# The Effect of Surface Heterogeneity on the Structure Parameters of Temperature and Specific Humidity: A Large-Eddy Simulation Case Study for the LITFASS-2003 Experiment

- First Online:

- Received:
- Accepted:

## Abstract

We conduct a high-resolution large-eddy simulation (LES) case study in order to investigate the effects of surface heterogeneity on the (local) structure parameters of potential temperature \(C_T^2\) and specific humidity \(C_q^2\) in the convective boundary layer (CBL). The kilometre-scale heterogeneous land-use distribution as observed during the LITFASS-2003 experiment was prescribed at the surface of the LES model in order to simulate a realistic CBL development from the early morning until early afternoon. The surface patches are irregularly distributed and represent different land-use types that exhibit different roughness conditions as well as near-surface fluxes of sensible and latent heat. In the analysis, particular attention is given to the Monin–Obukhov similarity theory (MOST) relationships and local free convection (LFC) scaling for structure parameters in the surface layer, relating \(C_T^2\) and \(C_q^2\) to the surface fluxes of sensible and latent heat, respectively. Moreover we study possible effects of surface heterogeneity on scintillometer measurements that are usually performed in the surface layer. The LES data show that the local structure parameters reflect the surface heterogeneity pattern up to heights of 100–200 m. The assumption of a blending height, i.e. the height above the surface where the surface heterogeneity pattern is no longer visible in the structure parameters, is studied by means of a two-dimensional correlation analysis. We show that no such blending height is found at typical heights of scintillometer measurements for the studied case. Moreover, \(C_q^2\) does not follow MOST, which is ascribed to the entrainment of dry air at the top of the boundary layer. The application of MOST and LFC scaling to elevated \(C_T^2\) data still gives reliable estimates of the surface sensible heat flux. We show, however, that this flux, derived from scintillometer data, is only representative of the footprint area of the scintillometer, whose size depends strongly on the synoptic conditions.

### Keywords

Convective boundary layer Large-eddy simulation Scintillometry Surface heterogeneity Turbulent structure parameter## 1 Introduction

The turbulent surface fluxes of sensible and latent heat play an important role in the vertical transport of energy and water vapour in the atmospheric boundary layer (ABL). The measurement of the area-averaged fluxes at a regional scale is necessary both for a better understanding of the meteorological and hydrological processes as well as for the validation of parametrizations in numerical weather prediction (NWP) and climate models (De Bruin et al. 1993; Li et al. 2012; Beyrich et al. 2012; Braam et al. 2012). The grid resolution in NWP models is usually of the size of several km. For heterogeneous surfaces, it is thus essential to measure surface fluxes that are representative of an area of the order of a few km\(^{2}\). Over homogeneous terrain, point measurements using the eddy-covariance technique are the traditional and most common means of determining the surface fluxes (Andreas 1991; Lee et al. 2004; Braam et al. 2012). Natural landscapes rarely provide horizontally homogeneous conditions such that the grid boxes in NWP models often contain different surface patches of farmland, settlements, water, forest, etc. The local surface fluxes of the different surface types, however, might differ significantly (e.g. Bange et al. 2006; Beyrich et al. 2006) and local point measurements over all relevant land-use types would be necessary to derive area-averaged surface fluxes based on a suitable averaging strategy (Gottschalk et al. 1999; Beyrich et al. 2006). This requires considerable experimental effort.

As an alternative, scintillometry offers a technique that allows for estimating the surface sensible and latent heat fluxes by measuring turbulent density fluctuations in the surface layer in terms of the refractive index structure parameter \(C_n^2\) as a line average over horizontal distances of up to 10 km (e.g. Kohsiek et al. 2002; Meijninger et al. 2002b, a, 2006; Evans et al. 2012, inter alia). Hill (1978), for example, showed that \(C_n^2\) is related to the structure parameters of temperature, \(C_T^2\), humidity, \(C_q^2\), and the cross-structure parameter \(C_{Tq}\), since the density fluctuations are dominantly caused by fluctuations of temperature and humidity. The relative contribution of both depends on the electromagnetic wavelength. Temperature fluctuations are absolutely dominant for optical and near-infrared electromagnetic wave propagation, except for very moist conditions, while both temperature and humidity fluctuations are relevant for millimetre waves except for very dry conditions.

Provided that the ABL is neither extremely moist or dry, temperature fluctuations are absolutely dominant for optical and near-infrared electromagnetic wave propagation, while both humidity and temperature fluctuations are relevant for millimetre waves. As a consequence, an optical large-aperture scintillometer (LAS) can be used to determine path-averaged values of \(C_T^2\), and a microwave scintillometer (MWS) in combination with LAS can provide both \(C_T^2\) and \(C_q^2\) (e.g. Kohsiek and Herben 1983; Meijninger et al. 2002a, 2006; Lüdi et al. 2005). In order to determine the surface sensible and latent heat fluxes from the estimates of \(C_T^2\) and \(C_q^2\), respectively, Monin–Obukhov similarity theory (MOST) is applied.

Theoretically, the application of MOST is limited to horizontally homogeneous surfaces (Andreas 1991; Beyrich et al. 2012). The application of MOST for structure parameters over a heterogeneous surface thus requires that the height of the measurement is above a blending height for structure parameters, where any signal from the present surface heterogeneity is no longer visible in the sampled turbulence field (Wieringa 1976; Mahrt 2000; Meijninger et al. 2002b; Sühring and Raasch 2013). In this way, the measurement sees a virtually homogeneous surface that incorporates all heterogeneity effects. So far the blending height concept and its existence has been discussed controversially in the literature (e.g. Meijninger et al. 2002b; Bange et al. 2006; Sühring and Raasch 2013).

Meijninger et al. (2002b) estimated the blending height for \(C_T^2\) to be not higher than 14 m during the Flevoland field experiment based on measurements. In contrast, the large-eddy simulation (LES) study of Sühring and Raasch (2013) over the moderately heterogeneous LITFASS area in eastern Germany suggested a clear dependence of the turbulent fluxes on the underlying surface type up to the top of the convective boundary layer (CBL). We might hence consider surface heterogeneity to be a significant factor that might lead to a decorrelation between scintillometer-derived surface fluxes from a LAS or MWS path and the area-averaged surface fluxes for a domain of several km\(^{2}\) around this scintillometer path. This, however, has not been demonstrated so far, but Beyrich et al. (2006) found reasonable agreement between LAS/MWS and aggregated eddy-covariance surface fluxes over the LITFASS area.

While structure parameters are traditionally defined for statistically homogeneous and isotropic fields, they can also be defined as a local and time-dependent quantity within a chosen volume of air so long as an inertial subrange is present (e.g. Cheinet and Siebesma 2009). Recently, van den Kroonenberg et al. (2012) employed small unmanned aircraft to study the variability of the local estimates of \(C_T^2\) along a LAS path during the LITFASS-2009 experiment. They found a noticeable variability of \(C_T^2\) along the path and ascribed this to both temporal variations as well as to the underlying surface heterogeneity. However, Sühring and Raasch (2013) pointed out that sufficient independent flight measurements are required to obtain a significant estimate of a heterogeneity-induced effect in terms of turbulent fluxes. The question, as to whether there is a heterogeneity-induced effect on \(C_T^2\) and \(C_q^2\) and their local estimates in the surface layer and if this can be measured by aircraft (and scintillometers), is still open.

