Influence of coastal upwelling on micro-phytoplankton variability at Valparaíso Bay (~33oS), Central Chile

In this work 10 years of data (1986-1996) from a fixed station located in the northern part of Valparaíso Bay (33°00’S; 71°35’W) were analysed to study the influence of coastal upwelling activity on the temporal variation of micro-phytoplankton (20-200 μm) and their relationship with oceanographic conditions. The upwelling activity at the bay was associated to semi-annual wind regime with an intensification of upwelling-favourable S-SW winds from September to March followed by a decrease and the occurrence of downwelling events from April to August. Oceanographic conditions showed the ascent of cold, nutrient-rich salty water in spring (September-November). However, during summertime under highest upwelling index, thermal stratification conditions were registered. This stratification might be associated to either the solar radiation or the presence of an upwelling shadow area in the bay. The upwelling period had the highest micro-phytoplankton abundance mainly dominated by diatoms. This period was associated with an increase in biomass and richness in the bay. Meanwhile during non-upwelling period —under homogenous conditions of temperature, salinity and nutrients— an increase in diversity (but low abundance and richness) associated to dinoflagellates and silicoflagellates was noted. Therefore, the results suggest the presence of a bi-modal regime of micro-phytoplankton in the bay in response to changes in oceanographic conditions related to local wind forcing and mixing/stratification.


Introduction
The central-south Chilean coast (~30º-40ºS) is a highly dynamic and biological productivity area supported by the upwelling activity (Thiel et al. 2007, Escribano & Morales 2012. The phytoplankton in this coastal zone is characterized by the dominance of the micro-phytoplankton (20-200 μm) under upwelling active conditions, typical in spring-summer periods (Anabalón et al. 2007). In several bays along this region (Coliumo, Quintay and Valparaíso), in situ observations have shown a seasonal pattern of phytoplankton biomass (Chl-a) and cell abundance. These patterns are often coupled with seasonal wind-driven upwelling, mixing/stratification processes and solar radiation (Avaria 1971, Avaria et al. 1989, Anabalón et al. 2007, 2016González et al. 2007, Correa-Ramirez et al. 2012, Corredor-Acosta et al. 2015. The intensification of wind-driven upwelling usually observed in spring-summer (September-February) is considered a crucial factor for phytoplankton growth. During these months, upwelled nutrient-rich water fertilizes the photic layer in response to favourable S-SW wind intensification (Shaffer et al. 1999, Sobarzo et al. 2007, Pinochet et al. 2019. This triggers the phytoplankton response via an increment in biomass and cell abundance and impulse the primary production (PP) (González et al. 2007, Testa et al. 2018. detected changes in micro-phytoplankton related to the hydrological conditions on the inter-annual scale with an increment in dinoflagellate abundance after high river discharge events (Aparicio-Rizzo & Masotti 2019).
Studies of micro-phytoplankton in this bay have mostly focused on descriptive phytoplankton species composition (Avaria 1971(Avaria , 1975Avaria & Orellana 1975). Two types of phytoplankton community have been described: 1) a spring/ summer community characterized by frequent bloom events in chain-forming colonies of a few dominant diatom species with high abundance; and 2) an autumn/winter community composed of heterogeneous phytoplankton with low abundance (Avaria 1971(Avaria , 1975Avaria & Orellana 1975). However, physical-chemical oceanographic conditions have never been study at this bay and far from this the relationships between these conditions and the microphytoplankton variability remain unknown in this area. Thus, the aim of this study was to characterize the annual variation of micro-phytoplankton and their interaction with meteorological (wind), physical (temperature, salinity, UI), and chemical (inorganic nutrients) variables in Valparaíso Bay using an extended time series of in situ data (10 years).

Study area
The study area is located in Valparaíso Bay (~32º55'-33º30'S) -a N-NE open embayment flanked by two headlands (Fig. 1). A bio-oceanographic fixed station was located two nautical miles (~3.7 km) from the coast at the north of the bay (St. M; 32º58'2"S; 71º35'02"W); and a depth of ~90 m. Sampling cruises were carried out at this fixed station from October 1986 to December 1996 with intervals of 15 to 30 days. However, no data were registered in February along the time-series.
