiod. A PPENDIX E – B RÖTZ ET AL . (2013) 87 8Bj¨ ornBr¨ otz etal. ww* −1 probability density −2 −1 0 1 2 3 0.0 0.5 1.0 1.5 2.0 Sodar p1 LES updraught areas LES all points zzi −1 = 0.25 (a) −2 −1 0 1 2 3 0.0 0.5 1.0 1.5 2.0 −2 −1 0 1 2 3 0.0 0.5 1.0 1.5 2.0 ww* −1 probability density −2 −1 0 1 2 3 0.0 0.5 1.0 1.5 2.0 Sodar p1 LES updraught areas LES all points zzi −1 = 0.50 (b) −2 −1 0 1 2 3 0.0 0.5 1.0 1.5 2.0 −2 −1 0 1 2 3 0.0 0.5 1.0 1.5 2.0 ww* −1 probability density −2 −1 0 1 2 3 0.0 0.5 1.0 1.5 2.0 Sodar p1 LES updraught areas LES all points zzi −1 = 0.75 (c) −2 −1 0 1 2 3 0.0 0.5 1.0 1.5 2.0 −2 −1 0 1 2 3 0.0 0.5 1.0 1.5 2.0 Fig.2Probability densitiesofthenormalized verticalwind speedww−1 ∗forthree heights(zz−1 i= 0.25,0.5,0.75).Histograms arederivedfromSodarobservationsat theFußbachsiteduring thelow wind speed periodsp1on the23 selected days,while curvesarederivedfromthesimulation S2.Thedottedcurve showstheprobability densityfunction (pdf) forall pointsinthehorizontalplane,thedashedcurveshowsthe pdf forthe conditionallysampled updraughtareasonly. Thefinding discussedabove alsosupportsthe choice ofa constantheatflux forcing for thetransientsimulation R2(see Sect.2.2).Thesamerelativeforcing by QHisachieved forboth periodsp1and p2by adding (forsimplification)100%oftheresidual toQH.In thisway,theforcing ofperiod p1canalso beusedforperiod p2.Moreover,no significant relativeflux differencesofQEexist in both periods(see Fig.1b).Thus,forthequestions addressedinthis study,it appearsto be appropriatetoconcentrateon drymodelruns. 3.2Coherentstructuresinthevalleyimposed by surrounding orography In ordertoinvestigatethe early-morning CBLevolution insidethevalley,Sodardatafrom themorning period p1oftheselected dayswere chosentocreatethehistogramsofthe normalized verticalwind ww−1 ∗(see Fig.2).Theobserved distributionsofthis study deviate stronglyfromprobability densityfunctions(pdfs)observed overflat,homogeneousenvironmentsasreported by many studies(e.g.Deardorff and Willis,1985;Stull,1988).Inthese 88 A PPENDIX E – B RÖTZ ET AL . (2013) Early-morning flowtransition inavalleyinlow-mountainterrain 9 0 2 4 6 x (km) 0 2 4 6 y (km) -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 -0.05 0.05 0.1 0.2 0.3 0.4 0.5 0.6 Fig.3Ensemble and time mean oftheverticalwind speedinms−1at300ma.s.l.forsimulation R1(colourcoded).Blacksolidlinesmarktheorography instepsof50m. Greycontoursmarkintersection withthe orography.Theredframe representsthesection ofthevalleyshowninFig.5.Theposition oftheSodaris indicated by theblackcircle. studiesthepdfsareright-skewedand showanegativemaximum.Especiallyinthelowerpart oftheboundarylayer (zz−1 i=0.25)themaximumoftheobserved distribution atFußbach siteis shiftedtowardsweak positivevaluesinstead oftheweak negativevaluesknownfrom literature.Alsotheobserved histogramisfarless skewedat thisheight.Tofind the cause ofthisbehaviour,idealizedsimulationsofaCBLare carried out(Simulation S1and S2, describedinSect.2.2). To gainconfidence inthesimulationsthewell known pdfsoftheverticalwind ina convectiveboundarylayerare calculatedforthesimulated dataofset-up S1.Very good agreementwiththepublished valuesisfound (notshown).Moreover,thesimulation results