Atmospheric Forcing of Coastal Upwelling in the Southern Baltic Sea Basin

Home , Atmospheric circulation, Ekman spiral

Article Atmospheric Forcing of Coastal Upwelling in the Southern Baltic Sea Basin

Ewa Bednorz * , Marek Półrolniczak , Bartosz Czernecki and Arkadiusz M. Tomczyk

Department of Climatology, Institute of Physical Geography and Environmental Planning, Adam Mickiewicz University, 61-712 Pozna´n,Poland; [email protected] (M.P.); [email protected] (B.C.); [email protected] (A.M.T.) * Correspondence: [email protected]; Tel.: +48-618-296-267

Received: 26 April 2019; Accepted: 14 June 2019; Published: 17 June 2019

Abstract: This study analyzes the atmospheric forcing of upwelling occurrence along differently oriented coastlines of the southern Baltic Sea basin. The mean daily sea surface temperature (SST) data from the summer seasons (June–August) of the years 1982–2017 made the basis for the detection of upwelling cases. For the atmospheric part of the analysis, monthly indices of four macroscale circulation patterns were used: North Atlantic Oscillation (NAO), Scandinavian (SCAND), East Atlantic (EA) and East Atlantic/Western Russia (EATL/WRUS). In order to identify the local airflows and wind conditions, zonal and meridional regional circulation indices were constructed and introduced to the analysis. Within the southern Baltic Sea basin, upwelling occurs most frequently along the zonally oriented southern coasts of Sweden, and least frequently along the southern (Polish) and eastern (Lithuanian-Latvian) coasts. Among the macroscale circulation patterns, the SCAND has the strongest impact on the horizontal flow of surface sea waters in the southern Baltic, which triggers upwelling. The summer NAO and EA appeared to have a weak effect on upwelling occurrence, and EATL/WRUS have the smallest impact. Local circulation indices allowed us to recognize the atmospheric control of upwelling frequency better than the indices of the macroscale patterns. Anomalies in upwelling frequency are their highest at the positive/negative phase of the zonal circulation, particularly along the southern and eastern coast of the southern Baltic Sea basin.

Keywords: Baltic Sea; coastal upwelling; macroscale circulation patterns; local circulation indices

1. Introduction Upwelling occurs in oceanic and sea waters, and is defined as a process of “an ascending motion of subsurface water by which water from deeper layers is brought into the surface layer and is removed from the area of upwelling by a divergent horizontal flow” [1]. Upwelling has a strong impact on sea waters mixing, as well as on the bio-geochemical processes and phytoplankton development [2–4]. In summer, when the sea waters are thermally stratified, cold and nutrient-rich waters rise to the surface. The nutrients fertilize surface waters, and as a result the areas of upwelling often have high biological productivity. In this way upwelling can have economic relevance, providing good fishing grounds in areas where it is common. Upwelling may take place in the open ocean, or along coastlines. In the latter case, it appears where the wind, blowing alongshore, has the coast on its left [2,5,6]. This general rule works as described in the Northern Hemisphere, where the Coriolis Force turns the water or air-flows to the right. The surface water is deflected offshore by surface currents, and is replaced with the deep seawater [7]. The direction of sea surface currents under the particular wind conditions is explained by Ekman’s Theory [5]. The direction of the air and water current within the Ekman Spiral is rotating due to the effect of the Earth’s rotation and frictional forces, and it is integral with the current down.

Atmosphere 2019, 10, 327; doi:10.3390/atmos10060327 www.mdpi.com/journal/atmosphere Atmosphere 2019, 10, 327 2 of 15

However, according to Ekman [5], also off-shore wind, i.e., wind blowing directly from the land interior, can cause upwelling. This may happen on scales too small for the Coriolis forces to dominate, in the order of the Rossby radius. Thus, it can be observed in the Baltic Sea, where the upwelling scale is approximately 10–20 km offshore and 100 km alongshore [6]. Describing the wind and ocean currents relationship, Ekman explained how the spiral “unwinds” close to the shore, making water move with the wind if it blows offshore [5] (p. 38–39, Plate 1), and argued that the Earth’s rotation has a negligible effect in a small scale and in an enclosed sea basin. In summer, warm surface water pushed offshore by alongshore or by offshore winds is replaced by the rising colder water from deeper layers, usually from below the thermocline [8–10]. This strong relationship between seawaters and atmospheric circulation is one of the apparent examples of the sea-atmosphere coupling. The Baltic Sea is described as a relatively small and semi enclosed basin, with borders oriented in different directions. As a consequence, coastal upwelling may appear under almost any wind direction at some coast section, and is thus considered to be frequent in the Baltic Sea [2,6,11]. Many studies of the upwelling statistics in different regions of the Baltic Sea were conducted, based at first on in situ data and, since the 1980s, upon satellite data [12–18]. Using remote sensing enables us to recognize upwelling events via analyzing the sea surface temperature distribution, but it also allows us to construct the three dimensional models of coastal waters circulation. Upwelling patterns are recognized as front structures, separating two masses of water, which differ in density, temperature and salinity [19]. Upwelling streams are also related to the meandering coastal water jets, associated with the undersea and coastal topographic features [9]. Some modern studies were based on modeling carried out to identify the upwelling events and describe them statistically, mainly showing the frequency of occurrence in different locations [6,11,17]. Fewer contributions considered the role of atmospheric forcing on horizontal and vertical sea currents. Remote and local atmospheric forcing on Baltic seawater circulation and the occurrence of upwelling were analyzed by Lehmann et al. [20]. They studied variations of the horizontal transport of seawaters and the occurrence of upwelling in response to changes in the North Atlantic Oscillation, and variations in local atmospheric conditions. Bychkova et al. [21] defined synoptic situations associated with upwelling in different coastal regions of the Baltic Sea coast. Their findings for the southeastern coasts of the Baltic Sea (namely, the Lithuanian-Latvian and Polish coastlines) were verified and discussed by Bednorz et al. [22,23]. This study aims to contribute to the research of the sea-atmosphere coupling by finding out the atmospheric forcing of the sea waters’ circulation and sea surface temperature along the differently oriented coastlines of the southern Baltic Sea basin. The macroscale Northern Hemisphere teleconnection patterns, recognized as the most relevant for the Euro-Atlantic section, will be taken into consideration first, and then, the local circulation conditions will be provided to recognize the atmospheric impact on the upwelling frequency in different coastal sections.

