Information system for monitoring environmental ...

Environ Sci Pollut Res (2010) 17:1479–1490 DOI 10.1007/s11356-010-0334-y

RESEARCH ARTICLE

Information system for monitoring environmental impacts of genetically modified organisms Hauke Reuter & Ulrike Middelhoff & Frieder Graef & Richard Verhoeven & Thomas Batz & Martin Weis & Gunther Schmidt & Winfried Schröder & Broder Breckling

Received: 30 October 2009 / Accepted: 4 April 2010 / Published online: 23 April 2010 # Springer-Verlag 2010

Abstract Background, aim and scope European legislation stipulates that genetically modified organisms (GMO) have to be monitored to identify potential adverse environmental effects. A wealth of different types of monitoring data from various sources including existing environmental monitoring programmes is expected to accumulate. This requires an information system to efficiently structure, process and evaluate the monitoring data. Methods A structure for an Information System for Monitoring GMO (ISMO) was developed by a multidisciplinary research team. It is based on the requirement to organise all

relevant information in a logical, readily accessible and functional manner. Results For the ISMO, we present a combination of three interrelated components: Firstly, an ISMO should comprise a knowledge database structured according to information related to the different scale levels of biological organisation relevant to GMO monitoring and scientific hypotheses on cause–effects which should be validated by monitoring data. Secondly, a monitoring database should be part of an ISMO containing GMO-specific monitoring data and metadata. This monitoring database should be linked with monitoring data from other monitoring programmes which

Responsible editor: Elena Maestri H. Reuter : U. Middelhoff : B. Breckling Centre for Environmental Research and Sustainable Technology, University of Bremen, Leobener Str., 28359 Bremen, Germany F. Graef Federal Agency for Nature Conservation (BfN), Konstantinstr. 110, 14191 Bonn, Germany F. Graef Leibniz Centre for Agricultural Landscape Research (ZALF), Eberswalder Str. 84, 15374 Müncheberg, Germany R. Verhoeven Elisabethstr. 70, 28219 Bremen, Germany M. Weis Institute for Phytomedicine (360), Department of Weed Science, University of Hohenheim, 70599 Stuttgart, Germany

T. Batz Fraunhofer Institute for Information and Dataprocessing, Fraunhoferstraße 1, 76131 Karlsruhe, Germany G. Schmidt : W. Schröder Chair of Landscape Ecology, University of Vechta, P.O. Box 1553, 49364 Vechta, Germany

Present Address: H. Reuter (*) Department of Ecological Modelling, Centre for Tropical Marine Ecology (ZMT), Fahrenheitstr. 6, 28359 Bremen, Germany e-mail: [email protected]

Present Address: U. Middelhoff Federal Office of Consumer Protection and Food Safety, P.O. Box 100214, 10562 Berlin, Germany

1480

are relevant for GMO-related questions. Thirdly, an ISMO should encompass a database covering administrative and procedural data. Neither national nor international approaches to an ISMO exist yet. Conclusions An ISMO as designed in this paper could support competent authorities in both the GMO notification process and in post-market monitoring. This includes evaluating the environmental risks of experimentally releasing GMO and placing them on the market, assessing monitoring plans and evaluating monitoring results. The ISMO should be implemented on both the national and international level, preferably combining different administrative scales. Harmonisation approaches towards GMO monitoring data are at an initial stage, but they are a precondition to coordinated GMO monitoring and to successfully implementing an ISMO. It is recommended to set up a legal basis and to agree on common strategies for the data coordination and harmonisation. Keywords Cause–effect-hypotheses . Environmental impact . Environmental risk assessment . Genetically modified organism . Monitoring . Information system

