The role of forcing agents on biogeochemical variability along the southwestern Adriatic coast: The Gulf of Manfredonia case study

Accepted Manuscript The role of forcing agents on biogeochemical variability along the southwestern Adriatic coast: The Gulf of Manfredonia case study Antonietta Specchiulli, Francesco Bignami, Mauro Marini, Adele Fabbrocini, Tommaso Scirocco, Alessandra Campanelli, Pierluigi Penna, Angela Santucci, Raffaele D'Adamo PII:

S0272-7714(16)30530-3

DOI:

10.1016/j.ecss.2016.10.033

Reference:

YECSS 5292

To appear in:

Estuarine, Coastal and Shelf Science

Received Date: 10 February 2016 Revised Date:

20 October 2016

Accepted Date: 26 October 2016

Please cite this article as: Specchiulli, A., Bignami, F., Marini, M., Fabbrocini, A., Scirocco, T., Campanelli, A., Penna, P., Santucci, A., D'Adamo, R., The role of forcing agents on biogeochemical variability along the southwestern Adriatic coast: The Gulf of Manfredonia case study, Estuarine, Coastal and Shelf Science (2016), doi: 10.1016/j.ecss.2016.10.033. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT The role of forcing agents on biogeochemical variability along the southwestern Adriatic coast: the Gulf of Manfredonia case study

RI PT

AntoniettaSpecchiulli1*, Francesco Bignami2, Mauro Marini3, Adele Fabbrocini1,Tommaso Scirocco1, Alessandra Campanelli3, Pierluigi Penna3, Angela Santucci1, Raffaele D’Adamo1

1*

SC

National Research Council - Institute of Marine Science, Lesina section(FG),Via Pola 4, 71010

2

National Research Council - Institute of Atmospheric Sciences and Climate, Rome section, Via

Fosso del Cavaliere100, 00133Rome, Italy

National Research Council - Institute of Marine Science, Ancona section, Largo Fiera della Pesca,

EP

60122Ancona, Italy

TE D

3

M AN U

Lesina (FG), Italy

AC C

*Corresponding author: Phone: +39 0882992702; Fax +39 0882991352 E-mail address: [email protected] Complete correspondence address: National Research Council – Institute of Marine Science, Department of Lesina (FG) Via Pola, 4 - 71010 Lesina (FG), Italy

1

ACCEPTED MANUSCRIPT Abstract This study investigates how multiple forcing factors such as rivers, surface marine circulation and winds affect hydrology and biogeochemical processes in the Gulf of Manfredonia and the seas

RI PT

around the Gargano peninsula, in the south-western Adriatic Sea. The study adopted an integrated approach, using in situ and remote sensing data, as well as the output of current models. The data reveal variability in the area's hydrography induced by local freshwater sources, the Western Adriatic Current (WAC) flowing from the north along the Italian coast, and the current patterns

SC

under different wind regimes. Specifically, exchange with offshore waters in the gulf induces

M AN U

variability in salinity and biogeochemical content, even within the same season, i.e. winter, in our case. This strong dependence on physical and biogeochemical factors makes the ManfredoniaGargano ecosystem vulnerable to climate change, which could compromise its important role as a nursery area for the Adriatic Sea.

AC C

EP

TE D

Keywords: Coastal waters; biogeochemistry; CDOM; remote sensing; river input; Adriatic Sea

2

ACCEPTED MANUSCRIPT 1. Introduction Coastal areas are highly dynamic environments, where local-scale natural processes and anthropogenic pressures affect biogeochemical and optical processes, leading to continuous changes in the ecosystems and their services (Mancinelli and Vizzini, 2015). Mixing of terrigenous waters

RI PT

and seawater represents a key process for the biological productivity of marine coastal environments, with important implications for the functioning of the whole coastal system.

Recently, there has been an increase in studies of physical, biological and chemical processes in

SC

coastal systems over narrower time and space ranges (Cherukuru et al., 2014; Organelli et al., 2014; Tremblay et al., 2014). This interest reflects the rising frequency of short-term phenomena resulting

M AN U

from climate change, especially the increase in intense rainfall events and consequently in terrigenous loads entering coastal waters from rivers. Intensive regional research programs have also been carried out in coastal waters in order to study optical variability and its dependence on biogeochemical processes (Erga et al., 2012; Li et al., 2014; Retelletti Brogi et al., 2015; Sasaki et

TE D

al. 2005). In coastal systems affected by river discharge, physical processes such as wind and surface currents, as well as freshwater inflow, influence the input of organic matter, phytoplankton production and trophic transfer (Bowers and Brett, 2008; Philips et al., 2012). The consequences of

EP

altered patterns of freshwater input on the biogeochemistry of coastal systems have been reported in many studies (Vignudelli et al., 2004; Berto et al., 2010; Buzzelli et al., 2014; Marini et al., 2015).

AC C

Within the Mediterranean Sea area, the Adriatic Sea represents a complex basin characterised by diverse biogeochemical dynamics and circulation, from north to south. The general surface circulation pattern is cyclonic, flowing southward along the Italian coast as the Western Adriatic Current (WAC) and carrying fresh waters originating from the largest north Italian river, the Po (Artegiani et al. 1997; Marini et al. 2008). Several studies have highlighted a clear WAC signal extending as far south as the “spur” of the Italian boot, i.e. the Gargano Peninsula (Fig. 1) and beyond, into the Ionian Sea (Bignami et al. 2007; Lipizer et al. 2014; Marini et al. 2015). The Gargano peninsula forms the northern limit of the Gulf of Manfredonia, an area of oceanographic 3

ACCEPTED MANUSCRIPT and ecological relevance, as it is a nursery zone for small pelagic fish (Borme et al. 2013; Monticelli et al. 2014). The Gulf (Fig. 1) is affected by two main rivers, the Candelaro (67 Km long, catchment area 2200 Km2) and the Ofanto (165 Km long, catchment area 2790 Km2), which receive wastewaters from major towns (∼300,000 inhabitants) and together drain a large catchment

RI PT

area subject to intensive agriculture (De Girolamo et al., 2014; Focardi et al., 2009).

The presence of multiple forcing factors in the area (rivers, surface marine circulation, winds) and its ecological importance prompted us to investigate how these variables affect the

SC

area’s hydrology and biogeochemical processes and thus its productivity. To address this question, we used in situ and satellite observations as well as model output to assess: 1) the effects of small

M AN U

local rivers with season-dependent flood periods on the hydrological structure of the marine water masses; 2) the biogeochemical variability related to river discharges; 3) the influence of the WAC on regional coastal dynamics; 4) the ecological implications of the observed features.

2.1 Sampling design

TE D

2. Materials and Methods

Bi-monthly field samplings were performed in 2013 (February22nd, April 23rd, June 18th,

EP

August 29th, October 23rd, December 12th) at 6 stations (Fig. 1, Table 1). Site M (4 km from the

AC C

coast with a depth of 20 m) was located north of the Gargano as a reference point for WAC waters, and sites A-E were located in the Gulf of Manfredonia, along a transect from the Candelaro river delta eastwards. Site A was located 0.3 km from the river mouth, sites B-C-D were spaced about 7 km from each other and the outermost site E was 25 km from the coast, with depths of 9, 15, 17, 19, 20 m, respectively. Details on sampling dates and times are shown in Table 1. At each site, the hydrographic temperature (T) and salinity (S) profiles were recorded with a calibrated SBE 19Plus probe (Sea-Bird Electronics, Inc.). Two depths (surface and bottom) were sampled in duplicate(144 discrete seawater samples) for nutrient analyses (ammonia NH4+, nitrite NO2-, nitrate NO3-, soluble reactive phosphorous SRP, soluble reactive silicate SRSi, total 4

ACCEPTED MANUSCRIPT phosphorous TP), chlorophyll a (Chl a) and CDOM. Water samples for dissolved nutrients were directly filtered on board through a syringe with Whatman GF/F filters and stored at -20°C for about two weeks until laboratory measurements. For Chl a, 1000 ml water samples were drawn directly from Niskin bottles into dark polyethylene bottles and stored at 4°C until laboratory

RI PT

filtration. For CDOM, 200 ml samples were drawn into polyethylene syringes and filtered through sterile Whatman GD/X 0.2 µm filters. All glassware and syringes were previously soaked with 10% HCl and thereafter rinsed with Milli-Q water. Samples for phytoplankton analysis weredirectly

SC

drawn into dark polyethylene bottles (500 ml) with 4% formaldehyde.

