Exploiting oceanographic satellite data to study the small scale coastal dynamics in a NE Atlantic open embayment Noela Sánchez-Carnero a, ⁎, Elena Couñago a,b , Daniel Rodriguez-Perez c , Juan Freire a a Grupo de Recursos Marinos y Pesquerías, Universidade da Coruña, C/Alejandro de la Sota, 1, 15008, A Coruña, Spain b Fismare Innovación para la Sostenibilidad S.L. Rúa Castro de Elviña, n° 1, Bajo Izq, 15008, A Coruña, Spain c Departamento de Física Matemática y de Fluidos, Facultad de Ciencias, UNED, P° Senda del Rey 9, 28040 Madrid, Spain abstract article info Article history: Received 18 August 2010 Received in revised form 3 February 2011 Accepted 16 March 2011 Available online 22 March 2011 Keywords: Coastal zone Scientific satellites GLM Temporal variations Spatial variations Oceanographic variables An approach to the processing and analysis of oceanographic satellite data is presented suitable for its use in coastal areas. Satellite images of a small (300 km 2 ) open embayment in the NW of the Iberian Peninsula (Seno de Corcubión) are used to study a decade of time series of three important oceanographic variables: sea surface temperature (SST), chlorophyll a concentration ([Chlo]) and turbidity (K490). Their interannual and seasonal trends, together with the regional variability, are assessed using a Generalized Linear Model (GLM). We identify seasonal upwelling–downwelling processes, on top of interannual trends that show that coastal dynamics is not directly driven by the oceanic trend. To study spatial patterns the study area was divided in subareas; these subareas show significant differences among them in seasonal as well as in interannual scales. Using this information, some oceanographic hypotheses were tested based on the statistical significance of the variable residual dependencies: a possible relationship between SST and turbidity (found to be not significant in the study area), and the presence of short-term upwelling events reported by other researchers in this geographical region (not found statistically significant either). A database and an oceanographic atlas summarizing all the information used in the analyses have been generated and can be freely visualized online and downloaded. © 2011 Elsevier B.V. All rights reserved. 1. Introduction The coastal zone, comprising from 200 m above to 200 m below sea level, is of major economic and social importance. Despite being no more than 18% of the globe surface, it is the area where around a quarter of the global primary production occurs and it supplies approximately 90% of world fish catch (Bijlsma, 1997). For these reasons, the characterization of processes that take place in coastal ecosystems is essential to devise a suitable management strategy. Galicia (NW Spain) is a region highly dependent of the coastal area since it supports a large number of human settlements directly related to the sea (Freire and García-Allut, 2000). This region is characterized by a high productivity caused by the fertilization associated to frequent seasonal upwelling events of Eastern North Atlantic Central Water mass (ENACW). These events bring cold nutrient-rich water to the coastal area, especially in the spring–summer period (Fraga, 1981; Fiuza et al., 1998). Due to their environmental and economic relevance, upwelling in the Galician Rías has been widely studied by Fraga (1981), Blanton et al. (1984), Torres and Barton (2007) and Álvarez-Salgado et al. (1996), just to mention a few. The traditional direct methods to study these events, as well as other oceanographic processes, are very expensive in terms of time and money. Specific surveys performed to obtain field data require a high organizational effort for their implementation and don’t allow the acquisition of long homogeneous time series. Albeit monitoring networks were started recently, even nowadays these networks are scarce and, in the case of Galicia, the data are not easily accessible to the public even for research purposes. Contrary to these methodologies, oceanographic satellite data provide retrospective and homogeneous time series for oceanograph- ic variables (as Sea Surface Temperature, SST, chlorophyll a concen- tration [Chlo] and turbidity among others). For this reason, since the early 80s, the satellite imagery has been intensively used to characterize spatial patterns in the global ocean (McClain, 2009), but not in coastal areas due to their spatial resolution. As in the case of temporal series, an adequate statistical treatment of the imagery data is needed. In this sense, the use of empirical orthogonal functions (EOF) has attracted attention to describe spatially varying behaviors (Tew-Kai and Marsac, 2009; Nardelli et al., 2010); EOF are effective for pattern extraction, but difficult to interpret in terms of oscillatory normal modes. On the contrary, linear models are familiar to many scientists and easy to understand. Moreover, generalized linear models (GLM) combine this simplicity with their adaptability to non- linear functional dependences (Lindsey, 1997). Journal of Marine Systems 87 (2011) 123–132 ⁎ Corresponding author. Tel.: + 34 981167000x2208. E-mail address: noela.sanchez@udc.es (N. Sánchez-Carnero). 0924-7963/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jmarsys.2011.03.007 Contents lists available at ScienceDirect Journal of Marine Systems journal homepage: www.elsevier.com/locate/jmarsys