LES is a promising technique for studying the effect of surface heterogeneity on the structure parameters and their MOST relationships. Unlike in situ measurements, where the (heterogeneous) surface fluxes are basically unknown, they can be explicitly prescribed in the LES model. Previous LES studies have shown that vertical profiles and local estimates of \(C_T^2\) and \(C_q^2\) can be reliably derived from LES data for the homogeneously-heated CBL (Peltier and Wyngaard 1995; Cheinet and Siebesma 2009; Cheinet and Cumin 2011; Wilson and Fedorovich 2012; Maronga et al. 2013b). Cheinet and Siebesma (2009) studied the spatial variability of \(C_T^2\) and found a bimodal log-normal distribution in the lower boundary layer, which was previously also found in the Sodar measurements of Petenko and Shurygin (1999). Cheinet and Cumin (2011) studied the behaviour of \(C_T^2\) and \(C_q^2\) in the entrainment-drying CBL and found that the distribution of \(C_q^2\) was determined by entrained air parcels from the free atmosphere, whereas \(C_T^2\) was dominated by near-surface convective plumes. Cheinet and Siebesma (2007) used a wave propagation modelling framework to derive the scintillation rate and coherence length from virtual path measurements in their LES. They found that the variability of their virtual measurements of \(C_n^2\) increased with height, while the path mean \(C_n^2\) decreased. However, the coarse spatial resolution of their LES did not allow study of the wave propagation at realistic scintillometer heights on the order of 50 m, since they could not resolve the surface layer.

A first comparison of LES with in situ aircraft and LAS data was performed by Maronga et al. (2013b). They also employed virtual LAS (VLAS) measurements in their LES at realistic heights between 30 and 70 m above the ground to investigate the temporal and spatial variability of \(C_T^2\) measurements using a LAS, and confirmed the results of Cheinet and Siebesma (2007). Maronga (2014) calculated the MOST relationships for structure parameters from a set of LES runs ranging from near-neutral to the free convective ABL above a homogeneous surface and found that universal MOST functions exist for \(C_T^2\) that were well within the range of proposed functions. Maronga (2014) also showed that entrainment of dry air at the top of the mixed layer explains the dissimilarity between the turbulent transport of heat and moisture and leads to non-universal similarity functions for \(C_q^2\). Up to now, all LES studies have been performed for idealized conditions with homogeneous surfaces and under quasi-stationary conditions. The question whether MOST for structure parameters is also valid over heterogeneous terrain with different surface patches that provide different characteristics concerning surface fluxes and roughness has not been studied by means of LES so far.

In the present case study the CBL over the eastern part of the LITFASS-2003 area (see Fig. 1a, b, dominated by farmland) is simulated by high-resolution LES runs using the model PALM. The structure parameters are for the first time derived from surface-layer turbulence resolving LES over such an irregular surface heterogeneity, and they are compared with LAS measurements performed during the LITFASS-2003 experiment. The concept of a blending height for structure parameters is studied using a lagged two-dimensional correlation analysis method. Moreover, we discuss if MOST can be applied over such heterogeneous landscapes, and explore possible implications of surface heterogeneity on LAS observations.

The paper is organized as follows: Sect. 2 describes the LES model PALM, model set-up as well as data processing. Section 3 outlines the derivation of structure parameters from LES data, and a brief introduction to the MOST relationships for structure parameters is given. Simulation results are presented in Sect. 4, and Sect. 5 gives a summary.

## 2 LES Model, Set-up and Case Description

### 2.1 LES Model

The LES model PALM (revision 1105) (see e.g. Raasch and Schröter 2001; Riechelmann et al. 2012) has been used in the present study. It has been recently applied to study different flow regimes in the CBL over homogeneous (e.g. Raasch and Franke 2011; Maronga et al. 2013b) and heterogeneous terrain (e.g. Maronga and Raasch 2013a; Sühring and Raasch 2013). All simulations were carried out using cyclic lateral boundary conditions. The grid was stretched in the vertical direction well above the top of the ABL in the free atmosphere to save computational time. MOST was applied as a surface boundary condition locally between the surface and the first computational grid level (“local similarity model”, see also Peltier and Wyngaard 1995), including the calculation of the local friction velocity \(u_*\). A 1.5-order flux-gradient subgrid closure scheme after Deardorff (1980) was applied (see also Moeng and Wyngaard 1988; Saiki et al. 2000), which requires the solution of an additional prognostic equation for the subgrid-scale (SGS) turbulent kinetic energy. It should be noted that previous LES studies suggested that such static SGS models can lead to deficiencies near the transition boundaries of surface heterogeneity (e.g. Stoll and Porté-Agel 2006, 2008). We tried to minimize such effects by using a high grid resolution that allows for resolving the bulk part of the turbulence in the atmospheric surface layer (see below). A fifth-order advection scheme of Wicker and Skamarock (2002) and a third-order Runge–Kutta timestep scheme were used (Williamson 1980). A one-dimensional version of the model with fully parametrized turbulence, using a mixing-length approach after Blackadar (1997) and stationary temperature and humidity profiles, was used for precursor simulations to generate steady-state wind profiles as initialization for the LES runs.

### 2.2 Model Set-up and Case Description

^{1}dataset with a resolution of 100 m. The size of the area was about 20 km \(\times \) 20 km, containing different surface patches of forest, lakes and farmland, as was used in the previous LES studies. These studies, however, had relatively coarse grid resolutions of 40–100 m in all spatial directions. In order to resolve turbulence in the surface layer, a much higher grid resolution was required in the present study. Due to limited computing resources it was necessary to limit the model domain to an area of 5.3 km \(\times \) 5.3 km with \(1600 \times 1600\) grid points in the horizontal directions, resulting in a horizontal grid resolution of 3.3125 m (Fig. 1a, black box and Fig. 1b). In early test runs we found that the vertical profiles of \(C_T^2\) and \(C_q^2\) from different calculation methods (see below) converge at a grid resolution of 5 m, so that the grid resolution of the present study is an adequate choice for studying structure parameters. From the previous LES studies we found that the mean boundary-layer depth \(z_{i}\) reached about 1.8 km at 1400 UTC, and hence used a constant vertical grid resolution of 2 m up to a height level of 2 km. Above, the grid was vertically stretched by a factor of 1.02 for each plane so that the vertical model extension was about 3.7 km. Moreover, the simulation time was reduced to that part of the diurnal cycle from 0500 UTC to 1300 UTC. In this way one simulation still required about 120 h of real time on 4096 Intel Xeon Gainestown processors (2.93 GHz) on an SGI Altix ICE 8200 Plus cluster. In order to quantify effects of the surface heterogeneity, a second run with a homogeneous surface but temporally-varying surface forcing (hereafter case HOM, see Fig. 2 and below) was performed. A third simulation, hereafter referred to as LIT2E_W, was performed for sensitivity tests (see Sect. 4.4) where the model domain was shifted 1.5 km westwards (see Fig. 1a, white dashed box).

