CFD Model Estimates of the Airflow Distortion over Research Ships and the Impact on Momentum Flux Measurements

Abstract

Wind velocity and air-sea turbulent flux measurements made from shipborne instruments are biased due to the effect of the ship on the flow of air to the instruments. The presence of the ship causes the airflow to a particular instrument site to be either accelerated or decelerated, displaced vertically, and sometimes deflected slightly in the horizontal. Although recognized for some time, it is only recently that the problem has been addressed using three-dimensional computational fluid dynamics (CFD) models to simulate the flow over particular ships, quantify the effects of flow distortion, and hence correct the ship-based measurements. It has previously been shown that this improves the calculated momentum fluxes by removing disparities between data from different ships, or from instruments in different locations on the same ship. This paper provides validation of the CFD model simulations. Two research ships were instrumented with multiple anemometers located in both well-exposed and badly exposed sites. Data are compared to the results of model simulations of the flow at various relative wind directions and wind speeds. Except when the anemometers are in the wake of an upwind obstruction, the model and the in situ wind speed estimates typically agree to within 2%. Direct validation of the model-derived estimates of the vertical displacement of the flow was not possible due to the extreme difficulty of obtaining such measurements in the field. In this study, simulations of flows at 0Њ and 90Њ from the bow of the ship were made and displacements of about 1 and 5 m were found, respectively. These results were used to correct the in situ momentum flux data. In one case, the application of the different bow-on and beam-on corrections for vertical displacement successfully removed the disparity seen in the uncorrected data. In a second case, the beam-on vertical displacement overcorrected the flux results. This overcorrection could be caused either by uncertainties in the in situ estimate of the relative wind direction or by partial adjustment of the turbulence during the vertical displacement. The effects of flow distortion are found to vary only slightly with wind speed, but are very sensitive to the relative wind direction and, if uncorrected, can cause large biases in ship-based meteorological measurements (up to 60% for the drag coefficient). Model results are given for bow-on flows over 11 research ships (American, British, Canadian, French, and German).

AutoFlux: an autonomous system for the direct measurement of the air-sea fluxes of CO2, heat and momentum

“…The ID method has the advantage that a) the flux results are insensitive to the motion of the ship and b) they can be corrected for the effects of the ship distorting the air flow to the sensors using numerical models of the air flow around the ship. 20 Biases of up to 60% are possible in momentum flux measurements made via the ID method from well-exposed instruments on research ships, 1 but these biases can be removed using the results from the numerical models. Momentum and latent heat flux measurements have been successfully made using the ID method for a number of years.…”

Direct measurements of the CO2 flux over the ocean: Development of a novel method

“…[20] These are results are preliminary and may change with further refinements of the analysis procedure such as: sensor alignments [Brooks, 2008]; changes to the head deformation; impacts of air-flow distortion [e.g., Yelland et al, 1998Yelland et al, , 2002; possible effects of solar heating of the sensor [Burba et al, 2008] etc. The instrumentation has operated continuously for three years: once the entire data set is processed we expect to greatly decrease the uncertainty in the parameterisation of the gas transfer velocity, and extend the parameterisation to wind speeds above 15 m.s -1 .…”

Trends in ship wind speeds adjusted for observation method and height

“…These include differences in flow distortion due to ship geometry and anemometer location (Yelland et al, 2002;Moat et al, 2005Moat et al, , 2006a, changes in composition (by ship type) of the VOS fleet, and differences in wind sampling period and averaging method (Dobson, 1981;Lindau, 1995b;Gulev, 1999;Thomas et al, 2005). Cardone et al (1990) attempted to adjust individual estimated and measured wind reports from the period 1946-1984, but were hindered by limited information about the observation method and the anemometer heights of individual ships.…”