M AN U

Table 1. Latitude, longitude, sampling days and UTC times for the various sites. Note that sites A-E were always sampled on the same day. Dates are in MM/DD/YY format. Lat (°N)

Lon (°E)

02/23/13 time UTC

A B C D E Site

41.58 41.56 41.54 41.57 41.51 Lat (°N)

15.92 15.98 16.07 16.15 16.19 Lon (°E)

13.30 12:45 11:55 10:30 09:40 03/01/13 time UTC

M

41.94

15.42

10:00

04/23/13 time UTC

06/18/13 time UTC

08/21/13 time UTC

13:15 12:15 11:20 10:10 08:50 04/20/13 time UTC

13:10 12:30 11:20 10:30 09:20 06/19/13 time UTC

13:15 12:15 11:20 10:10 08:50 08/28/13 time UTC

09:30

09:30

10:10

10/20/13 12/12/13 time time UTC UTC 13:05 13:25 12:10 12:12 11:00 11:00 10:15 10:15 09:11 08:40 10/15/13 12/16/13 time time UTC UTC 10:20 09:10

AC C

EP

TE D

Site

2.2 River-wind data and current model output Data for the Candelaro and Ofanto rivers were measured just off their mouths at the observation stations of the Hydrographic and Mareographic Service of Puglia Regional Administration and were provided by the ”Centro Funzionale”. River data were supplied for the period from the 1st of January 2013 to the 31stof December 2013 and the general trend of their inputs was derived by plotting daily average data vs. time.

5

ACCEPTED MANUSCRIPT In situ wind data (speed and direction, April 24 - December 13, 2013) were obtained from a climatological station (Davis Vantage Pro2) mounted on the “Meda Gargano” fixed buoy located at site D (Fig. 1; see also http://rmm.fg.ismar.cnr.it/). Wind speed and direction were recorded every

to obtain a picture of local wind forcing during the sampling period.

RI PT

10 minutes and were analysed to derive the dominant wind statistics and daily mean winds, in order

The Meda Gargano wind data are complemented by wind data in the form of daily averaged gridded data for the Adriatic Sea provided by the ASCAT satellite scatterometer (IFREMER,

SC

Franceftp://ftp.ifremer.fr/ifremer/cersat/products/gridded/MWF/L3/ASCAT/Daily/Doc/DailyAscat Wind-Doc.pdf). The data consist of daily surface wind values in neutral conditions and have a

M AN U

spatial resolution of 0.25 × 0.25°. The ASCAT wind values are quality-controlled as described by the above documentation. Specifically, coastal pixels too close to the coast are already flagged in the wind products.

Model-based current data for the Mediterranean Sea were also obtained from the Marine

TE D

Environment Monitoring Service (CMEMS) database of the COPERNICUS project (marine.copernicus.eu). The currents are those of the Mediterranean Forecasting System (Oddo et al., 2009) in coupled hydrodynamic-wave mode, with a 1/16˚ (ca. 6-7 km) horizontal resolution and

EP

72 unevenly spaced vertical levels (CMEMS product id.: MEDSEA_ANALYSIS_FORECAST_PHYS_006_001). The model is forced by momentum, water

AC C

and heat fluxes computed by bulk formulae. The bulk formulae in turn utilise the 6-hour, 0.125° horizontal-resolution operational analysis and forecast fields provided by the European Centre for Medium-Range Weather Forecasts (ECMWF) and the model predicted surface temperatures (Tonani et al., 2008) as well as river runoff data. We present daily maps of the currents at the depths of 1.47 m and 15 m, i.e. near-bottom for the Gulf of Manfredonia, so as to illustrate the main dynamic regimes of the current field, determined by the wind and the WAC. We chose to examine the experimental ASCAT winds together with the model currents, and not the ECMWF model forcing fields, because (1) the ASCAT wind data are real data, (2) the very good agreement 6

ACCEPTED MANUSCRIPT observed in the current behaviour with what one might expect, given the corresponding ASCAT wind regime, represents an additional validation of the current model, (3) the ECMWF fields are also obtained by assimilating ASCAT winds (Hersbach et al., 2007).

RI PT

2.3 Laboratory analyses

Temperature and salinity data were measured by sensors periodically calibrated at Sea-Bird Electronics Inc. and processed using the SeaSoft suite of standard processing programs (SBE Data

SC

Processing, version 5.25).The accuracy of the data was ±0.02°C for temperature and ±0.02 for salinity.

M AN U

Dissolved nutrients concentrations were measured using a three-channel continuous flow analyser (Autoanalyzer 3, Bran+Luebbe, Norderstedt, Germany), following standard procedures reported in the manual (Bran and Luebbe, 2004). TP was analysed in the same way using unfiltered water, after persulphate digestion (Grasshoff et al., 1999). Instrument LOD (based on EPA

TE D

procedure pt. 136, app. B) were: 0.03 µM for ammonia; 0.008 µM for nitrite; 0.02 µM for nitrite + nitrate; 0.02µM for soluble silicate; 0.01 µM for soluble phosphorous. The reproducibility of the measurements was evaluated by an intercalibration exercise (Quasimeme). Dissolved inorganic

AC C

are reported in µM.

EP

nitrogen (DIN) was calculated as the sum of ammonia, nitrite and nitrate concentrations. All values

Chl a pigments were extracted from Whatman GF/F filters with 90% acetone and measured using the fluorometric method in accordance with EPA445.0 (1997). Pigment fluorescence was measured with a Trilogy Laboratory Fluorometer (Turner Designs, V. 1.2), after calibration with a commercially available Chl a standard (Sigma Aldrich). Also, phytoplankton cells were counted and identified by inverted microscope following Utermöhl’s (1958)method. CDOM absorbance was measured using a dual beam UV-VIS spectrophotometer (SHIMADZU 2600 Series)with a 10-cm quartz cuvette and the following settings: wavelength

7

ACCEPTED MANUSCRIPT range 250-750 nm; scan speed medium; sampling interval 0.5-1 nm; slit width 0.1 nm. Milli-Q water was used as a reference. A baseline correction, compliant to Babin et al. (2003),was applied to the absorbance measurements by subtracting the average absorbance value for the range 680-690 nm from all spectral values. All absorbance values were converted to absorption coefficients

RI PT

(αCDOM, m-1) as follows α (λ) = 2.303 * A(λ)/l, where A is the absorbance at wavelength λ; l is the path length of the optical cell in metres (in this case 0.10 m). The absorbance measurements considered here ranged from 250 to 450 nm, representing wavelengths where the signal was

SC

sufficiently high for a reliable estimation of αCDOM (Berto et al., 2010; Vignudelli et al., 2004).

(humic-like CDOM).

2.4 Satellite SST and ocean colour data

M AN U

Specifically, we calculated absorption coefficients at 280 nm (protein-like CDOM) and 355 nm

The satellite Sea Surface Temperature (SST) values shown in this study are based on L3S

TE D

SST data for the Mediterranean Sea obtained by merging MyOcean multi-sensor (MODIS, METOP, SEVIRI, etc.) and L3P single-sensor data and remapping the results over the Mediterranean Sea at ultra-high spatial resolution (UHR,0.01°i.e. about 1 km)

EP

(http://www.myocean.eu.org/; Buongiorno Nardelli et al., 2013).

AC C

Ocean colour data include chlorophyll (in mg m-3) measured by the MODIS AQUA satellite sensor; detrital and gelbstoff (CDOM or yellow substance) absorption coefficients at 412 nm (ADG, in m-1); particle backscattering coefficients at 547 nm (BBP, in m-1; Carder et al., 2005); and daily maps at high resolution (about 1km) of the Mediterranean Sea. The data were obtained from the MyOcean project data repository (http://www.myocean.eu.org/). Specifically, the chlorophyll maps were generated by means of the combined use of the open ocean (Case 1) Mediterranean Sea regional algorithm (MEDOC4, Volpe et al., 2007) and the coastal waters (Case 2) AD4 algorithm (D’Alimonte and Zibordi, 2003).Ocean colour parameter errors are about 20% of the measured value for chlorophyll (Volpe et al., 2012) and about 10% for ADG and BBP (Lee et al., 2010). 8

ACCEPTED MANUSCRIPT 2.5 Statistical treatment First, the degree of normality of the in situ measured variables’ distribution and the homogeneity of variance were calculated in order to choose whether to apply parametric or nonparametric statistics. The Kolmogorov-Smirnov and Shapiro-Wilk tests were used to verify the

RI PT

shape of the data distributions. Their statistics (W and D respectively) are based on the H0

hypothesis of no difference between normal and data distributions. The variables’ distribution in our study differed significantly from normality, requiring the adoption of non-parametric statistical

SC

methods.

Significant differences in variables among observation sites and periods were explored by

M AN U

means of the Kruskal-Wallis Test (Kruskal and Wallis, 1952), a non-parametric ANOVA for ranks, by calculating H (the test statistic) and p (the level of significance).Correlations between measured variables were tested by Spearman’s rank correlation coefficient R, first calculated for the entire data matrix and then for each sampling month. As variables had different units, all data were first

TE D

normalised to their maximum value and log-transformed (Clarke and Green, 1988). Statistical

3. Results

EP

analyses were performed with the STATISTICA 8.0 computer package (StatSoft Inc.).

3.1 River discharge, winds and currents

AC C

River discharge seems to be episodic, with isolated discharge peaks on Feb. 13-15, 2013 mainly for the Candelaro river, and Dec. 2-5, 2013 for both rivers (Figs. 2a, b). In addition, the Candelaro river maintains a more constant discharge "baseline" with levels of 0.25 to 0.5 m throughout the year (Fig. 2a). Finally the lowest values occur in June-August with the Candelaro water level < 0.3 m and the Ofanto flow< 5 m3 s-1. Next, the Meda Gargano weather station data indicate a predominance of easterly winds in 2013, i.e. the Bora winds blowing from the Balkans, followed by south-easterlies (Scirocco) and

9

ACCEPTED MANUSCRIPT westerlies (Fig. 2c), consistent with the dominant winds in the Adriatic Sea (Artegiani et al., 1997, Orlic et al., 1994). Examination of the entire 2013 ASCAT wind and current model datasets revealed that there are four basic dynamic configurations in the Gulf's current pattern, depending on the wind and

RI PT

stratification conditions (Fig. 3). Wind variability is described in the leftmost panels of Figure 3, in which one north-westerly wind (Fig. 3a), one very weak wind (Fig. 3b) and two southerly and south-westerly wind (Fig. 3c, d) situations are shown. We chose to present winds for the day before

SC

the day of the current maps, because of the time lag, in the order of at least one inertial period (i.e. about 18 hours at 40° latitude), of the wind’s effect on the ocean. Finally, for reasons of brevity,

M AN U

winds in the entire basin are not shown, although all wind patterns in Fig. 3 refer to the entire Adriatic Sea, save for the weak wind case in Fig. 3c, during which strong north-easterly (Bora) winds blew in the northern Adriatic, thus energising the WAC in its northern portion, and a southwesterly (Scirocco) regime present south of the area, which slows the WAC down.

TE D

The currents on February 19th were barotropic (Fig. 3a, centre and right) due to the absence of stratification in winter (see in situ data in Section 3.2 below). In addition, the NW wind regime (Fig. 3a. left) induced sea level rise near the coast, thus intensifying the WAC, which is seen to flow

EP

swiftly past the Gargano, but also to penetrate the Gulf, flowing around the peninsula. Therefore, this wind regime is favourable to the offshore migration of the Gulf's locally-produced riverine

AC C

waters. This barotropic current pattern also occurred in stratified conditions under a similar northwesterly wind regime, e.g. on June 13, associated with the June 11-13 NW wind event (not shown). Next, on Feb 21st (Fig. 3b, left) winds were weak in the area, but Scirocco winds affected the Adriatic Sea more to the south (not shown) and seemingly slowed the WAC down (compare currents in Figs. 3b and 3a). The Gulf seems isolated and stagnant in its entire water column (Fig. 3b centre and left). Finally, the current pattern in a southerly or south-easterly wind (Figs. 3c, d, left) reveals water renewal in the gulf, with surface waters flowing offshore to the north. This flow involved the entire water column in unstratified conditions on February 24th (Fig. 3c, centre and 10

ACCEPTED MANUSCRIPT right), probably with a compensation flow entering the Gulf from the eastern portion of the WAC, forming an anticyclonic pattern. This renewal does not occur in the same way when stratification is present, as on April 21st. In contrast, surface waters flow offshore (Fig. 3d, centre), but there is a compensation (inshore) flow from the south in the lower layer (Fig. 3d, right). To summarise,

RI PT

depending on wind and stratification, the Gulf's entire water column may be either isolated from offshore waters or "flushed" offshore by transiting (Fig. 3a) or recirculating (Fig. 3c) WAC waters, while in stratified conditions, only the surface waters may be "directly" removed by the wind, with

SC

a compensation flow involving salt-wedge intrusion of offshore waters.

M AN U

3.2 In situ hydrography

On February 22nd (Fig. 4a), the effect of riverine discharge (peaking on Feb. 13-15) was observed at inshore-most sites A and B, as a cold and relatively fresh (9.5 - 10.0 °C, S<36.5) surface layer, with values similar to those found at site M north of the Gargano peninsula (T = 10°C, 36.5

TE D


EP

stagnated there until Feb. 21st, the day before sampling (Figs. 3a, b). We now present the December data, in order to highlight winter variability. The peak

AC C

discharge of the Candelaro and Ofanto rivers on Dec. 2-5 (Fig. 2a) was higher than in February. This is probably why fresher and cooler waters (T<13°C, S<36.25, Fig. 4f) were observed along much of the transect on the sampling day. One reason for retention of fresh waters is that no significant flushing episodes occurred between Dec. 5 and the sampling day, according to the model current data (not shown, but similar to Fig. 3b). In addition, bottom waters may have penetrated the Gulf from offshore, given the warmer and saltier waters seen at the bottom of the more offshore sites (Fig. 4f). Lastly, the salinity of about 37 at station M, save for a thin fresher layer (Fig. 4f), indicates less freshwater in the water column. 11

ACCEPTED MANUSCRIPT In April (Fig. 4b), the saltier bottom-layer indicates the presence of intruding offshore waters, in agreement with subsurface (15 m) penetration by offshore waters seen in the model current pattern (Fig 3c, right). Surface layer salinities of around 37 indicate the presence of freshwater in this layer. At site M, a different situation was observed, with relatively fresh waters

RI PT

throughout the water column (S = 36.1 - 36.8) and a small salt wedge at the bottom (Fig. 4b). In June (Fig. 4c), strong thermal stratification was seen at both site M and along transect AE, with salinities again increasing with depth. The lowest salinities (S<37) are confined more

SC

inshore, with respect to April, in agreement with the lower river input in June (Fig. 4c). In addition, surface salinity at station M (S = 36; Fig. 4c, profile) was lower than in the Gulf, possibly indicating

M AN U

that the lagoons were still actively supplying fresh waters.

