BookPDF Available

ECOMAR: A data-driven framework for ecosystem-based Maritime Spatial Planning in Danish marine waters

Authors:
  • NIVA Denmark Water Research
  • NIVA Denmark Water Research

Abstract and Figures

We report the developments and results of the ECOMAR project, which have taken place 2018-2020 and was funded by THE VELUX FOUNDATIONS. ECOMAR has established state-of-the-art data sets for the distribution of human activities and pressures as well as ecosystem components in Danish marine waters. ECOMAR has mapped the combined effects of multiple human pressures, ranked pressures and analysed the potential effects of changes in pressure intensities and new human activities. ECOMAR has also outlined how zoning could be initiated in Denmark. Further, ECOMAR has modelled scenarios for 2030 and 2050 and concludes that agreed strategies and plans will probably not lead to reductions in human activities and pressures and accordingly unlikely to lead to improvements in environmental status.
Content may be subject to copyright.
REPORT SNO. 7563-2020
ECOMAR:
A data driven framework for ecosystem-based
Maritime Spatial Planning in Danish marine waters
ECOMAR:
A data-driven framework for ecosystem-based
Maritime Spatial Planning in Danish marine waters
REPORT S.NO 7562-2020
© Lars Gejl/Biofoto/Ritzau Scanpix
© Lars Geijl/Biofoto/Ritzau Scanpix
NIVA 7562-2020
NIVA Denmark Water Research
REPORT
Headquarter
NIVA Region East
NIVA Region West
NIVA Denmark
Gaustadalléen 21
Sandvikaveien 59
Thormøhlensgate 53 D
Njalsgade 76, 4th floor
NO-0349 Oslo, Norge
NO-2312 Ottestad, Norge
NO-5006 Bergen Norge
DK-2300 Copenhagen S, Denmark
Phone (47) 22 18 51 00
Phone (47) 22 18 51 00
Phone (47) 22 18 51 00
Phone (45) 39 17 97 33
Internet: www.niva.no
www.niva-danmark.dk
Title
ECOMAR: A data-driven framework for ecosystem-based Maritime Spatial Planning in Danish marine waters.
Results and conclusions from a development and demonstration project
Serial number
7562-2020
Date
11 December 2020
Authors
Jesper H. Andersen, Jørgen Bendtsen, Kathrine J. Hammer, E. Therese Harvey, Steen W. Knudsen & Ciaran J. Murray
NIVA Denmark Water Research
Jacob Carstensen, Ib Krag Petersen, Signe Sveegaard & Jakob Tougaard Department of Bioscience, Aarhus University
Karen Edelvang, Josefine Egekvist, Jeppe Olsen & Morten Vinther National Institute of Aquatic Resources, Technical University of
Denmark (DTU Aqua)
Zyad Al-Hamdani, Jørn Bo Jensen & Jørgen O. Leth Geological Survey of Denmark and Greenland
Berit C. Kaae & Anton Stahl Olafsson Department of Geosciences and Natural Resource Management, University of Copenhagen
Will McClintock, Chad Burt & Dan Yocum SeaSketch.org (NCEAS/UCSB)
Topic
Marine Spatial Planning
Distribution
Open
Geographical area
Denmark
North Sea
Baltic Sea
Pages
81 + annexes
Client(s)
THE VELUX FOUNDATIONS
Clients reference
Mikkel Klougart
Published by NIVA DK
Project number 180048
Summary
We report the developments and results of the ECOMAR project, which have taken place 2018-2020 and was funded by THE VELUX FOUNDATIONS. ECOMAR has established state-of-the-art
data sets for the distribution of human activities and pressures as well as ecosystem components in Danish marine waters. ECOMAR has mapped the combined effects of multiple human
pressures, ranked pressures and analysed the potential effects of changes in pressure intensities and new human activities. ECOMAR has also outlined how zoning could be initiated in
Denmark. Further, ECOMAR has modelled scenarios for 2030 and 2050 and concludes that agreed strategies and plans will probably not lead to reductions in human activities and pressures
and accordingly unlikely to lead to improvements in environmental status.
Fire emneord
Four keywords
1.
Havplanlægning
1.
Marine Spatial Planning (MSP)
2.
Økosystem-baseret forvaltning
2.
Ecosystem-based Management (EBM)
3.
Havplandirektivet (HPD)
3.
Maritime Spatial Planning Directive (MSPD)
4.
Havstrategidirektivet (HSD)
4.
Marine Strategy Framework Directive (MSFD)
This report is quality assured in accordance with NIVA's quality system and approved by:
Jesper H. Andersen
Jørgen Bendtsen
Chief scientist
Office manager
ISBN 978-82-577-7297-0
NIVA Report ISSN 1894-7948
© Norsk institutt for vannforskning/Norwegian Institute for Water Research.
The publication can be cited freely if the source is stated.
1
ECOMAR
A data-driven framework for
ecosystem-based Maritime Spatial Planning
in Danish marine waters
Results and conclusions from
a development and demonstration project
NIVA 7562-2020
2
NIVA 7562-2020
3
Preface
ECOMAR is an abbreviation of ‘Development and testing of a data-driven
framework for ecosystem-based marine spatial planning’.
The ECOMAR project, which has been funded by the VILLUM Foundation for
the period 2018-2020, is as indicated a project focusing on ecosystem-based
maritime spatial planning, especially on the development and demonstra-
tion of state-of-the-art methods informing future decision making.
The following institutions participated in ECOMAR: NIVA Denmark Water Re-
search (lead partner), the Department of Bioscience (BIOS) at Aarhus Uni-
versity, the Department of Geosciences and Natural Resource Management
(IGN) at the University of Copenhagen, DTU Institute for Aquatic Resources
(DTU Aqua) and the Geological Survey of Denmark and Greenland (GEUS).
Further, the International Council for the Exploration of the Seas (ICES) and
the National Centre for Ecological Analysis and Synthesis (NCEAS) have also
contributed to the project.
The ECOMAR project has developed, tested and applied different tools for
ecosystem-based maritime spatial planning and analysed in detail how Dan-
ish marine waters can be used sustainably, taking into account existing legis-
lation, both national and international, in particular the EU Maritime Spatial
Planning Directive and also the EU Marine Strategy Framework Directive,
the EU Water Framework Directive and the Nature 2000 Directives (Habitats
Directive and Birds Directive).
Copenhagen, 11 December 2020
Jesper H. Andersen
Please cite as:
Andersen, J.H., J. Bendtsen, K.J. Hammer, E.T. Harvey, S.W. Knudsen, C.J. Murray, J. Carstensen, I.K. Petersen, J. Tougaard, S. Sveegaard, K. Edelvang,
J. Egekvist, J. Olsen, M. Vinther, Z. Al-Hamdani, J.B. Jensen, J.O. Leth, B.C. Kaae, A.S. Olafsson, W. McClintock, C. Burt & D. Yocum (2020):
ECOMAR: A data-driven framework for ecosystem-based Maritime Spatial Planning in Danish marine waters.
Results and conclusions from a development and demonstration project.
NIVA Denmark Report, 81 pp.
ECOMAR partners:
NIVA 7562-2020
4
Table of contents
1 Introduction .............................................................................................................................................................................................................. 8
1.1 What is Maritime Spatial Planning? ................................................................................................................................................................... 8
1.2 Where are we now? ......................................................................................................................................................................................... 9
1.3 Where should we be going?............................................................................................................................................................................ 10
1.4 ECOMAR objectives ........................................................................................................................................................................................ 11
2 Methods .................................................................................................................................................................................................................. 12
2.1 Study area ..................................................................................................................................................................................................... 12
2.2 Data layers .................................................................................................................................................................................................... 15
2.2.1 Pressures and activities .................................................................................................................................................................... 15
2.2.2 Ecosystem components.................................................................................................................................................................... 17
2.2.3 Societal interests ............................................................................................................................................................................. 19
2.3 Combined effects of human activities .............................................................................................................................................................. 20
2.4 Analysis and scenarios .................................................................................................................................................................................... 22
2.4.1 SeaSketch........................................................................................................................................................................................ 22
2.4.2 Description of analyses and scenarios ............................................................................................................................................... 23
2.4.3 Linking MSPD, MSFD and WFD pressure analyses .............................................................................................................................. 27
2.5 Towards marine zoning establishing building blocks for a future zoning plan .................................................................................................. 28
2.6 Confidence and uncertainty ............................................................................................................................................................................ 29
2.6.1 Sources of uncertainty ..................................................................................................................................................................... 29
2.6.2 Assessing uncertainty in layers made from discrete data .................................................................................................................... 30
2.6.3 Combining kriging and splines for oxygen depletion........................................................................................................................... 31
2.6.4 Uncertainty assessment of aggregated products................................................................................................................................ 32
2.6.5 Initial uncertainty and data coverage assessment .............................................................................................................................. 32
3 Results .................................................................................................................................................................................................................... 33
3.1 Pressure map ................................................................................................................................................................................................. 33
3.2 Ecosystem map .............................................................................................................................................................................................. 33
3.3 Mapping of combined effects ......................................................................................................................................................................... 36
3.4 Ranking of pressure groups............................................................................................................................................................................. 38
NIVA 7562-2020
5
3.5 Analysis and scenarios .................................................................................................................................................................................... 41
3.5.1 Increase/decrease in existing activities and pressures ........................................................................................................................ 41
3.5.2 Placement of activities and pressures ............................................................................................................................................... 47
3.5.3 Introductions of new activities.......................................................................................................................................................... 48
3.6 Combining MSPD, MSFD and WFD pressures analyses into a coherent process .................................................................................................. 50
3.7 Combining analyses and scenarios .................................................................................................................................................................. 52
3.7.1 2030 scenario .................................................................................................................................................................................. 52
3.7.2 2050 scenario .................................................................................................................................................................................. 54
3.7.3 MSFD GES scenario .......................................................................................................................................................................... 54
3.8 Assessing data coverage, uncertainty of data layers and sensitivity of the CEA model ........................................................................................ 56
3.8.1 Uncertainty and sensitivity of the CEA model .................................................................................................................................... 58
3.8.2 Uncertainty of oxygen depletion maps .............................................................................................................................................. 59
3.9 Zoning steps toward a roadmap ................................................................................................................................................................... 64
4 Discussion and conclusions ...................................................................................................................................................................................... 71
4.1 Data layers .................................................................................................................................................................................................... 71
4.2 Mapping of combined effects ......................................................................................................................................................................... 71
4.3 Confidence and uncertainty ............................................................................................................................................................................ 72
4.4 Analysis and scenarios .................................................................................................................................................................................... 73
4.5 From data to zoning and beyond ..................................................................................................................................................................... 74
4.6 How to use ECOMAR’s data, tools and results .................................................................................................................................................. 74
4.7 Synthesis and outlook .................................................................................................................................................................................... 75
5 Acknowledgements ................................................................................................................................................................................................. 77
6 Author contributions ............................................................................................................................................................................................... 77
7 References............................................................................................................................................................................................................... 78
NIVA 7562-2020
6
Summary
Key objectives of the ECOMAR project, which took place between spring
2018 and autumn 2020 and was funded by THE VELUX FOUNDATIONS, were:
Compilation of distribution maps covering all relevant pressures and hu-
man activities and ecosystem components in Danish marine waters.
Estimation of the potential combined effects of multiple human pres-
sures, nationally and regionally, and ranking of pressures and also to ana-
lyse the potential effects of future changes in pressure intensities.
Mapping existing human activities, to identify potential conflicts between
spatially overlapping activities and to outline how zoning could be done.
ECOMAR has gathered state-of-the-art data sets regarding spatial distribu-
tions of human activities, pressures and ecosystem components in Danish
marine waters. A wide range of relevant pressures (n = 42) are included as
well as a broad range of ecosystem components (n = 56) covering pelagic
habitats, benthic habitats, fish, seabirds and marine mammals. ECOMAR has
applied existing tools such as SeaSketch and EcoImpactMapper, but also de-
veloped specific codes for postprocessing of results.
Further, ECOMAR has demonstrated how data and tools can be used, both
as analytical tools and as decision support tools in relation to not only the
Maritime Spatial Planning Directive, but also the Marine Strategy Frame-
work Directive as well as the the Water Framework Directive .
With this synthesis report, key ECOMAR results imply that: i) the combined
stress on marine ecosystems and their species, habitats and communities
will probably increase toward 2030 and 2050, and ii) there is no evidence
suggesting that the Danish implementation of the Maritime Spatial Planning
Directive will support the implementation of the MSFD and WFD and thus
lead to improved environmental conditions in Danish waters.
Reviewing the methods used for assessing the confidence of the individual
data layers in ECOMAR has revealed that uncertainty is quantified very dif-
ferently, and in many cases does not take all relevant uncertainty compo-
nents into account. This may affect the ECOMAR results. Therefore, ECO-
MAR has outlined and demonstrated a new general methodology for as-
sessing the uncertainty in data layers derived from discrete data points, ir-
regularly distributed in time and space.
Implementation of ecosystem-based Maritime Spatial Planning in Denmark
faces two obstacles. Firstly, it starts almost from scratch following decades-
long focus on pollution by nutrients and hazardous substances. Secondly, it
must include not only all sectors and pressures, including land-sea interac-
tions, but it must take into account the fact that the environmental condi-
tions are impaired leaving limited or no room for sustainable Blue Growth.
In fact, increasing pressures in the marine environment can potentially com-
promise environmental status goals both for the Danish open waters (the
MSFD domain) and coastal waters (the WFD domain).
ECOMAR has provided high-quality data sets, methods and cross-cutting
analyses in support of evidence-based implementation of the Maritime Spa-
tial Planning Directive. Further, ECOMAR demonstrates that methods and
tools are at hand and ready for use. Thus, the ECOMAR partnership hopes
that this work will be useful and encourages relevant authorities and other
potential end-users to make use of data sets, tools and results in the context
of evidence-based management or in projects following up on ECOMAR.
NIVA 7562-2020
7
Sammenfatning
ØKOMAR-projektet blev udført fra foråret 2018 frem til efteråret 2020 og
blev finansieret af VELUX Fonden. Projektets hovedformål var at i) kortlægge
alle relevante presfaktorer og menneskelige aktiviteter samt en række øko-
systemkomponenter i de danske farvande, ii) estimere potentielle akkumu-
lerede effekter (på engelsk ”Cumulative Effect Assessement” (CEA)) fra for-
skellige menneskelige presfaktorer, for derefter at rangordne disse og
undersøge effekter af mulige fremtidige ændringer i presfaktorernes inten-
sitet og iii) kortlægge eksisterende menneskelige aktiviteter med henblik på
at identificere potentielle konflikter mellem aktiviteter og for at undersøge
hvordan en såkaldt ’zonering kan udføres.
ØKOMAR har samlet state-of-the-art datalag af den rumlige fordeling af
menneskelige aktiviteter, presfaktorer og økosystemkomponenter i danske
havområder. En bred vifte af relevante presfaktorer (n = 42) er inkluderet
såvel som økosystemkomponenter (n = 56), der dækker pelagiske habitater,
bentiske habitater, fisk, havfugle og havpattedyr. ØKOMAR har anvendt
eksisterende værktøjer såsom SeaSketch og EcoImpactMapper, men har
også udviklet specifikke koder til efterbehandling af resultater.
ØKOMAR har desuden demonstreret, hvordan data og værktøjer kan
bruges, både som analytiske værktøjer og som beslutningsstøtteværktøjer,
ikke kun i forhold til Havplandirektivet (MSPD), men også Havstrategidirek-
tivet (MSFD) og Vandrammedirektivet (WFD). Således peges der med denne
synteserapport på, at det samlede pres på havets ressourcer, herunder
natur- og miljøforholdene, sandsynligvis vil stige mod 2030 og 2050, og at
der ikke er noget, der tyder på, at den danske implementering af
Havplandirektivet vil støtte implementeringen af Havstrategidirektivet og
Vandrammedirektivet og dermed føre til bedre miljø- og naturforhold.
Analyse af metoderne til bestemmelse af usikkerhed for de enkelte datalag
har vist, at der anvendes vidt forskellige metoder, og i mange tilfælde med-
tages ikke alle relevante usikkerhedskomponenter. Dette kan påvirke resul-
taterne i ØKOMAR. Der er derfor udviklet en ny metodik for bestemmelse af
usikkerheden i datalag som er estimeret på basis af diskrete datapunkter,
ofte indsamlet uensartet i tid og rum.
Implementering af økosystembaseret maarin arealforvaltning, ofiicielt be-
nævnt ’maritim fysisk planlægning’, i Danmark står over for to forhindringer.
For det første starter implementaringen næsten fra bunden efter årtiers
fokus på forurening fra næringsstoffer og farlige stoffer. For det andet skal
den ikke kun omfatte alle sektorer og presfaktorer, herunder interaktioner
mellem land og hav, men det skal tage højde for, at miljøforholdene svæk-
kes, hvilket efterlader begrænset eller potentielt ingen plads til bæredygtig
blå vækst. Faktisk kan øget pres havmiljøet føre til manglende opfyldelse
af miljøstatusmålene både for de åbne farvande (under Havstrategi-
direktivet) og kystområderne (under Vandrammedirektivet).
Med ØKOMAR er der nu etableret data, metoder og resultater som nu og
fremadrettet kan understøtte en evidens-baseret gennemførelse af EU’s
Havplandirektiv. Det er således ØKOMAR-partnernes håb at data, metoder
og eksempler i fremtiden vil blive inddraget i forskellige processer, først og
fremmet i arbejde for at gennemføre evidens-baseret forvaltning (MSPD,
MSFD, WFD), men også i andre relevante forsknings- og udviklingsprojekter.
Titel: ØKOMAR: en data-dreven metodik for økosystem-baseret havplanlægning i de danske farvande.
År: 2020.
Forfatter(e): J.H. Andersen, J. Bendtsen, K. Hammer, E. T. Harvey, S.W. Knudsen, C. Murray, J. Carstensen, I.K. Petersen, J. Tougaard, S. Sveegaard, K. Edelvang, J.
Egekvist, J. Olsen, M. Vinther, Z. Al-Hamdani, J.B. Jensen, J.O. Leth, B.C. Kaae, A.S. Olafsson, W. McClintock, C. Burt & D. Yocum.
Udgiver: Norsk Institut for Vandforskning (NIVA), ISBN 978-82-577-7297-0
NIVA 7562-2020
8
1 Introduction
Denmark is in the process of implementing Maritime Spatial Planning in Danish law and to develop the first national Maritime Spatial Plan.
This process is embedded in the 2014 EU Maritime Spatial Planning Directive, which follows from the 2002 Council Recommendation on
Integrated Coastal Zone Management. For decades, national strategies for the protection of Danish marine waters have focused on inputs of
pollutants (nutrients and contaminants) and to a lesser degree on how space is utilized and the effects caused by multiple human activities.
1.1 What is Maritime Spatial Planning?
Maritime Spatial Planning refers to a process established in the EU Integra-
ted Maritime Policy (Anon. 2007) and regulated by the EU Maritime Spatial
Planning Directive (MSPD), adopted 23 July 2014 (Anon. 2014).
MSP aims to reduce potential conflicts between sectors and activities com-
peting for marine space. At the same time, MSP also intends to protect the
marine environment. Further, MSP seeks to encourage investment by creat-
ing a level playing field between sectors and interests. MSP is by the UN In-
tergovernmental Oceanographic Commission (2020) defined as:
MSP is a public process of analysing and allocating the spatial and tem-
poral distribution of human activities in marine areas to achieve ecologi-
cal, economic, and social objectives that have usually been specified
through a political process. Characteristics of marine spatial planning in-
clude ecosystem-based, area-based, integrated, adaptive, strategic and
participatory.
MSP is not an end in itself, but a practical way to create and establish a
more rational use of marine space and to manage the interactions be-
tween its uses, to balance demands for development with the need to
protect the environment, and to deliver social and economic outcomes in
an open and planned way.
Ecosystem-based MSP works across borders and sectors. Therefore, land-
sea interactions should also be considered because human activities in up-
stream catchments may have significant impacts on environmental condi-
tion in downstream coastal and marine waters. For MSP implementation to
be ecosystem-based, the planning process should include all ecologically rel-
evant features and the human activities and pressures impacting these.
In Denmark, the MSPD is implemented in ‘Lov om maritim fysisk planlægn-
ing’ (Anon. 2016). In accordance with the MSPD, this law will establish a ba-
sis for national coordination of sea use and support sustainable Blue Growth
in Danish marine areas. The law shall contribute to sustainable development
of the off-shore energy sector (oil and gas as well as offshore wind farms),
shipping, transport infrastructure, fisheries and aquaculture, extraction of
marine raw materials, and environmental protection and improvements in-
cluding the resilience towards climate change.
However, other sectors such as tourism and recreation are mentioned as op-
tional in the marine spatial plan while underwater cultural heritage is not
mentioned. Areas will be designated for specific uses, e.g. for offshore ener-
gy production (oil and gas as well as well as offshore wind farms), shipping,
fisheries, aquaculture, deep sea mining and environmental protection to-
wards 2030. This upcoming national MSP plan is sometimes referred to as
‘Havplan Danmark’ and will enter into force in 2021. Havplan Danmark will,
according to the Danish Maritime Authority (2020), establish not only pre-
dictable circumstances for maritime activities but also predictable provisions
regarding the use of sea space. In a Danish context, implementtation of spa-
tial planning at sea has to start virtually from scratch (see section 1.2). Thus,
there is a need for compilation of spatial data sets and development and
testing of MSP tools.
