br Wind Pattern Change Along a Period of Coastal Occurrence Variation of a Stinging Medusa on a SW Atlantic Beach

Monte Hermoso (SW Atlantic, Argentina) is a coastal ecosystem highly modulated by wind forcing. Offshore winds have historically played a leading role in the tourist influx as they are associated with the occurrence of the endemic stinging medusa Olindias sambaquiensis. This species is closely related to warm summers but it could be favored by low temperatures in previous winters. Since summer 2013, a change in the wind pattern was locally perceived coincidently with the absence of the medusa in coastal waters. This work aimed at analyzing if wind pattern and sea surface temperature (SST) showed a measurable change along 2008-2015 and evaluating if this change can be associated to the observed variations in the occurrence pattern of O. sambaquiensis. Winds and SST data were analyzed in periods of high (2008-2012) and low (2013-2015) medusae occurrence, based on media and own observations. A significant change in the wind pattern in Monte Hermoso was detected. Wind speed decreased from 6.3 m s-1 in 2008 to 2.4 m s-1 in 2015 and calms and light winds increased by more than 44.2 and 7.5 %, respectively. This change implied an impediment of coastal upwelling, which is the physical process by which medusae reach the coast, and is associated with persistent (10 h or more) WNW to ENE winds blowing at ≥ 3.3 m s-1. Accordingly, in summer 2008 (period of high medusae occurrence), 45 coastal upwelling-days were detected in contrast to only 8 days in summer 2015 (period of absence of medusae). Also, positive SST anomalies were found over the 8-yr period, suggesting an increasing trend of 0.4 °C. Higher temperatures than normal during recent winters probably affect negatively the polyp reproduction. Our results demonstrate an environmental change in the ecosystem of Monte Hermoso and may explain, in part, the low occurrence of O. sambaquiensis in recent summers. The knowledge of the factors that regulate the occurrence of O. sambaquiensis is fundamental for developing a coastal management plan and/or for determining the first adaptation strategies.


