Cyberinfrastructure for eddy covariance systems: From site selection to data submission

Eddy covariance (EC) is one of the most direct methods to measure the vertical turbulence that drives the mass exchange of heat, water vapor, and carbon within the atmospheric boundary layer. EC is mathematically complex and challenging to implement, especially for a researcher with little experience. Therefore, significant care in setting up the EC equipment and processing data is required; failure to do so may cause numerous types of data errors. In addition, analysis of data from flux towers is site specific, and it is nearly impossible to use a uniform methodology across different ecosystems. Although, a large effort to unify different approaches is being done by national and international flux networks (e.g., FLUXNET, Ameriflux, CarboEurope, NEON), as well as the flux community in general; Implementation remains challenging especially for non-experience users and small labs striving to collaborate with large networks. Here, we introduce Semantic Abstract Workflows (SAWs) as a novel approach to document and manage the processes used for the implementation of an eddy covariance research site. From site selection, data collection, data processing, data preparation for submission to community networks. Our eddy covariance tower represents the Chihuahuan Desert and specifically aids the study of land-atmosphere interactions in a mixed creosote (Larrea tridentata) - mesquite (Prosopis glandulosa) shrubland located within the Jornada Experimental Range, USDA ARS, Las Cruces, NM. . Documented processes in the form of SAWs were constructed using the Wdo-it tool1. These SAWs are currently being used to annotate data as is being generated from the Jornada EC site. The Provenance Mark-up Language (PML) is being used to structure such annotations in a form that effectively captures provenance of data. Another tool called Probe-IT is then used to analyze data from a provenance perspective. SAWs provide a simple graphical notation intended to facilitate reuse of process documentation among communities of researchers to easily and effectively capture provenance about eddy covariance processes with minimal training. The resulting outputs support collaborative research by optimizing knowledge transfer within and between small labs or large networks.