Final Report Summary - EUFODOS (European Forest Downstream Services- Improved Information on Forest Structure and Damages)
Executive Summary:
European Forest Downstream Services - Improved Information on Forest Structure and Damage
Forests play a key role in the European economy and environment. This role incorporates ecological as well as economic functions which can be affected by the occurrence of insect infestations, storms, fire or snowbreak events. Local, regional and national authorities thus require detailed information on the degradation status of their forests to be able to take appropriate countermeasures against forest damage and to ensure sustainable forest management. The EUFODOS project has proven that the use of state-of-the-art satellite and laser scanning technology can provide forest authorities with cost-effective, timely and comprehensive information on forest structure and damage.
The main goals
The Forest Downstream Services (FDS) within EUFODOS include in particular the assessment of forest damage and the measurement of functional parameters for commercial and protective forests. These monitoring services have been developed to an operational level by a consortium of research organisations and commercial service providers from Austria, Germany, Finland, Italy, Bulgaria and Poland.
The benefits
The use of space- and air-borne sensor platforms allows data to be acquired at short time intervals, in a cost effective way and over large areas. EUFODOS Forest Downstream Services thus offer the derivation of fast and reliable information for effective damage assessment and sustainable forest management. The products present space-related information to be used by users for a wide range of applications, such as:
- Effective damage assessment and countermeasures: identification of damaged areas – due to storm, fire, snow-break or insect infestations – in order to enable proper countermeasures, compensation payments and reforestation planning
- Sustainable management of protective forests: targeted management of protection forests in order to maintain and enhance their protective function against natural hazards
- Sustainable management of commercial forests: wood procurement planning and strategic investment planning for commercial forests
- Reporting: revision of forest maps and inventories, compilation of regular reports and annual statistics (e.g. changes in forested area), establishment of forest damage information system.
The EUFODOS consortium has reported on the results to an extensive forest user community via the newsletter distribution and still maintains the project webpage www.eufodos.info
Project Context and Objectives:
The European Earth Observation Programme GMES, currently renamed to COPERNICUS, provides data useful in tackling a range of issues including climate change and citizens’ security. The purpose of COPERNICUS is to deliver information which corresponds to user needs. In EUFODOS regional forest downstream services have been developed which are focusing on degradation issues such as the assessment of damage caused by disastrous events from storm, fire, snow-break or insect infestations. These services, which have been successfully implemented in pilot testcases, are demanded strongly by the regional users in order to fulfil their mandate. The EUFODOS services were developed to an operational level.
The use of space- and air-borne sensor platforms allows data to be acquired at short time intervals, in a cost effective way and over large areas. The Forest Downstream Services thus provide fast and reliable information for effective damage assessment and sustainable forest management. The data can be utilized by users in a wide range of applications:
• Effective damage assessment and countermeasures: identification of damaged areas – due to storm, fire or insect infestations – in order to enable proper countermeasures, compensation payments and reforestation planning.
• Sustainable management of protective forests: targeted management in order to maintain and enhance their protective function against natural hazards.
• Sustainable management of commercial forests: wood procurement planning and strategic investment planning for commercial forests.
• Reporting: revision of forest maps and inventories, compilation of regular reports and annual statistics (e.g. changes in forested area), establishment of forest damage information systems.
EUFODOS involves an extensive user community well connected to other related GMES/COPERNICUS User Groups. A list of 392 potential users from 46 countries and 12 international organisations was compiled, enabling the Service Providers to contact potential users in the European countries.
Project Results:
1.1 Frame Conditions for DS Operations
1.1.1 Objectives and Results
This WP, managed by GAF AG, aimed to develop overarching organisational frame conditions for an effective beneficiary interaction, for utilisation of the Forest Core products and for effective implementation of Forest Downstream Services. This comprised the formation of a Service Network of Service Providers, Users and the Research Community, as well as the development of a Service Partnership Protocol (co-operation procedures and communications) and the organisation of a User Executive Body (UEB). Another focus was on the development of a consistent IPR Strategy for the consortium and the harmonisation of the EUFODOS spatial data and services with the INSPIRE standards.
These requirements have been addressed as foreseen in the first 9 months of the project and have been documented in the Deliverable D210.1 Service Architecture Components. This Deliverable provides a review and outline of all the main components that form the relevant building blocks, or service architecture components, for the EUFODOS project implementation. This comprises an assessment of the different requirements for implementing various Forest Downstream Services for different users in Europe, thus providing the EUFODOS service development frame conditions. This allows a further assessment of how an optimal trade-off between the Forest Downstream Service requirements, user needs and technology constraints can be achieved. The final version of the Deliverable D210.1 for Phase 1 has been delivered in April 2011.
Most of the concepts and issues presented in the Deliverable D210.1 have a legacy in the GMES Service Element Forest Monitoring (GSE FM) which successfully provided Forest Downstream Services (FDS) to a variety of users in Europe and internationally. The report first presented the Organisational Structure of EUFODOS, bringing together key players from the research community, Service Providers and the user community and further presents the User Executive Body and explains its role within the Consortium and the project, especially to contribute to report reviews, identify user needs and representing the user community. All procedures related to Management Issues such as Reporting, Decision Making and Conflict Resolution as well as collaboration procedures with associated projects have been described in the Deliverable. The Deliverable also described the EUFODOS dissemination strategy, which aimed at the assessment of new markets and the extension of the user community. The Deliverable described the primary tools for dissemination, i.e. the project website, user federation mechanisms (i.e. the UEB), brochures, publications and service related trainings and workshops.
The Service Architecture Components Deliverable (D210.1) presented also the relevant regulations related to the entry / exit for new / current members further described in the EUFODOS Consortium Agreement, as well as the description of production procedures for Forest Downstream services and products. Troubleshooting, Quality Control procedures and standards, adopted for the various steps of the production process, are also described. The important issue of Intellectual Property Rights (IPR) has also been extensively explored and is described in the Deliverable D210.1 in terms of identifying the main IPR aspects that will affect EUFODOS and which hitherto had been addressed only in a limited manner in the EUFODOS Consortium Agreement. The FP7 IP Rules have been examined in detail and recommendations on how to manage IPR related to Ownership, Protection of Foreground, Access Rights to Background, and Dissemination are provided as strategies for the EUFODOS project to implement. Another key component explored in WP 210 and presented in D210.1 is the INSPIRE relevant harmonisation process necessary for Service implementation. This WP had a significant impact in the project definition as it outlined all the main components that formed the building blocks or service architecture for the EUFODOS project implementation. The first issue had been released in Phase 1 and an update had been submitted to REA in September 2012 (Phase 2 of EUFODOS).
1.2 Co-ordination with GMES Core and Analysis of Core Products Existing Systems
1.2.1 Objectives and Results
The Work Package has been managed by GAF and started in May 2011 (Month 5).
During Month 18 and Month 27, the two deliverables, D220.3 Assessment Report of Downstream Inputs and the Report on Coordinated Actions with GMES Core (D220.4) have been successfully updated from Phase 1 and submitted to REA.
The Assessment Report of Downstream Inputs assessed the potential data sources available to the GMES Forestry Downstream Services as provided by EUFODOS. It considered GMES core services and other „core“-type data sources, as well as to a minor extent also national data. A special focus had been put on core Earth Observation (EO) data as well as on Forest core products as produced by the GIO-Land project, managed by EEA. Data access conditions and procedures for the relevant Forest Downstream service input data, which are available free of charge and without restrictions, have been described. Relevant information to access EO data via the GMES Data Warehouse has been updated (e.g. information on RSS Feeds from ESA) and the most recent version 2.6 of the DAP document has been introduced. The section on the upcoming Sentinel-2 mission, one of the most relevant (for the successful implementation of Downstream Services in the future) EO HR satellite sensors has been updated accordingly.
The section on the High Resolution Forest Core products got a major update related to data availability of Forest Core products, produced within geoland2 (geoland2 project ended Dec 2012) and the GMES Initial Operations (GIO Land) project (2012-2014), managed by EEA. The GIO-Land project will provide operational Core Land products such as the HR Forest Layers (Tree Cover Density, Forest Type Layer) for EEA-39 countries based on Image 2011/12/13 data. Compared to geoland2, different HR layer product specifications have changed (e.g. Tree Cover against Forest Cover in geoland2). For future implementation of Forest Downstream Services, the new product specifications have to be taken into account and if possible, tested by EUFODOS service providers.
Another update has been provided about the GMES Emergency Management Service (GIO-EMS) project, the operational follow-up project to the FP7 project SAFER. The GMES Emergency Management Service (EMS) is one of the first GMES services that became operational. It started in April 2012 and consists of a set of services which are funded by the European Commission in the frame of the GMES Initial Operations (GIO). The GIO-EMS services are providing mapping products based on satellite imagery to actors in the management of natural and man-made disasters, in particular Civil Protection Authorities and Humanitarian Aid actors. There is no direct interference between Forest Downstream Services and the GIO-EMS project. Nevertheless, coordination e.g. in the field of EO data access/acquisition via the GMES Data Warehouse could be envisaged in the near future.
Within report D220.4 the status and potential coordination activities with former FP7 (e.g. geoland2, SAFER) and GMES Initial Operation projects (e.g. GIO-Land, GIO-EMS) have been updated. Important to know is, that no more-recent information is available on the final GIO data access policy by early 2013, except that the European Commission generally pursues and prepares a “free and open” access policy to all GMES data, which is deemed to also cover the GIO-Land HR Layers. A direct link and further communication of all developments to the project EUFODOS has been described and has been ensured by the direct involvement of GAF as a service provider in GIO-Land. Since several restrictions apply to the further distribution of internal knowledge that GAF has from this involvement in GIO, separate communications have been established between GAF and EEA on which service and product specifications. This has also led to a EUFODOS newsletter publication (EUFODOS 2013), providing all the necessary insight.
The product and service portfolio of the GIO-EMS comprises also reference and delineation maps similar to the type of services provided by EUFODOS. Specifically, burnt area maps are mentioned which are likely to be related also to forest areas. Therefore an information exchange of EUFODOS with the future ERCSGIO-EMS follow-up services under GIO should be envisaged and has been established by the WP 220 lead to avoid potential duplications in storm damage assessment during this operational phase of the ERCSGIO-EMS from the beginning. It has been agreed that the GIO-EMS shall inform EUFODOS immediately in case of activations in relation to Forest damage (by storm, fires, insects, etc.) within Europe. This procedure is deemed to avoid duplications in EO data orders and damage mapping.
1.3 Method for Utilizing Core Service Products
1.3.1 Objectives and Results
The aim of this WP, “Method for Utilizing Core Service Products” is to realise a cost and time saving procedure when utilizing the Core Service (CS) products for the generation of the Downstream Service (DS).
The key findings are:
• Availability of the Core Service: Only 2 out of 6 services cases are completely covered by geoaland2 HR layers. For 2 other cases, HR-layers were produced by GAF in the scope of EUFODOS, 2 test sites are not covered by HR-layers and HR-Layers had to be simulated.
• Use of the HR Forest Layers in EUFODOS: All service providers have implemented the HR-layers (or simulated products) in their processing lines. In most cases, the HR-Layers were used for masking forest areas in various stages of processing (e.g. for change detection). Furthermore, the HR-layers were used for training the classification processes, validation or as an initial input for a forest density map.
• Cost and time saving has been seen as the main added value of the integration of the HR-layers into the processing chains. Additionally, the integration of the HR-layers in parallel with other input data can improve accuracy and robustness of various EUFODOS products.
• Related to the potential advantages of integrating Core Service forest HR-layers are a couple of bottlenecks which could be improved
o Missing coverage: in EUFODOS, not all areas were covered by the Core Service products. In case of a commercial DS this would create additional costs
o Update frequency: For most EUFODOS DS products an update frequency of 1 year would be preferred. For damage assessment up-to date data for pre-event (reference) as well as for post-event is necessary.
o Accuracy, resolution and Minimum Mapping Unit (MMU): to detect damages in forest, in particular small scale damages (e.g. due to insect infestations), a high accuracy (>85%), high resolution (<= 10m) and a relatively small MMU (< 0,5ha) are required.
o Documentation and Metadata: to be able to better assess the quality of the HR-Layers a better documentation would be necessary (how was the dataset produced? Which data was used for validation? Validation report).
o For damage assessment a fast delivery of data would be essential. A solution could be to consider the rush-mode of the ESA Data Warehouse.
A problem to evaluate the full potential of the Core Service HR-Layer data within EUFODOS is the fact that EUFODOS is working with the geoland2 HR-Layers, while in the meanwhile EEA has contracted the production of the GIO HR layers with technical specifications different from the geoland2 products. Therefore, processing lines would have to be adapted to the new products as soon as they are available.
1.4 Standardisation of Production Chains
1.4.1 Objectives and Results
The WP has been managed by GAF AG and had contributions from all EUFODOS Service Providers. Its core aim was to develop and document overall standards, quality assurance systems and methodological advancements which were providing an overarching framework for the production of Forest downstream service cases. This was considered crucial in order to harmonise the varying capacities and levels of technical development existing in the current consortium. In order to achieve this, the WP aimed to build on the rich legacy of capacities, technologies and processing chains for forest downstream services developed and established already within the GSE Forest Monitoring project financed by ESA. A further aim was the facilitation of links to the COPERNICUS Core services. The WP also focused on the incorporation of INSPIRE standards into the processing chains, as well as on the need for a harmonised approach towards metadata, quality assurance and validation for all Service Providers.
The majorities of these requirements have been addressed as foreseen within the project duration and have been documented in three Deliverables per Phase. The first Deliverable was the D320.1 Service Portfolio Specifications, which provided a comprehensive description of all the EUFODOS product and service specifications. The primary goal was to present the technical specifications of the services and products which are required for product generation and service provision including quality systems and standards used. The production processes have been specified in a structured manner in order to support and enable standardization, automation, enhancement and growth. Thus all descriptions of the production process or service chain have been based on the four main components which are input data sources (EO/in-situ/models/others), pre-processing, thematic processing and the analytical or modeling component which is finally required for output production. One major focus has been laid on the use of the Core Image data (e.g. 2006, 2009 and 2012) as well as the High Resolution Forest Core layers. The Products in the EUFODOS Portfolio have been finally grouped into three main categories which are Rush Mode for Forest Damage Assessment (e.g. Storm damages), Non-Rush Mode for Forest Damage Assessment and the assessment of Ecological Functions as well as different Forest Parameters. Quality assurance of EUFODOS products has been based on internal quality assurance (within the Service Provider as well the Service Network) and an external, independent user validation. Internal quality checking and reporting systems have been designed to conform to ISO standards. These procedures guaranteed traceable service production standards and generated standardized quantitative output which can be used for further assessment. User validation supervised the overall quality of service production and verified user needs of EUFODOS products. Specific mechanisms have been developed to support the potential for technology transfer in upgrading the Service Portfolio. One of the main inputs for improved service provision and evolution of services was the inclusion of R&D in the development of the Downstream Service Portfolio. Main objectives of the Service Evolution activities were to identify, test, transfer and integrate mature results of R&D from research partners and wider R&D domains into the Service Portfolio, and, thus improving operational efficiencies and evolution of the Service Portfolio to better meet user requirements.
The second Deliverable was the D320.2/D320.5 Service Network Validation Protocol, which presented the Validation Protocol for the EUFODOS Project. It provided the overall framework for service and product validation within EUFODOS and summarized the steps of quality assurance documentation during the production process (validation of Products and Services) based on criteria such as e.g. thematic accuracy, thematic completeness, etc. It also specified the validation approach for Input data such as Earth Observation data, reference data and GMES Core datasets, as well as the utility assessment of products by the end-user, which has been considered a main part of the validation process. The Deliverable D320.2 and D320.5 further provided recommendations how to handle/document potential non-compliance (if it should occur) and necessary product/service improvements.
The (D320.3/D320.6) Service Network INSPIRE Implementation Technical Note as the third deliverable within WP320 was based on a two-phase approach (Phase 1: OGC compliancy, Phase 2 extension to INSPIRE compliancy, as possible). It provided the results of the investigations on the background as well as technical recommendations how to implement INSPIRE standards and requirements in the EUFODOS Forest Downstream Services. Key INSPIRE Requirements on Metadata, Data Specifications and Network Services have been explained in detail and the possible approach how to technically include these relevant Implementing Rules in EUFODOS and the respective Forest Downstream Services has been described. The INSPIRE schedule/roadmap, information on the data specifications for Land Cover and Land Use have been revised and new specifications related to Discovery, Viewing and Download Services have been provided. Whilst the INSPIRE process is still evolving with full implementation in 2019/2020 and not all INSPIRE Implementing Rules have been finalized so far, it was recommended in EUFODOS to use a two-phase approach, where the usage of OGC compliant Server Software is planned for implementation. In case that INSPIRE compliant Software is available at a later stage, it is recommended that service providers upgrade their server software to these INSPIRE compliant software versions. The Deliverable concluded with the presentation of different software options (free and proprietary), which have been recommended to provide INSPIRE compliant web services for EUFODOS. The main objectives have been achieved to provide an overarching framework for the production incl. key aspects in terms of standards, quality assurance and methodological advancements. Also the incorporation of INSPIRE standards in the production work and related to quality assurance and metadata has been achieved.
