Application of open photoacoustic cell in an eddy covariance system for water vapor flux measurement

Water vapor flux plays a crucial role in surface-atmosphere exchange processes as evapotranspiration regulates the energy balance of the surface. Moreover, it transfers water vapor into the atmosphere and, as a result, shapes the hydrological cycle. The eddy-covariance (EC) technique is the most commonly applied method to directly measure water vapor flux over a wide variety of surfaces. An EC arrangement consists of a 3D sonic anemometer and a gas analyzer. To derive surface fluxes, wind components and the gas concentration (e.g. water vapor) have to be recorded with high-frequency (at least at 10 Hz). In the case of open-path (sampling-free) EC systems, infrared (IR) gas analyzers are used dominantly, which are still quite large so that e.g. they cannot be easily mounted on drones. In contrast, small and light sonic anemometers are available for flux measurements.

In this study, we present the application of a sampling-free photoacoustic (PA) sensor for water vapor flux measurement employing the EC technique. The fast response PA sensor records the water vapor concentration through an open cylindrical chamber having an overall size of less than 1 dm³. On the one hand, a previous first test showed that the vertical covariance functions obtained by the PA cell follow closely to those resulting from an accepted IR sensor. On the other hand, the PA system showed some underestimation at higher frequencies based on the analysis of co-spectra.  

To comprehensively test and evaluate the PA cell for flux measurements, a seven-week-long field measurement was performed over a plain grassland when a calibrated EC150 IR sensor (Campbell Sci.) was used as a reference gas analyzer. We analyze the accuracy of the PA system: (i) depending on the orientation of the cell or i.e. the wind direction, and (ii) for a broad range of meteorological conditions, such as different wind speeds and atmospheric stability. Furthermore, to overcome the high-frequency attenuation, we establish and apply empirical spectral transfer functions following the literature and standard EC postprocessing procedures. The characteristic response time of the PA sensor is also assessed.