NIVA 7562-2020
9
1.2 Where are we now?
In a Danish context, MSP is a new activity. This may sound unexpected, but
this current state-of-play is justified to some extent and can be explained.
For decades, the key environmental problem in Danish marine waters, espe-
cially in coastal waters and the Danish Straits, has been nutrient inputs and
eutrophication, i.e. the effects of elevated nutrient concentration stimulat-
ing excessive growth of algae in surface water, leading to reduced transpar-
ency of waters and subsequent loss of submerged aquatic vegetation, low
oxygen concentrations in bottom water due to sedimentation and minerali-
sation of organic matter from surface water causing fish kills due to the low
concentrations or even absence of oxygen (Ærtebjerg et al. 2003).
The problem of eutrophication has received considerable attention and
other problematic pressures have been addressed to a limited extent or not
at all. Discharges and emission of contaminants, dumping of dredged mate-
rial and physical modification have been addressed nationally sector by sec-
tor (e.g. shipping, industries) or case by case (i.e. the Great Belt fixed link
and the Øresund fixed link). Fishing activities have been dealt with through
the EU Common Fisheries Policy. For an overview of national environmental
policies, please confer with Miljø- og Energiministeriet (1999).
The problems with excessive algal blooms, oxygen depletion and occasional
fish kills have been overwhelming and therefore at the top of the political
agenda, leaving other issues with less attention. However, with the adoption
of the Marine Strategy Framework Directive (MSFD) in 2008, the focus has
been widened and evidence has emerged that pressures other than nutrient
inputs also impact Danish marine waters (Miljøministeriet 2012, Petersen et
al. 2018, Andersen et al. 2019, Miljø- og Fødevareministeriet 2019).
The concept of the human activities and pressures in Danish marine water
has evolved from a state with a single dominant pressure (nutrient inputs)
and a relatively low number of pressures of concern to today’s situation
with a few dominating pressures (nutrient inputs, fishing, contaminants, and
climate change) and more than a dozen of other ecologically relevant pres-
sures. Further, the MSFD has put focus on new emerging threats such as in-
troduction of non-indigenous species and inputs of marine litter as well as
noise leading to increased awareness on the political agenda.
Climate change is an important pressure affecting marine ecosystems, but it
is not included in the MSFD, which may seem peculiar as the MSFD is an-
chored in an ecosystem-based approach for managing human activities.
However, climate change is addressed in many other ways, nationally and
internationally (i.e., IPPC and EU). The reasons for not including climate
change in the MSFD, are somewhat obscure and this deficiency contradicts
the definition of the ecosystem-based approach and how it is supposed to
be implemented.
In Denmark, land-sea interactions are the main pressures in coastal waters
(i.e. nutrient inputs from upstream catchments). Accordingly, it is important
to include this key pressure to ensure a genuine ecosystem-based Danish
implementation of the MSPD.
NIVA 7562-2020
10
1.3 Where should we be going?
Evidence-based decision support is a prerequisite for ecosystem-based MSP
and ultimately for attaining sustainable Blue Growth and clean, healthy, and
productive seas with a good environmental status (EEA 2019a).
With the adoption of the MSPD in 2014 and the subsequent enactment in
Danish law in 2016, Denmark is committed to developing and adopting a na-
tional maritime plan for the Danish EEZ by 2021. Despite the current uncer-
tainties regarding the probable focus and context of ‘Havplan Danmark’, cur-
rently being developed by the Danish Maritime Authority, it could seem
from those parts of the plan that have been presented to key stakeholders
in the autumn 2020, that the plan most likely will focus on the minimum re-
quirements of the MSPD and to a lesser degree on a full range of features vi-
tal for ecosystem-based management.
Marine ecosystem-based management (EBM) has according to Long et al.
(2015) been defined as:
Ecosystem-based management is an interdisciplinary approach that bal-
ances ecological, social and governance principles at appropriate tem-
poral and spatial scales in a distinct geographical area to achieve sustain-
able resource use. Scientific knowledge and effective monitoring are used
to acknowledge the connections, integrity and biodiversity within an eco-
system along with its dynamic nature and associated uncertainties. EBM
recognizes coupled social-ecological systems with stakeholders involved
in an integrated and adaptive management process where decisions re-
flect societal choice.
It follows from the definition that prerequisites for genuinely ecosystem-
based management of human activities will have to include, as a minimum,
the following features:
Detailed information on the occurrence of all ecologically relevant eco-
system components ranging from planktonic organisms over benthic
communities to fish, seabirds and marine mammals.
Data on the occurrence of all human pressures, either as distribution
maps of the intensity or presence/absence of specific pressures.
Information on the effect distances, i.e. how far from the point a specific
pressure takes place to where it can be expected to have an impact.
Information on the sensitivity of a specific ecosystem component to a
specific pressure.
With the above information at hand, it is possible to map, analyse and plan a
broad range of human activities as well as their interactions within Danish
marine waters and thereby support the decision-making process and ulti-
mately ecosystem-based MSP.
Reflecting the many facets and interests covered by the MSPD process,
ECOMAR has been an interdisciplinary project covering a broad knowledge-
base, including biology, geology, chemistry as well as social sciences. Fur-
ther, ECOMAR has not only focused on establishing state-of-the-art data
sets but has also demonstrated the use of data and relevant tools. These
data and tools, as well as key synthesis results, are presented in this synthe-
sis report. We hope the report can enable a data-driven ecosystem-based
MSP and that relevant end users, for example national agencies and NGO’s,
will utilise the ECOMAR outputs in the same manner as they have been
demonstrated in this report.
ECOMAR alone will not improve environmental status of Danish marine wa-
ters, but it can support ecosystem-based MSP, if three key conditions are
met:
Firstly, a paradigm shift is required it should be recognized that MSP is
a data-driven process where the best available data is turned into
knowledge, which again is fed into decision-making processes.
Secondly, it should be acknowledged that the MSPD and MSFD are
closely connected and that MSP should not be used to increase pres-
sures, but to decrease or restrict pressures.
Thirdly, all relevant sectors should be included.
NIVA 7562-2020
11
1.4 ECOMAR objectives
The overarching purpose of the ECOMAR project has been to develop, test
and demonstrate the use of ecologically relevant data sets and tools for
data-driven and ecosystem-based maritime spatial planning in Danish ma-
rine waters.
In order to achieve this, we defined several concrete objectives to be
achieved. A key objective of ECOMAR has been to gather and organize the
best possible data sets describing human pressures and ecosystem compo-
nents within the Danish EEZ. A second, but no less important, objective has
been to map human activities and pressures and analyse both pressures and
their potential combined effects on a broad range of ecosystem compo-
nents, and the relative importance of individual groups of pressures on dif-
ferent scales (nationally, regionally and locally).
Further, it has been an objective to demonstrate the concept of zoning in
Danish marine waters through the following process:
Establish a baseline for activities, pressures and combined effects.
Identify potential conflicts between activities within the same areas.
Group and categorize existing and planned activities in different catego-
ries, ranging from multi-purpose zone (zone level 1) where permits are
not mandatory to a restricted access zone (zone level 4).
An additional objective of ECOMAR has been to carry out a broad range of
scenario analyses aiming to showcase potential consequences of increasing
or decreasing intensity of human activities and pressures. ECOMAR has also
analysed ways to combine the demands of the MSPD, MSFD and WFD re-
garding pressure analyses into a streamlined and cost-effective process.
ECOMAR also aims to communicate the results of the development and
demonstrations to a wider audience. For this reason, a list of abbreviation
and acronyms used is shown in Table 1.
Table 1: Abbreviations and acronyms used in this report.
BITS
Baltic International trawling Survey
CEA
Combined effect assessment (see CIA)
CIA
Cumulative impact assessment (see CEA)
DEA
Danish Energy Agency (Energistyrelsen)
DMA
Danish Maritime Authority (Søfartsstyrelsen)
EA
Ecosystem Approach
EBA
Ecosystem-based Approach
EBM
Ecosystem-based Management
EEA
European Environment Agency
EEZ
Exclusive Economic Zone
EMODnet
European Monitoring and Observation Data Network
EU
European Union
GES
Good Ecological/Environmental Status
HD
Habitats Directive
HELCOM
Helsinki Commission. See www.HELCOM.fi
IPCC
Intergovernmental Panel on Climate Change
MPA
Marine Protected Area
IBTS
International Baltic Trawling Survey
MFVM
Miljø- og Fødevareministeriet (Ministry of Environment and
Food of Denmark)
MSFD
Marine Strategy Framework Directive
MSP
Maritime/Marine Spatial Planning
MSPD
Maritime Spatial Planning Directive
MST
Miljøstyrelsen (Danish Environmental Protection Agency)
MSY
Maximum Sustainable Yield
N2000
Natura 2000
NIS
Non-indigenous species
NOVANA
National Monitoring Programme for Water and Nature
OSPAR
OSPAR Commission. See www.ospar.org
POPs
Persistent organic pollutants
SAR
Swept Area Ratio
SwAM
Swedish Agency for Water and Marine Management
VME
Vulnerable Marine Ecosystem
VMS
Vessel Monitoring System
WFD
Water Framework Directive
NIVA 7562-2020
12
2 Methods
This chapter describes the study area of the Danish Exclusive Economic Zone (EEZ) and introduces the data sets used. The Danish marine
waters are often divided into three sub-regions: i) the North Sea/Skagerrak area, ii) the Kattegat, and iii) parts of the western Baltic Sea
including the Danish Straits. Most parts of the Danish EEZ are well studied and monitored, especially regarding eutrophication, contaminants
and biodiversity. Further, the methods, concepts and models for analyses, mapping and uncertainty, developed or applied in the context of
the ECOMAR project are introduced.
2.1 Study area
The Danish Exclusive Economic Zone (EEZ) covers eastern parts of the North
Sea, southern parts of the Skagerrak, western and central parts of the Katte-
gat, the Little Belt, the Great Belt, the western parts of the Sound, parts of
the western Baltic Sea and also the marine waters around Bornholm and the
Ertholmene archipelago.
Together with the Swedish parts of the Kattegat and Sound, the Danish EEZ
forms the transition zone between the North Sea and Baltic Sea (see Fig. 1).
The environmental conditions of the Danish marine waters have been consi-
stently monitored since the mid-70’s, mostly via the national monitoring
program, see Svendsen et al. (2005) and Andersen et al. (2017) for details.
Most coastal waters in Denmark are classified as eutrophication problem ar-
eas (EEA 2019b), meaning that nutrient levels are elevated and primary ef-
fects, such as high concentrations of chlorophyll, or secondary effects, such
as loss of submerged aquatic vegetation or benthic invertebrates, are com-
mon. Offshore parts of the North Sea and Skagerrak are classified as non-
problem areas indicating that inputs of nutrients, i.e. nitrogen and phospho-
rus, are not an issue. Over the past three decades, eutrophication status has
improved, both in coastal waters (Riemann et al. 2016), but especially in the
open parts of the North Sea (Andersen et al. 2016 and OSPAR 2017).
Classification of eutrophication status in Danish marine waters has been
done as part of the Danish MSFD implementation (Naturstyrelsen 2012) and
for the OSPAR Common Procedure (Andersen et al. 2016) as well as part of
the HELCOM second holistic assessment (HELCOM 2018). An up-to-date
overview can be found in EEA (2019b), where ‘problem areas’ and ‘non-
problem areas’ have been mapped (see Fig. 2A - please note that although
large parts of the Danish waters are classified as problem areas, the off-
shore part of the North Sea and Skagerrak are classified as ‘non-problem ar-
eas’). For more information on nutrient inputs and eutrophication in Danish
marine waters, please see Ærtebjerg et al. (2003), Riemann et al. (2016),
HELCOM (2018), OSPAR (2017), Miljø- og Fødevareministeriet (2019) and
EEA (2019b).
Biodiversity in Danish waters is threatened and there is no evidence for ar-
eas with a good biodiversity status. The key pressures causing an impaired
biodiversity status are nutrient inputs, especially in coastal waters, and fish-
ing activities, mostly in offshore waters of the North Sea, Skagerrak and Kat-
tegat. Biodiversity status of Danish marine water was assessed as part of the
Danish MSFD implementation (Miljøstyrelsen 2012) and in recent HELCOM
and OSPAR assessments (HELCOM 2018, OSPAR 2017). An up-to-date as-
sessment of biodiversity status can be found in EEA (2019c) and Fig. 2B pro-
vides an overview for the Danish EEZ. For more information on status of the
biodiversity in Danish marine waters, please see Andersen et al. (2015),
OSPAR (2017b), HELCOM (2018), MFVM (2019) and EEA (2019c).
NIVA 7562-2020
13
Figure 1: Map of the Danish Exclusive Economic Zone (EEZ). The total area of the Danish EEZ is 105,000 km2, where internal marine waters make up 3,500
km2, the Territorial Zone (12 nautical miles) 40,000 km2 and the remainder 61,500 km2. The Danish EEZ can be divided into three sub-regions: i) the Danish
parts of the North Sea and Skagerrak, ii) the Danish parts of the Kattegat, and iii) the Danish parts of the western Baltic Sea including the marine waters
around Bornholm.
NIVA 7562-2020
14
Contamination by hazardous substances is also a challenge in Danish waters,
and most parts are classified as ‘problem areas’ , i.e. with an impaired envi-
ronmental status. Substances of concern are many, especially some heavy
metals, e.g., mercury (Hg), cadmium (Ca), copper(Cu) and lead (Pb), selected
groups of persistent organic pollutants (POP’s), e.g., polybrominated diphe-
nyl ethers (PBDEs) from flame-retardant chemicals, pest controls, leaded
petrol, PCBs and polycyclic aromatic hydrocarbons (PAHs) from burning
wood and fossil fuel.
Assessment of contamination status has also been carried out in connection
with the Danish MSFD implementation (Milstyrelsen 2012) and in recent
HELCOM and OSPAR assessments (see HELCOM 2018 and OSPAR 2017).
Recently, the EEA carried out an integrated assessment and mapped ‘prob-
lem areas’ and ‘non-problem areas’ with respect to contaminants and Fig.
2C shows the results for the Danish EEZ. For more information on contami-
nants in Danish marine waters, please see HELCOM (2010a), OSPAR (2017b),
Miljø- og Fødevareministeriet (2019) and EEA (2019d).
In summary, environmental conditions in most parts of the Danish marine
waters are impaired compared to the objectives and goals for the MSFD,
WFD and HD (e.g. MFVM 2019). For Danish marine waters, ‘ecosystem
health’ was assessed as part of Naturstyrelsen (2012), which concluded that
a good status (= GES) could only be found in the western-most parts of the
Danish sector of the North Sea, whilst the status in all other parts of the
North Sea, Skagerrak, Kattegat and the western Baltic Sea was not good
(sub-GES).
The latest EEA assessment of ecosystem health in Europes seas is based on
an unprecedented data set including a broad set of indicators. For the Dan-
ish marine waters, no evidence could be found for identification and subse-
quent classification of so-called ‘non-problem areas’ having a good status
(EEA 2019a). This can also be concluded from Fig. 2, showing that when all
assessments are overlaid using a ‘one out, all out’ principle, not a single one
of the combined assessment units is in a high or good status.
Panel A:
Eutrophication
status
Panel B:
Biodiversity
status
Panel C:
Contamination
status
Figure 2: Results of mapping of ‘eutrophication status (panel A), biodiver-
sity status (panel B), and contamination status (panel C) in Danish marine
waters, extracted from Europe-wide classifications of ‘problem areas’ and
non-problem areas’ using established assessment tools (i.e. HEAT+, BEAT+
and CHASE+; see EEA 2019b, 2019c and 2019d for details).
NIVA 7562-2020
15
2.2 Data layers
A key objective of ECOMAR has been to establish state-of-the-art data sets
representing i) human pressures and activities, ii) ecosystem components
and iii) a selection of societal interests, including recreational interests and
marine protected areas. Given the goal of MSP, to allocate space for human
activities, it is imperative to have high quality data sets for a variety of hu-
man activities and pressures. Hence, considerable effort has been spent by
the ECOMAR partnership to establish the best possible data sets for all eco-
logically relevant human activities and pressures. Similarly, with ECOMAR
aiming to develop and test tools and concepts for ecosystem-based MSP, it
is also imperative to gather the best possible data sets for the ecosystem
components.
2.2.1 Pressures and activities
The starting point for establishing the nation-wide data set on pressures and
activities is the EU MSFD (Anon 2008), especially Annex III, Table 1 ‘Pressu-
res and impacts’ which focuses on 8 themes and 19 individual pressures. In
ECOMAR, focus has been broadened in comparison with the MSFD and pre-
vious studies (e.g. HELCOM 2010b, Korpinen et al. 2012, Andersen & Stock
2013, Riemann et al. 2019, Andersen et al. 2020). Hence, a total of 13
themes, 42 groups and 98 distinct pressures and activities (see overview in
Table 2) are included in ECOMAR:
1. Aquaculture: Data on the location of marine aquaculture plants are
taken from The Danish Food Administration Agency (2020) and applied.
2. Climate change: Two existing data sets on sea surface temperature
anomalies and sea level increase have been adapted and applied.
3. Industry, energy and infrastructure: Georeferencing of the variety of ac-
tivities is based on multiple sources cf. the Supplementary Material, An-
nex A.
4. Marine litter: Data is collected under IBTS and BITS fish surveys, where
several stations are trawled in a standardized procedure and where
each ICES rectangle is swept representatively. In addition to the fish
caught, all litter is collected.
5. Noise and cooling water: Noise is included as pulse-block days (from
ICES impulsive noise register, 2016-2018) and continuous noise (ship
noise levels exceeding ambient noise).
6. Non-indigenous species: An existing index developed for MSFD report-
ing (Andersen et al. 2020) has been used for ECOMAR purposes.
7. Physical disturbance of the sea floor: Data set on this pressure group is
based on multiple sources cf. the Supplementary Material, Annex A.
8. Pollution, contaminants: A recent assessment of contaminants in Eu-
rope’s seas (EEA 2019d) has been rescaled and used as a proxy of inputs
of contaminants.
9. Pollution, nutrients: Concentrations of nutrients, i.e. nitrogen and phos-
phorus, are used as proxies for nutrient inputs.
10. Selective extraction of species: commercial fishing: International land-
ings in tons by the gear groups pelagic trawl, mobile bottom contacting
gears for industrial purposes, mobile bottom contacting gears for hu-
man consumption, longlines and set gillnets as a yearly average based
on the period 2015-2017. In ECOMAR, it is used as a pressure layer. VMS
is mandatory on vessels longer than 12 meters, so for those vessels, all
fishing activity is represented in the data layers.
11. Selective extraction of species: recreational fishing and bird hunting:
Two data layers based on an extract from two nation-wide recreation
and tourism surveys (Kaae et al. 2018), and activities including three
types of hunting and nine types of fishing, as well as mapping of tour
boats in Øresund (Angantyr & Holm-Hansen 2017).
12. Shipping and transportation: Data on the spatial distribution and inten-
sity of shipping intensities is taken from EMODnet (2020).
13. Recreation and tourism: Data come from two nation-wide surveys of 92
coastal and marine recreation activities grouped in 16 main types (Kaae
et al. 2018). Use frequencies are added to the approx. 16,000 mapped
recreation sites and data combined with AIS data for recreational boat-
ing. Data represents the annual participation by Danish citizens including
domestic tourists, but not international tourists.
Description of all individual data sets and their origin can be found in the
Supplementary Material (Annex A) to this report.
NIVA 7562-2020
16
Table 2: Overview of pressures and activities and the groups of pressures. A so-called Pressure map illustrating the spatial distribution of pressures can be
found in section 3.1. For detailed information about individual data layers, please confer with the Supplementary Material (Annex A).
Pressures and activities
1. Aquaculture
8. Pollution - contaminants
Fish farms
Contaminants
Shellfish farms
Dumped chemical munitions
2. Climate change
Oil spills
Sea surface temperature anomalies
9. Pollution - nutrients
Sea level rise trend
Nitrogen winter concentrations (DIN)
3. Industry, energy and infrastructurs
Phosphorus winter concentrations (DIP)
Coastal habitat modification
10. Selective extraction of species: Commercial fishing effort by gear group
Bridges and coastal constructions
Set gillnet
Dredging
Longlines
Disposal sites for construction, garbage and dredged material
Mobile bottom contacting gears, for industrial purposes (small mesh sizes)
Offshore oil and gas installations
Mobile bottom contacting gears, for human consumption (large mesh sizes)
Oil and gas pipelines
Pelagic trawl
Wind farms
Mussel dredging
Sea cables
11. Selective extraction of species: Recreational fishing and hunting
Lighthouses
Recreational fishing
Military areas
Bird hunting
4. Marine litter
12. Shipping and transportation
Marine litter
Shipping
5. Noise and energy
Industrial ports
Continuous noise (ship sound 125 Hz)
Harbours
Impulsive noise
13. Recreational activities
Energy production
Recreational boating
6. Non-indigenous species
Non-motorised water craft
Non-indigenous species
Coastal recreation sites
7. Physical disturbance to the sea floor
Scuba-diving recreational
Swept area ratio (SAR) from bottom trawling: Surface SAR
Swept area ratio (SAR) from bottom trawling: Sub-surface SAR
Extraction of material from the seafloor
NIVA 7562-2020
17
2.2.2 Ecosystem components
Inclusion of all ecologically relevant groups of organisms is a prerequisite for
ecosystem-based MSP. Hence, ECOMAR focuses on a substantial range of
organisms ranging from primary producers, i.e. from phytoplankton over a
broad range of benthic habitats to fish, seabirds and top predators such as
seals and harbour porpoise.