In the course of the LITFASS-2003 experiment, LAS observations were made along a 4.8-km long path at an effective height of 43 m above the ground (following topography), extending from the boundary-layer field site (GM) of the German Meteorological Service (DWD) near Falkenberg to the Lindenberg Meteorological Observatory (MOL) (see Beyrich and Mengelkamp 2006). We located the model domain around the LAS path as shown in Fig. 1b with the start and finish of the VLAS measurement path in the LES being located on grassland. The height of the VLAS was 41 m and thus equal to the height of the LAS (see above). The LAS path was nearly aligned in north-south direction. For simplification we decided to align the VLAS path exactly along the \(y\)-direction in the model, which should not introduce a significant error since the heterogeneity was still prescribed on a 100 m raster. A footprint analysis was carried out afterwards for LIT2E to retrospectively ensure that the chosen model domain was large enough to cover the footprint source area of the VLAS. The results of this footprint analysis will be discussed in Sect. 3.4. Figure 1b shows that the surface in the model domain mainly consists of different agricultural fields, where triticale makes up 39 % of the total area. The remaining area is covered by maize (17 %), rape (16 %), forest (11 %, including settlements, see Maronga and Raasch 2013a), grassland (10 %) and barley (7 %). No open water surfaces occur in the area chosen for this study. Therewith, the distribution of surface types is significantly different from the previous LES runs for the LITFASS area and hence also the area-averaged surface fluxes differ.

Close scrutiny of the LITFASS-2003 data, however, unfortunately revealed that we did not use the composite fluxes derived from all available energy balance stations over the same surface type by Beyrich et al. (2006) (see their Fig. 5), but rather the eddy-covariance fluxes from individual energy balance stations. These forcing data provide too strong a forcing compared to the composite fluxes and have been used for several years during LES of the LITFASS area (Uhlenbrock et al. 2004; Maronga and Raasch 2013a; Sühring and Raasch 2013). Depending on the land-use type and time of the day, the deviation between the composite fluxes and the used fluxes was up to 104 W m\(^{-2}\) for sensible heat (forest) and 69 W m\(^{-2}\) for latent heat (barley). The area-averaged surface flux displayed a difference in the course of the day of 10–50 W m\(^{-2}\) for both sensible and latent heat. The inconsistency in the data had not been discovered before, mainly because the simulated ABL still compared well with observations. We discuss the effect of this too strong a forcing later in Sect. 4. Due to high computational demands it was neither feasible to perform more than the present case study, nor was it possible to repeat a simulation with an improved set-up. In the present study this issue restricts the direct validation of VLAS measurements with the in situ LAS data. Moreover, we might expect that mean profiles e.g. of temperature and humidity, as well as those of structure parameters, can be altered slightly. Fortunately, the other analyses included in this and the former studies should not be significantly affected by the inaccurate fluxes.

Maronga and Raasch (2013a) showed that secondary circulations developed over the LITFASS area, depending on the geostrophic wind, and that a sufficiently large upwind buffer zone is required in order to resolve them. However, these secondary circulations only develop if the heterogeneity scales are at least of size \(z_{i}\), and if there are strong constrasts in the surface characteristics (as between forest and water) (e.g. Shen and Leclerc 1995). Over the eastern part of the LITFASS area (dominated by farmland patches with a size in the order of 1 km) the heterogeneity scales are usually too small and the heterogeneity amplitude is too low to trigger secondary circulations (see Maronga and Raasch 2013a, their Fig. 5a). Moreover, Maronga and Raasch (2013a) pointed out that secondary circulations are generally weak close to the surface. Effects of such circulations, such as a modification of the turbulent vertical fluxes of heat and moisture, are thus expected to be irrelevant for the present study. Local small-scale effects of the surface heterogeneity, however, remain and will be most prominent in the lower ABL.

The ABL depth was calculated from the \(\theta \) profile following Sullivan et al. (1998), and was found to increase in the course of the simulation up to 1.5 km. The top of the surface layer \(z_\mathrm{SL}\) was defined as \(z_\mathrm{SL} = 0.1 z_{i}\), following e.g. Brasseur and Wei (2010), noting that \(z_\mathrm{SL}\) indicates the height up to which MOST roughly should be valid. In order to study the MOST relationships for \(C_T^2\) and \(C_q^2\) we thus only used data from height levels below \(z_\mathrm{SL}\). Moreover, as pointed out by Khanna and Brasseur (1997), the lowest levels are always affected by the SGS model and thus cannot resemble the surface-layer dynamics correctly. We hence excluded the lowest seven grid points from the analysis of the MOST relationships as suggested by Maronga (2014). Due to the very high vertical resolution of 2 m used in our simulations, the entire surface layer is represented by more than \(40\) grid points (except the early morning hours before the morning inversion is eroded, see below), so that omitting the lowest levels does not restrict our analysis of the surface layer much.

## 3 Theory and Methods

### 3.1 Structure Parameters

*spectral method*) at a given height by

*dissipation method*and Eq. 2 then yields

### 3.2 Monin–Obukhov Similarity Relationships

*local free convection*(LFC) (e.g. Wyngaard et al. 1971a), and the dimensionless structure parameters are given by

### 3.3 LES Data Processing

\(C_T^2\) and \(C_q^2\) were derived from both the spectral and the dissipation methods (see Eqs. 1 and 3) as described in Maronga et al. (2013b). Cheinet and Siebesma (2009) and Maronga et al. (2013b) both found that the dissipation method yields structure parameters that are too small in magnitude (compared to semi-empirical profiles), whereas the spectral method gives reliable estimates of \(C_T^2\) and \(C_q^2\) in the study of Maronga et al. (2013b). The simulation data in the latter study, as well as the data in the present study, suggest a constant correction factor for the dissipation method of 1.7–1.9. In the present study we used the data from the spectral method to determine this correction factor, giving a value of 1.85 (evaluated at the height of the VLAS, see Appendix). We decided to use the corrected structure parameters from the dissipation method for our analysis, because it gives local estimates, which allows investigation of possible local effects of surface heterogeneity on the structure parameters and VLAS measurements.

Here, we use these VLAS data along the path from the GM site (transmitter) to the MOL site (receiver) (see Fig. 1b) for comparison with LAS data. We employed the dissipation method and calculated the path-weighted average of the structure parameters along horizontal paths in the LES model as suggested by Maronga et al. (2013b) (see their Fig. 6 and Appendix). They argued that, on the one hand, the scintillations seen by LAS are mainly determined by fluctuations at the scale of the aperture (here 0.15 m), whereas smaller scales are averaged out and larger scales lead to variability of the scintillations. On the other hand, the nominal truncation size (or nominal aperture of the VLAS) in the LES is \(\varDelta \approx 2.7\) m (the actual truncation happens at even larger scales up to \(6\varDelta \)), so that part of the variability in the structure parameters seen by the LAS is unresolved by the LES, and thus in the VLAS data.