Lastly, summer warming of the entire water column caused the August and October temperatures (Figs. 4d, e) to be everywhere vertically homogeneous. Salinities progressively grew from 37 to 37.9 (Figs. 4 d, e), again indicating a fall in freshwater content, both in the Gulf and at

TE D

site M.

3.3 In situ biogeochemical properties

EP

The observed ranges of biogeochemical variables in the study area are shown in Table 2 and the significant differences among sites and periods in Table 3. Sites A, E and M (Fig. 1) were

AC C

identified as those displaying the highest variability and are therefore used in the rest of the paper to discuss this variability.

The highest DIN levels were observed in the surface layers of site M in February (20.56 µM) and site A in February (17.05µM) and December (15.31µM) (Table 2, Fig. 5a). Next, site E generally had lower DIN levels throughout the year, and site M did not display a maximum in December (Fig. 5a). Figure 5a also shows that: (1) surface DIN is about twice the bottom value in February-April and October-December; (2) DIN is generally low in June-October; (3) the DIN spatial and temporal patterns (Fig. 5a) mirror the salinity patterns (Fig. 4, left panels).Salinity was seen to be relatively 12

ACCEPTED MANUSCRIPT low in the Gulf in February and December (Fig. 4a, f), but was low at site M only in February (Fig. 4a), with DIN being high in the presence of such low salinities.

Table 2. Parameter ranges (minimum-maximum) observed at sites A, B, C, D, E and M during six samplings in 2013. N

February

April

June

August

October

December

36 36 144 144 144 144 144 144 72 144 144

9.27-14.16 35.00-37.50 4.27-20.56 3.35-13.39 0.46-10.77 0.01-0.26 0.06-0.46 0.39-9.92 0.07-1.53 0.21-2.04 0.06-0.45

15.9-18.60 36.65-38.30 1.17-10.85 0.51-5.17 0.32-4.04 0.00-0.05 0.10-0.46 1.39-5.87 0.07-0.37 1.08-2.67 0.18-0.83

18.29-25.50 36.07-38.25 0.18-1.77 0.02-0.86 0.95-7.36 0.01-0.21 0.13-0.61 0.38-1.94 0.03-0.13 1.32-1.91 0.05-0.39

25.65-26.71 36.08-38.01 0.02-1.35 0.00-0.88 1.75-3.88 0.00-0.40 0.08-0.55 0.17-3.81 0.08-0.53 0.99-1.67 0.17-0.38

20.76-21.31 36.72-37.89 0.00-2.46 0.00-2.33 1.01-5.52 0.00-0.13 0.08-0.35 0.48-2.28 0.03-0.12 0.86-1.73 0.07-0.58

9.63-14.42 34.85-37.26 3.28-15.31 4.41-12.91 2.15-13.08 0.00-0.24 0.13-0.33 1.33-3.98 0.14-0.30 1.11-2.56 0.01-0.82

SC

N

M AN U

T °C S DIN µM N-NO3 µM SRSi µM SRP µM TPµM Chl amg m-3 Phytoplankton 106cell. L-1 α280m-1 α355 m-1

RI PT

indicates number of CTD casts and paired surface and bottom water samples.

Table 3. Significance of temporal and spatial variability for biogeochemical variables observed in the study area in 2013

TE D

(Kruskal-Wallis test H and p). NS = not significant.

Period H 49.80 48.94 22.04 16.40 5.07 35.52 13.94 22.26 14.87

AC C

EP

DIN N-NO3 SRSi SRP TP Chl a Phyto α280 α355

n 144 144 144 144 144 144 72 144 144

p-value <0.001 <0.001 <0.001 <0.01 NS <0.001 <0.05 <0.001 <0.05

n 144 144 144 144 144 144 72 144 144

Site H 11.32 12.11 14.29 10.31 29.35 6.31 6.53 12.89 12.39

p-value <0.05 <0.05 <0.05 NS <0.001 NS NS <0.05 <0.05

The reverse is also true, with low DIN and higher salinities in the rest of the cases (Fig. 4, left panels and Fig. 5a) and all this quantitatively translates into a general negative (p<0.05) correlation between DIN and salinity for the entire dataset (Fig. 5e, left). Lastly, nitrate was the

13

ACCEPTED MANUSCRIPT most abundant compound among all the nitrogen salts, accounting for 67% of the total, with temporal evolution very similar to DIN shown in Fig. 5a. Silicates also exhibited winter highs in February (10.77 µM) and December (13.08 µM) and summer lows in June-August (2.24 µM) at site A (Fig. 5b, left), while no winter maxima were

RI PT

observed at the outermost site E, nor at site M north of the Gargano (Fig. 5b, centre and right). Increasing surface silicate values from February to December were observed at sites E (1.36 µM to 5.99 µM; Fig. 5b centre) and M (0.46 µM to 3.61 µM; Fig. 5b right), with the February - June

SC

values of site E and February - April values of site M being higher at the bottom than at the surface. Similarly to DIN, the dataset as a whole shows a significant (p<0.05) inverse correlation with

M AN U

salinity (Fig. 5e, right).

Soluble reactive phosphorous showed neither a well-defined temporal pattern (Fig. 5c), nor any correlation with salinity (not shown), the latter fact suggesting supply mechanisms other than riverine input. At site A, concentrations oscillate between 0 and 0.1-0.2 µM, with the highs

TE D

occurring in February (0.209 µM), June (0.145µM) and October (Fig. 5c, left). At site E, concentrations close to zero were observed at all sampling times, except in June (0.049 µM, surface) and August (0.071 µM, bottom) (Fig. 5c, centre). Site M was characterised by near-zero

EP

SRP values from February to June and relatively high concentrations from August to December (about 0.130 µM; Fig. 5c, right). Similarly, TP reached its bottom higher concentrations at sites A

AC C

(0.61 µM) in June and site M (0.55 µM) in August, and it decreased with distance from the shore, from site A to site E.

The highest concentrations of Chl a were observed in February at site M (9.92 mg m-3) and April at site A (5.87 mg m-3) (Fig. 5d, left and right; Table 2), with significant (p< 0.001) temporal differences between sample times (Table 3). Site A also displayed a peak of Chl a in August (3.76 mg m-3) (Fig. 5d, left). A strict correspondence between Chl a highs and SRP lows and vice versa is observed at all sites (Figs. 5c, d), as also confirmed by significant negative relationships (p<0.01)

14

ACCEPTED MANUSCRIPT between Chl a and SRP. No correlation was observed between Chl a and salinity, except in December (R = -0.72). The protein-like CDOM absorption coefficient α280 (Fig. 6a) fell progressively from April (2.67 m-1) to October (about 1 m-1) at site A, rising again in December (2.56m-1). A similar pattern

RI PT

was observed at site M, while at site E, the seasonal drop was more attenuated, probably due to mixing with marine waters of more constant α280value. Interestingly, α280 values were generally higher in April than in February, possibly indicating an increase in CDOM related to primary

SC

production.

As for the humic-like CDOM absorption coefficient α355, higher values were measured in

M AN U

surface waters in December (0.57m-1) at site A and in April (0.59 m-1) and June (0.39 m-1) at site M (Fig. 6b). Both coefficients α280 and α355 were positively correlated with Chl a (p<0.05; compare Fig. 5d to Figs. 6a, b).

