Contents lists available at ScienceDirect

Progress in Oceanography

journal homepage: www.elsevier.com/locate/pocean

Interplay between freshwater discharge and oceanic waters modulatesphytoplankton size-structure in fjords and channel systems of the ChileanPatagonia

L. Antonio Cuevasa,b,⁎, Fabián J. Tapiac,d, José Luis Iriartec,e,g, Humberto E. Gonzálezf,g,Nelson Silvah, Cristian A. Vargasa,b,i

a Laboratorio de Funcionamiento de Ecosistemas Acuáticos, Departamento de Sistemas Acuáticos, Facultad de Ciencias Ambientales, Centro EULA-Chile, Universidad deConcepción, Chileb Centro para el Estudio de Forzantes Múltiples sobre Sistemas Socio-Ecológicos Marinos (MUSELS), Universidad de Concepción, Chilec Centro de Investigación Oceanográfica Copas Sur-Austral, Universidad de Concepción, Casilla 160-C, Concepción, Chiled Centro Interdisciplinario para la Investigación Acuícola (INCAR), Universidad de Concepción, Chilee Instituto de Acuicultura, Universidad Austral de Chile, Puerto Montt, Chilef Institute Ciencias Marinas y Limnológicas, Universidad Austral de Chile, Valdivia, Chileg Centro de Investigación Dinámica de Ecosistemas Marinos de Altas Latitudes (IDEAL), Universidad Austral de Chile, Valdivia, Chileh Escuela de Ciencias del Mar, Pontificia Universidad Católica de Valparaíso, Valparaíso, Chilei Instituto Milenio de Oceanografía (IMO), Universidad de Concepción, Chile

A R T I C L E I N F O

Keywords:Phytoplankton size structureChlorophyll-aMineral nutrientsPrimary productionFjordsChilean Patagonia

A B S T R A C T

Here, we participated in five independent research cruises that spanned almost the entire Patagonian fjordsregion (from 41.5 to 56.0°S) in order to determine the importance of the physical/chemical factors that influencephytoplankton size structure triggered by freshwater discharge and oceanic water intrusion. Throughout thePatagonian region phytoplankton biomass varies in association with freshwater discharge and mineral nutrientload, and to a lesser extent with surface solar radiation and photosynthesis. These correlations and the spatialdomains changed depending on which size fraction was analyzed. Fresh water discharge negatively correlatedwith salinity and density, positively correlated with stratification and silicic acid concentration, and representsthe primary influence on phytoplankton populations. More than 40% of the sites characterized as discharge-intensive locations exhibited lower total chlorophyll-a concentrations (chl-a) and phytoplankton size-structurethat was dominated by small cells (< 20 µm). Oceanic nutrients (nitrate and phosphate input) are the second-most important factor that control total chl-a, favoring total chlorophyll-a concentration in the southern half ofthe Patagonian region. Microphytoplankton contribute to more than 75% of the total community in high pro-ductivity waters with chl-a concentrations higher than 2 µg L−1 and picophytoplankton dominate when chl-a islower than 1 µg L−1. Thus, in this extensive area, the relative success of different phytoplankton size classes maybe sensitive to changes in hydrological cycles, continental runoffs, and potential anthropogenic eutrophication,modifying important ecological processes and the fate of organic matter.

1. Introduction

Primary productivity and autotrophic biomass varies widely inaquatic ecosystems, where coastal and estuarine waters account for asignificant fraction of the global primary productivity. Along the con-tinental margins of South America chlorophyll-a has increased from1998 to 2004 (Gregg et al., 2005), however, locally along the western

coast of South America the concentration has decreased (Lara et al.,2016). Despite their potential role in biological productivity, carboncycling, and nutrient remineralization, these coastal waters, includingthe fjords and channels of the Chilean Patagonia (references in Iriarteet al., 2007), are among the least studied ocean regions of the world. Itis well known that Patagonian fjords and inner channel areas south of41.5°S undergo strong seasonal fluctuations, mainly due to variations in

https://doi.org/10.1016/j.pocean.2019.02.012Received 12 June 2018; Received in revised form 19 February 2019; Accepted 28 February 2019

⁎ Corresponding author at: Laboratorio de Funcionamiento de Ecosistemas Acuáticos, Departamento de Sistemas Acuáticos, Facultad de Ciencias Ambientales,Centro EULA-Chile, Universidad de Concepcion, Chile.

E-mail addresses: lcuevas@udec.cl (L.A. Cuevas), ftapiaj@udec.cl (F.J. Tapia), jiriarte@uach.cl (J.L. Iriarte), hgonzale@uach.cl (H.E. González),nsilva@ucv.cl (N. Silva), crvargas@udec.cl (C.A. Vargas).

Progress in Oceanography 173 (2019) 103–113

Available online 28 February 20190079-6611/ © 2019 Elsevier Ltd. All rights reserved.

T

precipitations and solar radiation (Acha et al., 2004; Pickard, 1971).Since Patagonian fjords are major CO2 sink areas (Torres et al., 2011),understanding the factors that modulate the size structure of theplanktonic community will help to understand the factors that controlthe efficiency of the biological pump. A previous study (Iriarte et al.,2012) suggested that nitrogen sources (e.g. nitrate and ammonia) couldplay an important role in maintaining relatively large productivity rates(3 g Cm−2 d−1) in Patagonian fjords. Additionally, water columnstratification positively correlates to thin layers of high chlorophyll-aconcentrations (Ríos et al., 2016), indicating the importance of mixing/stability for autotrophic biomass in the southern part of the Patagonianfjords. Seasonal carbon fluxes varied widely in this ecosystem in termsof productivity and vertical fluxes of particulate organic carbon,reaching high values during the productive seasons (González et al.,2010, 2011, 2013, 2016). In terms of phytoplankton size structure, cellssmaller than 11 μm dominate when chl-a values< 1 μg L−1, and cellslarger than 20 μm dominate when chl-a > 1 μg L−1 for fjords andchannels in the Aysén area from 43° to 47° South (Paredes andMontecino, 2011). This pattern was previously observed in differentareas of the Patagonian fjords (Montecino and Pizarro, 2008; Pizarroet al., 2005). The area is also characterized by high precipitations(Garreaud et al., 2013), producing higher freshwater discharge into thefjord system (Kilian et al., 2013), and strong salinity gradients in thecoastal zone (Dávila et al., 2002; Saldías et al., 2018). Thus, highfreshwater runoff transports a large amount of energy and terrestrialcarbon (Vargas et al., 2011), high rates of terrestrial nutrient input,particularly dissolved silica (Torres et al., 2011, 2014) and essentialmicronutrients (Iriarte et al., 2014). Consequently, it will be ideal totest whether temperature, nutrient availability, and surface radiationtogether with other variables affect phytoplankton size structure in asensitive fjord systems such as those of the Chilean Patagonia.

Phytoplankton are often regulated by nutrient availability and lossprocesses such as mortality, grazing, viral lyses, and their importancehas been studied extensively in freshwater (McQueen et al., 1989;Staehr and Sand-Jensen, 2006) and marine ecosystems (Calbet andLandry, 2004; Chen and Liu, 2010). Factors such as hydrographicconditions, water column stability, light availability, and freshwaterinputs may influence the composition and productivity of phyto-plankton communities in coastal and estuarine ecosystems, thus af-fecting biogeochemical fluxes (Chisholm, 1992). In the upper photiczone, limitation of mineral nutrients (e.g. N, P, Si, Fe) and light avail-ability regulate autotrophic biomass and modify the structure of phy-toplankton communities; mainly in high-turbidity ecosystems such ascoastal zones and estuaries (Falkowski and Oliver, 2007; Marañónet al., 2012).

The size structure of phytoplankton determines the trophic organi-zation of pelagic ecosystems and the efficiency by which organic matterproduced by autotrophs is channeled towards upper trophic levels orexported through both the ‘classical’ and ‘microbial’ pathways. Smallerphytoplankton cells, due to higher surface-area-to-volume ratio, have acompetitive advantage over large cells in nutrient-impoverished en-vironments (Chisholm, 1992; Kiørboe, 1993). Conversely, large phy-toplankton cells are capable of sustaining higher rates of biomass-spe-cific production in nutrient-rich waters (Cermeño et al., 2005; Marañónet al., 2007), and are less strongly controlled by predators (Kiørboe,1993). This explains why picophytoplankton (< 2 μm in diameter)

dominates the autotrophic biomass and primary production in oligo-trophic areas, whereas microphytoplankton (> 20 µm diameter) isdominant in eutrophic areas (Agawin et al., 2000; Marañón et al.,2001). Two different scenarios can be expected depending on the pre-vailing phytoplankton size structure. If small cells (usually< 20 μm)are dominant, in terms of autotrophic biomass and primary production,little potential for carbon export can be expected due to the low sedi-mentation rates and intense recycling of matter through the microbialfood web (González et al., 2010, 2013; Marañón et al., 2012). Ad-ditionally, if phytoplankton biomass is dominated by large species,organic matter can be efficiently transferred as well as enhanced, pro-ducing downward export fluxes and biological CO2 drawdown througha short classical food chain favoring the biological pump and carbonsequestration (Legendre and Le Fevre, 1995; Turner, 2002).

