Caballero-Alfonso et al. 2015

Biogeochemical and environmental drivers of coastal hypoxia
Angela M. Caballero-Alfonso a,1
, Jacob Carstensen b
, Daniel J. Conley a
a
Department of Geology, Lund University, Sölvegatan 12, SE-223 62 Lund, Sweden
b
Department of Bioscience, Aarhus University, Frederiksborgvej 399, DK-4000 Roskilde, Denmark
a b s t r a c ta r t i c l e i n f o
Article history:
Received 9 January 2014
Received in revised form 31 March 2014
Accepted 11 April 2014
Available online 21 April 2014
Keywords:
Hypoxia
Coastal zone
Baltic Sea
Nutrients
Recent reports have demonstrated that hypoxia is widespread in the coastal zone of the Baltic Sea. Here we
evaluate the long-term trends of dissolved oxygen in bottom waters and of the drivers of coastal hypoxia. Eleven
of the 33 sites evaluated had increasing trends of bottom water dissolved oxygen, but only the Stockholm
Archipelago presents a consistent positive increasing trend in time. The vast majority of sites continue to
worsen, especially along the Danish and Finnish coasts, in spite of remediation efforts to reduce nutrients. Surface
temperatures were relatively comparable across the entire coastal Baltic Sea, whereas bottom water tempera-
tures varied more strongly among sites, most likely due to differences in mixing (or stratification) and water
exchange with the open Baltic Sea. Nutrient concentrations varied by factors 2–3 with highest levels at sites
with restricted water exchange and higher land based nutrient loading. None of the sites were permanently
stratified during the summer seasonal window although most of the sites were stratified more than half of the
time. The frequency of hypoxia was also quite variable with sites in Gulf of Bothnia almost never experiencing
hypoxia to enclosed sites with more than 50% chance of hypoxia. There are many factors governing hypoxia
and the complexity of interacting processes in the coastal zone makes it difficult to identify specific causes.
Our results demonstrate that managing nutrients can create positive feedbacks for oxygen recovery to occur.
In the absence of nutrient reductions, the recovery from hypoxia in coastal marine ecosystems is unlikely.
© 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
1. Introduction
Hypoxia has intermittently occurred in the Baltic Sea since its forma-
tion 8000 years ago with its occurrence constrained by hydrological
processes, climate and anthropogenic activities (Sohlenius et al., 2001;
Zillén et al., 2008). Since the 1950s the spatial distribution and intensity
of hypoxia in the open waters have increased due to anthropogenic
nutrient enrichment (Carstensen et al., 2014; Conley et al., 2009a,
2009b). Nutrient loads have increased by approximately 2.5 times for
nitrogen (N) and 3.7 for phosphorus (P) since the turn of the last century
(Gustafsson et al., 2012). Nutrients stimulate phytoplankton production
and the subsequent decomposition of organic matter creates a demand
for oxygen. The Baltic Sea is comprised of several connected basins with
permanent salinity stratification and restricted ventilation of the bottom
waters making it susceptible to hypoxia. Hypoxia reduces the abundance
of benthic fauna and alters sediment nutrient dynamics (Conley et al.,
2007; Karlson et al., 2007; Perus and Bonsdorff, 2004). Hypoxia also re-
stricts the reproduction of the bottom dwelling fish species, especially
cod, by reducing spawning area (Limburg et al., 2011; Vallin et al., 1999).
Much less is known regarding hypoxia in the coastal zone, where the
physical processes are generally more dynamic and complex. Globally, the
ratesofdecreaseinbottomwateroxygenconcentrationsarehigherin coast-
al areas than in the open oceans (Gilbert et al., 2010). Hypoxia has been ob-
served in the coastal zone of the Baltic Sea with 115 sites out of 613 sites,
assessed from monitoring data, having experienced hypoxia between
1955 and 2009 (Conley et al., 2011). In addition to the existing coastal ma-
rine sites worldwide that have reported hypoxia (Diaz and Rosenberg,
2008), these 115 sites make the Baltic Sea the largest region in the world
that suffers from hypoxia with over the 20% of the known areas worldwide
(Conley et al., 2011). In addition, there is an increasing trend in the severity
of hypoxia since the 1950s in many coastal sites of the Baltic Sea.
There are many factors governing hypoxia and the complexity of
interacting processes in the coastal zone makes it difficult to identify
specific causes. Anthropogenic drivers are primarily nutrient inputs,
one of the common features of coastal eutrophication. Eutrophication
is defined as the organic enrichment of a system (Nixon, 1995), which
has serious consequences for the marine environment including de-
clines in water transparency, increasing algal blooms, loss of benthic
vegetation and hypoxia among others (Nixon, 2009). There are also
natural drivers of hypoxia: 1) hydrology (reduced horizontal advection
enhancing bottom-water residence time, stratification reducing vertical
mixing), and 2) climate (temperature reducing the solubility of oxygen
and increasing respiration as well as changes in wind patterns, rainfall
and runoff affecting the hydrological processes). In the open ocean cli-
mate models predict declines in dissolved oxygen produced by global
Journal of Marine Systems 141 (2015) 190–199
E-mail address: angela.caballero-alfonso@geol.lu.se (A.M. Caballero-Alfonso).
1
Tel.: +46 462220449.
http://dx.doi.org/10.1016/j.jmarsys.2014.04.008
0924-7963/© 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Contents lists available at ScienceDirect
Journal of Marine Systems
journal homepage: www.elsevier.com/locate/jmarsys