Introduction
Coastal environments are dominated by complex interactions among atmospheric, terrestrial and marine factors. Circulation in neritic waters tends to be more active and complicated than in the open sea due to the efects of bottom and shoreline topography (Mc Connaughey, 1982;Barua, 2005) and meteorological phenomena (Ingle, 1999). For instance, ofshore winds tend to drive surface waters seaward and cause intermediate depth water to move shoreward and upwell near the coast. Thus, on coasts, winds usually account for seasonal dynamics as well as much of the daily and between-day variability in water circulation, nutrients input and the occurrence, abundance and spatial distribution of marine biota (e.g. Villate, 1994Villate, , 1997Gómez Erache et al., 2000). This may have an impact in many economic activities, including coastal management and ishing (Sousa et al., 2012).
Despite of the importance of local winds on the spatio-temporal distribution of organisms, in particular of planktonic ones, there are few studies considering this topic in the literature of coastal waters. Within zooplankton, gelatinous species like medusae are greatly afected by wind-driven events and/or hydrodynamic processes (Arai, 1992;Graham et al., 2001;Purcell & Decker, 2005). Dense aggregations of medusae frequently occur in the surface waters along coastal margins, inluenced by the direction and energy of prevailing winds and surface currents (Graham et al., 2001). Variation in medusae abundance can be a function of wind-driven circulation being wind dynamics more inluential to jellyish encounters with humans than any other abiotic variable (Decker et al., 2007;Kaneshiro-Pineiro & Kimmel, 2015). This is particularly important in tourist coastal areas where environmental information for the prediction of stinging species is employed to prevent harmful encounters (Kaneshiro-Pineiro & Kimmel, 2015;Keesing et al., 2015). Temperature is another factor that inluences blooming potential in medusae (Arai, 1992;Purcell, 2005;Schiariti et al., 2014), so it should be considered to understand variations in the dynamic of their populations.
The coastal area located on the northern Argentine continental shelf at 38°-42° S comprises highly relevant ecosystems from biological, economical and touristic viewpoints (Hofmeyer et al., 2009;Auad & Martos, 2012). This area is particularly responsive to atmospheric conditions due to its shallow depths and its extreme breadth (Piccolo, 1998;Lucas et al., 2005). Local inputs, tidal currents, seasonal thermal cycles and prevailing winds play crucial roles in inluencing the inner shelf dynamic and the biological communities living in it (Lucas et al., 2005;Hofmeyer et al., 2009;Auad & Martos, 2012). Furthermore, some biological patterns in the area are closely connected and forced by local wind stresses (Auad & Martos, 2012;Menéndez et al., 2012;Delgado et al., 2015).
One of the most enigmatic cases in the region of a biological pattern supposedly driven by winds takes place in Monte Hermoso and involves a stinging medusa species. Monte Hermoso is a touristic sandy beach mainly modulated by winds and tidal forcing (Martos & Piccolo, 1988;Delgado et al., 2012). Prevailing winds are from the N-NNW-NW directions and produce, particularly in summer, increases in the air temperature, luctuations of the atmospheric pressure and environmental dryness (Delgado et al., 2012). The presence of the stinging medusa Olindias sambaquiensis in coastal waters during summer is a determining factor in the tourism low to Monte Hermoso, and local people associate its occurrence to the inluence of persistent and intense N winds (ofshore direction) ( Figure 1). Varela (1982) and Mianzan & Zamponi (1988) studied this relationship and found higher probabilities of medusae encounters under persistent (18 to 24 h.) and intense (3.3 to 4.2 m s -1 ) N winds. This result was based on the coastline orientation (E to W) which allows that N winds cause coastal upwellings removing surface waters and producing the emergence of the bottom layer, promoting that medusae inhabiting in it reach the coast (Mianzan & Zamponi, 1988). Extremely high abundances has been commonly reported in the region (Mianzan et al., 2001;Chiaverano et al., 2004) but since the summer 2013, medusae of O. sambaquiensis have not been observed in the coastal waters of Monte Hermoso and surrounding areas (Dutto et al., 2017). Also, a change in the local wind pattern has been perceived by local visitors and reported by media but it has not been studied until now ( Figure 2). Thus, the aim of this work was to analyze if wind pattern showed a measurable change in terms of Wind Pattern Change Along a Period of Coastal Occurrence Variation of a Stinging Medusa on a SW Atlantic Beach Andrea Soledad Brendel ; María Sofía Dutto; María Clara Menéndez; María Andrea Huamantinco Cisneros & María Cintia Piccolo predominance (h d -1 ), persistence (consecutive h d -1 ) and speed (m s -1 ) between 2008 and 2015, and to compare within this period, the wind pattern during summers in which medusae were locally reported with those summers in which medusae were absent according to media reports and own data. Complementary, the Sea Surface Temperature (SST) pattern was also analyzed during the study period in order to look for variations that help to explain the reported changes in the medusa occurrence.
The region has a temperate climate characterized by warm summers, cold winters, moderate  Andrea Soledad Brendel; María Sofía Dutto; María Clara Menéndez; María Andrea Huamantinco Cisneros & María Cintia Piccolo springs and autumns, and the inluence of diverse air masses that gives variability in its weather condition (Delgado et al., 2012;Huamantinco Cisneros, 2012). Mean temperatures vary between 8 (winter) and 21.4 °C (summer), evidencing a marked seasonality (Huamantinco Cisneros, 2012). Prevailing winds throughout the year are from the N, NW and NNW directions with mean speed of 4 m s -1 while S, SE, SSE, SW and SSW winds are more intense (mean speed of 6.4 m s -1 ) and less frequent (Lucas et al., 2005;Delgado et al., 2012;Huamantinco Cisneros, 2012).

Wind Pattern Change Along a Period of Coastal Occurrence Variation of a Stinging Medusa on a SW Atlantic Beach
Monte Hermoso beach is associated to the homonymous coastal city. Tourism and secondly artisanal ishery, represent its main economic activities, therefore, there is an intense use of marine and coastal resources (Vaquero et al., 2007;Rojas et al., 2015). Monte Hermoso has evolved into a tourist center which shows the greatest population growth and expansion of the region due to it is one of the most selected tourist destinations of Argentina (Vaquero et al., 2007). During summers its population increases by more than 60,000 people (up to 1000 %) (Rojas et al., 2015).

