Sensible heat and momentum flux measurement with an optical inner scale meter

Path-averaged measurements of sensible heat (H) and momentum fluxes (u.) were made under unstable conditions (ζ ξ ζ < 0.1) above pasture using an inner scale meter (ISM), an instrument which employs in tandem both a diverging laser beam-point detector scintillometer and a second instrument using a large-aperture (15 cm) incoherent source and receiver of the same diameter. By distorting the beam, refractive turbulence along the optical path creates intensity fluctuations or scintillations at the receiver; eddies with spatial scales of the order of the beam diameter are most effective at this, so that whereas the laser responds to the smaller eddies in the flow, the large-aperture instrument is sensitive to inertial subrange turbulence. By combining signals from these two instruments, it is in principle possible to calculate H and u. explicitly and independent of assumptions about surface roughness. Pathlengths were 100 m for the laser and 292 m for the large-aperture instrument. The site was chosen with the expectation of providing a homogeneous test surface. Fluxes were first derived from a measurement of mean windspeed and either scintillometer operating independently, and, in line with previous experience, these agreed very closely with reference eddy correlation measurements. The worst case was when winds blew down the propagation path and H derived from the large aperture differed on average from reference measurements by 11%. We suspect that this systematic difference was real and that spatial non uniformities in the surface energy budget combined with the pathlength differences meant that the two scintillometer beams experienced slightly different path averaged refractive turbulence under some wind directions. Whatever the reason, this discrepancy engendered unrealistic ISM estimates for u. which depend very sensitively on the difference in the scintillometer output voltages. After excluding such data, ISM estimates and direct eddy correlation measurements of H agreed within 2%, but the error in u. was still unacceptably high (residual standard deviation 0.09 ms ^-1). A sensitivity analysis confirms that the ISM estimate of H is relatively robust compared with u ?, and we conclude that, although the ISM does provide an alternative technique for the measurement of sensible heat, it yields only a coarse estimate of the momentum flux.