A novel wavelet-based approach to characterize dynamic environmental factors controlling short-term soil surface CO2 flux: Application to a controlled CO2 release test site (EIT) in South Korea

Abstract

Soil surface CO2 flux (FCO2) fluctuates complicatedly in time and space according to hydro-meteorological and atmospheric conditions as well as physicochemical and biological conditions of soil. In this study, we developed a novel approach to identify and quantitatively evaluate the driving forces (DFs) that control FCO2 fluctuations in the near-surface environment with the combined use of dynamic factor analysis (DFA) and wavelet-based multiresolution analysis (WMRA). We focused on short-term (<16 days) periodic DFs using 6-hourly data blocks observed for 107 days. The procedures were as follows: first, potential DFs (PDFs) were examined using DFA for 11 types of data (i.e., solar radiation, air temperature, relative humidity, atmospheric pressure, wind speed, soil water content, soil temperature, soil EC, water vapor, and the concentration and flux of CO2). Then, major DFs (MDFs) were identified using multiresolution correlation analysis (MODCA) and wavelet transform coherence (WTC). MODCA, with the support of WTC, quantitatively estimated the multiscale similarities and lagged phase responses between MDFs and observations including precipitation by using a scale-localized maximum correlation coefficient (rmax). Finally, the contributions of MDFs to FCO2 were evaluated by calculating the effective dynamic efficiency (Def) based on dynamic factor loadings (α) of MDFs to FCO2 and cumulative wavelet energy ratios at main time-frequency scales. As a result, four MDFs and their maximum contributions to baseline FCO2 were characterized as follows: MDF1 represented irregular precipitation (scales of 1/2 to 2 days; α = 0.38; rmax = |−0.36|; Def ≤ 7.58%) and the effect of evapotranspiration (scales of 1/2 to 2 days; α = 0.38; rmax = 0.58; Def ≤ 21.57%); MDF2 addressed seasonal precipitation (scales of 8 to 16 days; α = 0.53; rmax = 0.91; Def ≤38.33%) and the soil moisture effect (scales of 8 to 16 days; α = 0.53; rmax = 0.79; Def ≤52.86%); MDF3 and MDF4 were the surface atmospheric stability effect (scales of 8 to 16 days; α = 0.09, rmax = |−0.85|; Def ≤ 8.25%) and the soil temperature effect (scales of 8 to 16 days; α = 0.44; rmax = 0.29; Def ≤41.79%), respectively. A comparison with conventional multiple regression showed that the combined application of DFA and WMRA supplements the conventional method by considering the dynamic properties of DFs and their contributions and by analyzing correlations at multiresolution time-frequency scales. The present method is expected to better characterize MDFs for other complex physicochemical and biological interactions in the environmental system.