Data Collection and Analysis
Wind data (speed and direction) for the period 2008-2015 were obtained from a meteorological monitoring station located in Monte Hermoso. The meteorological station integrates a coastal environmental monitoring stations network and is available on its website (EMAC, Estación de Monitoreo Ambiental Costero, http://emac.criba. edu.ar). EMAC provides, since 2007, continuous measurements of meteorological variables including air temperature, relative humidity, atmospheric pressure, wind speed, wind direction and solar radiation. All sensors were developed at the Instituto Argentino de Oceanografía and record measurements at 5-minute intervals.
SST data were obtained from the National Center for Environmental Prediction-National Center for Atmospheric Research (NCEP/NCAR) reanalysis data. NCEP/NCAR isa consistent long period data set and is one of the few available for oceanic model forcing (see Kalnay et al., 1996). It has been widely used over the Southern Hemisphere including several studies performed in the Argentine platform (Auad & Martos, 2012; Simionato et al., Figure 3 Location of Monte Hermoso (study area) in the southern region of Buenos Aires province (Argentina). Andrea Soledad Brendel;María Sofía Dutto;María Clara Menéndez;María Andrea Huamantinco Cisneros & María Cintia Piccolo 2005). The discussion on the quality of its use in the Southern Hemisphere can be found in Garreaud & Battisti (1999) and Simmonds & Keay (2009). The data provided for the NCEP/NCAR are within the range of data measured in situ during 2008-2010(Huamantinco Cisneros, 2012.

Wind Pattern Change Along a Period of Coastal Occurrence Variation of a Stinging Medusa on a SW Atlantic Beach
Wind and sea temperature analyses were irstly performed considering the whole 8-yr period (2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015). Thus, to visualize the wind pattern, the direction and speed of winds along with the air temperature were plotted throughout 2008-2015.
Winds were also classiied according to Beaufort Scale as calm (<0.5 m s -1 ), light winds (0.5-3.3 m s -1 ), moderate winds (3.3-10.8 m s -1 ) and strong winds (>10.8 m s -1 ), and the percentage of each Beaufort category was calculated monthly and plotted over the period. To analyze SST, monthly temperature means from NCEP/NCAR Reanalysis of the period 2008-2015 on a 2.5 x 2.5 latitude-longitude grid were used as the primary data. Annual and monthly SST means of the period 1948-2015 were then employed to estimate monthly and annual anomalies for the period 2008-2015.
The second time-scale focused on summers within the period 2008-2015, given that medusae typically show strong seasonality displaying maximum population values on the warmest months. Therefore, the direction and speed of winds were plotted for each summer, and the frequency of occurrence and the mean speed of winds were associated to each plot. To compare the wind patterns between contrasting periods in relation to medusae occurrence, two summers were chosen based on local media and published biological data ( Figure 2). Thus, December 2007, January and February 2008, henceforth summer 2008, was chosen as a representative period of high medusae occurrence, while December 2014, January and February 2015, henceforth summer 2015, was considered as a representative one of absence of medusae. Medusae-associated winds (WNW-ENE) (Mianzan and Zamponi, 1988) were selected and predominance (h d -1 ) and persistence (consecutive h d -1 ) were estimated and evaluated over those summers. The number of days of winds associated to medusae occurrence (WNW-ENE winds blowing for 10 h or more, and speed ≥ 3.3 m s -1 , Mianzan & Zamponi, 1988) were compared between summers 2008 and 2015.

Statistical Analysis
Wind variables (speed, predominance and persistence) and the number of days of winds associated to medusae occurrence were compared between summers 2008 and 2015 applying Mann-Whitney tests given that normality was not assumed even after data transformation. Principal Component Analysis (PCA) using Spearman's rank correlation matrix was used to arrange and visualize the data, and to detect relationships among variables (Beaufort categories of winds and SST). The irst two principal components (PC1 and PC2) were retained given that they explained a meaningful part of the total variation. Plotted variables showed a reconstruction percentage >50 % in the two-dimensional plot.

