SciELO - Scientific Electronic Library Online

Home Pagelista alfabética de revistas  

Servicios Personalizados




Links relacionados


Gayana (Concepción)

versión impresa ISSN 0717-652Xversión On-line ISSN 0717-6538

Gayana (Concepc.) v.68 n.2 supl.TIIProc Concepción  2004 

  Gayana 68(2): 611-614, 2004


Gabriel Yuras1,3, Osvaldo Ulloa1,3, Ricardo Letelier4 & Oscar Pizarro2,3

1. Departamento de Oceanografía, Universidad de Concepción. Casilla 160-C, Concepción, Chile.
2. Departamento de Física del Océano y de la Atmósfera (DEFAO), Universidad de Concepción. Casilla 160-C, Concepción, Chile.
3. Centro de Investigación FONDAP-COPAS, Universidad de Concepción
4. College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, USA.


Satellite - derived oceanographic parameters were used to explain the coastal and oceanic behavior of the chlorophyll concentration (CHL) in Chilean waters (18°-40°S, coast- 900 km offshore). CHL shows two different regimens of the annual cycle, the closer to the shore one with a austral summer maximum in concordance to the upwelling and wind regimes; and an offshore one with a CHL maximum in winter time. The annual cycle is stronger in a thin band (20 km) close to the shore and in the oceanic zone, where it counts for more of the 80% of the variability present. There is a band, which varies with latitude, located between 100-200 km from coast where the annual cycle is not well resolved by the 100 km data bins used in this study. A EOF analysis of 9km resolution data was performed and the first mode (29.5% of variance explained) represented clearly the annual cycle and its 6 months out of phase behavior among coastal and oceanic waters The previous idea of a weak annual cycle in northern Chile, changes to a strong annual cycle close to the shore with maximum CHL in summer, an intermediate zone in which it is no possible to resolve an annual cycle and an outer zone in which the annual cycle has a maximum in the winter months.


The study area is shown in Figure 1. It goes form 18 to 40S and from coast to 900 km offshore. Meridionally it has been divided in 1 degree fringes, centered in the corresponding degree. Each fringe has been zonally divided in 23 subareas ending at 20, 50, 100 and every 100 km to 900 km from shore. For every one of the 253 bins (23 latitudes by 11 distances from shore) SeaWiFS chlorophyll, Pathfinder SST, Quickscatt wind, ERS wind stress curl and Levitus temperature, salinity and density time series has been composed in order to evaluate the annual cycle behavior.

Figure 1: Study area, showing the bin arrangement for the zonal transects. Lines signal the 20, 50, 100, 200, 300, 400, 500, 600, 700, 900 and 1000 km from shore. Background image SeaWiFS Chlorophyll mean concentration for the 1998-2003 period.

The SeaWiFS L1a High Resolution Picture Transmition (HRPT) data was captured by the University of Chile and University of Concepcion ground receiving stations. They were retrieved from the NASA DAAC ( project, processed at Regional Programme for Physical Oceanography and Climate studies (PROFC), University of Concepcion using the SEADAS and the standard Mc Clain et al. (1998) procedures to derive OC4 chlorophyll (O' Reilly et al,1998) with a space resolution of ca 1 km at nadir. L3 9 km resolution mapped SeaWiFS images were obtained from NASA DAAC as well. Monthly averages were produced using all the high resolution data available. The data projection was changed from latitude-longitude to latitude-distance from shore, and all the monthly averages were fitted to the previously described grid in order to produce time series. For each time series an annual cycle was adjusted using harmonic analysis and a Montecarlo simulation was performed 5000 times to get an estimation of the confidence of fitted harmonic. The same approach was applied to wind stress derivated and wind stress curl from ERS and then the binning approach was used on those data also. SeaWiFS L3 data were interpolated using a 3D (latitude-longitude-time) approach to reduce missing data; they were then reshaped to a latitude v/s distance from shore array. Empirical Orthogonal Function (EOF) using the single value decomposition (SVD) was applied to those data.