### 3.4 Footprint Analysis

In our simulations, cyclic lateral boundary conditions were used, implying that the prescribed surface heterogeneity was repeated periodically. This does not occur in reality. In the vicinity of the horizontal domain boundaries the flow is hence exposed to an erroneous surface forcing. It is thus essential that a sufficiently large buffer zone is prescribed so that the VLAS signal is not affected by the cyclic boundary conditions. The surface area was chosen in such a way that the VLAS path was located close to the western boundary of the model. A footprint analysis for the VLAS path in the LITFASS area on May 30 2003 was conducted using the LES data as input for the analytical footprint model of Kormann and Meixner (2001), designed for a point source. We hence applied a convolution with the VLAS path-weighting function to obtain the LAS footprint as suggested by Meijninger et al. (2002b). As easterly winds were prescribed in the model, the footprint of the VLAS was located in the upstream direction; footprints for selected times are shown in Fig. 3. As expected the footprint source area was located in the area east of the path, with 90 % of the footprint lying within a distance of 200 m from the scintillometer path (see Fig. 1b). Only small variability of the footprints in time is observed that can be ascribed to the strongly convective character of the ABL. The fetch of 4 km was thus much larger than the required upstream domain according to the footprint analysis. Therefore, it is possible to compare VLAS measurements with the in situ LAS data obtained during LITFASS-2003. A typical contribution of the land use to the footprint was (e.g. at 1100 UTC): 56% triticale, 23% rape, 6% maize, 4% barley, 4% forest and 7% grassland. If no blending height exists below the VLAS, this composition suggests that the VLAS predominantly samples turbulence originating from triticale and rape. Figure 2a shows the flux averaged over the VLAS footprint area, derived from the weighted contribution of the individual land-use classes to the footprint area. It can be seen that the VLAS footprint is coincidentally in remarkable agreement with the area-averaged surface sensible heat flux used for case HOM throughout the day. A similar behaviour can be observed when considering the surface latent heat flux (see Fig. 2b).

## 4 Results

### 4.1 Mean Profiles

The profiles of \(\langle \theta \rangle \) reveal a warming of the ABL with a maximum mixed-layer temperature close to 297 K at 1300 UTC; \(\langle q \rangle \) is characterized by a continuous decrease in time after 0700 UTC, reaching \(4\hbox { g kg}^{-1}\) at 1300 UTC. One important process relates to the entrainment of dry air at the top of the CBL, which decreases the CBL humidity. This is shown in Fig. 4f: the moisture flux at the top of the CBL is a factor of 3–4 higher than the respective surface flux all day long (entrainment flux ratio of 3–4), leading to a net moisture loss in the CBL. In contrast, the entrainment flux ratio is only around \(-0.2\) for the kinematic heat flux (see Fig. 4e), showing that the relative importance of the surface flux is high for sensible heat and low for latent heat. Moreover, the available water vapour is reduced as the CBL grows in the course of the day. As the sum of both processes is not balanced by the released water vapour at the surface, the CBL dries out in the course of the day. This is often observed in the CBL, was a frequent feature during LITFASS-2003, and was also previously shown by Couvreux et al. (2005, 2007). Both the profiles of \(\langle \theta \rangle \) and \(\langle q\rangle \) are comparable with the previous LES runs for the LITFASS area (see Maronga and Raasch 2013a). It can also be seen from Fig. 4a, b that the mean profiles, and in particular the height of the temperature inversion (and thus \(z_{i}\)), are not modified significantly by the surface heterogeneity. While Maronga and Raasch (2013a) showed that secondary circulations developed that spanned the entire domain and affected the horizontally-averaged \(z_{i}\), the surface heterogeneity in the present study does not show such effects, most likely due to the fact that the heterogeneity scales and amplitudes are rather small (compared to the amplitude between forest and water patches that led to strong secondary circulations in Maronga and Raasch (2013a)).

Figure 4c, d shows the mean profiles of \(C_T^2\) and \(C_q^2\), respectively. Both \(\langle C_T^2\rangle \) and \(\langle C_q^2\rangle \) display a maximum near the surface, caused by the generation of small-scale turbulence due to rising thermals and wind shear. \(\langle C_T^2\rangle \) decreases in the mixed layer up to levels where entrainment processes become important, leading to temperature fluctuations at the interface between the mixed layer and free atmosphere and thus a secondary peak in \(\langle C_T^2\rangle \). The shape of the \(\langle C_T^2\rangle \) profiles is in agreement with previous LES studies (Peltier and Wyngaard 1995; Cheinet and Siebesma 2009; Maronga et al. 2013b) and semi-empirical profiles after Kaimal et al. (1976). The shape of \(\langle C_q^2\rangle \) is in agreement with previous studies as well (Fairall 1987; Peltier and Wyngaard 1995; Cheinet and Cumin 2011; Maronga et al. 2013b). \(\langle C_q^2\rangle \) increases throughout the mixed layer, suggesting that fluctuations of humidity in the CBL are controlled by the entrainment of dry air at the top of the mixed layer. This is supported by the entrainment flux ratio of 3–4 (see Fig. 4f), and in agreement with observations (e.g. Druilhet et al. 1983). Secondary peaks of \(\langle C_T^2\rangle \) and \(\langle C_q^2\rangle \) that can be observed at 800 m (at 0700 UTC) can be ascribed to turbulent fluctuations at the top of the residual layer during the morning transition. Peaks for \(\langle C_q^2\rangle \) around 2000 m are caused by the humidity jump from 1,900–2,000 m. For both \(\langle C_T^2\rangle \) and \(\langle C_q^2\rangle \) some differences between cases LIT2E and HOM are observed, even though they appear to be marginal, particularly for \(\langle C_T^2\rangle \). We consider whether these differences affect the structure parameters at the height of a LAS system, and the similarity relationships, in the following Sect. 4.2.

### 4.2 Monin–Obkuhov Similarity Relationships

Figure 5b shows that the dimensionless \(\langle C_q^2 \rangle \) does not collapse onto a single curve and it is evident that the data do not agree with the proposed function \(f_q\), and that there are only a few data points that can be considered to follow MOST. These are located very close to the surface (not shown). It is also obvious that no fitting function can be derived from the present data, and we must conclude that \(C_q^2\) generally does not follow MOST in the present study. This finding can be again attributed to the fact that the ratio of the entrainment flux to the surface flux of latent heat, for case LIT2E was around 3–4 (see Fig. 4f). As was shown by Maronga (2014), \(C_q^2\) no longer follows MOST if this ratio \(>\)1. The entrainment of dry air is then the dominant process that affects \(C_q^2\) in the surface layer, in agreement with the assumption of Li et al. (2012). The present study supports this finding so that it is impossible to derive a value for \(A_q\). As a direct consequence, MWS measurements should not be used to derive the surface flux of latent heat when the entrainment of dry air becomes too large.

### 4.3 Blending Height

MOST theoretically is only applied to derive surface fluxes from elevated LAS measurements if a blending height for structure parameters lies somewhere below the scintillometer path. Otherwise, the scintillometer “sees” the fluxes from different surface patches and a footprint analysis must be employed to relate the LAS signal to the surface fluxes of the different surface types (Meijninger et al. 2002b, 2006; Beyrich et al. 2012). Furthermore, in the latter case, MOST is violated, and likely will introduce additional errors. Then, the derived surface fluxes can at best be regarded only as representative for the footprint area. The spatial distribution of surface types and thus surface fluxes within this footprint area, however, might be completely different from the area-averaged fluxes at a larger scale.

### 4.4 Spatial and Temporal Variability of Structure Parameters and VLAS Measurements

In Sects. 4.1 and 4.3 we showed that surface heterogeneity affects the local structure parameters and their similarity relationships. Here we discuss the impact on LAS measurements. Figure 11 shows the time series of \(\langle C_T^2\rangle \) and \(\langle C_q^2\rangle \) at the VLAS (43 m) height. It appears that \(\langle C_T^2\rangle \) reflects the evolution of \(\langle \overline{w'\theta '}_0\rangle \) very well (cf. Fig. 2a), and it is plausible to consider MOST as an appropriate framework to derive the surface sensible heat flux from measurements of \(C_T^2\). Moreover, it is also clear that the value of \(\langle C_T^2\rangle \) is slightly higher for case LIT2E than for case HOM, suggesting that the surface heterogeneity generates additional temperature fluctuations that lead to a higher \(\langle C_T^2\rangle \).