The data includes a total of five groups and 56 individual ecosystem compo-
nents cf. Table 3:
1. Pelagic habitats: This ecosystem component group consist of two data
layers: Chlorophyll a concentrations in surface water and oxygen deple-
tion, expressed as areas with low oxygen concentrations in bottom wa-
ters (two thresholds). The data layer on phytoplankton is produced
based on data from ODA and ICES, while the data layer on areas deple-
ted of oxygen is based on the official reporting from the Danish Centre
for Environment and Energy (DCE).
2. Benthic habitats: This group of ecosystem components includes two
types of data layers, one for broad-scale benthic habitats and one for
the distribution of Eelgrass (Zostera marina). Information and maps on
the distribution of broad-scale benthic habitats originates from EMOD-
net covering Europe’s seas. In ECOMAR, we have combined the original
maps in eight groups, each constituting a separate map. The map of the
potential distribution of eelgrass in Danish coastal waters is based on
Stæhr et al. (2019).
3. Fish: Data from two sources show the distribution of fish species: one
dataset shows the yearly average Catch Per Unit Effort (CPUE) per spe-
cies for commercial MSFD species for the period 2015-2017. It is based
on VMS data, which is only available for vessels longer than 12 meters,
so species caught by smaller vessels cannot be presented using this
method (e.g. eel, blue mussels and cockles). Another dataset shows
Catch Per Unit Effort (CPUE) or presence derived from scientific trawl
surveys for the period 2009-2018, used as a proxy for abundance of
commercial MSFD species. This dataset shows the CPUE (number caught
per trawl haul, standardized with respect to haul duration, year, time of
the year and gear used) or presence (probability of catching at least one
individual in a standardized trawl haul). Spatial abundance indices are
derived from analysis of the data from the International scientific trawl
surveys, IBTS, BITS, BTS available from ICES and data from the Danish
Cod and Sole surveys.
4. Seabirds: The following bird abundance data layers are included in our
work: Razorbill/Guillemot (Alk/Lomvie: Alca torda/Uria aalge), Red-
throated Diver/Black-throated Diver (Rødstrubet Lom/Sortstrubet Lom:
Gavia stellate/Gavia arctica), Common Eider (Edderfugl: Somateria
mollissima), Long-tailed duck (Havlit: Clangula hyemalis), Red-breasted
Merganser (Toppet skallesluger: Mergus serrator), and Common Scoter
(Sortand: Melanitta nigra).
5. Marine mammals: There are data layers for three species: harbour por-
poise (Marsvin: Phocoena phocoena), grey seal (Gråsæl: Halichoerus
grypus), and harbour seal (Spættet sæl: Phoca vitulina) in the western
part of the study area. However, in the North Sea and Skagerrak spatial
distribution layers only exists for harbour porpoises.
Description of all individual data sets and their origin can be found in the
Supplementary Material (Annex B) to this report.
NIVA 7562-2020
18
Table 3: Overview of ecosystem component data and the groups. A so-called Ecosystem Map illustrating the spatial distribution of the amount of ecosystem
components can be found in section 3.2. For detailed information about individual data layers, please see the Supplementary Material (Annex B).
Ecosystem component
1. Pelagic habitats
3.2 Commercial fish species
Productive surface waters- chlorophyll a
Pelagic fish species
Oxygen depletion
Herring, Clupea harengus
2. Benthic habitats
Mackerel, Scomber scombrus
Broad scale benthic habitats: Infralittoral sand and muddy sand
Norway pout, Trisopterus esmarki
Broad scale benthic habitats: Infralittoral mud
Saithe, Pollachius virens
Broad scale benthic habitats: Infralittoral coarse sediments
Sprat, Sprattus sprattus
Broad scale benthic habitats: Infralittoral rocks and biogenic reefs
Demercial/benthic fish species
Broad scale benthic habitats: Infralittoral mixed sediments
Plaice, Pleuronectes platessa
Broad scale benthic habitats: Circalittoral sand and muddy sand
Sole, Solea solea
Broad scale benthic habitats: Circalittoral mud
Cod, Gadus morhua
Broad scale benthic habitats: Circalittoral coarse sediments
Haddock, Melanogrammus aeglefinus
Broad scale benthic habitats: Circalittoral rocks and biogenic reefs
Hake, Merluccius merluccius
Broad scale benthic habitats: Circalittoral mixed sediments
Sandeel, Ammodytes spp.
Broad scale benthic habitats: Upper bathyal sediments
Turbot, Psetta maxima
Eelgrass potential distribution, Zostera marina
Crustaceans
Stone reefs within ˈNatura2000ˈ areas
Shrimp, Crangon crangon
3.1 Sensitive fish species
Norwegian lobster, Nephrops norvegicus
Cartilaginous fish species
Pandalus, Pandalus borealis
School Shark, Galeorhinus galeus
4. Seabirds
Skates, Dipturus spp.
Auks, Alcidae (Razorbill/Guillemot)
Smooth-hound sharks, Mustelus spp.
Common scoter, Melanitta nigra
Spotted Ray, Raja montagui
Eider, Somateria mollissima
Starry ray, Amblyraja radiata
Fulmar, Fulmar spp.
Thornback Ray, Raja clavate
Red-breasted Merganser, Mergus serrator
Bony fish species
Red-throated/Black-throated diver, Gavia spp.
Atlantic wolffish, Anarhichas lupus
Long-tailed Duck, Clangula hyemalis
Atlantic Halibut, Hippoglossus hippoglossus
5. Marine mammals
Greater forkbeard, Phycis blennoides
Grey Seal, Halichoerus grypus
Ling, Molva molva
Harbour Seal, Phoca vitulina
Monkfish, Lophius piscatorius
Harbour Porpoise, Phocoena phocoena
Rabbit fish, Chimaera monstrosa
NIVA 7562-2020
19
2.2.3 Societal interests
A number of societal interests are also impacted by marine activities and are
included in the project as equivalent to ecosystem components. This ena-
bles a variety of novel analyses, e.g.:
To analyse which pressures (such as trawling, dredging, or construction)
may potentially disturb or destroy ancient shipwrecks and/or archaeolog-
ical sites or findings.
To estimate how coastal and marine recreation and tourism areas may
be impacted by factors such as low water quality, noise, and aesthetic
disturbances from large constructions.
To highly to which extent marine protected areas may be impacted by
pressures from marine activities.
Data layers treated as ecosystem compentents are:
1. Archaeological sites, findings and ancient shipwrecks: These were
downloaded from the Danish Agency for Culture and Palaces (Slots- og
Kulturstyrelsen 2020) and comprise information about the location, the
type of the artefact, the name of the location, the archaeological date as
well as the approximate age manifested as a range of years it can fall
into. The dataset was first filtered for age for all findings and ancient
wrecks, the findings and ancient shipwercks which are from the Stone
Age (ca. 12,800 ca. 3.500 BC), the Bronze Age (ca. 3500 ca. 500 BC),
the Iron Age (ca. 500 BC ca. 800 AD), the Viking Age (ca. 800 AD 1047
AD), and the Middle Age (1047 AD 1536 AD) were taken into considera-
tion.
2. Shipwrecks: Data was obtained from the Danish Agency for Culture and
Palaces (Slots- og Kulturstyrelsen 2020). Data includes the actual ship-
wrecks of historical importance including parts of the wreck and/or the
anchor. The metadata contains information about the location coordi-
nates, the name of the location, the type of the artefact, the archaeologi-
cal date as well as the approximate range of years it can fall into.
3. Recreational interests: National coastal and marine recreation data was
collected by the University of Copenhagen (Kaae at al. 2018) through two
studies: A crowdsource-based study using an online public participation
GIS (PPGIS) mapping tool allowing respondents to map places of marine
recreation and identify key facts about their activity and the site. Sec-
ondly, a national representative survey of the Danish adult population
with 10,291 valid responses and combined with mapping. These studies
provide new in-depth knowledge of 92 water-oriented recreation activi-
ties grouped in 16 main types as well as nation-wide spatial mapping. In
total approx. 16,000 recreation sites were mapped, and the two studies
supplement each other. Recreation and tourism are interlinked, as partic-
ipation in activities outside the local municipality is regarded as day-visits
by one-day tourists. Danish residents undertaking recreation activities
while staying overnight outside their municipality are classified as do-
mestic tourists (25 % of the data). Results show that marine recreation is
very widespread, and 77.6 percent of the adult population has partici-
pated in water-oriented recreation within the past year. Triangulation of
data with AIS data on recreational boating has shown to provide a solid
documentation and mapping at national level.
4. Marine Protected Areas (MPAs): Data comprises four types of marine
protected areas designated under i) the Habitats Directive (Anon. 1992),
ii) the Birds Directive (Anon. 2009), and iii) the Ramsar Conventions
(2020) plus iv) national offshore biodiversity protection areas under the
MSFD (Naturstyrelsen 2016).
NIVA 7562-2020
20
2.3 Combined effects of human activities
Applying the Ecosystem Approach or an Ecosystem-based Approach to man-
agement of human activities and pressures in Danish marine waters implies
that all ecologically relevant activities and pressures should be addressed. A
truly ecosystem-based MSP, or an MSP plan claiming to be ecosystem-
based, must accordingly address not only all relevant pressures but also
their potential combined effects. Otherwise, resulting policies and manage-
ment strategies cannot claim to be ecosystem-based.
Potential combined effects are estimated using a concept developed by
Halpern et al. (2008) and subsequently applied in northern Europe by Kor-
pinen et al. (2012) and Andersen & Stock (2013). To estimate the potential
combined effects, four types of information are required cf. Fig. 3: i) data on
the spatial distribution of pressures, ii) data on spatial distribution of all rele-
vant ecosystem components, iii) information on the effect distances, i.e.
how far from a point source the effect of a pressure can be expected, and iv)
information on how sensitive or susceptible a specific ecosystem compo-
nent is to impact from a specific pressure.
Using the EcoImpactMapper software developed by Stock (2016) in combi-
nation with the data sets compiled by ECOMAR, including effects distances
and sensitivity scores, we have not only mapped the potential combined ef-
fects of multiple human pressures in Danish marine waters but also ranked
pressures and carried out a wide range of downstream analyses on different
scales ranging from national to local (see section 2.6).
Figure 3: Conceptual illustration of the steps in mapping of combined effects of multiple human pressures. Step 1a: Scaling and log-transformation of individ-
ual pressure data layers including addition of effect distances for point data; Step 1b: Scaling and log-transformation of individual ecosystem component
data layers; Step 2: Calculation of mean impact; Step 3: Mapping and subsequent post-processing of results. Data sets, effects distances and sensitivity
scores are described in detail in the Supplementary Material (Annex A-C) to this report. Based on Andersen et al. (2019).
NIVA 7562-2020
21
The methods developed by Halpern et al. (2008) and further by Halpern et
al. (2009 and 2015) and Stock (2016) for an additive human impact index
have been used to calulate the combined effects of human activities as a
pressures index. Stock (2016) has developed the software EcoImpactMap-
per’, an open-source Java program that implements the Halpern model.
EcoImpactMapper has, although the method is rather simple (Halpern & Fu-
jita 2013), been used used widley in marine ecosystems for assessment of
cumulative human impacts (Korpinen & Andersen 2016) and now also in the
ECOMAR project.
EcoImpactMapper and the Halpern model require three kinds of input data:
Di, pressures and activities i, represented by its spatial distribution in a
regular grid; for example, fishing intensity with a given gear type.
ej, ecosystem component j, represented by its spatial distribution in a
regular grid; for example, different kinds of broad-scale habitats, eel-
grass or fish species, either as presence-absence or continuous data.
µi,j, sensitivity scores, a numerical representation of the sensitivity of
ecosystem component j to pressure or activity i, based on surveys.
The intensities of the pressures and activities and the ecosystem component
data were all normalized by log(x + 1)-transformation and rescaling to maxi-
mum 1. The reason for this was to enable comparisions between different
units for the layers, e.g. presence/absence, probabilities of presence, popu-
lation densities or concentrations (see Supplementary Material, Annex A
and B) for details of the data units for each layer).
For pressures or activities with a point distribution or decay from a restric-
ted area, effect distances were also estimated based on expert surveys, and
those data layers were pre-processed by adding this effect. A simple linear
decay function from the source and to the limit of the effect distance was
used.
Based on these data, we calculated the dimensionless combined effects of
multiple human pressures index for each cell in the regular grid (x,y) esti-
mated for n pressures/activities and m ecosystem components; IMean:
where
In ECOMAR, the combined effects are estimated as the mean of the impact
over all present ecosystem components, rather than the sum, because some
ecosystem component datasets did not cover the whole study area. This
mean model is also more applied in more recent publications, e.g. Halpern
et al. (2015) and Andersen et al. (2020). The mean model is used to avoid
conflating the effects of high-intensity pressures/activities with the number
of ecosystem components in a given grid cell. Besides the combined effects
indices, the contribution of each of the stressors to the total effect was also
calculated.
Sensitivity scores linking specific pressures with specific ecosystem compo-
nents as well as effect distances for specific point sources/activities have
been set through surveys with relevant experts. See the Supplementary Ma-
terial (Annex C1) for detailed information about the sensitivity scores and
Annex C2 for the effect distances used.
Mapping of combined effects of multiple human pressures and ranking of
pressures are reported in section 3.3 and 3.4, respectively.
 
μ


  

NIVA 7562-2020
22
2.4 Analysis and scenarios
An important objective of ECOMAR has been to analyse the potential effects
and consequences of changes in human activities and the management of
these. The analyses have focused on increases or decreases in pressure in-
tensities, as well as on the introduction of new activities, and have been
demonstrated on relevant scales, i.e. national, regional or local scale. When-
ever possible, the assumed changes are anchored in existing predictions and
scenarios. In some cases the levels of change have been set together with
the experts within the ECOMAR partnership. Nevertheless, the scenarios
should be looked upon as hypothetical and demonstrational.
2.4.1 SeaSketch
The analyses and scenarios are done using SeaSketch, a widely used MSP
tool developed by the National Centre for Ecological Analyses and Synthesis
(NCEAS) at University of California Santa Barbara (UCSB), USA.
SeaSketch is an online participatory mapping tool designed to allow plan-
ners and stakeholders to interact with data related to MSP processes (please
confer with https://www.seasketch.org/home.html).
The ECOMAR SeaSketch project (http://ecomar.seasketch.org) allows users
to visualize > 100 interactive maps depicting the distribution of ecosystem
components (e.g., habitats and species), human activities and pressures
(e.g., pollution, extraction, physical disturbance, climate change, etc.) and
other reference layers. Within the ECOMAR SeaSketch project, these layers
have been modified according to set up analyses and scenarios (described in
section 2.4.2 and 2.4.3), by icreasing or decreasing the pressures or by
sketching new activities (in SeaSketch called ‘sketches’) within the Danish
EEZ. The existing pressures within the new sketches was either modified or
new pressures were added to simulate future management scenarios.
SeaSketch provides analytical feedback about your sketched zone within
seconds. These reports identified if and where the combined human effects
had increased or decreased and which pressures that was contributing most
as well as which ecosystem components that were most affected. These
types of analyses can be used to e.g. inform the development of broadly
supported marine spatial plans.
The analyses and scenarios were analysed individually or as collections,
simulating a comprehensive marine spatial plan with a variety of sketches
including different human activities. The combined effects of all human ac-
tivities within the sketch was compared to the baseline state, pointing to the
differences between the current state and the scenarios. One might, for ex-
ample, calculate whether increasing shipping activities while removing oil
and gas installations within zones ultimately increases or decreases cumula-
tive effects.
The ECOMAR partners could share the designs (including any modified
stressors) within map-based discussion forums including commentary on the
justifications for their designs. Others with access to the forums could then
copy the designs, modify boundaries and stressors, and contribute their own
design concepts. In this way, users may collaboratively develop scenarios
with a quantitative understanding of how they impact ocean ecosystems.
NIVA 7562-2020
23
2.4.2 Description of analyses and scenarios
Analyses were done for the 13 pressure groups (see section 2.2.1) and spe-
cific new activities were included. Predicted changes were combined for two
future scenarios, one for 2030 and one for 2050. In addition, a scenario an-
chored in the ecosystem components and an improved conservation regime
in accordance with the MSFD will be undertaken (MSFD GES scenario). Most
of the scenarios are directly linked to the Blue Growth strategy from the Eu-
ropean Commission or National Danish plans and strategies, aiming to sup-
port a sustainable growth in the marine and maritime sectors. Others are
more general, based on the current intensities of pressures.
Pressure group 1: Aquaculture
Marine aquaculture for fish meat production in net cages is a small industry
in Denmark. However, this industry gives rise to environmental concerns
and has consequently been on the political agenda for decades, as it re-
leases excess nutrients and organic matter to the ecosystem and thus in-
creasing the problems with eutrophication. According to the national aqua-
culture strategy production is expected to increase, mostly for land-based
systems. However, increases in marine shellfish and kelp farms are also ex-
pected, which does not add but instead remove nutrients from the water
column. Several estuaries have been identified as suitable for potential new
shellfish farms to be established: Roskilde Fjord, Gamborg Fjord, Limfjorden,
Mariager Fjord, Vejle Fjord, Kolding Fjord, Åbenrå Fjord, Augustenborg Fjord
and Flensborg Fjord (Altinget 2020). Marine fish farms can expect a decrease
in nutrient losses due to improved technologies and environmental con-
cerns (Miljøstyrelsen 2020). Therefore, the 2030 scenario includes a de-
crease for marine aquaculture by 10% relative to present levels, but an in-
crease in the shellfish farm area by 5% placed within the estuaries men-
tioned. In 2050, the increase in shellfish farm areas is 10%. Kelp farms are
not considered.
Pressure group 2: Climate change
Sea Surface Temperature (SST) and marine heat waves are foreseen to rise
in the future if CO2 emissions are not lowered significantly (IPCC 2019). Fu-
ture scenarios of SST in the North Sea (Schrum et al. 2016) and the Baltic Sea
(Meier 2015) suggest an increase of about 2°C by the end of century, with
some variation between the various scenarios. The rate of SST anomalies is
therefore assumed to be 0.2°C per decade, which will lead to anomalies of
about 0.33-0.37°C in 2030 and 0.73-0.77°C in 2050. These temperature ano-
malies correspond to increases of 54% in the North Sea/Skagerrak, 59% in
the Kattegat and 60% in the Western Baltic in 2030, relative to baseline. In
2050, the SST anomalies will have increased by 0.6°C, which corresponds to
an increase of 78% increase in the North Sea/Skagerrak, 81% in the Kattegat
and 82% in the Western Baltic. The global rate of sea level rise is 3.3 mm per
year (EEA 2019e, IPCC 2019) which means 3.3 cm over a decade and thus 9.9
cm increase in 2050. The change in the sea level rise trend over 10 years
(2030) is assumed to be at least the same as the global rate, which then
gives an increase compared the current rate of 47% in the North Sea/Skag-
errak, 86% in the Kattegat and 55% in the Western Baltic. In 2050, the rate is
expected to be even steeper and we used the current global rate of 3.3
mm/year times 1.5, resulting in a rate of 4.65 mm/year. For 2050 the rate is
thus expected to increase compared to the current rate by 64% in the North
Sea/Skagerrak, 91% in the Kattegat and 70% in the Western Baltic.
Pressure group 3: Industry, energy and infrastructure
The long term forecast up to 2025 is that Denmark will be a net exporter of
energy until 2035 (DEA 2018). According to this report the production will
be about the same in 2030 as per today (2018), due to new findings and
technical advancements, except for a temporary drop in 2020/21 due to re-
building of a specific oil field. The wind power plants pressure is expected to
increase by new areas being taken into use. The wind power parks that al-
ready are approved will be assumed to be implemented by 2030. In 2050, all
areas that are “in pipeline” will be implemented (DEA 2018). An increase in
wind power plants will lead to an extension of the sea cables and accord-
ingly, dredging within the new areas will increase and disposal sites will in-
crease by 5% each year, which also will be implemented. An increase in ship-
ping will also lead to an increase in coastal protections and piers with 10% in
2030 and 20% in 2050 and dredging and disposal of material will both in-
crease by 5% in 2030 and 10% in 2050.
NIVA 7562-2020
24
Pressure group 4: Marine litter
Despite huge media attention and expression of political will to develop Eu-
ropean and national plastic management strategies, there is currently no ev-
idence supporting a decrease in the amount of plastic entering Europe’s
seas and Danish marine waters. On the contrary, we have assumed future
trends to be increasing with 10% by 2030 and 20% in 2050.
Pressure group 5: Noise and cooling water
If shipping intensity increases up to 2030, the continuous noise from ship-
ping will also increase. Noise is measured in a log10 scale (dB) and a direct
calculation of the increased noise is not possible. The 20% increase in ship-
ping is therefore estimated to give an increase by 2.3% on average for the
noise index used (based on an increase of the median level of 0.8 dB). As-
suming new technologies, we estimate that by 2050 the sound pollution will
decrease again and be of the same magnitude as per today. The average im-
pulsive noise of impulse-block days per year will increase, especially in the
areas where new wind power plants will be constructed, thus an increase by
20% in the new areas of wind power parks will be used for both 2030 and
2050. As no changes in inputs of cooling water to coastal waters is expected,
no modelling focusing on this specific input has been undertaken.
