Mapping AmeriFlux footprints: Towards knowing the flux source area across a network of towers

Abstract:

The AmeriFlux network collects long-term carbon, water and energy flux measurements obtained with the eddy covariance method. In order to attribute fluxes to specific areas of the land surface, flux source calculations are essential. Consequently, footprint models can support flux up-scaling exercises to larger regions, often based on remote sensing data. However, flux footprints are not currently being routinely calculated; different approaches exist but have not been standardized. In part, this is due to varying instrumentation and data processing methods at the site level. The goal of this work is to map tower footprints for a future standardized AmeriFlux product to be generated at the network level.

These footprints can be estimated by analytical models, Lagrangian simulations, and large-eddy simulations. However, for many sites, the datasets currently submitted to central databases generally do not include all variables required. The AmeriFlux network is moving to collection of raw data and expansion of the variables requested from sites, giving the possibility to calculate all parameters and variables needed to run most of the available footprint models. In this pilot study, we are applying state of the art footprint models across a subset of AmeriFlux sites, to evaluate the feasibility and merit of developing standardized footprint results.

In addition to comparing outcomes from several footprint models, we will attempt to verify and validate the results in two ways: (i) Verification of our footprint calculations at sites where footprints have been experimentally estimated. (ii) Validation at towers situated in heterogeneous landscapes: here, variations in the observed fluxes are expected to correlate with spatiotemporal variations of the source area composition. Once implemented, the footprint results can be used as additional information within the AmeriFlux database that can support data interpretation and data assimilation. Lastly, we will explore the expandability of this approach to other flux networks by collaborating with and including sites from the ICOS and NEON networks in our analyses. This can enable utilizing the footprint model output to improve network interoperability, thus further promoting synthesis analyses and understanding of system-level questions in the future.