Eddy Covariance Measurements Of Marine Atmospheric Boundary-Layer Turbulence And Air-Sea Fluxes

Turbulence characteristics of marine atmospheric boundary layer and air-seafluxes are important parts of air-sea interaction study, and are important for theunderstanding of global climate change. The eddy covariance is a direct method usedto measure the atmospheric turbulence and is expected to play a vital role for theunderstanding of air-sea interaction. However, due to the inherent difficulty in themarine atmosphere using the eddy covariance method, the measurement of air-seaCO₂flux has been shown to suffer from large uncertainties. Hence, further researchon the application and data processing of the eddy covariance method is necessary.In this study, the turbulent transport characteristics, and air-sea CO₂flux inmarine surface layer was analyzed using an open-path eddy covariance system on anoceanographic research vessel in the North Yellow Sea. Using the data set, wedeveloped a series of approaches to identifying cross sensitivity to determine thefrequency range less affected by cross sensitivity. In addition, we developed a simplemethod to reduce artifacts caused by cross sensitivity in estimating air-sea CO₂flux.The air-sea nanoparticle flux was measured in Qianliyan Island using eddy covariancesystem. The following are the main results:（1） The effect of the ship motion on wind velocity can be seen by comparing theraw and corrected spectra of the vertical wind velocity fluctuations. The effect of theship-motion is apparent as a spike, and the false spike can be effectively removed bythe ship motion correction. The spectrum from anchored platform did not change atother frequency range after the ship-motion correction. However, the spectrum frommoving platform changed over low frequency after ship-motion correction. Thismight be due to the low frequency translational motion of the platform was notremoved completely and added components on the other two velocity measurementsfollowing the instantaneous tilt corrections. （2） For measurement from moving platform, it was found that the relationshipbetween dimensionless wind speed and stability parameter followed1/3power law asexpected from similarity theory. It was also found that the coefficients of similarityfunction were comparable with the results from anchored platform and other surfaceconditions such as land and other sea area. The parameter of turbulence intensity wasdifferent between result from moving platform and that from anchored platform.However, the relationship between turbulence intensity and wind speed was consistent.And hence, the difference between values was due to the different measurementconditions. The approach measuring on moving platform is thereby suggested to bereasonable.（3） The eddy covariance method can obtain reasonable air-sea moment flux,sensible heat flux and latent heat flux. However, the air-sea CO₂flux measured by anopen-path eddy covariance system has been shown to suffer from large uncertaintiesand was higher than that estimated using other methods. The normalized spectra ofthe fluctuations of three scalars （CO₂, water vapor, and temperature） were analyzed.In11out of64samples, the normalized spectra of scalars showed similarity. However,in the remaining samples, the normalized CO₂spectra were observed to be greaterthan those of water vapor and temperature at low frequencies. In this paper, the noisedue to cross sensitivity was identified through a combination of intercomparisonsamong the normalized spectra of three scalars and additional analyses. Uponexamination, the cross sensitivity noise appeared to be mainly present at frequency<0.8Hz. Our analysis also suggested that the high-frequency fluctuations of CO₂concentration （frequency>0.8Hz） was probably less affected by the cross sensitivity.The cross sensitivity is expected to be the reason that cause the low frequencycontribute more to the variance in CO₂fluctuations.（4） To circumvent the cross sensitivity issue, the cospectrum in the highfrequency range0.81.5Hz, instead of the whole range, was used to estimate the CO₂flux by taking the contribution of the high frequency to the CO₂flux to be the same asthe contribution to the water vapor flux. The air-sea CO₂flux estimated by the eddycovariance method was0.23mg m^-2s^-1. However, the air-sea CO₂flux (0.039mg m^-2s^-1) estimated using the revised eddy covariance method was only one-fifth ofthat estimated using the direct eddy covariance method. The results suggest that thenoise caused by cross sensitivity could account for approximately80%of fluctuationsof the CO₂mixing ratio. The results also suggest that cross sensitivity could be thelargest artifact source and can be effectively removed by the revised eddy covariancemethod.（5） The nanoparticle concentration shows that there is a new particle formationevent in Qianliyan Island on morning of3May2011. The diameter of particles wasbelow20nm during the beginning period of new particle formation, and the particlesgrew gradually up to with diameter of50nm. During the period before the newparticle formation, the apparent change of meteorological condition was that thesoutherly wind became northerly wind. The new particle formation might be relatedwith the change of wind direction. The bin averaged nanoparticle flux during newparticle formation estimated using the eddy covariance method was6.7×10⁶m^-2s^-1 and the average of median was5.8×10⁶m^-2s^-1 The bin averaged nanoparticle fluxduring period without new particle formation was2.7×10⁶m^-2s^-1and the average ofmedian was2.1×10⁶m^-2s^-1)During measurement, the turbulent nanoparticle fluxeswere bidirectional, and the upward flux account for greater proportion. Comparing thenanaparticle flux during period of new particle formation with that during periodwithout new particle formation, it was found that the proportion of upward flux wasgreater and this might be an indication that the new particle formation occurredmainly in surface layer.