Lastly, analysis of α280, α355 vs. salinity relationship performed on the entire dataset once

TE D

again yields no correlation, except for December for which the few data available indicate a significant correlation for α280 and surface α355 (Fig. 6c). Indeed, the correlation between α280 and

EP

salinity is technically , R=0.629 (p<0.001). However, given the small number of data points (12), it only suggests a relationship and thus requires further investigation, e.g. via multi-year winter-

AC C

month datasets, to establish whether river discharge actually controls CDOM content in the Gulf and at site M. The relationship is nevertheless probable, because neither February, with a more modest freshwater presence in the Gulf, nor the lower river discharge months (Apr. - Oct.) show any qualitatively or quantitatively significant correlation α280, α355 .

3.4 Phytoplankton analysis The highest phytoplankton density(cellsL-1) was observed in February (Table 2), with diatoms being dominant in terms of both abundance (50-80%) and number of species at all 15

ACCEPTED MANUSCRIPT sampling sites. Site A showed a moderate phytoplankton peak in April (0.22 x 106 cells L-1), explaining the drop in DIN and SRSi (Figs. 5a and b), followed by another one in August (0.47x 106 cells L-1), with Chaetoceros spp. (0.16 x 106 cells L-1) and Pseudonitzschia spp. (0.2x 106cells L-1) as the dominant genera in April and August respectively. At sites E and M, diatom peaks

RI PT

occurred in February with Skeletonema sp. as the most abundant genus(0.24 x 106 cells L-1 and 1.3 x 106 cells L-1 respectively). Significant monthly differences in the phytoplankton community were

SC

observed (Table 3), with the community’s temporal distribution very similar to that of Chl a.

3.5 Satellite data

M AN U

In February, the satellite detected SSTs are coherent with in situ T values: comparing SST in Fig. 7a to in situ T in Fig. 4a one finds9.8
TE D

waters into the Gulf observed in the February current pattern, albeit partial and with low velocities (Fig. 3a, bottom).

On the other hand, the February satellite data show relatively chlorophyll-rich WAC waters

EP

flowing southwards and spilling over the Gargano peninsula but not entering the Gulf, (Fig. 7b). Both in situ and satellite-detected chlorophyll are highest at station M north of the Gargano, in the

AC C

WAC system (in situ Chl = 9 mg m-3, sat. Chl = 3.8 mg m-3) and lower in the Gulf (in situ Chl = 1-2 mg m-3 and sat. Chl = 0.8 - 1.0 mg m-3). Thus it seems that high WAC chlorophyll values are confined to the main stream of this current, while cold, but more oligotrophic, WAC waters are able to intrude into the Gulf. The variability of the winter hydrographic situation in the Gulf is revealed by the contrast between the cold, oligotrophic waters observed in February and the cold isolated plume observed on December 4 (Fig. 7c), the closest day to the sampling. This plume is possibly a signature of the above-mentioned December river discharge peaks (Fig. 7c), confirmed by the low salinity observed 16

ACCEPTED MANUSCRIPT in the Gulf's in situ data (Fig. 3f), and maintained within the Gulf in December by the dynamic isolation mentioned above. The April SST images show that the WAC was thermally “fragmented", with its cold core sometimes far from the coast (Fig.7e), in agreement with the south-easterly wind-induced

RI PT

dynamics, including offshore-oriented currents and WAC slowdown (Fig. 3d). Moreover a

sequence of chlorophyll images acquired around the in situ sampling day and exemplified in Figure 7f, highlights chlorophyll-rich WAC waters around the Gargano. However, a closer view (Fig. 7f,

SC

inset) also shows chlorophyll-rich plumes in the Gulf, seemingly isolated from the offshore waters. Next, no marked onshore-offshore differences in SSTs are visible in June (Fig. 7g), due to

M AN U

the summer thermal homogeneity of the WAC (Po river-rich) and offshore waters. The chlorophyll patterns (Fig. 7h) reveal many local instabilities and plumes along the WAC pathand that WAC values are slightly higher than those in the Gulf (Fig. 7h, inset).

In August and October, SST (not shown) and ocean colour data show warmer and

TE D

chlorophyll-rich waters (Fig. 8a) in both the Gulf and site M. The high chlorophyll values in the Gulf are found to extend offshore in a curved plume (Fig. 8a;see also Burrage et al., 2009). Given the observed current patterns with the main body of the WAC flowing over the Gargano peninsula,

EP

the plume may well have been caused by entrainment of the Gulf waters by the WAC itself. We conclude this description of remote sensing data by briefly describing the variability in

AC C

the area of two biogeochemical parameters which have been made available recently as satellite products, i.e. the 412 nm CDOM absorption coefficient (ADG) and particulate backscattering (BBP) patterns. The ADG and BBP patterns in August (Figs. 8b, c) show that the plume in Figure 8a is rich in CDOM and particulate. However, unlike chlorophyll, both parameters show much lower values in the offshore portion of the plume, suggesting that CDOM is retained in the Gulf and that particulate deposition occurs almost exclusively within the Gulf. The same occurs in October (Figs. 8d,e), thus suggesting that the plume is an important biogeochemical factor for the Gulf.

17

ACCEPTED MANUSCRIPT 4. Discussion 4.1 River input hydrographic effects in the gulf The observations and model output described above indicate that the winter hydrography of the Gulf may be highly variable, and that the composition of its waters is controlled by the intensity

RI PT

of river discharge events as well as winds and currents. That is, a more WAC-like hydrography is observed in February, when less intense river discharge events occur (Fig. 2) and the WAC is probably (slightly) more penetrating (Fig. 3a, bottom), than in December. In addition, the higher

SC

salinities in February (Fig. 4a) may also be due to the above-mentioned exchange events between the Gulf and offshore waters, two of which occur in February. In contrast, the last flushing events of

M AN U

the year occur in November, i.e. before the December peak river discharge, thus leaving the riverine fresh waters in the Gulf undisturbed (Fig. 4f).

In periods of the year that are not affected by peak discharge events, the influence of rivers declines and a more marine regime prevails, with contributions from the WAC and offshore waters.

(Berto et al., 2010).

TE D

These waters intrude in the bottom layer (April and June), which becomes progressively saltier

This qualitative/conceptual identification of the hydrodynamic agents conditioning the

EP

Gulf's ecosystem represents an initial effort. A more quantitative analysis needs to be implemented, assessing the Gulf's freshwater budget and local water renewal, regulated by riverine input and

AC C

exchange with the WAC system at the Gulf's offshore limit. This may be accomplished by the combined use of models and in situ data, as is the case here. In addition, the analysis may be enriched by data from an in situ current meter data (installed in April 2013; http://rmm.fg.ismar.cnr.it/index.php/meda-gargano), as well as high-frequency radar data for the surface currents in the area (Corgnati et al., 2015), available for the period June 2013-June 2015.

18

ACCEPTED MANUSCRIPT 4.2 River discharge and biogeochemical properties Dissolved nitrogen (DIN) and silicate (SRSi) concentrations are evidently influenced by peak riverine discharge in February and December and low input in the rest of the year, as occurs in other estuarine coastal zones (Alber, 2002; Buzzelli et al., 2014; Tremblay et al., 2014). This is

RI PT

shown by the greater differences between winter and summer at the more coastal sites A and M, more directly impacted by the Candelaro river and the Lesina Lagoon, respectively. It is also interesting to note that at the offshore-most site E, (1) bottom silicate is higher than at the surface,

SC

and (2) surface silicate increases until it reaches bottom values in August. This increase may be due to either remineralisation after the diatom-dominated spring blooms or to benthic supply (Marz et

M AN U

al., 2015).