The paucity of regional-scale information on the relative importanceof factors that may regulate the structure and productivity of phyto-plankton assemblages in the Chilean Patagonia is mainly due to theextent and geographic complexity of this region, as well as the limitedavailability of scientific vessels. However, annual research cruisessponsored by the Chilean Navy’s CIMAR program have been conductedover the past 12 years, providing a piecemeal coverage of the entireregion. Here we present a meta-analysis of data gathered from 5 suchcruises (Table 1), all carried out during different but consecutive years,which span ca.14 degrees of latitude (42–56°S). We discuss the patternsof phytoplankton size structure that emerge, and analyze their asso-ciation with hydrographic conditions (temperature, salinity, density,stability of the water column), surface radiation (satellite-derived PAR),and mineral nutrients (nitrate, phosphate, silicic acid), as well as withtotal phytoplankton biomass and productivity rates within the euphoticzone (0–25m depth).

Our goal is to establish whether freshwater discharges and oceanicwater intrusion into the fjord modify phytoplankton size structureleading to ecological changes that may impact carbon cycling andbiogeochemical fluxes in this region.

2. Materials and methods

2.1. Study area

The Chilean Patagonia constitutes one of the most important andextensive fjord and channel regions in the world, extending from theReloncaví Fjord (41.5°S) to Cape Horn (56.0°S) (Fig. 1). It spans a lineardistance of ca. 1600 km in length, and is composed of 84,000 km ofcoastline and 240,000 km2 of complex geomorphology and hydro-graphy, characterized by a highly fragmented and rugged coastline witha variety of environments such as embayments, fjords, channels, straits,and estuaries (Silva and Palma, 2008). Similar to other high latitudeecosystems, Chilean fjords receive freshwater discharges from rivers –many of them of glacial origin – as well as surface runoff and groundwater flows due to the strong influence of subpolar low-pressure sys-tems associated with high rainfall (2000–5000mmyear−1 (Calvete andSobarzo, 2011; Dávila et al., 2002). At the same time, these fjords andchannels are strongly influenced by tidal fluctuations and by sub-sur-face intrusions of oceanic waters, mainly Sub-Antarctic waters (SAAW)from the South Pacific Ocean (Aiken, 2012, 2008; Palma and Silva,2004).

Table 1Date, position and number of stations sampled in the different research cruisers covering almost the entire Patagonian fjord area.

Cruise Date Year Latitude Number of stations

CIMAR-12 4–14 November 2006 41°30′S–43°30′S 19CIMAR-13 30 October–15 November 2007 43°30′S–46°30′S 30CIMAR-14 28 October–27 November 2008 47°00′S–50°10′S 28CIMAR-15 12 October–20 November 2009 50°10′S–52°50′S 20CIMAR-16 13 October–20 November 2010 52°30′S–56°00′S 32

L.A. Cuevas, et al. Progress in Oceanography 173 (2019) 103–113

104

2.2. Physical, chemical, and biological data

To assess the relative importance of different phytoplankton sizeclasses across the region of interest, we compiled data on total and size-fractioned chlorophyll-a, together with CTD profiles, satellite-derivedsurface PAR (MODIS-Aqua, Level 3, 8-day composites), mineral nu-trients (nitrate, phosphate and silicic acid), and primary productionestimates from five independent research cruises (CIMAR-Fjords cruises12–16, from 2006 to 2010, Table 1) carried out during spring between41.5°S and 56.0°S (Fig. 1). Physical structure of the water column wasrecorded with a Seabird 19 CTD attached to a rosette equipped with 24Niskin bottles. Water column stability (Brunt-Väisälä Frequency, N) was

calculated as: =∂

∂N g

ρgz , using gravity (g) and seawater density (ρ)

differences among 4 layers (surface, 5, 10 and 25m depth). Densitycomputations were based on the atmosphere International Equation ofState of Seawater (UNESCO, 1981). Surface PAR values for each stationwere obtained from 4-km resolution MODIS-Aqua images gathered forthe 2006–2010 period from NOAA’s ERDDAP platform (http://

coastwatch.pfeg.noaa.gov/erddap/griddap/erdMHpar08day.html).Eight-day composite images were used because of the frequent lack ofdata over this region due to cloud coverage. Thus, PAR values for agiven sampling station and date were extracted from the nearest pixelin the weekly image that encompassed that date. When the nearest pixelfor a specific sampling site did not contain data, the monthly mean orspringtime (October-November) mean were used. To measure theconcentration of nitrate (NO3−), phosphate (PO4

−3), and silicic acid (Si(OH)4), discrete water samples (50mL) within the euphotic zone (sur-face, 2, 5, 10, and 25m depth) were stored at –20 °C in acid-cleaned,high-density polyethylene bottles and analyzed as described by Atlaset al. (1971). Size-fractionated chlorophyll-a (Chl-a) concentration wasalso measured within the euphotic zone (same previous depths) and forthree different size classes. For the nanophytoplankton fraction(2.0–20 µm), 125mL of seawater was pre-filtered using a 20 µm Nitexmesh and collected on a 2.0 µm Nuclepore filter. For the picophyto-plankton fraction (0.7–2.0 µm), 125mL of seawater was pre-filteredusing 2.0 µm Nuclepore and collected on a 0.7 µm GF/F glass fiberfilter. The microphytoplankton fraction was obtained by subtracting theprevious two fractions from the total chl-a concentration (collected over0.7 µm GF/F filters). Filters were immediately frozen (−20 °C) untillater fluorometric analysis, using acetone (90% v/v) for pigment ex-traction (Turner Design TD-700) as described by Parsons et al. (1984).The primary production process requires the availability of nutrientsand light as basic resources for phytoplankton growth. Hence, in thisstudy, this parameter was incorporated as an indicator of the resourceutilization rate by the phytoplankton (Marañón et al., 2012). All pri-mary production estimates were obtained using the 14C method(Steeman-Nielsen, 1952) within the euphotic zone (above 25m depth)during 4 h of daylight (earliest at 09:00 h an latest at 15:00 h) at thesame depths described for nutrients and chl-a samples. Two replicatesand two dark bottles were used for every depth. Since all cruises werecarried out during different years (Table 1), satellite-based chlorophyll-a data from 2006 to 2010 were extracted. MODIS-Aqua images (NOAA’sERDDAP platform) using eight-day composite images (http://coastwatch.pfeg.noaa.gov/erddap/griddap/erdMHchla8day.html)were compared between years using chlorophyll-a data of every sam-pling site to check whether the overall conditions between cruises werecomparable or not. Using average chlorophyll-a data (period of Oc-tober-November) at each station and year, a one-way ANOVA analysiswas performed to determine differences between years.

2.3. Data analyses

Although the samples collected on every research cruise were fromthe entire water column, the dataset used in this study was restricted tothe top 25m in order to avoid confounding effects of vertical varia-bility, and to better represent the effects of surface brackish water onphytoplankton productivity and biomass. Correlation analysis was usedto assess the collinearity of temperature, salinity, density, Brunt-Väisäläfrequency, and nutrients. This collinearity was then used to producestatistically independent environmental axes through a principal com-ponent analysis (PCA). Depth-averaged and standardized (log trans-formed) values of temperature, salinity, density, Brunt-Väisälä fre-quency, mineral nutrients, and surface PAR for each station were usedin the PCA. The association between the first two principal componentsof environmental variability and autotrophic biomass (total and size-fractionated) was analyzed via Pearson correlation coefficients. Giventhe extent, geographic complexity, and environmental heterogeneity ofthe Patagonian region (see Aracena et al., 2011; González et al., 2013;Jacob et al., 2014), it was assumed that any relationship betweenchlorophyll and environmental variables was unlikely to exist over theentire region. Therefore, a series of local-scale correlations, each cen-tered on a sampling station, were computed using all the observationsthat fell within a 150 km radius around each station. The radius of150 km for spatial neighborhoods was based on a smallest-sample

Fig. 1. Study area and stations in the Chilean Patagonia. Black circles: CIMAR-12 cruise (2006), white circles: CIMAR-13 (2007), black triangles: CIMAR-14(2008), white triangles: CIMAR-15 (2009), and black squares: CIMAR-16(2010).

L.A. Cuevas, et al. Progress in Oceanography 173 (2019) 103–113

105

criterion. To compute a correlation coefficient around each one of thesampling stations, neighborhoods with at least 10 pairs of values wereconsidered in the computation of correlation coefficients. Only thoseproducing a significant correlation (alpha= 0.05) were considered forgraphical representation. A radius of 150 km was the smallest neigh-borhood that met these requirements. The same procedure to computecorrelations was applied for both principal components and each of thefour biological variables: depth-integrated total chlorophyll-a and totalpercentages that corresponded to micro-, nano- and pico-phyto-plankton. Finally, we tested linear, logarithmic, and polynomial re-gressions to determine the relative importance of chlorophyll-a con-centrations and primary production (as resource use rate) in the controlof phytoplankton size structure. chlorophyll-a concentrations and pri-mary production data were log10-transformed prior to analysis.

