Spatio-temporal stability and abnormality of chlorophyll-a in the Northern South China Sea during 2002-2012 from MODIS images using wavelet analysis

Detecting a regular pattern of chlorophyll-a (Chl-a) in the ocean can provide a preliminary scientific understanding of regional environmental changes. The objective of this research was to identify the potential of a wavelet transform to capture and describe both the stationary level and anomalous variability of Chl-a. An 11-year time series (from July 2002 to December 2012) of the Moderate Resolution Imaging Spectroradiometer (MODIS) chlorophyll-a product in the Northern South China Sea (NSCS) was collected. The Data INterpolating Empirical Orthogonal Functions (DINEOF) was used to reconstruct the original MODIS data. The approximation and detailed components from the original series of the MODIS Chl-a data were considered to be a source of the stationary level and anomalous variability of Chl-a, respectively. The stationary level of the Chl-a concentration was characterized by the Chl-a concentration of the coastal areas that was higher than that of the open ocean area, as well as monthly, seasonal and annual averaged Chl-a concentrations concentrating on between 0.05 and 0.25 mg m(-3). The anomalous variability of Chl-a has a short-oscillating period of 0.5 years; specifically, the Chl-a negative amplitude occurred in spring and autumn, and the positive amplitude was recorded in winter and summer. Furthermore, a long-oscillating period of four years, that is, the inter-annual singularity of Chl-a, primarily appeared in May 2003, May 2007 and May 2011. The maxima of the Chl-a concentration were dominated by between 0.5 and 1 mg m(-3). The peak winter Chl-a concentration was mainly located in the open ocean area, and the peak summer Chl-a concentration was mostly limited to the coastal region. This study suggests that a wavelet transform is promising for detecting the anomalous and stationary variability of ocean parameters. (C) 2014 Elsevier Ltd. All rights reserved.