Periodic Reporting for period 1 - AQUISS (Air Quality Information Services)
Période du rapport: 2016-06-01 au 2017-02-28
Résumé du contexte et des objectifs généraux du projet
Over 70% of the European population lives in cities and this percentage grows over time. Air pollution is a threat to health and many cities do not satisfy regulatory requirements of sufficiently clean air. Air quality (AQ) forecasts with high spatial resolution usable at city level are an important component to cope with this global challenge. Today such forecasts are only available using expensive dispersion models, too complex to be operated by many city authorities and similar non-scientific users. Major obstacles are high cost and intricacies of daily operation.
The AQUISS innovation idea is to offer a service which delivers AQ forecasts for urban users, which are sufficiently accurate for practical application, avoiding the use of complex dispersion models. For this purpose, existing methods of spatial statistics are used to enrich two existing EU-wide data resources: Airbase (European Environment Agency) and CAMS (Copernicus).
Anticipated users are city councils, media (e.g. press, local radio stations), weather services and value-adding SME’s, who want to use AQ forecasts in their business without operating complex modelling services themselves. Service delivery will be in a cloud-based fashion, making it significantly simpler than today and using new business models. The commercial potential is up to several hundred users in Europe.
The AQUISS project was a stage-1 SME action in H2020-INSO-10-2015 to assess the feasibility of the innovation idea and to advance the initial business plan towards an elaborated business plan. All elements of the business idea existed at TRL7 to TRL9 at project start. The technical innovation is to combine existing EU-wide AQ data sources with statistical methods to provide high resolution forecasts. Innovation in service delivery is to use a cloud-environment relieving users from daily operation and offering interfaces for value-adding third parties. The project has concentrated on the feasibility assessment of the business model for different customers and practical viability of the innovation idea.
The results of the project are: a) a real live demonstrator in large German city, which was originally not planned but is used as a showcase to communicate with users, b) an elaborated business plan, and c) a number of dissemination and exploitation activities.
Both from a technical and business viewpoint, feasibility of the innovation idea has been shown. There are some concerns regarding the spatial resolution of base datasets, availability of enough AQ stations for spatial interpolation, forecast quality and trust of end users. All these concerns can be addressed, for instance by improved data preparation and continued support of end users.
The AQUISS innovation idea is to offer a service which delivers AQ forecasts for urban users, which are sufficiently accurate for practical application, avoiding the use of complex dispersion models. For this purpose, existing methods of spatial statistics are used to enrich two existing EU-wide data resources: Airbase (European Environment Agency) and CAMS (Copernicus).
Anticipated users are city councils, media (e.g. press, local radio stations), weather services and value-adding SME’s, who want to use AQ forecasts in their business without operating complex modelling services themselves. Service delivery will be in a cloud-based fashion, making it significantly simpler than today and using new business models. The commercial potential is up to several hundred users in Europe.
The AQUISS project was a stage-1 SME action in H2020-INSO-10-2015 to assess the feasibility of the innovation idea and to advance the initial business plan towards an elaborated business plan. All elements of the business idea existed at TRL7 to TRL9 at project start. The technical innovation is to combine existing EU-wide AQ data sources with statistical methods to provide high resolution forecasts. Innovation in service delivery is to use a cloud-environment relieving users from daily operation and offering interfaces for value-adding third parties. The project has concentrated on the feasibility assessment of the business model for different customers and practical viability of the innovation idea.
The results of the project are: a) a real live demonstrator in large German city, which was originally not planned but is used as a showcase to communicate with users, b) an elaborated business plan, and c) a number of dissemination and exploitation activities.
Both from a technical and business viewpoint, feasibility of the innovation idea has been shown. There are some concerns regarding the spatial resolution of base datasets, availability of enough AQ stations for spatial interpolation, forecast quality and trust of end users. All these concerns can be addressed, for instance by improved data preparation and continued support of end users.