3. Results

3.1. Latitudinal variability of physical-chemical data and chlorophyll-a

Annual satellite-based median chlorophyll-a showed similar valuesbetween 2006–2007 and 2008–2009, with a slight decrease with lati-tude (Fig. 2). No significant differences between years were observed (F(4670)= 1.7, p=0.149), which is a sign of low interannual variabilityin the region during the observed years (from 2006 to 2010). Upper andlower quartiles overlapped from 2006 to 2010, where only values from2010 showed the shortest range. The total range of satellite-basedchlorophyll-a overlapped from 2006 to 2008, but showed a lower rangeonly during the spring of 2009 (Fig. 2). Euphotic zone depth-averagedtemperatures decreased from north to south Patagonia during thespring, from 12.3 to 4.7 °C (Fig. 3A), with a stronger latitudinal gradientnear the transition between northern and southern Patagonian IceFields. The lowest values were observed next to the Southern Patago-nian Ice Field (50–50.5°S), and also at the inner stations of glacial fjordsin the Southern Strait of Magellan (∼54.5°S). Stations located withinfjords and channels exhibited the lowest salinity values, which is con-sistent with a stronger influence of freshwater discharges by local rivers(Fig. 3B). Average salinity values below 20 were observed at the Re-loncaví fjord (41.5°S), Aysén fjord (45.5°S), Baker channel (48.0°S), andÚltima Esperanza Sound (51.5°S), whereas oceanic waters were domi-nant at Boca del Guafo (44.0°S), Penas Gulf (47.5°S), and south of theBeagle channel (> 54.5°S). Brunt-Väisälä frequency exhibited a latitu-dinal pattern that mirrored patterns of surface salinity north of ∼52°S,and revealed strong horizontal gradients in vertical stratification, withthe highest values at fjords Reloncaví (middle and head of the fjord),Puyuhuapi (45°S), Aysén, and Baker (Fig. 3C). No latitudinal variabilityin phosphate and nitrate concentrations was observed (Fig. 3D and E),which did not strongly correlate with salinity or density (results notshown), although higher concentrations of both nutrients were

observed at oceanic sites (Fig. 3D and E). Average phosphate and ni-trate above 25m of depth varied from 0.2 to 1.6 μM, and from 0.1 to16.5 μM, respectively, resulting in the euphotic-zone average inorganicN/P ratio of 9.0 (standard error= 0.3) for the entire Patagonian eco-system. Silicic acid concentrations varied widely north of 48°S, andshowed a strong relationship with water column stratification(r= 0.68, p < 0.001) as evidence of freshwater discharges, whereaverage values ranged from 1.3 to 55.2 μM (Fig. 3F). South of 48°S,silicic acid varied over a much narrower range (0–9.9 μM) and wasnegatively correlated with salinity values (r=−0.63, p < 0.001).High values of depth-integrated chl-a (> 200mgm−2) were estimatedfor the Reloncaví fjord, and the gulfs of Ancud and Corcovado innorthern Patagonia. In contrast, low chl-a values (4.0–81.1mgm−2)were observed from 44 to 53°S regardless of mineral nutrient con-centrations, water-column stability, and hydrographic conditions(Fig. 3G). Depth-integrated chl-a increased again south of 53.0°S, fromthe northern section of the Magellan Strait to Cape Horn, with valuesranging from 25.7 to 821.8 mgm−2 (Fig. 3G). Phytoplankton sizestructure varied largely across the region under spring conditions, witha clear dominance of microphytoplankton (> 80%) from 41.5 to 44.0°S,and a more variable composition from 44.0 to 53.0°S, where picophy-toplankton appeared in higher proportion at sites Puyuhuapi Channel,Bakker Channel, and Última Esperanza, with low average salinity (i.e.higher fresh water discharges) (Fig. 3H–J). South of 53.0°S, the pico-phytoplankton contribution was consistently low (< 6%), whereas thenanophytoplankton fraction was highly variable (2.0–72.5%), and themicrophytoplankton fraction had the highest contribution of total chl-aas in the northern section of the Patagonia (59.0–98.0%) although withgreater spatial variability (Fig. 3H–J). Mean surface PAR for the period2006–2010 showed a latitudinal as well as a longitudinal gradient insurface light conditions (Fig. 4A), with dramatic differences betweeninner channels and fjords (< 20 Einstein m−2 s−1) and areas such asthe Inner Sea of Chiloe in northern Patagonia and the east of the Ma-gellan Strait in the south (> 25 Einstein m−2 s−1). This pattern ap-peared even clearer when surface PAR estimates corresponding to thespecific sampling sites and dates were examined (Fig. 4B). High values(> 34 Einstein m−2 s−1) were observed for stations located in thenorthern (42–45°S) and southern sections (53–56°S), whereas lowervalues (< 34 Einstein m−2 s−1) were observed in fjords such asPuyuhuapi and Aysén (45.0 and 45.5°S, respectively). The lowest sur-face PAR values (< 25 Einstein m−2 s−1) were found at stations locatedbetween 49.0 and 52.5°S, on the western side of the Southern Ice Field(Fig. 4B).

3.2. Main control factors on phytoplankton size structure

The first two principal components obtained from a PCA on depth-averaged hydrographic data, nutrient concentrations, and surface PARexplained 63% of the total variance (Table 2). The first principalcomponent (PC1) explained 36.0% of total variability and accountedmostly for the influence of low salinity, low density, and silicic acid richwaters that are discharged from Patagonian rivers. The second principalcomponent (PC2) accounted for 27.3% of total variability, and re-presented the zonal gradient in nutrient concentrations, which in thisregion is manifested mostly through the intrusion of nitrate- andphosphate-rich oceanic waters (Table 2).

The spatial distribution of PCA scores included a large spatial het-erogeneity at the different fjords and channels (Fig. 5A and B). Thus,along the PC1 axis (Fig. 5A), freshwater discharges directly affectedsites at the eastern continental margin (e.g. Reloncaví fjord, PuyuhuapiChannel, Aysén fjord, and Baker Channel). Along the PC2 axis, the PCAordination of physical-chemical conditions indicated a separation be-tween stations located north and south of 48°S (Fig. 5B and C). Non-glacial rivers with high silicon inputs and strong density stratificationnear the surface, and also oceanic waters rich in nitrate and phosphateinfluenced fjords and channels located in the northern section of

Fig. 2. Logarithmic satellite-based chlorophyll-a (μg L−1) extracted fromMODIS-Aqua imagery for all stations from 2006 to 2010. Median (50% of data)is the bold line. Upper and lower quartiles are expressed in the box. Total rangeis represented with lines.

L.A. Cuevas, et al. Progress in Oceanography 173 (2019) 103–113

106

Patagonia (Fig. 5B and C). Pearson correlations computed over the150 km spatial neighborhoods (Fig. 6) indicated that total and size-fractionated chl-a are significantly correlated with both river influence(PC1) and oceanic nutrients (PC2) at a number of sub-regions. How-ever, the extent of these sub-regions and the sign of significant corre-lations change depending on which pair of biological variables andenvironmental features (i.e. principal component) are compared. Forinstance, while the correlation between total chl-a and PC1 was sig-nificantly negative for 43% of the sites influenced by the Southern IceField (48–54°S), as well as along the Gulf of Corcovado and MoraledaChannel (∼44°S), at the northernmost stations, located along the

Reloncaví fjord and the northern section of the Chiloé Inner Sea it waspositive (Fig. 6A). A very similar distribution was found for the mi-crophytoplankton fraction in PC1 (Fig. 6B), which is consistent with thefact that the largest contribution to phytoplankton biomass at 40% ofthe stations corresponded to large-cell taxa. The nanophytoplanktonfraction positively correlated with PC1 at only one site near the Bocadel Guafo area (Fig. 6C). Correlations for picophytoplankton and PC1were almost a mirror image of the correlations found for the micro-phytoplankton fraction. Positive correlations were found for the sta-tions near Boca del Guafo (∼44°S) and those influenced by SouthernIce Field glaciers (50–53°S); correlations were negative for stations lo-cated inside the fjords Reloncaví, Puyuhuapi, Aysén, and Quitralco(Fig. 6D). Spatial correlation patterns between depth-integrated chl-aand the influence of oceanic nutrients (i.e. nitrate and phosphate) re-presented by PC2 (Fig. 5) were clearly different from those found forPC1. There was a latitudinal change in the significant correlations be-tween total chl-a and PC2 (Fig. 6E), suggesting that phytoplanktonbiomass responds negatively to the influence of oceanic nutrients inchannels and fjords located in northern Patagonia (north of 48°S), but

Fig. 3. Latitudinal variability of temperature (°C), salinity, Brunt-Väisälä frequency (s−1), phosphate (PO4, μM), nitrate (NO3, μM), silicic acid (Si(OH)4, μM),integrated chl-a (mgm−2), and proportion of micro-, nano-, and picophytoplankton for the Chilean Patagonia. Latitude corresponds to the northern and southernlimits of each research cruise specified in Table 1.