1.5 Development of Operational Toolboxes
At the beginning of the project a document was elaborated encompassing information to the SPs, such as an overview of existing software packages, the identification of software elements within EUFODOS or a review of processing chains. Based on this overview the SPs SW needs were asked and collected from them. This preparation step was followed by the elaboration of a questionnaire called “Questionnaire on Processing Lines and Methodologies in the EUFODOS Test Cases”. The analysis of the questionnaire and distribution of its results to the beneficiaries as well as a review of literature and the outcome of different forestry related projects finalized this step. As a completion of this introductory phase a SW training was organized and took place at JR in Graz from 26th to28th September. Thus, based on these preparatory steps the SPs have got a thorough knowledge of all the SW capabilities in the consortium and had the opportunity to use the toolboxes according to their needs. Within Phase 2 the operational tasks have been initialized and several refinements on the toolboxes performed. JR and VTT were thus enabled to optimize the toolboxes and finalize them according to the experience made in both phases. As an outcome of EUFODOS the toolboxes are now in an operational mode and are commercially viable products.
1.5.1 Objectives and Results
This started in month 3 and its main actors were JR and VTT. JR and VTT offered already developed in-house algorithms and methodologies for the forest parameter assessment which was tested on-demand and adopted within the processing chains by the service partners in order to improve the cost-effectiveness of the production. WP 330 focused, therefore, on the investigation and dissemination of new tool boxes/algorithms to facilitate the production of the service cases. The operational toolboxes were at hand to all SPs and were used for the processing of the data in the testcases. The other partners were also participating in this WP with their engagement in the various development phases. Continuous updating has been performed on the toolboxes based on the experience made during the roll-out phase and the feedback to JR and VTT in order to optimise the performance of the workflow.
The following software algorithm /toolboxes have been developed during the EUFODOS project:
• EUFODOS Change Detection toolbox
• Envimon SW for Satellite Data Pre-processing
• Probability SW for Forest Variable estimation
• AutoChange SW for Change detection
• FoVEA model refinement by sub-clustering
• EUFODOS LiDAR toolbox
The “EUFODOS Change Detection toolbox” has been developed and tested in Phase 1 on existing data sets. Furthermore a pilot project was initiated and contracted by the Forestry Board of Salzburg/AT applying the toolbox for the assessment of changes within a period of 10 years on the entire forest area of the province. It could be proved that the toolbox is an effective instrument which can be used to obtain information over large coverage in a cost effective way. In the finalization phase of EUFODOS some more small refinements were applied in the toolbox.
“Proba_cluster: Unsupervised clustering”, “Proba_plot (future: proba_model): Cluster labeling software“, and “Proba_Estimates: Estimation software for target variables” are software packages which were implemented for EUFODOS. In the boreal testcase they were used for the processing of the economic parameters and can be considered as operational tools.
“FoVEA stand variable data screening” software was developed for the testcase implementation in the boreal zone. A certain forest stand in a stand-wise reference data can be assumed to be homogenous, i.e. the forest characteristics should not vary too much within the delineated stand boundaries. In practice this assumption does not hold for the real data: the stands may contain large natural variations, that the data collection process does not take into account but only at average level (details below certain size are neglected), or the reference data may be out-of-date due to forest management actions or natural causes (data inconsistent). These variations in the forest structure can naturally be seen as large variations in the satellite image reflectance data. This adds unwanted noise in the modelling process. To avoid introducing this type of noise into the models, the reference data stands are screened for the EO data reflectance variations. As a result a new reduced stand variable data set is produced with the most homogenous (in the sense of image reflectance) stand. The method was found useful in producing more consistent models with increased maximum stem volumes, and it has been included in the standard repertoire of the FoVEA software.
“FoVEA model refinement by sub-clustering” was developed for the modelling cases in which there is relatively large amount of reference data available (more typical with stand variable data). The Probability method results in a model where many ground reference data vectors are associated with most of the clusters that represent forest (in the spectral data). The produced model is robust and typically generalizes well, but the drawback is that the maximum values that can be estimated with the model will be biased downwards due to this averaging. A method to compensate this effect was developed in EUFODOS. This method cannot be applied for models for which there are only few reference data available, thus the default value of 50 reference data vectors per cluster has been set as minimum to enable the sub-clustering (may be selected by the user). The sub-clustering method has been implemented as model refinement (or tuning) tool in the FoVEA software.
The “EUFODOS LiDAR toolbox” is a special case in the project and was initialized by the Austrian user LFD-STMK who showed a strong need for the assessment of forest parameters in the protective forest environment. As the state authorities had already initiated a campaign for a LiDAR survey it was a strong interest to use the high detailed LiDAR for the ascertainment of the needed parameters. JR had already some software elements for the processing of LiDAR data in house and as a consequence the development of a powerful toolbox was the major goal in WP330. As the “EUFODOS LiDAR toolbox” is a very complex software, it can be stated, that both phases were highly needed to create an operational software package. A special custom-tailored workflow was created in order to make the whole processing user-friendly. Especially in the final roll-out phase some challenges had to be solved in terms of data volume and computing speed. Finally the toolbox is now a user-friendly software environment which can be offered on a commercial base.
1.6 Assessing New Sensor Data
1.6.1 Objectives and Results
The work of this WP focuses on the testing and assessment of the utility of new sensor data and to evaluate their capabilities to meet the various thematic and technical product specifications. All studies have the objective in common to contribute improvements and enhancements to the EUFODOS services.
The following studies have been executed
1. Test of TerraSAR-X data processing options for mapping of small sized clear cuts to conclude on storm damage mapping – case study located in Poland
2. TerraSAR-X pre & post storm intensity based change detection for storm damage mapping – study area located in Thuringia, Germany
3. TerraSAR-X & Tandem-X single pass interferometry surface model quality assessment as background information for storm damage mapping – study area located in Karlsruhe, Germany
4. Pre & post storm based TerraSAR-X & Tandem-X single pass interferometry change detection for storm damage mapping – study area located in Thuringia, Germany
5. RapidEye for storm damage mapping – study area located in Poland
6. RapidEye for snow damage mapping & assessment of forest operations – study area located in Thuringia, Germany
7. RapidEye for bark beetle mapping – study area located in Baden-Württemberg, Germany
8. Suitability of spectral variables in mapping insect defoliation using RapidEye data – study area located in Brandenburg, Germany
9. Mapping of Forest Changes – study area located in Brandenburg Saxony
10. Digital surface model generation by digital aerial photography – study area located in Austria
All studies contributed to the enhancement and improvements of the services. The following findings of the studies have already integrated in the portfolio and processing lines of EUFODOS:
Table 8: Studies integrated in the portfolio and processing lines
Study Integrated results
Test of TerraSAR-X data processing options for mapping of small sized clear cuts to conclude on storm damage mapping – case study located in Poland Processing options for SAR Data.
TerraSAR-X pre & post storm intensity based change detection for storm damage mapping – study area located in Thuringia, Germany Integration of the SAR Option into the portfolio in Phase 2.
TerraSAR-X & Tandem-X single pass interferometry surface model quality assessment as background information for storm damage mapping – study area located in Karlsruhe, Germany The developments made in these studies show very high potential of the TerraSAR-X & Tandem-X single pass interferometry option to monitor storm damages. Still this option cannot be integrated in the portfolio since this option is from EO-Data availability side restricted. Still the findings can contribute to the benefit identification of future operational single pass interferometry satellite configurations.
Pre & post storm based TerraSAR-X & Tandem-X single pass interferometry change detection for storm damage mapping – study area located in Thuringia, Germany
RapidEye for storm damage mapping – study area located in Poland The results of the study show that a regional storm damage event that is dominated by small damage areas can be mapped with Rapid Eye with high accuracy using an approach where damages are mapped with a minimum mapping unit of 0.2 ha. Still, to enable a realistic estimate of the amount of damaged wood in case of a storm of with the damage structure of the examined storm damage event in North-West-Poland in 2011 it is not sufficient to consider only damages on completely damaged areas with a minimum area of 0.2 ha but scattered damages have to be taken into consideration as well.
It is planned that further options for the estimation scattered damages will be developed and that the estimation of scattered damages will be evaluated based on damage data per compartment.
RapidEye for snow damage mapping & assessment of forest operations – study area located in Thuringia, Germany Integration of the assessment of forest operations into the portfolio in Phase 2.
RapidEye for bark beetle mapping – study area located in Baden-Württemberg, Germany The study still showed the general feasibility of monitoring of bark beetle damage expansions in natural forest that are maintained without harvesting inventions.
It became obvious that small damages can be more adequately mapped with aerial photography taken by planes or UAV. Still for events with substantial damage dynamic Earth observation could substantially contribute.
Suitability of spectral variables in mapping insect defoliation using RapidEye data – study area located in Brandenburg, Germany Processing options for RapidEye data processing: The results of this particular study suggested that the NDRE, being a vegetation index containing the RedEdge band, is the variable most suitable for mapping symptoms visible in the pine forest canopy. That’s why it was taken as a basis for the development of the BlackBridge Forest Vitality & Change Monitoring Layer Bundle – a service fully mature for operational application. As already shown in EUFODOS D430.4 the service consist of 4 Layers (products) which can be delivered.
Mapping of Forest Changes – study area located in Brandenburg Saxony Processing options for RapidEye data processing.
Digital surface model generation by digital aerial photography – study area located in Austria The study showed that using digital surface model generation by digital aerial photography has a great potential for operational use.
1.7 User Requirements
1.7.1 Objectives and Results
The main objective of this WP is explicitly to assess the actual user needs and standards of all the users which requires the documentation of the environmental/forest policies and/or reporting requirements for the user that the consortium can support, the working practices of the user, technical capacities and standards, the potential benefits of services to the user organization, the availability of in-situ data, and the finalisation of the Service Level Agreements with users.
This WP is divided in two issues (phases). Generally the necessary information for deliveries production was collected by the use of specialised Questionnaires. The questionnaire in issue1 was very comprehensive. The questionnaire in issue2 was more concise, updating the user requirements and collecting new, important information.
The time between the two issues was dynamic and to obtain the changes which occurred during that time, a new, very concise, template/questionnaire was prepared. All the users have answered to the same questions.
An additional job was done - all the services required by the Users and which were in the process of development by all SPs up to the time of User Needs and Standards Dossier for all Services Issue 2 provision, were listed in a special table (Annex 3, report for issue2). Each service was assigned to a unique identification code. It would help the reference to all downstream services, and would help better familiarisation with these services.
The deliverable of each issue is a comprehensive Dossier giving a comprehensive overview on the user needs for each service and comprises the relevant issues in standard dossiers with respect to user relevant services.
1.8 Adaptation, Integration and Testing of Processing Chains
1.8.1 Objectives
This WP deals with the intensive phase of implementation encompassing adapting, integrating and testing of production chains in the individual testcases. The envisaged service in the EUFODOS consortium is built on a strong expertise and wide experience of the beneficiaries for the realisation of operational forest applications. In order to build up a FDS it is a main task to use the single expertise of each beneficiary in a synergistic way. That means, existing knowledge, software and service experience are exchanged and modified to enhance the service delivery chains. In this WP the realization of the user requirements and needs were realized for the different applications and services. Thus main activities included the coordination, monitoring and generation of delivery of services, the utilization of common infrastructure, and implementation of the SN quality assurance plan. The 6 testcases are described in detail below:
1.8.2 Testcase Styria – Service Provider: JR
The focus in the testcase “Upper Styria” is put on two applications, namely “Forest degradation from storm damage assessment” and “Impact of Forest Degradation on forest Functional Indicators”.
The first assigned forest product is focused on the derivation of windfall areas, meaning that map information is produced which displays damaged forest areas. The detailed service specifications for forest degradation were based on the needs and requirements defined by EAA in close cooperation with the Austrian Federal Ministry for Agriculture, Forestry, Environment and Water Management/Department on Forestry. The service specifications for “Impact of Forest Degradation on forest Functional Indicators” are based on the needs and requirements defined by LFD-STMK. In the selected test site the derivation of functional parameters is focussing on the protective functions, which are closely related to degradation issues. A very powerful approach for the derivation of these parameters was the synergetic use of laserscanner (LiDAR) data providing 3D structural parameters (height, crown closure and vertical structure) with Core Service products, procured from GEOLAND II (LMCS: land monitoring core service) or GIO (GMES initial operations), providing information on tree species distribution. In the following the main issues of the roll-out phase are described in brief:
1.8.2.1 Forest degradation from storm damage assessment
The processing line for this service has been developed in WP330 “Development of Operational Toolboxes” and has been tested in WP420. The SP JR has successfully tested the new processing line for an operational service, which shall serve as a demonstration example in order to acquire further orders. In a pilot project the developed software Toolbox could be applied for an application called “Waldveränderung Salzburg”, which has its focus on the detection and delineation of forest degradation changes within a period of ten years. In this project the forest authorities of Salzburg have shown a strong interest in changes within this period in order to integrate them as important information layer into their forest management plan. In general, it can be stated that the performance of the processing line worked well and it is easy in the handling process.
1.8.2.2 Impact of Forest degradation on forest functional indicators
JR has chosen a series of LiDAR tools for the implementation of the EUFODOS LiDAR toolbox. The single modules have been tested and adapted to the requirements of the EUFODOS application. An on-going activity during this phase was the interaction with the user LFD-STMK which has also urged the necessity to adapt the toolbox continuously. Another important aspect during the processing phase was the optimisation of the processing chain especially in terms of computing speed, because the LiDAR data have huge data volume.
The goal of the LiDAR processing chain is the generation of forest functional indicators that will help to monitor degradation effects in Alpine forests in a very detailed manner. Following functional parameters, defined by the user LFD-STMK, were derived on stand level:
- Tree height
- Vertical stand structure
- Crown/canopy volume
- Crown cover density
- Development stages
- Understorage
- Forest mask (including the upper forest border)
For the implementation of this processing chain existing methods were adapted or if needed new ones developed. All methods were developed and tested at the EUFODOS small testsite Hohentauern in upper Styria. Meetings with the local users - LFD-STMK - were organized, to discuss their requirements. It was found that the definition of the alpine forest boundary as such was a very important issue. The derivation of the –critical – upper forest border line from LiDAR data proved to be highly accurate. The calculation of forest stand polygons by segmentation and the tree top detection are both fully automatic approaches leading to highly appreciated results for the user.
As already mentioned the direct use of the Core Service product P-EL-4B “Forest Types” is not possible, thus a method has been developed to obtain a product with an acceptable spatial resolution. In this context the already existing CS product was used to classify the 5m RapidEye imagery and thus saved time and costs which would have been needed for the ascertainment of appropriate training data sets.
1.8.2.3 Conclusions
The change detection processing line has been implemented into an easy following workflow, which enables the operator to apply it in fast and efficient way. The entire LiDAR processing chain has been implemented as a workflow which offers a guided tour throughout the processing line. The user LFD-STMK has validated the final LiDAR parameters especially with the focus for their processing within the forest GIS. It can be stated that from the current findings the use of LiDAR data is operationally applicable and leads to incremented value for the management of protective forest. Both toolboxes, the “EUFODOS Change Detection toolbox” and “EUFODOS LiDAR toolbox”, could be developed to an operational status and are now ready for commercial selling too.
1.8.3 Testcase Finland/Russia Service Provider: VTT
The VTT Technical Research Centre of Finland was developing EUFODOS Downstream Services for two test areas: one in South-Eastern Finland, the other one in Russia (Nizhniy-Novgorod). The VTT Test Case included the production of 1) Stem volume maps for selected test case areas in Finland and in Russia. The stem volumes were produced as total stem volume, and as species-wise products (pine, spruce, broadleaved). During the project a new product called 2) The Assessment of Storm Damage Economic Impact was included into the VTT service portfolio.
Development of the Production Chain
To set up the Forest Downstream Service production system for VTT Test case (P1 – P5, see SLA for the definition) an analysis of the production chain (called the Probability chain) was performed with the identification of major software development needs (no conceptual gaps were found).
The FDS production chain was subjected to operational tests (test per SW module and end-to-end test) with selected test case data in order to assure a consistent production flow. Three mutual working meetings with the user organization (Stora Enso Oyj) were organized for the elaboration and refinement of the user requirements.
The most important development needs from the user point of view were:
• To improve forest variable estimate accuracy, especially for the higher stem volumes
• To improve the separation of conifers (pine and spruce) in mixed conifer forest
• To produce indication of map reliability
• To obtain accurate information of storm damages
From the system level point of view the need to develop the existing in-house software tools into operational level was considered as major development driver.
The software development work included modification of the pre-processing software Envimon for the Image 2009 IRS data and RapidEye data. Other modifications were implemented to facilitate the Probability software usage in operative production, and a modification to facilitate the usage of CORE service data in map production. Implementation of batch-processing for EO data pre-processing phase was partially included in EUFODOS activities.
New software modules were written into FoVEA software - a GUI for running Probability chain programs. These included modules to screen reference data outliers (to improve model consistency) and to evaluate the product accuracy. Also tools for model cross-validation (K-fold cross-validation) and for model fine-tuning (sub-clustering method) were implemented.
Data Acquisition
The data acquisition plan that was originally written in SLA for the VTT test case areas in South-Eastern Finland and in Russia (Nizhniy Novgorod) was updated. The DWH was screened for suitable HR data sets for products P1 and P2. The primary set of optical EO data included four partly overlapping IRS-P6-LISS3 scenes from the South-East Finland area from summer 2010, and two additional image frames from 2008. The cloud cover of the scenes varied from 0% to 25%. Two scenes from year 2010 were selected for the production chain assembly and testing, and for the demonstration product generation. These scenes cover roughly ¾ of the AOI (Area of Interest) in Finland that was defined in the SLA, and included large area north to the AOI (). Three RapidEye image from June 2011 was selected for the first tests (P5) from the Finnish test site. From the second test area (Nizhniy Novgorod) two SPOT5 images from 2010 were acquired from the DWH. One of these was too cloudy to use for the production, but the other one, covering some 80% of the service case area, was used for demonstration.
To facilitate the production of VHR enhanced products (P3 and P4) the VHR image archives were screened for the years 2010 and 2011. For the Finnish test site the three RapidEye scenes were selected as the VHR data for the product P3 development, as they had a relatively good common coverage with the reference data set. For the Russian test area there was no VHR data available, and the product P4 was cancelled from the VTT’s portfolio in agreement with Stora Enso.
EO Data Pre-processing
The EO data selected for the demonstration was subjected to standard QA procedures and pre-processing steps (unpack, radiometric calibration, atmospheric correction). Relative calibration of the EO data with respect to the model image(s) was also performed where necessary. The production of cloud masks and water masks followed the automatic pre-processing steps. Topographic correction was not performed.
Preparation of Reference Data
The forest area in the Finnish test area is pine dominated forest with average stem volume of 116 m3/ha (stdev = 87 m3/ha; as calculated from the stand-wise reference data set). The reference data of year 2010 was extracted from the Stora Enso’s database. It has been interpreted from aerial imagery with ground verification and bi-annual updates. The data contained 16 variables from which the total stem volume and the stem volumes by tree species (pine, spruce and broadleaved trees) were selected for the demonstration products. The resulting stand variable data set consisted of 250 000 forest stands. The reference data set was screened to contain 52% of he most homogenous forest stands (as evaluated with the IRS imagery). A test reference data set of 504 field locations was extracted from this set using stratified random sampling, with total stem volume (ToV) as stratification variable. This test data set was excluded from the training data set, and was used throughout the production for model cross-validation and final product testing. It has to be noted, that the stand variable data contained significant errors (up to 30%; mostly due to forest management actions), which also had an impact for the FDS production chain software development.
The reference data for the Russian test site consisted of 156 field plots, having stem volumes for pine, spruce, birch, aspen and other broadleaved species. The reference data were converted to the reference coordinate system of the EO image (WGS84 NUTM38). Only 56 of the original set of 156 field plots were on the image area, and 20 additional plots had to be removed due to inconsistency, clouds or burnt areas. Thus only 26 reference data plots were available for modelling, which affected to the model quality.
Utilization of CORE products
To test the possible benefits of using CORE data in the downstream service production, a ‘simulated’ CORE data set was generated for the Finnish test site. The model for the CORE data production was built for Estonia in the GIO Land project, using the principles and experience gained in Geoland2 project. Even if the structure or tree species composition of the Estonian forest do not match exactly the ones in South-Eastern Finland, the model was considered to be accurate enough to be used in the testing. The training image was matched with the Estonian model training image reflectances with channel-wise relative calibration. The produced CORE data set included the layers crown cover percentage (CC) and proportion of broadleaved trees (BL%) as continuous variables (0 – 100%).
Conclusions
The implemented changes of the software tools for forest structural forest variables mapping (Probability, FoVEA and Envimon SW) improve the efficiency of the thematic map production considerably and will shorten the product delivery times. The achievements in the EUFODOS software development may be summarized as:
• Usage of CORE service data shortens the production time
• Batch pre-processing for large data sets shortens the production time (considerably)
• The developed new methods (model sub-clustering & reference data outlier removal) improve the product dynamic range by 25% - 35% without introducing bias or decreasing product accuracy
• Improvement in model consistency and dynamic range by the new reference data screening methods
• New tools for product accuracy assessment give user an indication of map reliability and support in decision making
The developed software packages are being commercialized at VTT and are available for forestry actors worldwide.
1.8.4 Testcase Thuringia Service Provider: GAF / ALU-FR
The Service Demonstration in Thuringia was based on a Service Level Agreement (including an update version from End of December 2012) between ThüringenForst in Gera (as user) and GAF AG in Munich (as Service Provider). The Service Level Agreement was updated in December 2012. GAF AG with support of ALU-FR (University of Freiburg, Department for Remote Sensing and Landscape Information System) delivered the following Downstream Services to ThüringenForst based on the SLA and within the Project lifetime of EUFODOS:
• Rush Service - Forest Damage Assessment
• Non-Rush Service – Forest Damage Assessment
• Assessment of Changes due to Forest Operations and Forest Degradation
Main contact point on User side was the Service and Competence Centre (SuK) of ThüringenForst located in Gera and represented in EUFODOS by S. Chmara and H. Sagischewski.
From beginning on, the HR Forest Core Layer were a pre-requisite for the Development of the above mentioned Downstream-Services. From the FP7 project geoland2, no Forest Core Layers were available for the test sites in Thuringia. Therefore, the HR Forest Core Layers had been produced for Thuringia within the first 4 Months of EUFODOS. The production was based on Image 2009 and from GIO-Land (lead by EEA) project; the actual product specifications were used. The Tree Cover Density and Forest Type (HR Core Layer) as well as a Tree Cover Map (derived from the Tree Cover Density Layer) and a Spruce Type Map built the basis for all further Downstream Service Development activities and product delivery.
Service Demonstration: Rush Service – Forest Damage Assessment
On 5th July 2012 a tornado caused damages in a forest near the City of Gera in Thuringia (Germany). Around 20 ha of an urban forest area covered with approx. 80% conifers and 20% deciduous trees was affected by this event. The event has been used to demonstrate the usability of the “Non-Rush- Damage Assessment Service”. The service included the use of EO (e.g. Image 2009) as the basis for reference mapping. In addition, SAR data (TerraSAR-X) from before and after the Tornado event was acquired to test a multi-temporal approach for the classification of the damaged areas. The classified damaged areas as well as all relevant additional data were delivered to the user ThüringenForst in September 2013. The acquired SAR data were analyzed and classified by ALU-FR. The developed processing chain was later transferred to and tested by GAF AG. A field mission already conducted in spring 2013provided the relevant figures for the thematic accuracy assessment. With an overall thematic accuracy of 87.5%, most of the damaged areas were identified successfully. Beside this, the shape of the areas found on ground differed strongly from the shape of the areas classified using SAR data. Nevertheless with the used method, a first and fast determination of damaged locations is possible with a high object related accuracy; that is of high interest for the user. The Rush Service is still not fully operational as the test case in EUFODOS was implemented under more or less optimal conditions (summer storm, flat area). As most of the heavy storm events in the past were taking place during the winter months (November - February/March), the user ThüringenForst and the Service Provider GAF AG decided after the last training session in November 2013, to implement a further test. It is planned to finally proof the service capability under extreme (snow, hilly terrain, bad weather) conditions in Thuringia during winter 2013/2014. Nevertheless, other Forest Administrations have shown high interest in this service and are thinking to implement this service in their plans for disaster management.
Service Demonstration: Non-Rush Service – Forest Damage Assessment
The Tornado on 5th July 2012 was also used to demonstrate the usability of the “Non-Rush Service – Damage Assessment”. The service includes the use of EO (e.g. Image 2009) as the basis for reference mapping as well as the derived HR Forest Core Layers (e.g. Tree Cover Density). EO data (RapidEye) after the Tornado has been acquired in order to facilitate damage assessment caused by the Tornado event. The classified damaged areas as well as all relevant additional data were delivered to the user ThüringenForst in December 2012. A field mission, supported by ALU-FR was implemented in spring 2013 and a very high thematic accuracy was achieved. All relevant service specifications were met. It was possible to identify damaged areas down to 0.1 ha.
Another case was used to further test the capability of the Non-Rush Service. In winter 2010/2011 heavy damages in forest caused by snow were reported in the forest district Stadtroda, Thuringia. Purpose of this study was, to compare the use of the HR Forest Core Layer and the use of RapidEye as reference data before the damages occurred. Therefore damaged forest areas (snow break) were classified based on RapidEye to RapidEye data without direct use of HR Image 2009 data, but in combination with the available Forest GIS data (mainly forest stand borders) from Thuringia and provided by ThüringenForst. The mapping activity achieved similar thematic accuracies compared to the Tornado test case described above.
In both cases, the overall workflow to derive the damaged areas included a segmentation based change classification algorithm. Additionally, the Tree Cover Density product (already produced in phase 1 of EUFODOS) was used to further minimize potential classification errors. The processing chain developed in EUFODOS has proven its usability. The development is finalized and the Service can be offered to a wider user group also outside of EUFODOS. Based on promotion activities, several forest administrations are already interested in this service.
Service Demonstration: Assessment of Changes due to Forest Operations and Forest Degradation
Federal forest administrations urgently need up-to-date information about the entire forest area and forest area change to carry out different tasks such as the implementation of political objectives pertaining to forestry, the protective functions of the forest, habitat mapping and recreational function of the forest as well as support of consulting services offered by the forest administration to private forest owners. Today, state forest administrations derive most information based on State Forest Inventory data with an update frequency of usually 10 years. Information about private and communal forest is most often excluded from State Forest Inventory and therefore not available on a regular basis and/or outdated. The developed Forest Downstream Service was aiming to close these information gaps by providing up-to-date information about the whole forest area and its changes on an annual basis. Within the test case, two areas were selected in Thuringia covering ¼ or ~ 4.000 km² of the Federal State Thuringia. As basis, the produced HR Core data and Core Image data, as well as RapidEye data from 2011 and 2013 were used. Pre-processing of the EO data (Image 2009, RapidEye from 210/2011 and 2013) comprised a full radiometric correction including also topographic normalization and cloud masking. The interactive segmentation based change detection to derive a map of potential changes between the EO images from different point in time was based on the software eCognition. Also manual editing and enhancement was done with eCognition. The use of the User Forest GIS as well as the HR Forest Core Layers was supporting the manual editing. All processing steps were followed by Quality Control documentation. Overall the production workflow was prepared in a way that a flexible scalability was possible. Depending on the available EO data and given reference data, small as well as large working units can be prepared and processed using the given workflow.
The service was successfully tested and implemented in the above mentioned test regions. Change Areas down to 0.1 ha were identified with an overall thematic accuracy of >90%. Within an operational production environment the service can deliver accuracy values between 90% -95% in dependence of the extent of the change areas. Additional manual enhancement and good knowledge about image interpretation can further improve the accuracy of the delineation of the damaged areas. The use of 5 to 10 m EO data such as RapidEye or upcoming Sentinel 2 data as base data for the change detection will further improve the thematic and geometric accuracy of the results. As the service was implemented successfully during EUFODOS, service development is finalized and the Service will be offered in the near future to interested user groups in Europe.
1.8.5 Testcase Swieradow Service Provider: REN
The service demonstration followed a service level agreement between the user Forest Inspectorate Świeradów, ul. 11-go Listopada 1, 59-850 Świeradów Zdrój, Poland, in partnership with FOREST RESEARCH INSTITUTE, 3 Braci Leśnej Street, Sękocin Stary, 05-090 Raszyn, Poland and RapidEye AG, RapidEye AG, Molkenmarkt 30, 14776 Brandenburg an der Havel, later changed into BlackBridge AG, Kurfürstendamm 22, 10719 Berlin, Germany as the Service Provider. It foresaw the delivery of the following Downstream Service Products:
• Product P-01 Time Series of Custom RapidEye Ortho-Image Maps,
• Product P-02 Forest Cover Map,
• Product P-03 Storm Damage Map,
• Product P-04 Forest Vitality Map,
for an area of about 70,000 ha in the Western Sudety mountains, near the country triangle formed by Germany, Poland and the Czech Republic (forest districts Świeradów and a part of Sklarska Poreba).
The main point of contact on users side was Mr. Radomir Balaczy, with whom a number of co-ordination meetings were held before and during the project execution period.
First the GMES core products were ordered for the test area and prepared by GAF AG. Based on this information layers the specific customer expectations were evaluated and laid out in an updated SLA with the user. Later RapidEye imagery was acquired over the area and the development of an enhanced Forest Cover Map product started. In a first step an advanced pre-processing procedure was developed and applied. It included the customized ortho-rectification based on user-provided data, special radiometric correction and mosaicking of the satellite images, necessary to cover the full area. During this step tools and technologies, provided by EUFODOS project partners were used and evaluated. The production process was executed for all following RapidEye image coverages and resulted in the Product P-01.
Using the first coverage of P-01 products a technology for production of alternative Forest Cover products was drafted, implemented and applied for test purposes. After a visual and statistical accuracy assessment the product was delivered to the user and field checked during a joint field campaign with the user. The sample forest cover map was well appreciated by the customer and accepted as EUFODOS product P-02.
An older storm case in Saxony was used for the development of a storm damage mapping technology. We made use of the previously developed technologies for products P-01 and P-02. In July 2012 a tornado event occurred in the Trzebciny area of Poland. With the help of our user we applied the developed processing chain to this event, produced and evaluated the storm damage product. Comparative accuracy analysis, using existent Forest GIS databases and data, collected during the removal of damaged timber from the area, was carried out by our user. Later he used the results for dissemination activities in Poland. The results were presented by Radomir Balaczy at National Foresters Conferences and Meetings and earned appreciation.
Meanwhile a forth product was developed for mapping of forest vitality in the test area. It first based on the multi-temporal coverage with RapidEye data. The maps showed just the seasonal vitality change. But there is a much higher value of such maps, if produced for current forest stress situation.
The RapidEye / BlackBridge project team identified a potential need for such products in the eastern German Federal States, which suffered from insect infestations in oak and pine stands. First the Brandenburg State Competence Center for Forestry became an additional user organization for EUFODOS (but without special SLA in the sense of the project), for which we produced a vitality map of infested stands. Together with specialists form the State Competence Center and local foresters the maps were checked in field surveys. They assessed the prototype map products as helpful and gave useful hints for further improvements. This led to a series of maps for infested regions, now called the Vitality Map Bundle.
Following the intention of the EUFODOS project additional potentially interested users of the developed DS were invited for service demonstration: the state forest authorities of the German Federal States of Saxony, Lower Saxony, Mecklenburg-Eastern Pommerania and Brandenburg. An overall number of 9 test cases were processed and successfully evaluated together with specialists from user side.
1.8.6 Testcase Bolzano Service Provider: EURAC
The testcase Bolzano is based on the Service Level Agreement (SLA) concluded between the service provider (EURAC, Institute for Applied Remote Sensing, Bolzano, Italy) and the user (Department of Forest Planning, Province of Bolzano, Italy). According to the SLA, the demonstration included the presentation and delivery of the following Downstream Service Products:
• Product P-1 Base data on forest,
• Product P-2 After Event maps,
• Product P-3 After event monitoring product (recovery),
• Product P-4 Forest Damage Information System (FODIS).
Thereby, P-1 should cover the entire area of the Province of Bolzano with a total area of 7,400 km², while P-2 and P-3 were foreseen for a total coverage of about 300 km². P-4 is a spatially explicit information system on forest damages integrating the products P-1 to P-3.
The main point of contact on the users side was Mr. Günther Unterthiner, with whom regular co-ordination meetings were held before and during the project duration.
In the initial phase of the project, the Department of Forest Planning provided forest damage data to the EURAC which were used to start the work on P-2. These data mainly consisted of broad places and described impact of insect infestations. A timeseries interpretation technique was developed using broad-scale MODIS data to identify these areas, however, this attempt was not continued as the results were not sufficient enough. Moreover, in June 2011, a storm event occurred in the north-eastern part of the Province of Bolzano. This event was then used to develop and implement a forest damage mapping approach based on high-resolution data and a forest change map was derived (P-2). The accuracy of the map was visually and statistically assessed and compared to the field mapping results of the foresters. The results were delivered to the respective foresters who field checked the results. Subsequently, the product was presented in detail to the foresters who very positively evaluated the product and confirmed that the accuracy of the change product outperformed the field mapping attempts in almost all cases. Due to the very positive feedback and the planned operational integration of this product on an annual base in the FODIS, EURAC modified the processing chain and derived annual change maps for the period 2009-2013.
In parallel, product P-3 was developed to allow continuous monitoring of forest intervention measures. The user provided different afforestation areas and EURAC evaluated different options on how to provide relevant information. Timeseries of Landsat for 1984-2012 and RapidEye data for 2009-2013 were utilized and an approach was develop to show spatial-temporal changes of the vegetation cover.
Product P-1 was developed to map forest type and forest extent based on multi-temporal RapidEye data for 2012. A procedure was developed based on the existing Geoland tree type information to derive reference data and a reliable tree type layer was generated. This layer was transferred to a forest type layer by consulting the user and agree on a forest definition applicable to the data source.
All three products were delivered to the user and positively evaluated. Moreover, all layers were integrated in the FODIS which was developed in the project.
1.8.7 Testcase Sofia Service Provider: RESAC
The service demonstration of the products delivered by Remote Sensing Application Center – ReSAC Bulgaria followed a service level agreement between the user Executive Forest Agency (EFA), 55 Hristo Botev Blvd., 1000, Sofia, Bulgaria and ReSAC, 61 Tsar Asen Str., 1463, Sofia, Bulgaria (Service Provider).