Pressure group 6: Non-indigenous species
There is no evidence for any reduction in the number in new introductions
of non-indigenous species to Danish marine waters (Stæhr et al. 2016). In
European Seas an average of 28 new species per year were recorded be-
tween 2006-2011 (EEA 2019f). This number is hard to directly translate to
the NIS index used in the model. A reasonable and likely increase of the in-
dex by 25% over a decade and 50% for 2050, will be applied.
Pressure group 7: Physical disturbance of the sea floor
Due to the expected growth in the building industry and the increased need
for sand for coastal protection, the production of marine resources is ex-
pected to increase in the existing resource areas as well as in new developed
areas (NIRAS, 2018). The pressure intensity is in 2030 expected to increase
by 20% relative to current levels. In 2050, the areas currently designated as
approved, but not active, are included in the pressure layer.
Pressure group 8: Pollution - Contaminants
Oils spills are mainly from shipping and oil platforms. The risk of oil spills is
proportional to shipping so their intensity is assumed to increase corre-
spondingly with 20% in 2030 and 40% in 2050. Oil spills from oil platforms
are expected to remain the same (DEA 2018). Contaminant levels will re-
main the same in 2030 and be reduced by 5% in 2050. It should however be
noted, that there is some uncertainty regarding whether the degree of con-
tamination is increasing or decreasing. Discharges of some substances have
been reduced significantly, but at the same time, new substances are being
introduced at a high rate. The data layer represents many present sub-
stances which are degraded slowly in nature. They will therefore tend to
persist, even given a reduction in releases to the sea. No change in the levels
of dumped chemical munitions is forseen.
Pressure group 9: Nutrients
The most important pressure in Danish marine waters, especially estuaries
and coastal waters, is nutrient input from land and atmosphere. Dedicated
efforts have for decades focused on reducing losses from agriculture, dis-
charges from urban wastewater treatment plants and industries with sepa-
rate discharge. Since the late 1980s, significant reductions have been ob-
tained for both nitrogen (appr. 50%) and phosphorus (appr. 90%). However,
inputs to coastal waters have levelled out since 2001/2002 (Andersen et al.
2019) and nitrogen inputs may even have increased in the past decade. Cli-
mate change with increased rain leading to more run-off and changes in ani-
mal production in combination with less area used for agriculture might
have large impacts in the future. We assume a 10% decrease in 2030, ac-
cording to the current political goals, and a slightly more ambitious goal of a
decrease by 20% in 2050.
Pressure group 10: Selective extraction of species - commercial fishing
Assuming that more fish species will be sustainably managed in European
Seas based on the maximum sustainable yield (MSY) concept, it is expected
that intensity of fishing with bottom trawl for human consumption in the
Baltic Sea in 2030 will be 40% lower than at present and pelagic trawl in the
same area 30% lower. Bottom trawling for industrial use remains the same.
NIVA 7562-2020
25
In the North Sea and Kattegat, the fishing pressure by bottom trawl for hu-
man consumption and industry are decreased by 30% each and pelagic trawl
by 20%. In the 2050 scenario, the pressures are reduced with an additional
10%. The reduction of the fishing quotas for cod and sprat in 2020 and with
further reductions expected in the coming years the pressure from bottom
trawling will also decrease; in the Baltic Sea by 30% in 2030 and 40% in
2050; in the North Sea and Kattegat, the reduction in bottom trawling pres-
sure is analyzed with a decrease of 20% in 2030 and 30% in 2050. The pres-
sure from other fishing methods are also reduced by 10% in 2030 and 20%
in 2050. Mussel dredging occurs only in a few areas and is assumed to re-
main unchanged in future scenarios.
Pressure group 11: Selective extraction of species - recreational fishing
In the Baltic Sea, we expect a slight reduction in the fishing of cod allowed,
due to the critical condition of the population. A rise in recreational fishing
in general is expected, so we have projected an increase in the order of 15%
in general and a smaller increase of 10% in the Baltic Sea in 2030. For 2050
we anticipate increases of 25% and 20%, respectively. Bird hunting pressure
will decrease by 10% in the western Baltic Sea due to the condition of some
seabirds in 2030, in 2050 it will decrease by 20% in all regions.
Pressure group 12: Shipping and transportation
Overall shipping is expected to increase by 20% in 2030 and 40% in 2050, in-
cluding all types of ships from container industrial to large international
cruise ships. Denmark’s blue maritime strategy is investing in maritime deve-
lopments to facilitate, among other things, an increase in shipping capacity.
The industrial ports pressure will be increased by an equivalent of 20% and
30%, coastal protections and piers with 10% and 20% and dredging and dis-
posal of material will both increase by 5% and 10%.
Pressure group 13: Recreation and tourism
Future growth with respect to all types of tourism and recreational activities
is anticipated, especially in the coastal zone. For 2030 we have increased the
intensity in this pressure group with 20% for all categories and for 2050, we
have increased the group with 40%.
Addition of new activities
New activities in Danish marine waters, such as new offshore wind farms
and the construction of a bridge between Zealand and Jutland can be intro-
duced and tentatively analysed using SeaSketch:
Offshore wind farms: Based on information from the Danish Energy
Agency (DEA 2020) we can introduce new offshore wind farms, e.g. in the
North Sea and at Kriegers Flak east of the island Møn.
Kattegat Bridge: This proposed link is introduced assuming the route will
consist of two new bridges between Zealand (Røsnæs) and Samsø and
between Samsø and Jutland (Hov) (Ingeniøren 2018).
Scenarios for 2030 and 2050
In addition to the above pressure group specific analyses, we have com-
bined the results and established scenarios for the years 2030 and 2050:
2030 scenario: We have combined what we consider being the most re-
alistic scenarios for the year 2030 (see individual sections above) and re-
run the model and thus estimated how the expected combined effects
most likely will develop compared to the baseline in section 3.1.
2050 scenario: Similarly, we have re-run the model and estimated the
combined effects in year 2050 based on what we considered to be the
most likely scenario for individual pressures or group of pressures.
Interactions between pressure groups and ecosystem components
In addition to the analyses and scenarios described above, we have grouped
pressures and activities and ranked these with respect to the following key
groups of ecosystem components:
Pelagic habitats (Chlorophyll a and oxygen depletion).
Benthic habitats (Broad-scale benthic habitats and eelgrass).
Fish species.
Seabirds.
Marine mammals.
NIVA 7562-2020
26
Further, treating societal interests such as recreation & tourism, ancient
shipwrecks and stone-age dwellings as an analogue to ecosystem compo-
nents, we analyse and rank the human activities and pressures with poten-
tial impacts on these.
MSFD GES scenario
Based on the identification of key pressures affecting the ecosystem compo-
nent groups, a MSFD GES scenario was modelled. Here the environment was
prioritised by hypothetically reducing the current pressure intensities from
human activities with the aim of improving environmental status in accord-
ance with the EU MSFD. The GES scenario should be considered as a theo-
rethical ‘what if’ demonstration, not as a prediction of future conditions.
The MSFD scenario was modelled in two versions: one with climate change
remaining at current levels and one version with the increase as in the 2030.
Pressure group 1: Aquaculture: 20% decrease in current farms, and a de-
crease in renewed permissions. No new aquaculture introduced.
Pressure group 2: Climate change is not included.
Pressure group 3: Industry, energy and infrastructure Wind power re-
main as in 2030. Some oil and gas installations can be demounted, and
the areas can be restored, as well as the pipelines -> reduction 20%. Less
material dumped at sea -> reduction 30%. Dredging decrease by 20% in
accordance with less shipping. Military areas explosions are set to lower
levels and outside the breeding periods - reduction 20%.
Pressure group 4: Marine litter reduction of 20%, some of the litter can
be collected and removed, better information and better control of pol-
luters.
Pressure group 5: Noise and cooling water: Continuous noise is reduced
by 2.3% related to less shipping and technological developments.
Pressure group 6: Non-indigenous species: Better control legislation and
controls of ballast water can be implemented. No change as already in-
troduced species are present.
Pressure group 7: Physical disturbance of the sea floor: Sand dredging is
reduced by 50%. No new areas are changing status to active.
Pressure group 8: Contaminants: A decrease of 10%, oil spills reduced
5%. Areas with dumped chemical munitions can be sanitized and bombs
can be removed, reduction of 5%.
Pressure group 9: Nutrients: A decrease of 10% via changes of agricul-
tural practices and land use.
Pressure group 10: Commercial fishing: Impacts on fish species and im-
pacts on the seafloor caused by bottom trawling decreases 30%, other
fishing methods 20%. Mussel dredging reduced 10%.
Pressure group 11: Recreational fishing: A decrease of 25%, seabird hunt-
ing by 30%.
Pressure group 12: Shipping and transportation: Reduction in ship traffic
by 20% as well as industrial ports by 15 %. Recreational harbours remain
the same.
Pressure group 13: Recreation and tourism: A decrease by 15% due to
new regulations of restricted periods for human presence.
NIVA 7562-2020
27
2.4.3 Linking MSPD, MSFD and WFD pressure analyses
A total of three EU directives require Member States to map and assess hu-
man pressures in marine waters; the MSPD, MSFD and WFD, where the latter
focuses only on coastal and transitional water bodies. Despite some differ-
ences in the objectives and geographical coverage between these directives,
there is both as demand and an overlap in mapping of human pressures.
All three directives require assessment of current pressures. These pres-
sures can be either identical or overlapping between the three directives,
but, importantly, the data requirements usually are the same. However,
most EU Member States, including Denmark, carry out the prescribed as-
sessments as directive-specific activities in a manner that is usually not coor-
dinated with the activities of the other directives. The MSFD and WFD imple-
mentation processes are coordinated by the Ministry of Environment (MIM),
while the MSPD is in the hands of the Danish Maritime Authority under the
Ministry of Industry, Business and Financial Affairs (in Danish: Erhvervsminis-
teriet) and coordination between these two ministries appears to be limited.
Within MFVM, the practical implementation of MSFD and WFD is further di-
vided. The Danish Environmental Protection Agency produces WFD Initial
Assessments including pressure analyses, and the Department of the MFVM
produces MSFD Initial Assessments, including mapping and assessment of
combined/cumulative impacts. Even within the same ministry, coordination
seems limited, cf. the recent Initial Assessments published in 2019 (MSFD;
Miljø- og Fødevareministeriet 2019) and 2020 (WFD; Miljøstyrelsen, 2020).
ECOMAR demonstrates how the task ideally could be coordinated and car-
ried out on the basis of the same data and methods. Our take on this has
been straightforward and based on the data sets for pressures and ecosys-
tem components available (see Table 2 and 3). Analyses have already been
made in the context of the MSPD and MSFD (see section 3.6). For both the
MSFD and WFD analyses, pressures from climate change are not included as
it is considered an exogenic pressure and consequently not included.
In order to target the mapping and analyses for MSFD and WFD specific pur-
poses, we have focused on relevant ecosystem components as outlined in
Table 4 and below:
WFDDIR: Only quality elements and indicators directly used for assess-
ment of GES are considered, i.e. Chlorophyll-a and potential eelgrass dis-
tribution Eelgrass (used here as a proxy for depth limit).
WFDDIR+INDIR: In addition, quality elements and indicators indirectly re-
lated to WFD assessments of GES have also been included, i.e. oxygen
depletion and benthic habitats.
Table 4: Overview of ecosystem components in the analyses for i) MSFD, ii)
WFDDIR and iii) WFDDIR+INDIR.
Ecosystem component
MSFD1
WFDDIR
WFDDIR+INDIR
Pelagic habitats
- Productive surface waters
X
X
X
- Oxygen depletion
X
X
Benthic habitats
- Infralittoral coarse sediments
X
X
- Infralittoral rocks and biogenic reefs
X
X
- Infralittoral mixed sediments
X
X
- Infralittoral mud
X
X
- Infralittoral sand and muddy sand
X
X
- Circalittoral coarse sediments*
X
X
- Circalittoral rocks and biogenic reefs
X
X
- Circalittoral mixed sediments*
X
X
- Circalittoral mud*
X
X
- Circalittoral sand and muddy sand*
X
X
- Upper bathyal sediments
X
- Stone reefs within N2000 areas
X
X
- Eelgrass distribution
X
X
X
Fish2
X
Sea birds2
X
Marine mammals2
X
1 Includes offshore areas. 2 all ecosystem component data sets are considered.* Indicates that
the circalittoral layer is merged with the corresponding deep circalittoral layer.
NIVA 7562-2020
28
2.5 Towards marine zoning establishing building
blocks for a future zoning plan
Marine zoning was originally developed in Australia and has played a key
role in the adaptive management of human activities within the Great Bar-
rier Reef Marine Park (Day et al. 2018) for more than 40 years. It is a con-
cept for resource management through designation of zones for specific hu-
man activities and uses. A key element of the concept is the nesting of hu-
man activities into groups of similar types and subsequently a zoning matrix
for a given area. In our case, the area in question is Danish marine waters.
Benefits of zoning can include: i) licencing of specific activities, which can de-
velop with certainty, ii) reduction of conflicts between users, and iii) protec-
tion of culturally important areas (e.g. archaeological sites) or environmen-
tally important areas (e.g. Natura 2000 areas or Bird Protection Areas). Es-
tablishing a baseline for the development of a draft zoning plan in Danish
waters is straightforward ECOMAR has the information on activities and
pressures and has used this for outling a so-called ‘Zoning matrix’, the per-
haps single most important step in the making of zoning and a national data-
driven ‘Zoning plan’.
ECOMAR has thus developed a tentative zoning matrix opting for four differ-
ent types of zones ranging from i) General Use Zone over i) Targeted Man-
agement Zone, iii) Exclusive Use Zone, and iv) Restricted Access Zone
(based on Ekebom et al. 2008):
Zone 1 ‘General Use’ is the least restrictive of the four zones, and the
largest, and covers all marine activities not covered by the other three
zones all types of human activites can take place, except those specifi-
cally prohibited by law. Activities that require permissioins or licences are
only allowed after permissions and the same general rules within this
zone are applied to all the three stricter zones.
Zone 2 ‘Targeted Management’ considers areas where there is a further
restriction. This can be areas where a permission or a licence has been
granted but it can also be nature protection areas which have certain
regulations. Activities can be allowed as long as they are not in conflict
with the targeted management within the zone.
Zone 3 ‘Exclusive Use’ are zones reserved for single use activites and
most other activities should be prevented. Recreational or research activ-
ities can be allowed if it is allowed in the management plan. This can e.g.
be fish farms, harbours, wind farms or vulnerable marine habitats.
Zone 4 ‘Limited Access’ is subject to rigorous regulations and restrictions
allowing only limited access and specific activities. Entry to this zone
should be prohibited, except in case of emergency or for scientific activi-
ties, e.g. research or environmental monitoring.
An important step towards a draft zoning plan is to find out where there
could be spatial conflicts between activities. A conflict map was made based
on a conflict matrix where the most likely conflicts could potentially occur,
for example commercial fishing vs. harbour or aquaculture areas, or place-
ment of disturbing industrial activities vs. areas used intensively for recrea-
tion or vice versa.
As another important step, we have estimated the sensitivity linkages of all
ecosystem components based on their sensitivity scores and created a pro-
visional vulnerability map showing the most and least vulnerable marine ar-
eas within the Dansish EEZ.
A fully-fledged data- and criteria-driven application of zoning in Danish ma-
rine water including all human activities and land-sea interaction will ensure
space for all human uses, involvement of all interested stakeholders while
taking care to minimise impacts and multiple pressures on the marine envi-
ronment. Hence, zoning is a necessary tool for planning multiple use of sea
areas in a way that balances human activities and sea use with environmen-
tal objectives.
Further, it should be noted, that a zoning plan must have a management
plan, including the specific permissions and regulations etc. for each zone.
But this is a task for Competent Authories, and also very complicated as vi-
sions and goals of at least MSPD, MSFD and WFD would have to be consid-
ered and aligned.
NIVA 7562-2020
29
2.6 Confidence and uncertainty
An essential component of every scientific result is the quantification of its
confidence and uncertainty. Whereas uncertainty expresses the random var-
iability associated with scientific results and has the same unit as the result,
the confidence expresses a probability that the result is within certain
boundaries, i.e. typically the probability that the result is above/below a cer-
tain threshold or the boundaries of the result associated with a fixed confi-
dence level (e.g. 95%). Quantifying the uncertainty of a given result is a pre-
requisite for quantifying the confidence. In practice, if regulatory limits are
available for data layers or aggregates of these, then we can quantify the
probability (confidence) of complying with these limits, provided that the
uncertainty of data layers and their aggregates are known. Alternatively, we
can fix the confidence level (typically to 95%) and then estimate the range of
variability for that confidence level, provided that (again) the uncertainty of
data layers and their aggregates are known.
2.6.1 Sources of uncertainty
Data layers are typically generated from some underlying data, which may
be associated with uncertainty in their acquirement (e.g. measurement
noise), but spatial and temporal variability may also influence the underlying
data and add to the overall uncertainty of the data layer, when discrete ob-
servations are compiled. Data layers are typically generated using: i) full-cov-
erage registrations, ii) simulation models, and iii) statistical models.
Full-coverage registrations cover the entire spatial domain and may or may
not have a temporal component. For instance, the mapping of oil and gas
pipelines in Danish waters can be considered complete and without tem-
poral variability, as stone reefs are static features the location of which is
considered known. Consequently, the uncertainty associated with such a
data layer is zero. In other cases, full-coverage registrations may exist for all
or a subset of years of the assessment period, which in the case of all years
implies no temporal variability (the entire temporal population is known)
and therefore zero uncertainty, and in the latter case implies that interan-
nual variability will contribute with uncertainty to the data layer because the
spatial distribution is not known and has to be estimated for those years
without data.
Simulation models can be used to produce spatial distributions that cover
the entire spatial domain for the entire assessment period. However, this
does not mean that model output can be equated with full-coverage regis-
trations and similarly be associated with zero uncertainty. Instead, model
output is influenced by uncertainties deriving from model input, model pa-
rameters, boundary conditions and inadequacy of the model structure (i.e.
all models are approximations). The uncertainty of the model output can be
quantified if the magnitude of these sources of uncertainty is known, but
this is seldom the case and consequently, the uncertainty of data layers
based on simulation models is generally not quantified.
Statistical models are typically used for data layers, where simulation mod-
els do not exist or are not considered sufficiently precise for describing spa-
tial and temporal variations. Statistical models are typically employed on
discrete observations that are heterogeneously distributed in time and
space and in many cases, the temporal dimension of the data is overlooked
when the objective is to describe spatial variability. For example, a data
layer produced from a single year’s data will only represent that given year
and not the mean spatial distribution for the entire assessment period. In
addition to the temporal variability, predictions from the statistical models
are also associated with uncertainty, typically increasing with distance from
the observations. Statistical models handle random variations differently
and some methods are better at describing uncertainty than others (see be-
low). In this respect, it should be noted that methods such as linear interpo-
lation, inverse distance weighting and nearest neighbour are mathematical
methods and do not incorporate any terms for the random variability.
An important question for correctly addressing the temporal variability in
the compilations of data layers is to delineate the relevant period. First, data
layers should preferably represent the same period before aggregation, as
artefacts may appear from combinations of pressure and ecosystem layers
covering different periods, when temporal variability is large. Second, the
NIVA 7562-2020
30
uncertainty associated with a data layer may be underestimated if based on
temporally limited data without considering the random temporal variation.
However, this temporal random component cannot be truly assessed if the
temporal limits of the assessment period are not defined. Consequently, it is
of outmost importance that the temporal limits are clearly defined and har-
monised across data layers.
ECOMAR has addressed confidence and uncertainty in the data input in two
ways. Firstly, we have assessed uncertainty in individual data layers and
evaluated the overall uncertainty in the aggregated product, (see section
3.8). Secondly, we outline and demonstrate, based on a case study, how a
data-driven methodology could be developed and implemented in the fu-
ture (see section 3.8.2).
Other sources of uncertainty can arise from the inherent model assumptions
for the CEA method (Stock & Micheli 2016, Stock et al. 2018). To assess the
model uncertainty we have applied the Monte Carlo simulations inbedded in
EcoImpactMapper described by Stock et al. (2018) and the Morris analyses
(so called Elementary effects, i.e. sensitivity and uncertainty estimations)
according to Stock & Micheli (2016) (see section 3.8.1).
2.6.2 Assessing uncertainty in layers made from discrete data
Data layers used in planning tools are typically generated from discrete ob-
servations () that have been sampled more or less irregularly in time
(t) and space (s). These discrete observations are integrated over the spatial
domain using different methods such as natural neighbors, inverse distance
weighting, splines and kriging. Common to all integration methods is that we
are interested in predicting the value of the spatio-temporal process
() in each point in time and space from the observations, i.e.:
   
implying that the true spatio-temporal process can be described as:
   
where is a spatio-temporal process describing the uncertainty of
the prediction at the point . The prediction error may comprise both
observation error and model error, i.e. the latter describing the inadequacy
of the spatial interpolation function (). For many of the commonly appli-
ed integration methods has the following properties:
  
  
which means that
is a central estimator (unbiased) with a variance
described by . Variance of the prediction error may exhibit both
spatially and temporally dependent variations. For a treatment of spatial
and spatio-temporal processes, see Cressie (1993) and Diggle (2013).
