Deployment and evaluation of a system for ground-based measurement of cloud liquid water turbulent fluxes

Direct interception of windblown cloud water by forests has been dubbed ''occult deposition'' because it represents a hydrological input that is hidden from rain gauges. Eddy correlation studies of this phenomenon have estimated cloud water fluxes to vegetation yet have lacked estimates of error bounds. This paper presents an evaluation of instrumental and methodological errors for cloud liquid water fluxes to put such eddy correlation measurements in context. Procedures for data acquisition, processing (including correction factors), and calibration testing of the particulate volume monitor (PVM) and forward-scattering spectrometer probe (FSSP) are detailed. Nearly 200 h of in-cloud data are analyzed for intercomparison of these instruments. Three methods of coordinate system rotation are investigated; the flux shows little sensitivity to the method used, and the difference between fluxes at different stations is even less sensitive to this choice. Side-by-side intercomparison of two PVMs and one FSSP leads to error bounds of 0.01-0.035 g m(-3) on half-hour mean cloud liquid water content (relative to typical values of 0.35 g m(-3)) and 2-3.5 mg m(-2) s(-1) on the surface-normal liquid water flux (typical magnitude of 7 mg m(-2) s(-1) for these data), depending on which instruments are compared.