In order to further investigate the phase of the annual cycle for chlorophyll in the study area, annual harmonics were fitted to all the 23 latitudes and 11 distances from shore data. Figure 2 shows the summary of the annual harmonic fitted to the data. At 20 km from shore (Figure 4A) the maximum of the annual cycle phase (MACP) it is located in late spring and summer months (between November and February). Roughly it can be signaled that in the northern Chile (18- 23S) MAC it is produced later in summer (January and February). In the northernmost part of this area (18-23S) R2 is more then than 40% and the Montecarlo simulation percentage (MSP) shows that 100% of the 5000 simulated permutations of the time series had and amplitude smaller than the recorded for the real time series; this index gives an idea of how well-fitted and harmonic is, a low (high) value indicates that there is a high (low) probability that a random permutation of the series produces a better fit of the harmonic thus indicating that the annual harmonics does not (does) fit well to the data. The annual cycle amplitude (ACA) decreases southwards from 5 mg m-3 to 2.5 mg m-3 . From 24 to 29S the MACP is in summer months too, mainly in January; MSP is close to 100; R2 is smaller than the ones in the northern part and are in the range of 20-40% and the ACA are smaller too, having a sharp decrease at 24S, being around 1 mg CHL m-3. Southward of 30S the MACP is located in December with exception of 31 and 40S), MSP are also close to 100 and R2 are in a similar range than the previous one (20-45). The ACA in this zone has a striking increase at 36-37S being those latitudes similar to the previous ones in MACP, MSP and R2.

Figure 2: Chlorophyll annual cycle at A) 20 km from shore and B) 400 km from shore. Diagrams represent the fitted annual harmonic for every latitude (18-40S) at the referred distance from shore, two years are shown for better visual effect. Right panels are the r2 of the fitted harmonic and the numbers are the Montecarlo Simulation percentage.

Between 20 and 50 km from shore (not shown) the MACP in the northern part of the study area remains in spring and summer, but both the ACA (less than 2 mg CHL mg m-3), MSP (as low as 72%) and R2 (5-30) are decreased when compared with the more coastal ones (up to 20 km). Between 24 and 29S MACP are located in spring (November and December) and at 26S is in Late July. ACA is smaller than in the northern area (less than 1 mg CHL m-3), MSP and R2 are reduced too indicating that the annual harmonic does not fit well to the data. From 30 and southward the MACP is located in late spring ­ early summer, ACA is higher than in the central part of the study area and increases toward the 35-37S zone (from 1 to 3.6 mg CHL m-3), coinciding with the Chilean widest continental shelf; MSP and R2 (up to 51%) are higher than in the central part too showing a better fit for the annual harmonic, mainly from 35S and southward. Between 50 and 100 from coast (not shown) the MACP is shifted towards winterÊ months (June-September) in the northern part of the study area (18-23S). Here the ACA remains below 1 mg CHL m ­3, MSP and R2 show that the annual harmonic does not fit well (R2 below 40%). From 24 to 29S MACP shifts from middle July to late October remaining R2 lower than 40% and MSP above 88%. From 30 to 36S MACP is located from early September (33S) to late December (36S); ACA remains around 1 mg CHL m-3 and increases sharply towards 36S (1.8 mg CHL m-3), south of that decreases to values in the range of 0.5 mg CHL m-3. MSP is higher than 90% with exception of 37 and 38S which also have the smallest R2 values (less than 5%).

The situation at 400 km offshore (figure 4b) illustrates what happens from there to 900 km off. There MACP is located only in winter months, being slightly closer to late august for middle latitudes (around 25S). For latitudes from 30 to 40S and distances from 400 to 900 km from coast MACP is located between June and July in middle winter. ACA is always lower than 1 mg CHL m ­3 for all latitudes; moreover, with the exception of 35-37S is always below 0.5 mg CHL m­3. MSP is 100% for all latitudes but one (40S). R2 remains over 40% for most of the latitudes at becomes higher as it goes offshore being higher than 50% at 500 km from coast) and higher than 60% for most of the latitudes from 600 to 900 km (not shown).