showthewell-knownspokepatternsofcoherentconvectivemotion knownfromnumerous numericalstudies(e.g.Schmidtand Schumann,1989).Thedatafromthesimulation of set-up S2isthen usedtocreatethepdfsforaCBLwith growing mixedlayer,see Fig.2 (situation more closeto observation period p1).Inthis set-up,spokepatternsevolveinthe simulated boundarylayerthatgrowslightlyinsize astheinversion layerheightgrowsin time.Aconditionalsampling isappliedto obtainthepdfsofthe coherentupdraughtareas. Theresulting pdfsresemblethepropertiesofthehistogramsfromtheSodardata(see Fig.2).Themaximumand theskewness ofthepdfsofthe conditionallysampled updraught areasmatchwell withthoseofthehistogramsoftheobservationaldataforall heights.Only the absolutenumbersoftheprobability density do notfullymatch.Apossible explanation A PPENDIX E – B RÖTZ ET AL . (2013) 89 10 Bj¨ ornBr¨ otz etal. forthisisthat theSodarinstrumentaveragesovera certain volume,sothat theprobability density ofvaluesaround zero getsincreased.Thiseffectbecomeslargerwith height. Weinterpret thesefindingsasfollows:It iswell knownthatcoherentconvectivemotions evolveinaCBLwhichremain quasi-stationaryinspace and time(see e.g.Stull,1988).This meansthat thelocation where a possibleSodarinstrument islocatedwill remain underan updraughtoradowndraughtarea foralong time.Asa consequence,it isverylikelythat aSodarmeasurementwill captureonlythestatisticalpropertiesofapartofthevelocity spectrum.Thefact that theresult fromFig.2stemsfroma compositeover23 periodswith similaroverall conditions suggeststhehypothesisthat theSodarinstrumentwaspreferentiallylocatedatan updraughtregion. To verifythishypothesis,anensembleofeight large-eddy simulationswascarried out withrealistictopography at thelowerboundary(simulation R1)and different initialnoise seeding forw(see Sect.2.2.2).Figure3showsthe ensemble and timemean ofthevertical wind speedat300ma.s.l.(approximately 130mabovethevalleyfloor).Although we applied bothatemporalmeanand anensemblemeantothesimulated data,coherentpatterns oftheCBLflowfieldinsidethevalleyremain.Thismeanflowfield hasalargeramplitude thanitsanaloguefromsimulationsoverflat terrain.Weinterpret thisfinding asfollows:The surrounding ridgesimpose coherentconvectivemotionstothevalleyflowatspecificlocationsduring the early-morning p1periods.Theirpositionsrelativetotheridgespersist in contrast tothe changing locationsofthe coherentstructuresintheflatCBLsimulation.The position oftheobservationalsiteismarked by ablackcircleinFig.3and showsthat thesite islocatedinan updraughtregion. 3.3Vertical transport inthe early-morning valleyatmosphere SpectralanalysisoftheECmeasurementsinthevalleyshowedanincreaseofspectralpower withinturbulentscalesofafewminutesduring thelow wind speed period p1inthemorning (see Eigenmann etal., 2009).Thesetimescalescould berelatedtothepresence oflarge coherentverticalstructures(e.g.plumesorupdraughts)withaspatialextent intheorder oftheboundary-layerheight,whichareknownto beresponsibleforthemajority ofthe transportwithintheCBL(see e.g.Stull,1988;Chandra etal., 2010).Theoccurrence of theseturbulentscalesintheground-basedECdataindicatesthatduring theperiod p1,air veryclosetotheground isableto betransported upwardsveryefficiently by non-local large-eddy