Fig. 4. Surface PAR (Einstein m−2 s−1) from Level-3 MODIS-Aqua imagery. (A)Mean field for 2006–2010, (B) point values extracted from 8-day compositeimages for sampling sites and dates closest to CIMAR cruises stations.

Table 2Principal component analysis (PCA) on oceanographic and chemical data col-lected for the entire Patagonian ecosystem. Columns show the weights for everyvariable. Weights> 0.4 are shown in boldface.

Variables PC1 PC2 PC3

Temperature (°C) 0.220 0.340 −0.380Salinity −0.441 0.272Density (kg m−3) −0.457 0.241Brunt-Väisälä Frequency (s−1) 0.423 −0.200Surface PAR −0.205 0.166 −0.586Phosphate (μM) 0.542 −0.113Nitrate (μM) 0.149 0.553Silicic acid (μM) 0.472 0.149N/P ratio 0.139 0.294 0.617Si/N ratio 0.241 −0.125 −0.192

% Total variance 36.0 27.3 12.6

L.A. Cuevas, et al. Progress in Oceanography 173 (2019) 103–113

107

Fig. 5. Principal component analysis (PCA) of the N=135 stations according to hydrographic and nutrient conditions as well as surface PAR measured at eachstation. Maps in (a) and (b) show the spatial distribution of PCA loadings for PC1 and PC2, respectively. The colored bar in (b) applies to both maps. (c) Ordination ofstations according to PC1 and PC2, which jointly explained 63% of the total variance (see Table 1), with symbols corresponding to the research cruises specified inFig. 1, and vector labels indicating the individual variables (Temp: temperature, sali: salinity, sigm: sigma-theta (density), BVF: Brunt-Väisälä frequency, par:photosynthetically active radiation, Si: silicic acid, N: nitrate, P: phosphate, N/P: nitrate:phosphate ratio, Si/N: silicic acid:nitrate ratio). (For interpretation of thereferences to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 6. Spatial variability of Pearson correlation between total chlorophyll-a (A, E) and size-fractionated (microphytoplankton: B, F; nanophytoplankton: C, G;picophytoplankton: D, H) and the first two principal components obtained from a PCA of hydrographic and chemical variables in the water column (see Fig. 5c). PC1:A, B, C and D. PC2: E, F, G and H) Red and blue symbols indicate positive and negative correlations, respectively. Only significant correlations are shown. (Forinterpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

L.A. Cuevas, et al. Progress in Oceanography 173 (2019) 103–113

108

positively in those located south of 48°S. Only few sites correlatedsignificantly with the microphytoplankton fraction with PC2 (Fig. 6F),whereas the nanophytoplankton fraction positively correlated withoceanic nutrients east of the southern Patagonian Ice Fields and at somesites of the Strait of Magellan (Fig. 6G). Finally, the picophytoplanktonfraction was positively correlated with oceanic nutrients at the Re-loncaví Fjord (middle and head of the fjord), northern section of theChiloé Inner Sea, and negatively correlated at sites near the BeagleChannel (Fig. 6H).

3.3. Resource supply rate and phytoplankton size structure

The extensive fjord ecosystem that was studied presented highlycomplex hydrographic conditions formed by the interaction of marineand fresh water, where we expected a large horizontal variation in theconcentration of nutrients. Thus, it is possible to observe sharp latitu-dinal and longitudinal gradients in nutrient concentrations even afterdepth-averaging nutrient data from 0 to 25m in depth (Fig. 3D–F).Some of this spatial variability in mineral nutrients can also be observedin the relative contribution of different size fractions to total auto-trophic biomass. The microphytoplankton fraction increases rapidlywith increasing total chl-a, reaching values> 75% when autotrophicbiomass is above 10 µg L−1 (Fig. 7A). A concomitant drop in the

contribution of total biomass was observed in the nanophytoplanktonfraction, with<40% when total chl-a levels were above 10 µg L−1

(Fig. 7B). A larger drop in total chl-a was observed in the picophyto-plankton portion, which tended to dominate when total chlorophyll-awas< 2 µg L−1 but contributed<10% of total biomass when total chl-a concentrations were above 5 µg L−1 (Fig. 7C). These patterns ap-peared to be a consequence of a strong association between size par-titioning and primary production rates (Fig. 8). Primary production isan indicator of the resource utilization rate since phytoplankton re-quires a supply of nutrients and light to grow. The contribution ofmicrophytoplankton increases rapidly with primary productionreaching values above 50% at rates of 50 µg C L−1 d−1, following alogarithmic relationship (Fig. 8A, y=6.712 ln(x)+ 26.819, R2= 0.16,p < 0.01, n=136). In contrast, nano- and picophytoplankton de-crease at higher productivity rates. At low productivity rates, nano-phytoplankton showed a large variation in its relative contribution(0–85%) and its contribution varies from 5 to 45% when primaryproduction increases above 200 µg C L−1 d−1 (Fig. 8B). Picophyto-plankton contribution decreases rapidly through the productivity range(Fig. 8C). At production rates above 100 µg C L−1 d−1 picophyto-plankton contributes only with 2–40% of the total autotrophic biomass.

Fig. 7. Total chlorophyll-a concentration versus the contribution of (A) mi-crophytoplankton, (B) nanophytoplankton, and (C) picophytoplankton.Symbols correspond to each research cruise specified in Fig. 1.

Fig. 8. Primary production versus the contribution of (A) microphytoplankton,(B) nanophytoplankton, and (C) picophytoplankton. Symbols correspond toeach research cruise specified in Fig. 1.

L.A. Cuevas, et al. Progress in Oceanography 173 (2019) 103–113

109

4. Discussion

Our data set was built from observations collected during five re-search cruises (CIMAR 12, 13, 14, 15 and 16) conducted during con-secutive spring seasons (Table 1). A logical concern with this kind ofdataset is how much of the spatial variability we present here is due totemporal changes in the physical-chemical conditions of the euphoticzone, rather than to persistent spatial patterns in hydrography andnutrient supply (Fig. 3). Comparison of satellite-based chlorophyll-afrom 2006 to 2010 (Fig. 2) showed an overlap in the overall conditionsrelated to autotrophic biomass, displaying low temporal variabilitybetween years (from 2006 to 2010). Thus our temporally-fragmentedobservations can be directed to address questions that pertain to thelarge Patagonian ecosystem, representing mainly three aspects: (i) thecontrasting hydrographic conditions along this large estuarine region,(ii) the inputs of mineral nutrients (marine versus terrestrial origin),and (iii) the entire range of productivity/resources available. The widerange of physical-chemical conditions (Fig. 3A–C) highlighted thestrong influence of freshwater discharges (e.g. Reloncaví fjord, Puyu-huapi channel, Aysén fjord, Baker channel, and channels near PuertoNatales) on the water column at different latitudes, adding the es-tuarine variability to our analysis. The entire region also exhibits stronggradients in mineral nutrient concentrations, such as the latitudinaldecrease in nitrate (1 order of magnitude) and silicate (2 orders ofmagnitude) from northern to southern Patagonia (Fig. 3D–F). In thesouthern section (south of the Patagonian Ice Fields) the fluvial influ-ence is considerably lower compared to the glacial influence, affectingthe total input of nutrients and suspended sediment load into the fjordecosystem (Aracena et al., 2011; Dávila et al., 2002), where surfacefjord waters contain low silicic acid levels (Fig. 2F, Torres et al., 2011).Furthermore, a latitudinal trend towards lower primary productionrates in southern Patagonia was also reported (Aracena et al., 2011).

At the ecosystem level, the results of this analysis stress the im-portance of two main estuarine features (river discharges and nutrientsfrom oceanic waters), rather than other physical control factors that arecommon in modulating the phytoplankton size structure of othercoastal systems.