In this region, the seasonality of micro-phytoplankton has been characterized by a peak of diatoms in the spring/ summer followed by dinoflagellates in the austral autumn and succeeded by high levels of flagellates in winter (Montecino et al. 2006, Anabalón et al. 2007, González et al. 2007. This seasonality in the micro-phytoplankton community has been related to changes in environmental factors (temperature, nutrient concentration and light availability). Small flagellates are usually efficient at acquiring nutrients under low nutrient and stratified conditions. Meanwhile, high nutrient concentrations and turbulence can sustain large phytoplankton cells that are better adapted to these conditions than the smaller ones (Margalef 1978, Smayda 1998.
The relationship between seasonal upwelling activity and phytoplankton response has been characterized by a bimodal regime of upwelling versus non-upwelling conditions along the coastal area of Central Chile (Anabalón et al. 2007(Anabalón et al. , 2016González et al. 2007). First, successive upwelling events induce an increase in phytoplankton biomass and promote diatom populations including chain-forming diatoms in spring/summer under mixing and eutrophication conditions (high nutrient content). Second, dinoflagellates are favoured under stratification (non-upwelling) and oligotrophic (low nutrient content) conditions during the autumn/winter period (Anabalón et al. 2007, González et al. 2007). However, this seasonal pattern has showed temporal and spatial heterogeneity related to episodic wind-forcing regimes and other hydrographic factors (Thiel et al. 2007, Anabalón et al. 2016, Testa et al. 2018. Changes in coastal upwelling systems can impact the phytoplankton community. These changes are related to wind-driven upwelling activity that promotes the arrival of nutrients to the surface. However, nutrients from rivers in Central Chile can also impact the phytoplankton (Iriarte et al. 2012, Anabalón et al. 2016, Aparicio-Rizzo & Masotti 2019. Seasonal freshwater discharge can impact coastal systems during weak upwelling activity periods when nutrient input from rivers can help maintain basal phytoplankton biomass (Chl-a) and PP (Léniz et al. 2012, Masotti et al. 2018, Testa et al. 2018. In these areas, with quasi-constant high nutrient concentrations year-round, changes in phytoplankton are not just linked to nutrients availability but also other factors (temperature, salinity, stratification/mixing conditions, or solar radiation).
In Central Chile, the coastal area off Valparaíso Bay has been historically defined by quasi-permanent S-SW upwelling-favourable winds (>50%) with no seasonal pattern in upwelling activity, and the stability of the water column connected to solar radiation (Pizarro 1973(Pizarro , 1976Avaria et al. 1989). Furthermore, recent studies have Daily data of Aconcagua river flow were collected from the Chilean Dirección General de Aguas (DGA) 2 at the San Felipe station (32º55'09"S; 71º30'28"W) ( Fig.  1, St. F). At the bio-oceanographic fixed station discrete water samples via a Niskin bottle were collected in the upper 60 m (0, 5, 15, 25, 40 and 55 m depth) to study the variability of temperature, salinity, and nutrients (Nitrate-NO 3 -, Phosphate-PO 4 3-, and Silicate-SiO 4 4-). Temperature (ºC) was measured in situ via protected reverse-calibrated thermometers arranged in pairs in the sampling bottle. A Tsurumi Seiki digital induction salinometer measured salinity. The Brunt-Väisälä frequency was obtained as a derivate variable using the Ocean Data View (ODV) program (Schlitzer 2019). Nutrient concentrations (μM) were determined via spectrophotometry. PO 4 3was measured following Koroleff (1983). The Grasshoff methodology (Grasshoff 1983) was employed to measure NO 3 -, and SiO 4 4according to Parsons et al. (1984).
The micro-phytoplankton (20-200 μm) was binned into three main classes: diatoms (Bacillariophyceae), dinoflagellates (Dinophyceae), and silicoflagellates (Dictyochophyceae). The total and relative percentage of abundance (cells L -1 ) and richness (number of taxa) were calculated at 0 and 10 m. In relation to the biodiversity analysis, different types of diversity indexes were explored (Supplementary Table 1). Menhinick's diversity index was the most statistically significant index in relation to microphytoplankton abundance (P < 0.05). Dominance was calculated (1-Simpson index) using the whole count data (Zar 1999) in tandem with the diversity index.