Note that this good agreement between the VLAS measurement and the area-averaged \(C_T^2\) over heterogeneous terrain is no universal feature, because it is based on one single case study, and we suppose that this is due to the fact that the footprint of the VLAS is coincidentally representative of the entire area as shown in Sect. 3.4 (see Fig. 2a). In order to prove this, we employed case LIT2E_W, where the model domain was shifted in a westward direction so that a large patch of forest is captured by the model. The slightly different surface heterogeneity in case LIT2E_W leads to a significantly higher area-averaged surface sensible heat flux. If our hypothesis holds that the VLAS is rather representative for its footprint area, then the VLAS measurement should be comparable to that in case LIT2E because the footprint area does not change and is located over the farmland area to the east of the VLAS path (see also Meijninger et al. 2006). In case LIT2E_W, however, \(\langle C_T^2\rangle \) should deviate significantly from case LIT2E as the surface heterogeneity distribution and thus the area-averaged surface flux has changed. Figure 12a shows \(\overline{\langle C_T^2\rangle }\) as well as \(\overline{[C_T^2]}\) for both cases LIT2E and LIT2E_W, and it is obvious that \(\overline{\langle C_T^2\rangle }\) is indeed significantly higher in case LIT2E_W, as a result of the increased area-averaged surface sensible heat flux. By the same token it is clear that \(\overline{[C_T^2]}\) shows comparable values for both cases, which clearly supports our hypothesis.

The comparison of \(\overline{[C_T^2]}\) with \(\overline{[C_T^2]}_\mathrm{LAS}\) shows that the expected increase of \(C_T^2\) in the course of the day is captured by the LES (see Fig. 12a). Nevertheless it is also clear that \(\overline{[C_T^2]}_\mathrm{LAS}\) is about a factor of two lower than the \(C_T^2\) derived from the LES. The surface flux derived from the LAS (using MOST, not shown) was around \(200\hbox { W m}^{-2}\) around noon, whereas the prescribed area-averaged flux in the LES was close to \(250\hbox { W m}^{-2}\) (see Fig. 2a). Meijninger et al. (2006) showed that the LAS data generally compare well with the aggregated fluxes based on several eddy-covariance stations. This discrepancy between LES results and LAS measurements is to a large part a direct consequence of inappropriate flux forcing data, as discussed in Sect. 2. The fluxes used in the present study are thus about 50 W m\(^{-2}\) too high. The difference between the area-averaged surface fluxes used in our LES and the composite fluxes were found to account for 70 % of the difference between VLAS and LAS data.

Path-averaged values and standard deviations of \(C_T^2\) along the VLAS path from the data shown in Fig. 13

Case | Averaging interval (min) | |||
---|---|---|---|---|

0 | 30 | 60 | 120 | |

LIT2E | ||||

Path-average | 0.0222 | 0.0238 | 0.0229 | 0.0210 |

Standard deviation | 0.0647 | 0.0135 | 0.0091 | 0.0068 |

HOM | ||||

Path-average | 0.0171 | 0.0222 | 0.0214 | 0.0200 |

Standard deviation | 0.0504 | 0.0068 | 0.0047 | 0.0030 |

Our results demonstrate not only that the VLAS measurements are made well below the blending height (if this exists at all), but also that signals from the local surface fluxes can be seen in the VLAS signal if an adequate time average (here 30 min) is applied. The local fluxes that have been derived by MOST are then in fairly good agreement with the prescribed surface values. Hence we conclude that MOST can also be a useful framework for deriving local surface fluxes from elevated LAS measurements.

As discussed above, the good agreement between VLAS measurements and area-averaged \(C_T^2\) is a fortunate coincidence to be ascribed to the fact that the heterogeneity in the footprint area is representative for the entire model domain (see Fig. 2a). This should be also considered when comparing the derived surface fluxes from the VLAS measurements. A comparison of the VLAS measurements from case LIT2E_W (see Fig.14a,c) reveals that for MOST the surface flux is slightly underestimated in case LIT2E_W, which is consistent with the higher area-averaged surface flux in this case. This underestimation, however, only leads to a very small increase in the error of 0.2 %. For LFC scaling it can be observed that the error in the derived surface flux is significantly reduced in case LIT2E_W, compared to LIT2E, which is counterintuitive. However, as discussed above, \(\overline{[H]}_\mathrm{LFC}\) was generally found to overestimate \(\overline{\langle H \rangle }_\mathrm{LFC}\) for \(\overline{\langle H \rangle } > 120\) W m\(^{-2}\) (see Fig. 14d). The increased value of \(\overline{\langle H \rangle }_\mathrm{LFC}\) in case LIT2E_W thus decreases the relative error to 3 %. In conclusion, Fig. 14 shows that, even though the VLAS measurements give rather reliable estimates of the footprint-averaged surface fluxes, they are also good estimates for the area-averaged surface fluxes in the case of moderate heterogeneity, with deviations from the area-averaged surface sensible heat flux in the order of only 5 %.

## 5 Summary

A case study for May 30 2003 of the LITFASS-2003 experiment was conducted using a set of large-eddy simulations (LES) that allow surface-layer turbulence to be resolved. The CBL over the heterogeneous LITFASS terrain was simulated from the early morning until early afternoon using prescribed surface fluxes of sensible and latent heat derived from eddy-covariance measurements over the major land-use classes. The model domain used was 5.3 km \(\times \) 5.3 km in area and represents the farmland portion of the LITFASS area. The data were compared with a homogeneous control run using spatially-averaged surface fluxes. \(C_T^2\) and \(C_q^2\) were derived from the simulation data over heterogeneous terrain and compared to data from the homogeneous simulation as well as with LAS data collected during the LITFASS-2003 experiment. Particular attention was given to the spatial distribution of \(C_T^2\) and \(C_q^2\) in the surface layer and to the MOST/LFC relationships, relating the structure parameters to the surface sensible and latent heat fluxes, respectively.

We found for the heterogeneous simulation that, compared to the homogeneous simulation, the mean \(C_T^2\) and \(C_q^2\) in the surface layer are about 4 and 12 % higher, respectively, and we showed that this difference slightly modifies the MOST/LFC relationships for \(C_T^2\). However, it transpires that the induced error did not exceed 1.5 % for both MOST and LFC scaling and is thus very small. It was also discussed that the VLAS-derived fluxes tend to overestimate the prescribed surface fluxes due to the non-linear relationship between structure parameters and fluxes, in agreement with previous scintillometer studies (Lagouarde et al. 2002; Meijninger et al. 2002b, 2006). We found that no MOST/LFC relationship can be obtained for \(C_q^2\) from the LES data as the entrainment of dry air at the top of the mixed layer was the dominant process generating humidity fluctuations. These fluctuations, in turn, affect the surface-layer dynamics in such a way that the surface latent heat flux is largely decoupled from the distribution of \(C_q^2\) in both the heterogeneous and homogeneous simulations.