warming due to a contribution of thermal, dynamical, and biogeochem-
ical factors (Stramma et al., 2008). However, further studies are needed
to understand and elucidate the underlying mechanisms for hypoxia in
coastal areas (Conley et al., 2009b; Rabalais et al., 2010).
We focused our investigation on evaluating the long-term trends of
bottom water oxygen concentrations and of the drivers of coastal hyp-
oxia in the Baltic Sea. Our goal was to examine relationships between
nutrients, physical factors and climate as important drivers for coastal
hypoxia in the Baltic Sea.
2. Material and methods
2.1. Study area and site descriptions
The Baltic Sea is a nearly landlocked sea located between 54 and
66°N having restricted water exchange with the oceans. It consists of
the main deeper basin (Baltic Proper) with three major gulfs in the
northeast (Gulf of Riga, Gulf of Finland and Gulf of Bothnia) and con-
nects to the North Sea through the Danish Straits. Most of the Baltic
Sea is brackish with salinities less than 10, but the entrance area (the
Danish Straits) is characterized by strong salinity gradients. The climate
is temperate with an annual rainfall of approximately 600 mm per year
(Omstedt et al., 1997). About 85 million people inhabit the Baltic Sea
catchment, which is dominated by agriculture and urbanization in the
southwest and forests towards the northeast. Nutrient loads from the
different regions are a reflection of the land use (Wulff et al., 1990).
The coastal zone surrounding the Baltic Sea is quite diverse, owing to
its geologic history. Shallow, sandy estuaries and coastal embayments
with complex geomorphology dominate in the southwest, changing to
an open coastline with lagoons along the southeastern coastline. Most
of the Swedish and Finnish coastline is archipelagic. Large rivers drain
an area approximately four times the Baltic Sea itself and discharge to
the Baltic Proper, Gulf of Riga, Gulf of Finland and Gulf of Bothnia. Rivers
draining into the Danish Straits are smaller.
A total of 33 sites were studied and grouped in 9 regions following
HELCOM designations: Kattegat (4 sites), Øresund (1 site), Belt Sea
(6 sites), West Gotland Basin (4 sites), Stockholm Archipelago (5 sites),
Bothnian Sea (1 site), Bothnian Bay (2 sites), Finnish Archipelago (5
sites) and Gulf of Finland (5 sites); see Table 1 for more details. These
sites were selected based on the results in Conley et al. (2011) and the
length of time series available in the Baltic Environmental Database
(BED). We chose sites that had experienced coastal hypoxia and were
representative of a region, and which had more than 10 years of data.
Some data-rich areas, e.g. the Stockholm Archipelago, had numerous
sites with long-term time series and in such areas the most representa-
tive sites were selected.
2.2. Data analyses
Seasonal windows for hypoxia-prone months were selected per
region to give a better description of low oxygen conditions (Conley
et al., 2011). For the Bothnian Bay, Stockholm and Finnish archipelagos,
the West Gotland Basin, the Kattegat and the Öresund the summer
months considered were August through October. For the Bothnian
Sea we considered only August and September; and for the Gulf of
Finland July through October. These data were organized by region
and sites with the aim to establish annual time series for each site.
In coastal areas periods of stratification and mixing can alternate
over short time scales, unlike the hydrographic conditions in the deeper
Baltic Sea basins (Conley et al., 2011). Stratification is a precondition for
hypoxia to occur and therefore only profiles characterized by stratified
conditions were selected to evaluate oxygen trends. We defined the
water column to be stratified if the density difference between surface
and bottom waters was larger than 0.5 g l−1
(see Conley et al., 2007).
The rate of oxygen depletion in the bottom water depends on the vol-
ume below the pycnocline and to characterize this volume we
calculated bottom water thickness as the profile maximum depth
minus the depth location of the pycnocline. This characteristic was cal-
culated on the more recent data, due to low resolution of older profiles,
and assuming that bottom water thickness remained the same through-
out the summer period.
Oxygen saturation (%) and Apparent Oxygen Utilization (AOU) were
calculated for the bottom waters. AOU is defined as the observed oxy-
gen concentration subtracted from the oxygen concentration at satura-
tion state calculated from temperature and salinity of the bottom water
(Hetland and Di Marco, 2008). AOU allows temperature and salinity
effects on oxygen solubility to be removed from the observations and
quantifies the loss of oxygen due to biological and chemical activities
(Bianchi et al., 2010), most importantly respiratory demands from
decomposition of organic material. The bottom oxygen trend was tested
against the following possible hypoxia triggers: (a) physical status:
temperature and stratification; (b) biological status: surface chlorophyll
a (chla); and (c) chemical status: total phosphorus (TP) and total nitro-
gen (TN). Additionally, to assess the putative effect of low oxygen con-
centrations on benthic communities the frequency of hypoxia was
assessed. Although the thresholds for faunal effects to occur vary widely
in the literature from 0.28 to 4 mg O2 l−1
(Steckbauer et al., 2011;
Vaquer-Sunyer and Duarte, 2010), we define hypoxia as concentrations
of dissolved oxygen equal to or below 2 mg O2 l−1
(Rabalais et al.,
2010). Hence, the frequency of hypoxia was calculated as the propor-
tion of profiles with bottom water oxygen concentrations below 2 mg
O2 l−1
.
Trends of seasonal means were estimated using Generalized Linear
Models (GLMs), separately for each site and variables of interest, taking
into account temporal variations (months and years as additive categor-
ical factors). After examining the distribution of different variables,
oxygen concentrations, AOU, and temperature were assumed to follow
a normal distribution; chla, TN and TP were assumed lognormal distrib-
uted; and frequencies of hypoxia and stratification were binomial. Using
the GLM approach accounted for changes in sampling across years by
weighting each profile according to the month it was taken (Carstensen
et al., 2006). Trends in oxygen concentrations were related to trends in
temperature, frequency of stratification, surface TN, TP and chla concen-
trations using Spearman correlations. Spearman correlation was chosen
since it is less sensitive to single years with strongly deviating annual
means that may otherwise influence standard correlations. All the data
processing and statistical analyses were programmed using SAS®.
3. Results
The studied sites varied from 10 to 117 m depth with bottom water
below the pycnocline ranging from 1.1 to 22.2 m (Table 1). The thick-
ness of the bottom water typically increased with water depth. Surface
temperatures were relatively comparable across the entire coastal Baltic
Sea, whereas bottom water temperatures varied more strongly among
sites, most likely due to differences in mixing (or stratification) and
water exchange with the open Baltic Sea. Nutrient concentrations var-
ied by factors 2–3 with the highest levels at sites with restricted water
exchange and higher connectivity to land. None of the sites were per-
manently stratified during the summer seasonal window although
most of the sites were stratified more than half of the time. The frequen-
cy of hypoxia was also quite variable with sites in Gulf of Bothnia almost
never experiencing hypoxia to enclosed sites with more than 50%
chance of hypoxia.
In the following trends in oxygen conditions and associated vari-
ables are presented, but the trends are only visualized for a few selected
sites for plotting, as we could not show all. However, directions of trends
are visualized geographically for all sites (Fig. 1). The sites presented
(Figs. 2–5) were those having the highest significant correlation for
bottom oxygen concentration versus at least one of the tested influenc-
ing variables (Table 2).
191A.M. Caballero-Alfonso et al. / Journal of Marine Systems 141 (2015) 190–199