2. Area, Data and Methods The area of the study encompasses the southern part of the Baltic Sea basin, and embodies several diﬀerently-oriented coastal zones. Upwelling regions considered in this study were chosen following Bychkova et al. [21], who delimited 22 areas of coastal upwelling occurrence in the entire Baltic Sea (Figure1). In the moderate climate zone, upwelling is best recognized in summer, when the strongest stratification is found, and warm surface waters overlay colder deeper water masses. Therefore only the summer months (June–August) were analyzed in this study. The mean daily sea surface temperature (SST) data available in the NOAA OI SST V2 High Resolution Dataset, provided by the NOAA/OAR/ESRL PSD, Boulder, Colorado, USA, (http://www.esrl.noaa.gov/psd) made the basis for the detection of upwelling cases [24]. This data has a relatively low temporal (1 day) and low spatial resolution (geographical grid 0.25 0.25 ), which is disadvantageous in analyzing the upwelling cases. Nonetheless, the chosen ◦ × ◦ Atmosphere 2019, 10, x FOR PEER REVIEW 3 of 15

(a) (b) Atmosphere 2019, 10, 327 3 of 15 Figure 1. Area of the study with pixels taken into consideration computing sea surface temperature (SST) differences (a); main upwelling regions in the Baltic Sea (redrawn from Bychkova et al. [20]) (b). dataset contains SST for over thirty years (1982–2017) and has no gaps, which is particularly important in climatologicalIn the moderate analysis. climate An alternative zone, upwelling SST dataset is best of betterrecognized resolution in summer, (0.03 0.03 when), but the of strongest shorter ◦ × ◦ periodstratification is available is found at: http:, and// warmmarine.copernicus.eu surface waters .overlay The NOAA colder OI deeper SST V2 water High masses. Resolution Therefore Dataset only SST datasetthe summer was used mont inhs previous (June–Aug studiesust) concerning were analyzed upwelling in this along study. the Polish The mean and Lithuanian-Latvian daily sea surface coaststemperature [22,23]. (SST) Geographical data available resolution in the NOAA of 0.25◦ OImeans SST V2 27.8 High km Resolution resolution alongDataset, the provided meridians by and the ca.NOAA/OAR/ESRL 15–16 km resolution PSD, alongBoulder, the Colorado, parallels. AccordingUSA, (http://www.esrl.noaa.gov/psd) to Lehmann et al. [6] the horizontal made the scalesbasis for of upwellingthe detection amounts of upwelling to 10–20 cases km offshore [24]. This and data 100 has km a alongshore relatively [low13,21 temporal]. This asserts (1 day) that and the low resolution spatial ofresolution the SST data(geographical used in this grid study 0.25° is × sufficient 0.25°), which to recognize is disadvantageous upwelling events in analyzing (examples the in upwelling Figure2). Nevertheless,cases. Nonetheless one must, the rememberedchosen dataset that contains smaller SST upwelling for over close thirty to theyears coast (1982 may–2017) be undetectable and has no usinggaps, Atmospherethiswhich 2019 NOAA, 10is, x particularly FOR OI SST PEER data. REVIEW important in climatological analysis. An alternative SST dataset of better 3 of 15 resolution (0.03 × 0.03), but of shorter period is available at: http://marine.copernicus.eu. The NOAA OI SST V2 High Resolution Dataset SST dataset was used in previous studies concerning upwelling along the Polish and Lithuanian-Latvian coasts [22,23]. Geographical resolution of 0.25° means 27.8 km resolution along the meridians and ca. 15–16 km resolution along the parallels. According to Lehmann et al. [6] the horizontal scales of upwelling amounts to 10–20 km offshore and 100 km alongshore [13,21]. This asserts that the resolution of the SST data used in this study is sufficient to recognize upwelling events (examples in Figure 2). Nevertheless, one must remembered that smaller upwelling close to the coast may be undetectable using this NOAA OI SST data. The two-steps procedure was undertaken to recognize the upwelling events. First, the pairs of grids along the coast were selected, and the SST differences between the grids closer to the coast and the farther ones were computed for each pair (see Figure 1a for the pixels engaged). A negative value indicated that the coastal waters were colder than open-sea waters, and an upwelling event probably took place. Secondly, maps of SST were plotted for each of the preliminary-selected days, and a visual detection method was applied to identify the phenomenon at each coastal region. Situations with a significant drop of SST along(a )the coast, in comparison with the surrounding(b) open waters, with SST isotherms bent towards the sea, were selected (examples in Figure 2). The seasonal (June–August) Figurenumber 1.Figure Area of 1.upwelling ofArea the ofstudy the cases study with in with each pixels pixels pixel taken taken indicated into into considerationconsideration in Figure 1a computing was computing calculated sea surface sea providing surface temperature a temperature basis for (SST) diﬀerences (a); main upwelling regions in the Baltic Sea (redrawn from Bychkova et al. [20]) (b). (SST)further differences analysis. (a ); main upwelling regions in the Baltic Sea (redrawn from Bychkova et al. [20]) (b).

In the moderate climate zone, upwelling is best recognized in summer, when the strongest stratification is found, and warm surface waters overlay colder deeper water masses. Therefore only the summer months (June–August) were analyzed in this study. The mean daily sea surface temperature (SST) data available in the NOAA OI SST V2 High Resolution Dataset, provided by the NOAA/OAR/ESRL PSD, Boulder, Colorado, USA, (http://www.esrl.noaa.gov/psd) made the basis for the detection of upwelling cases [24]. This data has a relatively low temporal (1 day) and low spatial resolution (geographical grid 0.25° × 0.25°), which is disadvantageous in analyzing the upwelling cases. NonethelessFigure 2., theExamples chosen of seadataset surface contains temperature SST (SST) for distribution over thirty showing years upwelling (1982–2017) in diff anderent has no gaps, which is particularlyregions of southern important Baltic Sea in based climatological on the NOAA OI analysis SST V2 dataset.. An alternative SST dataset of better resolution (0.03 × 0.03), but of shorter period is available at: http://marine.copernicus.eu. The NOAA The two-steps procedure was undertaken to recognize the upwelling events. First, the pairs of OI SST V2grids High along Resolution the coast were Dataset selected, SST and dataset the SST was differences used betweenin previous the grids studies closer concerning to the coast and upwelling along thethe Polish farther and ones Lithuanian were computed-Latvian for each coasts pair (see [2 Figure2,23].1 aGeographical for the pixels engaged). resolution A negative of 0.25° value means 27.8 km resolutionindicated along that the the coastal meridians waters were and colder ca. 15 than–16 open-sea km resolution waters, and along an upwelling the parallels. event probably According to Lehmanntook et place. al. [6 Secondly,] the horizontal maps of SST scales were plotted of upwelling for each of amounts the preliminary-selected to 10–20 km days, offshore and a visual and 100 km alongshoredetection [13,21 method]. This was asserts applied that to the identify resolution the phenomenon of the SST at each data coastal used region.in this Situationsstudy is withsufficient to a significant drop of SST along the coast, in comparison with the surrounding open waters, with SST recognizeisotherms upwellin bentg events towards (examples the sea, were in selected Figure (examples2). Nevertheless in Figure,2 one). The must seasonal remembered (June–August) that smaller upwellingnumber close of to upwelling the coast cases may in be each undetectable pixel indicated using in Figure this1 aNOAA was calculated OI SST providing data. a basis for Thefurther two-steps analysis. procedure was undertaken to recognize the upwelling events. First, the pairs of grids along the coast were selected, and the SST differences between the grids closer to the coast and the farther ones were computed for each pair (see Figure 1a for the pixels engaged). A negative value indicated that the coastal waters were colder than open-sea waters, and an upwelling event probably took place. Secondly, maps of SST were plotted for each of the preliminary-selected days, and a visual detection method was applied to identify the phenomenon at each coastal region. Situations with a significant drop of SST along the coast, in comparison with the surrounding open waters, with SST isotherms bent towards the sea, were selected (examples in Figure 2). The seasonal (June–August) number of upwelling cases in each pixel indicated in Figure 1a was calculated providing a basis for further analysis.

Atmosphere 2019, 10, x FOR PEER REVIEW 4 of 15

Figure 2. Examples of sea surface temperature (SST) distribution showing upwelling in different Atmosphere 2019, 10, 327 4 of 15 regions of southern Baltic Sea based on the NOAA OI SST V2 dataset.

For the the atmospheric atmospheric part part of of the the analysis, analysis, monthly monthly indices indices of of macroscale macroscale circulation circulation patterns patterns were were derived from the National Oceanic and Atmospheric Administration (NOAA) Climate Prediction Centre (CPC) datasets ( (ftp://ftp.cpc.ncep.noaa.gov/wd52dg/data/indices/nao_index.tim).ftp://ftp.cpc.ncep.noaa.gov/wd52dg/data/indices/nao_index.tim). These These are are PCAPCA-based-based circulation circulation patterns, patterns and, and the the principal principal component component analysis analysis (PCA) (PCA)is applied is appliedto the Northern to the NorthernHemisphere Hemi monthlysphere mean monthly standardized mean 500 standardized hPa height anomalies 500 hPa [25 height]. Four anomalies teleconnection [25 patterns]. Four teleconnectionare recognized patterns as having are the recognized largest impact as having on climatethe largest and impact environment on climate the and Euro-Atlantic environment region the Euro(Figure-Atlantic3). Among region them (Figure the 3) North. Among Atlantic them Oscillationthe North Atlantic (NAO) Oscillation is regarded (NAO) as the is most rega prominentrded as the mostcirculation prominent pattern circulation over the middlepattern andover high the middle latitudes and of thehigh Northern latitudes Hemisphere. of the Northern NAO Hemisphere. is observed NAOall over is observed the year; however,all over the its year; influence however, on the its climaticinfluence conditions on the climatic over the conditions Euro-Atlantic over the region Euro is- Atlanticmost pronounced region is most in winter. pronounced This is mainly in winter. because This ofis mainly the seasonal because variation of the ofseasonal the oscillation variation strength of the oscillation(it explains strength 36.7% of (it the explains variance 36.7% of seaof the level variance pressure of sea in winter,level pressure but only in 22.1%winter, in but summer), only 22.1% and inbecause summer) of the, and seasonal because changes of the seasonal in the spatial changes patterns in the of spatial the NAO patterns dipole of centerthe NAO locations dipole [25center–29]. locationsIn summer [25 both–29]. NAO In summer centers both are shifted NAO centers northeast are in shifted comparison northeast to the in wintercomparison counterpart, to the winter which counterpart,means that thewhich summer means southern that the dipolesummer is southern located overdipole northwestern is located over Europe, northwestern rather than Europe over, ratherthe Azores–Spain than over the region Azores [27,–28Spain]. Regardless, region [27 not,28] only. Regardless, winter but not also only summer winter fluctuations but also summer of the fluctuationsNAO phases of and the theirNAO intensity phases and affect their the intensity weather affect and climate the weather in Europe, and climate mainly in by Europe, modifying mainly the bydirection modifying and intensitythe direction of air and circulation intensity [of28 –air32 ].circulation Therefore, [2 NAO8–32]. isTherefore, likely to aNAOffect alsois likely the directionto affect alsoand the magnitude direction of and sea magnitude currents. Theof sea second currents. teleconnection The second patternteleconnection included pattern in the included analysis in is the analysisScandinavian is the (SCAND).Scandinavian It consists (SCAND). of aIt primary consists circulationof a primary center circulation over Scandinavia, center over Scandinavia, with weaker withcenters weaker of opposite centers sign of opposite over western sign over Europe western and eastern Europe Russia and eastern [25]. The Russia positive [25] phase. The positive of this pattern phase ofis associatedthis pattern with is associated positive airwith pressure positive anomalies, air pressure sometimes anomalies, reflecting sometimes major reflecting blocking major anticyclones, blocking anticyclones,over Scandinavia over and Scandinavia north-western and Russia. north-western The East Russia. Atlantic The (EA) East pattern Atlantic is considered (EA) pattern as the is consideredsecond prominent as the second mode ofprominent low-frequency mode variability of low-frequency over the variability North Atlantic. over the It is North structurally Atlantic. similar It is structurallyto the NAO, similar and consists to the NAO, of a north-south and consists dipole of a north of anomaly-south centers,dipole of spanning anomaly thecenters North, spanning Atlantic thefrom North east to Atlantic west. The from anomaly east to centers west. of The the EAanomaly pattern centers are displaced of the southeastward EA pattern are to displaced the NAO southeastwardcenters. The East to Atlantic the NAO/ West centers Russia. The (EATL East/WRUS) Atlantic/ pattern West is considered Russia (EATL/WRUS) as the third prominent pattern is consideredteleconnection as the pattern third that prominent affects teleconnection Eurasia throughout pattern year. that The affects EATL Eurasia/WRUS throughout patternconsists year. The of EATL/WRUSthree main anomaly pattern centersconsists located of three over main Euro-Atlantic anomaly cen region.ters located The positive over Euro phase-Atlantic is associated region. withThe positivenegative phase height is anomaliesassociated locatedwith negative over the height central anomalies North Atlanticlocated over and the north central of the North Caspian Atlantic Sea, and northwith positiveof the Caspian height anomaliesSea, and with located positive over height western anomalies Europe. However,located over the western positive Europe. anomaly However, center in thethe positive positive phase anomaly is pronounced center in the only positive in winter, phase and is it pronounced weakens substantially only in winter in summer, and (Figureit weakens3). substantially in summer (Figure 3). NAO SCAND EA EATL/WRUS