Travail effectué depuis le début du projet jusqu’à la fin de la période considérée dans le rapport et principaux résultats atteints jusqu’à présent
Task 1: Management
Overview of task. The management in this rather small project was very lean and efficient as no additional partners were involved. The main activities in Task 1 were:
• Setting up of a project workplan based on the content of the technical Tasks 2-4 and on Task 5.
• Follow up of the project workplan. On a monthly basis the progress was compared with the workplan. Whenever necessary the workplan was revised.
• Revision of the work plan to accommodate the changed approach as explained in section 1.1.
• Selecting and contracting an expert with scientific background to strengthen the development of the methodology and the pilot application.
• Taking care of all contractual, financial, communication and reporting activities with the EU-Commission.
• Performing a contract amendment.
Overview of results. Successful project implementation; one contract amendment.
Artefacts. Amended contract; Final report.
Task 2: Technical feasibility assessment
Overview of task. As explained in section 1.1 this task was converted to the prototype implementation of a concrete technical pilot or „show case“, for the greater area of Munich, based on a product description which is part of the business plan. Details are shown in section 2.1.
Overview of results. From a technical viewpoint the AQUISS innovation idea is feasible, as shown by the Munich pilot. The necessary public data are available, in the necessary quality, quantity and frequency. Data fusion of the various input data sets works and has been implemented using standard statistical tools, namely a Neural Network (NN) from an R-library. The spatial mapping in the Munich pilot is rudimentary and maps stations to land use directly without inter- or extrapolation. This would have to be improved in a real-world product, for instance by Kriging methods, to generate a better spatial representation of the situation.
The Munich pilot has shown that the spatial interpolation may have to be implemented differently for each city, depending on the concrete situation. It looks unlikely that one method can be made so generic that it needs no adaptation at all. This will have effects on the business plan because this will create increased initial set-up costs that originally envisaged. Also, the quality of the land use data sets may be a concern and may cause some costs when customizing the product for a given site. The future COPERNICUS project will generate valuable information in this respect for the future evolvement of the product.
There is one concern, which is the availability of enough AQ stations in the chosen city to generate sufficient spatial resolution. While this would not impact the forecasts as such (i.e. there would still be a valid product), the “spatial content” of the product would be less effective. This situation could be overcome by using additional AQ stations (not represented in EEA airbase) or by deploying low-cost sensors in the city, which might lead to an additional business idea, namely the operation of such low cost sensor networks on behalf of customers.
Artefacts. Product description document; Munich pilot.
Task 3: Business model
Overview of task. The core of this task was the investigations to what extent the AQUISS approach and expected product would have a chance to be taken up by the market. The business model has been set-up based on the minimum viable product” (MVP) and the CANVAS business model approach.
Overview of results. The task has been performed in accordance with the plan. Two market segments have been addressed. The Urban Market Segment comprises European Cities with a population between 250,000 and 1,000,000 inhabitants who a facing environmental problems but are lacking resources and expertise that prevents them to establish complex modelling and forecasting systems. An analysis of the most recent data on European cities shows that there are 130 cities with a population of between 250,000-500,000 and 46 cities with a population between 500,000 and 1,000,000 inhabitants. We have set our target to reach 15% of these cities resulting in roughly 25 larger cities as potential clients.
The Media Market Segment is much more difficult to assess, as the market is rather fragmented. As a rough assessment we assumed that in each of the 25 cities are 3 newspapers and 3-5 regional online magazines. This will leave us with 75 newspapers and 75-125 online magazines as potential clients.
Due to the fact, that the Munich Demonstrator was developed with real data under real-life conditions it was possible to assess the efforts for customizing and operating the AQUISS services based on the work. This part of the business plan is more robust than expected.
The business plan has been developed using the CANVAS model. This is a well know approach applied by a broad community. It offers an excellent opportunity to not only visualize the various elements of a business model, it also stimulates thinking, discussions and iterations to refine the model. In addition, the approach gives evidence to risks and crucial assumptions and hypotheses which need to be investigated or tested with potential clients.
