Seasonal and interannual variability of chlorophyll-a and associated physical synchronous variability in the western tropical Pacific

Abstract

Based on long-term satellite-derived ocean data sets and methods of empirical orthogonal function and singular value decomposition, we investigated the spatiotemporal variability of the chlorophyll-a concentration (CHL) on seasonal and interannual timescales in the western tropical Pacific associated with physical ocean variables of sea surface temperature (SST), sea level anomaly (SLA) and sea surface wind (SSW), and the El Niño Southern Oscillation (ENSO) index. The bio-physical synchronous variation on interannual timescale was also confirmed in terms of the scales of variability and oscillation periods in the time-frequency space using the methods of Fourier transform, Morlet wavelet transform, and wavelet coherence analysis. On a seasonal timescale, the first two modes of the monthly mean CHL fields described the consecutive spatiotemporal variation in CHL in the western tropical Pacific. CHL reached the maximum during late winter–early spring and minimum during summer–early autumn with the exception of the northeast of Papua New Guinea and the Solomon Islands. The CHL bloom in boreal winter–spring was closely associated with cold SST, high sea level along the North Equatorial Countercurrent meanders, and strong wind. On an interannual timescale, the variability of CHL exhibited a close correlation with SST, SLA, SSW, and ENSO. During El Niño, CHL increased in the oligotrophic western basin of the warm pool associated with cold SST, low SLA, and strong westerly winds but decreased in the mesotrophic eastern basin of the warm pool in association with warm SST, high SLA, and weak easterly trade winds. There may exist time-lag for the bio-physical covariation, i.e., CHL and SST varied simultaneously within 1month, and CHL variations led SLA by approximately 0–3months but lagged wind speed by about 1month. In the time-frequency domain, the interannual variability in CHL and physical ocean variables had high common power, indicating that the variability scales and oscillation periods of CHL were significantly related to these of SST, SLA, and ENSO index. The significant anti-phase relationships were also shown between CHL and SST, CHL and SLA, and CHL and multivariate ENSO index through the wavelet coherence analysis.