−60 −45 −30 −15 15 30 45 60 Figure 3. The loading patterns for the four macroscale circulation patterns for July. The plotted value at Figure 3. The loading patterns for the four macroscale circulation patterns for July. The plotted value each grid point represents the temporal correlation between the monthly standardized height anomalies at each grid point represents the temporal correlation between the monthly standardized height at that point and the teleconnection pattern time series [33]. anomalies at that point and the teleconnection pattern time series [33]. The Northern Hemisphere teleconnection patterns reflect the macroscale circulation conditions. In orderThe Northern to find out Hemisphere the local airflows teleconnection and wind patterns conditions, reflect the zonal macroscale (Z) and meridionalcirculation conditions. (M) regional In ordercirculation to find indices out the were local introduced airflows in and the wind atmospheric conditions part, zonal of the (Z) analysis. and meridional To this end, (M) the regional mean monthly sea level pressure (SLP) data was used. There are many sources of atmospheric data of Atmosphere 2019, 10, x FOR PEER REVIEW 5 of 15 circulation indices were introduced in the atmospheric part of the analysis. To this end, the mean monthlyAtmosphere sea level2019, 10 pressure, 327 (SLP) data was used. There are many sources of atmospheric5 of 15 data of different resolutions (for example the newest product ERA5 of 0.25 × 0.25 resolution, available at https://cds.climate.copernicus.eudifferent resolutions (for example). However, the newest in product this study, ERA5 offor 0.25 computing◦ 0.25◦ resolution, the common available monthly × circulationat https: indices//cds.climate.copernicus.eu reanalysis (which). was However, quite in a this simple study, task for computing), data from the commonNational monthly Centers for Environmentalcirculation Prediction/National indices reanalysis (which Centre was for quite Atmospheric a simple task),Research data of from spatial National resolution Centers 2.5 for × 2.5 Environmental Prediction/National Centre for Atmospheric Research of spatial resolution 2.5 2.5 were fully sufficient [34]. 14 grids were used in the calculations and the monthly Z index was◦ × computed◦ were fully sufficient [34]. 14 grids were used in the calculations and the monthly Z index was computed as a normalized difference between SLP at 52.5 N (SLP averaged for 5 grid points indicated in Figure as a normalized difference between SLP at 52.5◦ N (SLP averaged for 5 grid points indicated in Figure4) 4) and and60 60N◦ (averaged,N (averaged, respectively) respectively) and and MM indexindex as as a normalizeda normalized diff erencedifference between between 22.5◦ E22.5 (SLP E (SLP averagedaveraged for 4 forgrid 4 grid points) points) and and 12.5 12.5 ◦EE (averaged, (averaged, respectively). respectively) Positive. Positive values valu ofes Z / ofM indicated Z/M indicated strongerstronger than usual than usual western/southern western/southern air air-flow-flow over over the the studied area. area. Lehmann Lehmann et al.et [al.20 ][ constructed20] constructed a similara local similar circulation local circulation index, index, called called the theBaltic Baltic Sea Sea Index Index (BSI), (BSI), usi usingng thethe normalized normalized SLP SLP anomalies anomalies at at grid points 59.5 N, 10.5 E and 53.5 N, 14.5 E. They applied the BSI index to analyze the local grid points 59.5 N, 10.5◦  E and◦ 53.5 N,◦ 14.5 ◦E. They applied the BSI index to analyze the local atmospheric forcing on water circulation in the Baltic Sea. atmospheric forcing on water circulation in the Baltic Sea.

Figure 4. Grid-points used to calculate zonal (blue and grey dots) and meridional (red and grey dots) Figurecirculation 4. Grid-points indices. used to calculate zonal (blue and grey dots) and meridional (red and grey dots) circulation indices. Finally, anomalies of the coastal upwelling frequency in positive/negative phases of both macroscale Finally,and local anomalies circulation patterns of the were coastal computed upwelling and plotted frequency for each upwelling in positive/negative region. They were phases computed of both macroscalefor each and pixel local separately circulation according patterns to the were equation: computed and plotted for each upwelling region. They were computed for each pixel separately accordingUp/n toU the equation: UFa = − 100% 푈푝/U푛 − 푈 푈퐹푎 = 100% where: UFa—anomaly of upwelling frequency; Up/n푈—mean monthly number of cases of upwelling U where:in UF thea— positiveanomaly/negative of upwelling phase of frequency; a given macroscale Up/n— ormean local monthly circulation number pattern; of—mean cases of monthly upwelling in number of upwelling cases. the positive/negative phase of a given macroscale or local circulation pattern; U—mean monthly number3. of Results upwelling cases.

3.1. Occurrence of Upwelling 3. Results Within the southern Baltic Sea region the number of days with upwelling is diversiﬁed along 3.1. Occurrencethe coastline, of Upwelling and is variable from year to year. The phenomenon is most frequent along the zonally oriented southern coasts of Sweden, where the mean number of upwelling events in the summer season Within(June–August) the southern exceeds Baltic 30 days Sea in region several pixelsthe number (Figure 5of). Nevertheless,days with upwelling the number is of diversified upwelling along the coastlineevents, along and thisis variable part of the from coast year can exceedto year 50. The days phenomenon in a single summer is most season, frequent like in thealong summers the zonally orientedof 1990,southern 1998 andcoasts 2000, of or Sweden, even 60 days, where like the in 2015 mean and number 2017. On of the upwelling other hand, events it may notin exceedthe summer season10 (June days,– likeAugust) in the summersexceeds of 30 1982, days 1995, in 1996 several and 2006.pixels (Figure 5). Nevertheless, the number of upwelling events along this part of the coast can exceed 50 days in a single summer season, like in the summers of 1990, 1998 and 2000, or even 60 days, like in 2015 and 2017. On the other hand, it may not exceed 10 days, like in the summers of 1982, 1995, 1996 and 2006.

Atmosphere 2019, 10, 327 6 of 15

Meridionally-oriented parts of the Swedish coast, including the eastern ridge of Oland island, experiences on the average 10–20 days with upwelling in a season, and again this number varies from above 40 days (in 1988, 2007, 2008, 2013, 2015, 2016, and 2017) to less than 5 days (in 1983, 1994, 1995, andAtmosphere 2003). 2019, 10, x FOR PEER REVIEW 6 of 15

Figure 5. SeasonalSeasonal (Jun (June–August)e–August) number number of days days with upwelling in different diﬀerent coastal regions.

CoastalMeridionally upwelling-oriented is least parts frequent of the Swedish along the coast, Polish including and Latvian-Lithuanian the eastern ridge coasts, of Oland where island the, numberexperience ofs upwelling on the average days 10 rarely–20 days exceeds with 10 upwelling (Figure5). in In a fact,season the, and phenomenon again this wasnumber not detectedvaries from at allabove in several 40 days summer (in 1988, seasons, 2007, 2008, like in2013, 1982, 201 1983,5, 2016, 1998, and 1999, 2017) 2000 to less and than 2012 5 along days the(in Polish1983, 1994, coast, 1995, and inand 1985, 2003). 1987, 1991, 2007, 2011, 2012 and 2017 along the Latvian-Lithuanian coast. On the other hand, moreCoastal than 30 upwelling days with upwellingis least frequent were identiﬁed along the in Polish 1992, 2002and andLatvian 2006-Lithuanian along the Polish coasts, coast, where and the in 1982,number 1992, of upwelling 1994, 1997 days and 2006 rarely along exceeds the Latvian-Lithuanian 10 (Figure 5). In fact, coast. the phenomenon was not detected at all inThe several mean summer seasonal season numbers, like of upwellingsin 1982, 1983, indicates 1998, 1999, theregions 2000 and that 2012 are along predestined the Polish to frequent coast, and an occurrencein 1985, 1987, of 1991, upwelling, 2007, 2011, while 2012 the year-to-year and 2017 along variability the Latvian of its-Lithuanian appearance coast. and persistence On the other suggests hand, thatmore the than external–presumably 30 days with upwelling atmospheric–conditions were identified in 1992, control 2002 the and circulation 2006 along of seawaters.the Polish coast, and in 1982, 1992, 1994, 1997 and 2006 along the Latvian-Lithuanian coast. 3.2. Impact of Macroscale Circulation Patterns on The Upwelling Frequency The mean seasonal number of upwellings indicates the regions that are predestined to frequent an occurrenceThe NAO pattern,of upwelling, being commonly while the yearconsidered-to-year as variability the most prominent of its appearance and recurrent and persistence circulation patternsuggests over that the external middle– andpresumably high latitudes atmospheric of the–conditions Northern Hemisphere,control the circulation particularly of seawaters. in winter, is strongly transformed in the summer. First of all, it is weaker than its winter counterpart and—which is3.2. crucial Impact in of thisMacroscale study—its Circulation southern Patterns center on is The shifted Upwelling north- Frequency and east-wards. This means that the summerThe southernNAO pattern, dipole being is located commonly over northwesternconsidered as Europe,the most and prominent encompasses and recurrent the western circulation part of thepattern Baltic over Sea the (Figure middle A1 ).and In thehigh positive latitudes phase of the of NAO,Northern in the Hemisphere, westernmost particularly part of the in anticyclone, winter, is namelystrongly in transformed the eastern in part the of summer. the southern First of Baltic all, it Sea is basin,weaker the than northern its winter airﬂow counterpart occurs, representingand—which theis crucial gradient in clockwisethis study— windsits southern around thecenter anticyclone. is shifted This north enhances- and east upwelling-wards. This frequency means along that the easternsummer ridge southern of the dipole southern is located Baltic, i.e.,over the northwest Lithuanian-Latvianern Europe, coast, and encompasses by approximately the western 30% (Figure part 6of). Athe weak Baltic impact Sea (Figure on the A upwelling1). In the positive frequency phase both of in NAO, positive in andthe westernmost negative phases part of of NAO the anticyclone, is observed inname otherly in coastal the eastern regions part of theof the southern southern Baltic. Baltic Sea basin, the northern airflow occurs, representing the gradient clockwise winds around the anticyclone. This enhances upwelling frequency along the eastern ridge of the southern Baltic, i.e., the Lithuanian-Latvian coast, by approximately 30% (Figure 6). A weak impact on the upwelling frequency both in positive and negative phases of NAO is observed in other coastal regions of the southern Baltic.