Wind Speed Throughout the Whole 8-yr Period (2008-2015)
The wind speed showed a decreasing trend  Figure 4A), and it seemed to follow a seasonal variation: N winds seemed to be more frequent in summers, while S winds seemed to dominate during winters ( Figure  4A). A change in wind speed was detected through the study period and this was clearly observed by the Beaufort categories ( Figure 4B Figure 4B). Strong winds showed a mean of 5.6 % and similarly to moderate ones, they decreased from 2008 (10.4 %) to 2015 (3.2 %), being the latter the lowest value found over the entire study period ( Figure 4B).

Principal Component Analysis (PCA): Relationships Among Variables
The irst two components explained     Wind direction is indicated according to the meteorological convention (i.e. the direction from which the wind is blowing) and speed is represented by the length of the arrows. Wind rose plots next to each vector plot show frequency of wind occurrence (%) and mean speed for each summer. Andrea Soledad Brendel; María Sofía Dutto; María Clara Menéndez; María Andrea Huamantinco Cisneros & María Cintia Piccolo

Wind Predominance and Speed Throughout the Summers of 2008-2015
Wind occurrence and speed patterns for each summer along the period 2008-2015 are shown in Figure 8. Winds from the whole N sector (N, NW, NE, NNW, NNE) dominated during all summers (Figure 8) and showed a decrease from summer 2008 (5.1 m s -1 ) to summer 2015 (1.6 m s -1 ) ( Figure  8, see wind roses). Particularly, N winds were more frequent in most of the summers, with frequencies of occurrence between 13.3 and 21.3 %. Exception was observed in summer 2009 and 2013 when NNE and NW winds dominated (42.2 and 11.9 %, respectively) ( Figure 8). Prevailing winds (N-NNW-NW) were signiicantly higher in summer 2008 than in 2015 (5.2 ± 0.2 and 2.6 ± 0.5 m s -1 respectively; Mann-Whitney U test, W=9292.50; p<0.0001) blowing at both summers for similar time duration (11h19m ± 05h52m in 2008 and 12h31m ± 6h13m in 2015, Mann-Whitney U test, W=6009; p=0.4796). This result is clearly evident in the vector plots, in which the length of the vectors decreased noticeably over the summers (Figure 8). The highest frequencies of occurrence of S winds were detected during the summers 2014 and 2015 (close to 10 %) (Figure 8 continuation). Comparing summers 2008 and 2015 in relation to whole S sector winds, it is evident that in 2008 there was a higher variation in the direction, whereas in 2015 S winds dominated most of the period (Figure 8, see vector plots). Winds from S sector were less frequent but more intense than N winds during both periods (means of 8.7 m s -1 in summer 2008 and 3.9 m s -1 in summer 2015).

Wind Associated to Coastal Upwelling: Analyses Between the Summers of 2008 and 2015
Ofshore winds (WNW-ENE) were higher in summer 2008 than 2015 (mean speed 5.5 m s -1 in 2008 and 1.6 m s -1 in 2015). As it was already mentioned, summer 2008 and 2015 corresponded to periods of high occurrence and absence of medusae, respectively ( Figure 2). The number of days of winds associated to coastal upwelling and consequently to medusae occurrence (winds blowing 10 hours straight from WNW to ENE at 3.3 m s -1 or more), were consistently higher in summer 2008 than in summer 2015 (Figure 9). In December 2007, 19 days showed that wind condition in contrast to 4 in December 2014 ( Figure 9A). In January 2008, 15 days of medusae-associated winds occurred while only 4 days happened during January 2015 ( Figure 9B). They accounted for up to 11 days in February 2008 while any day of medusae-associated winds was found in February 2015 ( Figure 9C).  Andrea Soledad Brendel; María Sofía Dutto; María Clara Menéndez; María Andrea Huamantinco Cisneros & María Cintia Piccolo

Discussion 4.1 Shifts in Wind and Temperature Patterns Along 2008-2015 in Monte Hermoso
A quantiiable change in wind speed was detected in Monte Hermoso throughout the period 2008-2015. Moderate and strong winds decreased importantly. In relation to wind predominance, N-NNW-NW winds were dominant all over the period, consistent with the typical pattern described for the region (Martos & Piccolo, 1988;Huamantinco Cisneros & Piccolo, 2010;Delgado et al., 2012).

