recoca final report - bonus portal

1

RECOCA Reduction of Baltic Sea Nutrient Inputs and Cost Allocation within the

Baltic Sea Catchment

Final Report January 2012

Covering 1 January 2009 to 31 December 2011

Co-ordinator: Professor Fredrik Wulff

Stockholm University

SE-10691 Stockholm

Sweden

Tel. +46707310026

Email: [email protected]

2

Project coordinator:

Fredrik Wulff, Baltic Nest Institute, Stockholm University, Sweden, E-mail: [email protected]

List of Principal Scientists:

New catchment database and nutrient accounting tool to estimate nutrient loading

Christoph Humborg, Department of Applied Environmental Science and Baltic Nest Institute, Stockholm University, Sweden, E-mail: [email protected]

Simulations of nutrient pathways and effects of mitigation measures for the Baltic Sea

Hans-Estrup Andersen, Department of Biological Science, Århus University, Denmark, E-mail: [email protected]

Carl-Magnus Mörth, Department of Geological Sciences, Stockholm University, Sweden, E-mail: [email protected]

Per Stålnacke, Norwegian Institute for Agricultural and Environmental Research (Bioforsk), Ås, Norway, E-mail: per.stå[email protected]

New Cost minimization models

Berit Hasler, Department of Policy Analysis, Århus University, Denmark, E-mail: [email protected]

Tomasz Zylicz, Warsaw Ecological Economics Center, Warsaw University. Poland, E-mail: [email protected]

3

1. Executive summary Implementation of the eutrophication section within the Baltic Sea Action Plan will require modelling tools to simulate the effects of various abatement strategies, allocate nutrient reduction to countries, and estimate related costs. RECOCA has created an array of hierarchically connected data bases and models that will allow managers and decision makers to view scenario analyses via the decision support system Nest. The databases, catchment models and cost minimization models have been closely integrated, assuring for the first time realistic estimates of anticipated nutrient reductions and their costs for each of the seven sub-basins.

We have created various model tools that allow users to calculate the effect of mitigation measures, view scenario analyses and .produce maps showing hot spots of nutrient leakage. To estimate cost-effective ways to reach nutrient reduction goals, RECOCA has developed economic models, which account not only for the costs at the sources, but also the total costs of various abatement measures in the individual riparian countries. These models take into account the natural conditions and reduction capacities in the different watersheds and the net effect of measures for the Baltic Sea.

1.1 Overview RECOCA is a clearly interdisciplinary research project featuring scientists from nine groups in five countries, specialising in e.g. ecology, environmental economics, hydrology, agronomy and biogeochemistry.

The implementation of the eutrophication segment within BSAP of HELCOM, with the strategic goal of a “Baltic Sea unaffected by eutrophication”, requires modelling tools in order to simulate the effects of various abatement strategies and to estimate the related costs.

The overall objectives of RECOCA are to

• Simulate possible future riverine nutrient loads to the Baltic Sea • Estimate cost effective reductions of these loads and corresponding improvements in

ecological indicators, and • Suggest cost allocation schemes for the riparian countries.

1.2 Key Research and Key Results

1.2.1 New catchment data base and nutrient accounting tool to estimate nutrient loading RECOCA scientists have compiled data on land use patterns and levels of economic activities. These gridded data are now available via the Nest decision support system (Figure 1). Data originate from many sources, including the EU Joint Research Centre (fertilizer use, crop types), EUROSTAT (livestock data), HYDE database (population), CORINE (land cover) and SMHI (hydrological and climate forcing). All these data have been compiled for 117 watersheds (82 major watersheds and 35 coastal areas) as well as for 8 type watersheds. The compiled data are organized into fertilizer use, atmospheric deposition, biological N-fixation, crops, animals and population distributions. RECOCA then calculates food and feed budgets for all watershed types using these data. The resulting nutrient budgets for all

4

watershed types are then addressed in RECOCA. Moreover we created a detailed list of point source data, i.e. statistics on municipal wastewater treatment including estimates on people connected and unconnected to sewage systems.

The data on land use patterns were organized to allow decision makers to be informed of all major drivers of nutrient emissions and estimate the total nutrient loading to the coast from a watershed. Anthropogenic nutrient loadings are determined by the use of fertilizers, atmospheric deposition, biological fixation and net feed and food import or export i.e. the difference between the nutrient requirements of humans and livestock and the crops and animal production in a watershed. A ”nutrient accounting toolbox” quantifies the individual drivers of nutrient loads and this toolbox forms the basis for the dynamic catchment simulation model soon to be available via the Nest decision support system.

Figure 1: The new catchment database accessible via the Nest system (www.balticnest.org)

Nutrient fluxes to the Baltic Sea will increase gradually if the man-made nutrient loading to the landscape increases. Synthetic fertilizers are the main sources of nutrient pollution. Other sources - such as nutrients in food and animal feed or atmospheric deposition - contribute and are dominant for some sub-regions.

5

1.2.2 Simulations of nutrient pathways and effects of mitigation measures for the Baltic Sea

RECOCA produced an array of dynamic simulation models to allow future users of the Nest decision support system to quantify the effects of mitigation measures or to undertake scenario analyses on likely trends. Mitigation measures simulated are e.g. changes in fertilizer use, livestock density and atmospheric deposition, the creation of wetlands and improved sewage treatment. However, to quantify the effects of these measures there is a need to describe the load reductions at the various sources, as well as the pathways and removal of nutrients in soils, groundwater, lakes and streams. Large amount of nutrients are retained in the drainage basins and do not reach the Baltic Sea and accurate estimates of nutrient retentions are essential for allocation of cost effective measures.

The dynamic models produced in RECOCA simulate

• Nutrient leakage from agricultural soils and point sources, • The natural removal of nutrients through pathways in the watersheds, and finally • The net effects of mitigation measures observed as changes in nutrient loads at the

river mouths and thus relevant for the Baltic Sea ecosystem.

Diffuse riverine leakage from agriculture is today the most important nutrient source to the Baltic Sea. Effluents from humans are decreasing and continue to do so by improved sewage treatment. The anticipated intensification of agricultural production in central and eastern European states (transitional countries) may however lead to increased nutrient leakages, counteracting improved sewage treatment. The ability to quantify nutrient leakages from different production systems and to identify regional hot spots is crucial when selecting abating measures of these detrimental inputs to the Baltic Sea. The strong heterogeneity in farm size and production intensity within the watersheds has been described by three different management strategies, which were parameterized individually for each riparian country and calibrated to national statistics with regard to consumption of fertilizer and manure. Nutrient leakages from agricultural land were calculated from the distributed management and land use data using a state of the art soil-vegetation-atmosphere model. The natural removal of nutrients along the flow path through groundwater and surface waters determines the cost efficiency of mitigation measures. Overall nutrient removal in ground and surface waters are calculated for the entire Baltic drainage basin, subdivided into 103 catchments.

The net effects of mitigation measures on nutrient loads reaching the river mouths, and thus relevant for the Baltic Sea ecosystem, is simulated using an updated catchment simulation model in Nest (CSIM). This dynamic model integrates the databases and accounting tools to predict leakage and nutrient removal along nutrient transport pathways through the watersheds and can thus be used for management. This model system is currently being used in the revision of the BSAP of HELCOM by estimating the effect of country-level compliance with the EU Waste Water Treatment Directive. This is as an essential step in the calculation of a country allocation scheme for nutrient abatement. The model shows that high

6

natural nutrient retentions within the watersheds will substantially reduce the effects on the Baltic Sea ecosystem from upstream measures. The model has already been used to simulate changes in fertilizer use. Fig 2b shows a scenario analysis of the effects of a likely increase in fertilizer use in transitional countries to levels comparable with Germany, Sweden or Denmark. Results suggest that in some watersheds the nutrient load to the Baltic Sea may increase significantly.

Figure 2. Hot spots of nutrient leakage from agriculture (left panel) and simulations of increased fertilizer use in transitional countries (right panel).

1.2.3 New Cost minimization models Two cost-minimisation models have been developed to assess the selection and distribution of cost-effective nutrient reductions to the Baltic Sea, e.g. by fulfilling the BSAP targets.

The BALTCOST model, developed jointly through RECOCA and the Baltic Nest Institute, is well suited for scenario modelling of cost-effective combinations of abatement measures to fulfil nutrient load reduction targets for the countries bordering the different Baltic Sea regions (7 regions in all). The model BALTCOST, when combined with marine

Dramatic increases in fertilizer use and manure are likely to occur in exactly those areas where nutrient leakage already is most intense, and there is a high risk that nutrient loads to the Baltic Sea from Poland, the Baltic States and Russia will increase in spite on improved water water treatment. Moreover, Compliance to the BSAP will require massive changes in the agricultural sector with drastic reductions in fertilizer use and livestock.

7

biogeochemical models, is well suited to modelling cost-allocations between sea-regions, countries and main catchments.

The higher spatial resolution RECOCA model is designed to address cost-effective spatial allocations of these abatement measures within countries and regions.

Both models provide results to help identify the most cost-effective measures for reducing nutrient loads. Both models can, for a given set of sea-basin-specific N and P reduction targets, specify to what extent each measure should be applied and in which locations, but at different resolutions and scales in order to reduce costs. Both models will eventually include abatement measures in the agricultural, energy and transport sectors, together with wetland restoration and improved wastewater treatment. Emissions to both air and water will be included in the cost-minimisation calculations. Special attention is paid to the natural removal and transport of nutrients from agricultural and wastewater sources through rivers and lakes to the sea, and also air-borne transport of emissions from energy, transportation and agriculture.

