Scale-Specific Controller of Carbon and Water Exchanges Over Wheat Field Identified by Ensemble Empirical Mode Decomposition
- Publication Type:
- Journal Article
- International Journal of Plant Production, 2018, 12 (1), pp. 43 - 52
- Issue Date:
© 2017, Springer International Publishing AG, part of Springer Nature. The exchange of carbon and water in the ecosystem is influenced not only by weather and climatic perturbations but also by vegetation dynamics. The relationship between carbon and water exchange and environment in agro-ecosystem across different temporal scales is not very often been quantified. Spectral analysis of eddy covariance measurements can identify the interactions between environmental and biological factors at multi-temporal scales. Here, we used a new method, ensemble empirical mode decomposition (EEMD), to study the temporal covariance between ecosystem exchange of carbon dioxide (NEE), latent heat flux (LE) and environmental factors in a winter wheat cropping system located at the North China Plain. The results showed that the NEE, LE and environmental factors can be decomposed into 12 significant quasi-period oscillations on various time-scales i.e. hourly, diurnal, weekly and seasonal timescales. Variance of NEE in diurnal, hourly, seasonal, weekly scale was 58.9, 29.6, 4.7, 0.6%, respectively. Variance of LE in diurnal, hourly, seasonal, weekly scale was 55.2, 15.5, 5.1, 1.8%, respectively. The largest of variance contribution is at diurnal time-scale from net radiation (Rn), wind speed (μ) and vapor pressure deficit (VPD) due to daily rhythms in solar radiation. The soil water content varied significantly at a relatively longer time-scale i.e. weekly and seasonal scale. Large variance contribution of ambient temperature (T) (63.4%) and VPD (33.6%) is in trend term due to the significant increasing seasonal trend from winter to summer. The correlation analysis indicated that NEE and LE was correlated highly with net radiation (Rn) at all time-scale, as well as with VPD, ambient temperature (T), and wind speed (μ) in diurnal scale and with soil water in seasonal time-scales. This implied that solar radiation contributed the main variation of carbon and water in short time-scale, i.e. hourly and diurnal. Soil water variation strongly correlated with the seasonal variation of NEE and LE. Furthermore, seasonal signals of NEE and LE synchronized with LAI, which indicated that carbon dioxide and water flux are also regulated by LAI in seasonal time-scale. The quantification of the variation explained by carbon and water fluxes and environmental factors across different temporal scales using EEMD improved the understanding of carbon and water process in a cropping system.
Please use this identifier to cite or link to this item: