B53A-0154:
Wavelet Analysis - A Building Block for NEON’s Ecosystem Exchange Data Products

Friday, 19 December 2014
David Durden1, Stefan Metzger1, Rommel C Zulueta1, Natchaya Pingintha Durden1, Ke Xu2, Natascha Kljun3 and Jeff R Taylor1, (1)National Ecological Observatory Network, Fundamental Instrument Unit, Boulder, CO, United States, (2)University of Wisconsin Madison, Madison, WI, United States, (3)Swansea University, Swansea, United Kingdom
Abstract:
The use of wavelet analysis has been increasing in geophysical sciences over the past 20 years. Its ability to decompose a time series into its frequency components while maintaining their localization in time provides valuable information on the mechanistic properties influencing turbulent exchange. Compared to Fourier transforms, Wavelet analysis is well suited to analyze non-stationary signals and provides high temporal resolution without dismissing longer wavelength contributions.

One goal of NEON is to provide high quality ecosystem exchange observations to the science community, and wavelet analysis is one tool that enables multiple data processing pathways, such as; (i) high- and low-frequency spectral corrections, (ii) comparison to spectral references, (iii) evaluation of flow characteristics in complex terrain, (iv) high-frequency EC flux processing and source area calculations (≈60 h−1), and (v) inferring the flux field around tower measurements.

Here, we provide an overview of how Wavelet analyses are integrated into NEON’s ecosystem exchange data processing framework. Preliminary results include: (i) Changes in soil CO2 concentration are dominated on timescales >0.5 h, which informs the frequency response correction of a very slow but robust, diffusion-based soil CO2 sensor; (ii) Source area calculations explain significant spatial variation when resolved at the integral timescale of atmospheric turbulence; and (iii) our companion presentations “Towards the spatial rectification of tower-based eddy-covariance flux observations” and “Assessing and correcting spatial representativeness of tower eddy-covariance flux measurements” demonstrate how Wavelet analysis facilitates inferring the flux field around tower measurements.