The RECOCA model adds to the present cost-modelling tools for the Baltic, as it is the first bottom-up model, building on intensive, detailed interdisciplinary work and data exchange between the natural scientists and the economists in the RECOCA project. For each 10 x 10 km grid cell, area-specific physiographic, cost and population data are used.

These results of the simplified RECOCA model cannot yet be used for policy analysis since only data for two agricultural measures are included. However, results relevant for policy analysis of cost-effective reductions of coastal loads, as decided in HELCOM BSAP, including comparisons between different regions can already now be explored with the BALTCOST model. Data from the other work-packages in RECOCA have been utilised to describe the capacities for nutrient load reduction in each of the 22 drainage basins and the 117 catchments as well as the nitrogen and phosphorus retention in these catchments. BALTCOST results suggest that it should be possible to achieve the BSAP load reduction targets for N and P in most Baltic Sea regions, with the exception of the P load target in the Baltic Proper and the N reduction target in the Danish Straits, where only 74% and 88% respectively, of the desired BSAP load reductions can be achieved. These results indicate that the targets are hard to achieve, and that more abatement measures are required to fulfil these targets. The minimised total cost of delivering these load reductions across the 9 Baltic littoral countries is estimated to be 4.69 billion Euros, annually (Table 1).

8

Country Annual cost million € Sweden 271.7 Finland 17.2 Russia 506.9 Estonia 32.3 Latvia 226.7

Lithuania 405.9 Poland 2385.5

Germany 472.1 Denmark 370.8 Total cost 4689.2

Table 1. Total annual cost of delivering the nutrient reduction targets using the lowest cost combination of basic specific abatement measures using the BALTCOST model.

1.3 Socio-economic relevance and policy implications

The socio-economic and policy implications of RECOCA and BALTCOST are multifold and can be structured in decision support tools (models), training workshops and stakeholder conferences.

Decision support

We anticipate that RECOCA will allow decision makers to evaluate how changes in land use will affect nutrient loads to the Baltic Sea. Most importantly the CSIM model (WP5), developed and supported by the SWAT, DAISY and NANI models (WP2-4) will be used during the update of the eutrophication section of the BSAP; the BALTCOST model (WP8) further developed in RECOCA and supported by the regional economic models (WP7) is a corner stone of the Baltic Stern Initiative as initiated by the Swedish Finnish and Danish EPAs. Both platforms (HELCOM, Baltic Stern) will allow that RECOCA results are disseminated and applied in an optimal way to achieve and plan Ecosystem Based Management Strategies for the Baltic Sea.

During the RECOCA starting phase in 2009, D.G. Enterprise, European Commission, and the Department of Environment, Spanish National Institute for Agriculture and Food Research

The costs for the BSAP, based on updated and more reliable calculations of costs at sources and nutrient retentions, are substantially higher than earlier estimates found in the literature. The total costs of achieving BSAP targets would undoubtedly be lower if the cost-effectiveness of different measures were taken into account in the allocation of the country specific quotas.

9

and Technology (INIA) consulted us for an assessment of the eutrophication risk of phosphates in detergents on the Baltic Sea. This was sponsored by CEEP (a CEFIC Sector Group). Later on, the European Task Force Group on the EU Marine Strategy Directive for the Quality Descriptor (QD) 5 (Eutrophication) also consulted us in 2009 and 2010. RECOCA data and tools have been applied for both tasks and can be regarded as the first highlight of the project

As a second highlight of the project, the database and hydrological-biogeochemical models (WP2-5) have been presented for the HELCOM TARGREV project in June 2010 and in follow up meetings in 2011. The results will update the environmental targets within the next phase of the BSAP. Our model tools and data will also be the basis for the “country allocation scheme”, i.e. the distribution of nutrient reductions needed per country based on the new targets. In total, RECOCA scientists have participated in >10 HELCOM meetings of various characters, i.e. HELCOM Load Expert Group, HELCOM TARGREV group and acted as observers and advisers in the HELCOM MONAS group. Moreover, the RECOCA tools will be used to modify relevant policy documents (Baltic Sea Action Plan, BSAP), i.e., to i) estimate the effect of the compliance of HELCOM contracting parties to the Waste Water Treatment Directive on river nutrient loads, ii) to estimate the effect of the compliance to the Nitrogen Ceiling Directive for atmospheric deposition over the Baltic Sea area and iii) to calculate country allocation scheme based on river and atmospheric loads within the updated BSAP. HELCOM acknowledged these tools as appropriate for the further BSAP revision and by this the RECOCA suggestions (i, ii, iii see above) for designing, implementing and evaluating the pertinent public policies and governance will be implemented.

The leaching of nutrient from intensive livestock farms to the Baltic Sea, has been identified e.g. by the Helsinki Commission as a major point source for eutrophication of the Baltic Sea. Fred Wulff is member of the steering committee of the private fund Baltic Sea 2020 (http://www.balticsea2020.org ) A high priority project of the Fund is to reduce the leaching, through technical development and spreading information on “Best Available Technologies” for cost efficient manure treatment. Focus is on pig manure at large industrial farms (>2000 slaughter pigs/750 sows). RECOCA has provided estimates on the importance of manure leaching for nutrient loads to the Baltic.

Training workshops

Within the RECOCA framework we organized five major courses, four on the hydrological-biogeochemical tools and one on economic tools developed. At all workshops senior scientists, Postdocs and PhD students from RECOCA but also from other related projects participated.

Workshop for SWAT and DAISY users ”How to decrease the pollution to the Baltic Sea?” 20-21 May 2009, Bioforsk, Ås, Norway.

Workshop SWAT and DAISY 14-15 December 2009, in Silkeborg, Denmark.

10

Workshop for SWAT users in RECOCA WP 4 - River basin model 9-10 September 2010 Warsaw, Poland.

Workshop for COST minimization models 15-17 February 2010, in Copenhagen, Denmark

Workshop for SWAT users in RECOCA WP 4 -Type river catchments in the Baltic Sea drainage basin, 14 - 16 March 2011, Tallinn, Estonia.

Scientific and Stakeholder Conferences

RECOCA results have been presented at numerous international conferences; here we report only the most significant conferences when it comes to scientific and stakeholder impacts.

The Swedish Ministry of the Environment of Sweden arranged a seminar “Building marine policy on best available knowledge”, on 25 August 2010 with participation of high level delegates including several ministers of environment (Finland, Sweden, Latvia, Poland, Estonia). The seminar contained three sessions including a high level ministerial segment addressing the latest assessment of the state of the environment in the Baltic Sea, scientific challenges, important policy processes, and new tools and measures. Christoph Humborg and Fred Wulff were invited and Wulff gave a plenary presentation ‘ Accounting and modelling: the BSAP nutrient reduction scheme where the catchment models of RECOCA were presented.

Other major conferences:

Presentations during the EUTRO-conference, 15-18 June 2011, Nyborg, Denmark.

Presentations during the Linking science and management section in the Baltic Sea ecoregion-conference, 9-10 September 2009

Presentations of the RECOCA data base and hydrological-biogeochemical models during a plenary talk of the Chesapeake Bay Modelling Conference in May 2010 in Washington DC.

Presentation of the RECOCA results at the N2010 conference in New Delhi, December 2010.

Presentation during the seminar ‘Hållbara Hav 2011 – Östersjön’ with Crown Princess Victoria, 22 September 2011, Stockholm, Sweden.

Presentation at the BONUS Forum Coinciding with the 2nd Annual Forum for the EU Strategy for the Baltic Sea Region and Baltic Development Forum, 24 October 2011, Gdansk, Poland.

Presentations during the joint ECOSUPPORT/RECOCA Final conference 5-6 December 2011, Stockholm, Sweden.

Presentations during the joint ECOSUPPORT/RECOCA Stakeholder conference, 7 December 2011, Stockholm, Sweden. The BONUS projects ECOSUPPORT and RECOCA

11

invited scientists and stakeholders to a conference on "An outlook to the future Baltic Sea: how can we reach the targets of the Baltic Sea Action Plan?" In total some 80 people participated with delegates from the office of the Swedish prime minister, the ministries of agriculture and environment from Sweden, agricultural NGOs from Sweden and other Baltic Sea riparian countries, and many more stakeholder and scientists. More information under: (http://www.balticnest.org/balticnest/activities/events/events/ecosupportandrecocastakeholderconference.5.50fc7668132ea9eba0a8000143.html

This final stakeholder conference was clearly the third highlight of the RECOCA project.

2. Gained scientific results during the reporting period 2009-2011

Note: All deliverables produced within the nine RECOCA working packages can be found under ftp://Recoca_delList:[email protected]

In this final report only an overview about the various tasks and the most important results produced within the WPs are given. For further detailed information on deliverables, see Table 1.

12

WP No

Deliverable No

Deliverable title Nature Publicity level

Direct Link

1 1 Set up RECOCA web site with updated work plans and implementation strategies WS PU ftp://Recoca_delList:[email protected]/WP1/d1

1 2 Documents of strategic guidance from the advisory group and stakeholder advice from the Liaison groups RS PU ftp://Recoca_delList:[email protected]/WP1/d2

1 3 Set up of updated river basin model and economic model in NEST PR? PU ftp://Recoca_delList:[email protected]/WP1/d3

2 1 Data on level of economic activities and land usage patterns for type river basins DB? ? ftp://Recoca_delList:[email protected]/WP2/d1

2 2 Data on levels of economic activities and land use patterns for 85 main river basins DB? ? ftp://Recoca_delList:[email protected]/WP2/d2

2 3 Map of nutrient export sensitive river basins around the Baltic Sea catchment a shown in NEST DB PU ftp://Recoca_delList:[email protected]/WP2/d3

2 4 Overall data base on land use and level of economic activities as shown in NEST DB PU ftp://Recoca_delList:[email protected]/WP2/d4

3 1 Data set on hydro-meteorological, physiographical, and statistical forcing data. DB PU ftp://Recoca_delList:[email protected]/WP3/d1

3 2 The natural scientific input to a detailed and consistent database relevant to calculation of abatement costs. DB PU ftp://Recoca_delList:[email protected]/WP3/d2

3 3 Library of model runs (scenarios) of type concentrations of nutrients for surface and ground waters as a function of the application regional measures in varying degree and combinations [Daisy].

DB PU ftp://Recoca_delList:[email protected]/WP3/d3

4 1 Data set on hydro-meteorological, physiographical, and statistical forcing data. DB PU ftp://Recoca_delList:[email protected]/WP4/d1

4 2 Estimates of ground water nutrient retention RS PU ftp://Recoca_delList:[email protected]/WP4/d2

4 3 Estimates of uncertainties in river basins RS PU ftp://Recoca_delList:[email protected]/WP4/d3

4 4 Library of model runs (scenarios) of type concentrations of nutrients for surface and ground waters as a function of the application regional measures in varying degree and combinations [SWAT].

DB PU ftp://Recoca_delList:[email protected]/WP4/d4

5 1 An updated CSIM model where measures on point and diffusive sources can be demonstrated RS PU ftp://Recoca_delList:[email protected]/WP5/d1

5 2 New CSIM spatial integration modes: 1) individual river basin mode, 2) Baltic Sea sub basin catchment mode 3) EU Water District mode

RS PU ftp://Recoca_delList:[email protected]/WP5/d2

5 3 New CSIM measure mode visualizing in NEST the effects of measures on loads (a combination of various measures in different regions (scenarios) will be given to achieve the nutrient reductions as recommended in the HELCOM BSAP).


6 1 Data set on regional retention characteristics for the various sources DB? PU? ftp://Recoca_delList:[email protected]/WP6/d1

6 2 Assessment of retention and uncertainty RS PU ftp://Recoca_delList:[email protected]/WP6/d2

6 3 Report on impact of different measures on coastal loads. RS PU ftp://Recoca_delList:[email protected]/WP6/d3

6 4 Report on uncertainty regarding the impact of different measures on coastal loads RS PU ftp://Recoca_delList:[email protected]/WP6/d4

7 1 Report on costs estimates for different measures, regions and locations. RS PU ftp://Recoca_delList:[email protected]/WP7/d1

7 2 Report on the full regional cost model structure. RS PU ftp://Recoca_delList:[email protected]/WP7/d2

7 3 Analyses of cost-effective reductions of coastal loads (scenarios), including reductions suggested by HELCOM BSAP, and comparisons between different regions.


8 1 Improved COST model in the NEST RS PU ftp://Recoca_delList:[email protected]/WP8/d1

8 2 Report the implications of uncertainty about nutrient transports for the cost-effective solutions. RS PU ftp://Recoca_delList:[email protected]/WP8/d2

8 3 Report on the implications of the informational situation for central and regional decision makers with regard to policy choice and strategic interaction.


9 1 Report on cost effective solutions for achievement of alternative marine and coastal targets as for example to reach the nutrient reductions recommended by HELCOM BSAP.


9 2 Report on role of uncertainty for the cost-effective allocation for alternative marine targets. RS PU ftp://Recoca_delList:[email protected]/WP9/d2

13

Table 1: RECOCA Deliverables

14

2.1. Work Package 1 – Management and Dissemination Lead Partner: Fredrik Wulff, SU-BNI

Researchers involved in the current work:

SU-ITM : Christoph Humborg

ÅU: Berit Hasler

2.1.1 Objective Overall scientific management of RECOCA, the establishment of a Project Advisory Board and Regional Liaison Groups and disseminations via NEST.

2.1.2 Methodology and scientific achievements Task 1.1 Review project implementation and set annual work plans

The project implementation and the follow up of the annual work plans have been presented and discussed during the kick-off meeting in Stockholm (February 2009) and the all partner annual meetings in January 2010 (Roskilde) and 2011 (Warsaw). Moreover, numerous meetings of the natural scientific (Warsaw, Stockholm, Oslo, Roskilde and Silkeborg) and the economic subgroups (Copenhagen, Roskilde, Stockholm) have set up more specific working plans of the individual disciplines.

Task 1.2 Receiving strategic advices from a Project Advisory Board and stakeholder advice

In the original proposal it was suggested to establish an official advisory board where HELCOM should be the given partner, because the CSIM and COST model are already and will be a cornerstone in the formulation of the final Baltic Sea Action Plan. A letter was formulated in this matter and addressed to the HELCOM executive officer. The response was negative; HELCOM had decided not to act in advisory boards within the BONUS projects, simply due to the fact that all of these projects have by definition strong links to the HELCOM activities. However, Fredrik Wulff and Christoph Humborg are both members of the revision group of the BSAP eutrophication segment and the advice from HELCOM on which processes should be parameterized in the NEST CSIM model were and are given throughout the BSAP update process, because the BSAP follow up will partly be built on CSIM results.

Strategic guidance from HELCOM

In particular, HELCOM asked the RECOCA researchers to develop their CSIM model to assist the “Calculation of Country-wise Allocation of Nutrient Load Targets” following the revision of the BSAP (HELCOM HOD22/2007); the detailed guidance report is given under deliverable 2.2 (D. 2.2 of WP1).

15

Strategic guidance from Baltic-Stern

The Swedish EPA established in 2009 the Baltic Stern Secretariat at the Stockholm Resilience Centre (http://www.stockholmresilience.org/balticstern) with the aim to coordinate the economic analyses within Sweden, Denmark and Finland concerning cost allocation schemes, cost minimization models and cost/benefit analyses towards the ecological targets as set by the actual and revised BSAP. Berit Hasler coordinating within RECOCA the development of the NEST BALTCOST model is affiliated to the Stern secretariat. Advice for the development of BALTCOST is given both by the Stern staff and the economic scientists in the other northern countries.

Thus, the BSAP implementation group and the Baltic Stern secretariat have proven to be valuable Liaison groups for the overall advice on the implementation of RECOCA overall aims.

The RECOCA web page

(http://www.balticnest.org/balticnest/research/ongoingprojects/bonus/recoca.4.2beb0a011325eb5811a8000125967.html ) was set up during the first project year and we are continuing to update and improve this website (D 1.1).

1.3 Designing new NEST modules showing effects and costs of measures RECOCA contributes to three major products that will be visualized in the Nest system:

• New catchment data base and nutrient accounting tool to estimate nutrient loading • New catchment model CSIM simulating nutrient pathways and effects of mitigation

measures for the Baltic Sea • New Cost minimization model BaltCost

The new catchment database has been implemented into Nest and the user can download all relevant data. The new CSIM and BaltCost models are operational and have been used to do scenario analysis on effects and costs of measures. However, these models are will implemented into Nest during late spring 2012, i.e., we are late with this parts of deliverable D. 1.3 (Set up of updated river basin model and economic model in Nest) by a few months.

Publications: http://www.balticnest.org/balticnest/research/ongoingprojects/bonus/recoca.4.2beb0a011325eb5811a8000125967.html

16

2.2. Work Package 2 – Data base and river basin nutrient budgets

Lead Partner: Christoph Humborg, SU-ITM


ÅU: Hans Estrup Andersen, Hans Thodsen

Bioforsk: Per Stålnacke

LLU: Viesturs Jansson

WEEC: Tomasz Zylicz

SGGW: Adam Was

Informal contribution from Dennis P. Swaney and Bongghi Hong (Cornell University, USA) and Erik Smedberg (SU)

2.2.1 Objective Forming a common database for river basin models and economic models and analyses. Calculating Net Anthropogenic Nutrient Inputs

2.2.2 Methodology and scientific achievements Task 2.1 Estimate land use patterns and level of economic activities in all riparian countries and Task 2.2 Detailed estimate of land use patterns and level of economic activities for type river basins

Land use patterns and levels of economic activities have been compiled during the initial phase of RECOCA using GIS technique and are stored in the major RECOCA database. We further analysed the data from the EU Joint Research Centre (fertilizer use, crop types), EUROSTAT (livestock data), HYDE database (population), CORINE (land cover) and SMHI (hydrological and climate forcing). All these data have been compiled for the 117 watersheds (82 major watersheds and 35 coastal areas) and also for 8 smaller, type watersheds. This extended data compilation (D 2.1 and 2.2) was made available to all other partners within RECOCA. The data have now been organized following the NANI calculations (see below) that are also retrievable for the other hydrological-biogeochemical (DAISY, SWAT) as well as economic models (BALTCOST and RECOCA). These data are organized into fertilizer data, atmospheric deposition data, biological N-fixation data, crop data, animal data, population data and resulting food and feed budgets for all watershed types addressed in RECOCA. Moreover we created for this deliverable a detailed list of point source data, i.e. statistics on WWTP including estimates on people connected and unconnected to sewage systems.