The two most common methods for integrating point observations into data
layer are ordinary kriging and splines. In short, ordinary kriging assumes a
constant mean field (
 ) and that the variance of the prediction
error depends on the distances to the observations () used for pre-
dicting
. The variance structure is described through a variogram
model, normally based on the assumption that observations are spatially
correlated up to a certain range (). Ordinary kriging is useful when the ob-
servation network is dense (distances less than ), utilizing that observations
are spatially correlated. Essentially, the prediction function ( 
) of ordinary kriging is a weighted average of the observations
(), where the weights are determined based on the variogram
model. However, ordinary kriging is less suitable when the observation grid
is less dense, with internal distances extending beyond the range . In such
cases, the predicted mean field will be constant. Therefore, a trend function
can be used to describe the mean field ( for stationary processes or
 when the mean field changes with time) and this mean field is formu-
lated as a parametric function of different co-variates, including geographic
position, depth, salinity or temperature fields, for which there are data lay-
ers covering the entire spatial domain. In such cases, the method is termed
universal kriging. However, universal kriging assumes that there is a para-
metric relationship between observations () and the explanatory var-
iables. Nevertheless, both ordinary kriging and universal kriging build on a
NIVA 7562-2020
31
spatial correlation structure, which can be useful for prediction purposes,
but these methods may also overestimate the prediction uncertainty if the
correlation structure, formulated as a random model, is governed by under-
lying mechanisms that are better described with a fixed model. Hence, alt-
hough ordinary kriging can indeed be useful and readily applied to dense ob-
servation grids, this method may also overestimate the prediction variance
for prediction points that are located further away from observation points.
Splines (2-dimensional, here referred to as splines) are also commonly used
to model data layers from discrete observations in time and space, where
there is no prior knowledge for parametric relationships between drivers
and the observations. In practice, splines are estimated within the frame-
work of generalized additive models (GAM). Splines are non-parametric
functions that aim at capturing an underlying spatial pattern in data through
smooth functions, which appropriately balance the spline’s goodness-of-fit
against the spline’s curvature (wiggliness), i.e. selecting smooth non-para-
metric functions that describe data reasonably well without overfitting the
observations. Typically, the mean field is described as a smooth function of
the coordinates (  ), but the spline model can also include co-
variates and time-dependencies. Although it is possible to formulate more
complex covariance structures as Generalized Additive Mixed Models
(GAMM), this is seldom done in practice due to the impracticability of esti-
mating both fixed and random model structures for the same data, unless
these models are very simple. Consequently, the spatial random variation
may not be adequately described using splines to model data layers.
Combining kriging and spline approaches could be advantageous, utilizing
the advantages of one method to accommodate the shortcomings of the
other. In short, splines are good in capturing variations attributable to un-
derlying but unknown relationships, whereas kriging is better at describing
small-scale random variations. In fact, both approaches are inadequate if
spatial variations are a combination of both large-scale fixed and small-scale
random. Therefore, the potential for combining the two approaches has
been investigated using oxygen depletion maps as an example.
2.6.3 Combining kriging and splines for oxygen depletion
Oxygen depletion maps are produced every month (August to October/Nov-
ember) as part of the Danish environmental reporting. The spatial distribu-
tion of oxygen depleted bottom waters is estimated from CTD casts (see
Carstensen & Erichsen 2003). In short, for each CTD profile the depths of 2
mg L-1 and 4 mg L-1 are found or predicted by regression analyses, in case
these thresholds are not observed in the profile (i.e. oxygen concentrations
are expected to fall below these thresholds at depths deeper than the pro-
file). The depths for the observed or expected occurrences of the two oxy-
gen thresholds are interpolated using ordinary kriging with a linear vario-
gram model, and the predicted oxyclines are combined with the bathymetry
to assess if oxygen concentrations below the thresholds are to be expected
in each grid point over the entire spatial domain. Maps from September, the
month when hypoxia peaks, were produced for the threshold of 2 mg L-1
based on the current approach using a linear variogram model as well as an
exponential variogram model fitted from empirical variograms for 13 years.
The latter was used to produce 95% confidence layers for the extent of oxy-
gen depletion with the ordinary kriging approach.
However, as stated above ordinary kriging assumes that the oxycline is con-
stant over the entire spatial domain, although it is locally adjusted to the ob-
served profiles. Therefore, the mean field of the oxycline was estimated on
data from multiple years (2002-2014) using a spline () to describe the
mean spatial variability across all years and an additive yearly factor () al-
lowing the oxycline mean field to move up and down for individual years
(denoted ). Hence, for the oxycline depth
  
where  is the prediction error in a given grid point and year. For
each individual year, the spatial variation of the residuals was modelled us-
ing ordinary kriging. This means that for a given year (), the oxycline depth
is distributed with a mean of   and a variance of  
     . The residuals from the GAM were analyzed with ordi-
nary kriging for each year, and the oxycline mean field was predicted by
combining the mean field with the ordinary kriging predictions.
NIVA 7562-2020
32
2.6.4 Uncertainty assessment of aggregated products
If all data layers are described with spatial distributions ( 
) then the resulting data layer following from an aggregation scheme
() will similarly be described as a spatial distribution. There are
several approaches to calculate the resulting spatial distribution from the
aggregation scheme, but the main three are: i) if the aggregation scheme is
linear then the distribution can be calculated in each grid point following
standard algebraic calculus, ii) if the aggregation scheme is non-linear, but
can be linearly approximated by a Taylor expansion then the same algebraic
calculations apply, and iii) otherwise determine the distribution by numeri-
cal integration using Monte Carlo simulations.
2.6.5 Initial uncertainty and data coverage assessment
All data are associated with some form of uncertainty and the more than
100 data layers within ECOMAR are very variable. Transforming 3-dimen-
sional information based on the actual existence into 2-dimensional maps
requires assumptions and may introduce biases. This is the case for all 2-di-
mensional maps but are assumptions that we accept.
Some of the ECOMAR data layers are based on direct observations, for ex-
ample stone reefs within Natura 2000 areas, archaeological sites and find-
ings, and recreational use. Other data layers are initially based on observa-
tions that have been used for further modelling of e.g. distributions of sea
birds, marine mammals, fish species, eelgrass distribution, broad scale ben-
thic habitat distributions and areas of oxygen depletion. The models are
based on a combination of observations and other explanatory factors, such
as seabed slope, currents, bathymetry, photic depth, salinity and sediments.
Within ECOMAR we have also assessed the overall uncertainty and data cov-
erage for the ecosystem components. The overall combined uncertainty of
the data layers was estimated using the different uncertainty variables of
the layers. In some cases the model error variables were used, such as the
coefficient of variation, standard deviations or the model exploratory per-
centage (derived from the r2 values). Uncertainty based on the amount of
fishing hauls, percent and landings of fish correlated to areas fished per year
(WMS data) were applied to the fish distribution maps. A categorical uncer-
tainty estimation was applied for the broad scale habitats, with 3 categories,
based on the uncertainty of the underlying models.
In other cases, where model estimates were not available categorical classi-
fication of the uncertainty was made from a scale of 0 = Observed data, 0.25
= Very good/validated model, 0.5 = Good model, 0.75 = Weak model/best
guess/extrapolation and 1 = No data.
Ideally, one should be able to provide a sufficient estimation of the associ-
ated uncertainty while creating a data layer, but this is not the present situa-
tion. Hence, the initial overall uncertainty assessment made in ECOMAR is a
first attempt of mapping the areas with high and low uncertainty and data
coverage, based on the uncertainty estimations available and the data cov-
erage maps, within the Danish EEZ. By this we are able to give a first indica-
tion of where to find areas with high quality data and where the data cover-
age is good, as well as to identify where more effort in collecting ecosystem
data should be undertaken.
Details regarding the models and uncertainties for each data layer are found
in the Supplementary Material (Annex A and B).
NIVA 7562-2020
33
3 Results
Using spatial data sets of the distributions of both human activities and pressures and ecosystem components, we have mapped and analysed
the potential combined effects in relation to the MSPD, MSFD and WFD. Subsequently, we have ranked pressures and identified potential key
pressures per group of ecosystem components. Based on comprehensive analyses of changes in the pressure groups as well as re-allocation
or introduction of new activities, we set up three scenarios for tentatively estimate the combined effects in 2030, 2050 and in a scenario
aiming to improve environmental status according to the MSFD goals of GES. There has been special focus on data coverage and uncertainty
including a future framework for assessment of uncertainty in spatial distribution maps. Last we have made a draft zoning matrix.
3.1 Pressure map
Calculated pressure intensities range between 1.3 and 12.8. The scale is ar-
bitrary, but important information about the spatial distribution of pres-
sures in Danish marine waters can be taken from Fig. 4. Areas without pres-
sures do not exist. However, areas with low pressure intensities can be
found in the central offshore parts of the Kattegat south of the island Læsø
and north and west of the island Anholt. Another offshore area with low
pressure intensity is located southwest of the island of Bornholm. Low pres-
sure intensities can also be found in some coastal waters, e.g. Hjelm Bay,
Jammer Bay and Sejerø Bay.
High pressure intensities are found mostly in coastal waters, e.g. areas in the
vicinity of major ports, coastal waters receiving discharges from cities, indu-
stries or upstream catchments, areas with high shipping intensity. Other ar-
eas with increased pressure intensities are also found where coastlines force
human activities to congregate for example in entrances to the Limfjorden
(both east and west), Isefjorden, Ringkøbing Fjord and in the areas between
the island of Funen and the Jutland peninsula and between Elsinore and Hel-
singborg in the Sound. Most of the shallow estuarine systems in Denmark
have high pressure intensities, e.g. Horsens Fjord, Limfjorden, Mariager
Fjord, Odense Fjord, Randers Fjord, Roskilde Fjord and Vejle Fjord. Further,
the Sound, the strait between Denmark and Sweden, and the southern parts
of the Little Belt and the Great Belt also have high pressure intensities.
3.2 Ecosystem map
Mapping the ecological components index illustrates a large spatial variation
in the number of ecosystem components in the different parts of Danish
marine waters (Fig. 5). Based on a total of 56 data sets representing a broad
range of ecosystem components), we identify areas with high ecological
value (high index values) as well as areas with lower ecological values or
missing data (low index values).
Areas with high index values are characterized by occurrence of many differ-
ent species, habitats and communities and are in Danish marine waters
found in the northern parts of the North Sea, Skagerrak and northern and
central parts of the Kattegat.
Areas with low index values are found in the Danish parts of the Baltic Sea,
especially in the water around the island of Bornholm and in the coastal wa-
ter south of the Little Belt, the Great Belt and the Sound.
Given the natural variability in salinity in Danish marine waters, and in the
Baltic Sea, we would expect to find low values in the Danish parts of the
southwestern Baltic Sea and to identify areas with high values in frontal ar-
eas in saline waters. This is also the case, as can bee seen in Fig. 5 lowest
values are found around Bornholm and in the Arkona Basin, whilst the high-
est values are found in the Skagerrak area.
NIVA 7562-2020
34
Figure 4: ECOMAR Pressure map a mapping of the current pressure and activities data sets. The colour scale shows the stretch for 2.5 standard deviations
from the mean, where the darker colour indicates a greater intensity/presence of pressure data sets.
NIVA 7562-2020
35
Figure 5: ECOMAR Ecosystem map mapping of the current ecosystem component data sets. The colour scale shows the stretch for 2.5 standard deviations
from the mean where the yellow colour indicates a greater intensity/presence of ecosystem component data sets.
NIVA 7562-2020
36
3.3 Mapping of combined effects
Potential combined effects of multiple human pressures, earlier denoted
‘cumulative impacts’, have been mapped in Danish marine water areas pre-
viously (Korpinen et al. 2012, Andersen & Stock 2013 and Andersen et al.
2020). ECOMAR builds on these studies but is based on broader data sets.
ECOMAR has established new sensitivity scores specifically for the pressures
and ecosystem components available. The median sensitivity scores pro-
vided by the expert survey ranged from 0 to 2 (see Supplementary Material,
Annex C1). This demonstrates that, according to the experts, some activities
and pressures may have no effects or insignificant effects upon specific eco-
system components whilst other pressures can potentially have much larger
effects on specific ecosystem components. By using the medians instead of
the means of the survey replies, we were able to avoid any possible large bi-
ases driven by either a few low or high replies.
The expert-derived effect distances are listed in Table 5 (see Supplementary
Material, Annex C2 for details). In general, pressures will have effects in the
precise location where they take place and also in their vicinity. The effect
distances applied do not indicate any long-range effects of pressures and is
assumed to be linear.
When combining information on pressures, ecosystem component, sensitiv-
ity scores and effect distances using the EcoImpacMapper software, we
have mapped the potential effects of multiple human pressures in relation
to the MSPD including climate change. Areas with high index values were
found both offshore and in coastal water (Fig. 6, an index map without cli-
mate change is found in Supplementary Material, Annex C3).
Offshore areas with high index values are found in parts of the North Sea, in
the northern parts of the Skagerrak, in the southwestern parts of the Katte-
gat and around the island of Bornholm, both to the west and east. The back-
ground for these is often a combination of high intensities of shipping, fish-
ing and contaminants.
Offshore areas with low index values are found in some parts of the North
Sea, in the southwestern parts of the Skagerrak, in the central parts of the
Kattegat and south of the island Bornholm. These are attributable to low in-
tensities of offshore activities (fishing and oil and gas).
Coastal areas with high index values are widespread and found from west to
east, e.g. in the Wadden Sea, along the west coast of Jutland, in almost all
estuaries, and in large parts of the Little Belt, the Great Belt and the Sound.
The causes are in general inputs of polluting substances from land.
Coastal waters with low index values are areas without major discharges
from land and are found in the Jammer Bay, Tannis Bay, Sejerø Bay, Mus-
holm Bay, north of Zealand, south of Læsø and around Anholt.
Table 5: Effect distances (medians in km) used in the calculations of the po-
tential combined effect index within the Danish EEZ.
Pressure
Effect distance
Dumped chemical munitions
5
Aquacultures: fish and shellfish farms
5
Sea cables
0
Offshore oil and gas installations
1
Oil and gas pipelines
0
Heat and power plants
1
Disposal sites for construction and dredged material
5
Dredging in harbours and shipping lanes
5
Dredging sites in production
1
Offshore wind turbines
1
Bridges and costal constructions
1
Coastal habitat modification (coastal protection, piers)
1
Lighthouses
0
Military areas
7.5
Marine ports: industrial
5
Marine ports and marinas: recreational
3
Mussel dredging
1
Dumped chemical munitions
5
NIVA 7562-2020
37
Figure 6: Map of intensities and spatial variations in the estimated combined effects of human pressures and activities including climate change. The colour
scale shows the stretch for 2.5 standard deviations from the mean, where red indicates a higher effect impact and blue lower. Please note that the values are
unit less and that the magnitude is defined by the models data inputs, which here are normalised between 0-1. The index is calculated using EcoImpactMap-
per developed by Stock (2016).
NIVA 7562-2020
38
3.4 Ranking of pressure groups
Based on the mapping of combined effects in Danish marine waters, we
have ranked and grouped pressures both nationally and regionally as well as
for coastal waters (the WFD domain) and offshore waters (the Danish EEZ
minus WFD coastal waters).
The analyses are based on 42 individual pressures and activities, which have
been combined in 13 pressure groups, similarly to what was done in the
context of HELCOM (Korpinen et al. 2012), the HARMONY project (Andersen
& Stock 2013), the two published Danish MSFD Initial Assessments (Na-
turstyrelsen 2012 and Miljø- og Fødevareministeriet 2019) and the Havplan
Øresund project (Riemann et al. 2019). The climate change pressure group is
included in the results shown in Fig. 6, the results without climate change
can be found in Supplementary Material (Annex C3).
On a national scale, ranking of pressures within the Danish EEZ (Fig. 7A) re-
veals that the potential governing pressures are: i) commercial fishing, ii) nu-
trient enrichment, iii) contaminants, iv) climate change and v) marine litter.
Other pressures, which are important, are vi) noise and energy, and vii) rec-
reational activities. Pressures, which may have local importance are iix) non-
indigenous species, ix) shipping, x) industry including energy production, xi)
recreational fishing and hunting, xii) physical disturbance and xiiv) aquacul-
ture.
In the North Sea/Skagerrak region (Fig. 7B), the main pressures groups are:
i) commercial fishing, ii) nutrients enrichment, iii) climate change, iv) con-
taminants, and v) marine litter. Other pressures, which are important on a
regional level, are vi) noise and energy, and vii) non-indigenous species.
Pressures, which may have local importance are iix) recreational activities,
ix) industry incl. energy production (the latter not entailing losses of contam-
inants), x) shipping and transportation, xi) physical disturbance, and xii) rec-
reational fishing and hunting.
For the Danish parts of the Kattegat (Fig. 7C), potential dominant pressure
groups are: i) nutrients enrichment, ii) recreational activities, iii) climate
change, iv) contaminants, and v) marine litter. Other pressures, which are
important on a regional level, are vi) commercial fishing and vii) noise and
energy. Pressures, which may have local importance, are iix) shipping, ix)
non-indigenous species, x) industry including energy production, xi) recrea-
tional fishing and hunting, xii) aquaculture, and xiiv) physical disturbance.
In the Danish parts of the western Baltic Sea (Fig. 7D), potential dominant
pressure groups are: i) contaminants, ii) nutrient enrichment, iii) marine lit-
ter, iv) climate change and v) recreational activities. Other pressures, which
are important on a regional level are vi) commercial fishing, and vii) noise
and energy. Pressures, which may have local importance are iix) non-indige-
nous species, ix) shipping and transportation, x) industries including energy
production xi) recreational fishing and hunting, xii) aquaculture, and xiiv)
physical disturbance.
By only including the ecosystem components relevant for the MSFD (See Ta-
ble 4) and excluding the pressures of climate change (Fig. 8A) , the dominant
pressure groups are: i) commercial fishing, ii) nutrients enrichment, iii) con-
taminants, iv) marine litter and v) noise and energy. Other pressures, which
are important are vi) recreational activities, vii) non-indigenous species and
iix) shipping and transportation. Pressures, which may have local impor-
tance are ix) industry including energy production, x) recreational fishing, xi)
physical disturbance, and xii) aquaculture.
Focusing on coastal waters under the WFD domain (see Table 4 and Fig. 8B
and C), the ranking of the pressure groups differs depending on the ecosys-
tem components included. Common is that nutrient enrichment has the po-
tential largest impact together with contaminants, commercial fishing and
marine litter. Recreational activities may have a larger impact when as-
sessing the direct ecosystem components within the WFD, whearas non- in-
digenous species seems to be of higher importance for the assessment of
indirect ecosystem components.
NIVA 7562-2020
39
Panel A: Danish EEZ
Panel B: North Sea/Skagerrak
Panel C: Kattegat
Panel D: Western Baltic Sea
Figure 7: Ranking of pressures in the Danish EEZ (panel A). Results for sub-divisions, i.e. the Danish parts of the North Sea and Skagerrak, the Kattegat and
the western Baltic Sea, are shown in panels B, C and D, respectively. Results of the same analyses but without the pressure group ‘climate change, can be
found in the Supplementary Material (Annex C3).
NIVA 7562-2020
40
Panel A: MSFD
Panel B: WFDDIR
Panel C: WFDDIR+INDIR
Figure 8: Ranking of pressures in the Danish EEZ for the MSFD (panel A) and the WFD. For WFD, panel B shows ‘direct ecosystem components’, whilst panel C
shows ‘direct and indirect ecosystem components’. See Table 4 for details on the ecosystem components included. Note that climate change is not included
for the MSFD and WFD analyses.
NIVA 7562-2020
41
3.5 Analysis and scenarios
A key objective for ECOMAR has been to analyse how changes in pressure
intensities may potentially change the combined effects and subsequently
lead to improvements or worsening of the environmental status in Danish
marine waters.
3.5.1 Increase/decrease in existing activities and pressures
In the following section, we describe the possible consequences of modify-
ing the pressure intensities of the 13 groups of pressures that have been al-
tered, i.e. reduced or increased according to the considerations and descrip-
tions in section 2.4.2.
Pressure group 1: Aquaculture
The ecosystem groups most likely to be affected by aquaculture farms are
pelagic habitats (in ECOMAR: phytoplankton and oxygen concentration in
bottom waters), benthic habitats and recreational interests. Other ecosys-
tem components potentially affected are various species of fish and birds.
The estimated impacts in the 2030 and 2050 scenarios are as follows: In
2030 and 2050 and in the MSFD scenario, improved condition may be ex-
pected regarding pelagic habitats, i.e. chlorophyll a concentration and to a
lesser degree for benthic habitats and recreational interests (Fig. 9A). The
MSFD scenario indicates that significant declines in impacts could be within
reach. The relation between aquaculture and plankton is well known, but it
is interesting to find a relation with recreational interests as well. In the
MSFD scenario, improvement could be up to three to five times greater,
probably even more pronounced on a regional or local scale.
Pressure group 2: Climate change
Based on the results of the mapping of combined effects and ranking of
stressors, the ecosystem components likely to be affected most by climate
change are pelagic habitats (increased phytoplankton biomass and lower ox-
ygen concentration in bottom waters) and benthic habitats (e.g. submer-ged
aquatic vegetation, biogenic reefs, and the composition of benthic inverte-
brates).
Panel A: Aquaculture
Panel B: Climate change
Figure 9: Differences in impacts on key ecosystem components cf. the base-
line and the 2030, 2050 and MSFD scenarios. Panel A: Aquaculture. Panel B:
Climate change. Please note the differences in ecosystem components and
the scale on the y-axises.