1 Background, aim and scope Environmental monitoring of genetically modified organisms (GMO) is legally required to identify the occurrence of adverse effects of GMO application on the environment after being placed on the market. The objective is to confirm or to rebut hypotheses regarding the occurrence of potential adverse effects in the Environmental Risk Assessment (ERA) which is an obligatory part of the notification procedure. As the ERA cannot anticipate all potential effects, monitoring also has the objective to detect unexpected, delayed, combinatory and cumulative adverse effects (EC 2001). Undesirable effects would be, e.g. changes in biodiversity or ecosystem functioning on different scales. GMO monitoring is a complex task because it deals with living modified organisms that reproduce, may persist, disperse and also evolve in natural and semi-natural environments. Not all detectable changes in the environment are caused by GMO. Therefore, it is reasonable to relate the monitoring to rational and explicit cause–effect chains that employ state-of-the-art ecological knowledge. The ongoing development of new GMO may lead to changing requirements for environmental monitoring, in particular those coping with potential combinatory effects. This requires inferences of monitoring results across larger spatial and temporal scales. The post-market monitoring plan for a specific transgenic event is an essential part of the notification procedure according to the Directive 2001/18 EC. The European Union (EU)-legislation distinguishes two GMO

Environ Sci Pollut Res (2010) 17:1479–1490

monitoring approaches, case-specific monitoring and general surveillance (EC 2001). Case-specific monitoring addresses possible adverse environmental effects which have been identified during the ERA but could not be assessed due to experimental limitations. Its purpose is also to evaluate any assumptions on possible risks detected in the ERA. General surveillance addresses unanticipated adverse effects that were not foreseen during the ERA (EFSA 2004). According to the EU-legislation, the notifier is responsible for the monitoring. The borderlines of case-specific monitoring and general surveillance cannot be fully distinguished, and relevant topics may relate to either case-specific monitoring or general surveillance. Such topics are, for instance (a) combinatory effects of several genetic modifications accumulating in individual plants of a crop species such as multiple resistances in oilseed rape (Beckie et al. 2003; Knispel et al. 2008), (b) effects of different Bt-toxins on susceptible butterfly populations (Losey et al. 1999; Dively et al. 2004; Lang and Vojtech 2006) or (c) long-term effects due to changes in farming practices (Graef 2009). Some of these adverse effects may have been anticipated during the ERA, but they cannot be entirely assessed. Moreover, they may go beyond the case of one specific GMO. Monitoring must therefore focus on the structural and functional relevant environmental compartments being exposed to GMOs. In this paper, we argue that GMO monitoring should be systematic and science-based in terms of ecosystems research (Fränzle et al. 2008). Therefore, it should apply a hypothesis-based approach which to some extent can be operationalised by involving ecological indicators. The term hypothesis-based is consistent with the precondition that monitoring should be operated on the level of scientific standards (EC 2002) because any scientific investigation is necessarily based on an explicitly stated hypothesis that is decided as the starting point of the investigation. The complexity of the task to indicate environmental changes on small as well as large scales makes it reasonable to link the GMO monitoring with national or international surveillance programmes (Graef et al. 2008). This will require a coordinating unit integrating large amounts of data resulting from different environmental surveillance networks. Furthermore, it requires adequate methodologies to detect GMO cultivation effects and other environmental changes by relating them to possible underlying causal structures. With an increasing number of GMO (AGBIOS Company 2009), many different types of monitoring data from various sources will be accumulated. Thus, an information and data management system is needed which helps structuring, processing and evaluating monitoring data. In order to efficiently access and analyse GMO monitoring data, we propose an Information System for MOnitoring GMO (ISMO) which must be able to integrate data from different sources and to analyse biogeographic

Environ Sci Pollut Res (2010) 17:1479–1490

variability. This paper focuses on the conceptual background as well as on the components and structure of such an information system.

