...

 Mass balance analysis of phosphorous in Motala Ström River Basin Linköpings Universitet

by user

on
Category: Documents
1

views

Report

Comments

Transcript

 Mass balance analysis of phosphorous in Motala Ström River Basin Linköpings Universitet

Water and Environmental Studies
Department of Thematic Studies
Linköping University
Mass balance analysis of phosphorous
in Motala Ström River Basin
– A focus on lake Roxen and Glan
Nathalie Stärner
Master’s programme
Science for Sustainable Development
Master’s Thesis, 30 ECTS credits
ISRN: LiU-TEMAV/MPSSD-A-12/012--SE
1
Linköpings Universitet

Water and Environmental Studies
Department of Thematic Studies
Linköping University
Mass balance analysis of phosphorous
in Motala Ström River Basin
– A focus on lake Roxen and Glan
Nathalie Stärner
Master’s programme
Science for Sustainable Development
Master’s Thesis, 30 ECTS credits
Supervisor: Hans Bertil Wittgren
2012
2
Upphovsrätt
Detta dokument hålls tillgängligt på Internet – eller dess framtida ersättare – under 25 år från
publiceringsdatum under förutsättning att inga extraordinära omständigheter uppstår.
Tillgång till dokumentet innebär tillstånd för var och en att läsa, ladda ner, skriva ut enstaka kopior
för enskilt bruk och att använda det oförändrat för ickekommersiell forskning och för undervisning.
Överföring av upphovsrätten vid en senare tidpunkt kan inte upphäva detta tillstånd. All annan
användning av dokumentet kräver upphovsmannens medgivande. För att garantera äktheten,
säkerheten och tillgängligheten finns lösningar av teknisk och administrativ art.
Upphovsmannens ideella rätt innefattar rätt att bli nämnd som upphovsman i den omfattning som
god sed kräver vid användning av dokumentet på ovan beskrivna sätt samt skydd mot att dokumentet
ändras eller presenteras i sådan form eller i sådant sammanhang som är kränkande för
upphovsmannens litterära eller konstnärliga anseende eller egenart.
För ytterligare information om Linköping University Electronic Press se förlagets hemsida
http://www.ep.liu.se/.
Copyright
The publishers will keep this document online on the Internet – or its possible replacement – for a
period of 25 years starting from the date of publication barring exceptional circumstances.
The online availability of the document implies permanent permission for anyone to read, to
download, or to print out single copies for his/her own use and to use it unchanged for noncommercial research and educational purpose. Subsequent transfers of copyright cannot revoke this
permission. All other uses of the document are conditional upon the consent of the copyright owner.
The publisher has taken technical and administrative measures to assure authenticity, security and
accessibility.
According to intellectual property law the author has the right to be mentioned when his/her work
is accessed as described above and to be protected against infringement.
For additional information about Linköping University Electronic Press and its procedures for
publication and for assurance of document integrity, please refer to its www home page:
http://www.ep.liu.se/.
© Nathalie Stärner
3
Table of Contents
2
Introduction .................................................................................................................................... 6
3
Background .................................................................................................................................... 7
4
3.1
P cycling in lakes .................................................................................................................... 7
3.2
Mass Balances ........................................................................................................................ 8
3.3
Motala Ström River Basin ...................................................................................................... 9
3.4
The Lakes Roxen and Glan................................................................................................... 11
Materials and Methods ................................................................................................................. 12
4.1
4.1.1
Phosphorous concentration data ................................................................................... 13
4.1.2
Modelled water flow data ............................................................................................. 15
4.2
5
6
Data ...................................................................................................................................... 12
Calculation methods ............................................................................................................. 16
4.2.1
Quality control of phosphorus concentration data......................................................... 16
4.2.2
Seasonal Mann Kendall trend analysis ......................................................................... 16
4.2.3
Interpolation of phosphorus concentration data ............................................................ 18
4.2.4
Phosphorus mass balance calculations .......................................................................... 18
4.2.5
Visualization of methodology ....................................................................................... 19
Results and Analysis .................................................................................................................... 20
5.1
Deviations in phosphorus concentration data ....................................................................... 20
5.2
Seasonal Mann Kendall trend analysis ................................................................................. 25
5.3
Phosphorus mass balances .................................................................................................... 26
5.3.1
Lake Roxen................................................................................................................... 26
5.3.2
Lake Glan ..................................................................................................................... 28
5.3.3
Visualization of the mass balances ............................................................................... 30
Discussion .................................................................................................................................... 32
6.1
Lake Roxen .......................................................................................................................... 32
6.2
Lake Glan ............................................................................................................................. 34
6.3
Critical considerations .......................................................................................................... 35
6.3.1
6.4
Mass balance calculations and models .......................................................................... 35
Further studies ...................................................................................................................... 36
7
Conclusion ................................................................................................................................... 37
8
Acknowledgements ...................................................................................................................... 37
9
References .................................................................................................................................... 37
4
Abstract
Phosphorous (P) has been found to be the limiting nutrient in freshwater systems, directly
affecting rates of planktonic growth. The P circulation is very complex, and its pathways
through lake systems are difficult to determine. Motala Ström is the biggest watercourse in the
south-east of Sweden and an important source of P to the Baltic Sea. The aim of this study is
to conduct a P mass balance analysis of the lakes Roxen and Glan over a period of time. The
analysis will also include a quality control of the concentrations data. The P concentration
data used in this investigation were collected from the Motala Ström River Association,
consisting of seasonal or monthly concentration data of Tot-P during the period 1960-2010.
Daily water flow data used in this study were modelled by the Swedish Meteorological and
Hydrological Institute (SMHI) using the S-HYPE model. P concentration deviations from
monthly averages at each sampling station were calculated, followed by a seasonal Mann
Kendall trend analysis. At five out of eight sampling stations, negative trends were detected,
indicating decreasing concentrations. The exception was the outflow from lake Glan, Stångån
and Finspångsån. Linear interpolation of P concentration data was performed to create daily
data for the period 1980-2010. Following interpolation, daily transport values were calculated
and summed up to annual values. Lake Roxen has acted as a source of P during the whole
period 1980-2010, except for one year. Lake Glan has acted as a source during 22 of the 31
years. There is a tendency of Glan to become more of a source over the years, which is in line
with the deviation observations, but variation between years makes it necessary to analyse
also future data in order to establish any possible trend in P transports. Before construction of
wastewater treatment plants, the lakes were certainly sinks of phosphorus. But at least for
Roxen, the switch from sink to source was completed before 1980.
Key words: Phosphorous, mass balance, deviation, seasonal Mann Kendall trend analysis,
Motala Ström, Roxen, Glan
5
2 Introduction
Natural ecosystems can be affected by increased nutrient input, which may lead to an increase
in primary productivity and eutrophication (Lau & Lane, 2001). Eutrophication changes lake
system properties as plant community, chemical functions, ecosystem services etc. beyond
recognition (Bennett et al. 2002, Lau & Lane, 2001, Sollie et al. 2008). Nuisance algal blooms
are the most severe symptoms of algal blooms in an eutrophic event and can eradicate
macrophytes, kill animals and cause illness to humans (Lau & Lane 2001). Shallow lakes tend
to have smaller dilution capacity and provide more circulation than deep water lakes, which
can seem more prone of eutrophication (Lau & Lane, 2001). The nutrients phosphorous (P)
and nitrogen (N) are widely recognised as the most important nutrients in aquatic systems,
affecting plankton growth and causing eutrophication (Asaoka et al. 2011, Bennett et al. 2002,
Dapeng et al. 2011, Maguire et al. 2000, Wan et al. 2010). P has been found to be the limiting
nutrient in freshwater systems, directly affecting rates of planktonic growth (Lau & Lane,
2001, Xiang & Zhou, 2011). P circulation is very complex, and its pathways through lake
systems are difficult to determine (Ulén & Kalisky, 2005). The amount of P available in a
freshwater system is depending on both internal and external sources (Xiang & Zhou, 2011).
External sources of P may be both anthropogenic and natural from either point or diffuse
sources (Zhong et al. 2008). Internal sources of P could be sediment releases as part of the
retention process, as P is adsorbed in sediments and could be resuspended back into the water
column. Hence, a freshwater system may still be fed with P even if external sources are
reduced, creating a continuous eutrophication event and the future of a lake system depends
on the internal impact and the rate the internal impact declines (Naturvårdsverket, 2006,
Xiang & Zhou, 2011). There are very few studies on effective actions towards P impacts,
investigations regarding internal turnover on P, qualitative and quantitative studies of the
adsorption and desorption of phosphorous under different oxygen concentrations and how the
processes are regulated and how the P is available for plankton (Naturvårdsverket, 2006).
Motala Ström is the biggest watercourse in the south-east of Sweden and one of the main
sources of P to the Baltic Sea from the Baltic southern water district (Vattenmyndigheterna,
2009). The outlet of the Motala Ström mainstream is to Bråviken, which is directly connected
to the Baltic Sea. As the internal releases from sediments make it difficult to evaluate the
effectiveness of actions on lowering external sources of P into the Motala Ström mainstream,
it is of importance to investigate the P transport into Bråviken more thoroughly and over a
larger timescale than in previous investigations. The internal source of P may be of significant
input to coastal zones, creating a need for a more thorough quantitative investigation of P
fluxes within the Motala Ström mainstream.
The aim of this study is to investigate the phosphorus (P) mass balances of the lakes Roxen
and Glan over time, to determine if Roxen and Glan acts as sources or sinks of P and if this
has changed over time. The study will also include an analysis of the quality of the P
concentration data series from the sampling sites in the Motala Ström River Basin. The
timescale 1980-2010 was evaluated as proper timescale due to the amount of available data.
The specific research questions are:
6
-
How can eventual deviations in the P concentration data series be explained, and how
can they affect mass balance calculations?
-
Are there any trends in annual transport of P in the different sampling sites during the
time period 1980-2010 and how can the eventual trends be explained?
-
Are the lakes Roxen and Glan to be classified as sinks or sources for P and how has
their character as sinks/sources changed over time during 1980-2010?
3 Background
3.1 P cycling in lakes
Phosphorous (P) is a mineral which is slowly circulating in a cycle which takes millions of
years to complete (Bennett et al. 2002). The mineral is present in continental rock and is
released by weathering or erosion, naturally occurring in soil and water (Brandt et al. 2006). P
has been recognized as the limiting nutrient in lakes and has major impact on ecological
functioning of the ecosystem (Bennett et al. 2002, Dapeng et al. 2011, Håkanson et al. 2003).
It exists in several different forms, as organic or inorganic, and as loosely-fixed, particulate or
strongly chemical bound (Andersson, 2006, Brandt et al. 2006). Specific forms of P are
bioavailable as PO4 or adsorbed loosely-fixed or easily degradable P in organic materials
(Brandt et al. 2006, Du et al. 2011). These P fractions are important factors in lake
eutrophication. The amount of bioavailable P is determined by external load and internal
processes within the water body. Nitrogen (N) and carbon (C) may leave a lake ecosystem in
gaseous states through respiration and denitrification, however P continues to circle within the
water body in a retention process as does not have a gaseous state (Lau & Lane, 2001).
Retention is a collective term for a multiple natural biogeochemical processes which takes
place in watercourses and lakes (Brandt et al. 2009). P retention processes are adsorption,
desorption, redox reactions, biological uptake, erosion, sedimentation, mineralization and
resuspension and these processes may vary depending on the physical and chemical properties
of a lake; temperature, dissolved oxygen concentration, nitrates, sulphates, bacteria activity,
salinity, runoff, residence time of the water, residence time of the nutrients, sediment type
capacity, pH etc. (Brandt et al. 2006, Brandt et al. 2009, Perrone et al. 2008). The lake
sediment has an important role in the internal P cycle and retention processes as it
accumulates in sediments and is storing previous external input of P, and it can be
resuspended from the sediment, acting as an internal source of P to the water body (Asaoka et
al. 2011, Bennett et al. 2002, Brandt et al. 2006, Dapeng et al. 2011, Nizzoli et al. 2011,
Perrone et al. 2008, Zhong et al. 2008). The concentration of P in the sediment is often higher
than in the water column, and the release depends on the distribution and concentration in the
sediment, the degree of saturation of exchangeable P, the intensity of the biological processes
and the hydrological patterns of the water body (Xiang & Zhou, 2011). Sediment
resuspension involves both abiotic and biotic processes and different types of available P
forms causes different eutrophication potential (Du et al. 2011, Xiang & Zhou, 2011).
Turbulence is one of the more important factors regulating the particulate settling velocity in
7
sediments and resuspended particles have a shorter distance to settle back into the sediment
than primary particles in lake surface waters (Malmaeus & Håkanson, 2004). The
resuspension may cause high internal P load especially in shallow, holmitic lakes as the
sediment is more likely to be exposed to different physical environment which may change
mobilization and release of P (Rydin et al. 1998). The main inorganic forms of P are
associated with aluminium (Al), iron (Fe) and manganese (Mn) oxides and hydroxides, where
iron phosphate (FePO4) complexes are strongly associated with P in sediments (Perrone et al.
2008, Xiang & Zhou, 2011). These compounds are easily desorbed. Redox conditions impact
the mobility of P in sediments and in anaerobic conditions, where a Fe reduction may
mobilize Fe bound P to slowly increase the P concentrations and bio available P to the bottom
waters (Brandt et al. 2006, Kokfelt et al. 2010, Malmaeus & Håkanson, 2004). Therefore the
amount of available Fe in the sediment to some extent controls some large sediment releases
of P (Perrone et al. 2008., Xiang & Zhou, 2011). However, sediment releases of P may even
at oxic conditions contribute up to 99% of total P input in some shallow lakes (Xiang & Zhou,
2011). The important mechanisms of transporting the P to the upper water column are bio
turbation, wind induced turbulence and diffusion (Brandt et al. 2006). Concentrations of P in
water bodies and sediments heavily vary in time and space. The chemical processes occurring
at the sediment interface are operating in synergy with the dynamics of the water body, with
the in- and outflows, currents, water mixing, and this complex combination of all the
processes makes every lake a unique case (Perrone et al. 2008). As time passes, P in the
sediment is mineralized back into the continental rock (Bennett et al. 2002). The cycle is
continued by the tectonic movement and continuous erosion of bed rock.
The different chemical, biological and physical processes in combination with anthropogenic
activities in P input determine the concentrations of different fractions of P within the soil,
sediment and water column (Brandt et al. 2006). Anthropogenic activities are continuously
disturbing the natural ecological cycle by adding P both from point and non-point sources as
runoff, soil leaching, rainfall, industrial end urban effluents etc. (Bennett et al. 2002, Dapeng
et al. 2011, Håkanson et al. 2003, Perrone et al. 2008, Zhong et al. 2008). Nutrient reductions
have mainly focused on P reductions from point sources as P is known to play a key role in
freshwater eutrophication and reduction measurements are easily introduced in existing waste
water treatment systems (Grimvall et al, 2000). A Swedish waste water treatment plant
normally reduces 90% of the P, and larger plants reduce even more (Naturvårdsverket, 2006).
The proportion of fractions in a watercourse depends on soil/sediment type, and the source
and forms of P which are added to the system (Reddy et al. 2011).
3.2 Mass Balances
Mass balance calculations have been incorporated into some models in order to enable
quantification of source contribution (Grimvall & Stålnacke, 1996). In general, a water
balance or mass balance equation, states that all water or concentrations of a substance which
is entering into a lake/basin, during any particular period of time, must either be placed as
storage, consumed, or leaving the lake within that certain time period (Singh et al. 2009). The
general water balance equation is hence the change in storage, and could be calculated as:
Storage = Inflow – Outflow
8
where the storage is the sedimentation and retention in the lake, inflow is everything added to
the system, and outflow is the added outflows of the system (Grimvall & Stålnacke, 1996,
Singh et al. 2009). Wahlin & Grimvall, (2010) conducted a mass balance investigation by
using concentration and discharge data to investigate monthly riverine loads, by normalising
the data and computed adjusted the annual flow weighted concentrations. They investigated in
spatiotemporal trends in annual summaries for all of the rivers with the same recipient. In this
investigation of P transport in the Motala Ström River Basin (MSRB), P entering and leaving
the lakes through a quantitative mass balance analysis has been an important statistical
approach.
3.3 Motala Ström River Basin
MSRB is the largest catchment area in the southeast of Sweden (MSV, 2010). The
watercourse is built up of the mainstream, reaching from Tiveden forests in Västergötland to
Bråkviken in Östergötland, and four larger branches, Stångån and Svartån are entering the
mainstream from the south, and Finspångsån and Ysundaån are entering from the north (see
Figure 1). Subdivided, the MSRB drainage areas are downstream Vättern, Storåns drainage
area, Söderköpingsåns drainage area, Vindåns drainage area, Östergötlands coastal regions,
Bråviken coastal water and the Southern Östergötland coastal waters (Naturvårdsverket,
2003a). Land use practices in the northern and southern parts of the MSRB are dominated by
forestry, while the central parts consist of distinctive plains with large agricultural practises
(MSV, 2010, Vattenmyndigheterna, 2009). The water course varies from deep, nutrient-poor
lakes in the southern parts of the MSRB, to shallow, nutrient-rich plain lakes in the central
parts along the mainstream (MSV, 2010). The two shallow and nutrient rich plain-lakes
Roxen and Glan are situated along the central mainstream (Vattenmyndigheterna, 2009). Glan
is located close to the mainstream outlet in Bråviken, and Roxen is connected to Glan without
any major branches affecting the water between the lakes. The outlet of MSRB in Bråviken
connects the MSRB to the Baltic Sea (MSV, 2010). The water turnover in Bråviken is
relatively fast with the water volume being replaced within a month. A high nutrient inflow
from the MSRB therefore causes a direct impact on eutrophication in Bråviken and the Baltic
Sea.
In the 1950s, concerns grew about the quality of the water in Motala Ström (MSV, 2010). The
watercourse supporting several communities and cities with freshwater and was also polluted
with untreated waste water. Motala Ström River Association (Motala Ströms
Vattenvårdsförbund, MSV), was developed in 1955 replacing local governments and
individual companies management of the river basin; instead the drainage area is monitored
by the MSV. An action plan was developed for the MSRB, with the goal to achieve “good
status” and in order to meet the goal restrictions of P releases; it is suggested to reach a
maximum of 25μg/l total P concentrations (Naturvårdsverket, 2003a, Vattenmyndigheterna,
2009).
9
Figure 1. Map over the Roxen and Glan section of MSRB with names of connecting waterways. Map constructed with info
from Hitta.se (2012), and SMHI (2012).
There are 23 water supply stations in the MSRB mainstream and 14 protected areas
(Vattenmyndigheterna, 2009). The MSRB mainstream supplied water to around 270 000
inhabitants in the cities Linköping and Norrköping in 2009. The MSRB branches are also
used for power generation, with several power stations along the watercourse (MSV, 2010,
Tibblin, 2011). Göta Kanal runs along the MSRB and boat traffic on the channel impact the
river basin (MSV, 2010). Previous P impact on the MSRV and the plain-lakes has caused
absorption of P to the sediment (MSV, 2010).
The non-point sources are mainly driven by precipitation and fertilizer applications, where
anthropogenic activities as application rates of fertilizers, livestock quantity and population
size are important factors influencing dissolved non-point sources of P to freshwater systems
(Wang et al. 2011). The single largest point sources of P emissions to the MSRB have been
local sewage treatment plants, large scale waste management and industrial sites within the
central mainstream (MSV, 2010). There are 23 local waste water treatment plants (WWTPs)
along the MSRB catchment area. Four of these are considered large with a load from more
than 10 000 people; Linköping, Mjölby, Motala and Finspång. There are a few which have a
few thousand people connected, but most of the WWTPs are very small, below 1000 people.
Many of the WWTPs were built in the 1970s even if some existed before, and most were
constructed with mechanical, biological and chemical cleaning steps. The release of P to
Roxen was large in the 1970s before the WWTPs installed biological and chemical treatment
and it has been suggested to have been absorbed in Roxen sediment (Naturvårdsverket,
2003b). One large point source emission reduction was the Linköping WWTP Nykvarn,
10
which introduced the treatment step with chemical residue in 1974, highly decreasing the P
load to Roxen. There are 9 WWTPs in Svartån, which transport to Roxen. P emissions
affecting lake Glan from WWTPs were and still are from Axäter, Skärblacka and Vånga
WWTPs, however Skärblacka WWTP was connected to the Norrköping city WWTP in
February 2010 and is not releasing emissions to MSRB since then (Norrköping Vatten, 2009).
Most of the current WWTPs have a cleaning coefficient of 90-95% where larger plants
usually have very good P reduction and low release concentrations and smaller plants usually
have slightly worse cleaning coefficient due to the size of the plants.
There are three larger industrial sites which cause significant releases of P into the MSRB
mainstream; Skärblacka Bruk, Aspa Bruk and Munksjö AB (Naturvårdsverket, 2003b). Aspa
Bruk and Munksjö AB are emitting to lake Vättern and Skärblacka Bruk is a paper mill
situated 3km upstream of the Glan inflow releasing into the mainstream (Billerud, 2007,
Naturvårdsverket, 2003b). Skärblacka Bruk modern facility was developed in the 1960s,
constructing a sulphate factory and two new paper machines. Large investments have been
implemented of environmental purpose, which has resulted in large emission reductions.
External treatment facilities have been constructed to treat the process water before it is
released. The wastewater from the paper mill, sulphate factory, fluting and cleaning
compartments are lead to a pre-sedimentation plant which consist of two ponds, 1420m2 large,
and here 85% of the particles are sedimented. The sludge from these ponds is pumped to a
dewatering plant where it is dried and combusted, and the water is transported to the longterm-active-aerated-sludge-plant (LAS), which was developed in 1998-1999 (Billerud, 2007,
Naturvårdsverket, 2003b). LAS is constructed with one machine house with a warmregulating system for cooling incoming water to 40 oC, tree dams and one sedimentation basin
(Billerud, 2007). The first dam is a 20 000m3 large anoxic dam and the second dam is an
86 000m3 large aerated dam. In the second dam, microbes reduce oxygen consummative
substances, and at the same time adsorbing P and N. After the second dam, the water is lead
into the sedimentation basin where the microorganisms are sedimented. The sludge is
removed, dried and combusted, and the water is lead to the last dam, as 150 000m3 cooling
dam. As the water passes through this dam, the temperature gradually decreases. The
outgoing water from the dam is used to cool incoming water to LAS, and 40 oC are the best
conditions for bacteria. The water from the cooling dam is then transported to two sediment
ponds, where a smaller amount of bio sludge is collected. The sludge is pumped back to the
pre-sedimentation tank and the water is released into Motala Ström (Billerud, 2007).
3.4 The Lakes Roxen and Glan
The Roxen plain-lake is situated in the central of Östergötland (Figure 1; Tibblin, 2011). It is
the largest lake within the county covering 97 km 2, and is 33 meters above mean sea levels.
The lake is shallow with average depth around 5 meters and max depth around 8 meters, and
stratification rarely occurs (Naturvårdsverket, 2003a, Tibblin, 2011, Vattenmyndigheterna,
2009). Roxen lake has a major function for the Motala Ström mainstream as it brings together
three major watercourses within the MSRB (Tibblin, 2011). Motala Ström mainstream is the
largest inflow to the lake coming from Vättern and Boren, Svartån is the second largest inflow
and originates from Småland and through Sommen, and Stångån is the third largest inflow to
11
Roxen, also originating from Småland. There are also 15 smaller inflows to the lake. Many of
the watercourses connecting to Roxen are impacted by several power stations.
The Roxen lake is positioned in a natural setting of a “fault precipice” reaching from east to
west from Bråviken to Vättern across Östergötland, dividing the landscape in 2 distinct
natural settings (Tibblin, 2011). The landscape north of the lake is dominated by pine forest
and the southern landscape is dominated by agricultural land, called Östgötaslätten or Östgöta
plain. There is an expanded shore protection, to 150 meters from the lake, and the western
part of Roxen is classified as a Natura 2000 area as the lake has a rich birdlife and fish fauna
with over 250 bird species and 21 naturally reproducing fish species (Naturvårdsverket,
2003a, Tibblin, 2011, Vattenmyndigheterna, 2009). Both lakes are classified of national
interest for professional fishing.
The Roxen lake is largely affected by eutrophication and has been considered of risk for long
term algal blooms in the future (Tibblin, 2011). It has also been evaluated that Roxen needs a
lowered concentration of P in order to achieve a good ecological status (Naturvårdsverket,
2003a, Vattenmyndigheterna, 2009).
Glan is a large shallow lake in the Östergötland County, situated close to the city of
Norrköping (Naturvårdsverket, 2003a). It is the second largest lake within the county after
Roxen covering 79 km2 and reaches 22 meters over the mean sea level
(Nationalencyklopedin, 2012). The lake is shallow with average depth 10 meters and max
depth is 23 meters (SMHI, 2009). The largest inflow to the lake is Motala Ström mainstream
from Roxen, and the two branches Finspångsån and Ysunda/Lotorpsån which enters from the
northern parts of the MSRB area. Glan is impacted with high concentrations of P both in inand outlet water as well as in the sediment (Naturvårdsverket, 2003a).
Previously, Glan had been considered a natural nutrient treatment plant due to large
adsorption of the sediments (MSV, 2010). Large amounts of P were previously added to the
lake from a paper mill, Skärblacka Bruk, and this release has been reduced by 16 tons of P per
year (Naturvårdsverket, 2003a). A simple mass balance of P in Glan conducted by MSV
showed a higher concentration of P in the outflow water than in the inflow (MSV, 2010). Due
to this, a suspicion has grown that during the last few years a desorption processes has
occurred within the lake sediment, releasing more P into the water column from the sediment.
4 Materials and Methods
This investigation has focused on Motala Ström River Basin mainstream and the lakes Roxen
and Glan with their branches Stångån and Svartån entering Roxen and Finspångsån and
Ysundaån entering Glan. The P concentrations data has been collected from the Motala Ström
catchment area, and the flow data has been received from SMHI.
4.1 Data
P emissions from Glan outflow to Bråviken are an important release of P to the Baltic Sea
(MSV, 2010). The nutrient transport depends on water flow, where a high flow of water may
12
cause a higher nutrient transport. MSV has previously in their annual reports conducted
simple mass balance analyses over short periods of time, where the transports of N and P has
been investigated using daily average flows from SMHIs S-HYPE model and interpolated
daily concentrations. Present investigation is aiming to perform a more quantitative analysis
using longer data series of a 30 year old time frame, and considering trends by investigating in
deviations, in addition to inspecting annual transport by mass balance calculations. The data
used within the investigation is presented below. There are areas in the catchment area which
have not been included in SMHI flow models or considered in the sampling sites, and these
areas could potentially add small amounts of P to Roxen or Glan lakes. The catchment areas,
which are not included in the data, were calculated and it can be assumed that 5.4% of the
Roxen catchment area and 1.8% of the Glan catchment area, including Sviestadsån, Kumlaån
and lakeside areas.
4.1.1 Phosphorous concentration data
Concentrations data from the MSRB is available from 1960-2010 in the MSV database, and
the sampling has been performed by MSV consultants (MSV, 2010). The P concentration data
used in this investigation consisted of seasonal or monthly concentration data of Total P (TotP) concentrations and was collected from the MSV webpage (see Figure 2 for a map over the
sampling sites and Table 1 for more detailed information on the sampling information). At
some sampling sites some values have been represented as “less than” values within the
concentrations data files, not presenting an exact data value. This was handled by presenting
half of the given “less than” value (ex. Tot-P conc. in the data file was presented as: <5 µg/l,
and was hence given the value: 2.5 µg/l). Data points which did not have concentration value
presented were removed from the dataset.
Figure 2. Map of sampling stations for the study site with site IDs, also included are the point sources Nykvarn WWTP
(blue star) and Skärblacka bruk (green star) (MSV, 2012). see table 1 for IDs and sampling sites.
13
Table 1. Sampling information of the concentration-data (MSV, 2012).
StID
Li12
Sampling sites
Motala Ström, Roxen inflow
Li05
Stångån, Nykvarn power station
Li13
Svartån, Svartåfors power station
Li11
Motala Ström, Roxen outflow
Gb02 Motala Ström, Glan inflow
Fi07
Finspångsån, Dovern outflow
Fi09
Ysundaån, Åmlången outflow
Gb06 Motala Ström, Glan outflow
Sampling frequency (samples/year)
1966-74: 1-2
1975-79: 4
1980-2010: 12 (3 missing data)
1966-74: 1-2
1975-84: 4 (1 missing data)
1985-2010: 12 (2 missing data)
1966-70: 1
1971-74: 2 (1 missing data).
1975-79: 4
1980-2010: 12 (4 missing data)
1966-70: 1
1971-75: 2-4
1976-2010: 12 (5 missing data)
1966-69: 1
1971-75: 2-4 (No data in 74)
1976-2010: 12 (1 missing data)
1974: 2
1975-81: 4
1982-2010: 12 (1 missing data)
1974: 2
1975-84: 4 (1 missing data)
1985-2010: 12 (2 missing data)
1969-2010: 12 (5 missing data).
2011: 4
Some point sources are not included in the measured concentrations in the phosphorous
concentrations data. The Nykvarn WWTP is situated downstream from the Li05 sampling
station in Stångån not recording its release of P to the watercourse. Roxen outflow has also
been used as Glan inflow data instead of the sampling station for Glan inflow for the mass
balance calculations (see 5.1 deviations in concentrations data), and Skärblacka bruk is
situated downstream from Roxen outflow, Li11, which means the paper mill emission is not
included either. The recorded emissions of annual P transport from these point sources has
been added to the annual transports in the data sheets; Nykvarns WWTP releases were added
to Stångån transport data and Skärblacka bruk releases were added to Roxen outflow transport
data in the cases where Roxen outflow act as Glan inflow in the analyse. The two point
sources have been marked out in Figure 2, and the annual releases of P into the waterways
from these two point sources are described in Table 2.
14
Table 2. Point source releases 1980-2010 in ton/year from the two point sources which are not included in the
concentrations data (Billerud, 2007, Göransson et al. 1991, Naturvårdsverket, 2003b, Tekniska Verken, 2012).
Point
source
Nykvarn
WWTP
Skärblacka
Bruk
19801989
6.2
19901998
4.9
1999
2000
4.9
3.9
15
22.4
14.6
6.2
20012008
3.9
20092010
3.4
4-5
(4.5)
4-5
(4.5)
The entire P concentrations dataset 1960-2010 were used in the deviation-calculations. For the
mass balance calculations the files were structured to only contain data from 1980-2010. This
was due to 12 month data sampling frequency was available for 5 of 8 sampling stations from
1980 and the rest in the beginning of the 1980s, whilst the sampling frequency before 1980s
were around 4 occasions per year.
4.1.2
Modelled water flow data
The development of the hydrologic drainage area model Hydrological Predictions for the
Environment (HYPE) which simulates water flows and substances from precipitation through
soil, rivers and lakes was carried out in 2005-2007 (Lindström et al. 2010, SMHI, 2012b). The
HYPE model is capable of describing major hydrological features in Sweden and S-HYPE
includes the drainage areas in the Swedish Water Archives (SVAR, Svenskt Vattenarkiv).
Vattenweb is giving access to simulations of water flows (m3/s) for the years 1990-2010. The
water flow data used in this study was supposed to be collected from vattenweb, however as
data were not available for the time period 1980-2010, new modelled data were ordered from
SMHIs S-HYPE covering the years 1979-2010 including a sampling-station correction as the
water is heavily regulated. The S-HYPE1 was recalibrated from regional parameter set to
better adjust for Motala Ström local conditions. The calibration was done in the fall of 2011,
and the model was run in April 2012 to produce the modelled data used in this research. All
values of flow data was collected from the same source and model-run and the sites, station
ID (StnID) and Area is presented in Table 3. Simulated concentrations data was not used from
the S-HYPE as real concentration values were available from MSV (2010).
Table 3. Water flow data, ID and Accumulated area upstream including bi-flows (SMHI, 2012a).
NAME
Stångån, Nykvarn power station
Stångån
Svartån, Svartåfors power station
Motala Ström, Roxen inflow
Motala Ström, Roxen outflow
Göta
kanal
Motala
Ström, Glan inflow
Motala Ström, Glan outflow
Ysundaån, Åmlången outflow
Finspångsån, Dovern outflow
StnID, coordinates
647837-148955
648105-148196
648648-148439
648785-150474
649645-150445
650049-150957
650990-150196
650622-149865
Area km2
2462
3429
6631
13244
13371
15362
424
1288
1
Used model: HYPE version 3.5.3. Used model setup S-HYPE 2012 1_1_0 recalibrated from regional parameter
setup to better adjust for Motala Ström local conditions.
15
4.2 Calculation methods
4.2.1 Quality control of phosphorus concentration data
Data quality is crucial for the reliability of investigations on riverine loads (Stålnacke et al.
1999). Inconsistency in data can be due to temporal problems in the quality assurance of the
reported data (Wahlin & Grimvall, 2010). Statistical procedures can be used to identify
inconsistencies in data sets which are of questionable quality (Stålnacke et al. 1999). The
uncertainty of a calculated load of P within a lake is a function of the raw data, the fulfilment
of the monitoring programs and the statistical methods used to combine observed nutrient
concentration with runoff data into load estimations. During a quality control of a data set, the
data time-series may first be visually inspected to reveal remarks in shifts in measured
concentrations and outliers (Stålnacke et al. 1999). Coefficients of variation may hence be
calculated. Where remarkable shifts in concentration are observed, a more detailed
investigation could be conducted involving calculation of variation after removal of linear
trends, and comparison with concentrations observed in other sampling campaigns. When a
quality control of the data has been completed, analytical errors or improper handling of the
data may be removed.
In this study, P concentration data were represented from the mid1960s until 2011 and the
consulting companies performing the sampling changed several times during the period. This
means data for different time periods and sites have been collected and analysed with
different methods and equipment. Due to this, a quality control was conducted for the P
concentrations data for the different stations, by using the method from Stålnacke et al.
(1999). The sampling sites can be viewed in Table 2. First, the average concentration of Tot-P
for each month 1960-2010 per one sampling station was calculated. As only one sample per
month was available, the average was calculated for a specific month using all values for that
month over the whole time period 1960-2010. The deviation was then calculated for all datasamples; calculating the deviation of every data point within the specific month in regard to
the average concentration of P for that month:
Devation = observed value – monthly average
The deviations were calculated for all data points and all sampling stations. When all
deviations were calculated, all sampling stations deviations were plotted separately against
time. The deviation plots make it possible to remove the seasonal variation, detect anomalies
in the data series and visually detect anomalies over the time period. The deviation
scatterplots were accompanied with a trend analysis in order to find out if there was a
presence of a trend and how district it was.
4.2.2 Seasonal Mann Kendall trend analysis
Procedures in trend analysis are generally built on regression and/or hypothesis testing (Helsel
& Hirsch, 2002). The variable in focus is generally time, however spatial and directional
trends may also be of interest. A trend is determined in terms of if it is increasing or
decreasing over time. In statistical terms, trends define whether there is a probability of the
distribution of a certain variable has changed over time and the amount or rate of which the
16
distribution has changed. There are several statistical methods existing to distinguish random
fluctuations and trends in time series data sets, and one of these is a widely used rank-based,
non-parametric test for detecting monotonic trends called Mann Kendall test (Hirono et al.
2009). The test is measuring the degree of correspondence between two variables and assesses
the significance of this correspondence (Modarres et al. 2012, Shadmani et al. 2012). Mann
Kendall has usually an advantage of the non-parametric tests as it is robust against departures
from normalities (Helsel & Hirsch, 2002). The Mann Kendall test may be used generally to
determine whether the values of the variable (Y) tend to increase or decrease over time.
Typically, the test is used for a specific purpose, as an example determining if the central
value or median changes over time. The test is commonly used to assess the significance of
trends in hydro-meteorological time series as water quality, stream flow, temperature,
precipitation, etc (Cheng et al. 2011).
In the time series of observed data, a null hypothesis and an alternative hypothesis are used to
describe non-existence and existence of a trend (Shadmani et al. 2012). A null hypothesis
describes no trend exist, and the outcome of a trend analysis is a decision whether or not to
reject the null hypothesis (Helsel & Hirsch, 2002). If the tested series has been determined to
be statistically significant at 5% significance level, the null hypothesis cannot be rejected if
the p value is >0.05 (Modarres et al. 2012, Sileika et al. 2006). A positive value of a
standardized Mann-Kendall statistic indicates an increasing trend while negative values
indicate decreasing trends (Cheng et al. 2011, Modarres et al. 2012, Shadmani et al. 2012).
Within a dataset, changes of different seasons may cause a source of variation in the data
variables Y (Helsel & Hirsch, 2002). Hence, seasonal variation must be incorporated,
compensated or removed in order for possibilities to analyse trend over time. The seasonal
Mann Kendall test is a refined version of the Mann Kendall test, which can perform
calculations representing the seasonal influence on periods (Cheng et al. 2011) and the
statistical significance of the p-value of the seasonal Mann Kendall may be slightly more
sensitive than of a Mann Kendall due to the seasonal variation. The seasonal Mann Kendall
test accounts for seasonality by computing the Mann Kendall test in each season separately
(Helsel & Hirsch, 2002). The seasons should be divided with enough length for data to
available for most seasons during the years of record, hence, if data are frequently collected
monthly, seasons may be divided into 12 months, but if data is collected quarterly seasons
should be divided quarterly. In 12 months seasonality of the Seasonal Kendall test, January is
compared only with January over the time period, February is compared with February only
etc.
The statistical method for performing the Mann Kendall trend analysis in this investigation
used an Excel-based programme called MULTITEST, which is focusing on univariate and
multivariate Mann Kendall tests for determining monotone trends (Grimvall et al. 2011). The
MULTITEST facilitates testing for trends in multiple time series to detect monotone
(increasing or decreasing) trends in time series of data. The programme performs trend test on
individual time series as well as testing overall trends in the data series. Hence, if data has
been collected over several seasons and sites, test is performed both for the individual site and
seasons as well as computing trends slopes.
17
The tested series has for this Mann Kendall trend analysis been regarded as statistically
significant at 5% significance level. The null hypothesis, describing ‘no trend exist’, cannot
be rejected if the p value is >0.05. The p-values <0.05 have been further divided in
MULITTEST in different grades of negative trends which is used in the results of the trend
analysis (see Table 4 in ch. 5.2). As no positive trends were detected, the grades only focus on
negative grades of trends.
4.2.3 Interpolation of phosphorus concentration data
When conducting mass balances daily values of the concentrations in between the sampling
dates must be estimated if using measured concentrations of monthly values (Göransson et al.
1991). There are several different methods to conduct this and linear interpolation was found
in a comparison with other methods by Göransson et al. (1991) to give the most robust
estimations. Linear interpolation is a curve fitting method where two known data points are
connected with a straight line, which is used to fill gaps within a table (Stålnacke et al. 1999).
It may be used to produce daily values of the data from monthly values, as long as the gaps
between data is not to large. Larger gaps may need a quality control of the data, decomposing
time series under consideration of trend, seasonal regulation and irregular variation before a
linear interpolation can be conducted.
The P concentrations data consisted of one concentration for one day within a month, but the
flow data was available in daily values for the time period. In order to calculate the transport,
concentrations were needed for every day during the whole time period. Linear interpolation
was used in order to produce data values for all days within a month. The P concentrations
data was interpolated after the quality control was conducted, and as the interpolated data
should include all concentrations data from January 1 st 1980- December 31st 2010. The tool
used for the interpolation was constructed by the country administration board in
Östergötland. The interpolation had to start in 1979 in order to get a complete dataset of daily
values Jan 1st 1980 to Dec 31st 2010, due to the structure of the interpolation tool. The
parameters gained from the interpolation tool daily values were P concentration, water flow
and transport of P.
4.2.4 Phosphorus mass balance calculations
After the data had been interpolated, the annual P transport was calculated for each individual
station by adding all daily transport values within one year, for one sampling station and
calculate the transport of P for each year from 1980-2010 for that station. The annual
transport of P was calculated for every station and for all years and presented in linear graphs
against time.
The difference between inflow was calculated for each individual lake by using the general
water balance equation (Storage = Inflow – Outflow) described by Singh et al. (2009). If the
inflows were higher than the outflow, the value of storage would be positive and could give
indication that the lake acts as a sink, and if the outflow is bigger than the inflows the value of
storage should be negative and could give indication that the lake act as a source.
18
A flow diagram was constructed for the mass balances in order to visually present the
information conducted from the study, using the software STAN. The mass balances were
calculated as annual mean of every 10 years within the time period (1981-1990, 1991-2000,
2001-2010). Added within the figure were except of the known inflows, also arrows within
the boxes representing internal input or output in regard to sediment absorption/desorption.
An arrow going into the box represents a source of P, and an arrow going out from the box is
representing a sink of P.
4.2.5 Visualization of methodology
There are several methodological steps within this investigation. In order to make it easier to
interpret the whole picture, Figure 3 has been developed to visually show how the analysis
has been conducted.
Figure 3. Methodological steps used in this investigation. Boxes represent a product and arrows represent analysing
steps.
19
5 Results and Analysis
The structure of the results and analysis section is first to present the deviations in
phosphorous concentration data, followed by the Mann Kendall trend analysis. Thereafter
phosphorous mass balances of the two lakes will be presented.
5.1
Deviations in phosphorus concentration data
The P concentration data were represented 1960-2011 and different years were sampled by
different consulting companies. The deviations were calculated for the concentration data in
order to remove the seasonal variation, to find eventual anomalies and if these could be
explained by changes in sampling analysing company. The deviation plots were also used for
visual investigation for possible trends in the data (Figure 4).
The plots were structured to show all data samples from all stations in white circles, named
“all cases” in Figure 4, as well as showing different sampling consultants for the
concentrations plots and only the deviations of the deviation plots. By visually inspecting the
plots of the concentrations data and the deviations, it was determined that there was larger
variation of the samples in the first years (1960-1980) of the time period for the sampling
stations. Some consultant companies’ seem to detect lower P limits than others, as example
VIAK AB in Figure 4. Generally negative deviations are more common in recent years for
many of the sites and there are some similarities in the data among the stations as a visual
detection of a slight negative trend occurs at many sampling sites. In order to further analyse
the existence and non-existence of trends, seasonal Mann Kendall trend analysis was
performed and is presented in section 5.2.