Atmosphere 2019, 10, 327 7 of 15 Atmosphere 2019, 10, x FOR PEER REVIEW 7 of 15

–

FigureFigure 6.6. AnomaliesAnomalies ofof upwellingupwelling frequencyfrequency inin positivepositive (+(+)) andand negativenegative ((−)) phasesphases ofof macroscalemacroscale − circulationcirculation patterns.patterns.

Atmosphere 2019, 10, 327 8 of 15

Among the macroscale teleconnection patterns considered in this study, the SCAND revealed the strongest influence on upwelling frequency. The positive phase of this pattern is associated with positive air pressure anomalies—sometimes reflecting blocking anticyclones—over Scandinavia (Figure A1). The anticyclone, accompanied with lower-than-normal pressure over western Europe and eastern Russia, induces geostrophic winds of clockwise direction around the high pressure center over Scandinavia. These bring about coast-parallel winds of eastern direction along the Polish coast and coast-perpendicular offshore winds at the Lithuanian-Latvian shore. These winds make favorable conditions for the rising of water along these coastlines, thereby causing upwelling. Consequently, the positive phase of the SCAND pattern exceeds the frequency of upwelling by more than 60% along the Lithuanian-Latvian coast and by more than 45–50% along the Polish coast (Figure6, SCAND +). On the other hand, the pressure patterns affiliated with the positive phase of SCAND are inconvenient to upwelling along the entire Swedish coast, and they diminish the probability of its occurrence by approximately 20–30%. The opposite pressure pattern under the SCAND negative phase (lower than normal pressure over the Scandinavia) indicates the opposite distribution of upwelling frequency anomalies (SCAND– in Figure6) with values exceeding 30 or even 40% along the Lithuanian-Latvian − − and Polish coasts. The EA pattern is structurally similar to the NAO, but its dipole centers are displaced southeastwards to the NAO ones. This means that the air circulation over northern and central Europe in the positive EA phase has a significant south-western component (Figure A1). This enhances upwelling along the meridionally-oriented Swedish coast, and increases its frequency by approximately 20–30%. A southern air-flow component in the positive EA phase modifies the frequency of upwelling in the westernmost part of the Polish coast (increased by 20–30%), while it does not interfere in its central and eastern part. The EATL/WRUS pattern, which in the positive phase characterizes with a weak center of positive SLP anomalies over southwest Scandinavia and the Norwegian Sea, does not reveal any significant impact on the frequency of coastal upwelling in the southern Baltic Sea basin.

3.3. Impact of Local Circulation Conditions on The Upwelling Frequency It was hypothesized that the local circulation patterns would better reflect the conditions which are favorable/unfavorable for coastal upwelling in the southern Baltic Sea basin. To this end zonal (Z) and meridional (M) regional circulation indices were introduced and anomalies of upwelling frequency were computed for positive/negative phases of Z and M. First, the correlations between macroscale and local circulation indices were computed. Statistically significant values were obtained for SCAND and Z as well as for EA and M (Table1). Strong negative correlation between SCAND and Z indices means that under the positive SCAND phase the anticyclonic conditions over Scandinavia substantially suppress westerly zonal airflow. Positive correlation between EA and M indices indicates significant southern air-flow component during the positive EA phase (Figures A1 and A2). Local circulation indices are last related to the fourth macroscale circulation pattern, namely EATL/WRUS.

Table 1. Pearson’s correlation coeﬃcients between indices of the macroscale circulation patterns (North Atlantic Oscillation (NAO), Scandinavian (SCAND), East Atlantic (EA), East Atlantic/Western Russia (EATL/WRUS).) and local circulation indices (Z—zonal, M—meridional).

Circulation Patterns NAO SCAND EA EATL/WRUS Z 0.295 0.767 ** 0.204 0.188 − − − M 0.328 0.010 0.349 * 0.038 − − − * values statistically signiﬁcant at p 0.05; ** values statistically signiﬁcant at p 0.01. ≤ ≤

Local circulation pattern Z, being strongly correlated with SCAND, revealed a similar to SCAND spatial pattern of the upwelling frequency control. In the positive phase of Z, the ampliﬁed zonal Atmosphere 2019, 10, x FOR PEER REVIEW 9 of 15

The frequency of upwelling is lowered by 50–65% along the Lithuanian-Latvian coast, and by 40–57% along the Polish coast. At the same time upwelling is slightly enhanced on the opposite side of the southern Baltic; i.e., upwelling frequency increases by up to 34% along the Swedish coast, where offshore and alongshore western winds blow. In the negative phase of Z, when the easterly airflow prevails, the opposite pattern of upwelling frequency is observed, namely there are much Atmosphere 2019, 10, 327 9 of 15 more-than-usually upwelling events along the Lithuanian-Latvian coast (increase by 58–76%) caused by the offshore winds, and along the Polish coast (40–67% more days with upwelling), caused by the easternwestern winds airﬂow; at the suppresses same time the coastalthe phenomenon upwelling in is the inhibited southern by and 20 eastern–40% along parts ofthe the Swedish southern coast (FigureBaltic 7). Sea basin (Figure7).