2 Materials and methods—systematising information on GMO The structure of an ISMO was developed by a multidisciplinary research team in the framework of research projects funded by the German Federal Agency for Nature Conservation and by the Federal Ministry for Education and Research. The project teams elaborated issues related to ecology and risk assessment, the analysis of spatial data in environmental information systems, compiled available information collected from environmental surveillance networks and included technical expertise in the development of information systems. Expertise on the legal situation was included but is not within the scope of this article. Information relevant for GMO monitoring was retrieved from peer-reviewed literature and research reports and was organised to make it readily accessible, verifiable and easily storable in an information system. Suitability for usage in monitoring plans was assessed with a set of criteria including availability of information, indicators and degree of verification. The approach to structure and retrieve information already constitutes an important result of the ISMO-project because this process has to be partly repeated when new GMO or traits are assessed. Existing environmental surveillance networks and further information in Germany were analysed with respect to the information they can provide in the context of GMO monitoring. The information system was then designed according to the functional requirements resulting from the developed information and data structure. Furthermore, it bases on the requirements derived from the development of monitoring plans and the evaluation of monitoring data. To evaluate the quality and completeness of GMOrelated monitoring plans, we established a stepwise procedure of information processing that is connected by a feedback flow of information between the different steps. 1. Compilation of a scientific knowledge database to assess potential effects of specific GMO including direct and indirect and combinatory effects on different spatial–temporal scales. 2. Structuring of data using a hypothesis-based approach which is organised according to ecological integration levels and aiming at relating potential environmental effects to respective causes. 3. Deduction of observation targets such as checkpoints, priorities and safeguard measures.

1481

4. Combination of a knowledge-based approach to select relevant indicators for the implementation and evaluation of monitoring schemes and a data-based approach relying on the availability of monitoring data.

3 Results The stepwise procedure described above yielded a concept for an ISMO encompassing three components: a knowledge database, a monitoring database and a database covering administrative and procedural data (Fig. 1). The knowledge database is structured according to information related to the different levels of biological organisation relevant for monitoring, from the molecular and cellular context up to landscape and regional data. This database comprises information on scientific hypotheses regarding ecological effects of GMO which should be applied for designing monitoring plans and for evaluating GMO monitoring data. The monitoring database provides GMO-specific survey data which serve to validate the GMO-related hypotheses. The database on administrative issues covers the admission and notification-related information and other data relevant in this context. This overall structure reflects scientific and administrative requirements and aims at supporting competent authorities with their tasks in the notification process and in the post-market monitoring. This includes: (a) evaluating the environmental risks of experimentally releasing GMO and placing them on the market, (b) assessing monitoring plans, (c) the continuing evaluation of monitoring results after approval and (d) issuing legal restraints. 3.1 Compiling the scientific knowledge database The ISMO knowledge database provides information on the GMOs including scientific data and further information relevant to ecological issues dealing with GMO effects in the environment. It requires a linkage with the further data processing steps as it is designed to inform about the overall GMOs’ context. This database ought to include information comparable to the Summary Notification Information Format (SNIF) of the European Union (JRC 2009), which concentrates on specific data relating to the plant species and to the introduced or deleted genes (EC 2001, Appendix IV, A7). However, additional information with respect to GMO use and environmental implications should be integrated as well. Thus, finally, the knowledge database encompasses (for details see Table 1): 1. A general description of the GM plants including data from the approval process

1482

Environ Sci Pollut Res (2010) 17:1479–1490

Fig. 1 Organisation of the Information System for Monitoring GMO (ISMO). Three specific databases relate to distinct tasks to (1) compile existing knowledge and develop cause–effect-hypotheses, (2) organise

the monitoring and evaluate monitoring data and (3) administer the approval and post-approval monitoring processes

2. Data on inserted genetic sequences, detection and identification methods. This includes nucleotide sequences, their sources and their intended functions as well as known additional functions 3. Information on intended uses of the GMO including information on potential exposition areas and other specific information such as areas with target pest infestation 4. Information on environmental effects identified in the ERA and the results of the hypothesis analysis. This includes those aspects not addressed in the ERA such as long-term, large-scale and combinatory effects 5. Information on monitoring plans related to the original suggestions of the license holder as well as to checkpoints and plans derived from the information system

