Hostname: page-component-7b7567c8c-fvjsw Total loading time: 0 Render date: 2026-11-17T09:20:16.598Z Has data issue: false hasContentIssue false

Scanning Doppler lidar measurements of drag force on a solitary tree

Published online by Cambridge University Press:  28 April 2021

Nikolas Angelou*
Affiliation:
Department of Wind Energy, Technical University of Denmark, Roskilde, 4000, Denmark
Jakob Mann
Affiliation:
Department of Wind Energy, Technical University of Denmark, Roskilde, 4000, Denmark
Ebba Dellwik
Affiliation:
Department of Wind Energy, Technical University of Denmark, Roskilde, 4000, Denmark
*
†Email address for correspondence: nang@dtu.dk

Abstract

Trees are known to reduce the wind momentum efficiently. Yet, firm quantitative estimates of their contribution to the land surface drag have remained elusive, partly because trees have complex shapes that consist of elastic multi-scale elements. This structural complexity makes trees inherently difficult to scale for wind tunnel studies. Here, we test a new method for quantifying the drag force on a solitary mature tree in its natural environment. The method is based on the application of mass and momentum conservation over a control volume that encloses the tree. For this control volume, the drag force is estimated through the momentum deficit in the wake. For the characterisation of the heterogeneous and high-gradient wind field in the wake, spatially distributed measurements of the wind vector were acquired using three synchronously scanning wind lidar instruments in a vertical plane encompassing the wake. The resulting drag force estimate is compared to a reference measurement from a tree-mounted sensor at the base of the stem. We find that the drag force in both methods shows a dependence on the wind speed raised to an exponent of 1.8 and that the drag force, based on the momentum deficit method, is consistently underestimated by 1 %–10 %. Potential reasons for this bias are discussed in light of the accuracy of both methods. The relatively close agreement between the two methods indicates that scanning Doppler lidar measurements can be used to determine the drag force on complex objects in their natural environment, such as trees.

Information

Type
JFM Papers
Copyright
© The Author(s), 2021. Published by Cambridge University Press
Figure 0

Figure 1.Elevation map of a $100\times 40\ \textrm {m}^2$ area around the tree used in this study showing the locations of the sonic anemometers at the M1 and M2 meteorological masts using the symbol ‘$\times$’ and ‘$+$’, respectively, the three short-range WindScanner instruments (denoted WS1, WS2 and WS3, respectively) using the symbol $\bullet$ and the horizontal projection of the vertical scanning pattern using a white line. The positive $x$-axis is pointing towards the direction of $110^\circ$, relative to the geographic North.

Figure 1

Figure 2.Photographs of the oak tree during the winter, autumn and summer of 2017. The photographs correspond to different states of the crown: (a) with a bare crown, (b) during the time of the field experiment, when the leaves were in the abscission phase and (c) with a fully developed foliage.

Figure 2

Figure 3.Top (a) and side (b) views of the control volume used in this study. The tree (black), the scanning plane (grey solid line) and the up- and downwind sonic anemometers (denoted as x and +, respectively). The minor and major axes of the inlet and outlet areas are denoted $a_{inlet}$, $a_{outlet}$ and $b_{inlet}$ and $b_{outlet}$.

Figure 3

Figure 4.Mean measured profiles of (a) the horizontal wind speed in five different heights (black dots) and the three models of vertical profile (A: logarithmic profile ($z \geqslant 1.5$ m) (dashed grey), B: logarithmic profile ($z\geqslant 4$ m) (solid grey) and C: profile based on the model suggested by Högström (1988) (solid black). (b) The vertical component of the wind vector, (c) the tilt angle and (d) of the friction velocity. The dashed line corresponds to the lower and higher limits of the tree crown ($H_c = 4.5$ m).

Figure 4

Figure 5.Histograms of the mean wind direction (a) over the time period examined in this study (left) and the mean wind speed of the direction sector $285^\circ \leqslant a < 295^\circ$ (b). The averaging period corresponds to the iteration time of a scanning plane. The dashed line in the wind speed histogram represents the threshold criterion applied to the selection of the cases used in this study.

Figure 5

Table 1.Wind conditions during the selected case studies, where $N$ is the number of the scans realised, $U_{{bin}}$ is the $1\text {-}\textrm {ms}^{-1}$ range between the minimum and maximum wind speeds in a bin, $u$ and $\sigma _{{u}}$ are the mean and standard deviation of the streamwise wind and $\alpha$ and $\sigma _{\alpha }$ are the mean and standard deviation of the wind direction, $d_y$ the expected displacement of the center of the wake; $A$, $B$ and $u_o$ are the fit parameters of the upwind profile using (2.7).

Figure 6

Figure 6.Vertical profile of the mean streamwise component of the six cases of different wind speeds as measured by the short-range WindScanners (blue) and up- and downwind sonic anemometers (red and black, respectively). A model of the up-wind profile is also presented in a thick red line. The error bars and the light blue region correspond to one standard deviation about the mean for the case of the sonic and WindScanner measurements, respectively. The grey line corresponds to the tree height.

Figure 7

Figure 7.Scatter plot of the mean streamwise component measured in the wake of the tree by the short-range WindScanners ($u_{ws}$) versus the corresponding sonic anemometer ($u_{sonic}$) measurements. The dashed line corresponds to the identity line. The linear fit over the ensemble of all the cases is presented with a black line.

Figure 8

Figure 8.Example (case II) of the mean (a) measured $u$ and (b) normalised by the upwind profile $\hat {u}$ streamwise component of case II. The vertical and horizontal profiles that intersect at the aerodynamic centre, depicted by white dashed lines in (b) are presented in (c) and (d), respectively.

Figure 9

Table 2.Ratio between the inlet and outlet areas of the control volume ($A_{inlet}/A_{outlet}$) for each of three streamline divergence patterns defined in figure 10 and for each of the wind speed cases presented in table 1; $\{y_c,z_c\}$ are the spanwise and vertical coordinates of the location of the wake centre, $d_o$ is the scaling parameter of the inlet area for which the conservation of the mass rate is satisfied.

Figure 10

Figure 9.Three patterns of streamline divergence caused by the presence of the tree, according to which the streamlines are considered to (a) diverge homogeneously radially from the centre of the tree, (b) diverge mainly above and below the tree and (c) diverge mainly on the sides of the tree.

Figure 11

Figure 10.Measured (a,c,e) and normalised (b,d,f) by the mean wind speed, difference of the mass rate between the inlet and outlet vertical elliptical planes, for different dimensions of the inlet plane for three different streamline divergence patterns; P1: (a)–(b), P2: (c)–(d) and P3: (e)–(f).

Figure 12

Figure 11.Estimated values of the wind drag based on (a) the momentum deficit for three different patterns (P1, P2 and P3) of streamline divergence and (b) the 1 min mean bending moment measurement. Comparison between the drag estimation using the P1 streamline pattern and (c) the bending moment and (d) the wind load on the same tree during the summer (long dashed grey line) and the winter (short dashed grey line) based on the results presented in Angelou et al. (2019).

Figure 13

Figure 12.(a) Vertical profile at $y/H=0$, of the $\langle uu \rangle$ measured at the upwind distance of $x/H=-2.3$ (red) and at the downwind distances of $x/H=-1.5$ (black) and $x/H=-1.3$ (blue, scanning plane). (b) Linearly interpolated plane of the upwind $\langle u'u' \rangle$ based on the fit presented in (a) (solid red line) and (c) measurement of $\langle u'u' \rangle$ in the wake of the tree.