Figure 7. Anomalies of upwelling frequency in positive (+) and negative ( ) phases of zonal (Z) and Figure 7. Anomalies of upwelling frequency in positive (+) and negative− (−) phases of zonal (Z) and meridional (M) circulation. meridional (M) circulation. The frequency of upwelling is lowered by 50–65% along the Lithuanian-Latvian coast, and by 40–57%Meridional along local the Polish-scale circu coast.lation At the has same a weaker time upwelling impact on is slightlythe occurrence enhanced of on upwelling the opposite within the southernside of the Baltic southern Sea Baltic;basin i.e.,than upwelling the zonal frequency one. In the increases positive by upphase to 34% of M, along the the intensified Swedish coast,southern airflowwhere and o ffaccompanyingshore and alongshore offshore western winds winds enhance blow. the In upwelling the negative frequency phase of Z,mainly when along the easterly the Polish coastairflow (by 20 prevails,–30% in theits oppositecentral and pattern eastern of upwelling part, and frequency by more is than observed, 50% namelyin the westernmost there are much part). Undermore-than-usually the negative phase upwelling of M events the frequency along the Lithuanian-Latvian of upwelling along coast the (increase Polish bycoast 58–76%) is substantially caused lowered.by the Some offshore signals winds, of andchanges along in the upwelling Polish coast frequency, (40–67% more under days the with positive upwelling), and negative caused by phase the of M, areeastern observed winds; also at the along same the time northern the phenomenon part of the is inhibitedSwedish bycoast 20–40% (changes along up the to Swedish +30/−30% coast in the (Figure7). positive/negative M). In the other coastal regions only the weak influence of meridional airflow is Meridional local-scale circulation has a weaker impact on the occurrence of upwelling within the recognized, which is expressed by the small changes in the upwelling frequency (0–20%) under the southern Baltic Sea basin than the zonal one. In the positive phase of M, the intensified southern airflow positiveand accompanyingand negative p ohaseffshore of windsM. enhance the upwelling frequency mainly along the Polish coast (by 20–30% in its central and eastern part, and by more than 50% in the westernmost part). Under the 4. Discussionnegative phase of M the frequency of upwelling along the Polish coast is substantially lowered. Some signals of changes in upwelling frequency, under the positive and negative phase of M, are observed also along the northern part of the Swedish coast (changes up to +30/ 30% in the positive/negative M). −

Atmosphere 2019, 10, 327 10 of 15

In the other coastal regions only the weak inﬂuence of meridional airﬂow is recognized, which is expressed by the small changes in the upwelling frequency (0–20%) under the positive and negative phase of M.

4. Discussion It was confirmed that within the southern basin of the Baltic Sea, coastal upwelling may occur almost constantly along the differently oriented coastlines, as almost each pressure pattern develop favorable conditions, i.e., alongshore blowing wind having the particular coast on its left or offshore winds blowing from the land interior. The number of days with upwelling is diversified along the coastline and extremely variable from year to year. The phenomenon is most frequent along the zonally oriented southern coasts of Sweden, where the average frequency of upwelling conditions exceeds 30%. A similar frequency of upwelling in these regions was proved by Lehmann et al. [6], and Myrberg and Andrejev [11]. However, they postulated a similarly high frequency for the Oland eastern coast as well, which was not reaffirmed in this study. Coastal upwelling is least frequent along the Lithuanian-Latvian and the Polish coast (up to 10%). Kowalewski and Ostrowski [17] emphasized that downwelling of the coastal waters is more frequent along the Polish coast due to the prevailing westerly winds [11]. The homogenous daily set of data used in this study allowed to evaluate the year-to-year variability of the upwelling occurrence. The frequency of upwelling may amount to over 60% of summer days along the Swedish coast, and may drop to 0% along the Polish and Latvian-Lithuanian coasts during single summer seasons. This study’s results proved that both local and macroscale circulation patterns control the occurrence of coastal upwelling within the southern Baltic Sea basin. However, the impact strength varies, depending on the coastal region; the strongest relationships were found in the southeastern part of the basin, i.e., Polish and Lithuanian-Latvian coasts. Among the macroscale circulation patterns, the SCAND appeared to have the strongest control of the horizontal flow of surface sea waters in the southern Baltic. The positive phase of SCAND is represented by a high pressure system located in the vicinity of the Baltic Sea, i.e., over Scandinavia. The clockwise air circulation around the Scandinavian anticyclone brings eastern flow over the southern Baltic Se basin which means coast-parallel winds along the southern coastline and offshore winds at eastern coastline, which—in line with Ekman’s theory—triggers upwelling [2]. The summer NAO, being weaker than its winter counterpart and, moreover, having its centers of action shifted to the northeast, appeared to have unexpectedly weak impact on upwelling occurrence. Some signals were found along the eastern Oland coast and southern Swedish coast, where the positive/negative phase enhanced/inhibited upwelling. An opposite reaction was found out along Lithuanian-Latvian coast. In the numerous previous studies NAO was proved to modify Baltic Sea conditions including ice extent, SST, surface currents, freshwater runoff, salinity and oxygen concentrations [20,35–39]. The relevance of NAO, however, seems to be definitely more substantial in winter than in summer, when the EA pattern (which is often described as a shifted southwards NAO) seems to be more important than NAO and it modifies the frequency of upwelling along the eastern Swedish coast. Not surprisingly, the fourth macroscale circulation pattern considered as relevant over Euro-Atlantic region—EATL/WRUS pattern reveals no impact on sea water circulation in the southern Baltic Sea basin. Local circulation indices allowed to better recognize the atmospheric impact on upwelling frequency than the indices of the macroscale patterns. The anomalies in upwelling frequency are highest at positive/negative phase of the zonal circulation, identified at the regional pressure field. The Baltic Sea Index (BSI) similar to zonal local circulation used in this study was introduced by Lehmann et al. [20] in their research of atmospheric forcing on upwelling and sea water circulation in Baltic Sea. The BSI was computed as the difference of normalized SLP anomalies in two grid points located at about Szczecin (Poland) and Oslo (Norway). Authors proved that the positive BSI triggered inflow conditions in the Baltic Sea, and negative BSI values were correlated with outflow Atmosphere 2019, 10, 327 11 of 15 and a corresponding drop of the mean sea level. At the same time the correlation between the NAO and the volume exchange or mean sea level elevation of the Baltic Sea appeared to be small. Authors ascertained that possible relevance the NAO in the Baltic Sea region must first become evident in the local atmospheric conditions, which, furthermore, constitutes the direct impact [20]. It was reaffirmed in this study, that the local circulation patterns indicate the atmospheric forcing of coastal upwelling in the Baltic Sea better than the macroscale patterns. Among the latter, the SCAND was proved to have the strongest influence on upwelling frequency in summer, while the summer NAO revealed a weak relationship with the upwelling frequency. Due to a strong fluctuation of pressure patterns in the Baltic Sea region and, consequently, large variability of circulation conditions, the analysis of the atmospheric forcing of coastal upwelling occurrence should be analyzed in different spatial scales. Furthermore, the future daily time-scale research should be carried out, which would allow to study the process of formation, persistence and disappearance of upwelling in reliance to fluctuations of atmospheric conditions.