17

Task 2.3 Calculating Net Anthropogenic Nutrient Inputs for 82 main river basins by means of atmospheric deposition, fertilizer use and food and feed spread sheets

We have calculated the Net Anthropogenic N and P inputs (NANI, NAPI) to all 82 main watersheds using the data on land use patterns and levels of economic activities and created a map of nutrient sensitive watersheds, i.e., those that export large amounts of nutrient to the Baltic Sea in relation to the net inputs (D. 2.3). We also developed the regional settings of the NANI budgeting tool that will address the significant variation in agricultural practices and resulting nutrient accountings among European countries. NANI, first introduced by Howarth et al. (1996) for US watersheds, estimate the human-induced nitrogen inputs to a watershed and have been shown to be a good predictor of riverine nitrogen export at a large scale, multi-year average basis. NANI has been calculated as the sum of four major components: atmospheric N deposition, fertilizer N application, agricultural N fixation, and net food and feed imports, which in turn are composed of crop and animal N production (negative fluxes removing N from watersheds) and animal and human N consumption (positive fluxes adding N to watersheds). Within WP 2 we created a NANI toolbox that is an automatic nutrient accounting tool extracting and compiling the GIS data relevant for nutrient budgets from the RECOCA database to a watershed level. The toolbox is described in detail together with a description of the overall database on land use and level of economic activities in deliverable 2.4 of WP2.

Publications: Hägg, H., C. Humborg, C. - M. Mörth, M. Rodriguez Medina, F. Wulff, Scenario Analysis on Protein Consumption and Climate Change Effects on Riverine N Export to the Baltic Sea. Environmental Science and Technology, 2010, 44 (7) 2379–2385.

Hong, B., Swaney, D.P., Mörth, C.-M. Smedberg, E., Eriksson Hägg, H., Humborg, C., Howarth, R.W., Bouraoui, F. Evaluating regional variation of net anthropogenic nitrogen and phosphorus inputs (NANI/NAPI), major drivers, nutrient retention pattern and management implications in the multinational areas of Baltic Sea basin. Ecological Modelling, 227 (2012) 117-135.

Eriksson Hägg, H., C. Humborg, D.P. Swaney, C.-M. Mörth. Riverine nitrogen export in Swedish catchments dominated by atmospheric inputs. Biogeochemistry, 2011. doi 10.1007/s10533-011-9634-7.

18

2.3 Work Package 3 – Farm Scale Model Lead Partner: Hans Estrup Andersen, ÅU


ÅU: Hans Estrup Andersen, Gitte Blicher-Mathiesen, Peter Mejlhede Andersen

LLU : Viesturs Janssons

Bioforsk: Johannes Deelstra, Per Stålnacke

SU-IGG: Erik Smedberg, Magnus Mörth, Christoph Humborg

2.3.1 Objective Provide type concentrations for nutrients in surface and ground waters to be used in CSIM. Describe the effects of changes in agriculture including mitigation measures for type river basins on a farm scale. Supply the natural scientific inputs to WP 7 and 8.

2.3.2 Methodology and scientific achievements Task 3.1 Compilation of hydro-‐meteorological, physiographical, and statistical forcing data was achieved and delivered in month 6 (D 3.1; see also WP2) and made available to all RECOCA partners.

Task 3.2 Defining type farms as a function of regional climate, soils, and agricultural management. We divided the Baltic Sea drainage basin (1.7 106 km2) into 10 x 10 km grid cells and described land use and agricultural practices for each cell from a comprehensive dataset combining national and regional statistics and published surveys (D 3.2). The strong heterogeneity in farm size and production intensity within the drainage basin was described by three different farm types, which were parameterized individually regarding livestock and inputs per crop of fertilizer and manure for each riparian country and calibrated to national statistics on consumption of fertilizer and manure. Within each country the farm types were distributed at the NUTS2/Oblast/Voblast level using livestock production as a key forming a consistent and very detailed description of agricultural production throughout the Baltic Sea drainage basin.

Task 3.3 Model calibration and validation in data-rich mini catchments nested in or adjacent to meso-scale type river basins. The state of the art soil-vegetation-atmosphere model DAISY was selected to calculate agricultural N losses. The model was calibrated to monitored root-zone N losses in mini catchments and to regional and national statistical data on crop yields.

Task 3.4 Quantifying the effect of eco-engineering approaches such as wetland formation based on existing experimental data was achieved by consulting national experts on the effects of wetlands on water quality. A differentiation was made between restored and constructed wetlands.

19

Task 3.5 Simulating type concentrations for surface and ground waters as a function of agricultural measures used in CSIM. We identified the ranges found in the Baltic Sea drainage basin regarding important drivers: precipitation, temperature, soil types, farm types and levels of inputs of fertilizer and manure to crops. We constructed more than 11,000 combinations of these drivers to describe the variations found in the Baltic Sea drainage basin and applied the DAISY model to these combinations running a time series of climate data (1995 – 2006). From the data set of DAISY simulations we identified the most important variables by multiple regression statistics and developed a statistical N leaching function for N losses from the root-zone of agricultural land in the Baltic Sea drainage basin (deliverable 3.2). With this N leaching function and the data sets developed in tasks 3.1 and 3.2 we were able to calculate N losses for the entire Baltic Sea drainage basin and effects of mitigation measures (D 3.3) and to identify agricultural hot spots in a consistent way and at a level of detail not hitherto seen for this area. During the project WP 3 was expanded from originally working only with representative farm types in selected river basins to encompass agriculture in the entire drainage basin. This has greatly added to the value of the products from the WP.

Publications: Blicher-Mathiesen, G. et al. (in prep.). Constructing a detailed and consistent data set on agricultural management in the Baltic Sea drainage basin.

Andersen, H.E. et al. (in prep.). Agricultural N loading of the Baltic Sea – effects of mitigation measures.

20

2.4 Work Package 4 – Type river basin models Lead Partner: Hans Estrup Andersen, ÅU


ÅU: Hans Estrup Andersen, Hans Thodsen

Bioforsk: Johannes Deelstra, Per Stålnacke, Csilla Farkas

LLU: Viesturs Janssons, Kaspars Abramenko, Ainis Lagzdins

SGGW: Jaroslaw Chormanski, Adam Was

SU-IGG: Magnus Mörth, Christoph Humborg

Informal contribution from Tallinn Technical University (Arvo Iital and Peeter Ennet) and Water Management Institute of the Lithuanian University of Agriculture (Aušra Šmitienė)

2.4.1 Objective Describe the effect of measures on point and diffuse sources for meso-scale type river basins. Provide estimates of ground water retention of nutrients to be used in CSIM. Provide type concentrations for nutrients in surface and ground water to be used in CSIM. Supply estimates of uncertainties in river basin data to be used in the cost model.

2.4.2 Methodology and scientific achievements Task 4.1 Data set on hydro-meteorological, physiographical and statistical forcing data. Seven meso-scale catchments representative of the total drainage basin from the boreal north to the intensively cultivated southeast were selected. All the forcing data necessary for running the SWAT model assembled, compiled and delivered in month 6 (D 4.1). Calibration and validation data were collected from European, national and local sources. Agricultural management data and point source information was collected in close cooperation with local data owners.

Task 4.2 Estimates of ground water nutrient retention. Groundwater nutrient retention was estimated for the entire Baltic Sea drainage basin subdivided into 117 watersheds as the difference between model based estimates of root-zone N leaching and measured riverine N loads, taking point source contributions directly to surface waters into account and correcting for surface water retention. Estimates of surface water retention were calculated with the MESAW model (task 6). Reported as D 4.2 and 6.3.

Task 4.3 Sensitivity analyses to assess the influence of uncertain input parameters and input data on important output data. Datasets covering the entire Baltic Sea drainage basin are generally of a lower quality compared to national and local data sources. Sensitivity analyses and uncertainty of important input parameters are addressed in D 4.3 and 6.4.

21

Task 4.4 Simulating type concentrations for surface and ground waters as a function of measures on point and diffuse sources used in CSIM. Finalization of this task has been delayed since the expansion of WP3 partly was done on the expense of WP4. Final results are expected in February 2012. The modelling work on the type river basins is performed by four model groups (from AU, Bioforsk, LLU and SGGW). During the project period four workshops have been held between the modellers in order to harmonize the work and to educate young scientists. The models are all calibrated to local monitoring data. The calibrated models are giving estimates on how a number of defined scenarios will affect nutrient concentrations in shallow groundwater and surface water in the regions.

Publications: Lagzdins A., Jansons V., Sudars R., Abramenko K. (2012). Scale issues for assessment of nutrient leaching from agricultural land in Latvia. Accepted for publication in Hydrology Research.

Thodsen, H. Csilla Farkas, Jaroslaw Chormanski, Adam Was, Kaspars Abramenko, Ainis Lagzdins, Hans E. Andersen (in prep). Modelled nutrient load changes from agricultural management scenarios in seven type watersheds around the Baltic sea.

Andersen, H. E., Csilla Farkas, Jaroslaw Chormanski, Adam Was, Kaspars Abramenko, Ainis Lagzdins, Hans Thodsen, Ausra Smitiene, Peeter Ennet, Erik Smedberg (in prep). Catchment functioning along gradients in climate and anthropogenic pressure in the Baltic Sea drainage basin.

Farkas, C., Hans Thodsen, Alexander Engebretsen, Jaroslaw Chormanski, Adam Was, Kaspars Abramenko, Ainis Lagzdins, Hans E. Andersen (in prep). Impact of projected land use changes on water quality in the Baltic Sea region.

22

2.5 Work Package 5 – Baltic Catchment Model Lead Partner: Carl-Magnus Mörth, SU-IGG


ÅU: Hans Estrup Andersen

Bioforsk: Per Stålnacke

LLU: Viesturs Jansons

SU-ITM: Christoph Humborg

Informal contribution from Cornell University, US (Dennis P. Swaney and Bongghi Hong)

2.5.1 Objective Up scaling the effect of measures as gained from NANI, DAISY and SWAT models runs for the type river basins by using type concentration libraries (simulations).