Table 1
Characteristics of the 33 coastal study sites in the Baltic Sea. Site names are from the national monitoring programs. Number of profiles and descriptive statistics characterizing these were calculated for the seasonal window as described in the Material
and methods section. (Abbreviations: Max. depth — maximum depth (m); bottom layer thickness (m); temperature (°C); TN — total nitrogen (μM); TP — total phosphorus (μM); Surf — surface water samples; Bot — bottom water samples; AOU —
Apparent Oxygen Utilization; Freq. — frequency).
Region Site Max. depth # of profiles # of bottom
oxygen samples
Bottom layer
thikness
Temperature TN TP Bottom AOU Freq. of
stratification (%)
Freq. of
hypoxia (%)
Surf Bot Surf Bot Surf Bot
Gulf of Bothnia U-5 Uudenkaarlepyyn 23 551 550 10.4 11.6 10.2 33.4 31.4 0.7 0.6 1.1 16.5 0.2
U-6 Uudenkaarlepyyn 15 103 99 7.6 14 12.3 40.1 27.8 0.9 0.6 1.1 40.8 4.9
Pome 115 Preiviikinlahti 26 277 274 9.8 14.7 11 20.6 18.1 0.4 0.4 2.3 36.8 4.0
Finnish Archipelago Hala N 22 114 114 1.3 14.4 9.3 38 50.8 1.1 1.6 – 32.5 48.3
Kamsholmsfjärden 19 178 176 2.5 17.7 8.4 27.8 50.9 0.9 7.6 10.5 89.9 80.3
Pala 47 223 223 18.5 17.6 12.7 30.2 31.7 1 1.3 4.5 52.9 3.1
Turm 290 Kuparivuori 25 83 82 1.2 15.7 9.5 30.5 39.7 1 1.4 5.6 54.2 22.9
Uki 245 Vähä-Seikomaa 26 1007 1002 10.5 13.7 12.5 28.7 29 1.2 1.6 3 17.8 4.2
Gulf of Finland Orrenkylänselkä 34 356 354 10.1 16.3 10.6 40.2 35.4 1.9 3 4.9 61.0 4.8
Skatafjärden 28 90 90 3.2 16.4 8.9 25.6 27.9 0.9 2 4.5 88.9 2.2
Suomenl Varissaari 17 193 193 2.1 15.9 10.9 34.5 30.9 1.1 1.6 3.5 65.8 2.6
Svartbäckinselkä 42 267 265 9.7 15.7 7.4 29.3 29.4 1.2 1.9 4.7 75.3 1.5
UUS-18 Sandöfjärden 31 301 301 8 15.9 8.7 26.7 34.5 1.1 4.4 7.2 71.1 38.5
Stockholm Archipelago Blomskär 27 153 136 2.9 13.8 5.2 49.5 52.8 2 5.4 11 90.2 72.6
Kanholmsfjärden 117 164 135 17.4 13.7 5.4 20.4 34.6 0.6 4.8 10.3 79.9 59.2
Karantänbojen 22 157 153 4.1 13.4 6.1 56.2 57.4 2.2 4.6 9.6 91.7 59.9
Solöfjärden 47 357 312 14 13.4 7.5 39.4 32.3 1.4 2.4 6.5 84.3 16.8
Vaxholmsfjärden 29 240 226 5.3 14 8.9 40.6 50 1.5 3.9 8.3 72.1 61.3
West of Gotland Basin Blankaholm 24 28 28 2.4 15.3 5.9 29.8 33.7 0.9 5.4 9.4 75.0 46.4
Bråvik 33 70 69 11.1 16.7 11.1 36 29.9 1.1 1.9 5.1 77.1 4.3
Västervik 64 90 89 22.2 15.9 4.8 30.5 50.6 0.9 5.4 9.3 83.3 45.6
Västrum 17 28 28 1.4 15.3 7.5 32 45.6 1.1 9 9.4 78.6 71.4
Belt Sea Aabenraa Fjord 35 235 235 5.1 16.2 11 27.8 34.8 1.4 4.5 7.3 98.3 69.4
Det Sydfynske Øhav 40 821 808 15 15.5 13.6 22.1 27.4 0.9 2.3 3.5 64.2 28.6
Eckernförder Bucht 31 80 72 5.2 15.5 11.6 21.3 26.9 1.2 4.1 7.8 82.5 73.8
Flensborg Fjord 31 345 343 7.7 16.2 11.1 27.4 40.6 1.6 5.4 7.6 91.6 69.6
Horsens Fjord 23 497 489 11.8 15.2 14.5 28.4 28.3 1.8 2 2.4 63.2 6.4
Århus Bugt 18 333 333 1.7 15.1 13 15.9 17.9 0.8 1.5 4.6 95.8 13.2
Kattegat and Öresund Hevring Bugt 10 151 151 1.1 14.6 14 16.4 17.5 0.8 1.2 2.2 78.1 2.7
Laholmsbugten 28 127 123 11.3 16.4 13 20.5 20.3 0.5 0.9 3.7 82.7 16.5
Middelgrund 19 198 198 2.2 14.9 12.5 19.5 21 0.6 1.1 4.2 95.0 7.1
Skälderviken 29 98 93 14.2 16.5 13.9 25 21.9 0.6 1.1 2.8 79.6 10.2
Valö 24 74 72 2.6 16.1 13.5 18.1 19.7 0.6 1.1 3.1 90.5 4.1
192A.M.Caballero-Alfonsoetal./JournalofMarineSystems141(2015)190–199

3.1. Gulf of Bothnia region
Three sites were investigated in this region (U-5 Uudenkaarlepyyn,
U-6 Uudenkaarlepyyn and Pome 115 Preiviikinlahti); two of these are
geographically close but have different nutrient levels and frequency
of stratification (Table 1). Pome 115 Preiviikinlahti is a relatively deep
site (26 m) with temperature stratification of ~4 °C separating a rela-
tively large volume of bottom water. Compared to the other sites
frequencies of stratification and hypoxia were both relatively low. The
three sites had no clear distinctive trends in oxygen conditions, temper-
ature, nutrient levels or stratification patterns, but there were strong
interannual variations in temperature and frequency of stratification
Fig. 1. Maps summarizing the trends for the bottom oxygen concentrations, AOU, temperature, frequency of stratification, TN and TP. The regions are grouped as: 1) Gulf of Bothnia,
2) Finnish Archipelago and 3) Gulf of Finland, 4) Stockholm Archipelago, 5) West Gotland Basin, 6) Belt Sea and 7) Kattegat and Öresund. Increasing and decreasing trends are shown with
upward and downward facing triangles, green or red colors indicating improving or worsening conditions, and open and filled triangles signify non-significant and significant trends,
respectively.

(Fig. 2). However, Pome 115 Preiviikinlahti experienced a period with
low oxygen concentrations in the 1980s, but these low concentrations
could not be linked to any of the potential explanatory variables
(Table 2). Temperature was important in explaining trends in oxygen
conditions at the two Uudenkaarlepyyn sites. For U-6 Uudenkaarlypyyn
changing stratification pattern was also important (Table 2), whereas
this factor was not significant at U-5 Uudenkaarlepyyn that generally
had a higher mixing frequency (Table 1).
3.2. Finnish Archipelago region
This region includes five sites (Hala N, Kamsholmsfjärden, Pala,
Turm 290 Kuparivuori and Uki 245 Vähä-Seikomaa) located along the
inner coastal archipelago with relatively high nutrient levels compared
to the other sites (Table 1). Overall, the frequency of stratification aver-
aged 44%, but there were large differences in stratification between
sites. Temperature differences between surface and bottom were large
at all sites except Uki 245 Vähä-Seikomaa, and bottom water tempera-
tures have been increasing for the last two decades (Fig. 3B). Frequency
of hypoxia was even more variable ranging from 3.1% to 80.3% with an
average of 31.8%. Three sites had a small bottom water volume and
these three sites were also the sites with the highest frequency of hypoxia.
There were no distinctive trends in the variables (most were not signif-
icant, Fig. 1), although there were large interannual variations (Fig. 3).
Significant correlations were obtained between the bottom oxygen con-
centration and the bottom temperature (Table 2).
3.3. Gulf of Finland region
Five sites were investigated in this region (Orrenkylänselkä,
Skatafjärden, Suomenl Varissaari, Svartbäckinselkä and UUS-18
Sandöfjärden) and all of them were stratified by temperature through-
out most of the summer period (Table 1). Stratification patterns varied
substantially between years. Temperature differences were large sepa-
rating a cold bottom water mass (7–11 °C) of several meters from a
warmer surface layer (~16 °C). The last two decades have generally
been warmer. Nutrient levels were relatively high and did not change
over the years (Fig. 3C). Despite strong stratification the frequency of
hypoxia was low, except for UUS-18 Sandöfjärden (Table 1). This site
is a semi-enclosed embayment where bottom water ventilation could
be low. Oxygen concentrations correlated significantly with tempera-
ture and/or nutrient concentrations at the five sites, displaying worsen-
ing oxygen conditions with increasing temperature and nutrients
(Table 2).
0
1
2
3
4
5
6
70
2
4
6
8
10
12
AOU(mg/l)
BottomOxygen(mg/l)Bottomtemperature(°C)
U-5 Uudenkaarlepyyn
Bottom oxygen
Pome 115 Preiviikinlahti
Bottom oxygen
U-5 Uudenkaarlepyyn
AOU
AOU
A
0
10
20
30
40
50
60
70
80
90
100
0
2
4
6
8
10
12
14
16
18
20
Frequencyofstratification(%)
U-5 Uudenkaarlepyyn
Bottom temperature
Bottom temperature
U-5 Uudenkaarlepyyn
Freq. of stratification
Freq. of stratification
B
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0
10
20
30
40
50
60
1960
1963
1966
1969
1972
1975
1978
1981
1984
1987
1990
1993
1996
1999
2002
2005
2008
2011
TP(µM)
TN(µM)
Year
U-5 Uudenkaarlepyyn TN
Pome 115 Preiviikinlahti TN
U-5 Uudenkaarlepyyn TP
Pome 115 Preiviikinlahti TP
C
Fig. 2. Long-term trends in the Gulf of Bothnia exemplified by two sites. A: dissolved
bottom oxygen concentrations and Apparent Oxygen Utilization; B: bottom temperature
and frequency of stratifications; C: surface TN and TP. Yearly seasonal means were obtained
from the GLM approach.
BottomOxygen(mg/l)Bottomtemperature(°C)
TP(µM)
TN(µM)
0
2
4
6
8
10
120
1
2
3
4
5
6
7
8
9
AOU(mg/l)
Kamsholmsfjärden Bottom oxygen
Orrenkylänselkä Bottom oxygen
Kamsholmsfjärden AOU
Orrenkylänselkä AOU
A
0
10
20
30
40
50
60
70
80
90
100
0
2
4
6
8
10
12
14
16
Kamsholmsfjärden Bottom temperature
Orrenkylänselkä Bottom temperature
Kamsholmsfjärden Frequency of stratification
Orrenkylänselkä Frequency of stratification
B
0
1
2
3
4
5
6
7
0
10
20
30
40
50
60
70
80
1960
1963
1966
1969
1972
1975
1978
1981
1984
1987
1990
1993
1996
1999
2002
2005
2008
2011
Year
Kamsholmsfjärden TN
Orrenkynlänselkä TN
Kamsholmsfjärden TP
Orrenkynlänselkä TP
C
Fig. 3. Long-term trends in the Finnish Archipelago and the Gulf of Finland exemplified by
two sites. Details as for Fig. 2.
194 A.M. Caballero-Alfonso et al. / Journal of Marine Systems 141 (2015) 190–199