Phosphorus on the other hand seems to be independent from freshwater input, but oscillates in opposite phase with chlorophyll. This indicates removal due to phytoplankton uptake and subsequent release after the blooms, as well as P regenerated from sediments, accompanied by

TE D

vertical mixing, as also found in other coastal waters by Roselli et al. (2009). Winter nutrient enrichment leads to a typical spring phytoplankton peak in the Gulf, more marked in the coastal zone, with a prevalence of diatoms. It is interesting to note, however, that in

EP

the following warmer months, falls in DIN and SRSi concentrations did not affect densities of phytoplankton cells, especially at inshore site A, probably because levels of phosphorous were

AC C

sufficient to support a further algal peak (August). To summarise, phytoplankton is ecologically efficient at removing riverine-derived nutrients (especially DIN) in the study area, but also in maintaining itself in nutrient-depleted months (Tantanasarit et al., 2013). Finally, the inverse correlation between CDOM absorption coefficients and salinity in December in the Gulf, as well as the winter-spring decoupling of surface and bottom values of both a280 and a355, confirm riverine influence in the cold season, a situation widely observed in the northern Adriatic (Berto et al., 2010). In contrast, during the spring-summer months, a positive correlation between Chl a and the CDOM absorption coefficients in the Gulf suggests that CDOM 19

ACCEPTED MANUSCRIPT is directly related to marine biological processes (Blough and Del Vecchio, 2002). In summer, a decrease in CDOM is observed at all sites along the transect, indicating suitable conditions for the establishment of organic matter degradation processes enhanced by high temperatures.

RI PT

4.3 Ecological Implications

We have seen that the Gulf's dynamic and physical hydrographic situation is modulated by riverine input, wind-driven and WAC-driven water renewal/replacement ("flushing"), with further

SC

complexity introduced by the presence or absence of stratification. Therefore, the Gulf's ecological stability may well be compromised, from the physical point of view, by changes in the intensity of

M AN U

these forcing agents. That is, a marked increase in freshwater input in the Gulf may be induced by climatic changes in meteo-marine conditions, such as the presently observed increase in extreme rainfall events in the Mediterranean area (Krichak et al., 2016), which could change the salt budget the area. Changes in the wind pattern could also alter the present rate of water renewal in the Gulf,

TE D

for example due to the weakening and reduced frequency of the Scirocco-induced flushing events described above. Finally, these climatic alterations may change the Gulf's biogeochemical budget, e.g. by reducing the Gulf's capacity to rid itself of toxic agents and or excess nutrients released by

EP

rivers which could cause eutrophication.

Specifically, all these potential threats could the jeopardise the role of the Gulf of

AC C

Manfredonia as an important Adriatic Sea nursery zone for small pelagic fish (Borme et al. 2013). Indeed, this seems to be already occurring, in that Morello and Arneri (2009) have attributed the 75%decrease in the average total catch of sardine juveniles (“bianchetto”) from the 1980s to the 1990s to both biotic and abiotic factors.

5. Conclusions The hydrological and biogeochemical variability in the Gulf of Manfredonia and the coastal region around the Gargano peninsula (southwestern Adriatic Sea) during 2013 was investigated, in 20

ACCEPTED MANUSCRIPT order to assess the present status of the area's ecosystem, which is an important nursery site for the Adriatic Sea, as well as a key tourism asset for Italy. The examination of time series data on river input and wind-driven currents in the area revealed the main dynamic patterns in the study area, with relatively rare flushing events and frequent isolation of the Gulf from offshore waters, sealed

RI PT

off by the WAC.

The in situ and remotely sensed biogeochemical datasets examined in this work confirmed the close dependence of the inshore-most waters in the Gulf and off the Gargano on

SC

riverine/lagoonal nutrients and other biogeochemical input, especially via the nutrient salt patterns and humic-like CDOM absorption coefficient α355, which proved to be a good tracer of terrigenous

M AN U

fresh waters.

This study represents an initial exploration of the area. The follow-up to this work must entail a more quantitative/statistical assessment of the variability of forcing factors in the area. Future work should quantitatively address the variability of hydrographic and biogeochemical

TE D

budgets in the Gulf and off the Gargano, via multi-annual time series of the above-described data and model output, combined with the above-mentioned HF radar and in situ current meter data, as well as data on the upper levels of the trophic chain, such as sardine juvenile and adult stocks. This

EP

would provide a solid scientific basis for future monitoring activities (such as those linked to the

AC C

EU’s Marine Strategy Framework Directive), by establishing the present ecological status, status trend and "tolerance" to change of the Manfredonia-Gargano ecosystem, which in turn would provide adequate "alarm thresholds" for the marine parameters to be monitored.

Author contributions Conceived and designed the experiments: AS, RD, MM, AF. Performed the experiments: TS, ASan, PP. Analysed the data: AS, FB, AF,TS, AC. Contributed reagents/materials/analysis tools: AS, ASan, TS. Wrote the paper: AS, RD, FB, AF. Performed statistical analysis: AF,AS. Developed ocean colour maps: FB. Supported revision and made intellectual contribution: AS,RD, FB, MM. 21

ACCEPTED MANUSCRIPT

Acknowledgements This study is a contribution to the “SSD-Pesca” project on fisheries in Southern Italy (see http://mezzogiorno.cnr.it/), funded by the Italian Ministry of Finance and Economy, and to a pilot

RI PT

project funded by Puglia Regional Administration within the European Fisheries Fund (2007-2013), with a grant to Dr. Raffaele D’Adamo. Special thanks go to the “Centro Funzionale” of Puglia Regional Administration for supplying data on the Candelaro and Ofanto rivers. We are also very

SC

grateful to Maddalena Maselli (CNR – ISMAR) for her essential help in phytoplankton cell counts. Thanks are also due to the Ocean Sciences Group of CNR-ISAC (headed by Dr. R. Santoleri) and to

M AN U

the NASA Jet Propulsion Laboratory (JPL, http://podaac.jpl.nasa.gov) for providing SST and ocean

AC C

EP

TE D

colour data for the Mediterranean Sea and the ASCAT wind data respectively.

22

ACCEPTED MANUSCRIPT References Alber, M., 2002. A conceptual model of estuarine freshwater inflow management. Estuaries 25, 1246-1261. Artegiani, A., Bregant, D., Paschini, E., Pinardi, N., Raicich, F., Russo, A., 1997. The

RI PT

Adriatic Sea general circulation. Part I: Air-sea interactions and water mass structure. J. Phys. Oceanogr. 27, 1492-1514.

Babin, M., Stramski, D., Ferrari, G. M., Claustre, H., Bricaud, A., Obolensky, G.,

SC

Hoepffner, N., 2003. Variations in the light absorption coefficients of phytoplankton, nonalgal

(C7), 3211 doi:10.1029/2001JC000882.

M AN U

particles, and dissolved organic matter in coastal waters around Europe. J. Geophys. Res., p. 108

Berto, D., Giani, M., Savelli, F., Centanni, E., Ferrari, C.R., Pavoni, B., 2010. Winter to spring variations of chromophoric dissolved organic matter in a temperate estuary (Po River, northern Adriatic Sea). Mar. Environ. Res. 70, 73-81.

turbid

waters

from

TE D

Bignami, F., Sciarra, R., Carniel, S., Santoleri, R., 2007. Variability of Adriatic Sea coastal SeaWIFS

10.1029/2006JC003518.

imagery.

J.

Geophys.

Res.