A two-dimensional correlation analysis was used in order to investigate the blending height concept for structure parameters, and we showed that the structure parameters at a height of a few tens of m exhibit significant signals of the prescribed surface heterogeneity all day long. This result was more pronounced for \(C_T^2\) since the entrainment of dry air affected the humidity fluctuations in the surface layer and thus decreased the correlation between the surface latent heat flux and \(C_q^2\). However, signals of the surface heterogeneity were also visible in the \(C_q^2\) pattern. Consequently, horizontal homogeneity of turbulence as required by MOST cannot generally be expected, at least under highly convective conditions. As a result we conclude that, under such conditions, LAS observations are not representative of an area of several km\(^{2}\) (in the order of NWP grid boxes, e.g. 100 km\(^2\)), but only for their footprint area of very limited size (here in the order of 1 km\(^2\)). As this footprint might be composed of different surface patches with differing surface properties, the obtained surface flux from an LAS system does not necessarily have to be representative of a larger area. Our results are, of course, based on a single case where the LAS footprint was much smaller than would be the case for synoptic conditions with higher geostrophic winds and lower buoyant forcing at the surface. For the simulated case the area-averaged surface flux was in remarkable agreement with the footprint-integrated surface flux.

Simulated VLAS measurements in the LES runs were performed along a 4.8-km long path, in agreement with the LAS set-up during LITFASS-2003. A direct comparison showed that the VLAS measurements did overestimate the path-weighted \(C_T^2\) as observed by the LAS. A revision of the prescribed surface-flux data revealed that the flux forcing data used previously in several LES studies for the LITFASS-2003 experiment were too high in comparison with the processed composite eddy-covariance data from all available energy balance stations given in Beyrich et al. (2006). However, we believe this basically affects the direct evaluation of VLAS with the LAS, while all generalized results from this and former studies remain valid.

Moreover, we employed the VLAS measurements in order to investigate the effect of surface heterogeneity on LAS observations. It was found that sufficient time-averaging (here 39 min) is required to obtain representative fluxes and that signals of the underlying surface heterogeneity affect the VLAS observations. Whereas the VLAS measurement converged to the area-averaged value over homogeneous terrain, the VLAS observations in the heterogeneous simulation were found to retain a rather local character and showed a mean deviation of 5 % from the area average. This was confirmed by a sensitivity simulation with a different surface heterogeneity distribution. Due to the fact that the surface fluxes changed in the course of the day, such that stationary conditions were not reached in the simulation, the averaging interval was limited to 30 min. Therefore, 50 % of the variability of \(C_T^2\) along the VLAS path was caused by the remaining noise from random turbulent fluctuations and not related to surface heterogeneity. This shows that it is difficult for in situ measurements, such as low-level aircraft flights, to capture and quantify effects of the surface heterogeneity in the \(C_T^2\) signal.

Even though the blending-height analysis clearly suggested that the VLAS measurements cannot be considered to yield representative estimates of the area-averaged surface sensible heat flux, it would be necessary to perform a more sophisticated footprint analysis for the LES to confirm that the VLAS “sees” only the surface fluxes in its footprint area. In a future study we will employ a Lagrangian footprint model, embedded into the LES (see Steinfeld et al. 2008), over the heterogeneous LITFASS-2003 terrain in order to determine more elaborate footprints that take into account the surface heterogeneity.

## Footnotes

- 1.
Abbr. “Coordinated Information on the European Environment”, published by the European Environment Agency (http://www.eea.europa.eu/publications/COR0-landcover).

## Notes

### Acknowledgments

This study was supported by the German Research Foundation (DFG) under grants BE 2044/3-3, RA 617/20-1 and RA 617/20-3. All simulations were performed on the SGI Altix ICE at The North-German Supercomputing Alliance (HLRN), Hannover/ Berlin. NCL (The NCAR Command Language (Version 5.2.1) [Software]. (2010). Boulder, CO: UCAR/NCAR/CISL/VETS. 10.5065/D6WD3XH5) has been used for data analysis and visualization. We appreciate the four anonymous reviewers for their constructive and valuable comments that helped to improve the manuscript.