The main point of contact on users side were Mr. Nikolay Pironkov, Mr. Vladi Konstantinov and Mr. Spas Tumbev, with whom a number of co-ordination meetings were held before and during the project execution period.
During the lifetime of the project and after the first phase when the demo products were delivered, the User drafted its user requirements, hence the final services which were agreed changed to: Forest Inventory Service divided to: P-01-1 Forest Area Map; P-01-2 Forest Type Map; P-02 Forest Change Map; P-03 Forest Density Map. The services for damage assessment changed as follows: Forest Disasters Service; P-05-1 Reference Map (Forest Cover Map and Forest Type Map); P-05-2 Damage Extent Map (Fire Extent Map, Storm Extent Map and Insect Infestation Extent Map); P-05-3 Detailed Damage Map (Crown/Surface Fire Map, Storm/Snowbreak Damage Map and Insect Infestation Damage Map); P-05-4 Additional products (Combination of the EO products with cadastral information).
On the other hand the test area which was preliminary chosen was extended to other territories in Bulgaria where natural disasters in the forest territories has occurred: Area 1 - Gostun Village, West Rhodopy Mountain; Area 2 – Tran, Breznik, Zemen Forestries, West Bulgaria; Area 3 – Sredec Forestry, Strandzha Mountain; Area 4 – Nature Park Vitosha, Vitosha Mountain; Area 5 – Koprivstitsa Town; Sredna Gora Mountain; Area 6 – Silistra Municipality, North East Bulgaria; and Area 6 – Danube Plain.
The first products delivered were the three types of the Forest Disaster Services. The products were developed for a small fire event in the Area 2. During the field check the EUFODOS project and services were presented to the local forestry. Together with the P05-1 and P05-3 products for the test case were generated also products P05-2, which were more detailed products based on User demands to combine the Forest cadastre with the disaster products.
The second products delivered were the ones from the Forest Inventory Services. These products were delivered for Area 2. The main image source used was Rapid Eye. The data were atmospherically corrected, geometrically corrected and using the methodology developed by ReSAC combining different vegetation indices the products P0-1, and P0-3 were delivered. Both services were delivered to the EFA and training on the products was performed.
The products for snowfall/snowbreak were delivered for the Area 3. The main image source used was Rapdi Eye data. The first image requested was in early spring and no good results were obtained because the leafs of the deciduous trees were not developed. A second image was requested for early summer which was combined with the reference image from the previous year and P0-2 Forest Change Map was produced. All the P05 products developed were quality checked and the accuracy was within the 90% specifications. During the field trip the EUFODOS services were presented on the administration of the local Forestry Sredec, and they were the user to which the products were delivered. ReSAC has delivered also Quantum GIS open source software, and conduct one day training on the software and data delivered.
For insect infestation ReSAC requested VHR data for the test area – Ikonos, WV2. In addition orthophoto from the airborne sensor was used. The processing was focused this time on working with VHR data and using the 8 channels of the WV2 data. The same for the storm event the methodology developed for the snowbreak event (area 3) was used. The spectral data of the WV2 data were used for mapping the insect infestation in the national reserve Bistrishko Branishte in the Vitosha Mountain. All the products were quality checked by a field trip campaign.
In the spring of 2013 a storm event occured in Area 5. The local forestry requested through EFA F05 products in order to map the main damaged areas. ReSAC informed the EUFODOS partners for the event and Terra SAR data were delivered. The processing of the data was not successful due to the big time interval between the two scenes.
During the same time (spring 2013) P05-4 were tested (area 6). The main source this time was LIDAR. The data were combined with orthophoto and Rapid Eye data and differentiation between tree species and forest height was done. The products were presented to the user as a option for further development as it was not in the initial SLA.
After the meeting with EEA in Copenhagen ReSAC has obtained the FMCS developed in the frame of GIO, together with the CORE3 data from Rapid Eye. For the test Area 7, for which ReSAC had additional auxiliary information such as LPIS, and FMP all the Forest Inventory Services were developed using the updated methodology. The products were delivered in the final seminar (17th of Dec 2013) to the EFA and presented to the 6 Forest Enterprises, and regional directions of the EFA.
The products developed by ReSAC were presented in 9 international events (conferences and workshops) and 8 events in Bulgaria. ReSAC delivered to the user 4 reports, more than 100 maps, around 100 thematic products (including reference GIS layers) and close to 30 orthorectified EO images.
1.9 Demonstration
1.9.1 Objectives and Results
During the EUFODOS project prototype demonstration products were produced in accordance with the user requirements and the prototype processing chains developed before. Thus main activities included the coordination, monitoring and generation of delivery of services, the utilisation of common infrastructure, and implementation of the SN quality assurance plan.
The activities of this WP had been undertaken by all SPs.
It was the task to execute all procedures to generate the information products, and to perform all required routine calibration and verification procedures (quality assurance). Some of the developments were reviewed, modified, adjusted or changed as a result of new findings, new ideas or changes in the user requirements. In some cases, new test areas were adopted and new users integrated, for instance for verification of results.
1.10 WP440 User Validation of Core and DS Services
1.10.1 Objectives and Results
To assess the service utility for the user, EURAC has designed an interactive questionnaire which was discussed with the EUFODOS user board in advance. All 6 users answered the questionnaire after they received the demonstrator products (D440.1) and after receiving the final products (D440.2) for the EUFODOS Downstream Services by their respective service provider. The results in short:
• State of work
o 95% of all services are completed or almost completed, 86% are completed.
• Functionality and utility of the DS-services
o All users rated the demonstration products very positively
o The DS-services are in most cases not replacing existing services but offering complementary or new information
o The level of integration was mostly rated as high and the ease of integration as easy
o Most agreed benefits were time saving and easy comprehensibility of the products. In most cases, the DS-services added improved monitoring capabilities.
o Potential constraints include the dependency of the users towards the service provider and the potentially higher costs if additional data is needed
• Sustainability of the service.
o Users are very interested in sustaining the services and would be willing to pay for selected services.
o Important for all services based on High Resolution data would be the availability of SENTINEL 2 data, which would reduce costs and allow for a higher update cycle and a more rapid availability of the service.
o For more than 40% of the products updates are planned, for 30% considered, but not yet clear.
• Assessment of Procedures and recommended Improvements
o The negotiation phase and the production phase for the products were in general rated as very good.
• Assessment of Impacts and Value Statement
o 80% of the users stated, that the EUFODOS DS-products have a significant impact on their institution. Many users have already recommended the product to other potential users.
• Overall Evaluation and Outlook
o All users evaluated the products as good or very good.
• Suggestions for further improvments included
o Information on the state of forest with protection function
o Information on typical forest parameters (volume, …)
o Update services based on Sentinel 2 data
o Upper forest border
o clear-cutting, forest roads, damage induced changes
o Development of natural forest associations and/or species (climate change)
o Natural regeneration
o Assessment of hazard caused by the instability of forest stand
1.11 WP 510 User Utility Assessment
The User Utility Assessments (UUAs) of the FDS from the various users are compiled and collated in order to make an overall assessment on the sustainability of these products. Thus, this WP summarises all the main results of the UUAs and provide a Service Utility Report document which is an independent, comprehensive assessments of the value of the services delivered by the FDS. The Service Utility Report documents the complete use and life-cycle and impact of products and services delivered to a single end-user-organisation over a fixed period of time, analyse the overall results and recommendations such that a more comprehensive report is made on the overall sustainability of the FDS.
1.11.1 Objectives and Results
In this document the results of the UUAs are briefly reviewed and a synthesis has been elaborated focussing on the UUAs results. This synthesis reflects the state of the work and the functionality and utility of the FDS, including valuable feedback from the users. The latter issues encompass information gains from the DSS, technical issues such as the integration of products into the Forest GIS environment or the benefits from the DSS. On the other hand critics and constraints are always accompanied as important items in each of the production steps. The overall performance or cooperation of SPs together with users was a major topic in the realization of the DSS. It can be stated at the end of EUFODOS that the feedback to these questions showed a very good ‘climate’ between SPs and users.
However, it can be stated, that after Phase 1 the service prototypes have been prepared and ready for the role out in Phase 2 of the project. In order to continue with the service development in Phase 2 small adaptations and modifications were applied to the prototypes according to the feedback from the user’s side. The main task in Phase 2 was altogether the transfer – that means proof of operationality – from the small testsites into the roll-out (wall-to-wall) implementation onto the entire testcases. For this purpose the SPs showed high flexibility in the realisation and in shifting the thematic areas where appropriate or needed. During Phase 2 the cooperation between SPs and users were even intensified, because the finalisation of the products required more and intensive training sessions. This approach proofed to be a very efficient working procedure
The analysis of the life cycle and sustainability comprise the fields of service definition and service quality, data management, processing lines, product delivery and its validation. The impact of the services is discussed as well as sustainability and recommendations. In this context the SPs were busy to reach high standards in order to fulfil their obligations from the SLA. The outcome of EUFODOS proofed that this goal has been achieved for almost all derived products in the testcases. The only aspect to be mentioned in this context is that the service for assessing storm damage has been implemented from the technical point of view, but could only be realized in two small events in Germany and Poland.
During the user workshop the users have reported on their experience within EUFODOS from working procedure to the evaluation of the products. The EUFODOS users can be regarded to be representative for a large user community and if their statements to make use of the services will be transferred to other users or user groups, then it can be expected that an uptake by many other users will follow. From the finalized UUAs and the discussions during the meetings it can be concluded that the DSS are ready for the market. This is an important aspect as it supports the sustainability of the DSS and enables the SPs to be competitive on the market with the EUFODOS services.
Potential Impact:
1.1 Economic DS Service Model
1.1.1 Objectives and Results
This WP focuses on the overall economic model for the FDS. The objective was to define a Road Map for future technology and business development for the FDS industry based on information on the market structure, cost-benefit analysis and business modelling for FDS Services. The objective was also to evaluate the economic potential of each service case developed and demonstrated by the Service Providers.
The main results of this WP are the market analysis, the cost analysis, the value chain and business models, and the EUFODOS road map.
The study of the European market for FDS Services shows that forest monitoring based on Earth Observation technology serves a multi-layer market. Firstly, information on forest situation, activities and damages are needed by authorities on a global, national and regional level. In addition the same information is needed also in operational forest management and in planning and performing of wood harvesting. On a higher level information is needed for all forest areas, while forest management and harvesting focus on a local level. Yet, the value of the information provided by FDS Services for the customers is, in general, higher the closer you get to operative activities. Storm or insect damaged areas can be seen as a special niche in the forest monitoring market where high values are at stake. A special niche, where high values are involved, is monitoring of protective forest in Alpine areas.
Cost information of each demonstrated FDS Service was collected by the Service Providers to improve the understanding of cost structure of the services. The main challenge reported is that cost of satellite data, which is the biggest cost factor, varies significantly from case to case depending on availability, type and quality of the data. In practice, costs for a case analysis can be calculated only as the data is known. Availability of Sentinel 2 data can improve the situation significantly for the Service Providers.
The FDS Services can provide the Users with benefits of high value. In special cases like monitoring of protective forests in Alpine areas, the services can reduce risks on human lives and major infrastructure. In all cases, the services provide economic benefits through cost saving or improved performance. The economic values of the services vary from 10-15 euro up to several hundred euro per square kilometre depending on the environment and the service type. The challenge for the Service Provider is to develop a business model suited to capture these values through win-win co-operation with the Users. In deliverable 520.4 a set of potential business models are described.
The EUFODOS road map identify trend and events, that are likely to impact the development in the FDS industry in the near (2014-1016) and medium to distant future (2017-2020). The demand for FDS Services is expected to be driven by increases in exceptional weather conditions like storm and draught, and related insect infestations. The growing concern about the global environment will simultaneously increase the demand for information about the state of forest. This will affect both environmental and energy policies. Essential for the development of commercial FDS Service activities is EU’s future activities in space, space industry and research. Increasing demand for monitoring and cost cutting will require increased efficiency among regional authorities responsible for forest monitoring and management. Increased use of biomass in energy industry is expected to increase demand and competition for wood.
On the supply side, the EUFODOS road map emphasizes trends supporting availability and accessibility to EO data. Increasing numbers of satellites and satellite operators will improve the situation. Special focus in a near future will be on Sentinel 2. Also improvements in data storage and distribution will improve the conditions for providing services to the Users. The use of radar technology will improve information collection and the speed and reliability of service provision. Increased computing capacity will be essential as more and more information will be requested.
1.2 Concept Paper
Utilising the new emerging and developing opportunities based on the European COPERNICUS land core services the FP7 EUFODOS project has developed new and innovative downstream services, which are focusing on degradation issues in forests such as damage caused by storms, fire, snow-break or insect infestations. The downstream services are addressing needs of forest authorities, forest owners, forest companies and other stakeholders on national to local scale all over Europe.
On this background the EUFODOS team is now in addition proposing a PAN-EUROPEAN FOREST SERVICE that provides information on damage in forests for all member states in a uniform approach as a new Core Service. This service will provide continuous information on forest status for national and European policies. For this purpose a new WP530 was created in order to promote this idea.
1.2.1 Objectives and results
The EUFODOS consortium has shown a strong focus on user needs and hence the services have been developed in a custom-tailored approach, which representatively reflects the need of regional users. The purpose of EUFODOS is the development of Forest Downstream Services (FDS) in order to assess the extent of damage from forest degradation such as storm events, forest fires, snow-breaks or insect infestations. Additionally services are implemented for the derivation of functional parameters in protective as well as in managed forests. In this context technology is used from space- and airborne sensor platforms, which allows data to be acquired at short time intervals and in a cost effective way.
Further information on the user requirements, specifications or service deliveries can be found in the EUFODOS deliverable documents which are listed on the http://www.eufodos.info/content/eufodos-reports ; e.g. D210.2 report on service architecture component or D410.2 user needs and standard dossier for all services.
The methods and production chains developed by EUFODOS for these downstream services can built the basis and be used for the proposed Pan European Service.
During the course of the project several activities have evolved for the pursuit of the Pan- European idea. These encompass meetings, workshops and promotion activities. During these meetings the consortium members had the opportunity to present the concept and to include the feedback and proposals into the document. Several organisations and stakeholders have been contacted and the concept was presented. They are EEA, EFI, REA and from the Commission DG Enterprise and Industry, DG Environment, DG Agriculture and Rural Development, JRC.
1.3 Co-ordination of User Platform
1.3.1 Objectives and results
In this WP frame conditions for the development and coordination of the EUFODOS User Platform were established. A User Executive Body was founded in the first user workshop in January 2011.
The deliverable D610.2 contains an overview of the multiple activities taken to coordinate and inform the EUFODOS users. This comprises for example the organisation and conduction of four user workshops, four teleconferences and e-mail coordination of the reviews concerning the User Requirement template and of the User Utility template, collection of user input for the EUFODOS Road Map, the edition and dissemination of sixteen Newsletters as well as other promotion and information activities.
1.4 Web Portal / SDI
1.4.1 Objectives and results
The Work Package 620 has been managed by GAF and started in January 2011 (Month 1). Within the first half of the project (phase 1), the EUFODOS website as well as a Map-Portal (beta) has been established. During Month 19 to 27, several improvements and updates have been implemented on both Website and Map-Portal. New Key words, relevant for the different Forest Downstream Services in different languages (English, German, Polish, Finnish, Italian, Bulgarian and French) have been introduced in Meta description. Further activities have been made to get more reference links to EUFODOS from other sites. Both activities improved a bit the raking of EUFODOS in search engines such as “Google”. Nevertheless, themes such as storms, tornado, large insect infestations or fire are always main topics for all the News Sites in the internet and provide a big competition about high ranks in search engines. Another action has been proposed by GAF and finally implemented by the UEB: the new EUFODOS entry in Wikipedia, which hopefully improves the project and service dissemination activities. The News and Events sides, the Newsletter section as well as the Featured Content have been updated continuously. The programming of the EUFODOS Map-Portal (INSPIRE compliant as described in D320.6) has been finalized during the reporting period. Downstream Service products, results and relevant Metadata of all EUFODOS Service Providers have been implemented within the Map-Portal. A continuous content update of the Map-Portal is foreseen also after the end of the project (~ 2 years).
1.5 User Training
1.5.1 Objectives and Results
The objective of this work package was to organise and implement the interface between production and the user. On-site presentations of the service to the user will be made; this includes activities such as awareness raising workshops/meeting and/or specific training if required by the user.
It was decided to have at least one training session for the user receiving a service and to have in addition workshops with the intention to interest further users for the services offered by the beneficiaries. Training and workshop principles have been prepared and communicated internally (D630.1 & D630.4).
Training Sessions and workshops had been held during EUFODOS, see Training Activities (chapter 5.2) for an complete overview.
1.6 Point of contact for potential users of the EUFODOS products
1.6.1 Objectives and results
In this WP frame conditions for a Point of Contact for interested and potential EUFODOS users were established.
D.640.1 gives an overview on the promotion ad dissemination activities. This comprises the compilation of a list of 392 potential users from 46 countries and 12 international organisations to facilitate the Service Providers to contact potential users in the neighbouring countries. The EUFODOS Newsletters were distributed to all listed persons, to inform them about the various EUFODOS services and products. Also a questionnaire for the first contact with potential users was prepared. Further on, multiple dissemination activities have been done like contributions to forest or EO related Newsletters and Websites, contributions about EUFODOS in Wikipedia, publications, poster, flyer and presentations about EUFODOS, its services and products.