The details of the business plan and the main results are described in deeper detail in section 2.2.2. We assume that we reach the market penetration rate of 25 cities within 5 years; we also assume that we can sell the service in parallel to two media (either newspaper or on-line magazines) in the same city/region. This will result in a total revenue of 112,500€ in year five.
Artefacts. Elaborated business plan.
Task 4: ICT implementation plan
Overview of task. The original objective of this task was to develop an implementation plan for a concrete software configuration which can be readily put in place when the service is being rolled out, also as a preparation for a stage-2 SME action. With the change of project approach, this task became covered by the concrete implementation of the Munich pilot, as part of task T2. With the successful bid for the COPERNICUS tender (see also 1.1) it would have been a waste of resources to carry out this task as initially planned. Instead some efforts outside the AQUISS project were put into the development of an initial software architecture. Nevertheless, the efforts undertaken delivered a clear indication of how much efforts are needed to set-up and run the AQUISS service. In so far, the underlying figures of the business plan are more robust than in the initial approach.
Overview of results. The result of this task is now the ICT implementation of the Munich pilot, which is a rudimentary implementation that includes data preparation and fusion software, the training of the NN, the actual execution of the forecasts and spatial mapping. A robust assessment of efforts needed for initial set-up has been derived from this activity
Artefacts. ICT implementation of Munich pilot.
Task 5: Dissemination and implementation
Overview of task. To stimulate the commercialisation of the AQUISS innovation and the resulting product dissemination is the key to get in touch with the target communities. The main aim of this task was to develop a dissemination approach and dissemination material that conveys the key messages in a way that potential clients understand the added value of AQUISS. Task 5 has produced a project dissemination website (www.aquiss.eu) and a brief description of the AQUISS product (factsheet/flyer). Using this material mailings and interviews have been conducted to get in contact with potential clients and to get their initial feedback.
Project website: The project website (www.aquiss.eu) has been set up and maintained throughout the project lifetime. This goes beyond the creation of the website. Through the graphical design and the graphical elements (e.g. AQUISS logo) the project has created its identity. In a second step, the key-message “„Air Quality information service as simple and reliable as a weather forecast“ has been used to illustrate the aim of the project. Everybody knows what a weather forecast is, this association it easily understood by non-experts and gives the impression of what the product is going to offer. The site will be operational for the next 2-3 years to further promote the concept. The site will see updates as the AQUISS service develops.
Factsheet/Flyer: A fact sheet that describes the AUQISS approach and added value has been produced in electronical format. This has been used when potential clients have been approached.
Mailings & Interviews: At the beginning of the project we have prioritized two main target groups namely the media and smaller- and medium size cities that don’t have the necessary resources to have their own AQ forecast. In a first step a limited number of potential clients have been approached and interviewed. These interviews haven’t followed a pre-defined structure. The aim was to understand whether the contact was interested in AQ issues, whether AQ is seen as a priority theme and whether the AQUISS solution would trigger the interest to receive further information.
The main results from this first round were not very encouraging. It was obviously rather difficult to explain the innovation of AQUISS without having a concrete example in hands. Furthermore, some doubts regarding the credibility of the approach and the expected results became evident. It seems that AQ information is expected to be delivered with stronger scientific backing. As a result of this initial feedback we decided to change the approach of AQUISS. We decided to prepare a concrete pilot application for the greater area of Munich (see 1.1 and 2.1.6). The intention was that the added value of AQUISS could be underpinned by a graphical representation of the AQ situation in Munich.
Overview of results. Project dissemination Website; Project factsheet/flyer
Artefacts. Website, Product flyer.
Overview of task. The management in this rather small project was very lean and efficient as no additional partners were involved. The main activities in Task 1 were:
• Setting up of a project workplan based on the content of the technical Tasks 2-4 and on Task 5.