### References

- Andreas EL (1988) Estimating \(C_n^2\) over snow and sea ice from meteorological data. J Opt Soc Am 5:481–495CrossRefGoogle Scholar
- Andreas EL (1991) Using scintillation at two wavelengths to measure path-averaged heat fluxes in free convection. Boundary-Layer Meteorol 54:167–182CrossRefGoogle Scholar
- 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 55:2666–2689CrossRefGoogle Scholar
- Bange J, Spieß T et al (2006) Turbulent fluxes from Helipod flights above quasi-homogeneous patches within the LITFASS area. Boundary-Layer Meteorol 121:127–151CrossRefGoogle Scholar
- Beyrich F, Mengelkamp HT (2006) Evaporation over a heterogeneous land surface: EVA\_GRIPS and the LITFASS-2003 experiment: an overview. Boundary-Layer Meteorol 121:5–32CrossRefGoogle Scholar
- Beyrich F, Kouznetsov RD, Leps JP, Lüdi A, Meijninger WML, Weisensee U (2005) Structure parameters for temperature and humidity from simultaneous eddy-covariance and scintillometer measurements. Meteorol Z 14:641–649CrossRefGoogle Scholar
- Beyrich F, Leps JP, Mauder M, Bange J, Foken T, Huneke S, Lohse H, Lüdi A, Meijninger WML, Mirnonov D, Weisensee U, Zettel P (2006) Area-averaged surface fluxes over the LITFASS region based on eddy-covariance measurements. Boundary-Layer Meteorol 121:33–65CrossRefGoogle Scholar
- Beyrich F, Bange J, Hartogensis OK, Raasch S, Braam M, van Dinther D, Gräf D, van Kesteren B, van den Kroonenberg AC, Maronga B, Martin S, Moene AF (2012) Towards a validation of scintillometer measurements: the LITFASS-2009 experiment. Boundary-Layer Meteorol 144:83–112CrossRefGoogle Scholar
- Blackadar AK (1997) Turbulence and diffusion in the atmosphere. Springer, Berlin 185 ppCrossRefGoogle Scholar
- Braam M (2014) Aspects of atmospheric turbulence related to scintillometry. PhD Thesis, Wageningen University, Wageningen. ISBN 978-94-6173-938-4Google Scholar
- Braam M, Bosveld FC, Moene AF (2012) On the Monin–Obkuhov scaling in and above the atmospheric surface layer: the complexities of elevated scintillometer measurements. Boundary-Layer Meteorol 144:157–177. doi:10.1007/s10546-012-9716-7 CrossRefGoogle Scholar
- Brasseur JG, Wei T (2010) Designing large-eddy simulation of the turbulent boundary layer to capture law-of-the-wall scaling. Phys Fluids 22:021303CrossRefGoogle Scholar
- Cheinet S, Cumin P (2011) Local structure parameters of temperature and humidity in the entrainment-drying convective boundary layer: A large-eddy simulation analysis. J Appl Meteorol 50:472–481CrossRefGoogle Scholar
- Cheinet S, Siebesma AP (2007) The impact of boundary layer turbulence on optical propagation. In: Proceedings of SPIE 6747, Optics in atmospheric propagation and adaptive systems X, Florence, p 67470A. doi:10.1117/12.741433
- Cheinet S, Siebesma AP (2009) Variability of local structure parameters in the convective boundary layer. J Atmos Sci 66:1002–1017CrossRefGoogle Scholar
- Couvreux F, Guichard F, Redelsperger JL, Kiemle C, Masson V, Lafore JP, Flamant C (2005) Water-vapour variability within a convective boundary-layer assessed by large-eddy simulations and ihop\_2002 observations. Q J R Meteorol Soc 131:2665–2693. doi:10.1256/qj.04.167 CrossRefGoogle Scholar
- Couvreux F, Guichard F, Masson V, Redelsperger JL (2007) Negative water vapour skewness and dry tongues in the convective boundary layer: observations and large-eddy simulation budget analysis. Boundary-Layer Meteorol 123:269–294. doi:10.1007/s10546-006-9140-y CrossRefGoogle Scholar
- De Bruin HAR, Kohsiek W, van den Hurk BJJM (1993) A verification of some methods to determine the fluxes of momentum, sensible heat and water vapour using standard deviation and structure parameter of scalar meterorological quantities. Boundary-Layer Meteorol 63:231–257CrossRefGoogle Scholar
- De Bruin HAR, van den Hurk BJJM, Kohsiek W (1995) The scintillation method tested over a dry vineyard area. Boundary-Layer Meteorol 76:25–40CrossRefGoogle Scholar
- Deardorff JW (1980) Stratocumulus-capped mixed layers derived from a three-dimensional model. Boundary-Layer Meteorol 18:495–527CrossRefGoogle Scholar
- Druilhet A, Frangi JP, Guedalia D, Fonton J (1983) Experimental studies of the turbulence structure parameters of the convective boundary layer. J Clim Appl Meteorol 22:594–608CrossRefGoogle Scholar
- Evans JG, McNeil DD, Finch JW, Murray T, Harding RJ, Ward HC, Verhoef A (2012) Determination of turbulent heat fluxes using a large aperture scintillometer over undulating mixed agricultural terrain. Agric For Meteorol 166–167:221–233. doi:10.1016/j.agrformet.2012.07.010 CrossRefGoogle Scholar
- Fairall CW (1987) A top-down and bottom-up diffusion model of \(C_T^2\) and \(C_Q^2\) in the entraining convective boundary layer. J Atmos Sci 44:1009–1017CrossRefGoogle Scholar
- Gottschalk L, Batchvarova E, Gryning SE, Lindroth A, Melas D, Motovilov Y, Frech M, Heikinheimo M, Samuelsson P, Grelle A, Persson T (1999) Scale aggregation comparison of flux estimates from NOPEX. Agric For Meteorol 98–99:103–119CrossRefGoogle Scholar
- Hill RJ (1978) Spectra of fluctuations in refractivity, humidity, and the temperature-humidity cospectrum in the inertial and dissipation range. Radio Sci 13:953–961CrossRefGoogle Scholar
- Hill RJ (1989) Implications of Monin–Obukhov similarity theory for scalar quantities. J Atmos Sci 46:2236–2244CrossRefGoogle Scholar
- Hill RJ, Ochs GR, Wilson JJ (1992) Measuring surface-layer fluxes of heat and momentum using optical scintillation. Boundary-Layer Meteorol 58:391–408CrossRefGoogle Scholar
- Kaimal JC, Wyngaard JC, Haugen DA, Coté OR, Izumi Y, Caughey SJ, Readings CJ (1976) Turbulence structure in the convective boundary layer. J Atmos Sci 33:2152–2169CrossRefGoogle Scholar
- Khanna S, Brasseur JG (1997) Analysis of Monin–Obukhov similarity from large-eddy simulations. J Fluid Mech 345:251–286CrossRefGoogle Scholar
- Kohsiek W, Herben M (1983) Evaporation derived from optical and radiowave scintillation. Appl Opt 22:2566–2570CrossRefGoogle Scholar
- Kohsiek W, Meijninger WML, Moene AF, Heusinkveld BG, Hartogensis OK, Hillen WCAM, De Bruin HAR (2002) An extra large aperture scintillometer for long range applications. Boundary-Layer Meteorol 105:119–127CrossRefGoogle Scholar
- Kormann R, Meixner FX (2001) An analytical footprint model for non-neutral stratification. Boundary-Layer Meteorol 99:207–224CrossRefGoogle Scholar
- Kunkel KE, Walters DL, Ely GA (1981) Behavior of the temperature structure parameter in a desert basin. J Appl Meteorol 20:130–136CrossRefGoogle Scholar
- Lagouarde JP, Bonnefond JM, Kerr YH, McAneney KJ, Irvine M (2002) Integrated sensible heat flux measurements of a two-surface composite landscape using scintillometry. Boundary-Layer Meteorol 105:5–35CrossRefGoogle Scholar
- Lüdi A, Beyrich F, Mätzler C (2005) Determination of the turbulent temperature-humidity correlation from scintillometric measurements. Boundary-Layer Meteorol 117:525–550CrossRefGoogle Scholar
- Lee X, Massman W, Law B (2004) Handbook of micrometeorology. Kluwer, Dordrecht 250 ppGoogle Scholar
- Li D, Bou-Zeid E, De Bruin HAR (2012) Monin–Obukhov similarity functions for the structure parameters of temperature and humidity. Boundary-Layer Meteorol 145:45–67. doi:10.1007/s10546-011-9660-y CrossRefGoogle Scholar
- Lohou F, Druilhet A, Campistron B (1998) Spatial and temporal characteristics of horizontal rolls and cells in the atmospheric boundary layer based on radar and in situ observations. Boundary-Layer Meteorol 89:407–444CrossRefGoogle Scholar
- Lohou F, Druilhet A, Campistron B, Redelsperger K, Said F (2000) Numerical study on the impact of coherent structures on vertical transfers in the atmospheric boundary layer. Boundary-Layer Meteorol 97:361–383CrossRefGoogle Scholar