3.4. Stockholm Archipelago region
In this region five sites were analyzed (Blomskär, Kanholmsfjärden,
Karantänbojen, Sölofjärden and Vaxholmsfjärden). All sites were
relatively deep with Kanholmsfjärden as the deepest investigated of
all sites (Table 1). The bottom water volume below the pycnocline
was relatively large; however, all sites were hypoxic for most of the
summer period. Nutrient levels in the Stockholm Archipelago were
the highest of all sites. Stratification was strong and driven by both
salinity and temperature differences. Oxygen concentrations have in-
creased and nutrient concentrations have decreased over time (Figs. 1,
4A,C). The bottom water has generally warmed and the frequency of
stratification has decreased in Vaxholmsfjärden (Fig. 4B), but this pat-
tern is not consistent at all sites. Bottom oxygen concentrations were
significantly correlated with nutrients at Karantänbojen, Sölofjärden
and Vaxholmsfjärden, as expected, whereas oxygen conditions
were related to frequency of stratification in Kanholmsfjärden and
positively related to temperature in Vaxholmsfjärden. Whereas the
positive correlation to temperature in Vaxholmsfjärden could be an
artifact from nutrients showing similar trends, the positive correla-
tion to frequency of stratification in Kanholmsfjärden is somewhat
more surprising. Blomskär, which is an isolated branch in the
Stockholm Archipelago, had a significant decreasing trend of bottom
oxygen concentration with time, as opposed to all the other sites
displaying increasing trend (Fig. 1). The oxygen trend at Blomskär
was not significantly correlated to any of the explanatory variables
(Table 2).
3.5. West Gotland Basin region
Four sites were investigated (Blankaholm, Bråvik, Västervik and
Västrum) with depths between 17 and 64 m; the two deepest sites
also having a high bottom water thickness (Table 1). Nutrient levels
were moderate compared to the other sites, whereas the frequency of
stratification was high. The frequency of hypoxia was high at all sites,
except Bråvik which is dominated by the large Motala River. Freshwater
inputs to the other sites are relatively smaller. The monitoring data set
from this region is relatively weak compared to other regions with
fewer years of data. Consequently, it is more difficult to assess the trends
and correlate oxygen conditions with other factors. Nevertheless, oxy-
gen conditions have improved at Blankaholm and Västervik (Figs. 1,
4C) and these trends correlated significantly with TP levels and temper-
ature (Table 2).
BottomOxygen(mg/l)Bottomtemperature(°C)TN(µM) 0
2
4
6
8
10
120
1
2
3
4
5
6
7
8
9
AOU(mg/l)
Vaxholmsfjärden Bottom oxygen
Västervik Bottom oxygen
Vaxholmsfjärden AOU
Västervik AOU
A
0
10
20
30
40
50
60
70
80
90
100
0
2
4
6
8
10
12
14
16
Vaxholmsfjärden Bottom temperature
Västervik Bottom temperature
Vaxholmsfjärden Frequency of stratification
Västervik Frequency of stratification
B
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
0
10
20
30
40
50
60
70
1960
1963
1966
1969
1972
1975
1978
1981
1984
1987
1990
1993
1996
1999
2002
2005
2008
2011
TP(µM)
Year
Vaxholmsfjärden TN Västervik TN
Vaxholmsfjärden TP Västervik TP
C
Fig. 4. Long-term trends in the Stockholm Archipelago and West Gotland Basin exemplified
by two sites. Details as for Fig. 2.
BottomOxygen(mg/l)Bottomtemperature(°C)TN(µM)
0
1
2
3
4
5
6
7
8
9
100
1
2
3
4
5
6
7
8
9
10
AOU(mg/l)
Aabenraa Fjord Bottom oxygen
Horsens Fjord Bottom oxygen
Aabenraa Fjord AOU
Horsens Fjord AOU
A
0
10
20
30
40
50
60
70
80
90
100
0
2
4
6
8
10
12
14
16
18
Aabenraa Fjord Bottom temperature
Horsens Fjord Bottom temperature
Aabenraa Fjord Frequency of stratification
Horsens Fjord Frequency of stratification
B
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
0
10
20
30
40
50
60
70
80
90
100
1960
1963
1966
1969
1972
1975
1978
1981
1984
1987
1990
1993
1996
1999
2002
2005
2008
2011
TP(µM)
Year
Aabenraa Fjord TN
Horsens Fjord TN
Aabenraa Fjord TP
Horsens Fjord TP
C
Fig. 5. Long-term trends in the Danish strait exemplified by two sites. Details as for Fig. 2.