• Follow up of the project workplan. On a monthly basis the progress was compared with the workplan. Whenever necessary the workplan was revised.
• Revision of the work plan to accommodate the changed approach as explained in section 1.1.
• Selecting and contracting an expert with scientific background to strengthen the development of the methodology and the pilot application.
• Taking care of all contractual, financial, communication and reporting activities with the EU-Commission.
• Performing a contract amendment.
Overview of results. Successful project implementation; one contract amendment.
Artefacts. Amended contract; Final report.
Task 2: Technical feasibility assessment
Overview of task. As explained in section 1.1 this task was converted to the prototype implementation of a concrete technical pilot or „show case“, for the greater area of Munich, based on a product description which is part of the business plan. Details are shown in section 2.1.
Overview of results. From a technical viewpoint the AQUISS innovation idea is feasible, as shown by the Munich pilot. The necessary public data are available, in the necessary quality, quantity and frequency. Data fusion of the various input data sets works and has been implemented using standard statistical tools, namely a Neural Network (NN) from an R-library. The spatial mapping in the Munich pilot is rudimentary and maps stations to land use directly without inter- or extrapolation. This would have to be improved in a real-world product, for instance by Kriging methods, to generate a better spatial representation of the situation.
The Munich pilot has shown that the spatial interpolation may have to be implemented differently for each city, depending on the concrete situation. It looks unlikely that one method can be made so generic that it needs no adaptation at all. This will have effects on the business plan because this will create increased initial set-up costs that originally envisaged. Also, the quality of the land use data sets may be a concern and may cause some costs when customizing the product for a given site. The future COPERNICUS project will generate valuable information in this respect for the future evolvement of the product.
There is one concern, which is the availability of enough AQ stations in the chosen city to generate sufficient spatial resolution. While this would not impact the forecasts as such (i.e. there would still be a valid product), the “spatial content” of the product would be less effective. This situation could be overcome by using additional AQ stations (not represented in EEA airbase) or by deploying low-cost sensors in the city, which might lead to an additional business idea, namely the operation of such low cost sensor networks on behalf of customers.
Artefacts. Product description document; Munich pilot.
Task 3: Business model
Overview of task. The core of this task was the investigations to what extent the AQUISS approach and expected product would have a chance to be taken up by the market. The business model has been set-up based on the minimum viable product” (MVP) and the CANVAS business model approach.
Overview of results. The task has been performed in accordance with the plan. Two market segments have been addressed. The Urban Market Segment comprises European Cities with a population between 250,000 and 1,000,000 inhabitants who a facing environmental problems but are lacking resources and expertise that prevents them to establish complex modelling and forecasting systems. An analysis of the most recent data on European cities shows that there are 130 cities with a population of between 250,000-500,000 and 46 cities with a population between 500,000 and 1,000,000 inhabitants. We have set our target to reach 15% of these cities resulting in roughly 25 larger cities as potential clients.
The Media Market Segment is much more difficult to assess, as the market is rather fragmented. As a rough assessment we assumed that in each of the 25 cities are 3 newspapers and 3-5 regional online magazines. This will leave us with 75 newspapers and 75-125 online magazines as potential clients.
Due to the fact, that the Munich Demonstrator was developed with real data under real-life conditions it was possible to assess the efforts for customizing and operating the AQUISS services based on the work. This part of the business plan is more robust than expected.
The business plan has been developed using the CANVAS model. This is a well know approach applied by a broad community. It offers an excellent opportunity to not only visualize the various elements of a business model, it also stimulates thinking, discussions and iterations to refine the model. In addition, the approach gives evidence to risks and crucial assumptions and hypotheses which need to be investigated or tested with potential clients.
The details of the business plan and the main results are described in deeper detail in section 2.2.2. We assume that we reach the market penetration rate of 25 cities within 5 years; we also assume that we can sell the service in parallel to two media (either newspaper or on-line magazines) in the same city/region. This will result in a total revenue of 112,500€ in year five.