All statistical analysis were performed using data registered in first 15 m with PAST v 3.11 software (Hammer et al. 2001). These statistics were based on the uniform phytoplankton vertical distribution described in the bay with cellular density concentrated in the first meters of the water column (0-20 m; Avaria & Orellana 1975) as well as the availability of micro-phytoplankton abundance data only at 0 and 10 m during the study period. Cluster analysis evaluated the similarity in micro-phytoplankton and oceanographic conditions throughout the year. The Bray-Curtis coefficient was used to determine the similarity. Finally, a non-parametric permutational multivariate test of variance (one-way PERMANOVA) assessed the significance differences among groups using a bi-modal regime model based on similarity matrix cluster results.
The upwelling index (UI) at the bay showed persistent upwelling favourable conditions especially from September to March (Fig. 2). In fact, a progressive increase in monthly UI is seen from August to November (Mdn from ~4.5 to 35 m 3 s -1 , Table 3). The highest UI values were reached between October and January (>750 m 3 s -1 , Fig. 2). The lowest UI values were registered from April to August (Mdn ~1.7-4.5 m 3 s -1 , Table 3), with an increment in downwelling events occurrence (Fig. 2). Likewise, an increase in the frequency of poleward winds were seen during this period (N-NE; Mdn ~30%, Table 1, Fig. 3) along with relaxation periods (Fig. 2, Table 2). Besides there was an increment in weak wind speed values (~0.5-2.1 m s -1 ) during this period (Table 2; Mdn ~20%). On the seasonal scale, the UI showed the highest values in spring and summer. Minimum value was registered in winter with an UI median value almost ten-fold lower than summer (Table 3).

Hydrological and oceanographic conditions
The Aconcagua River displayed high discharge values from November to January (Mdn from ~32.4 to 24.5 m 3 s -1 , Table 3); lower values were seen from June to August (Mdn ~9 m 3 s -1 , Table 3) and the minimum was in March to May (Mdn ~4 m 3 s -1 , Table 3). In this sense, a low surface salinity (Mdn ~33.0-33.2) was seen from November-January and June-August at the fixed station with a sea surface salinity (SSS, Fig. 4A and B) associated with Aconcagua River discharge influence in the bay.
The water column's salinity minimum occurred in July (0-60 m), and the maximum was during September-December with the arrival of salty and cold deep water (Fig. 4B). However, not important differences between seasons were seen in the first 15 m of water column salinity (Table 3). The temperature was low during August-November (Mdn ~12.4-12.8 ºC, Table 3) (Fig.  4C). Meanwhile, progressive warming in the upper layer (~0-10 m) was observed from November to April (Fig. 4C) with a noticeable thermal stratification in January-April (Fig. 4C) when the highest temperatures were registered (Mdn ~13.5-14.8 ºC, Table 3). Summer and autumn had the highest temperature median values versus spring and winter when the lowest values were registered ( Table 3).
The Brunt-Väisälä frequency showed high values from October to January in the upper layer (<10 m) (Fig. 4D), whereas a vertical homogenization was detected from March to August (Fig. 4D). Therefore, autumn and winter had a mixed water column, and late spring and summer showed a more stratified condition (Fig. 4D).

Phytoplankton
Phytoplankton biomass (Chl-a) showed a sub-surface maximum (~5-10 m) especially from October to March (Mdn range ~2.60-4.60 mg m -3 ; Table 4); low Chl-a values were registered from April-August period (Mdn ~2-0.5 mg m -3 ). In fact, winter showed low Chl-a with biomass barely reaching 1 mg m -3 in the water column (Table 4); spring and summer had the highest biomass values (Table 4).
The total micro-phytoplankton abundance was high from September to March with the highest values in  x 10 4 cells L -1 at 0 and 10 m, respectively) (Fig. 6B, Table 4). The abundance was low from May to August (Fig. 6, Table 4). In relation to the total micro-phytoplankton taxa, similar values of richness were registered along the year with a slight decrease from June to August, in winter, when the lowest abundance and richness were registered (Fig. 6C, Table 4).
Dinoflagellate abundance was considerably lower than diatoms (10-220 fold lower) (Table 4). Diatoms abundance was highest from September to March; whereas dinoflagellates were highest from March to April (Table 4). The dinoflagellates displayed seasonal highest abundance in autumn, and diatoms peaks in spring/summer (Table 4).