List of Websites:
www.eufodos.info
European Forest Downstream Services - Improved Information on Forest Structure and Damage
Forests play a key role in the European economy and environment. This role incorporates ecological as well as economic functions which can be affected by the occurrence of insect infestations, storms, fire or snowbreak events. Local, regional and national authorities thus require detailed information on the degradation status of their forests to be able to take appropriate countermeasures against forest damage and to ensure sustainable forest management. The EUFODOS project has proven that the use of state-of-the-art satellite and laser scanning technology can provide forest authorities with cost-effective, timely and comprehensive information on forest structure and damage.
The main goals
The Forest Downstream Services (FDS) within EUFODOS include in particular the assessment of forest damage and the measurement of functional parameters for commercial and protective forests. These monitoring services have been developed to an operational level by a consortium of research organisations and commercial service providers from Austria, Germany, Finland, Italy, Bulgaria and Poland.
The benefits
The use of space- and air-borne sensor platforms allows data to be acquired at short time intervals, in a cost effective way and over large areas. EUFODOS Forest Downstream Services thus offer the derivation of fast and reliable information for effective damage assessment and sustainable forest management. The products present space-related information to be used by users for a wide range of applications, such as:
- Effective damage assessment and countermeasures: identification of damaged areas – due to storm, fire, snow-break or insect infestations – in order to enable proper countermeasures, compensation payments and reforestation planning
- Sustainable management of protective forests: targeted management of protection forests in order to maintain and enhance their protective function against natural hazards
- Sustainable management of commercial forests: wood procurement planning and strategic investment planning for commercial forests
- Reporting: revision of forest maps and inventories, compilation of regular reports and annual statistics (e.g. changes in forested area), establishment of forest damage information system.
The EUFODOS consortium has reported on the results to an extensive forest user community via the newsletter distribution and still maintains the project webpage www.eufodos.info
Project Context and Objectives:
The European Earth Observation Programme GMES, currently renamed to COPERNICUS, provides data useful in tackling a range of issues including climate change and citizens’ security. The purpose of COPERNICUS is to deliver information which corresponds to user needs. In EUFODOS regional forest downstream services have been developed which are focusing on degradation issues such as the assessment of damage caused by disastrous events from storm, fire, snow-break or insect infestations. These services, which have been successfully implemented in pilot testcases, are demanded strongly by the regional users in order to fulfil their mandate. The EUFODOS services were developed to an operational level.
The use of space- and air-borne sensor platforms allows data to be acquired at short time intervals, in a cost effective way and over large areas. The Forest Downstream Services thus provide fast and reliable information for effective damage assessment and sustainable forest management. The data can be utilized by users in a wide range of applications:
• Effective damage assessment and countermeasures: identification of damaged areas – due to storm, fire or insect infestations – in order to enable proper countermeasures, compensation payments and reforestation planning.
• Sustainable management of protective forests: targeted management in order to maintain and enhance their protective function against natural hazards.
• Sustainable management of commercial forests: wood procurement planning and strategic investment planning for commercial forests.
• Reporting: revision of forest maps and inventories, compilation of regular reports and annual statistics (e.g. changes in forested area), establishment of forest damage information systems.
EUFODOS involves an extensive user community well connected to other related GMES/COPERNICUS User Groups. A list of 392 potential users from 46 countries and 12 international organisations was compiled, enabling the Service Providers to contact potential users in the European countries.
Project Results:
1.1 Frame Conditions for DS Operations
1.1.1 Objectives and Results
This WP, managed by GAF AG, aimed to develop overarching organisational frame conditions for an effective beneficiary interaction, for utilisation of the Forest Core products and for effective implementation of Forest Downstream Services. This comprised the formation of a Service Network of Service Providers, Users and the Research Community, as well as the development of a Service Partnership Protocol (co-operation procedures and communications) and the organisation of a User Executive Body (UEB). Another focus was on the development of a consistent IPR Strategy for the consortium and the harmonisation of the EUFODOS spatial data and services with the INSPIRE standards.
These requirements have been addressed as foreseen in the first 9 months of the project and have been documented in the Deliverable D210.1 Service Architecture Components. This Deliverable provides a review and outline of all the main components that form the relevant building blocks, or service architecture components, for the EUFODOS project implementation. This comprises an assessment of the different requirements for implementing various Forest Downstream Services for different users in Europe, thus providing the EUFODOS service development frame conditions. This allows a further assessment of how an optimal trade-off between the Forest Downstream Service requirements, user needs and technology constraints can be achieved. The final version of the Deliverable D210.1 for Phase 1 has been delivered in April 2011.
Most of the concepts and issues presented in the Deliverable D210.1 have a legacy in the GMES Service Element Forest Monitoring (GSE FM) which successfully provided Forest Downstream Services (FDS) to a variety of users in Europe and internationally. The report first presented the Organisational Structure of EUFODOS, bringing together key players from the research community, Service Providers and the user community and further presents the User Executive Body and explains its role within the Consortium and the project, especially to contribute to report reviews, identify user needs and representing the user community. All procedures related to Management Issues such as Reporting, Decision Making and Conflict Resolution as well as collaboration procedures with associated projects have been described in the Deliverable. The Deliverable also described the EUFODOS dissemination strategy, which aimed at the assessment of new markets and the extension of the user community. The Deliverable described the primary tools for dissemination, i.e. the project website, user federation mechanisms (i.e. the UEB), brochures, publications and service related trainings and workshops.
The Service Architecture Components Deliverable (D210.1) presented also the relevant regulations related to the entry / exit for new / current members further described in the EUFODOS Consortium Agreement, as well as the description of production procedures for Forest Downstream services and products. Troubleshooting, Quality Control procedures and standards, adopted for the various steps of the production process, are also described. The important issue of Intellectual Property Rights (IPR) has also been extensively explored and is described in the Deliverable D210.1 in terms of identifying the main IPR aspects that will affect EUFODOS and which hitherto had been addressed only in a limited manner in the EUFODOS Consortium Agreement. The FP7 IP Rules have been examined in detail and recommendations on how to manage IPR related to Ownership, Protection of Foreground, Access Rights to Background, and Dissemination are provided as strategies for the EUFODOS project to implement. Another key component explored in WP 210 and presented in D210.1 is the INSPIRE relevant harmonisation process necessary for Service implementation. This WP had a significant impact in the project definition as it outlined all the main components that formed the building blocks or service architecture for the EUFODOS project implementation. The first issue had been released in Phase 1 and an update had been submitted to REA in September 2012 (Phase 2 of EUFODOS).
1.2 Co-ordination with GMES Core and Analysis of Core Products Existing Systems
1.2.1 Objectives and Results
The Work Package has been managed by GAF and started in May 2011 (Month 5).
During Month 18 and Month 27, the two deliverables, D220.3 Assessment Report of Downstream Inputs and the Report on Coordinated Actions with GMES Core (D220.4) have been successfully updated from Phase 1 and submitted to REA.
The Assessment Report of Downstream Inputs assessed the potential data sources available to the GMES Forestry Downstream Services as provided by EUFODOS. It considered GMES core services and other „core“-type data sources, as well as to a minor extent also national data. A special focus had been put on core Earth Observation (EO) data as well as on Forest core products as produced by the GIO-Land project, managed by EEA. Data access conditions and procedures for the relevant Forest Downstream service input data, which are available free of charge and without restrictions, have been described. Relevant information to access EO data via the GMES Data Warehouse has been updated (e.g. information on RSS Feeds from ESA) and the most recent version 2.6 of the DAP document has been introduced. The section on the upcoming Sentinel-2 mission, one of the most relevant (for the successful implementation of Downstream Services in the future) EO HR satellite sensors has been updated accordingly.
The section on the High Resolution Forest Core products got a major update related to data availability of Forest Core products, produced within geoland2 (geoland2 project ended Dec 2012) and the GMES Initial Operations (GIO Land) project (2012-2014), managed by EEA. The GIO-Land project will provide operational Core Land products such as the HR Forest Layers (Tree Cover Density, Forest Type Layer) for EEA-39 countries based on Image 2011/12/13 data. Compared to geoland2, different HR layer product specifications have changed (e.g. Tree Cover against Forest Cover in geoland2). For future implementation of Forest Downstream Services, the new product specifications have to be taken into account and if possible, tested by EUFODOS service providers.
Another update has been provided about the GMES Emergency Management Service (GIO-EMS) project, the operational follow-up project to the FP7 project SAFER. The GMES Emergency Management Service (EMS) is one of the first GMES services that became operational. It started in April 2012 and consists of a set of services which are funded by the European Commission in the frame of the GMES Initial Operations (GIO). The GIO-EMS services are providing mapping products based on satellite imagery to actors in the management of natural and man-made disasters, in particular Civil Protection Authorities and Humanitarian Aid actors. There is no direct interference between Forest Downstream Services and the GIO-EMS project. Nevertheless, coordination e.g. in the field of EO data access/acquisition via the GMES Data Warehouse could be envisaged in the near future.
Within report D220.4 the status and potential coordination activities with former FP7 (e.g. geoland2, SAFER) and GMES Initial Operation projects (e.g. GIO-Land, GIO-EMS) have been updated. Important to know is, that no more-recent information is available on the final GIO data access policy by early 2013, except that the European Commission generally pursues and prepares a “free and open” access policy to all GMES data, which is deemed to also cover the GIO-Land HR Layers. A direct link and further communication of all developments to the project EUFODOS has been described and has been ensured by the direct involvement of GAF as a service provider in GIO-Land. Since several restrictions apply to the further distribution of internal knowledge that GAF has from this involvement in GIO, separate communications have been established between GAF and EEA on which service and product specifications. This has also led to a EUFODOS newsletter publication (EUFODOS 2013), providing all the necessary insight.
The product and service portfolio of the GIO-EMS comprises also reference and delineation maps similar to the type of services provided by EUFODOS. Specifically, burnt area maps are mentioned which are likely to be related also to forest areas. Therefore an information exchange of EUFODOS with the future ERCSGIO-EMS follow-up services under GIO should be envisaged and has been established by the WP 220 lead to avoid potential duplications in storm damage assessment during this operational phase of the ERCSGIO-EMS from the beginning. It has been agreed that the GIO-EMS shall inform EUFODOS immediately in case of activations in relation to Forest damage (by storm, fires, insects, etc.) within Europe. This procedure is deemed to avoid duplications in EO data orders and damage mapping.
1.3 Method for Utilizing Core Service Products
1.3.1 Objectives and Results
The aim of this WP, “Method for Utilizing Core Service Products” is to realise a cost and time saving procedure when utilizing the Core Service (CS) products for the generation of the Downstream Service (DS).
The key findings are:
• Availability of the Core Service: Only 2 out of 6 services cases are completely covered by geoaland2 HR layers. For 2 other cases, HR-layers were produced by GAF in the scope of EUFODOS, 2 test sites are not covered by HR-layers and HR-Layers had to be simulated.
• Use of the HR Forest Layers in EUFODOS: All service providers have implemented the HR-layers (or simulated products) in their processing lines. In most cases, the HR-Layers were used for masking forest areas in various stages of processing (e.g. for change detection). Furthermore, the HR-layers were used for training the classification processes, validation or as an initial input for a forest density map.
• Cost and time saving has been seen as the main added value of the integration of the HR-layers into the processing chains. Additionally, the integration of the HR-layers in parallel with other input data can improve accuracy and robustness of various EUFODOS products.
• Related to the potential advantages of integrating Core Service forest HR-layers are a couple of bottlenecks which could be improved
o Missing coverage: in EUFODOS, not all areas were covered by the Core Service products. In case of a commercial DS this would create additional costs
o Update frequency: For most EUFODOS DS products an update frequency of 1 year would be preferred. For damage assessment up-to date data for pre-event (reference) as well as for post-event is necessary.
o Accuracy, resolution and Minimum Mapping Unit (MMU): to detect damages in forest, in particular small scale damages (e.g. due to insect infestations), a high accuracy (>85%), high resolution (<= 10m) and a relatively small MMU (< 0,5ha) are required.
o Documentation and Metadata: to be able to better assess the quality of the HR-Layers a better documentation would be necessary (how was the dataset produced? Which data was used for validation? Validation report).
o For damage assessment a fast delivery of data would be essential. A solution could be to consider the rush-mode of the ESA Data Warehouse.
A problem to evaluate the full potential of the Core Service HR-Layer data within EUFODOS is the fact that EUFODOS is working with the geoland2 HR-Layers, while in the meanwhile EEA has contracted the production of the GIO HR layers with technical specifications different from the geoland2 products. Therefore, processing lines would have to be adapted to the new products as soon as they are available.
1.4 Standardisation of Production Chains
1.4.1 Objectives and Results
The WP has been managed by GAF AG and had contributions from all EUFODOS Service Providers. Its core aim was to develop and document overall standards, quality assurance systems and methodological advancements which were providing an overarching framework for the production of Forest downstream service cases. This was considered crucial in order to harmonise the varying capacities and levels of technical development existing in the current consortium. In order to achieve this, the WP aimed to build on the rich legacy of capacities, technologies and processing chains for forest downstream services developed and established already within the GSE Forest Monitoring project financed by ESA. A further aim was the facilitation of links to the COPERNICUS Core services. The WP also focused on the incorporation of INSPIRE standards into the processing chains, as well as on the need for a harmonised approach towards metadata, quality assurance and validation for all Service Providers.
The majorities of these requirements have been addressed as foreseen within the project duration and have been documented in three Deliverables per Phase. The first Deliverable was the D320.1 Service Portfolio Specifications, which provided a comprehensive description of all the EUFODOS product and service specifications. The primary goal was to present the technical specifications of the services and products which are required for product generation and service provision including quality systems and standards used. The production processes have been specified in a structured manner in order to support and enable standardization, automation, enhancement and growth. Thus all descriptions of the production process or service chain have been based on the four main components which are input data sources (EO/in-situ/models/others), pre-processing, thematic processing and the analytical or modeling component which is finally required for output production. One major focus has been laid on the use of the Core Image data (e.g. 2006, 2009 and 2012) as well as the High Resolution Forest Core layers. The Products in the EUFODOS Portfolio have been finally grouped into three main categories which are Rush Mode for Forest Damage Assessment (e.g. Storm damages), Non-Rush Mode for Forest Damage Assessment and the assessment of Ecological Functions as well as different Forest Parameters. Quality assurance of EUFODOS products has been based on internal quality assurance (within the Service Provider as well the Service Network) and an external, independent user validation. Internal quality checking and reporting systems have been designed to conform to ISO standards. These procedures guaranteed traceable service production standards and generated standardized quantitative output which can be used for further assessment. User validation supervised the overall quality of service production and verified user needs of EUFODOS products. Specific mechanisms have been developed to support the potential for technology transfer in upgrading the Service Portfolio. One of the main inputs for improved service provision and evolution of services was the inclusion of R&D in the development of the Downstream Service Portfolio. Main objectives of the Service Evolution activities were to identify, test, transfer and integrate mature results of R&D from research partners and wider R&D domains into the Service Portfolio, and, thus improving operational efficiencies and evolution of the Service Portfolio to better meet user requirements.
The second Deliverable was the D320.2/D320.5 Service Network Validation Protocol, which presented the Validation Protocol for the EUFODOS Project. It provided the overall framework for service and product validation within EUFODOS and summarized the steps of quality assurance documentation during the production process (validation of Products and Services) based on criteria such as e.g. thematic accuracy, thematic completeness, etc. It also specified the validation approach for Input data such as Earth Observation data, reference data and GMES Core datasets, as well as the utility assessment of products by the end-user, which has been considered a main part of the validation process. The Deliverable D320.2 and D320.5 further provided recommendations how to handle/document potential non-compliance (if it should occur) and necessary product/service improvements.
The (D320.3/D320.6) Service Network INSPIRE Implementation Technical Note as the third deliverable within WP320 was based on a two-phase approach (Phase 1: OGC compliancy, Phase 2 extension to INSPIRE compliancy, as possible). It provided the results of the investigations on the background as well as technical recommendations how to implement INSPIRE standards and requirements in the EUFODOS Forest Downstream Services. Key INSPIRE Requirements on Metadata, Data Specifications and Network Services have been explained in detail and the possible approach how to technically include these relevant Implementing Rules in EUFODOS and the respective Forest Downstream Services has been described. The INSPIRE schedule/roadmap, information on the data specifications for Land Cover and Land Use have been revised and new specifications related to Discovery, Viewing and Download Services have been provided. Whilst the INSPIRE process is still evolving with full implementation in 2019/2020 and not all INSPIRE Implementing Rules have been finalized so far, it was recommended in EUFODOS to use a two-phase approach, where the usage of OGC compliant Server Software is planned for implementation. In case that INSPIRE compliant Software is available at a later stage, it is recommended that service providers upgrade their server software to these INSPIRE compliant software versions. The Deliverable concluded with the presentation of different software options (free and proprietary), which have been recommended to provide INSPIRE compliant web services for EUFODOS. The main objectives have been achieved to provide an overarching framework for the production incl. key aspects in terms of standards, quality assurance and methodological advancements. Also the incorporation of INSPIRE standards in the production work and related to quality assurance and metadata has been achieved.