5. Conclusions Within the Baltic Sea, which is a small semi-enclosed basin with borders oriented in different directions, the coastal upwelling is considered to be a frequent phenomenon, as it may appear at some coast sections almost under any wind direction. The number of days with upwelling during the summer season (June-August) is diversified along the coastline and variable from year to year. Both local and macroscale circulation patterns control the occurrence of coastal upwelling within the southern Baltic Sea basin. However, the impact strength varies, depending on the coastal region and the strongest relationships were found in the southeastern part of the basin, i.e., Polish and Latvian-Lithuanian coasts. Among the macroscale circulation patterns, the SCAND has the strongest impact on the horizontal flow of surface sea waters in the southern Baltic, which triggers upwelling. The summer NAO, appeared to have weak influence on upwelling occurrence, while the EA pattern (which is often described as a shifted southwards NAO) seems to be more important and modifies the frequency of upwelling along the eastern Swedish coast. Local circulation indices allow to recognize the atmospheric impact on upwelling frequency better than the indices of the macroscale patterns. The anomalies in upwelling frequency are highest at positive/negative phase of the zonal circulation, recognized at the regional sea level pressure field.

Author Contributions: Conceptualization, E.B., M.P., B.C.; methodology, E.B., M.P., B.C.; formal analysis, E.B., M.P., B.C.; data curation, E.B., M.P., B.C.; writing—original draft preparation, E.B., A.M.T.; writing—review and editing, E.B., A.M.T.; visualization, E.B., M.P., A.M.T., H.F.-Ł.; funding acquisition, E.B. Funding: This research was funded by the National Science Centre, Poland, grant number 2016/21/B/ST10/01440. Acknowledgments: Authors would like to thank Hanna Forycka-Ławniczak for the technical support in constructing some figures. Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. Atmosphere 2019, 10, 327 12 of 15

Appendix A Atmosphere 2019, 10, x FOR PEER REVIEW 12 of 15

FigureFigure A1. A1Anomalies. Anomalies of SLP of SLP in positive in positive (left) (left) and negative and negative (right) (right) phase ofphase macroscale of macroscale circulation circulation patternspatterns based based at the at ﬁrstthe (negativefirst (negative phase) phase) and the and third the (positivethird (positive phase) phase) quartile. quartile.

Atmosphere 2019, 10, 327 13 of 15 Atmosphere 2019, 10, x FOR PEER REVIEW 13 of 15

Figure A2. Anomalies of SLP in positive (left) and negative (right) phase of zonal (Z) and meridional Figure A2. Anomalies of SLP in positive (left) and negative (right) phase of zonal (Z) and meridional (M) local circulation pattern based on the first (negative phase) and the third (positive phase) quartile. (M) local circulation pattern based on the first (negative phase) and the third (positive phase) quartile. References References 1. American Meteorological Society. Glossary of Meteorology. 2017. Available online: http://glossary.ametsoc. 1.org /wikiAmerican/climatology Meteorological (accessed on 21 January Society, 2019). 2017. Glossary of Meteorology. Available online: 2. Lehmann,http://glossary.ametsoc.org/wiki A.; Myrberg, K. Upwelling in/climatology the Baltic Sea—A (accessed review. on 21J. Marine January Syst. 2019)2008 , 74, 3–12. [CrossRef] 3. 2.Vahtera, Lehmann, E.; Laanemets, A.; Myrberg, J.; Pavelson, K. Upwelling J.; Huttunen, in the Baltic M.; Kononen,Sea—A review. K. Effect J. Marine of upwelling Syst. 2008 on,the 74, pelagic3–12. 3.environment Vahtera, and E.; Laanemets, bloom-forming J.; Pavelson, cyanobacteria J.; Huttunen, in the western M.; KononenGulf of Finland,, K. Effect Baltic of Sea. upwellingJ. Marine on Syst the pelagic 2005environment, 58, 67–82. [CrossRef and bloom] -forming cyanobacteria in the western Gulf of Finland, Baltic Sea. J. Marine Syst. 4. Zalewski,2005, 58(1 M.; Ameryk,–2), 67–82. A.; Szymelfenig, M. Primary production and chlorophyll a concentration during upwelling events along the Hel Peninsula (the Baltic Sea). Oceanological Hydrobiological Stud. 2005, 34, 97–113. 4. Zalewski, M.; Ameryk, A.; Szymelfenig, M. Primary production and chlorophyll a concentration during 5. Ekman, V.W. On the Influence of the Earth’s Rotation on Ocean-Currents. Ark. Mat. Astr. Fys. 1905, 2, 1–52. upwelling events along the Hel Peninsula (the Baltic Sea). Oceanological Hydrobiological Stud. 2005, 34(2), 6. Lehmann, A.; Myrberg, K.; Höflich, K. A statistical approach to coastal upwelling in the Baltic Sea based on 97–113. the analysis of satellite data for 1990–2009. Oceanologia 2012, 54, 369–393. [CrossRef] 5. Ekman, V.W. On the Influence of the Earth’s Rotation on Ocean-Currents. Ark. Mat. Astr. Fys. 1905, 2(11), 7. Urbański,J.A. Upwellings of the Polish coast of the Baltic Sea. Prz. Geofiz 1995, 40, 141–153. (in Polish). 1–52. 8. Choiński,A. Examples of upwelling in KoszalińskaBay. Badania Fizjograficzne Seria A–Geografia Fizyczna

6.2011Lehmann,, 62, 7–16. (inA.; Polish). Myrberg, K.; Höflich, K. A statistical approach to coastal upwelling in the Baltic Sea based 9. Gurova,on the E.; Lehmann,analysis of A.; satellite Ivanov,A. data Upwelling for 1990 dynamics–2009. Oceanologia in the Baltic 2012 Sea, studied54(3), 369 by– a393. combined SAR/infrared 7.satellite Urbański, data and J.A. circulation Upwellings model of the analysis. PolishOceanologia coast of the2013 Baltic, 55 ,Sea. 687–707. Prz. Geofiz [CrossRef. 1995] , 40(2), 141–153. (in Polish). 10. 8.Karstensen, Choiński, J.; A. Liblik, Examples T.; Fischer, of upwelling J.; Bumke, in Koszalińska K.; Krahmann, Bay. G. Badania Summer Fizjograficzne upwelling atSeria the A Boknis–Geografia Eck Fizyczna time-series2011, 62 station, 7–16. (1982(in Polish) to 2012)—A. combined glider and wind data analysis. Biogeosciences 2014, 11, 9.3603–3617. Gurova [CrossRef, E.; Lehmann,] A.; Ivanov, A. Upwelling dynamics in the Baltic Sea studied by a combined 11. Myrberg,SAR/infrared K.; Andrejev, satellite O. data Main and upwelling circulation regions model in analysis. the Baltic Oceanologia Sea—A statistical 2013, 55(3) analysis, 687–707. based on 10.three-dimensional Karstensen, J.; modelling. Liblik, T.; BorealFischer, Environ. J.; Bumke, Res. 2003 K.; Krahmann,, 8, 97–112. G. Summer upwelling at the Boknis Eck time- 12. Horstmann,series station U. Distribution (1982 to 2012) patterns – a of combined temperature glider and and water wind colour data in analysis. the Baltic SeaBiogeosciences as recorded 2014 in satellite, 11, 3603–3617. 11.images: Myrberg, indicators K.; Andrejev, for phytoplankton O. Main growth.upwellingBer. regions Inst. Meeresk. in the 1983Baltic, 106 Sea, 147.—A statistical analysis based on three- 13. Gidhagen,dimension L. Coastalal modelling. upwelling Boreal in Environ. the Baltic Res Sea—Satellite. 2003, 8, 97–112. and in situ measurements of sea-surface 12.temperatures Horstmann, indicating U. Distribution coastal upwelling. patternsEstuar. of temperature Coast. Shelf and Sci. 1987 water, 24 colour, 449–462. in the [CrossRef Baltic] Sea as recorded in satellite images: indicators for phytoplankton growth. Ber. Inst. Meeresk. 1983, 106, 147. 13. Gidhagen, L. Coastal upwelling in the Baltic Sea—Satellite and in situ measurements of sea-surface temperatures indicating coastal upwelling. Estuar. Coast. Shelf Sci. 1987, 24(4), 449–462.