20
Motala Ström inflow to Roxen (Li12)
150
Motala ström inflow to Roxen (Li12)
Al l cases
100
ALcontrol AB
Okänt
100
KM Lab i Norrköpi ng; IVL(TOCanal yser)
Eurofi ns
Li mnoconsul t
75
Lst E-l än; MSV
Lst E-l än; VIAK AB
50
Mi l jökontrol l aboratori et i Uppsal a
VIAK AB
25
0
1965
80
Anomalies of total P µg/l
Total-P µg/l
Anal yCen Nordi c AB
125
IVL; SGAB (metal l anal yser)
60
40
20
0
-20
-40
-60
-80
KM Lab i Norrköpi ng
1975
1985
1995
2005
2015
-100
1965
KM Lab i Norrköpi ng/Li nköpi ng
MSV
1970
1975
1980
1985
Year
Stångån (Li05)
2010
2015
100
KM Lab i Norrköpi ng; IVL(TOCanal yser)
Eurofi ns
Li mnoconsul t
75
Lst E-l än; MSV
Lst E-l än; VIAK AB
50
Mi l jökontrol l aboratori et i Uppsal a
VIAK AB
25
1995
2000
2005
2010
2015
1995
2000
2005
2010
2015
80
Anomalies of total P µg/l
Okänt
Total-P µg/l
2005
100
Anal yCen Nordi c AB
125
IVL; SGAB (metal l anal yser)
60
40
20
0
-20
-40
-60
-80
KM Lab i Norrköpi ng
1975
1985
1995
2005
2015
-100
1965
KM Lab i Norrköpi ng/Li nköpi ng
MSV
1970
1975
1980
1985
Year
Svartån (Li13)
Svartån (Li13)
All cases
100
ALcontrol AB
AnalyCen Nordic AB
Okänt
100
KM Lab i Norrköping; IVL(TOCanalyser)
Eurofins
Limnoconsult
75
Lst E-län; MSV
Lst E-län; VIAK AB
50
Miljökontrollaboratoriet i Uppsala
VIAK AB
25
IVL; SGAB (metallanalyser)
KM Lab i Norrköping
1985
1995
2005
2015
KM Lab i Norrköping/Linköping
MSV
Year
80
Anomalies of total P µg/l
125
1975
1990
Year
150
Total-P µg/l
2000
Stångån (Li05)
Al l cases
ALcontrol AB
0
1965
1995
Year
150
0
1965
1990
60
40
20
0
-20
-40
-60
-80
-100
1965
1970
1975
1980
1985
1990
Year
Roxen outflow (Li11)
Roxen outflow (Li11)
All cases
150
100
ALcontrol AB
Okänt
100
KM Lab i Norrköping; IVL(TOCanalyser)
Eurofi ns
Limnoconsult
75
Lst E-län; MSV
Lst E-län; VIAK AB
50
Miljökontrollaboratoriet i Uppsala
VIAK AB
25
0
1965
80
Anomalies of total P µg/l
Total-P µg/l
AnalyCen Nordic AB
125
IVL; SGAB (metallanalyser)
60
40
20
0
-20
-40
-60
-80
KM Lab i Norrköping
1975
1985
1995
2005
2015
-100
1965
KM Lab i Norrköping/Linköping
MSV
1970
1975
1980
1985
Year
Motala Ström inflow to Glan (Gb02)
Okänt
2010
2015
100
KM Lab i Norrköping; IVL(TOCanalyser)
Eurofi ns
Limnoconsult
75
Lst E-län; MSV
Lst E-län; VIAK AB
50
Miljökontrollaboratoriet i Uppsala
VIAK AB
25
2000
2005
2010
2015
2000
2005
2010
80
Anomalies of total P µg/l
Total-P µg/l
2005
100
ALcontrol AB
AnalyCen Nordic AB
IVL; SGAB (metallanalyser)
60
40
20
0
-20
-40
-60
-80
KM Lab i Norrköping
1975
1985
1995
2005
2015
-100
1965
KM Lab i Norrköping/Linköping
MSV
1970
1975
1980
1985
Year
1990
1995
Year
Finspångsån (Fi07)
Finspångsån (Fi07)
All cases
150
100
ALcontrol AB
AnalyCen Nordic AB
125
Okänt
100
KM Lab i Norrköping; IVL(TOCanalyser)
Eurofi ns
Limnoconsult
75
Lst E-län; MSV
Lst E-län; VIAK AB
50
Miljökontrollaboratoriet i Uppsala
VIAK AB
25
IVL; SGAB (metallanalyser)
80
Anomalies of total P µg/l
Total-P µg/l
2000
Motala Ström inflow to Glan (Gb02)
All cases
125
0
1965
1995
Year
150
0
1965
1990
60
40
20
0
-20
-40
-60
-80
KM Lab i Norrköping
1975
1985
1995
2005
2015
KM Lab i Norrköping/Linköping
MSV
Year
-100
1965
1970
1975
1980
1985
1990
Year
1995
22
2015
Ysundaån (Fi09)
Ysundaån (Fi09)
All cases
150
100
ALcontrol AB
Okänt
100
KM Lab i Norrköping; IVL(TOCanalyser)
Eurofi ns
Limnoconsult
75
Lst E-län; MSV
Lst E-län; VIAK AB
50
Miljökontrollaboratoriet i Uppsala
VIAK AB
25
80
Anomalies of total P µg/l
Total-P µg/l
AnalyCen Nordic AB
125
IVL; SGAB (metallanalyser)
0
1965
60
40
20
0
-20
-40
-60
-80
KM Lab i Norrköping
1975
1985
1995
2005
2015
-100
1965
KM Lab i Norrköping/Linköping
MSV
1970
1975
1980
1985
Year
2000
2005
2010
2015
1995
2000
2005
2010
2015
Glan outflow (Motala Str
All cases
150
100
ALcontrol AB
AnalyCen Nordic AB
125
Okänt
100
KM Lab i Norrköping; IVL(TOCanalyser)
Eurofins
Limnoconsult
75
Lst E-län; MSV
Lst E-län; VIAK AB
50
Miljökontrollaboratoriet i Uppsala
VIAK AB
25
IVL; SGAB (metallanalyser)
80
Anomalies of total P µg/l
Total-P µg/l
1995
Year
Glan outflow (Motala Str
0
1965
1990
60
40
20
0
-20
-40
-60
-80
KM Lab i Norrköping
1975
1985
1995
2005
2015
KM Lab i Norrköping/Linköping
MSV
-100
1965
1970
1975
1980
Year
1985
1990
Year
Figure 4. Total P concentration (µg/l) and deviations ‘anomalies’ in the figures (measuring value – average value) for all of the sampling
stations. White circle named “all cases” represent all data from all stations in both the concentration and deviation plots.
23
As previously mentioned, differences can be detected in the deviations/concentrations plots of
Roxen outflow and the Motala Ström inflow to Glan. Therefore, a line diagram with both
stations was produced in order to investigate this relationship further and the deviation
between the sampling sites was found to be quite large (Figure 5). In some years, as 1986,
Glan inflow transport of P exceeds the Roxen outflow with more than 50 tons. Also a line
diagram showing the difference between Motala Ström inflow to Glan and Roxen outflow
was produced, showing a difference of up to 60 tons (Figure 6). There is no major
watercourses connecting to Motala Ström between these sampling sites, and Skärblacka bruk
does not have discharges of a magnitude to explain the difference. A reasonable explanation
for the differences in the data could be that there is some kind of error in the Glan inflow data.
This could be connected to the chosen sampling site at which Motala Ström inflow to Glan is
monitored. If the water downstream is regulated creating a damming effect, or if the
measuring site is located where the water circulation is poor, there is a risk of the water not
being well mixed and a higher concentration may be sustained. If the site is close to
Skärblacka Bruk those releases may as an example cause an increased P concentration at the
measuring site. Due to the uncertainty in the data of sampling site “Motala Ström inflow to
Glan” it was decided that this station is not reliable as representing Motala Ström inflow to
Glan and it was decided not to use this station further in this study. Instead, Roxen outflow
was used as representing the inflow of P entering Glan for the mass balance analysis, and
Skärblacka Bruk annual P releases were added to Roxen outflow in order to create appropriate
estimates.
MS Rox out
250
MS Glan in
Transport of P, ton/year
200
150
100
50
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
0
Figure 5. Transport of P in ton/year from Roxen outflow and Glan inflow.
70
50
40
30
20
10
0
-10
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Glan Inflow - Roxen outflow, ton/year
60
-20
Figure 6. The difference in P transport in ton/year between Roxen outflow and Glan inflow.
5.2 Seasonal Mann Kendall trend analysis
As the deviations plots (Figure 4) left some questions about whether trends could be visually
detected from the graphs, a seasonal Mann Kendall trend analysis was performed in order to
statistically determine whether the visual observations were correct. As previously mentioned,
the tested series has been regarded as statistically significant at 5% significance level. The
null hypothesis cannot be rejected if the p value is >0.05 and the p-values <0.05 have been
further divided in different grades of negative trends (Table 4).
Table 4. Presenting the different grades of trends used within the result.
Legend:
p > 0,05
0,01 < p < 0,05
no trend
-
negative trend
0,001 < p < 0,01 --
strong negative trend
p < 0,001
very strong negative trend
---
Table 5 shows the result of the seasonal Mann Kendall trend analysis. Three of the sampling
sites (Li05, Fi07 and Gb06) gave p-values >0.05 and the null hypothesis could not be rejected
for those sampling stations. For the rest of the sampling stations, the null hypothesis was
rejected as the p-value was <0.05. In all of these sampling sites negative trends could be
detected. As this seasonal Mann Kendall trend analysis is divided in one univariate test which
inspect the trends in seasonal variation for each month, it would be possible that positive
trends existed while the result still were negative as an overall result of p-value for the whole
sampling period and for all months. This did not occur within this analysis, but some monthly
p-values were different from the overall P value for the specific sampling station.
25
Table 5. Result of Seasonal Mann Kendall trend analysis.
Sampling site
Motala Ström, Roxen inflow
Stångån, Nykvarn power station
Svartån,
Motala Ström, Roxen outflow
Motala Ström, Glan inflow
Finspångsån, Dovern outflow
Ysundaån
Motala Ström, Glan outflow
St ID
Li12
Li05
Li13
Li11
Gb02
Fi07
Fi09
Gb06
p-value
0.0002
0.2963
0.0272
0.0028
< 0,0001
0.1163
0.0463
0.2180
Significance
code
------
5.3 Phosphorus mass balances
The annual transport of P was calculated for each year and each sampling stations. These
were presented as linear graphs for the separate lake systems in order to visually present the
inflows and outflows of the separate lakes. One extra variable was introduced in the mass
balances, which was the sum of all inflows to the lake. For Roxen, this was Motala Ström
inflow to Roxen, Stångån and Svartån, and for Glan this was Motala Ström inflow to Glan,
Finspångsån and Ysundaån.
5.3.1 Lake Roxen
The graphs for the Roxen lake system are presented in Figure 7. Overall, it may seem Roxen
outflow exceeds the value of Roxen total inflow except in 2009. The inflows to Roxen seem
to have slightly decreased over time. However, the major dip was in the period of 1985-1997,
thereafter some values have started to increase again in the later years. The annual transport of
P to Roxen from Stångån and Svartån seem to be quite similar. The transport of P between
different years seems to be very different; dropping or increasing with as much as 100 tonnes
over a few years at some sampling sites. It is not possible to see a correlation in the data
concerning Roxen lake total inflow and outflow over the time period 1980-2010 from this
graph.
26
160
140
Roxen out
120
Roxen in
(sum)
P transport, ton/yr
100
Svartån in
80
60
Stångån in
40
Motala
Ström in
20
0
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Year
Figure 7. Annual transport of P in ton/year in the Roxen lake
The difference between the annual inflow transports of P in relation to the annual outflow
transports of P was calculated by subtracting the outflow from all the inflows for Roxen lake
and was produced as a linear graph (Figure 8). This variable was represented with a solid line
in the graph. Another variable was also included in the graph, represented with a dashed line,
to show how the graph would look like if the inflow was slightly higher. This is due to the
knowledge that a part of the catchment area (5.4% of Roxen and 1.8% of Glan) has not been
included in the data, and the variable chosen was 10% higher than the calculated difference
value. A positive outcome is a result of a higher P inflow to the lake system than the outflow
and a negative number is due to higher outflow transport of P than inflow. The Roxen lake
only consists of one positive value in 2009, and hence all other values are result of the
outflow being larger than the measured inflows. This could be explained as the Roxen lake
acting as a source of P, releasing P which causes the higher outflow than inflow, and this has
not changed over the 1980-2010 time period. It is also notable even if a 10% higher inflow
would still not cause the values to be positive and Roxen would remain a source of P.
27
P transport difference (in-out), ton/yr
30
20
10
0
-10
-20
-30
-40
-50
-60
-70
-80
1980
1982
1984
1986
1988
1990
1992
1994
1996
Year
1998
2000
2002
2004
2006
Figure 8. The difference of the annual transport of P in ton/year in the Roxen lake. The solid line shows the difference
between inflow and outflow transport, and dashed lines shows the difference with an estimation of increased inflow of
10% due to excluded catchment.
5.3.2 Lake Glan
The transport graph for the Glan lake system is presented in Figure 9. The transport of P into
Glan lake is the same as the outflow from Roxen. The annual transport of P into Glan from
Finspångsån appears to be slightly decreasing over the period from previous investigation.