transportprocesses.Thefree convective conditionsdetectedsimultaneously by theECmeasurementsalsosupport thesefindings.Bytheonsetoftheup-valleywind these turbulentscalesdisappear fromthedataindicating that theturbulenttransportofnear-ground airbecameless effective.The effectivevertical transport in period p1isimportantbecause airmassesclosetothevalley bottomarehumid,haveacharacteristic chemicalcomposition, and may possibly bepolluted.The effectofthefree convectivereleaseofsurface layer airmassesfromthevalley bottomon ozonemeasurementsatamountain-top station was recentlyreported by Mayeretal.(2008). During thesefree convectivesituationsin period p1,theSodar/RASS observedstrong verticalupdraughtsintothestablystratified valleyatmosphere.Forillustration,Fig.4shows forCOPS IOP15b,i.e.13 August2007,themorning evolution ofverticalwind and virtualpotential temperature.InthelowerpanelofFig.4theobserved verticalwind is shown from0500 to 1300 UTC.Theperiod oflowhorizontalwind speedinthemorning ismarked by verticaldashedlines.ThetimesoftheprofilesplottedinFig.4a-dareindicatedinthe lowerpanel.At time(a)thestablestratification is shownshortlyaftersunrise.In(b)the 90 A PPENDIX E – B RÖTZ ET AL . (2013) Early-morning flowtransition inavalleyinlow-mountainterrain 11 50 100 150 200 250 300 350 400 19 21 23 25 height a.g.l. (m) (a) 50 100 150 200 250 300 350 400 19 21 23 25 (b) virtual potential temperature (°C) 50 100 150 200 250 300 350 400 19 21 23 25 (c) 50 100 150 200 250 300 350 400 19 21 23 25 (d) 60 140 220 300 380 460 540 620 700 height a.g.l. (m) time (UTC) (a) (b) (c) (d) Vertical wind speed (ms−1) >0.8 0.6 0.4 0.2 0.1 −0.1 −0.2 −0.4 −0.6 −0.8 <−0.8 0500 0600 0700 0800 0900 1000 1100 1200 1300 Fig.4Upperpanel(a-d):Profilesofthevirtualpotential temperatureobserved by theSodar/RASS inthe morning hoursofCOPS IOP15b (13 August2007).Thetimesoftheprofilesaremarkedinthelowerpanel. Lowerpanel(fromEigenmann etal., 2009,modified):Verticalwind speedsincolourmeasured by theSodar/RASS from0500-1300 UTC.Theblack dashed vertical linesindicatetheperiod ofvanishing horizontal wind speeds. profileisrepresentativeforaperiod ofstrong coherentverticalupdraughts.Nearly neutral stratification below160mabovethevalleyfloorwasobservedinthisperiod.A weakstable stratification above160mcan beseenwhilethe corresponding verticalwind speedsremain positive.Thestrong updraughtperiod isinterrupted by aperiod ofweakerverticalwinds. Theprofilein(c)showsthatduring this short interruption theoriginalstablestratification recovers.Afterthisinterruption theverticalwind isagain positive and theprofilein(d) showsaneutralorslightly unstableprofile.Inthelightofthepreviousanalysis,thisindividualsceneisinterpretedasfollows:The convection organizes,influenced by theorography, inawaythat theupdraughtand downdraughtareasremain quasi-stationaryat theirspatial location (see Fig.3).TheimmobileSodar/RASS instrumentobservedthisquasi-stationary updraughtarea foraperiod ofapproximatelytwoand ahalfhours(0800 until 1030 UTC). Thisperiod isinterrupted by ashortperiod ofweakerwindsataround 0920UTC,when thequasi-stationary updraughtarea slightlymovesoutoftheviewoftheSodar/RASS instrument,sothat thepropertiesofanattached downdraughtarea