4.1. Impact of fresh water discharges

A clear signal is obtained primarily for the influence of freshwaterinputs (low salinity, low density, strong stratification, and high siliconwaters), and secondarily for the influence of oceanic waters with highconcentrations of nitrate and phosphate (Table 2). The effects of riverdischarges on hydrographic conditions (i.e. stratification of the watercolumn) and silicon availability appear to limit mainly large auto-trophic cells (Fig. 6A), and to favor the small size fraction at sites southof 48°S, and also at Puyuhuapi Channel and Aysén fjord (Fig. 6D). Inthis respect, increased abundance of pico- and nanophytoplankton wasreported for coastal waters strongly influenced by rivers (Iriarte et al.,2018; Jochem, 2003; Liu et al., 2004; Vargas et al., 2012). The centralPatagonia transition at ∼48°S was described as an area with elevatedfreshwater discharges from large rivers such as the Baker and Pascua(1133 and 753m3 s−1, respectively; National Water Directorate, www.dga.cl). Additionally, the glaciers associated to the Northern andSouthern Patagonian Ice Fields influence the content of inorganicmatter (glacial silt) and light attenuation in this region (Montecino andPizarro, 2008; Pizarro et al., 2005). Together with glacial sediments,river-borne suspended material reduce light availability, hence surfacePAR values, south of 48°S (Fig. 4B), and in fjords such as Puyuhuapi andAysén (∼45°S), and rivers such as Baker with a high concentration ofsuspended sediments (González et al., 2013). Consistently, primaryproduction rates that are low compared to areas not influenced byglaciers were reported for this section of the Patagonia (e.g.91 mg Cm−2 d−1 at Caleta Tortel, (Aracena et al., 2011);422mg Cm−2 d−1 near Baker river mouth, (González et al., 2013). Our

results suggest that autotrophic biomass and its composition arestrongly impacted by river discharges at latitudes south of 48°S, andalso at Puyuhuapi Channel and Aysén fjord, with a substantial reduc-tion in depth-integrated chl-a and a mixed phytoplankton size dis-tribution, with cells< 20 μm (nano- and picophytoplankton) re-presenting>50% of total chl-a. A slightly different condition wasobserved at the Sub-Antarctic system of the Magellan Strait, Cape Horn,and the Beagle Channel where river discharges did not have a sig-nificant negative correlation with the chlorophyll-a concentration(Fig. 6A and B), and chl-a ranges were at the same magnitude comparedto northern Patagonian channels and fjords (i.e. Reloncaví fjord, AncudGulf, Corcovado Gulf, Fig. 3G). Although there are numerous glaciersnearby, salinity at the observed sites (surface waters above 25m) wassimilar to other estuarine sites and did not decrease compared to siteswhere river discharges have a strong influence (e.g. Puyuhuapi channel,Baker channel, Última Esperanza, Fig. 3B). Additionally, low silicate/nitrate ratio characterized this area (Fig. 3), and light availability wasnot limited in this area (Fig. 4B), resulting in a chl-a concentration si-milar to previous studies (González et al., 2016; Iriarte et al., 2001).

Hydrographic variability (i.e. temperature and salinity) modulatesautotrophic production by way of seasonal changes in stratification(strongly stratified in summer and weakly stratified in spring and au-tumn), and the balance between fluvial and marine influences on nu-trient conditions. Changes in these two conditions can lead to con-trasting phytoplankton community structures, depending on theavailability of blooming species and turbulence conditions (Smayda,2002). Stratification strongly affects light penetration and the avail-ability of nutrients in surface waters (Goebel et al., 2005). In the stra-tified condition, flagellate forms (i.e. dinoflagellates) or small diatomsmight be favored, however they may be controlled by nutrient inputssuch as nitrate and phosphate (Guo et al., 2014). In our analysis, stra-tified waters (high Brunt-Väisälä frequency, Fig. 2) favored autotrophiccells in the low-size range including mainly picophytoplankton cells(e.g. cyanobacteria, picoeukaryotes). A previous study in the regiondemonstrated that stratified conditions and high silicate levels posi-tively correlated with cells with low surface/volume ratios (Alves-de-Souza et al., 2008). In southern Patagonian fjords, areas with highfreshwater input and strong layering processes favored the formation ofthin layers with high chl-a concentration (Ríos et al., 2016). Thus,during spring conditions, a greater contribution of autotrophic nano-flagellates versus dinoflagellates was observed (González et al., 2011,2010). When stratification weakens, the microphytoplankton con-tribution reaches its highest values. Under this hydrographic and nu-trient scenario, chain-forming diatoms from the genera Thalassiosira,Chaetoceros, and Skeletonema dominates the microphytoplankton bio-mass (González et al., 2011, 2010).

Additionally in our extensive study area, river discharges areusually defined as oligotrophic carrying low nitrate and phosphateloads (Fig. 3D and E). Theoretical and experimental evidence demon-strate that small cells have higher rates of nutrient uptake per unitbiomass and lower half-saturation constants due to their higher surfacearea to volume ratio (Aksnes and Egge, 1991). Furthermore, smallercells have a lower minimum cellular metabolic requirement that se-lectively allows them to survive at lower resource concentrations thanlarger cells (Grover, 1991). Thus lower resource requirements by pico-and nanophytoplankton size cells may support higher population den-sities of these size classes. Consequently, pico- and nano-phytoplanktonappear to have advantages over microphytoplankton under higher riverdischarges that produce nutrient-limited conditions, giving them ap-parent benefits over larger cells, modifying the community size dis-tribution during the productive season analyzed.

4.2. Nutrient rich oceanic waters and secondary control factors

Opposite to the river discharges, the intrusion of oceanic waters richin mineral nutrients (i.e. nitrate and phosphate) promoted higher chl-a

L.A. Cuevas, et al. Progress in Oceanography 173 (2019) 103–113

110

concentrations at some sites in the area southern than 48°S (Fig. 6E). Aswe discussed previously, when nutrients are in low supply, phyto-plankton biomass and production rates are low and the community isdominated by small cells, which due to their large surface-area-to-vo-lume ratio are better prepared to withstand nutrient limitations (Aksnesand Egge, 1991; Kiørboe, 1993). Additionally, limitation of specificnutrients may lead to different autotrophic communities. Fjord watersrich in silicic acid and nitrate lead to a major contribution of large cells,where chain-forming diatoms are typically dominant (González et al.,2010). Silicic acid is important in the formation of diatom silica shells,hence their ratio with nitrate has been hypothesized to affect diatoms(Rabalais et al., 1996). Furthermore, nitrate can be stored in diatomvacuoles giving a high vacuole/cytoplasm ratio (Stolte and Riegman,1995) leading to large phytoplankton cells. The influence of riverineand glacier freshwaters substantially increase dissolved silicon, but notnitrate or phosphate, in the surface waters of these fjord ecosystems(Vargas et al., 2011). This nutrient ratio supports the high proportion ofsmall cells (pico- and nano-phytoplankton fractions) that may bedominated by small autotrophic flagellates and dinoflagellates, togetherwith picoeukaryotes.

Other factors, such as temperature did not show an important in-fluence on phytoplankton size structure as some previous studies hadpredicted. Warmer temperatures were expected to shift phytoplanktonsize structure towards a larger contribution of small cells to total au-totrophic biomass. For the eastern and western North Atlantic Ocean,(Morán et al., 2010) temperature alone explains 73% of the variabilityof picophytoplankton regardless of the trophic status or nutrient load.This pattern is partly explained by the temperature-size rule for protist(Atkinson et al., 2003), which predicts a decrease in the mean cell sizeas temperature increases. Contrastingly, a meta-analysis of tempera-ture, phytoplankton size structure, and productivity rates concludedthat resource availability rather than temperature is the main factorcontrolling the observed phytoplankton size structure (Marañón et al.,2012). Thus, the effects of water temperature and resource availability(e.g. mineral nutrients and productivity rates) must be evaluated se-parately. Our analysis uses a sufficiently large data set making it pos-sible to reject the hypothesis that temperature plays a direct role incontrolling phytoplankton size structure (Table 2, Fig. 5), at least forthis extensive fjord ecosystem.

In addition to estuarine features, our analysis indicated that surfaceradiation limits autotrophic size structure, but to a lesser degree thanriver discharges and oceanic nutrients (Table 2). Unfortunately, fewPAR data were available thus the effect of light in the euphotic zone(above 25m in depth) was poorly evaluated as a controlling factor forphytoplankton size structure compared to other physical and chemicalvariables. However, assuming that our extracted MODIS-Aqua imagesare good estimates of surface PAR values and represent a large part ofthe entire fjord ecosystem studied, the satellite-derived PAR valueswere used as our best estimates of the effect of light on phytoplanktonstructure size. As an example, the mean surface PAR for the 2006–2010period (Fig. 4A) includes not only the latitudinal variability, it alsoincludes the large effect of suspended material from rivers and in-organic matter and sediments from glaciers as elements of light at-tenuation, which has also been observed in previous studies (Marínet al., 2013; Pizarro et al., 2000, 2005). The use of mineral nutrientsand their transformation into algal biomass requires solar radiation asan energy source to drive photosynthesis, thus the importance of mi-neral nutrient loads can be constrained by the availability of sunlightfor phytoplankton (Cloern, 1999). The efficient capture of photons bysmall cells becomes an advantage when light is a limiting factor (Sundaand Huntsman, 1997). It seems that, at least during the spring monthswhen our data were collected, light was a limiting factor in some es-tuarine regions (mainly at glacier-derived freshwater discharges), andthe input of mineral nutrients exerted a stronger controlling effect onphytoplankton. Going one step further into our PCA results, a thirdprincipal component that accounts for 13% of total variability (76% of

the total accumulative variability) revealed an important control oversurface PAR along with N:P ratios (Table 2). Thus, the combined effectof light limitation and mineral nutrient ratios potentially play a role indetermining the size distribution of the autotrophic community in thistype of ecosystem. A likely example might be the Strait of Magellan area(south of 53°S), where the river influence and nutrient rich oceanicwaters are expected to limit chl-a concentration and micro-phytoplankton abundance. However, the high depth-integrated chl-aconcentration and the fraction of microphytoplankton are surprising(Fig. 3G). Previous studies (e.g. Iriarte et al., 2001) showed thatspringtime phytoplankton in the Strait of Magellan are dominated bylarge autotrophic cells (chain forming diatoms and thecate dino-flagellates). A potential explanation for the observed phytoplanktonsize structure in this area is the low silicate/nitrate ratio (Fig. 3) andhigh surface PAR observed at this latitude (Fig. 4B). Jacob et al. (2014)also reported light limitation in Patagonian fjords, especially for semi-enclosed coastal waters (influenced by glacial melting) but not forcontinental waters.