1123

(C03S10),

doi:

EP

Blough, N.V., Del Vecchio, R., 2002. Chromophoric dissolved organic matter (CDOM) in the coastal environment. In D. Hansell, C. Carlson, (Eds.), Biogeochemistry of Marine Dissolved

AC C

Organic Matter. Academic Press, San Diego, CA, pp. 509-546. Borme, D., Tirelli, V., Palomera, I., 2013. Feeding habits of European pilchard late larvae in a nursery area in the Adriatic Sea. J. Sea Res. 78: 8-17. doi: /10.1016/j.seares.2012.12.010. Bowers, D.G., Brett, H.L., 2008. The relationship between CDOM and salinity in estuaries: An analytical and graphical solution. J. Mar. Syst. 73(1-2), 1-7. Bran and Luebbe, 2004. QuAAtro Applications for nutrients analysis in water and seawater. Methods No. Q-033-04, No. Q-030-04, No. Q-035-04, No. Q-031-04, No. Q-038-04. Bran+Luebbe, Noderstedt, Germany, pp. 62. 23

ACCEPTED MANUSCRIPT Buongiorno Nardelli, B., Tronconi, C., Pisano A., Santoleri, R. , 2013. High and Ultra-High resolution processing of satellite Sea Surface Temperature data over Southern European Seas in the framework of MyOcean project. Rem. Sens. Env. 129, 1-16.

Current: analysis and prediction. J. Mar. Sys., 78, S205–S226.

RI PT

Burrage, D.M., Book, J.W., Martin, P.J., 2013. Eddies and filaments of the Western Adriatic

Buzzelli, C., Doering, P.H., Wan, Y., Sun, D., Fugate, D., 2014. Modeling ecosystem processes with variable freshwater inflow to the Caloosahatchee River Estuary, southwest Florida.

SC

I. Model development. Estuar. Coast. Shelf Sci.151, 256-271.

Cherukuru, N., Brando, V.E., Schroeder, T., Clementson, L.A., Dekker, A.G., 2014.

188-203.

M AN U

Influence of river discharge and ocean currents on coastal optical properties. Cont. Shelf Res.84,

Clarke, K.R., Green, R.H., 1988. Statistical design and analysis for a 'biological effects' study. Mar. Ecol. Prog. Ser. 46, 213-226.

TE D

Corgnati, L., Mantovani, C., Griffa, A., Bellomo, L., Carlson, D.F., Magaldi, M.G., Berta, M., Pazienza, G., D’Adamo, R., 2015. The ISMAR High Frequency Coastal Radar Network: Monitoring Surface Currents for Management of Marine Resources. Proceedings of the OCEANS

EP

MTS/IEEE Conference, Genova, Italy.

D’Alimonte, D., Zibordi, G., 2003. Phytoplankton Determination in an Optically Complex

AC C

Coastal Region Using a Multilayer Perceptron Neural Network. IEEE Trans. Geosci. Remote Sensing, 41, pp. 286.

De Girolamo, A.M., Lo Porto, A., Pappagallo, G., Tzoraki, O., Gallart, F., 2014. The hydrological status concept: application at a temporary river (Candelaro, Italy). River Res. Appl. doi: 10.1002/rra.2786. EPA Method 445.0, 1997. In vitro determination of chlorophyll a and pheophytin a in marine and freshwater Algae by fluorescence. National Exposure Research Laboratory, Office of Research and Development, U.S. Environ. Protection Ag., Cincinnati, Ohio, p. 22. 24

ACCEPTED MANUSCRIPT Erga, S.R., Ssebiyonga, N., Frette, Ø., Hamre, B., Aure, J., Strand, Ø., 2012. Dynamics of phytoplankton distribution and photosynthetic capacity in a western Norwegian fjord during coastal upwelling: Effects on optical properties. Estuar. Coast. Shelf Sci.97, 91-103. Focardi, S., Specchiulli, A., Spagnoli, F., Fiesoletti, F., Rossi, C., 2009. A combinated

RI PT

approach to investigate the biochemistry and hydrography of a shallow bay in the South Adriatic Sea: the Gulf of Manfredonia (Italy). Environ. Monit. Assess. 153:209–220.

Grasshoff, K., Kremling, K., Ehrhardt, M. (Eds.), 1999. Methods of seawater analysis.

SC

Wiley-VCH, Weinheim, pp. 203-205.

Hersbach, H.P. Janssen, 2007. Operational assimilation of surface wind data from the

M AN U

MetOp ASCAT scatterometer at ECMWF. ECMWF Newsletter, 113, 6-8.

Krichak, S.O., Feldstein, S.B., Alpert, P., Gualdi, S., Scoccimarro, E., Yano, J.-I., 2016. Discussing the role of tropical and subtropical moisture sources in cold season extreme precipitation events in the Mediterranean region from a climate change perspective, Nat. Hazards Earth Syst.

TE D

Sci., 16, 269-285, doi:10.5194/nhess-16-269-2016.

Kruskal, W.H., Wallis, W.A., 1952. Use of ranks in one-criterion variance analysis. J. Am. Stat. Assoc. 47 (260), 583–621.

EP

Lee, Z.P., Arnone, R., Hu, C., Werdell, P.J., Lubac,B.2010. Uncertainties of optical parameters and their propagations in an analytical ocean color inversion algorithm. Appl. Optics, 49

AC C

(3), 369-381.doi: 10.1364/AO.49.000369. Lipizer, M., Partescano, E., Rabitti, A., Giorgetti, A., Crise, A., 2014. Qualified temperature, salinity and dissolved oxygen climatologies in a changing Adriatic Sea. Ocean Sci. 10, 771-797. Mancinelli, G., Vizzini, S., 2015. Assessing anthropogenic pressures on coastal marine ecosystems using stable CNS isotopes: State of the art, knowledge gaps, and community-scale perspectives. Estuar.Coast. Shelf Sci. 156, 195-204.

25

ACCEPTED MANUSCRIPT Marini, M., Jones, B.H., Campanelli, A., Grilli, F., Lee, C.M., 2008. Seasonal variability and Po river plume influence on biochemical properties along western Adriatic coast. J. Geophys. Res.: Oceans (1978–2012) 113 (C5). doi: 10.1029/2007JC004370. Marini, M., Campanelli, A., Sanxhaku, M., Kljajic, Z., Betti, M., Grilli1, F., 2015. Late

RI PT

spring characterization of different coastal areasof the Adriatic Sea. Acta Adriatica, 56, 27-46. Marz, C., Meinhardt, A.K., Schnetger, B., Brumsack, H.J., 2015. Silica diagenesis and benthic fluxes in the Arctic Ocean. Mar. Chem. 171: 1-9.

SC

http://dx.doi.org/10.1016/j.marchem.2015.02.003

Monticelli, L.S., Caruso, G., Decembrini, F., Caroppo, C., Fiesoletti, F., 2014. Role of

M AN U

prokaryotic biomasses and activities in carbon and phosphorus cycles at a coastal, thermohaline front and in offshore waters (Gulf of Manfredonia, Southern Adriatic Sea). Microb. Ecol. 67, 501519.

Morello, E.B., Arneri, E., 2009. Anchovy and sardine in the Adriatic Sea — an ecological

TE D

review. In: Gibson, R.N., Atkinson, R.J.A., Gordon, J.D.M. (Eds.), Oceanogr. Mar. Biol. – an Annual Review, 47, 209–256.

Oddo, P., Adani, M., Pinardi, N., Fratianni, C., Tonani, M., Pettenuzzo, D., 2009. A Nested

EP

Atlantic-Mediterranean Sea General Circulation Model for Operational Forecasting. Ocean Sci. Discuss., 6, 1093-1127.

AC C

Organelli, E., Bricaud, A., Antoine, D., Matsuoka, A., 2014. Seasonal dynamics of light absorption by chromophoric dissolved organic matter (CDOM) in the NW Mediterranean Sea (BOUSSOLE site). Deep-Sea Res. I. 9, 172-85. Orlić, M., Kuzmić, M., Pasarić, Z., 1994. Response of the Adriatic Sea to the bora and sirocco forcing. Continental Shelf Research, 14(1), 91-116. Para, J., Coble, P.G., Charrière, B., Tedetti, M., Fontana, C., Sempéré, R., 2010. Fluorescence and absorption properties of chromophoric dissolved organic matter (CDOM) in

26

ACCEPTED MANUSCRIPT coastal surface waters of the northwestern Mediterranean Sea, influence of the Rhône River. Biogeosciences 7, 4083-4103. Penna, N., Capellacci, S., Ricci, F., 2004. The influence of the Po River discharge on phytoplankton bloom dynamics along the coastline of Pesaro (Italy) in the Adriatic Sea. Mar.