1.5 Development of Operational Toolboxes
At the beginning of the project a document was elaborated encompassing information to the SPs, such as an overview of existing software packages, the identification of software elements within EUFODOS or a review of processing chains. Based on this overview the SPs SW needs were asked and collected from them. This preparation step was followed by the elaboration of a questionnaire called “Questionnaire on Processing Lines and Methodologies in the EUFODOS Test Cases”. The analysis of the questionnaire and distribution of its results to the beneficiaries as well as a review of literature and the outcome of different forestry related projects finalized this step. As a completion of this introductory phase a SW training was organized and took place at JR in Graz from 26th to28th September. Thus, based on these preparatory steps the SPs have got a thorough knowledge of all the SW capabilities in the consortium and had the opportunity to use the toolboxes according to their needs. Within Phase 2 the operational tasks have been initialized and several refinements on the toolboxes performed. JR and VTT were thus enabled to optimize the toolboxes and finalize them according to the experience made in both phases. As an outcome of EUFODOS the toolboxes are now in an operational mode and are commercially viable products.
1.5.1 Objectives and Results
This started in month 3 and its main actors were JR and VTT. JR and VTT offered already developed in-house algorithms and methodologies for the forest parameter assessment which was tested on-demand and adopted within the processing chains by the service partners in order to improve the cost-effectiveness of the production. WP 330 focused, therefore, on the investigation and dissemination of new tool boxes/algorithms to facilitate the production of the service cases. The operational toolboxes were at hand to all SPs and were used for the processing of the data in the testcases. The other partners were also participating in this WP with their engagement in the various development phases. Continuous updating has been performed on the toolboxes based on the experience made during the roll-out phase and the feedback to JR and VTT in order to optimise the performance of the workflow.
The following software algorithm /toolboxes have been developed during the EUFODOS project:
• EUFODOS Change Detection toolbox
• Envimon SW for Satellite Data Pre-processing
• Probability SW for Forest Variable estimation
• AutoChange SW for Change detection
• FoVEA model refinement by sub-clustering
• EUFODOS LiDAR toolbox
The “EUFODOS Change Detection toolbox” has been developed and tested in Phase 1 on existing data sets. Furthermore a pilot project was initiated and contracted by the Forestry Board of Salzburg/AT applying the toolbox for the assessment of changes within a period of 10 years on the entire forest area of the province. It could be proved that the toolbox is an effective instrument which can be used to obtain information over large coverage in a cost effective way. In the finalization phase of EUFODOS some more small refinements were applied in the toolbox.
“Proba_cluster: Unsupervised clustering”, “Proba_plot (future: proba_model): Cluster labeling software“, and “Proba_Estimates: Estimation software for target variables” are software packages which were implemented for EUFODOS. In the boreal testcase they were used for the processing of the economic parameters and can be considered as operational tools.
“FoVEA stand variable data screening” software was developed for the testcase implementation in the boreal zone. A certain forest stand in a stand-wise reference data can be assumed to be homogenous, i.e. the forest characteristics should not vary too much within the delineated stand boundaries. In practice this assumption does not hold for the real data: the stands may contain large natural variations, that the data collection process does not take into account but only at average level (details below certain size are neglected), or the reference data may be out-of-date due to forest management actions or natural causes (data inconsistent). These variations in the forest structure can naturally be seen as large variations in the satellite image reflectance data. This adds unwanted noise in the modelling process. To avoid introducing this type of noise into the models, the reference data stands are screened for the EO data reflectance variations. As a result a new reduced stand variable data set is produced with the most homogenous (in the sense of image reflectance) stand. The method was found useful in producing more consistent models with increased maximum stem volumes, and it has been included in the standard repertoire of the FoVEA software.
“FoVEA model refinement by sub-clustering” was developed for the modelling cases in which there is relatively large amount of reference data available (more typical with stand variable data). The Probability method results in a model where many ground reference data vectors are associated with most of the clusters that represent forest (in the spectral data). The produced model is robust and typically generalizes well, but the drawback is that the maximum values that can be estimated with the model will be biased downwards due to this averaging. A method to compensate this effect was developed in EUFODOS. This method cannot be applied for models for which there are only few reference data available, thus the default value of 50 reference data vectors per cluster has been set as minimum to enable the sub-clustering (may be selected by the user). The sub-clustering method has been implemented as model refinement (or tuning) tool in the FoVEA software.
The “EUFODOS LiDAR toolbox” is a special case in the project and was initialized by the Austrian user LFD-STMK who showed a strong need for the assessment of forest parameters in the protective forest environment. As the state authorities had already initiated a campaign for a LiDAR survey it was a strong interest to use the high detailed LiDAR for the ascertainment of the needed parameters. JR had already some software elements for the processing of LiDAR data in house and as a consequence the development of a powerful toolbox was the major goal in WP330. As the “EUFODOS LiDAR toolbox” is a very complex software, it can be stated, that both phases were highly needed to create an operational software package. A special custom-tailored workflow was created in order to make the whole processing user-friendly. Especially in the final roll-out phase some challenges had to be solved in terms of data volume and computing speed. Finally the toolbox is now a user-friendly software environment which can be offered on a commercial base.
1.6 Assessing New Sensor Data
1.6.1 Objectives and Results
The work of this WP focuses on the testing and assessment of the utility of new sensor data and to evaluate their capabilities to meet the various thematic and technical product specifications. All studies have the objective in common to contribute improvements and enhancements to the EUFODOS services.
The following studies have been executed
1. Test of TerraSAR-X data processing options for mapping of small sized clear cuts to conclude on storm damage mapping – case study located in Poland
2. TerraSAR-X pre & post storm intensity based change detection for storm damage mapping – study area located in Thuringia, Germany
3. TerraSAR-X & Tandem-X single pass interferometry surface model quality assessment as background information for storm damage mapping – study area located in Karlsruhe, Germany
4. Pre & post storm based TerraSAR-X & Tandem-X single pass interferometry change detection for storm damage mapping – study area located in Thuringia, Germany
5. RapidEye for storm damage mapping – study area located in Poland
6. RapidEye for snow damage mapping & assessment of forest operations – study area located in Thuringia, Germany
7. RapidEye for bark beetle mapping – study area located in Baden-Württemberg, Germany
8. Suitability of spectral variables in mapping insect defoliation using RapidEye data – study area located in Brandenburg, Germany
9. Mapping of Forest Changes – study area located in Brandenburg Saxony
10. Digital surface model generation by digital aerial photography – study area located in Austria
All studies contributed to the enhancement and improvements of the services. The following findings of the studies have already integrated in the portfolio and processing lines of EUFODOS:
Table 8: Studies integrated in the portfolio and processing lines
Study Integrated results
Test of TerraSAR-X data processing options for mapping of small sized clear cuts to conclude on storm damage mapping – case study located in Poland Processing options for SAR Data.
TerraSAR-X pre & post storm intensity based change detection for storm damage mapping – study area located in Thuringia, Germany Integration of the SAR Option into the portfolio in Phase 2.
TerraSAR-X & Tandem-X single pass interferometry surface model quality assessment as background information for storm damage mapping – study area located in Karlsruhe, Germany The developments made in these studies show very high potential of the TerraSAR-X & Tandem-X single pass interferometry option to monitor storm damages. Still this option cannot be integrated in the portfolio since this option is from EO-Data availability side restricted. Still the findings can contribute to the benefit identification of future operational single pass interferometry satellite configurations.
Pre & post storm based TerraSAR-X & Tandem-X single pass interferometry change detection for storm damage mapping – study area located in Thuringia, Germany
RapidEye for storm damage mapping – study area located in Poland The results of the study show that a regional storm damage event that is dominated by small damage areas can be mapped with Rapid Eye with high accuracy using an approach where damages are mapped with a minimum mapping unit of 0.2 ha. Still, to enable a realistic estimate of the amount of damaged wood in case of a storm of with the damage structure of the examined storm damage event in North-West-Poland in 2011 it is not sufficient to consider only damages on completely damaged areas with a minimum area of 0.2 ha but scattered damages have to be taken into consideration as well.
It is planned that further options for the estimation scattered damages will be developed and that the estimation of scattered damages will be evaluated based on damage data per compartment.
RapidEye for snow damage mapping & assessment of forest operations – study area located in Thuringia, Germany Integration of the assessment of forest operations into the portfolio in Phase 2.
RapidEye for bark beetle mapping – study area located in Baden-Württemberg, Germany The study still showed the general feasibility of monitoring of bark beetle damage expansions in natural forest that are maintained without harvesting inventions.
It became obvious that small damages can be more adequately mapped with aerial photography taken by planes or UAV. Still for events with substantial damage dynamic Earth observation could substantially contribute.
Suitability of spectral variables in mapping insect defoliation using RapidEye data – study area located in Brandenburg, Germany Processing options for RapidEye data processing: The results of this particular study suggested that the NDRE, being a vegetation index containing the RedEdge band, is the variable most suitable for mapping symptoms visible in the pine forest canopy. That’s why it was taken as a basis for the development of the BlackBridge Forest Vitality & Change Monitoring Layer Bundle – a service fully mature for operational application. As already shown in EUFODOS D430.4 the service consist of 4 Layers (products) which can be delivered.
Mapping of Forest Changes – study area located in Brandenburg Saxony Processing options for RapidEye data processing.
Digital surface model generation by digital aerial photography – study area located in Austria The study showed that using digital surface model generation by digital aerial photography has a great potential for operational use.
1.7 User Requirements
1.7.1 Objectives and Results
The main objective of this WP is explicitly to assess the actual user needs and standards of all the users which requires the documentation of the environmental/forest policies and/or reporting requirements for the user that the consortium can support, the working practices of the user, technical capacities and standards, the potential benefits of services to the user organization, the availability of in-situ data, and the finalisation of the Service Level Agreements with users.
This WP is divided in two issues (phases). Generally the necessary information for deliveries production was collected by the use of specialised Questionnaires. The questionnaire in issue1 was very comprehensive. The questionnaire in issue2 was more concise, updating the user requirements and collecting new, important information.
The time between the two issues was dynamic and to obtain the changes which occurred during that time, a new, very concise, template/questionnaire was prepared. All the users have answered to the same questions.
An additional job was done - all the services required by the Users and which were in the process of development by all SPs up to the time of User Needs and Standards Dossier for all Services Issue 2 provision, were listed in a special table (Annex 3, report for issue2). Each service was assigned to a unique identification code. It would help the reference to all downstream services, and would help better familiarisation with these services.
The deliverable of each issue is a comprehensive Dossier giving a comprehensive overview on the user needs for each service and comprises the relevant issues in standard dossiers with respect to user relevant services.
1.8 Adaptation, Integration and Testing of Processing Chains
1.8.1 Objectives
This WP deals with the intensive phase of implementation encompassing adapting, integrating and testing of production chains in the individual testcases. The envisaged service in the EUFODOS consortium is built on a strong expertise and wide experience of the beneficiaries for the realisation of operational forest applications. In order to build up a FDS it is a main task to use the single expertise of each beneficiary in a synergistic way. That means, existing knowledge, software and service experience are exchanged and modified to enhance the service delivery chains. In this WP the realization of the user requirements and needs were realized for the different applications and services. Thus main activities included the coordination, monitoring and generation of delivery of services, the utilization of common infrastructure, and implementation of the SN quality assurance plan. The 6 testcases are described in detail below:
1.8.2 Testcase Styria – Service Provider: JR
The focus in the testcase “Upper Styria” is put on two applications, namely “Forest degradation from storm damage assessment” and “Impact of Forest Degradation on forest Functional Indicators”.
The first assigned forest product is focused on the derivation of windfall areas, meaning that map information is produced which displays damaged forest areas. The detailed service specifications for forest degradation were based on the needs and requirements defined by EAA in close cooperation with the Austrian Federal Ministry for Agriculture, Forestry, Environment and Water Management/Department on Forestry. The service specifications for “Impact of Forest Degradation on forest Functional Indicators” are based on the needs and requirements defined by LFD-STMK. In the selected test site the derivation of functional parameters is focussing on the protective functions, which are closely related to degradation issues. A very powerful approach for the derivation of these parameters was the synergetic use of laserscanner (LiDAR) data providing 3D structural parameters (height, crown closure and vertical structure) with Core Service products, procured from GEOLAND II (LMCS: land monitoring core service) or GIO (GMES initial operations), providing information on tree species distribution. In the following the main issues of the roll-out phase are described in brief:
1.8.2.1 Forest degradation from storm damage assessment
The processing line for this service has been developed in WP330 “Development of Operational Toolboxes” and has been tested in WP420. The SP JR has successfully tested the new processing line for an operational service, which shall serve as a demonstration example in order to acquire further orders. In a pilot project the developed software Toolbox could be applied for an application called “Waldveränderung Salzburg”, which has its focus on the detection and delineation of forest degradation changes within a period of ten years. In this project the forest authorities of Salzburg have shown a strong interest in changes within this period in order to integrate them as important information layer into their forest management plan. In general, it can be stated that the performance of the processing line worked well and it is easy in the handling process.
1.8.2.2 Impact of Forest degradation on forest functional indicators
JR has chosen a series of LiDAR tools for the implementation of the EUFODOS LiDAR toolbox. The single modules have been tested and adapted to the requirements of the EUFODOS application. An on-going activity during this phase was the interaction with the user LFD-STMK which has also urged the necessity to adapt the toolbox continuously. Another important aspect during the processing phase was the optimisation of the processing chain especially in terms of computing speed, because the LiDAR data have huge data volume.
The goal of the LiDAR processing chain is the generation of forest functional indicators that will help to monitor degradation effects in Alpine forests in a very detailed manner. Following functional parameters, defined by the user LFD-STMK, were derived on stand level:
- Tree height
- Vertical stand structure
- Crown/canopy volume
- Crown cover density
- Development stages
- Understorage
- Forest mask (including the upper forest border)
For the implementation of this processing chain existing methods were adapted or if needed new ones developed. All methods were developed and tested at the EUFODOS small testsite Hohentauern in upper Styria. Meetings with the local users - LFD-STMK - were organized, to discuss their requirements. It was found that the definition of the alpine forest boundary as such was a very important issue. The derivation of the –critical – upper forest border line from LiDAR data proved to be highly accurate. The calculation of forest stand polygons by segmentation and the tree top detection are both fully automatic approaches leading to highly appreciated results for the user.
As already mentioned the direct use of the Core Service product P-EL-4B “Forest Types” is not possible, thus a method has been developed to obtain a product with an acceptable spatial resolution. In this context the already existing CS product was used to classify the 5m RapidEye imagery and thus saved time and costs which would have been needed for the ascertainment of appropriate training data sets.
1.8.2.3 Conclusions
The change detection processing line has been implemented into an easy following workflow, which enables the operator to apply it in fast and efficient way. The entire LiDAR processing chain has been implemented as a workflow which offers a guided tour throughout the processing line. The user LFD-STMK has validated the final LiDAR parameters especially with the focus for their processing within the forest GIS. It can be stated that from the current findings the use of LiDAR data is operationally applicable and leads to incremented value for the management of protective forest. Both toolboxes, the “EUFODOS Change Detection toolbox” and “EUFODOS LiDAR toolbox”, could be developed to an operational status and are now ready for commercial selling too.
1.8.3 Testcase Finland/Russia Service Provider: VTT
The VTT Technical Research Centre of Finland was developing EUFODOS Downstream Services for two test areas: one in South-Eastern Finland, the other one in Russia (Nizhniy-Novgorod). The VTT Test Case included the production of 1) Stem volume maps for selected test case areas in Finland and in Russia. The stem volumes were produced as total stem volume, and as species-wise products (pine, spruce, broadleaved). During the project a new product called 2) The Assessment of Storm Damage Economic Impact was included into the VTT service portfolio.
Development of the Production Chain
To set up the Forest Downstream Service production system for VTT Test case (P1 – P5, see SLA for the definition) an analysis of the production chain (called the Probability chain) was performed with the identification of major software development needs (no conceptual gaps were found).
The FDS production chain was subjected to operational tests (test per SW module and end-to-end test) with selected test case data in order to assure a consistent production flow. Three mutual working meetings with the user organization (Stora Enso Oyj) were organized for the elaboration and refinement of the user requirements.
The most important development needs from the user point of view were:
• To improve forest variable estimate accuracy, especially for the higher stem volumes
• To improve the separation of conifers (pine and spruce) in mixed conifer forest
• To produce indication of map reliability
• To obtain accurate information of storm damages
From the system level point of view the need to develop the existing in-house software tools into operational level was considered as major development driver.
The software development work included modification of the pre-processing software Envimon for the Image 2009 IRS data and RapidEye data. Other modifications were implemented to facilitate the Probability software usage in operative production, and a modification to facilitate the usage of CORE service data in map production. Implementation of batch-processing for EO data pre-processing phase was partially included in EUFODOS activities.
New software modules were written into FoVEA software - a GUI for running Probability chain programs. These included modules to screen reference data outliers (to improve model consistency) and to evaluate the product accuracy. Also tools for model cross-validation (K-fold cross-validation) and for model fine-tuning (sub-clustering method) were implemented.
