Introduction

The OHATS sonic array design was changed from that used during the HATS field project, based on the experience gained during the data analysis phase of HATS (seeHorst et al, 2004: HATS: Field Observations to Obtain Spatially Filtered Turbulence Fields from Crosswind Arrays of Sonic Anemometers in the Atmospheric Surface Layer", J. Atmos. Sci., 61, 1566-1581). These changes were the use of nine sonics in both of the two, vertically-displaced lines of sonics and a smaller vertical spacing between the two lines.

Nine sonics in both arrays

The HATS sonic array permitted finite difference estimates of vertical gradients to be made from the spatially-filtered measurements at two heights, which are applicable at the vertical mid-point of the two horizontal lines of sonics. On the other hand, finite difference estimates of crosswind gradients were made from the spatially-filtered measurements made at 5 crosswind distances by the nine sonic array at one of the heights. Thus the vertical and crosswind gradient estimates apply at two different heights, the vertical mid-point of the two horizontal lines of sonics and the height of the nine-sonic line. An improvement was made for the OHATS project by using nine sonics in both of the vertically-displaced lines of sonics. Thus crosswind (and streamwise) gradient estimates can be made at both heights and averaged to provide horizontal gradient estimates at the same height as the vertical gradient estimates, the vertical mid-point of the two horizontal lines of sonics.

Equal horizontal and vertical sonic spacing

Secondly, the vertical spacing between the two lines of sonics during the HATS field project was equal to the sonic spacing for the first sonic array geometry, but increased to twice the sonic spacing for the other three array geometries. It was found that the equal spacing of the first HATS array provided a better match between the spatial response of the vertical and horizontal gradient estimates. Thus for OHATS, the vertical spacing of the two lines of sonics was made equal to the horizontal sonic spacing, 58 cm.

Marshall flow distortion test

The possibility of flow distortion by the sonic arrays was monitered during the HATS field project by deploying two sonics at the same heights as the horizontal sonic arrays, but on an independent tower separated by 10 m from the horizontal arrays. The turbulence statistics measured by each of the sonics within the sonic arrays were then compared to the measurements by the independent sonics. For all four of the HATS sonic array geometries, the turbulence statistics within the array agreed with those measured by the independent sonics.

However, the proposed OHATS array has a closer vertical spacing and more sonics than used during any of the HATS arrays Therefore a pre-project test was made at the Marshall, CO, field site to examine the possibility of flow distortion by the OHATS array. The Marshall array consisted of two horizontal ASTER masts at nominal heights of 4.75 m and 5.27 m. (Because of interference between the cross-bracing of the vertical ASTER masts and the attachment hardware, we could not separate the horizontal masts by the desired 58 cm and had to settle for 52 cm.) The horizontal masts are oriented east-west so that winds from the south are perpendicular to them. About 6 m to the west, we erected a reference tower. Nine sonic yokes were placed on each of the two horizontal masts, with a spacing of 58 cm, and CSAT electronics were mounted on the back of the masts in the intervening spaces. Sonics were mounted at the center three locations on both horizontal masts and at the same two heights on the reference tower. Photos of the Marshall array can be seen at the bottom of the OHATS photo page.

The sonics are designated as, e.g. u3, l2, etc, with u and l referring to the upper and lower arrays and the ordinal number increasing from west to east. Thus u1 and l1 are the reference sonics and u3 and l3 are the center sonics in the horizontal arrays. The specific sonics are:

Data were collected from May 1-12 and the results are summarized below as linear fits of the flow statistics from the horizontal array sonics to the reference sonics. The data used for the comparisons were restricted to wind directions within ±45° of normal to the array (nominally SW to SE). Here nmr denotes the mean absolute deviation from the linear fit, normalized by the standard deviation of data. The conclusion is that the flow distortion caused by the sonic array is minimal.

Wind Speed (942 5-min averages):

Spd_u2 = 0.987 Spd_u1 + 0 cm/s; nmr = 0.042 

Spd_u3 = 0.983 Spd_u1 + 1 cm/s; nmr = 0.044 

Spd_u4 = 0.988 Spd_u1 + 3 cm/s; nmr = 0.043 

Spd_l2 = 0.994 Spd_l1 + 0 cm/s; nmr = 0.045 

Spd_l3 = 0.992 Spd_l1 - 1 cm/s; nmr = 0.044 

Spd_l4 = 0.984 Spd_l1 + 1 cm/s; nmr = 0.044

Turbulent Kinetic Energy (942 5-min averages; TKE < 5 m^2/s^2):

TKE_u2 = 0.984 TKE_u1 + 1e-3 m^2/s^2; nmr = 0.069 

TKE_u3 = 0.978 TKE_u1 + 2e-3 m^2/s^2; nmr = 0.070 

TKE_u4 = 0.993 TKE_u1 + 1e-3 m^2/s^2; nmr = 0.071 

TKE_l2 = 0.994 TKE_l1 + 0e-3 m^2/s^2; nmr = 0.072 

TKE_l3 = 0.995 TKE_l1 + 0e-3 m^2/s^2; nmr = 0.073 

TKE_l4 = 0.984 TKE_l1 + 0e-3 m^2/s^2; nmr = 0.074 

Friction velocity (359 15 min averages):

u*_u2 = 0.999 u*_u1 + 1 cm/s; nmr = 0.202 

u*_u3 = 0.995 u*_u1 + 1 cm/s; nmr = 0.209 

u*_u4 = 0.965 u*_u1 + 1 cm/s; nmr = 0.215 

u*_l2 = 1.01 u*_l1 + 1 cm/s; nmr = 0.204 

u*_l3 = 0.986 u*_l1 + 1 cm/s; nmr = 0.222 

u*_l4 = 0.960 u*_l1 + 1 cm/s; nmr = 0.208