4.3. Consequences of small-cell versus large-cell dominance

The relation between increased resource utilization rate (i.e. pri-mary production) and a greater dominance of large cells is a commonfeature in our analysis (Figs. 7 and 8). Thus, large cells contribute tomore than 75% of the autotrophic biomass in high productive waters(Fig. 7A). Irigoien et al. (2005) showed the importance of the relationbetween cell size and growth rates or growing conditions, thus acombination of controlling factors might modulate the resultingdominant phytoplankton size distribution. Stratified waters with non-limiting conditions of silicic acid, nitrate, and phosphate lead to highautotrophic biomass dominated by large cells, as observed from theReloncaví fjord to Boca del Guafo (Fig. 3), resulting in high primaryproductivity rates (González et al., 2010), and increasing the potentialfor carbon sequestration. Our results also reveal that not only largeautotrophic cells (> 20 μm) dominate the euphotic zone during theproductive season (i.e. spring), but also small cells represent a highproportion of the total phytoplankton biomass (Fig. 3I and J), with lowproductivity rates (Fig. 8B and C). The dominance of small nano- andpicophytoplanktonic groups leads to a microbial-type food web thatpartitions the autotrophic autochthonous organic matter, which cannotbe transferred to higher trophic levels, supporting the recycling of or-ganic matter. The efficiency of the biological pump, i.e. how the organicmatter is sequestered in the ocean if microbial or the classical food webtype dominate (Legendre and Le Fevre, 1995), can present largevariability in Patagonian fjords. Thus, the shifting balance betweenhydrography and resource control phytoplankton size structure mayhave important implications in the region.

Other factors such as grazing (i.e. top-down control, not included inour analysis), may have a major role in structuring the phytoplanktoncommunity in aquatic ecosystems such the Patagonian fjords.Nevertheless, previous studies, which were carried out during the sameresearch cruises used in this analysis (Table 1), have estimated that alow proportion of the primary productivity can be consumed by het-erotrophic plankton in this study area. In the northern area(41.5–44.0°S) less than 2% of the primary productivity can be con-sumed by euphausiids and calanoid copepods (González et al., 2010).Similar evidence found in the Aysén Fjord (45.3°S) indicates that het-erotrophic nanoplankton consumes less than 2% of the total primaryproductivity, where heterotrophic flagellates consume 1.1% and otherheterotrophs (i.e. ciliates, copepods and euphausiids) reach even lowervalues (González et al., 2011). In river-influenced areas, such as Bakerand Pascua rivers (47.0–48.0°S), heterotrophs consume 33% of theprimary productivity, where heterotrophic flagellates and small cala-noid copepods each have an impact of 14% in relation to primaryproductivity (González et al., 2013). Far south (52–56°S), Euphausiavallentini consumed 3% of the gross primary production in sites near the

L.A. Cuevas, et al. Progress in Oceanography 173 (2019) 103–113

111

Magellan Strait and Beagle Channel (González et al., 2016). Mixo-trophic flagellates can also be important in fjord ecosystems (Havskumand Riemann, 1996) and may represent an important control factor onsmall autotrophs. In fact, in northern Patagonian fjords, mixotrophicflagellates were reported as a dominant trophic group showing a strongcontrol during winter conditions (Czypionka et al., 2011), but notduring the productive season. Even though other potential grazers (e.g.larval invertebrates, picoheterotrophs and others) were not consideredin those estimates, top-down control does not seem to exceed more than50% of the autotrophic primary production, contrary to previous re-views from different ecosystems (e.g. Calbet and Landry, 2004).Therefore, the previous evidence seems to indicate that hydrographyand resource supply variables (i.e. inorganic nutrients, light avail-ability) are significant controlling factors for phytoplankton pro-ductivity and community structure in the Patagonian fjord ecosystem.

Elucidating the respective roles of control factors such as hydro-graphic variability and resource supplies (e.g. mineral nutrient loadsand light availability in the euphotic zone) are important for predictingthe response of pelagic ecosystems to local and global changes. Highlatitude ecosystems, such as fjords and channels in the ChileanPatagonia are projected to experience some of the largest climatechanges in the coming decades (IPCC, 2014). Thus environmental im-pact from climate change (i.e. warming, ice melting, and glacial retreat)and local/anthropogenic stressors (e.g. growing human population,fish/mussels farming) expected for this area (Iriarte et al., 2010; Iriarte,2018) will modify the hydrography, nutrient rates and light availabilityaffecting phytoplankton and the subsequent biological productivity. Itis expected that environmental changes will affect the basic function-ality of coastal ecosystems, thus understanding the effects of changes inresource availability will be central to predicting future alterations inthe phytoplankton size structure and its related processes.

Acknowledgements

We would like to thank the captains and crew of the AGOR VidalGormáz and BI Abate Molina for their invaluable support during thesampling program of each CIMAR cruise involved in this study. Theeffort and highly professional work of several technicians and studentsduring sampling and laboratory analyses are acknowledged. Data setcompilation, data analyzes, and manuscript preparations were fundedby FONDECYT project N° 3120089. Part of the data was analyzed in theframe of the FONDAP IDEAL 15150003 - CONICYT. L.A.C., J.L.I. andH.E.G. were also supported by European Commission (OCEAN-CERTAIN, FP7-ENV-2013-6.1-1, N° 603773). Partial funding for FJTwas provided by COPAS Sur-Austral (CONICYT PIA APOYO CCTEAFB170006) and INCAR (CONICYT FONDAP grant 15110027), and toJ.L.I. L.A.C by CONA Grant N° CONA-C15F 09-12. L.A.C and C.A.V.were supported by Millennium Nucleus Project MUSELS funded byMINECON NC1200286. Additionally, C.A.V was also supported by theMillennium Institute of Oceanography (IMO), MINECON IC120019.

Appendix A. Supplementary material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.pocean.2019.02.012.

References

Acha, E.M., Mianzan, H.W., Guerrero, R.A., Favero, M., Bava, J., 2004. Marine fronts atthe continental shelves of austral South America: physical and ecological processes. J.Mar. Syst. 44, 83–105. https://doi.org/10.1016/j.jmarsys.2003.09.005.

Agawin, N.S., Duarte, C.M., Agusti, S., 2000. Nutrient and temperature control of thecontribution of picoplankton to phytoplankton biomass and production. Limnol.Oceanogr. 45, 591–600.

Aiken, C.M., 2008. Barotropic tides of the Chilean Inland Sea and their sensitivity to basingeometry. J. Geophys. Res. Oceans 113https://doi.org/10.1029/2007JC004593. n/a–n/a.

Aiken, C.M., 2012. Seasonal thermal structure and exchange in Baker Channel, Chile.

Dyn. Atmospheres Oceans 58, 1–19. https://doi.org/10.1016/j.dynatmoce.2012.07.001.

Aksnes, D.L., Egge, J.K., 1991. A theoretical model for nutrient uptake in phytoplankton.Mar. Ecol. Prog. Ser. 70, 65–72. https://doi.org/10.3354/meps070065.

Alves-de-Souza, C., González, M.T., Iriarte, J.L., 2008. Functional groups in marinephytoplankton assemblages dominated by diatoms in fjords of southern Chile. J.Plankton Res. 30, 1233–1243. https://doi.org/10.1093/plankt/fbn079.

Aracena, C., Lange, C.B., Iriarte, J.L., Rebolledo, L., Pantoja, S., 2011. Latitudinal patternsof export production recorded in surface sediments of the Chilean Patagonian fjords(41–55°S) as a response to water column productivity. Cont. Shelf Res. 31, 340–355.https://doi.org/10.1016/j.csr.2010.08.008.

Atkinson, D., Ciotti, B.J., Montagnes, D.J.S., 2003. Protists decrease in size linearly withtemperature: ca. 2.5% C-1. Proc. R. Soc. B Biol. Sci. 270, 2605–2611. https://doi.org/10.1098/rspb.2003.2538.

Atlas, E.S., Hager, S., Gordon, L., Park, P., 1971. A Practical Manual for Use of theTechnicon Autoanalyser in Sea Water Nutrient Analyses (Technical Report). OregonState University, Department of Oceanography.

Calbet, A., Landry, M.R., 2004. Phytoplankton growth, microzooplankton grazing, andcarbon cycling in marine systems. Limnol. Oceanogr. 49, 51–57. https://doi.org/10.4319/lo.2004.49.1.0051.