2.5.2 Methodology and scientific achievements Task 5.1 Compilation and updating of hydro-meteorological and statistical forcing data

The statistical forcing data (land use and waste water treatment), for the new CSIM model has been taken from the overall RECOCA database (WP2) and are described in the NANI tool box (D 2.4). Daily data on precipitation and temperature were aggregated from Haylock et al. (2008). Added to the CSIM model are now also output data on the contribution from WWTP. In the CSIM model calculations are made to estimate load of N and P from WWTPs according to official EUROSTAT data. These data were added to the output and compared with the total load (after subtracting river retention), giving a percentage of the load from WWTPs. In this way we fulfilled D. 5.1 “An updated CSIM model where measures on point and diffusive sources can be demonstrated”.

Task 5.2 Model calibration and validation for 105 river basins

Temperature and rainfall is the driving force for the hydrological model within CSIM. The necessary forcing data (Haylock et al 2008) were extracted from synoptic stations in the whole Baltic drainage area. Synoptic precipitation and temperature observations for the time period 1950 to 2008 are available in an interpolated 1x1 degree grid database at SMHI, compiled into a database. The model is also capable of dealing with snow and snowmelt. The snow melt routine is one of the modifications done compared to the original model. In CSIM the temperature threshold where snow is melting can be set to any temperature as well as the amount of snow melting each day. The original CSIM model had two types of boxes, one soil box for each land use and one groundwater box common to the whole catchment. In the new CSIM model developed in RECOCA one more groundwater box has been added in order to better simulate hydrology and a source for nutrients in the northerly drainage basins. The upper soil box, linked to the land use, is from now on called runoff. Further efforts have been

23

taken in the spatial resolution of CSIM. The new CSIM model is able to simulate nutrient fluxes to the Baltic Sea from 1) individual river basin mode, 2) Baltic Sea sub-basins catchment mode 3) EU Water District mode (D. 5.2). Detailed maps on this new spatial integration mode can be found on the RECOCA server.

Task 5.3 Linking NANI/NAPI, Daisy, SWAT and CSIM models

The NANI-related variables in the Baltic Sea catchments, calculated by the NANI Calculator Toolbox, are used to estimate runoff N concentrations, which then are used as input to CSIM. Specifically, the following procedure was applied: Ten land use types used in CSIM hydrology simulation were aggregated into three types, “AG” for Cultivated Areas, “URBAN” for Artificial Surfaces and “DEP” for deposition type, i.e. forests etc. Watershed-based NANI-related variables (calculated by NANI Calculator Toolbox, WP2) were distributed into these three types of lands. The AG type received fertilizer N application, agricultural N fixation, livestock N excretion, and atmospheric N deposition. The URBAN type received human N consumption and atmospheric N deposition. The DEP type received atmospheric N deposition only. The AG and DEP type runoff N concentrations were calculated from these N inputs by applying the response functions reported in scientific literature, respectively. N inputs to the “URBAN” type lands were directly added to the stream, after applying some percent denitrification loss (see below). Groundwater N concentration was estimated as an area-weighted average of AG and DEP type runoff N concentrations estimated above. 75 percent of denitrification loss (i.e., 25 percent retention) of groundwater N was assumed, approximated from general watershed N retention estimates suggested by nutrient accounting studies (Howarth et al., 1996; Howarth et al., 2011). The same percent loss was assumed for the direct input of URBAN N to the stream. With this new model setup we were able to fulfil deliverable D 5.3 “New CSIM measure mode visualizing in NEST the effects of measures on loads (scenarios how to achieve the nutrient reductions as recommended in the HELCOM BSAP)”. The most important results from this deliverable indicate that the country wise nutrient reductions as foreseen by the BSAP can only be achieved with substantial (50-70%) of fertilizer and/or livestock reductions in southern cultivated part of the Baltic Sea catchment. A detailed report is available on the RECOCA ftp site.

Publications Hong, B., Swaney, D.P., Mörth, C.-M. Smedberg, E., Eriksson Hägg, H., Humborg, C., Howarth, R.W., Bouraoui, F. Evaluating regional variation of net anthropogenic nitrogen and phosphorus inputs (NANI/NAPI), major drivers, nutrient retention pattern and management implications in the multinational areas of Baltic Sea basin. Ecological Modelling, 227 (2012) 117-135.

24

2.6 Work Package 6 – Nutrient Reduction Effectiveness and Potential of Measures Lead Partner: Per Stålnacke, Csilla Farkas, Johannes Deelstra, Annelene Pengerud, BIOFORSK


LLU: Viesturs Jansons, Kaspars Abramenko, Latvia University of Agriculture, Latvia

SU-IGG: Carl-Magnus Mörth, Erik Smedberg, Christoph Humborg and Hanna Eriksson Hägg, Stockholm University, Sweden

ÅU: Hans Estrup Andersen, Gitte Blicher-Mathiesen and Hans Thodsen, Århus University, Denmark

SGGW: Adam Wąs, Warsaw University of Life Sciences, Poland

Informal contribution from Tallinn Technical University (Arvo Iital, Peeter Ennet and Anatoli Vassiliev) and Water Management Institute of the Lithuanian University of Agriculture (Aušra Šmitienė)

2.6.1 Objective Quantification of retentions from source emissions to river mouths. Estimation of impact of mitigation measures in regional and Baltic-wide models on coastal loads

2.6.2 Methodology and scientific achievements Task 6.1 Assessment of retention by using the developed retention software in the FP5-funded project EUROHARP plus data from the type catchment, experimental data hold by the consortium and modelling results (input from WP 3, 4)

The EUROHARP NutRet software has been tested for calculating the retention of nutrients (N and P) in surface water bodies (streams, rivers, lakes and wetlands) for 4 type catchments and for 81 watersheds within the Baltic Sea catchment. The software strongly overestimated the surface water retention, probably because it was developed for smaller river systems, differing from those used in the RECOCA project.

Therefore, the MESAW model was tested and further developed for RECOCA retention calculations. This model-approach uses nonlinear regression for simultaneous estimation of source strength (export coefficients to surface waters) for the different specified land-use or soil categories and retention coefficients for pollutants in a drainage basin. Nutrient retention results, obtained for the 4 type catchments using the MESAW model were in good agreement with those, calculated by other methods. Therefore, the MESAW model was selected for estimating regional nutrient retentions in the different Baltic Sea river basins.

25

Task 6.2 Regional retention characteristics for the various sources in the various river basins

The MESAW model was applied for calculating the retention of nutrients (N and P) in surface water bodies in the Baltic Sea river basins. Input data consisted of land use (distinguishing cultivated areas, wetlands, lakes and others), the total drainage area and point source emissions (WWTP and industry). Nutrient retentions were estimated for both, ISPRA and updated Corine/GLC land use data.

Results, obtained for 117 (N) and 76 (P) river basins (D 6.1) showed that the MESAW model was able to simulate the riverine loads with a very good accuracy. The MESAW model was able to estimate the nitrogen load at the river mouth of 88 Baltic Sea rivers for which we had observed data with a sufficient degree of precision and accuracy. Coefficients of determination between the observed and modelled data varied between 0.94 and 0.99 (D 6.2). The estimated retention parameters were also statistically significant for the nitrogen model. For phosphorus the MESAW model fit was not satisfactory despite a reasonable good fit between the observed and estimated loads. For the phosphorus model, all parameters for the diffuse losses from 3 land cover classes and also the two retention parameters were non-significant.

Our results showed that around 380,000 tons of nitrogen is annually retained in surface waters (streams, rivers, reservoirs and lakes). In comparison, the total riverine load to the Baltic Sea was estimated to 570,000 tons N/yr. This means that the nitrogen surface water retention is around 40%. The results for phosphorus indicate retention of 12,000 tons compared to an estimated river load of 18,000 tons P/yr for 76 Baltic Sea basins with measured P load. The estimates for phosphorus are, however, highly uncertain and should be interpreted and used with caution.

The obtained results will hopefully enable HELCOM and BNI to refine the nutrient load targets in the BSAP as well as to better identify cost-efficient measures to reduce nutrient loadings to the Baltic Sea. These issues are described and discussed in more detail in other RECOCA reports.

Task 6.3 Assessment about uncertainty regarding retention. Qualitative assessment supplemented with CSIM simulations models on coastal loads (input to WP 7 and 8), including estimation of uncertainties regarding these impacts.

The advantage of the applied MESAW model in comparison with the NutRet tool is that the MESAW model also calculates uncertainty of retention estimates. During the model application, statistically significant N estimates were obtained (Deliverable 6.2) that showed very good correlations. This enabled confidence intervals calculations.

The MESAW model was also able to estimate the phosphorus load at the river mouth of 76 Baltic Sea rivers, for which we had observed data with a sufficient degree of precision and accuracy (D 6.2). Still, the estimated parameters for the 3 land cover classes (i.e. cultivated, wetland and other) were not statistically significant. Moreover, the two retention parameters

26

were not statistically significant. However, drainage area as an explanatory variable was close to being significant (p<0.06). When using the obtained relationships for calculating the unit-area load, the model underestimated the observed load and was not very precise.

Our results are in agreement with those, reported in the literature, which states that nitrogen is generally easier to model compared to phosphorus with fewer clearly defined explanatory variables.