are also observed.Inthis shortperiod,it can beseenthat thestratification ofthevalleyatmosphereoutsideofthe updraughtsis still stable(Fig.4c). To betterunderstand thestateoftheboundarylayerinwhichtheseobservationswere made,thetransientsimulation R2(see Sect.2.2)wascarried outand analysed.InFig.5, snapshotsofthefield oftheverticalwind speedareshownfortwo heightsand forthetime A PPENDIX E – B RÖTZ ET AL . (2013) 91 12 Bj¨ ornBr¨ otz etal. Fig.5Instantaneous situationsat thetime t1inthesimulatedlow wind period p1(a and b)and forthetime t2 inthesimulated up-valleywind period p2(c and d)ofthesimulation R2.Verticalwind componentwinms−1 iscolour-coded.The cross sectionsareplacedat210m(a and c)and 300ma.s.l.(band d),respectively.The area shown hereismarked by theredframe inFig.3. 98 A PPENDIX E – B RÖTZ ET AL . (2013) Early-morning flowtransition inavalleyinlow-mountainterrain 19 PrusaJM,Smolarkiewicz PK,WyszogrodzkiAA (2008)Eulag,a computationalmodelformultiscaleflows.ComputFluids37(9):1193–1207,DOI 10.1016/J.Compfluid.2007.12.001 RaaschS,HarbuschG(2001)Ananalysisofsecondarycirculationsand theireffectscaused by small-scalesurface inhomogeneitiesusing large-eddy simulation.Boundary-Layer Meteorol101:31–59,DOI10.1023/A:1019297504109 RotachMW,AndrettaM,Calanca P,WeigelAP,Weiss A(2008)Boundarylayercharacteristicsand turbulentexchangemechanismsin highlycomplexterrain.ActaGeophys 56(1):194–219,DOI10.2478/s11600-007-0043-1 SchmidtH,Schumann U(1989)Coherentstructureofthe convectiveboundary layerderivedfromlarge-eddy simulations.JFluidMech 200:511–562,DOI 10.1017/S0022112089000753 Schumann U(1987)The countergradientheatflux inturbulentstratifiedflows.NuclEng Des100(3):255–262,DOI10.1016/0029-5493(87)90078-1 SegalM,Arritt RW(1992)Nonclassicalmesoscale circulationscaused by surface sensibleheat-flux gradients.Bull AmMeteorolSoc73(10):1593–1604,DOI10.1175/15200477(1992)073<1593:NMCCBS>2.0.CO;2 Smolarkiewicz PK,MargolinLG(1993)Onforward-in-timedifferencing forfluids: Extension toa curvilinear framework.Mon WeatherRev 121(6):1847–1859,DOI 10.1175/1520-0493(1993)121<1847:OFITDF>2.0.CO;2 Smolarkiewicz PK,MargolinLG(1998)MPDATA:afinite-difference solver forgeophysicalflows.JComputPhys140(2):459–480,DOI10.1006/jcph.1998.5901 Smolarkiewicz PK,GrubisicV,MargolinLG(1997)Onforward-in-timedifferencing forfluids:Stopping criteriaforiterativesolutionsofanelastic pressure equations.Mon WeatherRev 125(4):647–654,DOI10.1175/15200493(1997)125<0647:OFITDF>2.0.CO;2 Smolarkiewicz PK,SharmanR,Weil J,PerrySG,HeistD,BowkerG(2007)Building resolving large-eddy simulationsand comparison withwind tunnelexperiments.JComput Phys227:633–653,DOI10.1016/j.jcp.2007.08.005 SorbjanZ(1996)Numericalstudy ofpenetrative and ”solidlid”nonpenetrative convectiveboundarylayers.JAtmosSci53:101–112,DOI10.1175/15200469(1996)053<0101:NSOPAL>2.0.CO;2 Stoy P,MauderM,FokenT,MarcollaB,Boegh E,IbromA,ArainM,ArnethA,Aurela M,BernhoferC,Cescatti A,DellwikE,Duce P,GianelleD,vanGorselE,KielyG, KnohlA,MangolisH,McCaugheyH,MerboldL,MontagnaniL,PapaleD,Reichstein M,Serrano-OrtizP,SottocornolaM,SaundersM,Spano D,VaccariF,VarlaginA(2013) Adata-drivenanalysisofenergy balance closure across FLUXNET researchsites:The