Calvete, C., Sobarzo, M., 2011. Quantification of the surface brackish water layer andfrontal zones in southern Chilean fjords between Boca del Guafo (43°30′S) and EsteroElefantes (46°30′S). Cont. Shelf Res. 31, 162–171. https://doi.org/10.1016/j.csr.2010.09.013.

Cermeño, P., Marañón, E., Rodríguez, J., Fernández, E., 2005. Large-sized phytoplanktonsustain higher carbon-specific photosynthesis than smaller cells in a coastal eutrophicecosystem. Mar. Ecol. Prog. Ser. 297, 51–60.

Chen, B., Liu, H., 2010. Relationships between phytoplankton growth and cell size insurface oceans: interactive effects of temperature, nutrients, and grazing. Limnol.Oceanogr. 55, 965–972. https://doi.org/10.4319/lo.2010.55.3.0965.

Chisholm, S.W., 1992. Phytoplankton size. In: Primary Productivity and BiogeochemicalCycles in the Sea. Plenum Press, New York, pp. 213–237.

Cloern, J.E., 1999. The relative importance of light and nutrient limitation of phyto-plankton growth: a simple index of coastal ecosystem sensitivity to nutrient enrich-ment. Aquat. Ecol. 33, 3–15. https://doi.org/10.1023/A:1009952125558.

Czypionka, T., Vargas, C.A., Silva, N., Daneri, G., González, H.E., Iriarte, J.L., 2011.Importance of mixotrophic nanoplankton in Aysén Fjord (Southern Chile) duringaustral winter. Cont. Shelf Res. 31, 216–224. https://doi.org/10.1016/j.csr.2010.06.014.

Dávila, P.M., Figueroa, D., Müller, E., 2002. Freshwater input into the coastal ocean andits relation with the salinity distribution off austral Chile (35–55°S). Cont. Shelf Res.22, 521–534. https://doi.org/10.1016/S0278-4343(01)00072-3.

Falkowski, P.G., Oliver, M.J., 2007. Mix and match: how climate selects phytoplankton.Nat. Rev. Microbiol. 5, 813–819. https://doi.org/10.1038/nrmicro1751.

Garreaud, R.D., Lopez, P., Minvielle, M., Rojas, M., 2013. Large-scale control on thePatagonian climate. J. Climate 26, 215–230. https://doi.org/10.1175/JCLI-D-12-00001.1.

Goebel, N.L., Wing, S.R., Boyd, P.W., 2005. A mechanism for onset of diatom blooms in afjord with persistent salinity stratification. Estuar. Coast. Shelf Sci. 64, 546–560.https://doi.org/10.1016/j.ecss.2005.03.015.

González, H.E., Calderón, M.J., Castro, L., Clement, A., Cuevas, L.A., Daneri, G., Iriarte,J.L., Lizrraga, L., Martinez, R., Menschel, E., Silva, N., Carrasco, C., Valenzuela, C.,Vargas, C.A., Molinet, C., 2010. Primary production and plankton dynamics in theReloncaví Fjord and the Interior Sea of Chiloé, Northern Patagonia, Chile. Mar. Ecol.Prog. Ser. 402, 13–30. https://doi.org/10.3354/meps08360.

González, H.E., Castro, L., Daneri, G., Iriarte, J.L., Silva, N., Tapia, F., Teca, E., Vargas,C.A., 2013. Land–ocean gradient in haline stratification and its effects on planktondynamics and trophic carbon fluxes in Chilean Patagonian fjords (47–50°S). Prog.Oceanogr. 119, 32–47. https://doi.org/10.1016/j.pocean.2013.06.003.

González, H.E., Castro, L., Daneri, G., Iriarte, J.L., Silva, N., Vargas, C.A., Giesecke, R.,Sánchez, N., 2011. Seasonal plankton variability in Chilean Patagonia fjords: Carbonflow through the pelagic food web of Aysen Fjord and plankton dynamics in theMoraleda Channel basin. Cont. Shelf Res. 31, 225–243. https://doi.org/10.1016/j.csr.2010.08.010.

González, H.E., Graeve, M., Kattner, G., Silva, N., Castro, L., Iriarte, J.L., Osmán, L.,Daneri, G., Vargas, C.A., 2016. Carbon flow through the pelagic food web in southernChilean Patagonia: relevance of Euphausia vallentini as a key species. Mar. Ecol.Prog. Ser. 557, 91–110. https://doi.org/10.3354/meps11826.

Guo, S., Feng, Y., Wang, L., Dai, M., Liu, Z., Bai, Y., Sun, J., 2014. Seasonal variation inthe phytoplankton community of a continental-shelf sea: the East China Sea. Mar.Ecol. Prog. Ser. 516, 103–126. https://doi.org/10.3354/meps10952.

Gregg, W.W., Casey, N.W., McClain, C.R., 2005. Recent trends in global ocean chlor-ophyll. Geophys. Res. Lett. 32. https://doi.org/10.1029/2004GL021808.

Grover, James P., 1991. Resource competition in a variable environment: phytoplanktongrowing according to the variable-internal-stores model. Am. Nat. 138, 811–835.

Havskum, H., Riemann, B., 1996. Ecological importance of bacterivorous, pigmentedflagellates (mixotrophs) in the Bay of Aarhus, Denmark. Mar. Ecol. Prog. Ser. 137,251–263. https://doi.org/10.3354/meps137251.

IPCC, 2014. Climate change 2014: synthesis report. In: Pachauri, R.K., Meyer, L.A. (Eds.),Contribution of Working groups I, II and III to the Fifth Assessment Report of theIntergovernmental Panel on Climate Change. IPCC, Geneva, pp. 151.

Iriarte, J.L., 2018. Natural and human influences on marine processes in Patagoniansubantarctic coastal waters. Front. Mar. Sci. 5, 360. https://doi.org/10.3389/fmars.2018.00360.

Iriarte, J.L., Cuevas, L.A., Cornejo, F., Silva, N., González, H.E., Castro, L., Montero, P.,Vargas, C.A., Daneri, G., 2018. Low spring primary production and microplankton

L.A. Cuevas, et al. Progress in Oceanography 173 (2019) 103–113

112

carbon biomass in Sub-Antarctic Patagonian channels and fjords (50–53°S). Arctic,Antarctic, Alpine Res. 50 (1), e1525186. https://doi.org/10.1080/15230430.2018.1525186.

Iriarte, J.L., González, H.E., Liu, K.-K., Rivas, C., Valenzuela, C., 2007. Spatial and tem-poral variability of chlorophyll and primary productivity in surface waters ofsouthern Chile (41.5–43° S). Estuar. Coast. Shelf Sci. 74, 471–480. https://doi.org/10.1016/j.ecss.2007.05.015.

Iriarte, J.L., González, H.E., Nahuelhual, L., 2010. Patagonian fjord ecosystems insouthern Chile as a highly vulnerable region: problems and needs. AMBIO 39,463–466. https://doi.org/10.1007/s13280-010-0049-9.

Iriarte, J.L., Kusch, A., Osses, J., Ruiz, M., 2001. Phytoplankton biomass in the sub-Antarctic area of the Straits of Magellan (53 S), Chile during spring-summer 1997/1998. Polar Biol. 24, 154–162.

Iriarte, J.L., Pantoja, S., González, H.E., Silva, G., Paves, H., Labbé, P., Rebolledo, L.,Ardelan, M., Häussermann, V., 2012. Assessing the micro-phytoplankton response tonitrate in Comau Fjord (42°S) in Patagonia (Chile), using a microcosms approach.Environ. Monit. Assess. 185, 5055–5070. https://doi.org/10.1007/s10661-012-2925-1.

Iriarte, J.L., van Ardelan, M., Cuevas, L.A., González, H., Sánchez, N., Myklestad, S.M.,2014. Size-spectrum based differential response of phytoplankton to nutrient andiron-organic matter combinations in microcosm experiments in a Chilean Patagonianfjord. Phycol. Res. 62, 136–146. https://doi.org/10.1111/pre.12050.

Irigoien, X., Flynn, K.J., Harris, R.P., 2005. Phytoplankton blooms: a “loophole” in mi-crozooplankton grazing impact? J. Plankton Res. 27, 313–321. https://doi.org/10.1093/plankt/fbi011.

Jacob, B.G., Tapia, F.J., Daneri, G., Iriarte, J.L., Montero, P., Sobarzo, M., Quiñones, R.A.,2014. Springtime size-fractioned primary production across hydrographic and PAR-light gradients in Chilean Patagonia (41–50°S). Prog. Ocean. 129, 75–84. https://doi.org/10.1016/j.pocean.2014.08.003.

Jochem, F.J., 2003. Photo- and heterotrophic pico- and nanoplankton in the MississippiRiver plume: distribution and grazing activity. J. Plankton Res. 25, 1201–1214.https://doi.org/10.1093/plankt/fbg087.

Kilian, R., Lamy, F., Arz, H., 2013. Late Quaternary variations of the southern westerlywind belt and its influences on aquatic ecosystems and glacier extend within thesouthern-most Andes. German J. Geosci. 164, 279–294. https://doi.org/10.1127/1860-1804/2013/0027.