Task 6.4 Estimation of impact of measures in regional (river-basin) and Baltic-wide models on coastal loads (input to WP 7 and 8), including estimation of uncertainties regarding these impacts

A modelling system was developed and applied to the whole Baltic Sea drainage basin to assess agricultural nitrogen losses at a very fine scale and to calculate the effects of various measures to mitigate these losses. The system links modelling of root-zone nitrogen losses to catchment model estimates of retention in groundwater and surface waters yielding estimates of coastal nitrogen loading (D 6.3).

The basin was divided into 10 x 10 km grid cells. Land use and agricultural practices for each cell was derived from a comprehensive dataset combining national and regional statistics and published surveys (Deliverable 3.2). Average annual N losses from the root-zone were estimated using the soil-vegetation-atmosphere DAISY model for more than 11,000 combinations of the main drivers (precipitation, temperature, soil types, farm types and levels of inputs of fertilizer and manure to crops; D 6.3, Appendix). Based on the simulation results, multiple regression statistics was applied to identify the most important factors controlling N losses from the root-zone of agricultural land in the study area.

Retention of N along the flow path through groundwater and surface waters was calculated for the drainage basin subdivided into 117 catchments for which time series of monitored riverine N losses at the outlets were available (documented in D 6.2). Overall N retention per catchment was found as the difference between riverine N losses and catchment root-zone N leaching taking point sources into account (D 6.3). Independent estimates of surface water retention by the MESAW model (D 6.2) were used to split overall catchment N retention into estimates for respectively groundwater retention and surface water retention.

The effect of the following different measures on coastal nutrient loads was evaluated: i) increased utilisation of N in manure; ii) catch crops and iii) afforestation of 10% of the agricultural area. Sensitive areas - areas where the effect of mitigation measures is the largest and thus mitigation measures can be implemented most cost effective – were identified.

Uncertainties involved in the modelling system were evaluated by estimating the uncertainty (standard error) of the predicted riverine N losses at the catchment outlet (D 6.4). This was achieved by assigning estimates for the uncertainty (standard errors) to all variables of the N balance equation, used to estimate N losses at the catchment outlet to the sea.

Case studies from three Baltic Sea river basins gave an indication of the uncertainties on coastal load from five different mitigation measures. The results show that the standard errors

27

for the various scenarios were in the range of 24-60%. This might as a first glance be regarded as high uncertainty. However, it should be stressed that the entire exercise and in fact the entire RECOCA project had the intention to model all the 117 Baltic Sea drainage basins all with a uniform data set and a similar uniform modelling. In light of this, the given results and approach is really promising.

Publications: Per Stålnacke, Annelene Pengerud, Anatoli Vassiljev, Erik Smedberg, Carl-Magnus Morth, Hanna Eriksson Hägg, Christoph Humborg and Hans Estrup Andersen. Nitrogen surface water retention in the Baltic Sea drainage basin (in prep)

2.7 Work Package 7 – Regional cost effectiveness models Lead Partner: Tomasz Żylicz, WEEC


WEEC: Mikolaj Czajkowski

SLU: Katarina Elofsson

ÅU: Berit Hasler, Maria T.H. Konrad, Sisse L. Brodersen, K. Munck

SGGW: Adam Was

2.7.1 Objective Develop regional cost effectiveness models with higher spatial resolution than the Baltic-wide model. Analysis of cost-effective reduction scenarios taking into account uncertainty. Development of methods for linking regional and Baltic-wide cost minimization models.

2.7.2 Methodology and scientific achievements The aim of Work Package 7 (WP7) was to develop regional cost models for nutrient reduction measures that can later be aggregated and used for developing cost minimizing allocations for Baltic-wide nutrient reduction scenarios. All major tasks were addressed during the RECOCA project. However, we decided to restructure the time plan for the individual tasks that deviates from the numerical order given in the original proposal.

In order to do this, we designed and collected data for the RECOCA model – the first fully regionalized high resolution economic model for identifying cost-efficient measures to reduce nutrient loadings to the Baltic Sea. The model specifies how much of each measure should be applied and where, so that the aggregated costs are minimized, for a given set of basin-specific nutrient reduction targets.

The RECOCA model is a static model. It uses partial analysis and the time unit is one year. We note, however, that it is possible to extend the model to take dynamic effects into account; something that is especially significant when the mixing between the Baltic Sea basins is taken into account.

28

An important contribution of the RECOCA approach is that it is the first bottom-up model on such a disaggregated scale. The operational scale of the model is 10 x 10 km grid cells. For each grid cell the area-specific physiographic, population and cost data are used. This allows us to take the high variations in characteristics of the Baltic Sea drainage basin into account. In addition, area-specific potential and effectiveness of each measure is used, as well as area-specific cost coefficients. Finally, the RECOCA model utilizes grid cell specific surface water retention coefficients. All these components play an important role in determining the regionally disaggregated cost-effective solutions for reaching a specified nutrient reduction target.

Task 7.5 Development of consistent methods to link regional and Baltic-wide cost models.

Our approach allowed designing the regional cost minimization model in such a way that it is fully capable of solving nutrient reduction allocation schemes for each sub basin. The regional cost-effectiveness model was be constructed in such a way that it is be possible to identify a cost-efficient mix of measures to reach any set of targets set for each of the Baltic Sea basins, while utilizing region-specific data. This bottom-up approach is definitely very demanding in terms of computational time and data requirements, however, it gives an opportunity to utilize region-specific data – something existing models are not capable of doing and thus likely to overestimated costs of reaching specified targets. At the same time, this allowed us to develop a consistent method to fully link regional and basin-wide cost-efficiency models (Tasks 5).

The model has been developed using mathematical programming and coded in GAMS. This deliverable describes general model structure.

The RECOCA model was designed to take the following measures into account:

• Combined agricultural measures • Reduction of mineral fertilization • Reduction of livestock • Application of catch-crops • Wetlands restoration • Wetlands reconstruction • Improving wastewater treatment efficiency from primary to tertiary level • Improving wastewater treatment efficiency from secondary to tertiary level • Introducing tertiary wastewater treatment for people currently not connected to

wastewater treatment system • Reducing NOx emissions from transport, energy and shipping

Task 7.1 Baseline scenario constructed in cooperation with WP 2

We collected data, and estimated functional relationships describing efficiency, potential and cost of each measure to allow their inclusion in the model. Establishing the baseline scenario

29

required calculating the reference level of nutrient loadings – specifying current riverine nutrient loadings from each watershed and collecting data on the scale the measures are currently applied in different regions of the Baltic Sea, thus allowing for calculating the extent to which they can be applied further (their potential). These goals were completed within Task 1 of this WP.

Task 7.2 Calculation of cost functions and Task 7.3 Integration of impact coefficients estimated in WP 6.

Calculation of cost functions required collecting region-specific cost data for different measures in the agricultural, energy and transports sectors, for wastewater treatment and for different eco-engineering methods, and including them in the model. For most measures the cost functions were not estimated directly but rather data and functional relationships were included in the optimization model, where cost-effective solutions for all the measures are then determined.

We have collected the data required for estimating costs of most measures. In particular, these are: Standard Gross Margin of crop and livestock production (as a way of including opportunity cost of their reductions) and prices of mineral fertilizers. We have estimated costs of municipal wastewater collection and treatment and estimated costs of connecting new people to sewage collection systems. We have proposed ways to adjust these costs between countries, based on labour and capital prices and possibly other region-specific factors. These goals were completed within Task 2 and Task 3 of this WP.

Due to missing or delayed inputs, and other problems with the data it was not possible to make the full-scale RECOCA model operational. Instead, we present the results of a simplified version of the model, with only two agricultural measures included – reduced mineral fertilization and reduction of livestock.

Task 7.4 Analysis of cost-effective nutrient reduction scenarios.

These results of the simplified RECOCA model cannot be used for policy analysis. They are presented here as demonstration of how the RECOCA model works, and to some extent, to investigate how the bottom-up approach utilized by the model (even if only for two currently included measures) allows to identify the truly cost-efficient solution to reducing nutrient loadings to the Baltic Sea. For results relevant for policy analysis of cost-effective reductions of coastal loads as decided in HELCOM BSAP, including comparisons between different regions, we refer to the results from BALTCOST, described in deliverables 8.1 and 8.2 (Task 4).

In order to facilitate the comparison between the bottom-up approach utilized by the RECOCA model and the top-down approach utilized by other models, we present two sets of results. The simplified RECOCA model was used together with two sets of retention coefficients – one with grid-cell specific retention coefficients, and one with watershed specific coefficients. Note that the second approach is not truly a top-down approach, since we still use grid-cell specific effects, potentials and cost functions. However, the comparison

30

illustrates the potential of using the bottom-up approach to target retention-sensitive or cost-effective areas of the Baltic Sea drainage area, what allows for reducing the cost of reaching the reduction targets by at least an order of magnitude.

The measures included in the simplified RECOCA model turned out not to be capable of reaching HELCOM BSAP targets, even if fertilization rate and livestock were reduced to 0, if watershed level retention coefficients were used.1 If, however, the full spatial capability of the model is used (i.e. applying grid-cell specific retention coefficients to identify the most vulnerable grid cells) HELCOM BSAP targets can be achieved even with the two measures included in the model, at a cost of 96,182,805 EUR. It should be noted that these results only cover the nitrogen reduction targets, as RECOCA do not have P retention included in this first model version.