Figure 6. Monthly variability of: A) phytoplankton biomass (Chl-a), B) micro-phytoplankton groups relative (histogram) and total abundance (line) (0 and 10 m depth) and C) micro-phytoplankton groups relative (histogram) and total richness (line) (0 and 10 m depth). Median monthly values (line) are indicated in
The ecological indexes of the micro-phytoplankton community displayed a generally low diversity at both 0 and 10 m throughout the year (~ 0.02-0.11; Fig. 7A). However, there was some variability with high Menhinick index values in May and July and lower ones in November, December and March ( Fig. 7A  Finally, cluster analysis studied the relationship among environmental conditions and micro-phytoplankton in the bay at monthly scale considering the oceanographic variables of temperature, salinity, nutrient concentrations (NO 3 -, PO 4 3-, SiO 4 4-), and UI along with the phytoplankton biomass (Chl-a), and the abundance of diatoms and dinoflagellates (Fig. 8).
This cluster analysis showed relatively low similarity percentages, and two monthly periods can be distinguished: September to April and May to August (similarity of ~40%; Fig. 8). A one-way PERMANOVA test showed significant differences between these periods (pseudo-F= 12.35, P < 0.01).

Discussion
The results showed quasi-permanent S-SW favourable coastal upwelling winds along the year (~40-75%) similar to the wind temporal patterns previously described in Valparaíso Bay (Pizarro 1973, Reyes & Romero 1977. However, a detailed analysis identified two climatological periods with an annual signal in wind regime characterized by maximum in spring-summer and minimum during autumn-winter periods (Fig. 3) in agree with previous studies along central-south Chile (Shaffer et al. 1999, Narváez et al. 2004, Sobarzo et al. 2007. Intense S-SW winds and high wind speeds have been described during spring and summer. This is in contrast to the autumn/winter period when weak easterly winds increase in the bay with a shift from the southwest to the east winds. This temporal pattern in the winds at Valparaíso Bay contrast with the shift from the southwest to the north-northwest described by Sobarzo et al. (2007) at Concepcion Bay.
The wind regime at Valparaíso Bay (~33ºS) was connected to the upwelling activity (UI): The period from September to March (spring/summer) had the highest values followed by a decrease and downwelling events from April to August (autumn/winter). This coastal upwelling activity forced by the seasonally variable wind was also reflected in the oceanographic conditions of the bay. The oceanographic conditions from September to March -under intense upwelling activity-were characterized by the ascent of cold, nutrient-rich salty water in spring (September-November) followed by the development of core thermal stratification conditions in summer (December-March). These temporal temperature patterns have been described at other bays along central Chile including Concepcion (~36.5ºS) and Cartagena (~33.5ºS) where temperature has been linked to both upwelling activity and solar radiation (Narváez et al. 2004, Sobarzo et al. 2007, Testa et al. 2018. However, the temperature increase in summer under active upwelling has been associated not only with solar radiation but also with local warm-water pockets called upwelling shadows (Narváez et al. 2004, Thiel et al. 2007. Factors related to the rise of these warm water features include the coastal orientation of the bay (N-NE) and the location of the study area downstream of headlands. These factors correspond well to the study area at Valparaíso Bay. Nevertheless, the definition of an upwelling shadow area cannot be confirmed due to the bay's unknown physical dynamics.
The development of a nutricline has also been observed at Valparaíso Bay in the summer. This pattern is related to both the thermal stratification conditions preventing the arrival of nutrient-rich water to the upper layer (0-15 m) as well as the acquisition of spring upwelled nutrients by phytoplankton. In contrast, the autumn/winter period has homogenous conditions in the first few meters of the water column (temperature, salinity, and nutrients); this homogeneity is associated with weakening of upwelling activity.
This pattern coincided with the typical cycle of upwelling versus downwelling conditions (named bimodal regime) described in coastal systems governed by favourable upwelling winds (Anabalón et al. 2007(Anabalón et al. , 2016Thiel et al. 2007). The coastal region of central Chile (~30º-40ºS) has an intensification of upwelling-favourable S-SW winds during austral spring/summer period. This is accompanied by a combination of high temperature, salinity, and nutrient concentrations. Meanwhile, a nonupwelling period was seen by colder and less saline water in the surface layer (Montecino et al. 2006, Anabalón et al. 2007, 2016Thiel et al. 2007).