Diferent changes in atmospheric processes
have been previously reported for the southwestern Atlantic Ocean (Gregg & Conkright, 2002;Severov et al., 2004;Simionato et al., 2005;Dragani et al., 2013;Pescio et al., 2015), and in particular for the Argentine platform (Martos & Piccolo, 1988;Lucas et al., 2005;Auad & Martos, 2012;Delgado et al., 2012). The passage of cyclones produced from baroclinic and lee topographic instability of the westerlies seemed to cause the high temporal variability of the winds over the southeastern South Atlantic (Gan & Rao, 1991). The wind variability in this region features a seasonal cycle that is strongly linked to the position and strength of the South Atlantic highpressure system (Martos & Piccolo, 1988;Castro & Miranda, 1998;Auad & Martos, 2012). In summer, the atmospheric mean circulation changes as the South Atlantic high-pressure cell migrates southward (~32°S), generating E-NE winds become the prevailing winds between 18°S and 40°S (Pimenta et al., 2008). Therefore, the changes in the wind pattern detected in Monte Hermoso could be associated to changes in the position and speed of this high-pressure system. Given that the dynamic of this coastal ecosystem is strongly modulated by tidal and wind forcing (Auad & Martos, 2012;Delgado et al., 2014), signiicant changes in the wind pattern would have considerable efects on the ecosystem dynamic.
Shifts in the sea surface temperature were also suggested in Monte Hermoso, revealing that the sea surface become especially warmer in the last two years. The SST showed an increase over the study period of 0.4°C, in accordance with the warming trend of the sea and land surface detected globally (Casey & Cornillon, 2001;Stocker, 2013). Particularly, for the Argentine platform (34°-55°S), positive temperature anomalies and trends have been reported by several authors. For instance, Agosta and Campugnucci (2008) found a signiicant positive trend of SST along the period 1948-1977, and Martinez Avellaneda (2005 recorded positive anomalies (+0.4 °C between 1985-1999) and a positive tendency throughout the whole Argentine platform. Lucas et al. (2005) suggested diference of +3 °C in the SST during spring/summer (1993)(1994)(1995)(1996)(1997)(1998)(1999) between north and south of 38°S and Auad & Martos suggesting, although minimal, a positive trend for the 8-yr period. Finally, it is important to highlight that there are conclusive evidences that the Argentine platform is afected by El Niño-Southern Oscillation (ENSO) (Severov et al., 2004;Martínez Avellaneda, 2005;Delgado et al., 2015), which would inluence the SST adding an extra climatological forcing acting on the modulation of the temperature variability.