In order to facilitate a comparison of total costs if a bottom-up vs. top-down approach is used to identify efficient measures and locations of their applications, we provide two other scenarios in which reduction targets have been set to 10% of current N loading to each sea basin. The results show that using a spatially disaggregated approach to identify the most economical measures, locations and scale of their applications allows the same nutrient reduction targets at a fraction of the total cost that the top-down approach requires. If the most sensitive locations are targeted and the best measures are used (this is our grid cell level model) the 10% reduction of current N loadings to the Baltic Sea can be reached at the cost of 6,173,310 EUR. The same reduction if the policy was not spatially targeted (watershed level retention coefficients) would cost 8,910,819 – that is 44 % more. At the same time, the 10% nutrient reduction would be 36.76% lower, in absolute terms.2Again, it should be noted that only N targets are concerned, and not the P targets. The costs are therefore very illustrative for the differences in spatial modelling, but not usable for policy purposes.

Publications: Berbeka, K., M. Czajkowski i A. Markowska. (forthcoming), "Municipal Wastewater Treatment in Poland as a Means of Reducing Nutrient Loadings – Efficiency, Costs and Returns to Scale" Water Science and Technology.

1In the case of grid cell specific retention coefficients the model predicts the total possible reduction in nutrient loadings to the Baltic Sea to be 36.76% higher than in the case of using watershed specific retention coefficients. 2 This is because the estimated total nutrient loadings, when using grid cell specific retention coefficients are overall larger than in the case of watershed specific nutrient loadings, hence 10% reduction in the first case is larger (in absolute terms) than in the second.

31

2.8 Work Package 8 – Cost Minimization Model Lead Partner: Berit Hasler, ÅU


ÅU: Berit Hasler, James (Jim) C.R. Smart, Anders Fonnesbech-Wulff, Maria T.H. Konrad, Sisse L. Brodersen, K.Munck

SLU: Katarina Elofsson

WEEC: Mikołaj Czajkowski

SGGW: Adam Was

Objective Improve the former COST model with respect to measure coverage and data quality. Update all measures with respect to costs, as well as the data defining the regional capacity of the measure with respect to nutrient reduction capacity as well as the coverage of the measure (e.g. the potential of wetland restoration within a drainage basin). Analyse cost-effective solutions at a Baltic-wide scale and to analyse the informational situation for international and regional decision-makers.

Methodology and scientific achievements Task 8.1 The baseline scenario establishment (same baseline as in WP 7).

The baseline scenario year is 2005; prices and costs retrieved from other years than 2005 are adjusted to the price level of 2005. Data from the other work-packages in RECOCA haves been utilised to describe the capacities, the nutrient load reduction in each of the 22 drainage basins and the 117 catchments as well as the nitrogen and phosphorus retention in these catchments. The retentions have been used to estimate measure specific retentions in each of the 117 catchments and hereby the data and information from the other parts of the project has led to a novel integrated and interdisciplinary model set-up, where the cost-effectiveness is calculated using and utilising data on differences in unit costs and effects on load reductions at a regional scale. The data used for all measures are described in the annexes to D 8.1.

Task 8.2 Improvement of cost functions and extended measure coverage.

The modelling of the measures and the use of regional data has been extended for the new EU-member states and Russia: The cost-functions and the effects in the former COST model were to a large extent built on Danish experimental and empirical data. In the new BALTCOST model all cost data as well as the data for capacities and effects are revised, so that the BALTCOST model is a fundamentally new model (D 8.1), building on the data retrieved in the other parts of RECOCA. We have chosen to pay much attention to i) achieve as good as possible links between the natural scientific data for e.g. retention and nutrient load reductions and the measures applied on each of the drainage basins. ii) We have also

32

paid much attention to the modelling of wastewater treatment, as it was possible to improve this measure as a joint work in RECOCA. The modelling of the potential of these measures has been improved, especially for the eastern countries and new member states, utilising available new regional data estimates of the potential to extend the wastewater treatment by connecting households that were not connected in the baseline, or improving wastewater treatment from primary and secondary levels to tertiary treatment. Detailed cost data from Poland and Denmark has been used to estimated cost-functions and elasticities, and these cost-functions has been adjusted to fit the prices of e.g. labour and energy in the other countries. iii) All other cost-data have been revised and country specific costs are used for the agricultural measures as well as the other measures. All these measures are described in deliverable 8.1.

Detailed work improving the existing measures and the data used to model them has been chosen instead of including more measures. More measures will be included now that the integrated framework has been built up. New measures will include measures to reduce atmospheric nitrogen from power plants and ships, as well as manure handling. Improved integration between livestock production and crop production, enabling nutrients to be utilised more effectively, could also be adopted as a potential additional measure to reduce nutrient loads into the Baltic. Extending the measures could probably further reduce the modelled total costs of achieving the targets for the Baltic Sea regions. In deliverable 9.1 a detailed analysis shows that over 50% of the total costs of delivering the phosphorus load reductions in the Baltic Proper are used to achieve the last 12.5 % of the reductions. This high cost arises because the available capacity of the cheapest measures is used and more expensive, less effective, measures have to be used to deliver the reductions. Extending the model with more measures can be relatively easily be done using the present model structure and data for livestock production localisation etc., but require data that are not easily available. Examples of this are the use/costs of manure spreading equipment, storing etc. in countries where this utilisation is poor (Poland, Russia, Lithuania, Latvia, Estonia and parts of Germany).

The second part of D 8.1, i.e., the implementation in of the new BALTCOST model in the decision support system NEST is currently on going.

Task 8.3 Integration of catchment model (same as WP 7).

The BALTCOST model has been integrated with the catchment models; the particular implementations used to achieve this are explained for each of the abatement measures in the annexes to deliverable 8.1. The integration has been done by these steps: i) Data on crop allocation, fertiliser applications, livestock production have been retrieved from the same data bases and used in both catchment models and the economic model, on 117 catchment level as well as on the 10 k grid cell level. ii) Data on retention on the 117-catchment level have been used to estimate specific retentions for each of the 22 drainage basins. iii) Information of the distribution of organic soils has been used to estimate the catchment specific potential for wetland restoration. For the economic model BALTCOST the catchment data are aggregated

33

to the 22 drainage basins, but the information at the 117 catchments is retained and can be utilised in the future in future developments of the BALTCOST and RECOCA models.

Task 8.4 Investigate cost-effective solutions to improvements in different ecological indicators.

The marine model is integrated for the seven sea regions and for sea region specific targets for nutrient loads specified in the Baltic Sea Action Plan (BSAP). The marine transport matrix (between basins) in the former SANBALT model, which was previously integrated in the COST model, has been withdrawn from the BALTCOST model. Thus the BALTCOST model now minimises costs for delivering the required load reduction targets for each sea region separately, rather than for the Baltic Sea as a whole. The reason for this is that the transport of nutrients between sea regions is already taken into account in setting the targets of HELCOM BSAP. Further work integrating dynamic flows of nutrient between the sea regions, as well as the dynamics of nutrient stocks, will be a potential improvement of the model but is outside the scope of this project. Further cooperation with marine modellers is required to achieve this.

The cost-effective distribution of measures within and between drainage basins to achieve the BSAP targets for each Baltic Sea sub basin is investigated. BALTCOST results suggest that it should be possible to achieve the BSAP load reduction targets for N and P in most Baltic Sea regions, with the exception of the P load target in the Baltic Proper and the N reduction target in the Danish Straits, where only 74% and 88% respectively, of the desired BSAP load reductions can be achieved. These results indicate that the targets are hard to achieve, and that more abatement measures are required to fulfil these targets. The minimised total cost of delivering these load reductions across the 9 Baltic littoral countries is estimated to be 4.69 billion Euros, annually. Comparison of N and P reductions at source and at sea shows that nutrient retention is highly influential for the costs fulfilling the load reduction targets. In the (unrealistic) situation in which nutrient retention in the drainage basins is assumed to be zero the estimated costs of delivering the BSAP nutrient reduction targets are much lower (in total 714 million Euro per annum). (See D 8.2.). The comparison made using the RECOCA model (WP 7) identifies the very considerable influence of the spatial resolutions at which the nutrient retentions are modelled. Disregarding retention is clearly unrealistic, and the detailed modelling of nutrient retention is one of the major differences between BALTCOST, RECOCA and earlier models developed to assess cost-minimising strategies for nutrient reductions in the Baltic Sea, and the inclusion of N and P retention estimates is one of the major reasons for the higher costs predicted by BALTCOST than former models.

Task 8.5 Analysis of cost-effective scenarios at the Baltic-wide level under certainty and uncertainty.

Sensitivity analysis of the implications of nutrient retentions within the drainage basins is conducted, using two sets of retention coefficients produced in the catchment modelling by MESAW and the HL method. The BALTCOST model is run with these two sets of retentions, as well as without nutrient retentions, and the results (presented in D 8.2

34

addressing the implications which uncertainty regarding nutrient transport carries for the cost-effective solutions) shows that the inclusion of nutrient retentions has an extremely significant effect over the minimised cost of achieving the required nutrient load reductions. The total costs without retentions are 714 million Euros per annum, compared to the total costs of 4.6 billion Euros per annum using the MESAW retentions.

For most of the 9 Baltic littoral countries individually, the method of retention calculation (MESAW and the HL) makes only a modest difference to the minimised cost of achieving the required load reductions: Sweden and Denmark are exceptions, however. The results presented shows that most measures are undertaken in Poland under all three assumptions (MESAW retentions, HL retentions and no retention), but also that a much larger share of the reduction effort is undertaken in Russia when no retention is assumed, because the retention is lower here compared to the other regions (Stålnacke et al 2011). The effect of the spatial modelling of differences in retention is further explored in D 7.3 with the regional model RECOCA, and these model results also indicate that the spatial modelling at a disaggregated level is important for the model results, and indicates that attention should be paid to this area of research when improving the models for policy analysis.

