The influence of the Guadalquivir river on spatio-temporal variability in the pelagic ecosystem of the eastern Gulf of Cádiz

Abstract

This study examines the spatio-temporal variability of the turbidity plume and phytoplankton biomass (in terms of chlorophyll) in the marine region influenced by the Guadalquivir estuary using ocean colour images over a period of 11 years (2003-2013). The area of the turbidity plume was calculated using water-leaving radiance at 555 nm (nLw555). Climatologic and monthly averages showed recurrent high nLw555 levels in winter and high chlorophyll in spring. Similar variability was confirmed by Empirical Orthogonal Function (EOF) analysis of 8-day composite images, illustrating the existence of different regions with similar behavior. The first EOF mode explained 60.7% and 31% of the variability in nLw555 and chlorophyll, respectively, and was associated with enhanced Total Suspended Solids (TSS) in autumn-winter and phytoplankton blooms in winter-spring periods. The results confirmed that the development of the turbidity plume and subsequent phytoplankton blooms were strongly regulated by river discharges and precipitation. Indeed, interannual variation in nLw555 was consistent with changes in the large-scale climate index, the North Atlantic Oscillation, a proxy for regional rainfall patterns. In the case of phytoplankton biomass, the second EOF mode revealed differentiation between offshore and nearshore areas, the latter characterized by delayed development of phytoplankton bloom due to light limitation by high TSS. This suggests that the stability of the water column, via its influence on phytoplankton light-limitation, influenced also the timing and magnitude of phytoplankton bloom events. The dynamic of the Guadalquivir estuary turbidity plume is a crucial factor for the pelagic ecosystem of the Eastern Gulf of Cadiz, governing phytoplankton productivity.

A high-resolution hydrodynamic-biogeochemical coupled model of the Gulf of Cadiz – Alboran Sea region.

“…This is also true for the GoC where satellite estimates of Chla concentrations seems to fit very well with in-situ measures in the open sea regions (e.g. while quite a consistent bias has been described for coastal, shallow regions (Caballero et al, 2014). Henceforth, the large differences between observed and modelled Chla values along the coast of the GoC could be due to a combination of an overestimation of satellite data and an underestimation of model simulation, the true value being somewhere in between.…”

“…This light limitation is mainly caused by extremely high levels of suspended sediments in this coastal region (e.g. Caballero et al, 2014). Not incorporating the effect of suspended material on the light environment of the model could be partially responsible for the larger differences between observed and modelled Chla levels in the coastal regions of the GoC.…”

High-Chlorophyll-Area Assessment Based on Remote Sensing Observations: The Case Study of Cape Trafalgar

“…Chl-a concentration was derived from the OC4Me ocean colour algorithm [26]. The MERIS-OC4Me algorithm yielded reasonable chl-a retrievals in the GoC (r = 0.96; n = 27; p < 0.0001; bias of 0.31 mg m −3 and Root-Mean-Square Error of 0.41 mg m −3 ; [25]), although with a systematical overestimation [25] also found in this region using Moderate-Resolution Imaging Spectroradiometer (MODIS) products [27]. Finally, as extremely high chlorophyll values are not usually detected in the area of interest, only pixels with chl-a < 10 mg m −3 were considered [16,28].…”

Functional regression on remote sensing data in oceanography

“…The surface samples taken for analysis were collected with a rosette sampler (5 m below the water surface) at a distance of between 1km and 25 km from the coast. The amount of TSS concentration in each sample is measured according to the protocols given in Caballero et al (2014a).…”

“…In oceanography, Gong et al (2015) used Functional Principal Components Analysis (FPCA) to model high-dimensional temperature curves and temperature surfaces of Lake Victoria. Nevertheless, there are still only a few studies that use the FDA approach to remote sensing satellite data in the field of oceanography (Lahet et al, 2001;Gong et al, 2015), even though there are many applications of multivariate analysis techniques in this field (Clarke et al, 2006;Caballero et al, 2014a).…”