Kiørboe, T., 1993. Turbulence, phytoplankton cell size, and the structure of pelagic foodwebs. In: Blaxter, J.H.S., Southward, A.J. (Eds.), Advances in Marine Biology.Academic Press, pp. 1–72.

Lara, C., Saldías, G.S., Tapia, F.J., Iriarte, J.L., Broitman, B., 2016. Interannual variabilityin temporal patterns of chlorophyll-a and their potential influence on the supply ofmussel larvae to inner waters in northern Patagonia (41–44°S). J. Mar. Syst. 155,11–18. https://doi.org/10.1016/j.jmarsys.2015.10.010.

Legendre, L., Le Fevre, J., 1995. Microbial food webs and the export of bioge nic carbon inoceans. Aquat. Microb. Ecol. 9, 69–77.

Liu, H., Dagg, M., Campbell, L., Urban-Rich, J., 2004. Picophytoplankton and bacter-ioplankton in the Mississippi River plume and its adjacent waters. Estuaries 27,147–156. https://doi.org/10.1007/BF02803568.

Marañón, E., Cermeño, P., Latasa, M., Tadonléké, R.D., 2012. Temperature, resources,and phytoplankton size structure in the ocean. Limnol. Oceanogr. 57, 1266–1278.https://doi.org/10.4319/lo.2012.57.5.1266.

Marañón, E., Cermeño, P., Rodríguez, J., Zubkov, M.V., Harris, R.P., 2007. Scaling ofphytoplankton photosynthesis and cell size in the ocean. Limnol. Oceanogr. 52 (2),2190–2198. https://doi.org/10.4319/lo.2007.52.5.2190.

Marañón, E., Holligan, P.M., Barciela, R., González, N., Mouriño, B., Pazó, M.J., Varela,M., 2001. Patterns of phytoplankton size structure and productivity in contrastingopen-ocean environments. Mar. Ecol. Prog. Ser. 216, 43–56. https://doi.org/10.3354/meps216043.

Marín, V.H., Tironi, A., Paredes, M.A., Contreras, M., 2013. Modelin suspended solids in aNorthern Chilean Patagonian glacier-fed fjord: GLOF scenarios under climate changeconditions. Ecol. Model. 264, 7–16. https://doi.org/10.1016/jecolmodel.2012.06.017.

McQueen, D.J., Johannes, M.R.S., Post, J.R., Stewart, T.J., Lean, D.R.S., 1989. Bottom-upand top-down impacts on freshwater pelagic community structure. Ecol. Monogr. 59,289. https://doi.org/10.2307/1942603.

Montecino, V., Pizarro, G., 2008. Productividad primaria, biomasa y tamaño del fito-plancton en canales y fiordos australes: patrones primavera-verano. In: Avances En ElConocimiento Oceanográfico de Las Aguas Interiores Chilenas, Puerto Montt a Cabode Hornos. Comité Oceanográfico Nacional - Pontificia Universidad Católica deValparaíso, Valparaíso, pp. 93–97.

Morán, X.A.G., López-Urrutia, Ángel, Calvo-Díaz, A., Li, W.K.W., 2010. Increasing im-portance of small phytoplankton in a warmer ocean. Glob. Change Biol. 16,1137–1144. https://doi.org/10.1111/j.1365-2486.2009.01960.x.

Palma, S., Silva, N., 2004. Distribution of siphonophores, chaetognaths, euphausiids andoceanographic conditions in the fjords and channels of southern Chile. Deep Sea Res.Part II Top. Stud. Oceanogr. 51, 513–535. https://doi.org/10.1016/j.dsr2.2004.05.001.

Paredes, M.A., Montecino, V., 2011. Size diversity as an expression of phytoplanktoncommunity structure and the identification of its patterns on the scale of fjords andchannels. Cont. Shelf Res. 31, 272–281. https://doi.org/10.1016/j.csr.2010.07.012.

Parsons, T.R., Maita, Y., Lalli, C.M., 1984. Counting, media and preservation. In: AManual of Chemical and Biological Methods for Seawater Analysis, pp. 184.

Pickard, G.L., 1971. Some physical oceanographic features of inlets of Chiloe. J. Fish. Res.Board Can. 28, 1077–1106.

Pizarro, G., Iriarte, J.L., Montecino, V., Blanco, J.L., Guzmán, L., 2000. Distribución de labiomasa fitoplanctónica y productividad primaria máxima de fiordos y canales aus-trales (47–50°S) en Octubre 1996. Cienc. Tecnol. Mar. 23, 25–48.

Pizarro, G., Astoreca, R., Montecino, V., Paredes, V.A., Alarcón, G., Uribe, P., Guzmán, L.,2005. Patrones espaciales de la abundancia de la clorofila, su relación con la pro-ductividad primaria, y la estructura de tamaños del fitoplancton en Julio yNoviembre de 2001 en la región de Aysén (43°-46°S). Cienc. Tecnol. Mar. 28 (2),27–42.

Rabalais, N.N., Turner, R.E., Justic', D., Dortch, Q., Wiseman, W.J., Sen Gupta, B.K., 1996.Nutrient changes in the Mississippi River and system responses on adjacent con-tinental shelf. Estuaries 19 (2), 386–407. https://doi.org/10.2307/1352458.

Ríos, F., Kilian, R., Mutschke, E., 2016. Chlorophyll-a thin layers in the Magellan fjordsystem: the role of the water column stratification. Cont. Shelf Res. 124, 1–12.https://doi.org/10.1016/j.csr.2016.04.011.

Saldías, G.S., Sobarzo, M., Quiñones, R., 2018. Freshwater structure and its seasonalvariability off western Patagonia. Progr. Oceanogr. https://doi.org/10.1016/j.pocean.2018.10.014.

Silva, N., Palma, S., 2008. The CIMAR Program in the austral Chilean channels and fjords.In: Progress in the Oceanographic Knowledge of Chilean Interior Waters, from PuertoMontt to Cape Horn. Comité Oceanográfico Nacional - Pontificia UniversidadCatólica de Valparaíso, Valparaíso, pp. 11–15.

Smayda, T.J., 2002. Turbulence, watermass stratification and harmful algal blooms: analternative view and frontal zones as “pelagic seed banks”. Harmful Algae 1, 95–112.

Staehr, P.A., Sand-Jensen, K., 2006. Seasonal changes in temperature and nutrient controlof photosynthesis, respiration and growth of natural phytoplankton communities.Freshw. Biol. 51, 249–262. https://doi.org/10.1111/j.1365-2427.2005.01490.x.

Steeman-Nielsen, E., 1952. The use of radiocarbon (14C) for measuring organic pro-duction in the sea. J. Cons. Perm Intern. Explore Mer. 18, 117–140.

Stolte, W., Riegman, R., 1995. Effect of phytoplankton cell size on transient-state nitrateand ammonium uptake kinetics. Microbiology 141, 1221–1229.

Sunda, W.G., Huntsman, S.A., 1997. Interrelated influence of iron, light and cell size onmarine phytoplankton growth. Nature 390, 389–392. https://doi.org/10.1038/37093.

Torres, R., Pantoja, S., Harada, N., González, H.E., Daneri, G., Frangopulos, M., Rutllant,J.A., Duarte, C.M., Rúiz-Halpern, S., Mayol, E., Fukasawa, M., 2011. Air-sea CO2

fluxes along the coast of Chile: from CO2 outgassing in central northern upwellingwaters to CO2 uptake in southern Patagonian fjords. J. Geophys. Res. 116. https://doi.org/10.1029/2010JC006344.

Torres, R., Silva, N., Reid, B., Frangopulos, M., 2014. Silicic acid enrichment of sub-antartic surface water from continental inputs along the Patagonian archipielagointerior sea (41–56°S). Prog. Oceanogr. 129, 50–61. https://doi.org/10.1016/j.pocean.2014.09.008.

Turner, J.T., 2002. Zooplankton fecal pellets, marine snow and sinking phytoplanktonblooms. Aquat. Microb. Ecol. 27, 57–102. https://doi.org/10.3354/ame027057.

UNESCO, 1981. Background papers and supporting data on the International Equation ofState of Seawater 1980 (No. 38), UNESCO Technical Papers in Marine Science.Division of Marine Science, UNESCO, Paris, France.

Vargas, C.A., Contreras, P.Y., Iriarte, J.L., 2012. Relative importance of phototrophic,heterotrophic, and mixotrophic nanoflagellates in the microbial food web of a river-influenced coastal upwelling area. Aquat. Microb. Ecol. 65, 233–248. https://doi.org/10.3354/ame01551.

Vargas, C.A., Martinez, R.A., San Martin, V., Aguayo, M., Silva, N., Torres, R., 2011.Allochthonous subsidies of organic matter across a lake–river–fjord landscape in theChilean Patagonia: implications for marine zooplankton in inner fjord areas. Cont.Shelf Res. 31, 187–201. https://doi.org/10.1016/j.csr.2010.06.016.

L.A. Cuevas, et al. Progress in Oceanography 173 (2019) 103–113

113