Both the 2030 and 2050 scenarios (Fig. 9B) show that climate change will
lead to an increase in impact on marine ecosystems in Danish waters. This
may jeopardize potential improvements likely to be obtained through reduc-
tions of other pressures. Within the MSFD GES scenario, climate change is
considered as an exogenic pressure and included (see section 2.4.3).
NIVA 7562-2020
42
Pressure group 3: Industry, energy and infrastructure
The ecosystem components and cultural interests most impacted by pres-
sures anchored in industries, energy production and infrastructure are rec-
reational interests, birds and benthic habitats. In the 2030 scenario, an in-
crease in impacts can be expected due to an increase in the intensity of
these activities (Fig. 10A). Groups of ecosystem components and societal in-
terests most affected are recreational interests and benthic habitats. Other
ecosystem groups to be affected are fish and crustaceans, birds and pelagic
habitats. In the 2050 scenario, the most impacted groups are recreational
interests, benthic habitats and birds. Even in the MSFD scenario, negative ef-
fects are envisaged, mostly regarding marine mammals, benthic habitats,
birds and recreational interests. Hence, the planned activities in 2030 and
2050 and in the MSFD scenario, will result in increased impacts on marine
ecosystems and most likely contribute to a further deterioration, directly or
indirectly, of environmental status in Danish marine waters.
Pressure group 4: Marine litter
The ecosystem groups most likely to be impacted by marine litter are birds
and fish, while recreational interests, marine mammals and both benthic
and pelagic communities may also be impacted. For the 2030 and 2050 sce-
narios (Fig. 10B), the difference between the years is directly related to the
expected increase in pressure intensity. However, the MSFD reveals that re-
duction in pressure intensity would probably reduce the effects and subse-
quently lead to improvements in environmental status.
It should be noted that marine litter is a rather diverse group spanning sev-
eral types of litter, e.g. microplastic, plastic of different sizes and ghost nets.
These sub-groups may impact different ecosystem groups in different ways
macrolitter is known to be eaten by animals, e.g. birds, while microplastic
may be by eaten by filter feeders or deposited at the seafloor. Knowledge
about the effects for the various types of marine litter is scarce for the mo-
ment and more research on this is required to not only better understand
the relationships between the effects but also to estimate potential impacts.
Panel A: Industry, energy and infrastructure
Panel B: Marine litter
Figure 10: Differences in impacts on key ecosystem components cf. the base-
line and the 2030, 2050 and MSFD scenarios. Panel A: Industry, energy and
infrastructure. Panel B: Marine litter. Please note the differences in ecosys-
tem components and the scale on the y-axises.
NIVA 7562-2020
43
Pressure group 5: Noise and cooling water
Reduction in noise levels is mostly linked to the reduction of local impulsive
noise and will lead to reductions in the impacts on marine mammals, sea-
birds, and recreational interests. Minor reductions in impacts can be found
for fish and benthic communities.The highest increase in impacts in the
2030 scenario is found for marine mammals, while some increase is also
found for birds and recreational interests (Fig. 11A). In the 2050 scenario,
only minor increases from todays levels are expected. However, the MSFD
scenario indicates potentially large reductions in impacts, especially for ma-
rine mammals, fish, birds and also recreational interests. These results
should be seen as provisional and require more detailed analyses and stud-
ies if proven correct, there is an untapped potential for measures, for reg-
ulation the impulsive noise, that may ultimately improve environmental
conditions for higher trophic levels, in particular marine mammals and fish.
Further, the relations between reduced levels of noise and recreational in-
terests should be scrutinized at a variety of spatial scales, e.g. sub-regionally
and locally, as this pressure group may have a large influence.
Pressure group 6: Non-indigenous species
The introduction of non-indigenous species (NIS) to the Danish marine envi-
ronment may potentially have a large influence on its structure and functi-
oning, as well as its species, communities and populations. In some cases,
NIS can become invasive thereby acting as a significant pressure on endemic
species. Substantial impacts from NIS in some parts of Danish marine waters
are well-known and considered an emerging risk as the rate of newly intro-
duced species is relatively constant (Stæhr et al. 2016).
The key ecosystem components expected to be impacted by NIS are benthic
habitats (including crustaceans) and recreational interests. Assuming an un-
changed rate of new introductions (Fig. 11B), the 2030 scenario reveals in-
creased impacts, while the 2050 scenario, being based on improved mana-
gement practises and at the same level as today, the impacts are not sur-
prisingly matching todays estimated impacts. No changes are seen in the
MSFD scenario where already introduced species are present, and the pres-
sure was not altered.
Panel A: Noise and cooling water
Panel B: Non-indigenous species
Figure 11: Differences in impacts on key ecosystem components cf. the base-
line and the 2030, 2050 and MSFD scenarios. Panel A: noise and cooling wa-
ter. Panel B: Non-indigenous species. Please note the differences in ecosys-
tem components and the scale on the y-axes.
NIVA 7562-2020
44
Pressure group 7: Physical disturbance of the sea floor
Physical disturbance from a broad range of human activities is widespread in
Danish marine waters, e.g. from dredging and maintance of shipping lanes in
shallow waters, exploitatation of natural resources such as sand and gravel
or from smothering from activities such as dumping of dredged materials
from harbous and shipping lanes. With a reduction in dredging of sand and
gravel in 2030 compared to today activities, a significant decrease in impacts
can be expected (Fig. 12A), especially for the following ecosystem compo-
nents: fish and benthic habitats including crustaceans. Some reduction in
impacts are likely for pelagic habitats (reduced resuspension) as well as rec-
reational interests. The 2050 scenario indicates a slight increase in pressure
intensities impacting the ecosystem components in question almost equally,
except the sensitive fish species.
Given that significant improvement are attained in a short term perspective
and not in a long term, both political focus and more research on the envi-
ronmental consequences of dredging of sand and gravel as well as smother-
ing is urgently required.
Pressure group 8: Contaminants
Discharges and losses of contaminants from Danish sources in combination
with long-range transport and deposition constitute important pressures for
the Danish marine environment (see section 3.3 and Fig. 8A). Multiple stra-
tegies and action plans have been adopted and implemented, presumably
with a variety of successes (Dahlöf & Andersen 2009). In the 2030 scenario,
where the pressure intensity is assumed to increase slightly, the impact will
increase with regard to fish and crustaceans, but also marine mammals. The
2050 scenario may, however, lead to reductions in the pressure intensity
and thus a lower impact on marine mammals, fish, benthic habitats and
birds. The MSFD scenario, focusing on attaining a better environmental sta-
tus through a major reduction of pressure intensity, indicates lower impacts
on the following ecosystem component groups: fish, marine mammals, ben-
thic habitats, and seabirds (Fig. 12B). The latter indicates that reductions of
inputs of contaminants are essential for higher trophic levels to improve the
currently impaired conditions and to meet both the objectives of the MSFD
and WFD as well as the so-called Generation Target.
Panel A: Physical modification
Panel B: Contaminants
Figure 12: Differences in impacts on key ecosystem components cf. the base-
line and the 2030, 2050 and MSFD scenarios. Panel A: Physical modification.
Panel B: Contaminants. Please note the differences in ecosystem compo-
nents and the scale on the y-axises.
NIVA 7562-2020
45
Pressure group 9: Nutrients
Nutrient inputs resulting in elevated nutrient concentration and eutrophica-
tion effects have for decades been a crucial pressure in Danish marine wa-
ters, especially in estuaries and coastal waters (see also section 3.3 and Fig.
8A). Significant efforts have been made to reduce nutrient inputs from agri-
culture, urban wastewater treatment plants and from industries with sepa-
rate discharge (see Andersen 2012 and Riemann et al. 2016).
The 2030 and MSFD scenarios are identical and the groups of ecosystem
components most likely to face reduced impacts are pelagic habitats (i.e.
chlorophyll a concentration in surface water and oxygen concentrations in
bottom waters) and benthic habitats including the key species eelgrass. In
the 2050 scenario is based in reduction in nutrient inputs and thus lower nu-
trient levels is surface waters, highlights that significant reduction in pres-
sure intensity and subsequently impacts on pelagic and benthic habitats
(Fig. 13A). This indicates that improvement in both coastal waters (WFD do-
main) and offshore water (MSFD domain) can be expected this should sup-
port implementation of additional measures and reduction in nutrient in-
puts. Follow up analyses focusing on specific coastal waterbodies, specific
ecosystem components groups (related to WFD biological quality elements
or the MSFD D5 descriptor) are urgently needed.
Pressure group 10: Selective extraction of species: commercial fishing
A growing number of studies and reports on human activities and pressures
in Danish marine waters have indicated that fishing, especially bottom trawl-
ing, is an significant pressure (Miljø- og Fødevareministeriet 2019, HELCOM
2018, Andersen et al. 2020, EEA 2020). These results are confirmed by the
analyses done in the context of ECOMAR. In 2030, assuming a reduction in
fishing intensity, reduction in impact may be expected for the following eco-
system groups: fish, crustaceans and benthic habitats. Some effects, but to a
lesser extend, are foreseen for pelagic habitats, marine mammals (due to
bycatch) and recreational interests. The 2050 scenario is parallel to the 2030
scenarion with slightly higher reduction in pressure intensity, whilst the
MSFD scenario indicate that significant rediction in pressure intensity (Fig.
13B).
Panel A: Nutrients
Panel B: Selective extraction of species: commercial fishing
Figure 13: Differences in impacts on key ecosystem components cf. the base-
line and the 2030, 2050 and MSFD scenarios. Panel A: Nutrients. Panel B: Se-
lective extraction of species: commercial fishing. Please note the differences
in ecosystem components and the scale on the y-axises.
NIVA 7562-2020
46
Pressure group 11: Selective extraction of species: recreational fishing
Given the availability of information on recreational fishing, we have tenta-
tively estimated the potential effects of changes in the intensity of this spe-
cific activity. Assuming a slight increase in recreational fishing in 2030, an in-
creased impact is seen on fish populations, in bentich habitats and for recre-
ational interests. In the 2050 scenario, the ecosystem components esti-
mated to encounter reduced impacts are fish and crustaceans as well as sea
birds, mammals and bentic habitats. The MSFD scenario indicates reductions
in the pressures on the following ecosystem components: seabirds and fish
as well as recreational interests (Fig. 14A).
Although being a pressure of restricted importance on a national scale, rec-
reation can be of significant importance locally, for example in Øresund.
There is a need for more detailed studies, also linking the status and pres-
sures of target fish species to environmental conditions.
Pressure group 12: Shipping and transportation
Shipping can impact a broad range of ecosystem components through pres-
ence, resuspension of material at the seafloor or by generating waves etc.
Accordingly, the key ecosystem component groups impacted are seabirds,
marine mammals and benthic habitats. Recreational interest can also be af-
fected.
The 2030 scenario indicates elevated levels of impacts for seabirds, recrea-
tional interests, marine mammals and benthic habitats, the latter probably
through phycisal effects. In the 2050 scenario, the same ecosystem compo-
nent groups will be even more impacted. In the MSFD scenarios, the ecosys-
tem groups assumed to face a reduction in the impacts are the same (Fig.
14B).
Pressure group 13: Recreation and tourism
Recrational activities and tourism primarily have impacts on seabirds, ma-
rine mammals, benthic habits, and other recreational interests. In the 2030
and 2050 scenarios, the pressures are assumed to increase and so are the
potential impacts on the ecosystem groups and recreational interests (Fig.
15). Reductions of recreational activities and tourism, and envisage in the
MSFD scenario, will lower the impacts on seabirds, marine mammals, recre-
ational interests and benthic habitats.
Panel A: Recreational fishing
Panel B: Shipping and transportation
Figure 14: Differences in impacts on key ecosystem components cf. the base-
line and the 2030, 2050 and MSFD scenarios. Panel A: Recreational fishing.
Panel B: Shipping and transportation. Please note the differences in ecosys-
tem components and the scale on the y-axises.
NIVA 7562-2020
47
A key lesson learned from this straightforward analysis is that follow up
studies on the interlinkages between recreation and tourism and marine
ecosystem components (including other recreational interests) and vice
versa is required is order to achieve a better understanding as well as basis
for decision-making, e.g. in the context of MSPD, MSFD and WFD.
Recreation and tourism
Figure 15: Differences in impacts from recreation and tourism on key ecosys-
tem components cf. the baseline and the 2030, 2050 and MSFD scenarios.
3.5.2 Placement of activities and pressures
SeaSketch can be used to analyse the implication of reallocation of existing
activities to other areas. A real-life example of this is from the late 80s and
early 1990s, where several medium sized fish farms were reallocated from
shallow coastal waters, often inside fjord systems to more open waters. One
of the better examples was the fish farm in Nordby Bay, which previously
was located within Stavns Fjord, northeast of the island Samsø.
We are using the placement of two hypothetical aquaculture farms as a
demonstration example. Here, we show the consequences and differences
between two locations of a fish farm near Horsens Fjord, where one of the
selected locations is within an existing Natura 2000 area and the other out-
side (Fig. 16). When the aquaculture is placed within the Natura 2000 area
(alternative 1), and where there is a high potential of eelgrass coverage, the
impacts on the ecosystem components are higher compared to when the
aquaculture is placed outside the MPA (alternative 2). In alternative 2 the
relative impact on bentic habitats is less than half the impact in alternative 1
(~45%). This would mean that a placement of the aquaculture outside the
Natura 2000 area would be more beneficial from an ecosystem-based man-
agement perspective.
Other case studies of relevance could be: i) Natura 2000 areas, of which
some probably could be more suitable located, to ensure a higher ecological
protection, and ii) some shipping lanes could probably also be modified for
safety reasons. These issues are not considered in the ECOMAR project, but
thay are examples that it would be relevant to consider, as well as other
cases, at a later stage.
It should be mentioned that the aquaculture example used here is for
demonstration of the tool’s suitability. For undertaking a tangible environ-
mental impact assessment in a defined local assessment, higher spatial reso-
lution of ecological data and qualitative relationships between pressures
and ecosystem components (if present), would be more legitimate to use.
NIVA 7562-2020
48
Panel A: Study area
Panel B: Differences in impacts
Figure 16: Map (panel A) showing the example of the two alternative place-
ments of an aquaculture farm. Panel B shows the difference in the relative
impacts when placing the aquaculture within a protected area with sensitive
ecosystem components like eelgrass compared to in a less sensitive area.
3.5.3 Introductions of new activities
As with reallocation of existing activities, SeaSketch can analyse the impacts
of introducing new activities. To demonstrate this, we present two case
studies: i) introduction of new offshore wind farms in the North Sea and ii)
introduction of the planned Kattegat Bridge in the northern parts of the
Great Belt between Røsnæs on the west coast of Zealand and Hou on the
eastcoast of Jutland.
Denmark plans to expand the number and capacity of offshore winds parks
(DEA 2020). Thus, we have analysed the potential increase in impact for the
example of a new wind power park at Kriegers Flak, which is approved and
will be in place in a couple of years. When adding a wind park it means that
several ecosystem components will be affected (Fig. 17). There will be an in-
crease in the impulsive noise (affecting marine mammals) during the con-
struction as well as habitat loss for benthic species. Both recreational inter-
ests and birds species may be affected: i) birds as they have problems avoid-
ing the rotor blades and ii) recreational interests as there will be restrictions
in the access to the area for e.g. fishing, sailing etc, as well as an often per-
ceived negative impact on the scenery.
Another example is the addition of the planned Kattegat bridge between
Zealand over Samsø to Jutland (Ingeniøren 2018) (Fig. 18). Though the ferry
routes to and from Samsø are reduced, continuous noise remain the same
as there are many other ferry routes crossing the area. The pressure from
harbours decreases whereas all recreational activities will increase as the
area will be more accessible by the bridge. Commercial fishing is not a large
activity within the area but will decrease. Coastal constructions and disposal
of material will increase but dredging in relation to the ferry routes will de-
crease.
Other case studies of potential interest could be introduction of the so-
called ‘energy island’, additional Natura 2000 or MSFD areas, or new fish
farms, mussel farms or areas for production of macroalgae. These and many
other relevant cases could be explored and analysed.
NIVA 7562-2020
49
Wind power park at Kriegers Flak
Figure 17: Placement of the newly installed wind park at Kriegers Flak, east
of the island Møn. The graph shows the relative potential change in % (in-
crease or decrease) of the effects upon the ecosystem component groups.
The graph shows the relative change in % impact for adding an offshore
wind farm at Kriegers Flak, east of the island Møn.
Planned bridge across the Kattegat/Northern Great Belt
Figure 18: Location of the proposed Kattegat Bridge between Zealand (east)
and Jutland (west) the areas that potentially will be affected around it. The
graph shows the relative potential change in % (increase or decrease) of the
effects upon the ecosystem component groups.
NIVA 7562-2020
50
3.6 Combining MSPD, MSFD and WFD pressures
analyses into a coherent process
EU’s marine directives, i.e. the MSPD, the MSFD, both covering all marine
waters within the EEZ, and the WFD, covering coastal waters, all require that
pressures are assessed and reported as part of either a national maritime
spatial plan cf. the MSPD or Initial Assessments cf. the MSFD and WFD.
Given the quality and broad extent of the ECOMAR data sets, we demon-
strate that the above pressure assessments can be combined into a joint
process using the same data but with outputs targeted by addressing the di-
rective-specific requirements of the MSFD and WFD.
From running the EcoImpactMapper in combination with ECOMAR data
sets, we have produced the following four analyses: i) based on all pres-
sures including climate change and ecosystem components representing di-
rect WFD quality elements (Fig. 20A), ii) same as i), but without climate
change (Fig. 20B), iii) based on all ecosystem components representing both
direct WFD quality elements such as oxygen depletion and benthic habitats
and indirect elements (see Table 4) in coastal waters (Fig. 20C), and iv) same
as iii), but without climate change (Fig. 20D). All four maps demonstrate that
mapping of pressures in coastal waters is achievable with the EcoImpact-
Mapper software or similar tools. The map for the MSFD relevant ecosystem
components are presented in the Supplementary Material (Annex C4).
Previous studies (Korpinen et al. 2013, Andersen et al. 2020 and many more
cf. Korpinen & Andersen (2018) and ECOMAR have documented that the
methodology is applicable and widely used for open waters. The maps, de-
spite the variations in data on which they are based on, all identify areas
with high, intermediate and low potential pressure effects. Although the
maps can be a useful prioritization tool and identify which areas are prone
to high levels of pressures and identify the potential dominating pressure
groups in these areas, they cannot be used to quantify reductions in indivi-
dual pressures. It can be argued that climate change should be disregarded
when mapping pressures in coastal waters as it is an exogenic pressure and
that the management of that takes place at a global scale. If so, the scenar-
ios to consider are those presented in Fig. 22B & D.
Determining which of the analyses that would be most useful, WFD direct
(using only the ecosystem components equivalent to WFD biological quality
elements) or WFD direct + indirect’ (using also ecosystem component indi-
rectly linked to the WFD biological quality elements), is difficult. In practice,
it relates to the confidence required and what the acceptable uncertainty is.
The analysis being broadest in terms of ecosystem components is therefore
regarded as the most reliable and should be a demonstration of the applica-
bility of CEA mapping in support of WFD pressure analyses.
The CEA method is widely used today by both researchers and authorities
(e.g. MFVM, MST, SwAM, HELCOM, and EEA), probably because correlations
between environmental status and pressures are well-documented, e.g. for
‘ecosystem health’ and ‘ecological status’ (Fig. 19). For more information
about validation of the CEA method, please see EEA (2019a) and Korpinen et
al. (submitted).
Figure 19: Validation of the CEA index. The example correlates the cumula-
tive effects index to two different Ecological Status Classes of marine ecosys-
tems within European seas (EEA 2019a, Korpinen et al. submitted)).
NIVA 7562-2020
51
Panel A: Direct effects including climate change
Panel B: Direct effects wihtout climate change
Panel C: Indirect effects including climate change
Panel D: Direct effects wihtout climate change
Figure 20: Mapping of potential combined effects of multiple human pressures within the WFD domain, i.e. coastal waters defined as the baseline plus one
nautical mile. Effect index map based on WFD biological quality element only (WFD direct), including pressure group climate change (Panel A) and without
‘climate change’ (Panel B). Effect index map based on WFD biological quality element and associated indicators (WFDin direct), including pressure group
‘climate change’ (Panel C) and without‘climate change’ (Panel D). Note that the max and min impact changes with the model, which is reflected in the scales.
NIVA 7562-2020
52
3.7 Combining analyses and scenarios
Based on the analyses described in section 3.5 as well as the introduction of
specific new activities, we have combined these into two theoretical future
scenarios for 2030 and 2050 as well as a hypothetical MSFD scenario with
focus on improvements in environmental status through reduction of key
pressures.
The 2030 scenario represents the most likely expected developments in
pressures in Danish marine waters based on the present knowledge and
agreed policies, strategies, plans and measures. Similarly, the 2050 scenario
represents most likely changes in pressure intensities according to available
information.
However, the MSFD GES scenario differs from the above 2030 and 2050 sce-
narios as it is anchored in a hypothetical reduction in a broad range of pres-
sures assumed to be needed to achieve a good environmental status.
An overview of the changes in pressure intensities is given in Fig. 21 and the
consequential ecosystem responses are summarized in Fig. 22.
3.7.1 2030 scenario
Comparing the pressure-specific effects in the 2030 scenario reveals
changes in the relative impacts on the national level and shows that some
pressure groups have reduced contributions while other have increased con-
tributions (Fig. 22).