RI PT

Pollut. Bull. 48: 321-326.

Philips, E.J., Badylak, S., Hart, J., Haunert, D., Lockwood, J., O'Donnell, K., Sun, D., Viveros, P., Yilmaz, M., 2012. Climatic influences on autochthonous and allochthonous

SC

phytoplankton blooms in a subtropical estuary, St. Lucie Estuary, Florida, USA. Estuaries Coasts 35 (1), 335-352.

M AN U

Retelletti Brogi, S., Gonnelli, M., Vestri, S., Santinelli, C., 2015. Biophysical processes affecting DOM dynamics at the Arno river mouth (Tyrrhenian Sea). Biophys. Chem. 197, 1-9. Roselli, L., Fabbrocini, A., Manzo, C., D’Adamo, R., 2009. Hydrological heterogeneity, nutrient dynamics and water quality of a not-tidal lentic ecosystem (Lesina lagoon, Italy). Estuar.

TE D

Coast. Shelf Sci. 84, 539-552. doi:10.1016/j.ecss.2009.07.023.

Sasaki, H., Miyamura, T., Saitoh, S., Ishizaka, J., 2005. Seasonal variation of absorption by particles and colored dissolved organic matter (CDOM) in Funka Bay, southwestern Hokkaido,

EP

Japan. Estuar. Coast. Shelf Sci. 64, 447-458.

Tantanasarit, C., Englande, A., and Babel, S., 2013. Nitrogen, phosphorus and silicon uptake

AC C

kinetics by marine diatom Chaetoceros calcitrans under high nutrient concentrations. J. Exp. Mar. Biol. Ecol. 446, 67–75. doi: 10.1016/j.jembe.2013.05.004. Tonani, M., Pinardi, N., Dobricic, S., Pujol, I., Fratianni, C., 2008. A high-resolution freesurface model of the Mediterranean Sea. Ocean Sci., 4, 1-14." Totti, C., Civitarese, G., Acri, F., Barletta, D., Candelari, G., Paschini, E., Solazzi, A., 2000. Seasonal variability of phytoplankton populations in the middle Adriatic sub-basin. J. Plankton Res. 22(9), 1735-1756.

27

ACCEPTED MANUSCRIPT Tremblay, J.É, Raimbault, P., Garcia N., Lansard, B., Babin, M., Gagnon, J., 2014. Impact of river discharge, upwelling and vertical mixing on the nutrient loading and productivity of the Canadian Beaufort Shelf. Biogeosciences 11, 4853-4868.

Mitt. Int. Verein. Theor. Angew. Limnol. 9, 1-38.

RI PT

Utermöhl, H., 1958. Zur Vervollkommnung der quantitativen Phytoplankton-Methodik.

Vignudelli, S., Santinelli, C., Murru, E., Nainnicini, L., Seritti, A., 2004. Distributions of dissolved organic carbon (DOC) and chromophoric dissolved organic matter (CDOM) in coastal of

SC

the northern Tyrrhenian Sea (Italy). Estuar. Coast. Shelf Sci. 60 (1), 133-149.

Volpe, G., R. Santoleri, V. Vellucci, M. Ribera d’Alcalà, S. Marullo and F. D’Ortenzio

M AN U

(2007). The colour of the Mediterranean Sea: Global versus regional bio-optical algorithm sevaluation and implication for satellite chlorophyll estimates. Remote Sens. Environ. 107, 625638.

Volpe, G., Colella, S., Forneris, V., Tronconi, C., Santoleri, R. 2012. The Mediterranean

EP AC C

883.

TE D

Ocean Colour Observing System – system development and product validation. Ocean Sci., 8, 869–

28

ACCEPTED MANUSCRIPT

Figure captions

RI PT

Figure 1 -The study area and its position in the Adriatic Sea (inset, with main surface circulation redrawn, after Fig. 1 of Lipizer et al., 2014); red circles: sampling station M north of the Gargano peninsula and the stations A-E along the transect in the Gulf of Manfredonia. The Meda Gargano

SC

buoy (light blue square, for in situ wind measurements) is located at site D; the agricultural area is indicated by the white arrows and the river mouths with yellow arrows.

M AN U

Figure 2 - (a) Mean daily water level (m) of the Candelaro river in 2013. (b) Mean daily flow (m-3 s-1) of the Ofanto river in 2013. Ellipses and rectangles highlight periods of lowest and highest discharge respectively. The Po river flow is also reported in the background in (a) and (b) for comparing the flows of the major regional rivers. (c) Distribution of wind intensity and direction

TE D

frequency: “radar” plot from April to December 2013 measured at the Meda Gargano buoy. (d) ASCAT winds in the Adriatic and Ionian seas on February 22nd and April 23rd, 2013. Figure 3 - ASCAT wind (left) and model current at 1.47 m (center) and at 15 m (right) in the

EP

Manfredonia-Gargano area, illustrating the main current dynamics patterns observed in the entire 2013 wind and current dataset. (a) February 18-19, (b) February 21-22 (in situ sampling occurred

AC C

on February 22), (c) February 23-24, (d) April 20-21 (sampling occurred on April 22). Figure 4 - Vertical distribution of T and S at the northern site M (profiles on the left) and along the hydrographic transect in the Gulf of Manfredonia (right), for each sampling month. Figure 5 - Time series of surface (S) and bottom (B) values of DIN (a), SRSi (b), SRP (c) and chlorophyll a (d) at inshore (st. A) and offshore (st. E) sites in the gulf (left and centre, respectively) and site north of Gargano (st. M; right), for the six 2013 sampling months; e) DIN (left) and SRSi (right) vs. salinity plots (all data). Gray circles: surface samples; white circles: bottom samples.

29

ACCEPTED MANUSCRIPT Figure 6 - Time series of surface (S) and bottom (B) values of absorption coefficients α280 (a) and α355 (b) at inshore and offshore sites A and E in the gulf (left and centre) and site M north of Gargano (right), for the six 2013 sampling months; α280 (left) and α355 (right) vs. salinity plots (c) for December. Gray circles: surface samples; white circles: bottom samples.

RI PT

Figure 7 - SST and surface chlorophyll maps of the Adriatic Sea, closest to sampling days: (a) and (b) February 20, 2 days before sampling; (c) and (d) December 4 (SST) and 11 (Chl), 8 days and 1day before sampling; (e) and (f) April 24 and 19, 1 day after and four days before sampling; (g)

SC

and (h) June 18. Insets in chlorophyll maps: study area detail, same chlorophyll range as main image. Please note different temperature ranges in the maps, so as to optimize thermal feature

M AN U

rendering.

Figure 8 -. (a) Chlorophyll, (b) ADG and (c) BBP on the August29th2013 sampling day. d)

AC C

EP

TE D

Chlorophyll, e) ADG and f) BBP on the October 23rd 2013 sampling day.

30

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT

EP

TE D

M AN U

SC

RI PT

The Western Adriatic Current is a nutrient source north of the Gargano peninsula. Hydrography of the Gulf is controlled by river discharge, winds and currents Small river discharge is an important nutrient and CDOM source in the Gulf of Manfredonia in winter. High riverine discharge events confine the WAC influence to the gulf’s offshore waters. Biological production and degradation are a source of CDOM in spring –summer.

AC C

• • • • •