The major decrease seem to be during the end of the 1980s and the beginning of the 1990s,
thereafter it seems like the curve is starting to increase again. As previously stated, Ysundaån
seem to be the least anthropogenic impacted watercourse with smaller amounts of transported
P than the other branches. Ysundaån visually seem to correlate with Finspångsån, as the larger
decrease seems to be in the late 1980s and beginning of 1990s, and in recent years it may
seem to increase again. Both these branches contribute to a very small amount of the total
inflow. The transport of P out from Glan lake was tested in the Mann Kendall test to have no
trend, which may be interesting as the major inflows to Glan lake seem to have slightly
decreased over the period. This could mean that something is happening in the lake
compensating for the decrease of P by adding P from another source. It could be determined
from the deviation plots and the Mann Kendall test, that a negative trend over the period in
several of the plots could mean concentrations has been decreasing over time from Roxen to
Glan.
28
2008
2010
250
Glan out
200
P transport, ton/yr
Glan in
(sum)
150
Motala
Ström in
Finspångså
n in
100
Ysundaån
in
50
0
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Year
Figure 9. Annual transport of P in ton/year in the Glan lake
Contrary to Roxen lake, the difference between Glan lake total inflow and outflow show
many years in which the inflow exceeds the outflow (see Figure 10). Especially in the 1980s,
most values were above zero showing the P being contained in the lake. This could be due to
the sediment acting as a sink for the P. During the 30 year time period, it may seem that the
values is turning closer towards negative values which may describe Glan lake to start
releasing more P in the outflow than in the inflow. This could be a suggestion of the lake
commute to become a source rather than a sink of P over the time period.
29
120
P transport difference (in-out),
ton/yr
100
80
60
40
20
0
-20
-40
-60
-80
1980
1982
1984
1986
1988
1990
1992
1994
1996
Year
1998
2000
2002
2004
2006
2008
Figure 10. . The difference of the annual transport of P in kg/year in the Glan lake. The solid line shows the difference
between inflow and outflow transport, and dashed lines shows the difference with an estimation of increased inflow of
10% due to excluded catchments.
5.3.3 Visualization of the mass balances
Following figure was produced to visually present the mass balances of the Motala Ström
River Basin Roxen and Glan lakes (Figure 11). A flow diagram was constructed for the mass
balances in order to visually present the information conducted from the study, using the
software STAN. The mass balances were calculated as annual mean of every 10 years within
the time period (1981-1990, 1991-2000, and 2001-2010). Added within the figure were except
of the known inflows, also arrows within the boxes representing internal input or output in
regard to sediment absorption/desorption. An arrow going into the box represents a source of
P, and an arrow going out from the box is representing a sink of P.
30
2010
Figure 11. Mass balances of Motala Ström, lakes Roxen and Glan. I generally stand for imported substance and E generally stands for exported substance. Within the lake, sediment working as a source is named E for export
out from sediment, and sediment working as a sink of P is named I for inflow to sediment.
6 Discussion
Glan and Roxen are two shallow lakes, both are connected with several watercourses and both
have difficulties with eutrophication (Naturvårdsverket, 2003a, Tibblin, 2011,
Vattenmyndigheterna, 2009). The two lakes and their connected waterways are heavily
regulated by several power stations, which alter the river basin structure. Pinardi et al. (2011)
described damming as altering hydrological patterns, changing the physical, chemical and
biological properties in a freshwater system, as well as causing alterations in water turnover,
nutrient dynamics, thermal stratification, sedimentation, primary production rates, primary
production communities and food chain structure. Agricultural land has been shown to be one
main contributors of source appointment (Ulén & Kalisky, 2005). Non-point sources are
mainly driven by precipitation and fertilizer applications, according to Wang et al. (2011), and
anthropogenic activities as fertilizer application rates, livestock quantity and population size
are important factors influencing dissolved non-point sources of P to freshwater systems.
Göransson et al. (1991) found that the areas in the Glan/Roxen region with most agricultural
plains were responsible for the largest P transport from diffuse sources, reaching higher
releases than some of the point sources. The contribution of P from single household’s sewage
was found by Ulén & Kalisky (2005) to be quite important in comparison to contributions
from WWTPs.
The P impact on the MSRV may have previously caused absorption of P to the sediment,
according to MSV (2010). The lakes were investigated in 2010 by MSV (2010) and it was
concluded that lake Roxen has had a declined ecological status from 2006. They also found
the P transport to be larger in 2010 than in 2009 from Roxen to Glan, and the transport of P
from Roxen into Glan represented 81% of the total P transport into Glan.
6.1 Lake Roxen
The Roxen lake is a lake which rarely gets stratified, as the shallow depth makes it possible
for the water to be continuously mixed (Naturvårdsverket, 2003a, Tibblin, 2011,
Vattenmyndigheterna, 2009). Lau & Lane (2001) described shallow lakes to tend to have
smaller dilution capacity and provide more circulation than deep-water lakes, making it easier
for a shallow lake to be impacted by eutrophication events. The concentrations data,
deviations data and the transport of P for the stations in the Roxen lake system show a very
strong negative trend of P, both in Roxen inflow and outflow. Both Stångån and Svartån
behave a little different with the major dip in the middle of the 30 year period, and the P
transport has started to increase in later years, however only Svartån shows a trend. The
Roxen lake has previously been affected by large P input from Linköping WWTP
(Naturvårdsverket, 2003a, Vattenmyndigheterna, 2009). Tertiary treatment was introduced at
all major Swedish waste treatment sites in the mid-1970, including Linköping WWTP which
introduced a new treatment step in 1974 which reduced the P emissions to MSRB and could
be one reason for the negative trend over the time period (Grimvall et al, 2000,
Naturvårdsverket, 2003a, Vattenmyndigheterna, 2009). Detergents containing polyphosphates
have also been phased out in recent years and could have caused a decrease in P availability to
lake systems. However, Grimvall et al. (2000) explained MSRB as a lake system were
decreases due to these changes have not been recorded. As advanced waste water treatment
and the banning of P in detergents have generally lowered P loading, fertilized soils have
become increasingly important source of P and could be an explanation of the increase of P
transported into Roxen lake in recent years as agriculture expands (Gächter et al. 1998).
Possible releases of P from the Roxen sediment could depend on several factors as;
distribution and concentration of P in the sediment, degree of saturation of exchangeable P,
intensity of the biological processes and the hydrological patterns of the lake (Xiang & Zhou,
2011). As Rydin et al. (1998) describes the possibility of resuspension to cause high internal P
load especially in shallow lakes as the sediment is more likely to be exposed to different
physical environment, which may change mobilization and release of P. Redox and anaerobic
conditions may also impact the mobility of P in sediments, however in shallow lakes as
Roxen and Glan Xiang & Zhou, (2011) claim sediment releases of P may even at oxic
conditions contribute up to 99% of total P input. Göransson et al. (1991) investigated the P
transports of Roxen and Glan in the time period 1975-1989 and described the adsorbed load
of P in the Roxen sediment to be around 1000 tons from 1950s-1990s. As long as the lake has
decent oxygen supply in its bottom waters, adsorbed P will not be released into the overlying
water column Tibblin, (2011) states. However, as the lake is largely affected by
eutrophication, oxygen concentrations sometimes decrease, releasing P into the water through
resuspension. As of the results of the graphs, Roxen outflow is larger than the Roxen total
inflows. The difference of the annual inflow/outflow transport of Roxen lake (Figure 8) varies
with quite large changes from year to year over the whole time period. The largest jump
visually determined is from 2007-2009 where it spans from –70 tons to +15 tons, changing
the annual transport of ~85 tons over 2 years. The reason for these large changes has not been
possible to evaluate. All possible inflows into the Roxen lake system, as the 15 smaller
branches connecting to the lake as an example, has not been included in this investigation, but
even considering these impacts (as in Figure 8), these inputs are still very small and would not
change the state of the lake. The most reasonable explanation for the higher releases of P in
the outflow is sediment releases. Göransson et al. (1991) found the transport of P out of
Roxen to be higher than the amount of inflows. They claimed other studies also has found
Roxen to be an exporter of P since the Linköping WWTP changed the treatment structure on
their plant, as the reduction in P inflow after the restructure caused an increase of sediment
release. Hence, it could be suggested that Roxen lake has been acting as a source of P over the
whole time period of 1980-2010 as there is nothing in the data of the time period 1980-2010
which suggest Roxen has changed from being a sink to a source. Rather, it could be suspected
that Roxen lake has been a source of P maybe even long before the WWTs changed their
treatment in 1970s. For only one year, in 2009, the inflows of transported P exceeded the
outflows to the Roxen lake. The reason for this has been difficult to determine. Göransson et
al. (1991) further suggested Roxen may continue to leak P for a long time as no decrease of
leakage could be found during the 15 year period of their investigation.
33
6.2 Lake Glan
Lake Glan has previously been considered a natural nutrient treatment plant due to large
adsorption of P in the Glan sediments. During time period of the Göransson et al. (1991)
investigation (1975-89), Glan had a larger inflow than outflow and was also suggested to
adsorb P to the sediments.
Lake Glan may seem to previously have had a stabilising effect, as there are larger scatter in
the inflow concentrations and deviations to Glan than in the outflow of the lake. Changes in
Finspångsån and Ysundaån are very small in comparison to Glan inflow; however, these
waterways also consist of lower concentrations. It could be determined from the Mann
Kendall trend analysis and deviations plots of negative trend over the period in several of the
plots, meaning concentrations has been decreasing over time from Roxen to Glan and from
Ysundaån. Glan outflow sampling station is located close to Norrköping city and this
deviation plot act very different from the other stations with no clear trend to be detected and
dramatic changes may be detected over time. A change in lake system over the period 19802010 can also be visually determined due to the Glan inflow and outflow graphs (Figures
9;10). Especially in the 1980s, most annual transport values showed the P being contained in
the lake, and as suggested by MSV (2010) it could have a previous role of a treatment plant
and a sink of P. Contrary to lake Roxen, lake Glan show a shift in data starting to occur during
the 1990s and several times during the last 15 years, where the outflow values are exceeding
the inflows of the lake (Figure 10). The difference of the annual transport in Glan (Figure 10)
does not seem to correlate with Roxen lake with large variations between years over the
whole time period, however the beginning of the time period shows large changes between
years. As an example, 1981 has a value of -70 tons and 1983 has a value of almost +80 tons.
The difference from these two years is around 150 tons, a very high figure. In the end of the
time period, the variation of annual transport difference between in and outflow slows down,
staying within a limit of -20 ton to +20 ton (or 40 ton difference). The reason for these
variations has not been evaluated. It is not possible to detect a decrease in P concentration in
Glan outflow, but all inflows but one seem to have decreased over the years. This causes
suspicion of something happening within the lake, compensating for the lower inflow by
adding P from another source. Perrone et al. (2008) described lake sediment can lose its
ability to absorb P at certain stages, and a lake which has been operating as a sink of P may be
starting to operate as a source of P. The turning point according to Perrone et al. (2008)
coincides with the evolution of the lake enveloping towards eutrophic conditions, developing
anoxic conditions, and related releases of Fe-bound P. A recently conducted simple mass
balance of P in Glan by MSV (2010) also showed a higher concentration of P in the outflow
water than in the inflow water, and a suspicion was presented that desorption processes could
be taking place within the lake sediment. As the trend of Roxen outflow and the measured
values of Glan inflow downstream of Skärblacka Bruk both showed a negative trend in the
Mann Kendall test, Skärblacka Bruk has not been affecting the change in Glan lake, and the
most proper evaluation is that Glan sediment has changed is state from acting as a sink to
starting to act as a source of P in recent years.