Artefacts. Elaborated business plan.
Task 4: ICT implementation plan
Overview of task. The original objective of this task was to develop an implementation plan for a concrete software configuration which can be readily put in place when the service is being rolled out, also as a preparation for a stage-2 SME action. With the change of project approach, this task became covered by the concrete implementation of the Munich pilot, as part of task T2. With the successful bid for the COPERNICUS tender (see also 1.1) it would have been a waste of resources to carry out this task as initially planned. Instead some efforts outside the AQUISS project were put into the development of an initial software architecture. Nevertheless, the efforts undertaken delivered a clear indication of how much efforts are needed to set-up and run the AQUISS service. In so far, the underlying figures of the business plan are more robust than in the initial approach.
Overview of results. The result of this task is now the ICT implementation of the Munich pilot, which is a rudimentary implementation that includes data preparation and fusion software, the training of the NN, the actual execution of the forecasts and spatial mapping. A robust assessment of efforts needed for initial set-up has been derived from this activity
Artefacts. ICT implementation of Munich pilot.
Task 5: Dissemination and implementation
Overview of task. To stimulate the commercialisation of the AQUISS innovation and the resulting product dissemination is the key to get in touch with the target communities. The main aim of this task was to develop a dissemination approach and dissemination material that conveys the key messages in a way that potential clients understand the added value of AQUISS. Task 5 has produced a project dissemination website (www.aquiss.eu) and a brief description of the AQUISS product (factsheet/flyer). Using this material mailings and interviews have been conducted to get in contact with potential clients and to get their initial feedback.
Project website: The project website (www.aquiss.eu) has been set up and maintained throughout the project lifetime. This goes beyond the creation of the website. Through the graphical design and the graphical elements (e.g. AQUISS logo) the project has created its identity. In a second step, the key-message “„Air Quality information service as simple and reliable as a weather forecast“ has been used to illustrate the aim of the project. Everybody knows what a weather forecast is, this association it easily understood by non-experts and gives the impression of what the product is going to offer. The site will be operational for the next 2-3 years to further promote the concept. The site will see updates as the AQUISS service develops.
Factsheet/Flyer: A fact sheet that describes the AUQISS approach and added value has been produced in electronical format. This has been used when potential clients have been approached.
Mailings & Interviews: At the beginning of the project we have prioritized two main target groups namely the media and smaller- and medium size cities that don’t have the necessary resources to have their own AQ forecast. In a first step a limited number of potential clients have been approached and interviewed. These interviews haven’t followed a pre-defined structure. The aim was to understand whether the contact was interested in AQ issues, whether AQ is seen as a priority theme and whether the AQUISS solution would trigger the interest to receive further information.
The main results from this first round were not very encouraging. It was obviously rather difficult to explain the innovation of AQUISS without having a concrete example in hands. Furthermore, some doubts regarding the credibility of the approach and the expected results became evident. It seems that AQ information is expected to be delivered with stronger scientific backing. As a result of this initial feedback we decided to change the approach of AQUISS. We decided to prepare a concrete pilot application for the greater area of Munich (see 1.1 and 2.1.6). The intention was that the added value of AQUISS could be underpinned by a graphical representation of the AQ situation in Munich.
Overview of results. Project dissemination Website; Project factsheet/flyer
Artefacts. Website, Product flyer.