3.6. Belt Sea region
The Belt Sea region includes 6 sites (Aabenraa Fjord, Det Sydfynske
Øhav, Eckernförder Bucht, Flensborg Fjord, Horsens Fjord and Århus
Bugt) with depths ranging from 18 to 40 m (Table 1). It is a region
(Table 1) with a relatively high frequency of stratification (83%), due
to the location of a halocline around 15–25 m. Nutrient levels are
relatively low, compared to sites in the more brackish part of the Baltic
Sea, but there are differences among sites in the Belt Sea region associ-
ated with proximity to land-based nutrient sources. Three of the sites in
the Belt Sea region (Aabenraa Fjord, Eckernförder Bucht and Flensborg
Fjord) have high frequencies of hypoxia (~70%) and these sites are
connected to the Little Belt where bottom water ventilation is lowest.
The three other sites display more episodic events of hypoxia. Despite
decreases in both nitrogen and phosphorus concentrations over the
last many decades (Figs. 1, 5C) oxygen conditions have not generally
improved, although there is a tendency for improvement at some sites
in the most recent years (Fig. 5A). As a consequence, a general pattern
of declining nutrients associated with improving oxygen conditions
was not found, and only Horsens Fjord and Århus Bugt showed signifi-
cant correlations (Table 2). The lack of significant correlations in this
region could be due to the lack of variability in oxygen concentrations
over time making the identification of such relationships difficult.
3.7. Kattegat and Öresund region
Five sites are included in this region (Hevring Bugt, Laholmsbugten,
Middelgrund, Skälderviken and Valö); the three deepest of these
located on the Swedish coast. The sites are generally shallower than
those in other regions and the thickness of the bottom layer depends
on the permanent halocline in the Kattegat, which is typically found
around 15–20 m depth (Andersson and Rydberg, 1988). However,
windy conditions mix down the surface layer and may tilt the halocline.
The sites therefore have a high frequency of stratification (Table 1). The
sites in this region have the lowest nutrient levels of all Baltic Sea coastal
sites, because they are more influenced by water exchanges with the
North Sea where both TN and TP levels are lower than in the Baltic.
The frequency of hypoxia is generally low and these events have an
episodic nature (Conley et al., 2011). In this region oxygen concentra-
tions have displayed non-uniform trends (Fig. 1), but major changes
have not been observed despite decreasing nitrogen levels (Fig. 5). Due
to the episodic nature of low oxygen conditions and the lack of large
changes in oxygen conditions no significant correlations were found
between oxygen concentration and explanatory variables (Table 2).
3.8. Comparison of sites across regions
The 33 coastal sites were exposed to large differences in physical
forcing and behaved differently to nutrient enrichment. We examined
the effect of stratification patterns on hypoxia by plotting AOU versus
the frequency of stratification and thickness of bottom layer (Fig. 6).
Oxygen consumption was clearly related to frequency of stratification
with hypoxia rarely occurring at sites that were stratified less than
50% of the seasonal window (Table 1). There were regional differences
in AOU versus frequency of stratification. Sites in Kattegat, Öresund,
Belt Sea and Gulf of Finland generally had lower AOU compared to
sites in archipelagic regions (Finnish Archipelago, Stockholm Archipelago
and West Gotland Basin) with the same frequency of stratification
(Fig. 6A). The relationship between AOU and bottom water thickness
was more complex, although at some sites it appears as if oxygen
consumption could be higher if the bottom water volume below the
pycnocline was relatively low (Fig. 6B).
Table 2
Spearman correlations between trends of oxygen concentrations and physical, chemical and biological conditions believed to influence oxygen conditions. The significance of the
correlations is indicated as *: p b 0.05; **: p b 0.01; ***: p b 0.001. Trends used for the correlations are shown for U-5 Uudenkaarlepyyn and Pome 115 Preiviikinlahti (Fig. 1),
Kamsholmsfjärden and Orrenkylänselkä (Fig. 2), Vaxholmsfjärden and Västervik (Fig. 3), and Aabenraa Fjord and Horsens Fjord (Fig. 4). The symbol ‘–’ indicates that there were no or
too few annual values to calculate and test the correlation.
Region Site Surface chla Bottom temperature Frequency of stratification Surface TN Surface TP
Gulf of Bothnia U-5 Uudenkaarlepyyn 0.5 −0.83*** −0.02 0.24 −0.17
U-6 Uudenkaarlepyyn −0.5 −0.55*** −0.7*** 0.06 0.22
Pome 115 Preiviikinlahti – −0.1 0.14 0.36 −0.09
Finnish Archipelago Hala N – −0.47*** 0.24 −0.01 −0.32
Kamsholmsfjärden – −0.1 −0.19 −0.14 0.03
Pala −0.38 −0.62*** 0.1 0.09 0.17
Turm 290 Kuparivuori – −0.54*** 0.03 −0.25 −0.03
Uki 245 Vähä-Seikomaa – −0.41** 0.22 0.12 0.09
Gulf of Finland Orrenkylänselkä −0.5 −0.07 0.00 −0.21 −0.48***
Skatafjärden – −0.73*** 0.03 −0.38** −0.33*
Suomenl Varissaari – −0.69*** −0.05 −0.16 −0.1
Svartbäckinselkä – −0.29 0.15 −0.15 −0.34**
UUS-18 Sandöfjärden – −0.05 0.37 −0.61*** −0.68***
Stockholm Archipelago Blomskär −0.4 −0.25 0.18 −0.11 0.07
Kanholmsfjärden −0.2 0.00 0.34** 0.15 0.02
Karantänbojen −0.5 −0.04 0.19 −0.44*** −0.65***
Solöfjärden −0.4 −0.08 −0.02 −0.48*** −0.45***
Vaxholmsfjärden 0.3 0.5*** −0.1 −0.36** −0.47***
West of Gotland Basin Blankaholm 0.04 0.24 −0.39 0.24 −0.61**
Bråvik – −0.6** 0.13 −0.23 −0.21
Västervik −0.09 −0.08 0.02 0.12 −0.31
Västrum −0.25 0.42 0.06 0.07 −0.33
Belt Sea Aabenraa Fjord 0.11 0.31 −0.32 0.21 0.3
Det Sydfynske Øhav −0.06 −0.17 −0.13 −0.24 −0.02
Eckernförder Bucht – 0.07 −0.05 0.28 0.05
Flensborg Fjord 0.11 0.18 −0.23 0.36 0.15
Horsens Fjord −0.28 −0.23 −0.41** −0.36** −0.16
Århus Bugt −0.22 −0.04 −0.1 −0.63*** −0.46*
Kattegat and Öresund Hevring Bugt 0.12 −0.06 −0.11 −0.45 −0.21
Laholmsbugten −0.17 −0.16 −0.16 0.13 −0.01
Middelgrund −0.62 0.14 0.07 −0.33 −0.36
Skälderviken 0.05 0.03 0.24 0.11 0.05
Valö 0.002 −0.26 −0.05 0.003 0.38