- Mahrt L (2000) Surface heterogeneity and vertical structure of the boundary layer. Boundary-Layer Meteorol 96:33–62CrossRefGoogle Scholar
- Maronga B (2014) Monin–Obukhov similarity functions for the structure parameters of temperature and humidity in the unstable surface layer: Results from high-resolution large-eddy simulations. J Atmos Sci 71:716–733. doi:10.1175/JAS-D-13-0135.1 CrossRefGoogle Scholar
- Maronga B, Raasch S (2013) Large-eddy simulations of surface heterogeneity effects on the convective boundary layer during the LITFASS-2003 experiment. Boundary-Layer Meteorol 146:17–44. doi:10.1007/s10546-012-9748-z CrossRefGoogle Scholar
- Maronga B, Moene AF, van Dinther D, Raasch S, Bosveld FC, Gioli B (2013) Derivation of structure parameters of temperature and humidity in the convective boundary layer from large-eddy simulations and implications for the interpretation of scintillometer observations. Boundary-Layer Meteorol. doi:10.1007/s10546-013-9801-6
- Meijninger WML, Green AE, Hartogensis OK, Kohsiek W, Hoedjes JCB, Zuurbier RM, De Bruin HAR (2002a) Determination of area-averaged water vapour fluxes with large aperture and radio wave scintillometers over a heterogeneous surface - Flevoland field experiment. Boundary-Layer Meteorol 105:63–83CrossRefGoogle Scholar
- Meijninger WML, Hartogensis OK, Kohsiek W, Hoedjes JCB, Zuurbier RM, De Bruin HAR (2002b) Determination of area-averaged sensible heat fluxes with a large aperture scintillometer over a heterogeneous surface - Flevoland field experiment. Boundary-Layer Meteorol 105:37–62CrossRefGoogle Scholar
- Meijninger WML, Beyrich F, Lüdi A, Kohsiek W, De Bruin HAR (2006) Scintillometer-based turbulent fluxes of sensible and latent heat over a heterogeneous land surface—a contribution to LITFASS-2003. Boundary-Layer Meteorol 121:89–110CrossRefGoogle Scholar
- Moeng CH, Wyngaard JC (1988) Spectral analysis of large-eddy simulations of the convective boundary layer. J Atmos Sci 45:3573–3587CrossRefGoogle Scholar
- Muschinski A, Frehlich R, Balsley BB (2004) Small-scale and large-scale intermittency in the nocturnal boundary layer and the residual layer. J Fluid Mech 515:319–351CrossRefGoogle Scholar
- Obukhov AM (1946) Turbulence in an atmosphere with a non-uniform temperature. Tr Inst Teor Geofiz Akad Nauk SSSR 1:95–115 (translation. In: Boundary-Layer Meteorol, 1971, 2:7–29)Google Scholar
- Peltier LJ, Wyngaard JC (1995) Structure–function parameters in the convective boundary layer from large-eddy simulations. J Atmos Sci 52:3641–3660CrossRefGoogle Scholar
- Petenko IV, Shurygin EA (1999) A two-regime model for the probability density function of the temperature structure parameter in the convective boundary layer. Boundary-Layer Meteorol 93:381–394CrossRefGoogle Scholar
- Raasch S, Franke T (2011) Structure and formation of dust-devil-like vortices in the atmospheric boundary layer—a high resolution numerical study. J Geophys Res 116:D16120. doi:10.1029/2011JD016010
- Raasch S, Harbusch G (2001) An analysis of secondary circulations and their effects caused by small-scale surface inhomogeneities using large-eddy simulation. Boundary-Layer Meteorol 101:31–59CrossRefGoogle Scholar
- Raasch S, Schröter M (2001) PALM—a large-eddy simulation model performing on massively parallel computers. Meteorol Z 10:363–372CrossRefGoogle Scholar
- Riechelmann T, Noh Y, Raasch S (2012) A new method for large-eddy simulations of clouds with Lagrangian droplets including the effects of turbulent collision. New J Phys 14:065008. doi:10.1088/1367-2630/14/6/065008
- Saiki EM, Moeng CH, Sullivan PP (2000) Large-eddy simulation of the stably stratified planetary boundary layer. Boundary Layer Meteorol 95:1–30CrossRefGoogle Scholar
- Shen S, Leclerc MY (1995) How large must surface inhomogeneities be before they influence the convective boundary layer structure? A case study. Q J R Meteorol Soc 121:1209–1228CrossRefGoogle Scholar
- Sühring M, Raasch S (2013) Heterogeneity-induced heat flux patterns in the convective boundary layer: can they be detected from observations and is there a blending height? - A large-eddy simulation study for the LITFASS-2003 experiment. Boundary-Layer Meteorol. doi:10.1007/s10546-013-9822-1
- Shuttleworth W, Yang ZL, Arain MA (1997) Aggregation rules for surface parameters in global models. Hydrol Earth Syst Sci 1:217–226CrossRefGoogle Scholar
- Sreenivasan KR (1996) The passive scalar spectrum and the Obukhov–Corrsin constant. Phys Fluids 8:189–196CrossRefGoogle Scholar
- Steinfeld G, Raasch S, Markkanen T (2008) Footprints in homogeneously and heterogeneously driven boundary layers derived from a Lagrangian stochastic particle model embedded into large-eddy simulation. Boundary-Layer Meteorol 129:225–248CrossRefGoogle Scholar
- Stoll R, Porté-Agel F (2006) Dynamic subgrid-scale models for momentum and scalar fluxes in large-eddy simulations of neutrally stratified atmospheric boundary layers over heterogeneous terrain. Water Resour Res 42(W01):409. doi:10.1029/2005WR003989 Google Scholar
- Stoll R, Porté-Agel F (2008) Large-eddy simulation of the stable atmospheric boundary layer using dynamic models with different averaging schemes. Boundary-Layer Meteorol 126:1–28. doi:10.1007/s10546-007-9207-4 CrossRefGoogle Scholar
- Sullivan PP, Moeng CH, Stevens B, Lenschow DH, Mayor SD (1998) Structure of the entrainment zone capping the convective atmospheric boundary layer. J Atmos Sci 55:3042–3064CrossRefGoogle Scholar
- Tatarskii VI (1971) The effects of the turbulent atmosphere on wave propagation. Kefer Press, Jerusalem 472 ppGoogle Scholar
- Thiermann V, Grassl H (1992) The measurement of turbulent surface-layer fluxes by use of bichromatic scintillation. Boundary-Layer Meteorol 58:367–389CrossRefGoogle Scholar
- Uhlenbrock J, Raasch S, Hennemuth B, Zittel P, Meijninger WML (2004) Effects of land surface heterogeneities on the boundary layer structure and turbulence during LITFASS-2003: large-eddy simulations in comparison with turbulence measurements. In: 6th Symposium on boundary layers and turbulence. Am Meteorological Society, Portland (Maine), paper 9,3Google Scholar
- van den Kroonenberg AC, Martin S, Beyrich F, Bange J (2012) Spatially-averaged temperature structure parameter over a heterogeneous surface measured by an unmanned aerial vehicle. Boundary-Layer Meteorol 142:55–77CrossRefGoogle Scholar
- Wesely ML (1976) A comparison of two optical methods for measuring line averages of thermal exchanges above warm water surfaces. J Appl Meteorol 15:1177–1188CrossRefGoogle Scholar
- Wicker LJ, Skamarock WC (2002) Time-splitting methods for elastic models using forward time schemes. Mon Weather Rev 130:2088–2097CrossRefGoogle Scholar
- Wieringa J (1976) An objective exposure correction method for average wind speeds measured at a sheltered location. Q J R Meteorol Soc 102:241–253CrossRefGoogle Scholar
- Williamson JH (1980) Low-storage Runge–Kutta schemes. J Comput Phys 35:48–56. doi:10.1016/0021-9991(80)90033-9 CrossRefGoogle Scholar
- Wilson C, Fedorovich E (2012) Direct evaluation of refractive-index structure functions from large-eddy simulation output for atmospheric convective boundary layers. Acta Geophys 60:1474–1492CrossRefGoogle Scholar
- Wyngaard JC, LeMone MA (1980) Behavior of the refractive index structure parameter in the entraining convective boundary layer. J Atmos Sci 37:1573–1585CrossRefGoogle Scholar
- Wyngaard JC, Coté OR, Izumi Y (1971a) Local free convection, similarity, and the budgets of shear stress and heat flux. J Atmos Sci 28:1171–1182CrossRefGoogle Scholar
- Wyngaard JC, Izumi Y, Collins SAJ (1971b) Behavior of the refractive-index-structure parameter near the ground. J Opt Soc Am 61:1646–1650CrossRefGoogle Scholar

## Copyright information

**Open Access**This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.