Pressures groups where the changes in effects are numerically negative, and
thus appear to contribute to improved environmental conditions are: i)
commercial fishing (-8.2%), ii) nutrients (-3.4%) and iii) aquaculture (-1.9%).
For the following 10 pressure groups, we see numerically positive responses
to changes in pressure intensity and may therefore expect a negative impact
on the environmental conditions and status: i) climate change (20.1%), ii)
non-indigenous species (8.4%), iii) societal and recreational interests (6.7%),
iv) shipping and transportation (6.2%), v) physical disturbance (6.6%), vi) ma-
rine litter (3.4%), vii) industry, energy and infrastructure (2.3%), viii) recrea-
tional fishing and hunting (3.3%), ix) noise and energy (0.2%) and x) contami-
nants (0.3%).
These changes are relative to the baseline impact estimated for each pres-
sure group. Since Commercial fishing accounts for 18% of the combined
impact (Fig. 9A), an 8.2% change in the impact due to this pressure group is
approximately four times larger in absolute terms than the 8.4% change in
impact attributable to non-indegenous species, which accounts for 4.4% of
the baseline impact (Fig. 9A). Overall, the scenario for 2030 shows an in-
crease in impact of 8%.
When comparing the spatial differences (Supplementary Material, Annex
C5, Fig. C5.1) between the CEA baseline map and the changes of the 2030
scenario (Fig. 22), areas that potentially might be more affected can be iden-
tified. Only some small scattered areas in the North Sea and outside the
west coast of Denmark show a small decrease of up to 10% in the potential
cumulative pressure index when the changes in the 2030 scenario are imple-
mented. More generally, for the off-shore areas there is no change or a
small increase of up to 5% as the dominant change.
The exceptions are in the western and southern North Sea as well as south
of Bornholm, where there is an increase of 5-15% in the CEA index. Areas
with a more pronounced increase in the CEA index of 15% to above 20 % are
found within the fjords, the Kattegat, southern Baltic Sea as well as in all the
coastal areas. The large difference in the Kattegat is found in an area where
the current CEA index is relatively low. This means that in the future there
will only be a limited possibility to increase the pressures and that some of
the largest negative differences according to the changes in the 2030 sen-
ario will happen in areas which today show a relatively low CEA index.
NIVA 7562-2020
53
Figure 21: Overview of changes in pressure intensities. The plusses indicate ‘+’ = small increase, ‘++’ = moderate increase and ‘+++’ = large increase, while
the minusses indicate ‘-‘ = small decrease, ‘- -’ = moderate decrease, and ‘---’ = large decrease. ‘na’ = not applicable/ no change.
NIVA 7562-2020
54
3.7.2 2050 scenario
Adding another 20 years to the 2030 scenario and taking into account plan-
ned developments and expected changes in human activities (summarised
in Fig. 21), reveals a further increase in pressure intensities compared to the
present baseline (Fig. 22).
Pressure groups contributing to the increase in pressure intensities encom-
pass the following groups: i) climate change (25.8 %), ii) non-indigenous spe-
cies (16.8 %), iii) societal and recreational interests (13.3 %), iv) shipping and
transportation (12.0%), v) physical disturbance (10.2%), vi) marine litter
(6.7%), vii) industry, energy and infrastructure (4.6%) and viii) recreational
fishing and hunting (4.4 %). There is no change for ix) noise and energy.
Pressures groups where the effects are numerically negative, and thus tend
to contribute to improved environmental conditions are: i) commercial fish-
ing (-11.5 %), ii) nutrients (-6.7 %), iii) aquaculture (-1.9 %) and iv) contami-
nants (-1.1 %).
The changes revealed in the 2050 scenario provide a total negative impact
of 10.7%. The spatial differences between the baseline map and the 2050
scenario changes are shown in Supplementary Material, Annex C5, Fig.
C5.2.
Similar to the 2030 scenario there is a general low increase with 0-5% differ-
ence for the off-shore areas. However, the areas with a decrease (although
low at 0-10%) are relatively larger and are found spread across the North
Sea and again outside the Danish west coast. The difference in the CEA index
is larger and increasing with 10 to 20% or above in the same areas as in the
2030 sceario, but with larger extent and higher intensities. The largest dif-
ference is found in the southwestern Baltic Sea and the Kattegat with domi-
nating differences by an increase above 20%. As seen in the 2030 scenario,
areas with relatively low combined human pressures (low CEA index) can
potentially be disturbed, so these areas could be given extra attention in the
MSP plan, in addition to already intensively affected areas (with high CEA in-
dex), which might not be able to sustain more disturbance by different hu-
man pressures.
3.7.3 MSFD GES scenario
The MSFD GES scenario is, as mentioned, a hypothetical scenario but it does
however demonstrate the potential effects of reducing pressure intensities
across the board in order to improve environmental status of Danish marine
waters and thus fulfil the overarching objective of the MSFD, i.e. attaining a
good environmental status (GES) or at least move closer to fulfilling this ob-
jective.
Pressure groups with reduced effects on ecosystem component are: i) physi-
cal disturbance (-49.7%), ii) commercial fishing (-29.4%), iii) recreational fish-
ing and hunting (-26.1%), iv) aquaculture (-20%), v) marine litter (-19.9%),
vi) noise and energy (-18.5%), vii) shipping and transportation (-18.0%),
viii) industry, energy and infrastructure (-17.7%), ix) societal and recreational
interests (-14.9%), x) nutrients (-10%) and xi) contaminants (-9.6%).
The pressure groups ‘Non-indigenous species and ‘Climate change and the
contribution to the pressure intensites are cf. Fig. 22 assumed to be at the
same levels as today. With all these effects in combination, the MSFD GES
scenario results in a reduced impact of 14.7% overall (cf. the summary in Ta-
ble 8).
The spatial differences between the baseline and the MSFD GES scenario is
shown in Supplementary Material (Annex C5, Fig. C5.3I). A general de-
crease in the CEA index values in the order of 10%-20% is seen over the full
EEZ with large areas in the North Sea showing a difference by more than a
20% decrease.
In the MSFD GES scenario, the only areas with an increase in CEA index are
the areas where new wind power farms are currently being implemented or
will be in in the coming years.
The MSFD GES scenario shows that there is room for improvement and that
a potential decrease in the combined effects of human pressures can be
achived by reduction of pressure levels.
NIVA 7562-2020
55
Figure 22: Differences between the baseline and the 2030, 2050 and MSFD GES scenarios, where O indicates no change, positive values an increase in CEA
per pressure group and a negative values indicates a decrease in CEA per pressure group. See section 2.4 and Fig. 23 for description of the pressure group
specific changes for each scenario.
NIVA 7562-2020
56
3.8 Assessing data coverage, uncertainty of data lay-
ers and sensitivity of the CEA model
The results for two evaluations are presented below: i) an initial estimate of
the data coverage of the ecosystem components and the overall uncertainty
of the data layers, and ii) a case study on the quantitative estimation of un-
certainty in the oxygen depletion data layer.
The ecosystem component layers form the basis of the cumulative effect as-
sessment. The assessment of data coverage shows the variation in the data
availability for ecosystem components within the Danish EEZ and gives a
measure of the quantity of ecosystem component data available for the CEA
assessment.
The initial uncertainty assessment shows the result of combining the differ-
ent uncertainty estimates, qualitative, categorical or quantitative, associated
with each ecosystem component and pressure data layer, to give a simple
measure of the quality of all data layers used in the CEA assesment.
The spatial coverage of the ecosystem component data layers was assessed
by calculating an index showing the fraction of ecosystem component layers
having data within a grid cell (Supplementary Material, Annex C6). It should
be noted that it is not an index of presence or absence of the ecosystem
component itself, rather an index of availability of data describing the eco-
system component. In this respect, data indicating the absence of an ecosys-
tem component is as valuable as data indicating its presence.
A high coverage index means that the ecosystem components are well as-
sessed within the area and a low index shows that it there is less infor-
mation forming the basis for the assessment. In the map it is seen that there
is a large difference between the eastern parts of the Danish EEZ (the Skag-
errak/Kattegat and the southwestern Baltic regions) and the North Sea re-
gion and east of Bornholm area, the latter having a lower data coverage.
The extensive coastal area along the west coast of Jutland is also very much
underrepresented by data on ecosystem components. The inner Danish wa-
ters have the best coverage of ecosystem component data layers. This can
can be explained by the boundary limits imposed by many of the models
used to describe species distributions (e.g. marine mammals and sea birds).
These limits are either based on the lack of in situ observations or defining
habitats not considered relevant for investigation for species. National
monitoring also tends to revisit the same spots to ensure high quality time
series. However, areas with few observations may be important for other
species.
The uncertainty of the assessment was estimated by aggregating the uncer-
tainty data of all the individual data layers (pressures and ecosystem compo-
nents), giving a spatially varying indicator of assessment uncertainty. The
spatial uncertainty is based on normalised data layers varying from 0 and 1,
where 0 represents the lowest uncertainty and 1 the highest. The results are
shown in Fig. 23. The variation in aggregated uncertainty follows the same
pattern as data coverage, with the best assessments found in the inner Dan-
ish waters.
The maps indicate that the robustness of the assessment varies spatially
when evaluating the results of the CEA models and scenarios. Areas with
low data coverage might be underestimated in the CEA models as data on
several ecosystem components are missing. In the same manner the uncer-
tainty map provides an indication of areas with less confidence on the distri-
bution of data layers and thus assessment result.
Overall, the uncertainty map reflects a well-known bias, where mapping and
monitoring in the inner Danish waters (i.e. Kattegat and western Baltic Sea,
the latter including the water around the island of Bornholm) and in the
coastal waters along the west coast of Jutland have, in general, a high qual-
ity and an adequate spatial coverage. In contrast, mapping and monitoring
of the Danish parts of the North Sea and Skagerrak have had relatively low
priority for decades, both in terms of spatial and temporal coverage.
NIVA 7562-2020
57
Figure 23: Mapping uncertainty for the ECOMAR data layers. Red indicates a higher level of uncertainty of the data layers and the yellow a lower uncertainty.
NIVA 7562-2020
58
3.8.1 Uncertainty and sensitivity of the CEA model
Like all models, the Halpern CEA model also has assumptions that may affect
the model results in different ways. To address the uncertainty and the sen-
sitivity of the CEA model, alternative model assumptions and data quality
problems were tested in two ways: i) Monte Carlo (MC) simulations based
on 1000 runs and ii) Elementary Effect model (EE) with 500 simulations con-
ducted according to the methods and method adjustments (of the EE
model) described in Stock & Micheli (2016) and Stock et al. (2018). In the
simulations for both of the tests seven different factors/model assumptions
were tested: F1) Missing pressure data, F2) Data quality/errors of sensitivity
scores, F3) Linearity/Ecological thresholds, F4) Reduced analysis resolution,
F5) Impact model (mean or sum of impacts), F6) Transformation type, and
F7) Multiple stressor effects model (Additive, Dominant pressure or Antago-
nistic model). The factors were randomly selected for each MC run or EE
simulation. The factors Linear decay and Reduced stressor resolution were
not tested due to the pre-transformations of the input layers.
The uncertainty of the CEA model results were tested by the 1000 MC runs,
and showed that the three pressure groups that were ranked in the top 3 as
being the potentially most important pressures affecting the ecosystem
components are: Commercial fishing, Pollution nutrients’, and Pollution
contaminants. The three pressures grops that consistently were ranked in
the bottom are: Aquaculture, Physical disturbance of the sea floor’, and
Recreational fishing and hunting (Table. 6).
The sensitivity of the model was tested by the EE model and it showed that
the three factors that had the most overall effect on the ranking of pres-
sures, denoted μ* in Stock & Micheli (2016), were F6, F1 and F2. The factors
that had the least impact on the pressure ranks were F4 and F5. The factors
that were most and least influenced by interactions with other factors and
random components in the model, denoted σ* in Stock & Micheli (2016),
were the same as for μ*.
Factors affecting the rank of the ecostytem components (μ*) were primarily
F7, F5 and F2. The interaction and stochasticity measure σ* was dominated
by F7 followed by F1 and F3. The least influential factors affecting the model
results (μ*) and interaction and stochastisity (σ*) for the ecosystem compo-
nents were F4 and F6.
The results show that the uncertainty of the ranking is low and that the
model results are robust. The MC simulated runs showed that the top three
and bottom three pressure groups corresponds exactly with the ranking of
the pressures for the baseline presented in Fig. 6. The areas that are consist-
ently under potential high and low CEA index were also robust showing the
same pattern as the spatial distribution of the CEA in Fig. 6. The EE model
results indicates that there is a combination of the assumptions in the fac-
tors affecting the model that influence the results and not just a single fac-
tor, although some factors have a larger influence than others. The ones
with large influence in our model are all known and some solution to ad-
dress those in coming models are presented in section 3.8.2, like e.g., which
transformation type to use on the data layers. There are also more sophisti-
cated ways to conduct the expert interviews that can be used for collecting
the sensitivity scores (see e.g. Doubleday et al. 2017, Gissi et al. 2017 and
Jones et al. 2018), which can be considered in other studies. One of the
most influential factors was Missing pressure data’.
Table 6: Results of the rankings of pressure groups from the the model as-
sumptions and factors applied randomly in the 1000 Monte Carlo simula-
tions. Adopted from Stock & Micheli (2016). For detailed descriptions see
Stock & Micheli (2016) and Stock et al. (2018).
Pressure group
Top 3
Bottom 3
Commercial fishing
89.1%
0%
Pollution - nutrients
66.3%
0%
Pollution - contaminants
62.1%
0%
Aquaculture
0%
99.9%
Physical disturbance of the sea floor
0%
97.5%
Recreational fishing and hunting
0%
75.3%
NIVA 7562-2020
59
3.8.2 Uncertainty of oxygen depletion maps
The current approach to describe the areal extent of oxygen depletion in the
Danish Straits does not consider the uncertainty or confidence associated
with the maps. However, this uncertainty can be assessed by estimating the
variograms for the depth location of the oxyclines, computed from oxygen
profiles in a number of discrete monitoring positions scattered around the
spatial domain. The variogram describes the variation between two points in
space as function of their inter-distance. The empirical variograms for all the
years (2002-2014) showed the same tendency of increasing variance up to
around 100 km, where it reached a maximum plateau (Fig. 24). An exponen-
tial variogram model was found suitable to describe these empirical vario-
grams, with a nugget parameter of 0.05 (intercept), a scale parameter of
0.35 and range parameter of 40 km. The nugget effect of 0.05 indicates that
the estimated depth location of the oxycline itself was associated with un-
certainty (±25%). Although the uncertainty of the oxygen sensor is probably
less, the measured profile is essentially a single snapshot that should repre-
sent a period of approximately 10 days. Furthermore, changes in water lev-
els may also add to this observation error in the pycnocline depth. The
maximum variance (0.4; nugget+scale) was attained at distances beyond
120 km, implying that no spatial correlation remains for profiles that are
more than 120 km apart and that the prediction uncertainty of a grid point
more than 120 km away from the nearest measurement point is approxi-
mately ±88%.
The current algorithm for mapping the oxygen depletion uses ordinary
kriging with a linear variogram without a nugget effect, which implies that
the mean field has a trend (Fig. 25A). Using the exponential variogram
model (Fig. 24) produces a similar mean field by ordinary kriging (Fig. 25B),
although it should be noted that there is no trend with this model, i.e. pre-
dictions at far distances will all tend to the same constant value (“global
mean”). The relative uncertainty of the predictions strongly depended on
distances to monitoring points, ranging from ±25% near monitoring stations
(nugget effect) up to ±88% when all monitoring points were more than 120
km away (Fig. 25C). Using the prediction error of the ordinary kriging to cal-
culate a 95% confidence interval for the oxycline depth and combining these
with the bathymetry provides upper and lower confidence maps for the ar-
eal extent of oxygen depletion (Fig. 25D).
In September 2005, the estimated area and volume with hypoxia (ordinary
kriging with exponential variogram) were 493 km2 and 2.2 km3, respectively
(Fig. 26). However, the confidence intervals were extremely broad, ranging
from 0.6 to 19400 km2 for the areal extent. This high degree of uncertainty
would lead to quite different interpretations on the spread of hypoxia in the
Danish Straits, whether you consider the lower, mean or upper extent (Fig.
25D). The reason for broad confidence intervals is because the ordinary
kriging approach does not attempt to model the underlying mean field, and
therefore overestimates the magnitude of the random variation.
Figure 24: Empirical variograms estimated for each year in the study period
and the fitted exponential variogram model with parameters inserted.
NIVA 7562-2020
60
Panel A
Panel B
Panel C
Panel D
Figure 25: The mean field for the oxycline depth using a linear variogram model (panel A) and the estimated exponential variogram model (panel B). The
relative uncertainty of the oxycline depth from the exponential variogram model (panel C). The resulting 95% confidence mappings of oxygen depletion
(panel D). Monitoring stations used for spatial interpolations are shown as green dots. The maps represent September 2005.
NIVA 7562-2020
61
Panel A
Panel B
Figure 26: Interannual variation in the area and volume of bottom water
with oxygen concentration below 2 mg L-1. Data represent September.
Estimating the mean field of the oxycline depth for all years combined with
the GAM produces slightly smoother contours (Fig. 29) compared to those
obtained with ordinary kriging (Fig. 25A+B), because the spline is generally
smoother by nature and because it is based on data from 13 years.
The residuals from the GAM exhibited substantially lower variation (Fig. 28)
than the constant mean field model employed in ordinary kriging (Fig. 24).
The nugget effect corresponded to an “observation error” of ±33% and
Figure 27: The mean field estimated by a GAM with a 2-dimensional spline
and a year factor based on all oxycline depths from monitoring stations
(2002-2014).
spatial correlation only remained at distances less than 120 km. Obviously,
variations interpreted as random correlated variability by ordinary kriging
was converted into fixed variability by employing the GAM to describe the
mean field of the pycnocline depth.
NIVA 7562-2020
62
Figure 28: Empirical variograms estimated on residuals from the GAM. The
fitted exponential variogram model is shown with a solid black line and pa-
rameters inserted.
Although the predicted oxycline depth from the combined approach (Fig.
30A) did not appear much different from neither the ordinary kriging (Fig.
25B) nor the GAM mean field (Fig. 27), the predictions changed up to ±40%
from the GAM mean field due to the ordinary kriging adaptation for the spe-
cific year (Fig. 30B). In September 2005, the oxycline was relatively higher in
the water column in the western Kattegat, the Great Belt and the Sound,
whereas it was relatively deeper in the northern Little Belt, Fehmarn Belt
and Arkona Sea. The relative uncertainty of the predictions from the com-
bined approach varied with distances to the nearest monitoring station,
however noteworthy, with considerably lower uncertainty (Fig. 30C). Finally,
the confidence mapping of the oxygen depletion was also less variable, dis-
playing a more realistic range for oxygen depleted areas (Fig. 30D).
The reduced random variability was also apparent in the assessment of areal
extent and volume of hypoxia (Fig. 29). According to the combined ap-
proach, the predicted areal extent of hypoxia was 836 km2 in September
2005 with a 95% confidence interval of 208-6343 km2.
The uncertainty of the oxygen depletion maps of individual years is still con-
siderable, but this uncertainty is reduced if we consider the mean oxygen
depletion in September for the entire period (2002-2014). The average oxy-
cline for the 13 years (not shown) was similar to the mean field estimated
with the GAM approach (Fig. 27), but the relative uncertainty was generally
low, ranging up to 8.3% (Fig. 30E). The reduced uncertainty also had implica-
tions for the confidence interval of the oxygen depletion maps, which
showed less variability for the mean oxygen depletion (Fig. 30F).
This example demonstrates that the standard ordinary kriging overestimates
the uncertainty when predicting oxygen depletion maps. However, the un-
certainty can be reduced, and narrower confidence intervals produced by in-
corporating more of the variability in the data into a model for the mean
spatial field. It is possible that continued efforts to improve the mean field
model may reduce the random variation even further, which will constrain
the estimates of oxygen depletion area and volume.
Panel A
Panel B
Figure 29: Interannual variation in the area (A) and volume (B) of bottom
water with oxygen concentration below 2 mg L-1 from the combined ap-
proach (GAM and ordinary kriging). Data represent September.
NIVA 7562-2020
63
Panel A
Pane lB
Panel C
Panel D
Panel E
Panel F
Figure 30: The estimated oxycline depth using GAM and ordinary kriging (A) and the relative change of this to the GAM mean field (cf. Fig. 27) (B). The rela-
tive uncertainty of the oxycline depth from the exponential variogram model (cf. Fig. 25) (C). The resulting 95% confidence mappings of oxygen depletion (D).
Maps in A-D represent September 2005. The relative uncertainty of the mean oxycline depth (E) and the 95% confidence mappings of oxygen depletion (F).
Maps in E and F represent September for the years 2002-2014. Monitoring stations used for spatial interpolations are shown as green dots.
NIVA 7562-2020
64
3.9 Zoning steps toward a roadmap
The first step toward zoning in Danish marine waters has been a baseline
mapping of existing activities and pressures. This mapping is probably as
good as it gets, given the spatial coverage of monitoring and mapping activi-
ties. However, coexistence is not the same as a lack of conflict, but rather
the lack of principles for marine spatial planning. Hence, ECOMAR has car-
ried out a criteria-based evaluation of potential conflicts between activites
and subsequently mapped conflict zones (Fig. 31, page 69). Most of these
are to be considered hard conflicts and detrimental to both activities for
example the dumping of dredged material nearby an aquaculture plant or
dredging of sand within an important fishing area. Although the analysis is
preliminary, it is also an eye-opener in a MSP context, and can be used to
guide decision making on how to avoid or reduce the number of conflicting
activities within an area. To improve the values of these analyses, a revision
of the conflict criteria and considerations of not only spatial but also tem-
poral conflicts is recommended.