4. Discussion
4.1. Overall trends
Reports of hypoxia in coastal marine systems have increased during
the last few decades with the primary cause of low oxygen conditions
linked to increases in nutrient loads (Conley et al., 2011; Diaz and
Rosenberg, 2008; Rabalais et al., 2014). There are numerous studies
showing that eutrophication is promoting the expansion, duration, and
intensity of hypoxic events (Diaz and Rosenberg, 2008; Kemp et al.,
2009; Rabalais and Gilbert, 2008). Policies and measures have been im-
plemented to reduce the input of nutrients to coastal ecosystems
(Duarte et al., in press) and in the Baltic Sea as well (HELCOM, 2007).
The coastal waters around Denmark were one of the first areas to
document significant decreases in nutrient concentrations on a large
regional scale resulting from an active management strategy to reduce
nutrients from both diffuse and point sources (Carstensen et al., 2006).
Our analysis of Baltic Sea coastal ecosystems shows that many countries
in the watershed have reduced their nutrient loads resulting in lower
nutrient concentrations in coastal waters (Fig. 1). However, our results
also show that improvements in bottom water dissolved oxygen con-
centrations are not always observed (Fig. 1). From the 33 evaluated
sites only 11 show an increasing trend in dissolved oxygen of bottom
waters with time and only the Stockholm Archipelago had consistent
trends. However, the vast majority of sites continue to worsen, especially
along the Danish and Finnish coasts (Figs. 1, 3 and 5), in spite of remedi-
ation efforts to reduce nutrients. From a larger perspective, only a small
percentage (~4%) of 400 sites globally suffering hypoxia have shown
trends of increasing bottom water oxygen concentrations (Diaz and
Rosenberg, 2008).
The recovery of coastal and estuarine ecosystems also takes place in
a context of global climate change (Duarte et al., 2009) with dissolved
oxygen in oceanic oxygen minimum zones decreasing due to changes
associated with global warming (Stramma et al., 2008). Warming
plays an important role by reducing the solubility of oxygen and in-
creasing respiration. It also alters wind patterns, rainfall and runoff
affecting hydrological processes. Temperature increases stratification
and a stronger pycnocline reduces ventilation through vertical mixing
promoting bottom water oxygen depletion (Hagy et al., 2004). Increases
in bottom water temperature were observed in the Finnish Archipelago
and along the Danish coast (Figs. 1, 3 and 5) where increases in AOU
were also evident. In addition, coastal ecosystems are more vulnerable
to nutrient loads with temperature increases (Justić et al., 1996; Kemp
et al., 2005) as seen by us in the Bothnian Bay (Table 2, Fig. 2). Warming
can induce changes in periodical biological processes such as the timing
of the spring bloom and overall productivity (Nixon, 2009).
Wind speed and wind direction affect the efficiency of mixing
between surface and bottom waters, thus changing the ventilation of
bottom waters. Relationships between temperature increases, decreases
in wind speed and increases in oxygen demand have been observed in
the Narragansett Bay and Long Island Sound (Nixon, 2009) and Danish
coastal waters (Conley et al., 2007). In addition, climate change will
also modify rainfall patterns and snow melt generally increasing water
discharges from land (freshwater, pollutants and nutrients) into Baltic
coastal ecosystems (Meier et al., 2012). Although climate change will
affect the wind field (IPCC, 2007), we were not able to evaluate this
effect on trends in bottom water dissolved oxygen.
4.2. Differences between systems
The recovery of bottom water oxygen concentrations following
nutrient reduction measures is rare with increases in hypoxia continu-
ing to be observed globally (Conley et al., 2011; Diaz and Rosenberg,
2008; Gilbert et al., 2010). Our regional evaluation of coastal ecosystems
in the Baltic Sea has shown that different regions are responding sepa-
rately to reductions in nutrients. Area-specific ecological responses to
coastal eutrophication have been previously observed in the Baltic Sea
(Carstensen et al., 2011; Rönnberg and Bonsdorff, 2004). In the concep-
tual model of the coastal eutrophication problem by Cloern (2001) it is
explicitly recognized that system-specific attributes modulate the
responses to nutrient enrichment, leading to large differences among
coastal systems in their sensitivity. Our response parameter used is
bottom water hypoxia, whereas there are a complex suite of direct
and indirect responses to coastal eutrophication including changes
in water transparency, distribution of bottom vegetation, in nutrient
cycling processes, phytoplankton community composition, and changes
in ecosystem functions. The result is a wide range of estuary-specific
responses to the effects of nutrient loading on coastal eutrophication
and hypoxia (Cloern, 2001).
The coastal zone of the Baltic has a complex topography with
numerous islands and many basins separated by shallow sills resulting
in many distinct isolated basins. The coastline of the Western Gotland
Basin, including the Stockholm Archipelago, and Gulf of Finland, includ-
ing the Finnish Archipelago, is more prone to hypoxia than other parts
of the more unrestricted shoreline (Conley et al., 2011). These systems
become thermally stratified during the summer months. Because the
complex topography reduces circulation and increases the residence
time of bottom waters, these ecosystems are consequently more sensi-
tive to enhanced nutrient inputs from land and regularly experience
hypoxia. Our results confirm these patterns.
Stratification is a precursor for hypoxia to develop, but even perma-
nently stratified coastal systems can remain normoxic provided that the
bottom waters are sufficiently exchanged through bottom water venti-
lation. Hypoxia may also develop in coastal systems that stratify for
short periods only, provided that the bottom layer is so thin that respi-
ratory processes, mainly from the sediments, consume all oxygen
during such short-lived events. Our results show that it is difficult to
encapsulate the complex and highly dynamical physical processes
governing oxygen supply by means of monitoring data. Such processes
can be described more simply in open water systems, where changes in
oxygen conditions are much slower (Conley et al., 2007; Carstensen
et al., 2014).
0
2
4
6
8
10
12
0 20 40 60 80 100
AOU(mg/l)
Gulf of Bothnia
Finnish Archipelago
Gulf of Finland
Stockholm Archipelago
West Gotland Basin
Belt Sea
Kattegat & Öresund
A
0
2
4
6
8
10
12
0 5 10 15 20 25
AOU(mg/l)
Bottom layer thickness (m)
B
Frequency of stratification (%)
Fig. 6. Relationships between the AOU and the frequency of stratification (A) and thick-
ness of the bottom layer (B).