Variations in the Coastal Occurrence of O. sambaquiensis in Monte Hermoso: Explanations Associated to the Observed Environmental Change
The relationship between wind dynamic and the geomorphology of the coast can inluence the distribution of medusa causing them to concentrate or swarm near to the shore (Kaneshiro & Kimmel, 2015). For instance, the occurrence of Chrysaora quinquecirrha (Scyphozoa) into the Neuse River Estuary has been reported as a function of winddriven circulation (Kaneshiro & Kimmel, 2015). In Cable Beach (north-western Australia), on the other hand, stranding events of the scyphozoan Crambione mastigophora occurred when winds blow onshore (Keesing et al., 2015).
Ofshore winds can trigger coastal upwelling, producing that nutrients and plankton which originally were in the deeper water layer go upward Wind Pattern Change Along a Period of Coastal Occurrence Variation of a Stinging Medusa on a SW Atlantic Beach Andrea Soledad Brendel ;María Sofía Dutto;María Clara Menéndez;María Andrea Huamantinco Cisneros & María Cintia Piccolo (Barua, 2005). This phenomenon explains why, during summers of high abundances of medusae, they reach the coastal waters of Monte Hermoso when local winds blow ofshore for a long enough period (Varela, 1982;Mianzan & Zamponi, 1988). A number of 500 to 1000 stingings per day used to be reported in Monte Hermoso during those summers (Mosovich & Young, 2012). However, as far as we know, no one stinging accident has been reported in recent summers coinciding with the wind-pattern change. It is evident that wind is an important factor that inluences medusae-human encounters in Monte Hermoso. Therefore, a change in wind pattern will have a relevant impact not only in the dynamic of the coast but also in the tourist inlux to this beach.
On the other hand, as it is widely known, sea temperatures is one of the main environmental factors inluencing the distribution and abundance of medusae communities (Arai, 1992;Purcell, 2005;Primo et al., 2012;Schiariti et al., 2014). Although diferent medusae species may respond diferently to changing environmental conditions and same species may also difer by region (Lucas et al., 2012), the general trend for temperate gelatinous species is that warm temperatures lead to greater and more rapid production (Purcell, 2005). Regardless of the fact that blooms may be deined phylogenetically by the asexual modes each species develops (Schiariti et al., 2014), temperature (and salinity) both may afect positively or negatively the asexual reproduction rates of medusae directly through metabolism, and indirectly through prey capture (Purcell, 2005;Decker et al., 2007). In the case of Olindias genera, Patry et al. (2014) found that temperature interacts diferently with the polyp and the medusa.
According to their results, Olindias formosus hydroids persisted only at 15°C, with regular asexual reproduction, and were lost at temperatures higher than 20°C. Medusae of this species were produced at temperatures higher than 15°C, growing larger and in better health condition at 20°C (Patry et al., 2014). The results correlated with the seasonal inluence of the Kuroshio Current of coastal Japan where the research was experimentally conducted, revealing a close relation with the factors that may afect the environment (Patry et al., 2014). Based on all these indings, we could think that changes in the environment as a rise in temperature in winter may have a direct negative efect on the polyp asexual reproduction of O. sambaquiensis, which it could have negative consequences in the production rate of medusae during subsequent summers.
Given that in the last years, specimens of O. sambaquiensis were not found in plankton samples along the coastal area (Dutto et al., 2017;this study), an environmental regulation would be acting on the population dynamic of the species. If medusae were in the water column, the change in the wind pattern would probably prevent medusae reach the coast, but we do not have evidences that this change in wind magnitude is impacting on medusae reproduction or if it could impact it. The true is that the species has not been seen in the coastal waters of Monte Hermoso since summer 2013, and this may be probably in part due to an environmental regulation acting on the water column along with the lack of the physical phenomenon (i.e. coastal upwelling) that allow medusae reach the coast.
Our results evidence a signiicant decrease from 2008 to 2015 in the speed of ofshore winds which signiicantly impacted in the number of days of meteorological conditions associated to coastal upwellings over the period. Persistent, low-speed ofshore winds would not cause the phenomenon of upwelling. These results implied an important environmental change for this coastal region considering that days with intense and persistent N winds used to be the typical climatic condition for the area (Martos & Piccolo, 1988;Perillo & Cuadrado, 1990;Huamantinco Cisneros & Piccolo, 2011;Delgado et al., 2012). Under the current meteorological conditions, not conducive to the occurrence of upwellings, medusae that inhabit the deeper layers would not reach the coast if they were actually there. Although studies under controlled environmental condition (e.g. temperature and salinity) are needed to explain the variation in the population dynamic of O. sambaquiensis in the coastal ecosystem of Monte Hermoso and surroundings, our results could be closely associated to the observed change in the occurrence pattern of this species.

Acknowledgments
The authors are grateful to CONICET and UNS for the inancial support. Also, this work was carried out with the aid of a grant from the Inter-American Institute for Global Change Research