An important step toward production of a national zoning plan is the define-
tion of different types of zones ranging from a general use zone, where a lot
of activities can take place without any specific permit, over various inter-
mediate zones to a top zone type, where human activities are reducted to a
minimum. For ECOMAR, we have opted for four types of zones: zone 1 is a
‘general use zone’, zone 2 is a ‘targeted management zone’, zone 3 is an ‘ex-
clusive use zone’ and zone 4 is a ‘restricted access zone’. This approach is
adopted from Ekebom et al. (2007). A so-called zoning matrix, where human
activites are allocated to the four different use zones, together with indica-
tion whether a permit is required or not, has tentatively been developed
(Table 7). Such matrix enable a data-driven approach to zoning but further
consideration regarding the number of zone types and which activites al-
lowed to take place in the different zones is beyond the scope of ECOMAR.
Finally, we have ventured into creating a provisional ecological vulnerability
map (Fig. 32, page 70) which shows spatial variations in the vulnerability of
the ecosystem components in Danish marine waters. This vulnerability map
is calculated by EcoImpactMapper tool, based on the weighted mean sensi-
tivity scores of the ecosystem components present in each model point. The
areas tentatively identified as vulnerable should be regarded not only as
ecological ‘hot spots’ with high number of species, but as areas, where the
ecosystem components tend to be more sensitive to human pressures.
NIVA 7562-2020
65
Table 7: Zoning matrix for the Danish EEZ. Human activities are allocated to the four zones suggested, i.e. Zone 1: General Use, Zone 2: Targeted Manage-
ment, Zone 3: Exclusive Use, and Zone 4: Restricted Access. Based on Ekebom et al. (2008), but adapted according to ECOMAR data layers.
Human activities
Impacts
Zones
1: Physical loss
2: Physical damage
3: Other physical disturbance
4: Interference with hydrolog-
ical processes
5: Contamination by hazard-
ous substances
6: Systematic and/or inten-
tional release of substances
7: Nutrient and organic mat-
ter enrichment
8: Biological disturbance
Zone 1:
General Use
Zone 2:
Targeted Management
Zone 3:
Exclusive Use
Zone 4:
Restricted Access
Land-sea interactions
UWWTP discharge (point
source)**
-
-
x
-
x
x
x
x
Permit
Permit,
if no conflict
NO, except when in agree-
ment (then YES or Restricted)
NO
Discharges from industries***
-
-
-
-
x
x
x
-
Permit
Permit,
if no conflict
NO, except when in agree-
ment (then YES or Restricted)
NO
Pollution - Contaminants
Dumped chemical munitions**
-
x
x
-
x
-
-
-
NO
NO
Contracted if this is the exclu-
sive use
NO, Contracted if this is the
exclusive use
Maritime traffic/ Shipping and transportation
Large vessel traffic*
-
x
x
-
x
-
x
x
YES
YES,
if no conflict
NO or
Restricted
NO
Small vessel traffic*
-
x
x
-
x
-
x
x
YES
YES,
if no conflict
Restricted
NO
Industrial ports**
x
x
x
-
x
x
-
x
YES
YES,
if no conflict
NO, except when in agree-
ment (then YES or Restricted)
NO
Harbours**
x
x
x
-
x
x
-
x
YES
YES,
if no conflict
NO, except when in agree-
ment (then YES or Restricted)
NO
Seaplane traffic***
-
-
-
-
x
-
x
-
YES
YES,
if no conflict
NO, except when in agree-
ment (then YES or Restricted)
NO
Industry, energy and infrastructure
Coastal habitat modification**
x
x
-
x
-
-
-
-
Permit
Restricted + Permit
NO, except when part of the
exclusive use (Permit)
NO, unless part of the agreed
use (Permit)
Jetties, breakwaters*
x
x
-
x
-
-
-
-
YES
YES
NO, except when part of the
exclusive use (Permit)
NO, unless part of the agreed
use (Permit)
NIVA 7562-2020
66
Human activities
Impacts
Zones
Shoreline buildings*
x
x
-
x
-
-
-
-
Permit
Restricted + Permit
NO, except when part of the
exclusive use (Permit)
NO
Bridges & coastal constructions**
x
x
-
x
-
-
-
-
Permit
Restricted + Permit
NO, except when part of the
exclusive use (Permit)
NO
Dredging**
x
x
-
-
x
-
-
-
Permit
Restricted + Permit
NO, except when part of the
exclusive use (Permit)
NO
Disposal sites for construction,
garbage & dredged material**
x
x
-
x
x
-
-
-
Permit
Restricted + Permit
NO, except when part of the
exclusive use (Permit)
NO
Extraction of material from the
seafloor**
x
x
-
-
x
-
-
-
Permit
Restricted + Permit
NO
NO
Mining***
x
x
-
-
x
-
-
-
Permit
Restricted + Permit
NO, except when part of the
exclusive use (Permit)
NO
Offshore oil & gas installations**
x
x
x
-
x
x
-
-
Permit
Restricted + Permit
NO, except when part of the
exclusive use (Permit)
NO
Oil and gas pipelines**
-
x
-
-
x
-
-
-
Permit
Permit
NO, except when part of the
exclusive use (Permit)
NO
Power and heat plants**
x
x
-
x
-
-
-
-
Permit
Restricted + Permit
NO, except when part of the
exclusive use (Permit)
NO
Wind farms**
x
x
x
-
-
-
-
-
Permit
Restricted + Permit
NO, except when part of the
exclusive use (Permit)
NO
Sea cables**
-
x
-
-
-
-
-
-
Permit
Permit
NO, except when part of the
exclusive use (Permit)
NO
Lighthouses**
x
x
-
-
-
-
-
-
YES
YES
YES
YES, but can be restricted
Other nautical support struc-
tures***
x
x
-
x
-
-
-
-
YES
YES
YES
YES, but can be restricted
Selective extraction of species - Commercial and recrestional fishing, recreational hunting
Bottom trawling (small mesh
size)**
-
x
x
-
-
-
-
x
Permit
Permit, if no conflict
NO, except when in agree-
ment (then YES or Restricted)
NO
Bottom trawling (large mesh
size)**
-
x
x
-
-
-
-
x
Permit
Permit, if no conflict
NO, except when in agree-
ment (then YES or Restricted)
NO
Pelagic trawl**
-
-
x
-
-
-
-
x
Permit
Permit, if no conflict
NO, except when in agree-
ment (then YES or Restricted)
NO
Longlines**
-
x
x
-
-
-
-
x
Permit
Permit, if no conflict
NO, except when in agree-
ment (then YES or Restricted)
NO
Set gillnets**
-
x
x
-
-
-
-
x
Permit
Permit, if no conflict
NO, except when in agree-
ment (then YES or Restricted)
NO
Driftnet fishing***
-
-
x
-
-
-
-
x
Permit
Permit, if no conflict
NO, except when in agree-
ment (then YES or Restricted)
NO
NIVA 7562-2020
67
Human activities
Impacts
Zones
Mussel dredging**
-
x
x
-
-
-
-
x
Permit
Permit, if no conflict
NO, except when in agree-
ment (then YES or Restricted)
NO
Recreational net fishing*
-
x
x
-
-
-
-
x
YES + Permit
YES + Permit
NO, except when in agree-
ment (then YES or Restricted)
NO
Recreational angling/spinning*
-
x
x
-
-
-
-
x
YES + Permit
YES
NO, except when in agree-
ment (then YES or Restricted)
NO
Recreational fishing tour boats*
-
x
x
-
-
-
-
x
YES + Permit
YES + Permit
NO, except when in agree-
ment (then YES or Restricted)
NO
Bird recreational hunting**
-
-
-
-
-
-
-
x
Permit
Permit, if no conflict
NO, except when in agree-
ment (then YES or Restricted)
NO
Seal hunting***
-
-
-
-
-
-
-
x
NO
NO
NO
NO
Aquaculture
Fish farms**
-
x
-
x
x
x
x
x
NO
NO
Permit
NO
Shellfish farms**
-
x
-
x
x
x
x
x
NO
NO
Permit
NO
Military activities
Military practice areas*
x
x
x
-
x
x
-
-
NO
NO
Contracted if this is the exclu-
sive use
Contracted, if it is the reason
for protection
Military base areas*
x
x
x
-
x
x
-
-
NO
NO
Contracted if this is the exclu-
sive use
Contracted, if it is the reason
for protection
Recreation and leisure activities
Kayaking*
-
-
x
-
-
-
-
-
YES
YES, if no conflict
YES, unless in disagreement
with the exclusive use (then
NO or Restricted)
NO
Surfing*
-
-
x
-
-
-
-
-
YES
YES, if no conflict
YES, unless in disagreement
with the exclusive use (then
NO or Restricted)
NO
Rowing*
-
x
x
-
-
-
-
-
YES
YES, if no conflict
YES, unless in disagreement
with the exclusive use (then
NO or Restricted)
NO
Walking, visiting*
-
-
x
-
-
-
-
-
YES
YES, if no conflict
YES, unless in disagreement
with the exclusive use (then
NO or Restricted)
NO
Observing nature*
-
-
x
-
-
-
-
-
YES
YES, if no conflict
YES, unless in disagreement
with the exclusive use (then
NO or Restricted)
NO
Collecting*
-
x
x
-
-
-
-
x
YES
YES, if no conflict
YES, unless in disagreement
with the exclusive use (then
NO or Restricted)
NO
NIVA 7562-2020
68
Human activities
Impacts
Zones
Swimming/Bathing*
-
-
x
-
-
-
x
-
YES
YES, if no conflict
YES, unless in disagreement
with the exclusive use (then
NO or Restricted)
NO
Bathing sites**
x
x
x
-
-
-
x
-
YES
YES, if no conflict
YES, unless in disagreement
with the exclusive use (then
NO or Restricted)
NO
Coastal recreation sites**
x
-
x
-
-
-
x
-
YES
YES, if no conflict
YES, unless in disagreement
with the exclusive use (then
NO or Restricted)
NO
Boating**
-
x
x
-
x
-
-
-
YES
YES, if no conflict
YES, unless in disagreement
with the exclusive use (then
NO or Restricted)
NO
Jetskiing*
-
x
x
-
x
-
-
-
YES
YES, if no conflict
YES, unless in disagreement
with the exclusive use (then
NO or Restricted)
NO
Scuba diving**
-
x
x
-
x
-
-
x
YES
YES, if no conflict
YES, unless in disagreement
with the exclusive use (then
NO or Restricted)
NO
Other
Research***
x
x
x
x
x
-
-
x
YES
YES if no conflict, NO/Permit if
conflict
YES if no conflict, NO/Permit if
conflict
NO, if not contracted
Public arrangements (museum,
guided tours)*
-
x
x
-
-
-
-
x
YES
YES if no conflict, NO/Permit if
conflict
YES if no conflict, NO/Permit if
conflict
NO, if not contracted
Upsteream management activi-
ties***
-
x
x
x
x
x
x
x
YES
YES if no conflict, NO/Permit if
conflict
YES if no conflict, NO/Permit if
conflict
YES, if no conflict,
NO if conflict
Marine protection
Natura 2000 sites, HD*
-
-
-
-
-
-
-
-
NO
YES
YES
YES
Natura 2000 sites, BD*
-
-
-
-
-
-
-
-
NO
YES
YES
YES
Nature 2000 sites, Ramsar sites*
-
-
-
-
-
-
-
-
NO
YES
YES
YES
MSFD MPA*
-
-
-
-
-
-
-
-
NO
NO
YES
YES
Stone reefs within N2000 areas**
-
-
-
-
-
-
-
-
NO
NO
YES
YES
* Part of an ECOMAR data layer; ** A fully-fledged ECOMAR data layer, *** Not included in ECOMAR.
NIVA 7562-2020
69
Figure 31: Current conflict zones identified on the basis of provisional conflict criteria. Red areas indicates the highest 10th percent quantile and where there
are most potential conflicts. Blue areas indicate low potential for conflicts between pressures and activities.
The map of conflict zones is tentative and should be seen as a demonstra-
tion of the power of data and how data can support the development of evi-
dence-based MSP. If implemented, conflict criteria should be critically reas-
sessed and adapted for the marine region or sub-region in question. If so,
a criteria-based identification of conflicts and subsequent analysis of how to
avoid conflicts would potentially lead to few conflict, lower combined pres-
sures and finally improvement of ecological status in coastal waters (WFD
domain) and environmental status in offshore waters (MSFD domain)
NIVA 7562-2020
70
Figure 32: Provisional vulnerability mapping of the Danish marine waters. Red areas indicates the highest 10th percent quantile and where the ecosystem
components have the highest mean vulnerability and the blue areas indicate where there is a lower mean vulnerability.
The vulnerability map shows that specific areas, especially coastal, are more
prone to disturbance than most offshore areas. Given the relationship be-
tween biodiversity status (Andersen et al. 2015) and ecological status (EEA
2019a) on the one hand and human pressures on the other, the addition of
more human pressure is not in accordance with the overall visions and goals
of the MSFD and the MSPD.
NIVA 7562-2020
71
4 Discussion and conclusions
Based on the analyses and scenarios presented in Chapter 3, we discuss these as well as the strengths, weakness and potential implications
of the results. Special focus is put on the combined effects analyses and on the 2030, 2050 and MSFD scenarios, since these indicate that
pressures may neither be reduced in a short term (2030) or long-term perspective (2050). We also conclude that a better coordination be-
tween ‘marine’ authorities is needed and that it is imperative to make use of data-driven approaches, to consider uncertainty and to con-
sider all activities and pressures, including land-based, and to include all ecosystem components.
4.1 Data layers
The data sets and maps compiled by ECOMAR represent a leap forward to-
wards a more complete spatial analysis based on more extensive data sets,
especially with regard to the activity and pressure layers. All sectors, activi-
ties, and pressures mentioned in the MSPD, MSFD and WFD are covered,
and are supplemented with data representing climate change, recreational
activities and tourism. The data sets and maps representing ecosystem com-
ponents cover all key ecosystem groups, e.g. pelagic habitats, benthic habi-
tats, fish, birds and marine mammals. In addition, data layers representing
societal interests, i.e. areas important for recreation, shipwrecks and arche-
ological sites (e.g. stone age dwellings), are included and analysed. Although
ECOMAR has made significant progress and developed comprehensive
state-of-the-art data sets (see Supplementary material, Annex A and B), we
have also identified weaknesses in some of the data sets with room for im-
provement.
The following themes are of key relevance, when it comes to improving the
data layers, especially those related to ecosystem components: i) spatial
coverage (all data sets and model should cover the entire Danish EEZ), ii)
confidence (data sets with low confidence should be improved and made
more accurate and precise), and iii) updating (all data should be up-to-date).
This important task is evidently an assignment for relevant competent au-
thorities, i.e. the Danish Maritime Agency (regarding the MSPD) and the
Danish Ministry for the Enviroment (regarding the MSFD and WFD).
4.2 Mapping of combined effects
The estimation and mapping of potential combined effects of multiple hu-
man pressures in Danish marine waters is based on a well-documented (Hal-
pern et al. 2008, Stock 2013) and widely used method (Korpinen & Andersen
2016, HELCOM 2018, Miljø- og Fødevareministeriet 2019, EEA 2020).
Given the data availability in ECOMAR, we have been able to demonstrate
the power of combining data and demonstrate how CEA can support ecosys-
tem-based management in the context of multiple directives, e.g. MSPD,
MSFD and WFD. A key result is the map of potential combined effects of hu-
man pressures in Danish marine waters (see Fig. 6). It significantly advances
the performance of previous studies by Miljø- og Fødevareministeriet
(2019), HELCOM (2018), Andersen et al. (2020), EEA (2020) and the Swedish
Agency for Water Management (Hammar et al. 2018 and 2020).
The CEA method is admittedly simple but it provides a good overview of
where hot spots are located and also enables a broad range of analyses.
Weaknesses of the CEA method, e.g. the setting of sensitivity scores and ef-
fect distances as well as the assumption of linear responses between the
pressure and ecosystem components, normalization of data layers and the
assumed additive nature of the pressures are well known (Halpern & Fujita
2013). Other methods for assessing cumulative impacts exist and many of
them include some of these assumptions whereas others, such as mechanis-
tic, Bayesian models or marchine learning methods, can include both inter-
NIVA 7562-2020
72
actions and non-linear reponses between pressures and ecosystem compo-
nents (Petersen et al. 2020). These other types of models are usually heavy
to run and can demand a high quality of quantified relationships between
the data layers. The models are therefore not an alternative and useful for
MSP applications, but can be used on smaller scales or to a limited amount
of data layers. As in the CEA model used by ECOMAR, they can also include
expert knowledge when data or knowledge about the relationships are miss-
ing, which can in some situations be problematic (Halpern & Fujita 2013).
There are other ways of including an uncertainty estimate of the sensitivity
scores gathered from expert studies described by e.g. Doubleday et al.
(2017), Gissi et al. (2017) and Jones et al. (2018). Unfortunately, this com-
prehensive type of study was beyond the scope of the ECOMAR project, as
the setting of the sensitivity scores was planned for the early phase of ECO-
MAR. An ad hoc solution for testing of alternative model assumptions and as
well as data quality problems are presented by Stock & Micheli (2016) and
Sock et al. (2018) and have been applied in several previous studies, like An-
dersen et al. (2020) and also within the ECOMAR project (see section 3.8.1).
Although the various assumptions can affect the model results, it has also
been shown that they are robust (Stock & Micheli 2016, Stock et al. 2018,
Andersen et al. 2020). This is also the case for ECOMAR as seen from the re-
sults of the Monte Carlo simulations and the elementary effects model de-
scribed in section 3.8.1. These model results show that ECOMAR estimates
are robust and the factors affecting the model output are known restrictions
and in line with previous studies. The results from ECOMAR can therefore
with confidence be used in a Danish MSP context but can also be relevant to
others, e.g. neighbouring countries as well as HELCOM and OSPAR.
4.3 Confidence and uncertainty
At the beginning of ECOMAR, data layer providers were asked to deliver an
uncertainty estimate with their data layer. Due to the diversity of different
approaches and the lack of a general framework for assessing uncertainty, a
multitude of different uncertainty layers were produced. The experiences
gained from ECOMAR on assessing uncertainty in data layers will be useful
for further development of tools such as EcoImpactMapper and SeaSketch.
For aggregating data layers and their associated uncertainties, it is impor-
tant that these use the same currency before aggregation. This can be
achieved through transformation to a common scale, as was done for the
data layers in EcoImpactMapper. Essentially, the same transformation can
be applied to the uncertainty of the data layers, if the uncertainty is repre-
sented on the same scale as the data layer, i.e. as absolute uncertainty of
the value in the grid cell. In case the uncertainty is expressed relative to the
value of the grid cell, the absolute uncertainty can be computed by scaling
with the grid cell value. However, uncertainty expressed qualitatively using
different confidence classes or using a scale (e.g. 0-100) that is not directly
related to the grid cell value do not conform to the “common currency” ap-
proach that enables aggregation of uncertainties in a similar manner to ag-
gregation of data layers. Consequently, there is a strong need for harmoniz-
ing the reporting of uncertainty in absolute values, using the same unit as
the data layer itself. Therefore, a first key lesson from ECOMAR is that un-
certainty reporting should be done in the same unit as the data layer.
Quantification of uncertainty in data layers is a difficult task and therefore it
is often overlooked or sometimes even ignored. ECOMAR is one of the first
projects to emphasize confidence assessments in marine spatial planning,
which has led to the realisation that a general statistical method should be
developed to quantify the uncertainty. ECOMAR has contributed to this by
demonstrating the usefulness of combining two commonly used statistical
methods for obtaining realistic uncertainty estimates. Therefore, a second
key lesson from ECOMAR is that statistical methods, capable of quantifying
uncertainty, should be employed generically.
However, the background data for compiling data layers are associated with
multiple sources of uncertainty (e.g. observational, model, parameter er-
rors), which are not all relevant for expressing the uncertainty of the data
layer. Whereas model and parameter errors express uncertainty with the
approach used for compiling the data layer and should be included, observa-
tional errors pertain to uncertainty associated with the measurements and
thus not relevant for the data layer.
NIVA 7562-2020
73
Furthermore, data layers are mostly dynamic and should, preferably, repre-
sent the same assessment period before aggregation, but in reality they are
more likely based on data from different years. This adds another source of
variability; temporal variation, which should be quantified and included in
the data layer uncertainty, provided that the data layer is not based on data
from all years of the assessment period. For example, if the data layer is
compiled on a single year’s survey of a data type exhibiting larger interan-
nual variability, then that data layer is an uncertain representation of a
longer assessment period. Therefore, a third key lesson from ECOMAR is
that a standard assessment period should be chosen and the multiple sour-
ces of uncertainty affecting the data layer estimate representing that period
should be quantified and included in the uncertainty of the data layer.
4.4 Analysis and scenarios
With ECOMAR, it has been demonstrated that reductions or increases in in-
dividual pressure intensity also affects the intensity of the combined pres-
sure and ultimately on the ecosystem components. All pressure groups have
been addressed and