Atmosphere 2019, 10, 327 14 of 15

14. Siegel, H.; Gerth, M.; Rudolff, R.; Tschersich, G. Dynamic features in the western Baltic Sea investigated using NOAA-AVHRR data. Dt. Hydrogr. Z. 1994, 46, 191–209. [CrossRef] 15. Kahru, M.; Hakansson, B.; Rud, O. Distributions of the sea-surface temperature fronts in the Baltic Sea as derived from satellite imagery. Cont. Shelf Res. 1995, 15, 663–679. [CrossRef] 16. Lass, H.-U.; Schmidt, T.; Seifert, T. Hiddensee upwelling field measurements and modelling results. ICES Coop. Res. Rep. 2003, 257, 204–208. 17. Kowalewski, M.; Ostrowski, M. Coastal up- and downwelling in the southern Baltic. Oceanologia 2005, 47, 453–475. 18. Uiboupin, R.; Laanemets, J. Upwelling characteristics derived from satellite sea surface temperature data in the Gulf of Finland, Baltic Sea. Boreal Environ. Res. 2009, 14, 297–304. 19. Marghany, M. Three-Dimensional coastal front visualization from RADARSAT-1 SAR satellite data. In Proceedings of the 12th International Conference Computational Science and Its Applications—ICCSA, Salvador de Bahia, Brazil, 18–21 June 2012; Springer: Berlin/Heidelberg, Germany; pp. 447–456. 20. Lehmann, A.; Krauss, W.; Hinrichsen, H.H. Effects of remote and local atmospheric forcing on circulation and upwelling in the Baltic Sea. Tellus 2002, 54A, 299–316. [CrossRef] 21. Bychkova, I.; Viktorov, S.; Shumakher, D. A relationship between the large scale atmospheric circulation and the origin of coastal upwelling in the Baltic. Meteorol. Gidrol. 1988, 10, 91–98. (in Russian). 22. Bednorz, E.; Półrolniczak, M.; Czernecki, B. Synoptic conditions governing upwelling along the Polish Baltic coast. Oceanologia 2013, 55, 767–785. [CrossRef] 23. Bednorz, E.; Półrolniczak, M.; Czernecki, B.; Tomczyk, A.M. Atmospheric forcing of upwelling along the south-eastern Baltic coast. Baltica 2018, 31, 73–85. [CrossRef] 24. Reynolds, R.W.; Smith, T.M.; Liu, C.; Chelton, D.B.; Casey, K.S.; Schlax, M.G. Daily High-Resolution-Blended Analyses for Sea Surface Temperature. J. Climate 2007, 20, 5473–5496. [CrossRef] 25. Barnston, A.G.; Livezey, R.E. Classification, seasonality and persistence of low-frequency atmospheric circulation patterns. Mon. Wea. Rev. 1987, 115, 1083–1126. [CrossRef] 26. Portis, D.H.; Walsh, J.E.; El Hamly, M.; Lamb, P.J. Seasonality of the North Atlantic Oscillation. J. Climate 2001, 14, 2069–2078. [CrossRef] 27. Hurrel, J.W.; Deser, C. North Atlantic climate variability: The role of the North Atlantic Oscillation. J. Marine Syst. 2003, 78, 28–41. [CrossRef] 28. Folland, C.K.; Knight, J.; Linderholm, H.W.; Fereday, D.; Ineson, S.; Hurrell, J.W. The summer North Atlantic Oscillation: Past, present, and future. J. Clim. 2009, 22, 1082–1103. [CrossRef] 29. Pokorná, L.; Huth, R. Climate impacts of the NAO are sensitive to how the NAO is defined. Theor. Appl. Climatol. 2015, 119, 639–652. 30. Shabbar, A.; Huang, J.; Higuchi, K. The relationship between the wintertime North Atlantic oscillation and blocking episodes in the North Atlantic. Int. J. Climatol. 2001, 21, 355–369. [CrossRef] 31. Trigo, R.; Pozo-Vázquez, D.; Osborn, T.; Castro-Díez, Y.; Gámiz-Fortis, S.; Esteban-Parra, M. North Atlantic oscillation influence on precipitation, river flow and water resources in the Iberian peninsula. Int. J. Climatol. 2004, 24, 925–944. [CrossRef] 32. Scherrer, S.; Croci-Maspoli, M.; Schwierz, C.; Appenzeller, C. Two dimensional indices of atmospheric blocking and their statistical relationship with winter climate patterns in the Euro-Atlantic region. Int. J. Climatol. 2006, 26, 233–249. [CrossRef] 33. Climate Prediction Center. Available online: http://www.cpc.ncep.noaa.gov/data/teledoc/nao.shtml (accessed on 21 January 2019). 34. Kalnay, E.; Kanamitsu, M.; Kistler, R.; Collins, W.; Deaven, D.; Gandin, L.; Iredell, M.; Saha, S.; White, G.; Woollen, J.; et al. The NMC/NCAR 40—Year Reanalysis Project. B. Am. Meteorol. Soc. 1996, 77, 437–471. [CrossRef] 35. Koslowsky, G.; Loewe, P. The western Baltic Sea ice season in term of mass related severity index 1879–1992. Part I: temporal variability and association with the North Atlantic Oscillation. Tellus 1994, 46A, 66–74. [CrossRef] 36. Koslowsky, G.; Glaser, R. Variations in reconstructed ice winter severity in the western Baltic from 1501 to 1995, and their implications for the North Atlantic Oscillation. Climate Change 1999, 41, 175–191. [CrossRef] Atmosphere 2019, 10, 327 15 of 15

37. Omstedt, A.; Chen, D. Inﬂuence of atmospheric circulation on the maximum ice extent in the Baltic Sea. J. Geophys. Res. 2001, 106, 4493–4500. [CrossRef] 38. Hanninen, J.; Vuorinen, I.; Hjelt, P.; Climatic Killworth, P.; Stainforth, D.; Webbs, D.J.; Paterson, S.M. A free surface Bryan–Cox–Semtner model. Inst. Oceanogr. Sci. Deacon Lab. Rep. 1989, 270, 184. 39. Zorita, E.; Laine, A. Dependence of salinity and oxygen concentrations in the Baltic Sea on large-scale atmospheric circulation. Clim. Res. 2000, 14, 25–41. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).