However, this bi-modal regime is highly heterogeneous due to changes in topography, solar radiation, and freshwater river discharge (Sobarzo et al. 2007, Anabalón et al. 2016, Testa et al. 2018. Therefore, recent studies have suggested changes in the upwelling versus nonupwelling cycle at Concepcion Bay (~36.5ºS) due to the key role of river discharge in coastal systems (Testa et al. 2018). In this bay high nutrient concentrations have been described along the year associated to Itata river input and coastal upwelling activity with peaks in winter and spring, respectively (Léniz et al. 2012, Anabalón et al. 2016, Masotti et al. 2018. At Valparaíso Bay the Aconcagua river plume arrives as seen via salinity changes. However, the river discharge did not impact nutrient availability. Nutrients concentrations showed a bi-modal regime connected essentially with local wind-driven upwelling. Coastline orientation, distance from river mouth to fixed station (~6 km; Fig. 1), low river flow (~4-32 m 3 s -1 ), and the coincidence in time of maximum river discharge and upwelling activity (spring) could explain the river's weak influence on the seasonal scales in terms of the bay's oceanographic conditions. Changes (both biomass and cell abundance) in phytoplankton at Valparaíso Bay at the annual scale are related to upwelling activity. These changes are not related to river discharge as seen in other coastal upwelling systems, (e.g., California/Oregon, NW Iberian-Portugal coasts), or even in central-south Humboldt area. In those areas, freshwater input leads to changes in phytoplankton structure, biomass (Chl-a) and PP (Warrick et al. 2005, Silva et al. 2009, Iriarte et al. 2012, Léniz et al. 2012, Masotti et al. 2018, Testa et al. 2018, Aparicio-Rizzo & Masotti 2019).
The intensification of wind-driven upwelling in spring/ summer (September/February) at Valparaíso Bay is a key factor for phytoplankton temporal variability. Here, 10 years of data processing characterized the annual cycle of phytoplankton biomass (Chl-a) with a sub-surface maximum (5-10 m) from October to March (spring/ summer). This agrees with the highest micro-phytoplankton total abundance values which are triggered by the upwelled nutrient-rich water into the upper layer. The May-August period has the lowest micro-phytoplankton total abundance and biomass. This relationship between coastal upwelling activity and increased phytoplankton biomass (Chl-a) and abundance in the spring/summer period has been described previously along central Chile (Avaria et al. 1989, Thiel et al. 2007, Correa-Ramirez et al. 2012.
Differences in micro-phytoplankton are also detected between upwelling and non-upwelling periods. The micro-phytoplankton community in Valparaíso Bay is largely constituted of diatoms and dinoflagellates with very scarce abundance of silicoflagellates. Although diatoms clearly dominate year-round, (Avaria 1971(Avaria , 1975Avaria & Orellana 1975, Avaria et al. 1989, maximum abundance is detected during the upwelling-active period in coincidence with highest biomass and richness values. An opposite pattern in terms of low biomass, total abundance, and richness has been described during the non-upwelling period (autumn/winter). Here, increases in diversity were observed in relation to an increment in dinoflagellates and silicoflagellates. These results confirmed a bi-modal regime of phytoplankton in the bay as postulated previously (Yáñez 1948, Avaria 1971, 1975Avaria & Orellana 1975, Pizarro 1973, 1976. Changes in phytoplankton from diatoms to dinoflagellates are related to upwelling activity and stability conditions in the water column. This is similar to other studies along coastal upwelling systems in California, N-NW Africa, and even in Humboldt (Margalef 1978, Avaria et al. 1989, Smayda 1998, González et al. 2007, Ochoa et al. 2010. In summary, this study provides evidence supporting the plausible causal relationship between meteorological (wind) and oceanographic variability in Valparaíso Bay. The local wind-forced upwelling regulates oceanographic conditions that trigger micro-phytoplankton response in the bay. In this sense the results suggest that a bi-modal regime of upwelling versus non-upwelling conditions regulates micro-phytoplankton in Valparaíso Bay.