Task 8.6 Analysis of the differences in informational situation to international and regional decision makers and the implication it has with regard to policy choice and the incentives for strategic decision-making.

This task is analysed in D 8.3: Different nutrient abatement activities jointly determine water quality. Policies are determined by governments at central and local levels, implying that decisions can be affected by strategic considerations. In D 8.3, decentralization of wetland policies is analysed with regard to the environmental and economic consequences. A two-stage game is used to investigate strategic abatement decisions regarding nitrogen fertilizer reductions, Waste Water Treatment Plant (WWTP) phosphorus reductions and wetlands, assuming that wetland decision can be decentralized. This report explains why local governments often hesitate to take on additional responsibilities for environmental management, and identifies conditions under which local governments make smaller losses or even gain from delegation. The results also contribute to an understanding how strategically optimal matching grants are chosen when governments only take into account their own direct costs of abatement and the central government needs to satisfy the local government’s participation constraint.

Publications: Maria Konrad, Hans Estrup Andersen, Hans Thodsen, Mette Termansen, Berit Hasler: Cost-efficient reductions in nutrient loads; optimal spatial policy measures to meet water quality targets at multiple locations. Paper presented at the Danish Environmental Economic Conference at Skodsborg in sept. 2012, In press as scientific article.

Berit Hasler, Jim Smart, Anders Fonnesbech Wulff, Hans Estrup Andersen, Hans Thodsen, Mikolaj Chajkowskij,Christoph Humborg, Erik Smedberg, Per Stålnacke, Cost-effective

35

achievement of reductions in marine eutrophication in the Baltic Sea - the effect of retention and regional scales (Working title, in prep)

James C.R. Smart, Erik Smedberg, Mikolaj Chajkowskij, Anders Fonnesbech Wulff, Berit Hasler Estimating the cost of improving wastewater treatment around the Baltic Sea. Working title, in prep.

Berit Hasler. James C.R. Smart, Anders Fonnesbech Wulff, Maria Konrad, Sisse Jørgensen: BALTCOST – a cost-minimisation model for the Baltic Sea. Documentation report. AU. In prep.

2.9 Work Package 9 – Country Allocation Schemes Lead Partner: Fredrik Wulff, SU

2.9.1 Objective Investigation of the implications of different target formulation for the allocation of abatements and costs. Investigation of the role of uncertainty about catchment nutrient transports for the allocation of abatement and costs

2.9.2 Methodology and scientific achievements Task 9.1 Different alternative targets, such as basin targets, coastal load targets and improvements in environmental indicators are analysed with respect to the cost-effective allocation of abatement resulting from these targets.

Different alternative targets are analysed with BALTCOST in terms of different sea basin targets since the current version of the model (BALTCOST) cannot distinguish between coastal load reductions and open sea reductions. A more comprehensive cooperation with marine modellers is necessary to achieve this. Furthermore we do not currently have the information required to analyse improvements in indicators other than nutrient loads. Such information is not currently available from other partners in RECOCA or from other sources. Further development of BALTCOST could, however, address these types of issues if changes, and improvements in a range of other environmental indicators could be translated in terms of different load reduction requirements from land (e.g. agricultural sources, waste water), air (e.g. emissions from traffic, industry, power plants, other sources) and sea sources (e.g. ships).

In D 9.1 BALTCOST has been used to investigate the cost-effective implementation of the BSAP compared to other targets such as a uniform distribution of the load reduction requirements on all sea regions. The minimised annual total cost of delivering the near BSAP nutrient load reductions across the 9 Baltic littoral countries is estimated to be 4.7 billion Euros. The scenario analyses reported in this work package indicate that around 52% of this total abatement cost is incurred in delivering the final 1165 tonnes (12.5%) of the P load reduction in the Baltic Proper sea region. These disproportionately high costs arise because the available capacity of the more cost effective P abatement measures (improving WWT and restoring wetlands) has been fully utilised in the river basins draining into the Baltic

36

Proper. As a consequence, the more expensive, less effective livestock reduction measures have to be used to deliver the final 1165 tonnes of P abatement.

The costs of uniform reductions are lower than the total costs of the sea region wise targets in the BSAP, but the targets are not fulfilled using a uniform scheme. These results point at the fact that the BSAP distribution between sea regions in itself is not cost-effectively distributed, but also indicate that it is important to pay attention to the distribution of load reductions between the sea regions, and to the transport between them. Further research is needed in this area.

Task 9.2 The role of uncertainty about nutrient transports in the drainage basin for the cost-effective allocation is analysed.

The role of uncertainty regarding nutrient transport in the drainage basins for the cost-effective allocation has been analysed in connection to D 8.2. Therefore D 9.2. has been included in D 8.2. Comparison of N and P reductions at source and at sea shows that nutrient retention is highly influential over the costs incurred in fulfilling the load reduction targets. In the (unrealistic) situation in which nutrient retention in the drainage basins is assumed to be zero the estimated costs of delivering the BSAP nutrient reduction targets are much lower (in total 714 million Euro per annum). The comparison made using the RECOCA model (WP 7) identifies the very considerable influence of the spatial resolution at which the nutrient retentions are modelled exerts over the minimum cost configuration of abatement measures. Disregarding retention is clearly unrealistic, and the detailed modelling of nutrient retention is one of the major differences between BALTCOST, RECOCA and earlier models developed to assess cost-minimising strategies for nutrient reductions in the Baltic Sea.

Publications: Berit Hasler, James C.R. Smart, Anders Fonnesbech Wulff, Fred Wulff et al. Comparison of the cost-effectiveness of ecosystem based and country based allocation schemes for nutrient reductions to the Baltic Sea-. Working title, work in progress. Scientific article.

3. Practical implementation of project outputs (performance statistics 1-4) Number of times your project has contributed to consultations carried out by European Commission.

In total, RECOCA has contributed to 4 consultations by the European Commission. During the RECOCA starting phase in 2009 we were consulted by D.G. Enterprise, European Commission, and the Department of Environment, Spanish National Institute for Agriculture and Food Research and Technology (INIA) for an assessment of ‘The eutrophication risk of phosphates in detergent in the Baltic. This was sponsored by CEEP (a CEFIC Sector Group). Later on in 2009 and 2010 we were also consulted by the European Task Force Group on the EU Marine Strategy Directive for the Quality Descriptor (QD) 5 (Eutrophication). In the final project phase during 2011 RECOCA was presented at the BONUS Forum Coinciding with

37

the 2nd Annual Forum for the EU Strategy for the Baltic Sea Region and Baltic Development Forum, 24 October 2011, Gdansk, Poland and during the presentation of BONUS+ highlights to the European community, 8 November 2011.

Number of times the scientists working in your Project have served as members or observers in stakeholder and scientific committees.

The data base and hydrological-biogeochemical models (WP2-5) have been presented for the HELCOM TARGREV project in June 2010 and in follow up meetings in 2011 that will update the environmental targets within the next phase of the BSAP. Our model tools and data will be the basis for the “country allocation scheme”, i.e. the distribution of nutrient reductions needed per country based on the new targets. In total, RECOCA scientists have participated in >10 HELCOM meetings of various characters, i.e. HELCOM LOAD Expert Group, HELCOM TARGREV group and acted as observers and advisers in HELCOM MONAS group.

Number of times the effort of your Project has resulted in modifications made to relevant policy documents and action plans (in particular, Baltic Sea Action Plan)

RECOCA scientist participated in the 15TH MEETING OF HELCOM MONITORING AND ASSESSMENT (see D 1.3) The Meeting noted that the HELCOM LOAD group will now actively begin the revision of the nutrient reduction targets and stressed the importance of all countries to participate in this work. The Meeting took note of the proposals for country-wise allocations principles for reduction targets and the draft roadmap towards new-country-wise principles for allocation schemes and welcomed the allocation scenarios illustrating the application of principles as presented by BNI scientists presenting the RECOCA data base (WP2) and CSIM model (WP5) as developed in RECOCA as a major tool to calculate the necessary scenarios.

Number of suggestions for designing, implementing and evaluating the efficacy of pertinent public policies and governance originating from the work of your Project

During the final ECOSUPPORT/RECOCA Stakeholder conference, 7 December 2011, Stockholm, Sweden RECOCA scientist presented “An outlook to the future Baltic Sea: how can we reach the targets of the Baltic Sea Action Plan?" to some 80 invited scientists and stakeholders. The stakeholder comprised of delegates from the office of the Swedish prime minister, the ministries of agriculture and environment from Sweden, agricultural NGOs from Sweden and other Baltic Sea riparian countries, and many more stakeholder and scientists. Detailed consequences with respect to necessary fertilizer and livestock reductions and necessary costs to fulfil the BSAP have been presented to these stakeholders. More information under: http://www.balticnest.org/balticnest/activities/events/events/ecosupportandrecocastakeholderconference.5.50fc7668132ea9eba0a8000143.html

38

4. Comparison with the original research and financial plan, Overall, all major tasks as formulated in the research plan have been addressed during the project period although the interface to NEST of some components are not ready. Likewise all possible measures/costs have not yet been implemented in the economic models (WP7 and WP8). These tasks will be completed during spring 2012.

5. Statement if the research plan and schedule of deliverables had to be adapted No. Only minor changes see point 4 above. 6. Further research and exploitation of the results (only the Final report) The databases, the catchment model CSIM, which integrates major results from the dynamic models, and results from the cost-optimization model, will be made available via the Nest system (www.balticnest.org ). Interactive on-line visualization through the Nest system will allow policy makers to view the nutrient load reductions and cost allocations and compare result from different abatement strategies and target allocations.