Post-process eddy covariance data with ease using R package openeddy

Eddy covariance is one of the most precise and direct methods for measurement of fluxes of matter and energy at the ecosystem scale. It is instrumental in the development of our understanding of carbon and water cycles. It allows us to examine local conditions at the given site or provide global picture of the interaction of the terrestrial ecosystems with the overlaying atmosphere through data integration in modelling frameworks. The application of the method is multifaceted, and the data processing consists of multiple steps (i.e. raw data processing, quality control, gap-filling, flux partitioning, aggregation) that are dependent and result in a processing chain. Especially for new teams applying the method and not being connected to station networks, it might be a daunting process to set up the processing chain without available tooling. Fortunately, in this respect, a lot of publicly available software is already available, especially for raw data processing and gap-filling and flux partitioning. Although quality control is a required step before gap-filling, the tools simplifying the process and making it reproducible have not received equivalent attention yet. The purpose of the R package openeddy is to fill this gap and support independent researchers with a software infrastructure for eddy covariance data post-processing that improves the reproducibility of the results. A set of tutorials is prepared within this contribution that helps to exemplify different features of openeddy software (https://github.com/lsigut). These include:

• loading and saving of general tabular data including the support of units placed below the header
• remapping of variable names including aggregation across multiple variables
• automated merging of multiple EddyPro full output files
• scatter plot for the whole year of data with a treatment of outliers
• automated extraction of quality control information from coded columns included in EddyPro full output files
• general purpose functions for data filtering according to specified thresholds
• despiking function for removing outliers in the time series showing both daily and yearly variability
• functions to define the region of interest of the ecosystem station and perform footprint filtering based on 1D footprint output from EddyPro software
• assisted manual data exclusion
• combining multiple quality control filters with different properties into one quality control column
• summarization of quality control results in a tabular form or as a figure
• visualization of the time series of selected variable and auxiliary (meteorological) variables; the plots are optimized for viewing of half-hourly data in weekly and monthly intervals but any resolution is supported
• time series aggregation into various defined intervals including unit conversions
• barplots for plotting of aggregated results
• evaluation of aggregated uncertainty of flux measurements
• computation of Griebel et al. 2020 space-time equitable budgets with uncertainty estimation
• computation of spatio-temporal sampling coverage

This work was supported by the Ministry of Education, Youth, and Sports of the Czech Republic within the National Infrastructure for Carbon Observations—CzeCOS (No. LM2018123) and SustES—Adaptation Strategies for Sustainable Ecosystem Services and Food Security under Adverse Environmental Conditions (CZ.02.1.01/0.0/0.0/16_019/0000797).