Progrès au-delà de l’état des connaissances et impact potentiel prévu (y compris l’impact socio-économique et les conséquences sociétales plus larges du projet jusqu’à présent)
There are two comprehensive sources of Air Quality (AQ) Data in Europe available. First, the Member States operate large and complex monitoring networks that measure AQ. Since decades, concentrations of pollutants such as SO2, NO2, CO, PM10, PM2.5 and Ozone have been automatically monitored by governmental agencies throughout the EU. These data are accumulated in national AQ databases and through reporting obligations of the Member States are further integrated at European level at European Environment Agency (EEA) in the so-called AirBase . This data are high quality point data, i.e. it shows data at the location of the monitoring stations only. Second, EU-wide AQ modelling services have been made available by the research community, initially stemming from the MACC project (now part of the COPERNICUS service CAMS ). Due to the complexity of the dispersion modelling process and the required computational power these data are provided as wide-scale grid data (grid sizes vary between 50 * 50 km and 15*15 km). Although both data sets are freely available through the legal framework of INSPIRE , COPERNICUS and EEA data policy , traditional business does not deliver what is needed in cities and urban agglomerations, as important elements of the value chain are missing and key market requirements (particularly spatial resolution) are not met.
The monitoring data provided by AirBase delivers only point-related information missing any spatial representation of the AQ situation. Furthermore, no forecast is possible by means of monitoring. The large-scale data provided by CAMS is of very limited use on city level because of its coarse spatial resolution. Another possibility would be to set-up and run a complex dispersion model for a given area (e.g. city agglomeration). This calls for expertise and resources medium-size city (e.g. between approx. 250,000 – 1,000,000 inhabitants) usually don’t have.
The forecast methodology tested and applied for the Munich region overcomes these limitations. By combining freely available data sources in an intelligent manner through the deployment of Neuronal Network. The data sets used were:
• historical CAMS data,
• historical weather data (Global Forecasting System GFS)
• historical AQ station monitoring data (AirBase)
• Land use data (CORINE Land Cover)
• Street network (Open Street Map)
The methodology is in the position to deliver an air quality map for an urban area with reasonable accuracy. The demonstrator for the Munich area is the proof of concept. In addition it allowed assessing the efforts needed which are much less than setting up a complex dispersion model.
Conclusions
Although it became evident that the AQUISS innovation idea as presented in the project proposal needs further enhancements to be fit for the market, we have demonstrated that the methodology is valid and robust enough to continue both technical and business development.
The interim results of AQUISS were convincing enough to win a tender of ECWMF for a new project that will be part of the COPERNICUS program, and thus gain high visibility. The project will implement a pilot in Central Europe and will further elaborate on the methodology, add a systematic validation and will continue to approach the market. The start date will be 1st July 2017 and the duration 18 months.
The monitoring data provided by AirBase delivers only point-related information missing any spatial representation of the AQ situation. Furthermore, no forecast is possible by means of monitoring. The large-scale data provided by CAMS is of very limited use on city level because of its coarse spatial resolution. Another possibility would be to set-up and run a complex dispersion model for a given area (e.g. city agglomeration). This calls for expertise and resources medium-size city (e.g. between approx. 250,000 – 1,000,000 inhabitants) usually don’t have.
The forecast methodology tested and applied for the Munich region overcomes these limitations. By combining freely available data sources in an intelligent manner through the deployment of Neuronal Network. The data sets used were:
• historical CAMS data,
• historical weather data (Global Forecasting System GFS)
• historical AQ station monitoring data (AirBase)
• Land use data (CORINE Land Cover)
• Street network (Open Street Map)
The methodology is in the position to deliver an air quality map for an urban area with reasonable accuracy. The demonstrator for the Munich area is the proof of concept. In addition it allowed assessing the efforts needed which are much less than setting up a complex dispersion model.
Conclusions
Although it became evident that the AQUISS innovation idea as presented in the project proposal needs further enhancements to be fit for the market, we have demonstrated that the methodology is valid and robust enough to continue both technical and business development.
The interim results of AQUISS were convincing enough to win a tender of ECWMF for a new project that will be part of the COPERNICUS program, and thus gain high visibility. The project will implement a pilot in Central Europe and will further elaborate on the methodology, add a systematic validation and will continue to approach the market. The start date will be 1st July 2017 and the duration 18 months.