Oxygen depletion at a site might be imported by the advective trans-
port of saltier, low oxygen bottom water from an adjacent body; this
could be the case of what has been observed in the present study
at Horsens Fjord (Danish Straits) and Kanholmsfjärden (Stockholm
Archipelago). These intrusions are characterized by nutrient-rich low-
oxygen waters coming from the Belt Sea (Danish Straits) or from the
bottom waters of the Baltic Proper.
4.3. Potential regime shifts
Bottom water hypoxia is not only a particularly severe disturbance
on ecosystems because it causes death of biota, changes in the habitat
of living resources, but also the biogeochemical processes that control
nutrient concentrations in the water column (Conley et al., 2009a). Nu-
merous studies in other coastal marine ecosystems have demonstrated
that repeated hypoxic events can help to sustain future hypoxic condi-
tions (Hagy et al., 2004; Kemp et al., 2005, 2009; Turner et al., 2008).
These changes suggest that a regime shift occurs in coastal marine eco-
systems affected by large-scale hypoxia (Conley et al., 2007). Regime
shifts are often rapid transitions that occur in an ecosystem state with
the recovery of ecosystems delayed by internal feed-back mechanisms
(Scheffer et al., 2001). The recovery of ecosystems from hypoxia may
occur only after long periods of time (Diaz, 2001; Galloway et al.,
2008; Kemp et al., 2009) or with further reductions in nutrient inputs.
Perhaps the slow response of the coastal systems in the Baltic Sea to re-
duced nutrient inputs is due to feed-back mechanisms such as en-
hanced P release (Conley et al., 2009a, 2009b) and decreases in
denitrification from sediments (Testa and Kemp, 2012) that keep eco-
systems hypoxic. In addition, it may also take time for nutrients that
have accumulated in the sediments during the years of excessive nutri-
ent loading to be buried out of the reactive zone or flushed out of the
ecosystem.
Previous studies have shown that a trend reversal is possible when
reducing nutrient loads (Carstensen et al., 2006, 2011; Nixon, 2009;
Norkko et al., 2012). What is clear from our results is that there are
many coastal systems with decreasing nutrients, but not many showing
decreasing hypoxia. However, it is probable that the ecosystem charac-
teristics, composition, components and biogeochemical processes have
changed. Most coastal marine ecosystems are impacted by multiple fac-
tors that act to stress the ecosystem (Halpern et al., 2008; Micheli et al.,
2013). The removal of one pressure alone, such as nutrient reduction,
may prove insufficient for ecosystem recovery when other pressures
(e.g., overfishing, dredging, climate change, or pollution) have not
been removed (Duarte et al., in press). The classification of anthropo-
genic pressures shows that the highest cumulative impacts over the
Baltic Sea area were in the southern and south-western areas and in
the Gulf of Finland where hypoxia is still increasing and the lowest
index values found in the Gulf of Bothnia (Korpinen et al., 2012).
4.4. Implications for ecosystem management
Reductions in nutrient loadings to coastal marine ecosystems are the
only realistic measure to improve ecosystem health and reduce hypoxia
(Conley et al., 2009b). However, confident predictions of recovery tar-
gets with nutrient reductions are problematic due to the unique behav-
ior of ecosystems during the recovery process (Carstensen et al., 2011).
Unexpected hypoxia mitigation is proposed to have occurred with
the spread of an invasive species, e.g. the worm Marenzelleria spp.
(Norkko et al., 2012). Their bioirrigation activities acted as a driver of
ecological change affecting sediment P dynamics in the Stockholm
Archipelago, oxygenating the sediments and reducing nutrient leakage
facilitating the recovery of hypoxia. Other currently unidentified nonlin-
ear interactions and feedbacks may play a positive role in ecosystem
remediation if nutrient reductions are continued.
5. Conclusion
The important lesson learned is that the recovery of coastal marine
ecosystems from hypoxia with nutrient reductions is indeed possible
(Conley et al., 2011; Diaz and Rosenberg, 2008; this study). However,
nutrient reductions are not always enough and restoration efforts
should include actions to catalyze recovery and reduce the thresholds
to recovery (Duarte et al., in press). In addition, despite the multiple
stressors driving hypoxia, especially the global rise in temperatures,
our results demonstrate that managing nutrients can create positive
feedbacks for oxygen recovery to occur. In the absence of nutrient
reductions, there will not be recovery from hypoxia in coastal marine
ecosystems.
Acknowledgments
This paper is a contribution from the Multistressors Project funded
by a FORMAS Strong Research Environment and the COCOA Project
funded under the BONUS program.
References
Andersson, L., Rydberg, L., 1988. Trends in nutrient and oxygen conditions within the
Kattegat: effects of local nutrient supply. Estuar. Coast. Shelf Sci. 26, 559–579.
Bianchi, T.S., DiMarco, S.F., Cowan Jr., J.H., Hetland, R.D., Chapman, P., Day, J.W., Allison,
M.A., 2010. The science of hypoxia in the Northern Gulf of Mexico: a review. Sci.
Total Environ. 408, 1471–1484.
Carstensen, J., Conley, D.J., Andersen, J.H., Ærtebjerg, G., 2006. Coastal eutrophication and
trend reversal: a Danish case study. Limnol. Oceanogr. 51, 398–408.
Carstensen, J., Sánchez-Camacho, M., Duarte, C.M., Krause-Jensen, D., Marbà, N., 2011.
Connecting the dots: responses of coastal ecosystems to changing nutrient concen-
trations. Environ. Sci. Technol. 45, 9122–9132.
Carstensen, J., Andersen, J.H., Gustafsson, B.G., Conley, D.J., 2014. Deoxygenation of the
Baltic Sea during the last century. Proc. Natl. Acad. Sci. 111, vol. 15, 5628–5633.
http://dx.doi.org/10.1073/pnas.1323156111.
Cloern, J.E., 2001. Our evolving conceptual model of the coastal eutrophication problem.
Mar. Ecol. Prog. Ser. 210, 223–253.
Conley, D.J., Carstensen, J., Ærtebjerg, G., Christensen, P.B., Dalsgaard, W., Hansen, J.L.S.,
Josefson, A.B., 2007. Long-term changes and impacts of hypoxia in Danish coastal
waters. Ecol. Appl. 17, S165–S184.
Conley, D.J., Björck, S., Bonsdorff, E., Carstensen, J., Destouni, G., Gustafsson, B., Hiatanen, S.,
Kortekaas, K., Huosa, H., Meier, H.E.M., Müller-Karulis, B., Nordberg, K., Norkko, A.,
Nürnberg, G., Pitkänen, H., Rabalais, N.N., Rosenberg, R., Savchuk, O.P., Slomp, C.P.,
Voss, M., Wulff, F., Zillén, L., 2009a. Hypoxia-related processes in the Baltic Sea. Envi-
ron. Sci. Technol. 43, 3412–3420.
Conley, D.J., Bonsforff, E., Carstensen, J., Destouni, G., Gustafsson, B.G., Rabalais, N., Voss,
M., Zillén, L., 2009b. Tacking hypoxia in the Baltic Sea: is engineering a solution?
Environ. Sci. Technol. 43, 3407–3411.
Conley, D.J., Carstensen, J., Aigars, J., Axe, P., Bonsdorff, E., Eremina, T., Haahti, B.-M.,
Humborg, C., Jonsson, P., Kotta, J., Lännegren, C., Larsson, U., Maximov, A., Rodriguez
Medina, M., Lysiak-Pastuszak, E., Remeikaite-Nikiene, N., Walve, J., Wilhelms, S.,
Zillén, L., 2011. Hypoxia is increasing in the coastal zone of the Baltic Sea. Environ.
Sci. Technol. 45, 6777–6783.
Diaz, R.J., 2001. Overview of hypoxia around the world. J. Environ. Qual. 30, 275–281.
Diaz, R.J., Rosenberg, R., 2008. Spreading dead zones and consequences for marine ecosys-
tems. Science 321, 926–929.
Duarte, C.M., Conley, D.J., Carstensen, J., Sánchez-Camacho, M., 2009. Return to Neverland:
shifting baselines affect eutrophication restoration targets. Estuar. Coasts 32, 29–36.
Duarte, C.M., Borja, A., Carstensen, J., Elliott, M., Krause-Jensen, D., Marbà, N., 2014. Para-
digms in the recovery of estuarine and coastal ecosystems. Estuar. Coasts. http://dx.
doi.org/10.1007/s12237-013-9750-9.
Galloway, J.N., Townsend, A.R., Erisman, J.W., Bekunda, M., Cai, Z., Freney, J.R., Martinelli, L.A.,
Seitzinger, S.P., Sutton, M.A., 2008. Transformation of the nitrogen cycle: recent trends,
questions, and potential solutions. Science 320, 889–892.
Gilbert, D., Rabalais, N.N., Diaz, R.J., Zhang, J., 2010. Evidence for greater oxygen decline
rates in the coastal ocean then in the open ocean. Biogeosciences 7, 2283–2296.
Gustafsson, B.G., Schenk, F., Blenckner, T., Eilola, K., Meier, H.E.M., Müller-Karulis, B.,
Neumann, T., Ruoho-Airola, T., Savchik, O.P., Zorita, E., 2012. Reconstructing the
development of Baltic Sea eutrophication 1850–2006. Ambio 41, 534–548.
Hagy, J.D., Boyton, W.R., Keefe, C.W., Wood, K.V., 2004. Hypoxia in Chesapeake Bay,
1950–2001: long-term change in relation to nutrient loading and river flow. Estuaries
27, 634–658.
Halpern, B.S., McLeod, K.L., Rosenberg, A.A., Crowder, L.B., 2008. Managing for cumulative
impacts in ecosystem-based management through ocean zoning. Ocean Coast.
Manag. 51, 203–211.
HELCOM, 2007. The Baltic Sea action plan; Helsinki, Finland. www.helcom.fi/BSAP/.
Hetland, R.D., Di Marco, S.F., 2008. How does the character of oxygen demand control
the structure of hypoxia in the Texas–Louisiana continental shelf? J. Mar. Syst. 70,
49–62.