34
6.3 Critical considerations
A quantitative study may be conducted in terms where possible biased results are considered
in order to evaluate the quality of the study performed. According to Bouraoui & Grizzetti
(2011) trends should be investigated over a long period of time in order to separate impacts of
climate change etc. However, Malmaeus & Håkanson, (2004) claim using data for 20-30
years seem more unattractive, as natural trends and changes in ecology may alter the lake
characteristics for such long periods. They claim the guideline is to aim at closest possible
agreement between empirical data and modelled predictions, in terms of annual mean values
and seasonal variation of P. This investigation has taken into consideration all available data
with monthly measurements; hence a 30 year time period was suitable as the data was
available and the results could give some trend indications due to a longer time period.
One thing noted in the construction of the deviations was the differences between Roxen
outflow and Motala Ström inflow to Glan, where the only thing affecting the water in between
the lakes would be Skärblacka Bruk and Skärblacka WWTP (Vattenmyndigheterna, 2009).
Both the values of P transport from Roxen outflow and Motala Ström inflow to Glan showed
a negative trend, but Skärblacka was evaluated not to have caused a large impact on the
transport of P into Glan inflow. Skärblacka WWTP was not causing any large emissions
previously and is no longer an active treatment plant. The difference of Roxen outflow and
Motala Ström inflow to Glan still reached up to 60 tons some years, and the Skärblacka
release figures were not close to these, hence the Glan inflow sampling site was replaced by
Roxen outflow due to the risk of biased results (Göransson et al. 1991, Naturvårdsverket,
2003b).
6.3.1 Mass balance calculations and models
Due to P has caused excessive primary productivity in lake systems, implementations to
decrease P inflows have been on the agenda for several years (Gächter et al. 1998). As P
pathways in lakes are many and independent, diffuse P sources from agricultural land have
caused difficulties in modelling and understanding P circulation (Malmaeus & Håkanson,
2004). According to Malmaeus & Håkanson, (2004) all important internal processes as
sedimentation, burial, mineralization, resuspension, diffusion, bio uptake, bio production,
sediment bio-turbation and mixing must be considered to be able to evaluate P concentrations
within a lake. They further claim the empirical inflow concentrations of P not always being
representative of the P supply of the lake as plenty of P is transported when the accumulated
storage of eroded P is released during spring flood and this peak is most likely to be missed
by monthly or weekly sampling, which are data often used in mass balance calculations. This
could have been an issue in this investigation as well, as the investigation is relying on
monthly concentrations. However as Motala Ström being a large river basin, large flows have
a compensative effect of this peak. Malmaeus & Håkanson (2004) solved this difficulty by
multiplying inflow concentrations by eight during the month of peak discharge, however in
this investigation the measured values were used as they were measured.
In general a mass balance equation states all water or concentration entering into a lake must
either be placed as storage, consumed, or leaving the lake within that certain time period
35
(Singh et al. 2009). In this particular investigation, mass balance has been used as a
calculation measure and not as a modelling measure. Most mass balance models for P has
according to Håkanson et al. (2003), been miscalculating the abiotic fluxes of P, and some has
been compensating for the mistakes by making more mistakes when handling the retention of
P in the lake to get a proper correspondence between empirical data and modelled values. A
simple relationship between water turnover, sediment and P was identified by Vollenweider,
who according to Håkanson et al. (2003) claimed water turnover to be the important factor in
regulating nutrient loading on lakes. According to Malmaeus & Håkanson, (2004)
Vollenweider also described an approach to predict P concentrations by using a basic massbalance model and regression analysis. Several other studies have also used this method, and
though it has been successful to predict P concentrations in some lakes, it has failed in others.
As many mass balance models of P seem to have difficulties of reaching results which are of
relevant and proper values, a calculation measure seem to be more efficient in terms of
determining P transport and whether the lakes could be considered sources or sinks.
Malmaeus & Håkanson (2004) believes it is important to quantify the whole system and the
internal processes, but as more compartments are added to a model, the numbers of unknown
parameters will most likely increase, as well as the total uncertainty of the model. This means
it must not necessarily be the best choice in these types of investigations.
6.4 Further studies
The movement of water in rivers and streams play a large role in P transport. P has a
numerous of processes linked to its cycle and affecting its retention, and several of these
processes are just partly understood (Ulén & Kalisky, 2005). Several experimental and
comparative studies have been carried out for whole lake ecosystems, deriving loads for lake
management decisions (Håkanson et al. 2003). Sorption and desorption in lake sediments
have been evaluated by scientific research to regulate P concentration in water bodies,
however P fluxes caused by resuspension is not as widely investigated (Wan et al. 2010).
Impact of hydrodynamic conditions affecting the resuspension is poorly investigated in terms
of source strength and magnitude of the sediment (Wan et al. 2010). Jin et al. (2006) explain
releases of P from sediment as one of the most important reasons for high P concentration in
some lakes. Xiang & Zhou (2011) believes it is of importance to studying various forms of P
present in sediments and the effect these have on the environment. They claim it is hence
necessary to know the contents off different P fractions in the sediments and the forms of P in
the sediment which can help elucidate trends of P release in the lake water, as well as the TotP concentration. Perrone et al. (2008) also states the chemical behaviour of P and its fractions
is more important of investigation instead of Tot-P considering concentration. Further
investigations along the MSRB suggest analysing what happens with the P cycle within the
lakes; the retention process, sediment fractionation, eutrophication, external factors as weather
is affecting resuspension as the lakes are very shallow.
36
7 Conclusion
A P mass balance analysis of the lakes Roxen and Glan over the period of 1980-2010 has
been conducted in this study, also including a quality control of the concentrations data and a
Mann Kendall trend analysis. It was evaluated that 5 out of 8 sampling stations showed
different grades of negative trends, indicating decreasing concentrations of P. The exception
was Glan Outflow, Stångån and Finspångsån, where trends could not be detected.
This investigation has also concluded that Roxen lake has acted as a source of P during the
whole period 1980-2010, except for one year, and the lake is expected to continue to leak for
years ahead. Lake Glan has acted as a source during 22 of the 31 years of investigation and
has a tendency to become more of a source of P over the latter years. But the variation
between years in Glan lake makes it necessary to analyse the data further in order to establish
the reason for the lake behaviour. The lakes were certainly sinks of phosphorus before
construction of wastewater treatment plants, but at least for Roxen, the switch from sink to
source was completed before 1980.
8 Acknowledgements
I would like to thank my supervisor Hans Bertil Wittgren, Sofia Bastviken at the country
administration board in Östergötland and Anders Grimvall from Linköping University for
constructive discussions.
9 References
1. Andersson, K. (2006) Statistisk undersökning av näringsämnen i Svartån. Examensarbete
Miljövetarprogrammet, Linköpings Universitet. Institutionen för Tema. [Electronic]
Available: liu.diva-portal.org/smash/get/diva2:22038/FULLTEXT01 (accessed: 2012-03-02)
2. Asaoka, S, Yamamoto, T. (2011) Phosphorus mass balance in a highly eutrophic semienclosed inlet near a big metropolis: A small inlet can contribute towards particulate organic
matter production Marine Pollution Bulletin. Vol. 63, pp. 237-242.
3. Bennett, E, Carpenter, S.R. (2002) P soup: Humanity's impact on the phosphorus cycle World
Watch. Vol. 15, No. 2, pp. 24-32.
4. Billerud (2007) Billerud Skärblacka AB Verksamhets- och miljöbeskrivning 2007.
[Electronic] Available:
http://www.billerud.com/upload/pdf/Miljo/Billerud_Miljored%20Sk%C3%A4rblacka%2007.
pdf (Accessed: 2012-05-14)
5. Billerud (2011) Miljörapport för verksamhetsåret 2011. Billerud Skärblacka AB.
6. Billerud (2012) Om oss, Historia, Skärblacka. [Electronic] Available:
http://www.billerud.se/sv/About/Historia/Skarblacka/ (Accessed 12-05-14)
7. Bouraoui, F, Grizzetti, (2011) Long term change of nutrient concentrations of rivers
discharging in European seas. Science of the Total Environment. Vol. 409, No. 23, pp. 48994916.
8. Brandt, M, Larsson, M, Wallin, M, Rosberg, J, Pers, C, Wallenberg, P. (2006) Modellering av
fosforflöden med olika modellsystem, SMED rapport, 37 s. In: Rosberg, J, Pers, L, Brandt, M,
Johnsson, H, Larsson, M. (2006) Framtagning av nytt fosforberäkningssätt för
37
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
beräkningssystem för diffus belastning, retention och tillförsel till havet för PLC5
rapporteringen 2007. Svenska MiljöEmissionsData. SMED Rapport Nr 16. [Electronic]
Available: http://www.smed.se/wp-content/uploads/2011/05/SMED_Rapport_2006_16.pdf
(page 109-129. Accessed: 2012-03-02)
Brandt, M, Ejhed, H, Rapp, L. (2009) Nutrient loads to the Swedish marine environment in
2006: Sweden's report for HELCOM's fifth pollution load compilation. Swedish
Environmental Protection Agency. Report 5995. [Electronic] Available:
http://www.naturvardsverket.se/Documents/publikationer/978-91-620-5995-8.pdf (Accessed:
2012-03-02)
Cheng, B.-Y., Liu, T.-C., Shyu, G.-S., Chang, T.-K., Fang, W.-T. (2011) Analysis of trends in
water quality: constructed wetlands in metropolitan Taipei. Water Science and Technology.
Vol. 64, No. 11, pp. 2143-2150.
Dapeng, L, Yong, H, Chengxin, F, Yan, Y. (2011) Contributions of phosphorus on
sedimentary phosphorus bioavailability under sediment resuspension conditions Chemical
Engineering Journal. Vol. 168, No. 3, pp. 1049-1054.
Du, S.T, Shentu, J.L, Luo, B.F, Shamsi, I.H, Lin, X.Y, Zhang, Y.S, Jin, C.W. (2011)
Facilitation of phosphorus adsorption onto sediment by aquatic plant debris. Journal of
Hazardous Materials. Vol. 191, No. 1-3, pp. 212-218.
Finspångs tekniska verk (2009) Miljörapport 2009. Axsäters Reningsverk. Finspångs tekniska
verk AB.
Google Maps (2012) Google Maps, search: “Roxen, Linköping”. [Electronic] Available:
http://maps.google.se/maps?hl=sv&tab=wl (Accessed: 2012-05-18)
Grimvall, A, Stålnacke, P. (1996) Statistical methods for source apportionment of riverine
loads of pollutants. Environmetrics. Vol. 7, No. 2, pp. 201–213.
Grimvall, A, Stålnacke, P, Tonderski, A. (2000) Time scales of nutrient losses from land to
sea — a European perspective. Ecological Engineering. Vol. 14, pp. 363–371.
Grimvall, A, Libiseller, C, Wahlin, K. (Program developers) (2011) MULTITEST - an Excelbased program for univariate and multivariate Mann-Kendall tests for monotone trends.