Data Acquisition
The data acquisition plan that was originally written in SLA for the VTT test case areas in South-Eastern Finland and in Russia (Nizhniy Novgorod) was updated. The DWH was screened for suitable HR data sets for products P1 and P2. The primary set of optical EO data included four partly overlapping IRS-P6-LISS3 scenes from the South-East Finland area from summer 2010, and two additional image frames from 2008. The cloud cover of the scenes varied from 0% to 25%. Two scenes from year 2010 were selected for the production chain assembly and testing, and for the demonstration product generation. These scenes cover roughly ¾ of the AOI (Area of Interest) in Finland that was defined in the SLA, and included large area north to the AOI (). Three RapidEye image from June 2011 was selected for the first tests (P5) from the Finnish test site. From the second test area (Nizhniy Novgorod) two SPOT5 images from 2010 were acquired from the DWH. One of these was too cloudy to use for the production, but the other one, covering some 80% of the service case area, was used for demonstration.
To facilitate the production of VHR enhanced products (P3 and P4) the VHR image archives were screened for the years 2010 and 2011. For the Finnish test site the three RapidEye scenes were selected as the VHR data for the product P3 development, as they had a relatively good common coverage with the reference data set. For the Russian test area there was no VHR data available, and the product P4 was cancelled from the VTT’s portfolio in agreement with Stora Enso.
EO Data Pre-processing
The EO data selected for the demonstration was subjected to standard QA procedures and pre-processing steps (unpack, radiometric calibration, atmospheric correction). Relative calibration of the EO data with respect to the model image(s) was also performed where necessary. The production of cloud masks and water masks followed the automatic pre-processing steps. Topographic correction was not performed.
Preparation of Reference Data
The forest area in the Finnish test area is pine dominated forest with average stem volume of 116 m3/ha (stdev = 87 m3/ha; as calculated from the stand-wise reference data set). The reference data of year 2010 was extracted from the Stora Enso’s database. It has been interpreted from aerial imagery with ground verification and bi-annual updates. The data contained 16 variables from which the total stem volume and the stem volumes by tree species (pine, spruce and broadleaved trees) were selected for the demonstration products. The resulting stand variable data set consisted of 250 000 forest stands. The reference data set was screened to contain 52% of he most homogenous forest stands (as evaluated with the IRS imagery). A test reference data set of 504 field locations was extracted from this set using stratified random sampling, with total stem volume (ToV) as stratification variable. This test data set was excluded from the training data set, and was used throughout the production for model cross-validation and final product testing. It has to be noted, that the stand variable data contained significant errors (up to 30%; mostly due to forest management actions), which also had an impact for the FDS production chain software development.
The reference data for the Russian test site consisted of 156 field plots, having stem volumes for pine, spruce, birch, aspen and other broadleaved species. The reference data were converted to the reference coordinate system of the EO image (WGS84 NUTM38). Only 56 of the original set of 156 field plots were on the image area, and 20 additional plots had to be removed due to inconsistency, clouds or burnt areas. Thus only 26 reference data plots were available for modelling, which affected to the model quality.
Utilization of CORE products
To test the possible benefits of using CORE data in the downstream service production, a ‘simulated’ CORE data set was generated for the Finnish test site. The model for the CORE data production was built for Estonia in the GIO Land project, using the principles and experience gained in Geoland2 project. Even if the structure or tree species composition of the Estonian forest do not match exactly the ones in South-Eastern Finland, the model was considered to be accurate enough to be used in the testing. The training image was matched with the Estonian model training image reflectances with channel-wise relative calibration. The produced CORE data set included the layers crown cover percentage (CC) and proportion of broadleaved trees (BL%) as continuous variables (0 – 100%).
Conclusions
The implemented changes of the software tools for forest structural forest variables mapping (Probability, FoVEA and Envimon SW) improve the efficiency of the thematic map production considerably and will shorten the product delivery times. The achievements in the EUFODOS software development may be summarized as:
• Usage of CORE service data shortens the production time
• Batch pre-processing for large data sets shortens the production time (considerably)
• The developed new methods (model sub-clustering & reference data outlier removal) improve the product dynamic range by 25% - 35% without introducing bias or decreasing product accuracy
• Improvement in model consistency and dynamic range by the new reference data screening methods
• New tools for product accuracy assessment give user an indication of map reliability and support in decision making
The developed software packages are being commercialized at VTT and are available for forestry actors worldwide.
1.8.4 Testcase Thuringia Service Provider: GAF / ALU-FR
The Service Demonstration in Thuringia was based on a Service Level Agreement (including an update version from End of December 2012) between ThüringenForst in Gera (as user) and GAF AG in Munich (as Service Provider). The Service Level Agreement was updated in December 2012. GAF AG with support of ALU-FR (University of Freiburg, Department for Remote Sensing and Landscape Information System) delivered the following Downstream Services to ThüringenForst based on the SLA and within the Project lifetime of EUFODOS:
• Rush Service - Forest Damage Assessment
• Non-Rush Service – Forest Damage Assessment
• Assessment of Changes due to Forest Operations and Forest Degradation
Main contact point on User side was the Service and Competence Centre (SuK) of ThüringenForst located in Gera and represented in EUFODOS by S. Chmara and H. Sagischewski.
From beginning on, the HR Forest Core Layer were a pre-requisite for the Development of the above mentioned Downstream-Services. From the FP7 project geoland2, no Forest Core Layers were available for the test sites in Thuringia. Therefore, the HR Forest Core Layers had been produced for Thuringia within the first 4 Months of EUFODOS. The production was based on Image 2009 and from GIO-Land (lead by EEA) project; the actual product specifications were used. The Tree Cover Density and Forest Type (HR Core Layer) as well as a Tree Cover Map (derived from the Tree Cover Density Layer) and a Spruce Type Map built the basis for all further Downstream Service Development activities and product delivery.
Service Demonstration: Rush Service – Forest Damage Assessment
On 5th July 2012 a tornado caused damages in a forest near the City of Gera in Thuringia (Germany). Around 20 ha of an urban forest area covered with approx. 80% conifers and 20% deciduous trees was affected by this event. The event has been used to demonstrate the usability of the “Non-Rush- Damage Assessment Service”. The service included the use of EO (e.g. Image 2009) as the basis for reference mapping. In addition, SAR data (TerraSAR-X) from before and after the Tornado event was acquired to test a multi-temporal approach for the classification of the damaged areas. The classified damaged areas as well as all relevant additional data were delivered to the user ThüringenForst in September 2013. The acquired SAR data were analyzed and classified by ALU-FR. The developed processing chain was later transferred to and tested by GAF AG. A field mission already conducted in spring 2013provided the relevant figures for the thematic accuracy assessment. With an overall thematic accuracy of 87.5%, most of the damaged areas were identified successfully. Beside this, the shape of the areas found on ground differed strongly from the shape of the areas classified using SAR data. Nevertheless with the used method, a first and fast determination of damaged locations is possible with a high object related accuracy; that is of high interest for the user. The Rush Service is still not fully operational as the test case in EUFODOS was implemented under more or less optimal conditions (summer storm, flat area). As most of the heavy storm events in the past were taking place during the winter months (November - February/March), the user ThüringenForst and the Service Provider GAF AG decided after the last training session in November 2013, to implement a further test. It is planned to finally proof the service capability under extreme (snow, hilly terrain, bad weather) conditions in Thuringia during winter 2013/2014. Nevertheless, other Forest Administrations have shown high interest in this service and are thinking to implement this service in their plans for disaster management.
Service Demonstration: Non-Rush Service – Forest Damage Assessment
The Tornado on 5th July 2012 was also used to demonstrate the usability of the “Non-Rush Service – Damage Assessment”. The service includes the use of EO (e.g. Image 2009) as the basis for reference mapping as well as the derived HR Forest Core Layers (e.g. Tree Cover Density). EO data (RapidEye) after the Tornado has been acquired in order to facilitate damage assessment caused by the Tornado event. The classified damaged areas as well as all relevant additional data were delivered to the user ThüringenForst in December 2012. A field mission, supported by ALU-FR was implemented in spring 2013 and a very high thematic accuracy was achieved. All relevant service specifications were met. It was possible to identify damaged areas down to 0.1 ha.
Another case was used to further test the capability of the Non-Rush Service. In winter 2010/2011 heavy damages in forest caused by snow were reported in the forest district Stadtroda, Thuringia. Purpose of this study was, to compare the use of the HR Forest Core Layer and the use of RapidEye as reference data before the damages occurred. Therefore damaged forest areas (snow break) were classified based on RapidEye to RapidEye data without direct use of HR Image 2009 data, but in combination with the available Forest GIS data (mainly forest stand borders) from Thuringia and provided by ThüringenForst. The mapping activity achieved similar thematic accuracies compared to the Tornado test case described above.
In both cases, the overall workflow to derive the damaged areas included a segmentation based change classification algorithm. Additionally, the Tree Cover Density product (already produced in phase 1 of EUFODOS) was used to further minimize potential classification errors. The processing chain developed in EUFODOS has proven its usability. The development is finalized and the Service can be offered to a wider user group also outside of EUFODOS. Based on promotion activities, several forest administrations are already interested in this service.
Service Demonstration: Assessment of Changes due to Forest Operations and Forest Degradation
Federal forest administrations urgently need up-to-date information about the entire forest area and forest area change to carry out different tasks such as the implementation of political objectives pertaining to forestry, the protective functions of the forest, habitat mapping and recreational function of the forest as well as support of consulting services offered by the forest administration to private forest owners. Today, state forest administrations derive most information based on State Forest Inventory data with an update frequency of usually 10 years. Information about private and communal forest is most often excluded from State Forest Inventory and therefore not available on a regular basis and/or outdated. The developed Forest Downstream Service was aiming to close these information gaps by providing up-to-date information about the whole forest area and its changes on an annual basis. Within the test case, two areas were selected in Thuringia covering ¼ or ~ 4.000 km² of the Federal State Thuringia. As basis, the produced HR Core data and Core Image data, as well as RapidEye data from 2011 and 2013 were used. Pre-processing of the EO data (Image 2009, RapidEye from 210/2011 and 2013) comprised a full radiometric correction including also topographic normalization and cloud masking. The interactive segmentation based change detection to derive a map of potential changes between the EO images from different point in time was based on the software eCognition. Also manual editing and enhancement was done with eCognition. The use of the User Forest GIS as well as the HR Forest Core Layers was supporting the manual editing. All processing steps were followed by Quality Control documentation. Overall the production workflow was prepared in a way that a flexible scalability was possible. Depending on the available EO data and given reference data, small as well as large working units can be prepared and processed using the given workflow.
The service was successfully tested and implemented in the above mentioned test regions. Change Areas down to 0.1 ha were identified with an overall thematic accuracy of >90%. Within an operational production environment the service can deliver accuracy values between 90% -95% in dependence of the extent of the change areas. Additional manual enhancement and good knowledge about image interpretation can further improve the accuracy of the delineation of the damaged areas. The use of 5 to 10 m EO data such as RapidEye or upcoming Sentinel 2 data as base data for the change detection will further improve the thematic and geometric accuracy of the results. As the service was implemented successfully during EUFODOS, service development is finalized and the Service will be offered in the near future to interested user groups in Europe.
1.8.5 Testcase Swieradow Service Provider: REN
The service demonstration followed a service level agreement between the user Forest Inspectorate Świeradów, ul. 11-go Listopada 1, 59-850 Świeradów Zdrój, Poland, in partnership with FOREST RESEARCH INSTITUTE, 3 Braci Leśnej Street, Sękocin Stary, 05-090 Raszyn, Poland and RapidEye AG, RapidEye AG, Molkenmarkt 30, 14776 Brandenburg an der Havel, later changed into BlackBridge AG, Kurfürstendamm 22, 10719 Berlin, Germany as the Service Provider. It foresaw the delivery of the following Downstream Service Products:
• Product P-01 Time Series of Custom RapidEye Ortho-Image Maps,
• Product P-02 Forest Cover Map,
• Product P-03 Storm Damage Map,
• Product P-04 Forest Vitality Map,
for an area of about 70,000 ha in the Western Sudety mountains, near the country triangle formed by Germany, Poland and the Czech Republic (forest districts Świeradów and a part of Sklarska Poreba).
The main point of contact on users side was Mr. Radomir Balaczy, with whom a number of co-ordination meetings were held before and during the project execution period.
First the GMES core products were ordered for the test area and prepared by GAF AG. Based on this information layers the specific customer expectations were evaluated and laid out in an updated SLA with the user. Later RapidEye imagery was acquired over the area and the development of an enhanced Forest Cover Map product started. In a first step an advanced pre-processing procedure was developed and applied. It included the customized ortho-rectification based on user-provided data, special radiometric correction and mosaicking of the satellite images, necessary to cover the full area. During this step tools and technologies, provided by EUFODOS project partners were used and evaluated. The production process was executed for all following RapidEye image coverages and resulted in the Product P-01.
Using the first coverage of P-01 products a technology for production of alternative Forest Cover products was drafted, implemented and applied for test purposes. After a visual and statistical accuracy assessment the product was delivered to the user and field checked during a joint field campaign with the user. The sample forest cover map was well appreciated by the customer and accepted as EUFODOS product P-02.
An older storm case in Saxony was used for the development of a storm damage mapping technology. We made use of the previously developed technologies for products P-01 and P-02. In July 2012 a tornado event occurred in the Trzebciny area of Poland. With the help of our user we applied the developed processing chain to this event, produced and evaluated the storm damage product. Comparative accuracy analysis, using existent Forest GIS databases and data, collected during the removal of damaged timber from the area, was carried out by our user. Later he used the results for dissemination activities in Poland. The results were presented by Radomir Balaczy at National Foresters Conferences and Meetings and earned appreciation.
Meanwhile a forth product was developed for mapping of forest vitality in the test area. It first based on the multi-temporal coverage with RapidEye data. The maps showed just the seasonal vitality change. But there is a much higher value of such maps, if produced for current forest stress situation.
The RapidEye / BlackBridge project team identified a potential need for such products in the eastern German Federal States, which suffered from insect infestations in oak and pine stands. First the Brandenburg State Competence Center for Forestry became an additional user organization for EUFODOS (but without special SLA in the sense of the project), for which we produced a vitality map of infested stands. Together with specialists form the State Competence Center and local foresters the maps were checked in field surveys. They assessed the prototype map products as helpful and gave useful hints for further improvements. This led to a series of maps for infested regions, now called the Vitality Map Bundle.
Following the intention of the EUFODOS project additional potentially interested users of the developed DS were invited for service demonstration: the state forest authorities of the German Federal States of Saxony, Lower Saxony, Mecklenburg-Eastern Pommerania and Brandenburg. An overall number of 9 test cases were processed and successfully evaluated together with specialists from user side.
1.8.6 Testcase Bolzano Service Provider: EURAC
The testcase Bolzano is based on the Service Level Agreement (SLA) concluded between the service provider (EURAC, Institute for Applied Remote Sensing, Bolzano, Italy) and the user (Department of Forest Planning, Province of Bolzano, Italy). According to the SLA, the demonstration included the presentation and delivery of the following Downstream Service Products:
• Product P-1 Base data on forest,
• Product P-2 After Event maps,
• Product P-3 After event monitoring product (recovery),
• Product P-4 Forest Damage Information System (FODIS).
Thereby, P-1 should cover the entire area of the Province of Bolzano with a total area of 7,400 km², while P-2 and P-3 were foreseen for a total coverage of about 300 km². P-4 is a spatially explicit information system on forest damages integrating the products P-1 to P-3.
The main point of contact on the users side was Mr. Günther Unterthiner, with whom regular co-ordination meetings were held before and during the project duration.
In the initial phase of the project, the Department of Forest Planning provided forest damage data to the EURAC which were used to start the work on P-2. These data mainly consisted of broad places and described impact of insect infestations. A timeseries interpretation technique was developed using broad-scale MODIS data to identify these areas, however, this attempt was not continued as the results were not sufficient enough. Moreover, in June 2011, a storm event occurred in the north-eastern part of the Province of Bolzano. This event was then used to develop and implement a forest damage mapping approach based on high-resolution data and a forest change map was derived (P-2). The accuracy of the map was visually and statistically assessed and compared to the field mapping results of the foresters. The results were delivered to the respective foresters who field checked the results. Subsequently, the product was presented in detail to the foresters who very positively evaluated the product and confirmed that the accuracy of the change product outperformed the field mapping attempts in almost all cases. Due to the very positive feedback and the planned operational integration of this product on an annual base in the FODIS, EURAC modified the processing chain and derived annual change maps for the period 2009-2013.
In parallel, product P-3 was developed to allow continuous monitoring of forest intervention measures. The user provided different afforestation areas and EURAC evaluated different options on how to provide relevant information. Timeseries of Landsat for 1984-2012 and RapidEye data for 2009-2013 were utilized and an approach was develop to show spatial-temporal changes of the vegetation cover.
Product P-1 was developed to map forest type and forest extent based on multi-temporal RapidEye data for 2012. A procedure was developed based on the existing Geoland tree type information to derive reference data and a reliable tree type layer was generated. This layer was transferred to a forest type layer by consulting the user and agree on a forest definition applicable to the data source.
All three products were delivered to the user and positively evaluated. Moreover, all layers were integrated in the FODIS which was developed in the project.
1.8.7 Testcase Sofia Service Provider: RESAC
The service demonstration of the products delivered by Remote Sensing Application Center – ReSAC Bulgaria followed a service level agreement between the user Executive Forest Agency (EFA), 55 Hristo Botev Blvd., 1000, Sofia, Bulgaria and ReSAC, 61 Tsar Asen Str., 1463, Sofia, Bulgaria (Service Provider).
The main point of contact on users side were Mr. Nikolay Pironkov, Mr. Vladi Konstantinov and Mr. Spas Tumbev, with whom a number of co-ordination meetings were held before and during the project execution period.