IPCC (Intergovernmental Panel on Climate Change, Climate Change), 2007. The Physical
Science Basis, Contribution of Working Group I to the Fourth Assessment Report of
the Intergovernmental Panel on Climate Change. Cambridge Univ. Press, New York.
Justić, D., Rabalais, N.N., Turner, R.E., 1996. Effects of climate change on hypoxia in coastal
waters: a doubled CO2 scenario for the northern Gulf of Mexico. Limnol. Oceanogr.
41, 992–1003.
Karlson, K., Bonsdorff, E., Rosenberg, R., 2007. The impact of benthic macrofauna for
nutrient fluxes from Baltic Sea sediments. Ambio 36, 161–167.
Kemp, W.M., Boynton, W.R., Adoli, J.E., Boesch, D.F., Coicourt, W.C., Brush, G., Cornwell, J.C.,
Fisher, T.R., Gilbert, P.M., Hagy, J.D., Harding, L.W., Houde, E.D., Kimmel, D.G., Miller,
W.D., Newell, R.I.E., Roman, M.R., Smith, E.M., Stevenson, J.C., 2005. Eutrophication
of Chesapeake Bay: historical trends and ecological interactions. Mar. Ecol. Prog. Ser.
303, 1–29.
Kemp, W.M., Testa, J.M., Conley, D.J., Gilbert, D., Hagy, J.D., 2009. Temporal responses
of coastal hypoxia to nutrient loading and physical controls. Biogeosciences 6,
2985–3008.
Korpinen, S., Meski, L., Andersen, J.H., Laamanen, M., 2012. Human pressures and their
potential impact on the Baltic Sea ecosystem. Ecol. Indic. 15, 105–114.
Limburg, K.E., Olson, C., Walther, Y., Dale, D., Slomp, C.P., Hoie, H., 2011. Tracking Baltic
hypoxia and cod migration over millennia with natural tags. Proc. Natl. Acad. Sci.
108, E177–E182.
Meier, H.E.M., Andersson, H.C., Arheimer, B., Bleckner, T., Chubarenko, B., Donnelly, C.,
Eilola, K., Gustafsson, B.G., Hansson, A., Jonathan, H., Höglund, A., Kuznetsov, I.,
MacKenzie, B.R., Müller-Karulis, B., Neumann, T., Niiranen, S., Piwowarczyk, J.,
Raudsepp, U., Reckermann, M., Ruoho.Airola, T., Savchuk, O.P., Schenk, F., Schimanke,
S., Väli, G., Weslawski, J.-M., Zorita, E., 2012. Comparing reconstructed past variations
and future projections of the Baltic Sea ecosystem-first results from multi-model
ensemble simulations. Environ. Res. Lett. 7, 1–8.
Micheli, F., Halpern, B.S., Walbridge, S., Ciriaco, S., Ferretti, F., Fraschetti, S., Lewison, R.,
Nykjaer, L., Rosenberg, A.A., 2013. Cumulative human impacts on Mediterranean
and Black Sea marine ecosystems: assessing current pressures and opportunities.
PLoS One. http://dx.doi.org/10.1371/journal.pone.007988.
Nixon, S.W., 1995. Coastal marine eutrophication: a definition, social causes, and future
concerns. Ophelia 41, 199–219.
Nixon, S.W., 2009. Eutrophication and the macroscope. Hydrobiologia 629, 5–19. http://
dx.doi.org/10.1007/s10750-009-9759-z.
Norkko, J., Reed, D.C., Timmermann, K., Norkko, A., Gustafsson, B.G., Slomp, C.P.,
Carstensen, J., Conley, D.J., 2012. A welcome can of worms? Hypoxia mitigation by
an invasive species. Global Change Biol. http://dx.doi.org/10.1111/j.1365-2486.2011.
02513.x.
Omstedt, A., Meuller, A., Nyberg, L., 1997. Interannual, seasonal and regional variations of
precipitation and evaporation over the Baltic Sea. Ambio 26, 484–492.
Perus, J., Bonsdorff, E., 2004. Long-term changes in macrozoobenthos in the Åland
archipelago northern Baltic Sea. J. Sea Res. 52, 45–56.
Rabalais, N.N., Gilbert, D., 2008. Distribution and consequences of hypoxia. In: Urban, E.,
Sundby, B., Malanotte-Rizzoli, P., Melillo, J.M. (Eds.), Watersheds, Bays and Bounded
Seas. Island Press, Washington, DC, pp. 209–226.
Rabalais, N.N., Diaz, R.J., Levin, L.A., Turner, R.E., Gilbert, D., Zhang, J., 2010. Dynamics and
distribution of natural and human-caused hypoxia. Biogeosciences 7, 585–619.
Rabalais, N.N., Cai, W.-J., Carstensen, J., Conley, D.J., Fry, B., Hu, X., Quinones-Rivera, Z.,
Rosenberg, R., Slomp, C.P., Turner, R.E., Voss, M., Wissel, B., Zhang, J., 2014.
Eutrophication-driven deoxygenation in the coastal ocean. Oceanography 27, 66–77.
Rönnberg, C., Bonsdorff, E., 2004. Baltic Sea eutrophication: area-specific ecological conse-
quences. Hydrobiologia 514, 227–241.
Scheffer, M., Carpenter, S., Foley, J.A., Folke, C., Walker, B., 2001. Catastrophic shifts in eco-
systems. Nature 413, 591–596.
Sohlenius, G., Emeis, K.-C., Andrén, E., Andrén, T., Kohly, A., 2001. Development of anoxia
during the Holocene fresh-brackish water transition in the Baltic Sea. Mar. Geol. 177,
221–242.
Steckbauer, A., Duarte, C.M., Carstensen, J., Vaquer-Sunyer, R., Conley, D.J., 2011. Ecosys-
tem impacts of hypoxia: thresholds of hypoxia and pathways to recovery. Environ.
Res. Lett. 6. http://dx.doi.org/10.1088/1748-9326/6/2/025003.
Stramma, L., Johnson, G.C., Sprintall, J., Mohrholz, V., 2008. Expanding oxygen-minimum
zones in the tropical oceans. Science 320, 655–658.
Testa, J.M., Kemp, W.M., 2012. Hypoxia-induced shifts in nitrogen and phosphorus cycling
in Chesapeake Bay. Limnol. Oceanogr. 57, 835–850.
Turner, R.E., Rabalais, N.N., Justić, D., 2008. Gulf of Mexico hypoxia: alternate states and a
legacy. Environ. Sci. Technol. 42, 2323–2327.
Vallin, L., Nissling, A., Westin, L., 1999. Potential factors influencing reproductive success
of Baltic cod, Gadus morhua: a review. Ambio 28, 92–99.
Vaquer-Sunyer, R., Duarte, C.M., 2010. Sulfide exposure accelerates hypoxia-driven mor-
tality. Limnol. Oceanogr. 55, 1075–1082.
Wulff, F., Stigebrandt, A., Rahm, L., 1990. Nutrient dynamics of the Baltic Sea. Ambio 19,
126–133.
Zillén, L., Conley, D.J., Andrén, T., Andrén, E., Björck, S., 2008. Past occurrence of hypoxia in
the Baltic Sea and the role of climate variability, environmental change and human
impact. Earth Sci. Rev. 91, 77–92.

Caballero-Alfonso et al. 2015

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Viewers also liked

Viewers also liked (17)

Similar to Caballero-Alfonso et al. 2015

Similar to Caballero-Alfonso et al. 2015 (20)

Caballero-Alfonso et al. 2015