Version 6.1: Jan 19, 2011. [Electronic] Available:
http://www.ida.liu.se/divisions/stat/research/Software/index.en.shtml (accessed 2012-05-28)
Gächter, R, Ngatiah, J.M, Stamm, C. (1998) Transport of phosphate from soil to surface
waters by preferential flow. Environmental Science and Technology. Vol. 32, No. 13, pp.
1865-1869.
Göransson, B, Brandt, M, Wittgren, H.B. (1991) Markläckage och vattendragstransport av
kväve och fosfor i Roxen/Glan-systemet, Östergötland. Projekt Roxen/Glan. SMHI Hydrologi,
Nr. 34, 1991.
Helsel, D.R, Hirsch, R.M. (2002) Statistical Methods in Water Resources Techniques of
Water Resources Investigations, Book 4, chapter A3. U.S. Geological Survey. [Electronic]
Available: http://pubs.usgs.gov/twri/twri4a3/ (Accessed: 2012-08-28)
Hirono, Y., Watanabe, I., Nonaka, K. (2009) Trends in water quality around an intensive teagrowing area in Shizuoka, Japan. Soil Science and Plant Nutrition. Vol. 55, pp. 783–792.
Hitta (2012) mapinformation. Lantmäteriverket. [Electronic] Available: www.hitta.se
(Acessed 2012-06-05),
Håkanson, L, Ostapenia, A.P, Boulion, V.V. (2003) A mass-balance model for phosphorus in
lakes accounting for biouptake and retention in biota. Freshwater Biology. Vol. 48, No. 5, pp.
928-950.
38
24. Jin, X, Wang, S, Bu, Q, Wu, F. (2006) Laboratory experiments on phosphorous release from
the sediments of 9 lakes in the middle and lower reaches of Yangtze River region, China.
Water, Air, and Soil Pollution. Vol. 176, No. 1-4, pp. 233-251.
25. Kokfelt, U, Reuss, N, Struyf, E, Sonesson, M, Rundgren, M, Skog, G, Rosén, P, Hammarlund,
D. (2010) Wetland development, permafrost history and nutrient cycling inferred from late
Holocene peat and lake sediment records in subarctic Sweden. Journal of Paleolimnology.
Vol. 44, No. 1, pp. 327-342.
26. Lau, S.S.S, Lane, S.N. (2001) Continuity and change in environmental systems The case of
shallow lake ecosystems. Progress in Physical Geography. Vol. 25, No. 2, pp. 178-202.
27. Lindström, G, Pers, C.P, Rosberg, R, Strömqvist, J, Arheimer, B. (2010) Development and
test of the HYPE (Hydrological Predictions for the Environment) model – A water quality
model for different spatial scales. Hydrology Research. Vol. 41, No. 3-4, pp. 295-319.
28. Maguire, R.O, Sims, J.T, Coale, F.J. (2000) Division s-4 - Soil fertility & plant nutrition:
Phosphorus fractionation in biosolids-amended soils: Relationship to soluble and desorbable
phosphorus Soil Science Society of America Journal. Vol. 64, No. 6, pp. 2018-2024.
29. Malmaeus, J.M, Håkanson, L. (2004) Development of a Lake Eutrophication model.
Ecological Modelling. Vol. 171, pp. 35-63.
30. Modarres, R, Ouarda, T.B.M.J, Vanasse, A, Orzanco, M.G, Gosselin, P. (2012) Modeling
seasonal variation of hip fracture in Montreal, Canada. Bone. Vol. 50, (2012) pp. 909–916.
31. MSV (2010) Motala Ström Vattenvårdsförbund Årsrapport. Calluna 2010. Linköping.
[Electronic] Available: http://www.motalastrom.org/resultat/arsrapporter/arsrapport-2010
(Accessed: 2012-03-05)
32. MSV (2012) Motala ström Vattenvårdsförbund. Vattenkemi, kartsök. [Electronic] Available:
http://www.motalastrom.org/resultat/vattenkemi/kartsoek (Accessed: 2012-05-18)
33. Nationalencyklopedin (2012) Glan. [Electronic] Available: http://www.ne.se/lang/glan
(accessed 2012-05-19)
34. Naturvårdsverket (2003a) Åtgärds- och konsekvensanalys för införandet av
miljökvalitetsnormer för fosfor i sjöar. Rapport 5289. Underlagsrapport (1) till
Miljökvalitetsnormer för fosfor i sjöar – redovisning av ett regeringsuppdrag.
Naturvårdsverket (rapport 5288). [Electronic] Available:
http://www.naturvardsverket.se/Documents/publikationer/620-5289-6.pdf (Accessed: 201202-13)
35. Naturvårdsverket (2003b) Åtgärder och kostnader för minskade fosforutsläpp från enskilda
avlopp, industrier m.m. till sjön Glan. Rapport 5290. Underlagsrapport (2) till
Miljökvalitetsnormer för fosfor i sjöar – redovisning av ett regeringsuppdrag (NV rapport
5288). [Electronic] Available: http://www.naturvardsverket.se/Documents/publikationer/6205290-X.pdf (Accessed: 2012-05-13)
36. Naturvårdsverket (2006) Övergödningen av Sveriges kuster och hav - Naturvårdsverkets
ställningstaganden med anledning av en internationell expertutvärdering av kväve/fosforproblematiken i våra omgivande hav. Naturvårdsverket rapport 5587. [Electronic] Available:
http://www.naturvardsverket.se/Documents/publikationer/620-5587-9.pdf (Accessed: 201203-02)
37. Nizzoli, D, Welsh, D.T, Viaroli, P. (2011) Seasonal nitrogen and phosphorus dynamics during
benthic clam and suspended mussel cultivation. Marine Pollution Bulletin. Vol. 62, No. 6, pp.
1276-1287.
38. Norrköping Vatten (2009) Miljörapport 2009 Skärblacka Reningsverk. Norrköping Vatten
AB.
39
39. Perrone, U, Facchinelli, A, Sacchi, E. (2008) Phosphorus dynamics in a small eutrophic
Italian lake. Water, Air, and Soil Pollution. Vol. 189, No. 1-4, pp. 335-351.
40. Pinardi, M, Bartoli, M, Longhi, D, Viaroli, P. (2011) Net autotrophy in a fluvial lake The
relative role of phytoplankton and floating-leaved macrophytes. Aquatic Sciences. Vol. 73,
No. 3, pp. 389-403.
41. Reddy, K.R, Newman, S, Osborne, T.Z, White, J.R, Fitz, H.C. (2011) Phosphorous cycling in
the greater everglades ecosystem Legacy phosphorous implications for management and
restoration. Critical Reviews in Environmental Science and Technology. Vol. 41, pp. 149-186.
42. Rydin, E, Welch, E.B. (1998) Aluminum dose required to inactivate phosphate in lake
sediments Water Research. Vol. 32, No. 10, pp. 2969-2976.
43. Shadmani, M., Marofi, S., Roknian, M. (2012) Trend Analysis in Reference
Evapotranspiration Using Mann-Kendall and Spearman’s Rho Tests in Arid Regions of Iran.
Water Resource Management. Vol. 26, pp. 211–224.
44. Sileika, A.S., Stålnacke, P., Kutra, S., Gaigalis, K., Berankiene, L. (2006) Temporal and
spatial variation of nutrient levels in the Nemunas River (Lithuania and Belarus).
Environmental Monitoring and Assessment. Vol. 122, pp. 335–354.
45. Singh, S, Kumar, B, Thakural, L.N, Galkate, R. (2009) A comprehensive study on water
balance, sedimentation and physico-chemical characteristics of Sagar Lake in India.
Environmental Monitoring and Assessment. Vol. 148, No. 1-4, pp. 265-276.
46. SMHI (2009) Svenskt Vattenarkiv. Sjödjup och Sjövolym. [Electronic] Available:
http://www.smhi.se/k-data/hydrologi/sjoar_vattendrag/sjodjup_SVAR_2009.pdf (accessed
2012-05-18)
47. SMHI (2012a) Vattenweb. [Electronic] Available: http://vattenweb.smhi.se/ ( Accessed 201204-23)
48. SMHI (2012b) Klimatinformation för huvudavrinningsområden. [Electronic] Available:
http://www.smhi.se/klimatdata/klimatscenarier/klimatanalyser/Klimatinformationhuvudavrinningsomraden#f=1;p=;i=start;kl=t_;ti=ar_; (Accessed 2012-06-05)
49. Sollie, S, Janse, J.H, Mooij, W.M, Coops, H, Verhoeven, J.T.A. (2008) The contribution of
marsh zones to water quality in Dutch shallow lakes. A modeling study. Environmental
Management. Vol. 42, No. 6, pp. 1002-1016.
50. Stålnacke, P, Grimvall, A, Sundblad, K, Tonderski, A. (1999) Estimation of riverine loads of
nitrogen and phosphorus to the Baltic Sea, 1970–1993. Environmental Monitoring and
Assessment. Vol. 58, No. 2, pp. 173-200.
51. Tekniska Verken (2012) Miljörapport 2011, Nykvarnsverket i Linköping. Tekniska Verken i
Linköping AB (publ). Division Vatten. 2012‐03‐23
52. Tibblin, P. (2011) Fiskevårdsplan Roxen 2011. Länsstyrelsen I Östergötland. Länsstyrelsen
rapport: 2011:17. [Electronic] Available:
http://www.lansstyrelsen.se/ostergotland/SiteCollectionDocuments/Sv/djur-ochnatur/fiske/FISKEVARDSPLAN-ROXEN-webb.pdf (Accessed: 2012-03-05)
53. Ulén, B.M, Kalisky, T. (2005) Water erosion and phosphorus problems in an agricultural
catchment - Need for natural research for implementation of the EU Water Framework
Directive. Environmental Science and Policy. Vol. 8, No. 5, pp. 477-484.
54. Vattenmyndigheterna (2009) Förslag till åtgärdsprogram för Södra Östersjöns vattendistrikt
– Underlagsmaterial Remiss av samrådsmaterial, 1 mars - 1 september 2009. [Electronic]
Available: http://www.vattenmyndigheterna.se/SiteCollectionDocuments/sv/sodraostersjon/publikationer/forslag-ap-omrade-motalastrom.pdf (Accessed: 2012-02-13)
55. Wahlin, K, Grimvall, A. (2010) Roadmap for assessing regional trends in groundwater
quality. Environmental Monitoring and Assessment. Vol. 165, No. 1-4, pp. 217-231.
40
56. Wan, J, Wang, Z, Yuan, H. (2010) Characteristics of phosphorus fractionated from the
sediment resuspension in abrupt expansion flow experiments Journal of Environmental
Sciences. Vol. 22, No. 10, pp. 1519-1526.
57. Wang, X, Hao, F, Cheng, H, Yang, S, Zhang, X, Bu, Q. (2011) Estimating non-point source
pollutant loads for the large-scale basin of the Yangtze River in China. Environmental Earth
Sciences, Vol. 63, No. 5, pp. 1079-1092.
58. Xiang, S.L, Zhou, W.B. (2011) Phosphorus forms and distribution in the sediments of Poyang
Lake, China. International Journal of Sediment Research. Vol. 26, No. 2, pp. 230-238.
59. Zhong, J.C, You, B.S, Fan, C.X, Li, B, Zhang, L, Ding, S.M. (2008) Influence of Sediment
Dredging on Chemical Forms and Release of Phosphorus. Pedosphere. Vol. 18, No. 1, pp. 3444.
41
Fly UP