roleoflandscape-scaleheterogeneity.AgricForMeteorol,accepted Stull RB (1988)Anintroduction to boundarylayermeteorology.KluwerAcademicPublishers,Dordrecht ThomasSJ,HackerJP,Smolarkiewicz PK,Stull RB (2003)Spectralpreconditioners fornonhydrostatic atmosphericmodels.Mon WeatherRev 131(10):2464–2478,DOI 10.1175/1520-0493(2003)131<2464:SPFNAM>2.0.CO;2 TwineTE,KustasWP,NormanJM,Cook DR,HouserPR,MeyersTP,PruegerJH,Starks PJ,WeselyML(2000)Correcting eddy-covariance flux underestimatesoveragrassland. AgricForMeteorol103:279–300 Walko R,Cotton W,PielkeR(1992)Large-eddy simulationsofthe effectsofhillyterrain on the convectiveboundary-layer.Boundary-LayerMeteorol58(1-2):133–150,DOI 10.1007/BF00120755 A PPENDIX E – B RÖTZ ET AL . (2013) 99 20 Bj¨ ornBr¨ otz etal. WeckwerthTM,Wilson JW,WakimotoRM,Crook NA (1997)Horizontalconvectiverolls:determining the environmentalconditions supporting theirexistence and characteristics.Mon WeatherRev 125:505–526,DOI10.1175/15200493(1997)125<0505:HCRDTE>2.0.CO;2 WediNP,Smolarkiewicz PK(2004)Extending Gal-Chenand Somervilleterrain-following coordinatetransformation on time-dependentcurvilinearboundaries.JComputPhys 193:1–20,DOI10.1016/j.jcp.2003.07.034 WeigelA,ChowF,RotachM(2007)The effectofmountainoustopography on moisture exchangebetweenthe ”surface” and thefree atmosphere.Boundary-LayerMeteorol 125:227–244,DOI10.1007/s10546-006-9120-2 WhitemanCD(1990)Observationsofthermally developedwind systemsinmountainous terrain.In:BlumenW(ed)Atmosphericprocessesovercomplexterrain,Meteorological monographs,vol23 (45),AmericanMeteorologicalSociety,Boston,Massachusetts,pp 5–42 WhitemanCD(2000)MountainMeteorology:Fundamentalsand Applications.OxfordUniversityPress,USA WulfmeyerV,BehrendtA,KottmeierC,CorsmeierU,Barthlott C,CraigGC,HagenM,AlthausenD,AoshimaF,ArpagausM,BauerHS,Bennett L,BlythA,BrandauC,Champollion C,Crewell S,DickG,DiGirolamoP,DorningerM,DufournetY,Eigenmann R,Engelmann R,FlamantC,FokenT,GorgasT,GrzeschikM,HandwerkerJ,Hauck C,H¨ ollerH,Junkermann W,Kalthoff N,KiemleC,Klink S,K¨ onigM,Krauss L,Long CN,MadonnaF,MobbsS,NeiningerB,PalS,PetersG,Pigeon G,RichardE,Rotach MW,RusschenbergH,SchwitallaT,SmithV,SteinackerR,Trentmann J,TurnerDD, vanBaelenJ,VogtS,VolkertH,WeckwerthT,Wernli H,WieserA,WirthM(2011) TheConvective and Orographically-inducedPrecipitation Study (COPS): thescientific strategy,thefield phase,and research highlights.QJRoy MeteorolSoc137(S1):3–30, DOI10.1002/qj.752 ZardiD,WhitemanCD(2013)Diurnalmountainwind systems.In:ChowFK,DeWekker SF,SnyderBJ(eds)MountainWeatherResearchand Forecasting,SpringerNetherlands, Dordrecht,pp 35–119 Zhou D,Eigenmann R,BabelW,FokenT,MaY(2011)Thestudy ofnear-ground free convection conditionsatnamcostation on thetibetan plateau.TheorApplClimatol105:217– 228,DOI10.1007/s00704-010-0393-5 100 E RKLÄRUNG Erklärung Hiermit erkläre ich, dass ich die Arbeit selbständig verfasst und keine anderen als die von mir angegebenen Quellen und Hilfsmittel benutzt habe. Ferner erkläre ich, dass ich anderweitig mit oder ohne Erfolg nicht versucht habe, diese Dissertation einzureichen. Ich habe keine gleichartige Doktorprüfung an einer anderen Hochschule endgültig nicht bestanden. Bayreuth, den ____________________ ____________________ Rafael Eigenmann