Figure 3 shows the first mode of the EOF analysis performed in to the low resolution SeaWiFS data and shows clear the annual cycle in both the oceanic and coastal regions. The coastal zone has a high positive correlation with the first mode of the EOF, and the oceanic zone has a high negative correlation with this mode, showing that the phase when the chlorophyll is maximum in coastal waters it is minimum in oceanic waters.

Figure 3: UP: First mode of a Empirical Orthogonal Function (EOF) applied to SeaWiFS 9 km data. Down: Spatial structure of the first eigenvector.


The existence of an annual chlorophyll signal along the Chilean coasts has been noticed by several authors (Thomas et al. 1994, Thomas, 1999; Thomas et al. 2001). Comparing these 3 first works with the results presented here, differences arise in both the chlorophyll annual cycle phase and absolute concentrations. In order to explain such differences it is important to point out the different data sets used by the authors. CZCS were of a continuous nature and made extremely difficult to obtain a clear vision of a seasonal cycle. Moreover SeaWiFS and CZCS data can not be directly compared and let us to contrast the concentrations for Chilean coast just for the Thomas et al. (2001) and this work.

The answer of why having used the data produced by the same satellite and having such differences in both annual cycle maximum phase and chlorophyll concentration must be searched in the data process. Thomas et al (2001) used daily GAC (4 km) data produced with standard (OC4) NASA algorithm, and then they composite together the first 100 km from shore; Here the data used where LAC (1 km) L2 produced from L1A data with OC4 algorithm at PROFC, then the first 100 km have been divided in 3 regions (1-21, 21-50 and 51-100 km) and it is in those regions where the difference arises. There is a narrow coastal fringe where chlorophyll concentrations exceed 5 mg m-3; this fringe extends from shore to 20 km in the northern Chile (18-23S) and for the same area it drops to 1 mg m-3 and less than mg m-3 for 21-50 and 51-100 km from shore. This narrow high chlorophyll band can not be resolved in a shore ­ 100 km average, where it will be smeared by the comparatively much lower chlorophyll concentrations from 20 to 100 km.

It is possible to summarize with the data analyzed in this work that coastal (up 20 km from shore) Chilean waters have a chlorophyll annual cycle with maximum concentrations in summer months with is in agreement with the upwelling favorable wind stress. After that there is a zone (50 ­200 km depending on latitude) where the annual cycle it is not well defined. Offshore of 200 km the chlorophyll annual cycle has its maximum in winter months. Data availability does not allow us to elucidate which are the factors driving this oceanic annual cycle.


We thank the SeaWiFS Project (Code 970.2) and the Distributed Active Archive Center (Code 902) at the Goddard Space Flight Center, Greenbelt, MD 20771, for the production and distribution of the SeaWiFS L1a data, respectively.



McClain, CR, Cleave ML, Feldman GC, et al.1998. Science quality SeaWiFS data for global biosphere research. SEA TECHNOL 39(9):10-16. [1]

O'Reilly, J. S. Marticorena, B.G. Mitchell, . Siegel, K. Carder, S. Garver, M. kahru and C. McClain. 1998. Ocean Color chlorophyll algorithms for SeaWiFS. JGR vol 103, Num C11:24937:24954. [2]

Thomas, A... F. Huang, P. Strub y C. James. 1994 Comparison of the seasonal and interannual variability of phytoplankton pigment concentrations in the Peru and California Current systems. . Journal of Geophysical Research 99(C4):7355-7370. [3]

Thomas, A. 1999. Seasonal distributions of satellite-measured phytoplankton pigment concentration along the Chilean coast. . Journal of Geophysical Research 104(C11):25877-25890. [4]

Thomas, A C; Blanco, JL; Carr, ME; Strub, PT; Osses, J. 2001. Satellite-measured chlorophyll and temperature variability off northern Chile during the 1996-1998 La Nina and El Nino. Journal of Geophysical Research.. 106(C1):899-915. [5]


Creative Commons License Todo el contenido de esta revista, excepto dónde está identificado, está bajo una Licencia Creative Commons