During the lifetime of the project and after the first phase when the demo products were delivered, the User drafted its user requirements, hence the final services which were agreed changed to: Forest Inventory Service divided to: P-01-1 Forest Area Map; P-01-2 Forest Type Map; P-02 Forest Change Map; P-03 Forest Density Map. The services for damage assessment changed as follows: Forest Disasters Service; P-05-1 Reference Map (Forest Cover Map and Forest Type Map); P-05-2 Damage Extent Map (Fire Extent Map, Storm Extent Map and Insect Infestation Extent Map); P-05-3 Detailed Damage Map (Crown/Surface Fire Map, Storm/Snowbreak Damage Map and Insect Infestation Damage Map); P-05-4 Additional products (Combination of the EO products with cadastral information).
On the other hand the test area which was preliminary chosen was extended to other territories in Bulgaria where natural disasters in the forest territories has occurred: Area 1 - Gostun Village, West Rhodopy Mountain; Area 2 – Tran, Breznik, Zemen Forestries, West Bulgaria; Area 3 – Sredec Forestry, Strandzha Mountain; Area 4 – Nature Park Vitosha, Vitosha Mountain; Area 5 – Koprivstitsa Town; Sredna Gora Mountain; Area 6 – Silistra Municipality, North East Bulgaria; and Area 6 – Danube Plain.
The first products delivered were the three types of the Forest Disaster Services. The products were developed for a small fire event in the Area 2. During the field check the EUFODOS project and services were presented to the local forestry. Together with the P05-1 and P05-3 products for the test case were generated also products P05-2, which were more detailed products based on User demands to combine the Forest cadastre with the disaster products.
The second products delivered were the ones from the Forest Inventory Services. These products were delivered for Area 2. The main image source used was Rapid Eye. The data were atmospherically corrected, geometrically corrected and using the methodology developed by ReSAC combining different vegetation indices the products P0-1, and P0-3 were delivered. Both services were delivered to the EFA and training on the products was performed.
The products for snowfall/snowbreak were delivered for the Area 3. The main image source used was Rapdi Eye data. The first image requested was in early spring and no good results were obtained because the leafs of the deciduous trees were not developed. A second image was requested for early summer which was combined with the reference image from the previous year and P0-2 Forest Change Map was produced. All the P05 products developed were quality checked and the accuracy was within the 90% specifications. During the field trip the EUFODOS services were presented on the administration of the local Forestry Sredec, and they were the user to which the products were delivered. ReSAC has delivered also Quantum GIS open source software, and conduct one day training on the software and data delivered.
For insect infestation ReSAC requested VHR data for the test area – Ikonos, WV2. In addition orthophoto from the airborne sensor was used. The processing was focused this time on working with VHR data and using the 8 channels of the WV2 data. The same for the storm event the methodology developed for the snowbreak event (area 3) was used. The spectral data of the WV2 data were used for mapping the insect infestation in the national reserve Bistrishko Branishte in the Vitosha Mountain. All the products were quality checked by a field trip campaign.
In the spring of 2013 a storm event occured in Area 5. The local forestry requested through EFA F05 products in order to map the main damaged areas. ReSAC informed the EUFODOS partners for the event and Terra SAR data were delivered. The processing of the data was not successful due to the big time interval between the two scenes.
During the same time (spring 2013) P05-4 were tested (area 6). The main source this time was LIDAR. The data were combined with orthophoto and Rapid Eye data and differentiation between tree species and forest height was done. The products were presented to the user as a option for further development as it was not in the initial SLA.
After the meeting with EEA in Copenhagen ReSAC has obtained the FMCS developed in the frame of GIO, together with the CORE3 data from Rapid Eye. For the test Area 7, for which ReSAC had additional auxiliary information such as LPIS, and FMP all the Forest Inventory Services were developed using the updated methodology. The products were delivered in the final seminar (17th of Dec 2013) to the EFA and presented to the 6 Forest Enterprises, and regional directions of the EFA.
The products developed by ReSAC were presented in 9 international events (conferences and workshops) and 8 events in Bulgaria. ReSAC delivered to the user 4 reports, more than 100 maps, around 100 thematic products (including reference GIS layers) and close to 30 orthorectified EO images.
1.9 Demonstration
1.9.1 Objectives and Results
During the EUFODOS project prototype demonstration products were produced in accordance with the user requirements and the prototype processing chains developed before. Thus main activities included the coordination, monitoring and generation of delivery of services, the utilisation of common infrastructure, and implementation of the SN quality assurance plan.
The activities of this WP had been undertaken by all SPs.
It was the task to execute all procedures to generate the information products, and to perform all required routine calibration and verification procedures (quality assurance). Some of the developments were reviewed, modified, adjusted or changed as a result of new findings, new ideas or changes in the user requirements. In some cases, new test areas were adopted and new users integrated, for instance for verification of results.
1.10 WP440 User Validation of Core and DS Services
1.10.1 Objectives and Results
To assess the service utility for the user, EURAC has designed an interactive questionnaire which was discussed with the EUFODOS user board in advance. All 6 users answered the questionnaire after they received the demonstrator products (D440.1) and after receiving the final products (D440.2) for the EUFODOS Downstream Services by their respective service provider. The results in short:
• State of work
o 95% of all services are completed or almost completed, 86% are completed.
• Functionality and utility of the DS-services
o All users rated the demonstration products very positively
o The DS-services are in most cases not replacing existing services but offering complementary or new information
o The level of integration was mostly rated as high and the ease of integration as easy
o Most agreed benefits were time saving and easy comprehensibility of the products. In most cases, the DS-services added improved monitoring capabilities.
o Potential constraints include the dependency of the users towards the service provider and the potentially higher costs if additional data is needed
• Sustainability of the service.
o Users are very interested in sustaining the services and would be willing to pay for selected services.
o Important for all services based on High Resolution data would be the availability of SENTINEL 2 data, which would reduce costs and allow for a higher update cycle and a more rapid availability of the service.
o For more than 40% of the products updates are planned, for 30% considered, but not yet clear.
• Assessment of Procedures and recommended Improvements
o The negotiation phase and the production phase for the products were in general rated as very good.
• Assessment of Impacts and Value Statement
o 80% of the users stated, that the EUFODOS DS-products have a significant impact on their institution. Many users have already recommended the product to other potential users.
• Overall Evaluation and Outlook
o All users evaluated the products as good or very good.
• Suggestions for further improvments included
o Information on the state of forest with protection function
o Information on typical forest parameters (volume, …)
o Update services based on Sentinel 2 data
o Upper forest border
o clear-cutting, forest roads, damage induced changes
o Development of natural forest associations and/or species (climate change)
o Natural regeneration
o Assessment of hazard caused by the instability of forest stand
1.11 WP 510 User Utility Assessment
The User Utility Assessments (UUAs) of the FDS from the various users are compiled and collated in order to make an overall assessment on the sustainability of these products. Thus, this WP summarises all the main results of the UUAs and provide a Service Utility Report document which is an independent, comprehensive assessments of the value of the services delivered by the FDS. The Service Utility Report documents the complete use and life-cycle and impact of products and services delivered to a single end-user-organisation over a fixed period of time, analyse the overall results and recommendations such that a more comprehensive report is made on the overall sustainability of the FDS.
1.11.1 Objectives and Results
In this document the results of the UUAs are briefly reviewed and a synthesis has been elaborated focussing on the UUAs results. This synthesis reflects the state of the work and the functionality and utility of the FDS, including valuable feedback from the users. The latter issues encompass information gains from the DSS, technical issues such as the integration of products into the Forest GIS environment or the benefits from the DSS. On the other hand critics and constraints are always accompanied as important items in each of the production steps. The overall performance or cooperation of SPs together with users was a major topic in the realization of the DSS. It can be stated at the end of EUFODOS that the feedback to these questions showed a very good ‘climate’ between SPs and users.
However, it can be stated, that after Phase 1 the service prototypes have been prepared and ready for the role out in Phase 2 of the project. In order to continue with the service development in Phase 2 small adaptations and modifications were applied to the prototypes according to the feedback from the user’s side. The main task in Phase 2 was altogether the transfer – that means proof of operationality – from the small testsites into the roll-out (wall-to-wall) implementation onto the entire testcases. For this purpose the SPs showed high flexibility in the realisation and in shifting the thematic areas where appropriate or needed. During Phase 2 the cooperation between SPs and users were even intensified, because the finalisation of the products required more and intensive training sessions. This approach proofed to be a very efficient working procedure
The analysis of the life cycle and sustainability comprise the fields of service definition and service quality, data management, processing lines, product delivery and its validation. The impact of the services is discussed as well as sustainability and recommendations. In this context the SPs were busy to reach high standards in order to fulfil their obligations from the SLA. The outcome of EUFODOS proofed that this goal has been achieved for almost all derived products in the testcases. The only aspect to be mentioned in this context is that the service for assessing storm damage has been implemented from the technical point of view, but could only be realized in two small events in Germany and Poland.
During the user workshop the users have reported on their experience within EUFODOS from working procedure to the evaluation of the products. The EUFODOS users can be regarded to be representative for a large user community and if their statements to make use of the services will be transferred to other users or user groups, then it can be expected that an uptake by many other users will follow. From the finalized UUAs and the discussions during the meetings it can be concluded that the DSS are ready for the market. This is an important aspect as it supports the sustainability of the DSS and enables the SPs to be competitive on the market with the EUFODOS services.
Potential Impact:
1.1 Economic DS Service Model
1.1.1 Objectives and Results
This WP focuses on the overall economic model for the FDS. The objective was to define a Road Map for future technology and business development for the FDS industry based on information on the market structure, cost-benefit analysis and business modelling for FDS Services. The objective was also to evaluate the economic potential of each service case developed and demonstrated by the Service Providers.
The main results of this WP are the market analysis, the cost analysis, the value chain and business models, and the EUFODOS road map.
The study of the European market for FDS Services shows that forest monitoring based on Earth Observation technology serves a multi-layer market. Firstly, information on forest situation, activities and damages are needed by authorities on a global, national and regional level. In addition the same information is needed also in operational forest management and in planning and performing of wood harvesting. On a higher level information is needed for all forest areas, while forest management and harvesting focus on a local level. Yet, the value of the information provided by FDS Services for the customers is, in general, higher the closer you get to operative activities. Storm or insect damaged areas can be seen as a special niche in the forest monitoring market where high values are at stake. A special niche, where high values are involved, is monitoring of protective forest in Alpine areas.
Cost information of each demonstrated FDS Service was collected by the Service Providers to improve the understanding of cost structure of the services. The main challenge reported is that cost of satellite data, which is the biggest cost factor, varies significantly from case to case depending on availability, type and quality of the data. In practice, costs for a case analysis can be calculated only as the data is known. Availability of Sentinel 2 data can improve the situation significantly for the Service Providers.
The FDS Services can provide the Users with benefits of high value. In special cases like monitoring of protective forests in Alpine areas, the services can reduce risks on human lives and major infrastructure. In all cases, the services provide economic benefits through cost saving or improved performance. The economic values of the services vary from 10-15 euro up to several hundred euro per square kilometre depending on the environment and the service type. The challenge for the Service Provider is to develop a business model suited to capture these values through win-win co-operation with the Users. In deliverable 520.4 a set of potential business models are described.
The EUFODOS road map identify trend and events, that are likely to impact the development in the FDS industry in the near (2014-1016) and medium to distant future (2017-2020). The demand for FDS Services is expected to be driven by increases in exceptional weather conditions like storm and draught, and related insect infestations. The growing concern about the global environment will simultaneously increase the demand for information about the state of forest. This will affect both environmental and energy policies. Essential for the development of commercial FDS Service activities is EU’s future activities in space, space industry and research. Increasing demand for monitoring and cost cutting will require increased efficiency among regional authorities responsible for forest monitoring and management. Increased use of biomass in energy industry is expected to increase demand and competition for wood.
On the supply side, the EUFODOS road map emphasizes trends supporting availability and accessibility to EO data. Increasing numbers of satellites and satellite operators will improve the situation. Special focus in a near future will be on Sentinel 2. Also improvements in data storage and distribution will improve the conditions for providing services to the Users. The use of radar technology will improve information collection and the speed and reliability of service provision. Increased computing capacity will be essential as more and more information will be requested.
1.2 Concept Paper
Utilising the new emerging and developing opportunities based on the European COPERNICUS land core services the FP7 EUFODOS project has developed new and innovative downstream services, which are focusing on degradation issues in forests such as damage caused by storms, fire, snow-break or insect infestations. The downstream services are addressing needs of forest authorities, forest owners, forest companies and other stakeholders on national to local scale all over Europe.
On this background the EUFODOS team is now in addition proposing a PAN-EUROPEAN FOREST SERVICE that provides information on damage in forests for all member states in a uniform approach as a new Core Service. This service will provide continuous information on forest status for national and European policies. For this purpose a new WP530 was created in order to promote this idea.
1.2.1 Objectives and results
The EUFODOS consortium has shown a strong focus on user needs and hence the services have been developed in a custom-tailored approach, which representatively reflects the need of regional users. The purpose of EUFODOS is the development of Forest Downstream Services (FDS) in order to assess the extent of damage from forest degradation such as storm events, forest fires, snow-breaks or insect infestations. Additionally services are implemented for the derivation of functional parameters in protective as well as in managed forests. In this context technology is used from space- and airborne sensor platforms, which allows data to be acquired at short time intervals and in a cost effective way.
Further information on the user requirements, specifications or service deliveries can be found in the EUFODOS deliverable documents which are listed on the http://www.eufodos.info/content/eufodos-reports ; e.g. D210.2 report on service architecture component or D410.2 user needs and standard dossier for all services.
The methods and production chains developed by EUFODOS for these downstream services can built the basis and be used for the proposed Pan European Service.
During the course of the project several activities have evolved for the pursuit of the Pan- European idea. These encompass meetings, workshops and promotion activities. During these meetings the consortium members had the opportunity to present the concept and to include the feedback and proposals into the document. Several organisations and stakeholders have been contacted and the concept was presented. They are EEA, EFI, REA and from the Commission DG Enterprise and Industry, DG Environment, DG Agriculture and Rural Development, JRC.
1.3 Co-ordination of User Platform
1.3.1 Objectives and results
In this WP frame conditions for the development and coordination of the EUFODOS User Platform were established. A User Executive Body was founded in the first user workshop in January 2011.
The deliverable D610.2 contains an overview of the multiple activities taken to coordinate and inform the EUFODOS users. This comprises for example the organisation and conduction of four user workshops, four teleconferences and e-mail coordination of the reviews concerning the User Requirement template and of the User Utility template, collection of user input for the EUFODOS Road Map, the edition and dissemination of sixteen Newsletters as well as other promotion and information activities.
1.4 Web Portal / SDI
1.4.1 Objectives and results
The Work Package 620 has been managed by GAF and started in January 2011 (Month 1). Within the first half of the project (phase 1), the EUFODOS website as well as a Map-Portal (beta) has been established. During Month 19 to 27, several improvements and updates have been implemented on both Website and Map-Portal. New Key words, relevant for the different Forest Downstream Services in different languages (English, German, Polish, Finnish, Italian, Bulgarian and French) have been introduced in Meta description. Further activities have been made to get more reference links to EUFODOS from other sites. Both activities improved a bit the raking of EUFODOS in search engines such as “Google”. Nevertheless, themes such as storms, tornado, large insect infestations or fire are always main topics for all the News Sites in the internet and provide a big competition about high ranks in search engines. Another action has been proposed by GAF and finally implemented by the UEB: the new EUFODOS entry in Wikipedia, which hopefully improves the project and service dissemination activities. The News and Events sides, the Newsletter section as well as the Featured Content have been updated continuously. The programming of the EUFODOS Map-Portal (INSPIRE compliant as described in D320.6) has been finalized during the reporting period. Downstream Service products, results and relevant Metadata of all EUFODOS Service Providers have been implemented within the Map-Portal. A continuous content update of the Map-Portal is foreseen also after the end of the project (~ 2 years).
1.5 User Training
1.5.1 Objectives and Results
The objective of this work package was to organise and implement the interface between production and the user. On-site presentations of the service to the user will be made; this includes activities such as awareness raising workshops/meeting and/or specific training if required by the user.
It was decided to have at least one training session for the user receiving a service and to have in addition workshops with the intention to interest further users for the services offered by the beneficiaries. Training and workshop principles have been prepared and communicated internally (D630.1 & D630.4).
Training Sessions and workshops had been held during EUFODOS, see Training Activities (chapter 5.2) for an complete overview.
1.6 Point of contact for potential users of the EUFODOS products
1.6.1 Objectives and results
In this WP frame conditions for a Point of Contact for interested and potential EUFODOS users were established.
D.640.1 gives an overview on the promotion ad dissemination activities. This comprises the compilation of a list of 392 potential users from 46 countries and 12 international organisations to facilitate the Service Providers to contact potential users in the neighbouring countries. The EUFODOS Newsletters were distributed to all listed persons, to inform them about the various EUFODOS services and products. Also a questionnaire for the first contact with potential users was prepared. Further on, multiple dissemination activities have been done like contributions to forest or EO related Newsletters and Websites, contributions about EUFODOS in Wikipedia, publications, poster, flyer and presentations about EUFODOS, its services and products.
List of Websites:
www.eufodos.info