the ERA, but the results are relevant to the monitoring as well. Associating observed effects with the underlying causes is essential for any environmental monitoring and not specific for GMO monitoring. Monitoring the environmental effects of GM plants is necessarily based on the formulation of cause–effect-hypotheses in order to comply with the EU directives (EC 2001, 2003a, b). Detecting unexpected effects implicitly needs delimitation from those foreseen on the basis of existing knowledge and requires a systematic approach to avoid missing environmental issues that might be of immediate or later importance including specified interactions at all relevant levels of biological organisation. This systematic approach aims at reaching environmental protection goals and establishing a cause–effect relationship by applying specific hypotheses. The utility of a hypothesis-based approach was previously explained by Züghart and Breckling (2003), Breckling and Reuter (2006) and Snow et al. (2005). This approach uses available scientific knowledge on a specific GM plant and deduces a set of general cause– effect-hypotheses for different situations. The approach is highly dynamic: it can be extended to new knowledge and requirements by specifying and adding new hypotheses. Ecological systems are inherently complex, with changing components and a varying interaction structure that involves different interaction pathways. This makes the analysis of cause–effect relationships a special challenge which increases with investigated scale (EFSA 2004). We propose to structure the hypotheses according to the different biological and ecological organisation levels on which the analysed GMO are likely to have effects or on

The aspects 3–5 should be subject to a continuous update reflecting changes (a) during the application procedure, (b) of the hypotheses on potential environmental impacts, (c) of monitoring plans derived from the outcome of the information system itself or (d) because of new scientific information. 3.2 Structuring information according to ecological organisational levels The hypothesis-based approach uses physiological implications of the genetic transformation as a starting point to systematically derive cause–effect chains to survey potential relevant environmental effects. This is primarily part of

Environ Sci Pollut Res (2010) 17:1479–1490

1483

Table 1 Content topics to be collected in a GMO knowledge database Rubric

Specific informationa

1

General data on GM plant notifications submitted to the competent authorities and results of the notification procedure

2

Data on inserted transgenic sequences

3

Identification methods

4

Use of GMO

Registration number, date and duration of approval Trade names of products and GMO therein Recipient species Event Unique identifier Notifier Information for detecting the inserted nucleotide sequence Detailed map on genetic elements, coding and non-coding regions and references Donor organisms and used genetic elements, their original functions, eventual modifications of original sequences and the goal of their modifications Number of copies in target organism Information on transferred properties and known non-target changes in recipient organism Information on verification and identification methods as well as deposit of samples Intended use of the GMO Regional and/or national restrictions in Europe and worldwide Registered and used traits, as well as trait test reports

5

Environmental implications

6

Monitoring plans

a

Potential spatial distribution of commercial cultivation Specific information on infestation areas of target pest Seed cultivation areas (national and international) Climate information with respect to cultivation areas Information derived from the ERA of the application and on limited-scale GMO releases (EC 2003a, b) Hypotheses on environmental effects on specific levels of biological organisation (molecular and cellular, organismic, population, ecosystem, landscape and region) Monitoring plan of the applicant, contact data of responsible project leader, other links to the monitoring plan Overview on currently available monitoring reports, other links to reports List of checkpoints derived from the information system analysis

The specific information goes beyond the summary notification information format (SNIF) of the European Union

which effects have already been demonstrated. Hierarchy theory provides a coherent approach to analyse organisation levels in ecology (Allen and Starr 1982; Allen and Hoekstra 1991; Hölker and Breckling 2002). In the absence of relevant connections or interactions, ecological entities can be analysed in isolation from their context. In their presence, however, the resulting network can give rise to phenomena that are qualitatively different from the entities in isolation (emergent properties). Monitoring typically relates to the higher organisation levels like population, ecosystem and landscape. These levels define the spatio-temporal configuration required for the monitoring approach. We distinguish six relevant levels (Table 2), the first three of them (molecule, cell and organism) having properties that will mostly be investigated during the risk

analysis. They largely comprise characteristic plant properties and properties of the inserted transgene that trigger the environmental effects. The higher levels are of specific interest for hypotheses formulation and the subsequent development of the GMO monitoring approach (see 3.3 ff.). These hypotheses are testable and constitute the basis for structured evaluation and also for prioritisation criteria for checkpoints to be monitored. 3.3 Derivation of hypotheses and observation targets Table 2 lists example processes with their spatial and temporal dimension-related organisational levels. The first two steps refer mostly to compiling and structuring information. The knowledge base obtained for each GMO

1484

Environ Sci Pollut Res (2010) 17:1479–1490

Table 2 Overview of the basic structure and relevant processes and properties on different levels of biological and ecological organisation Organisation level

Relevant processes and GMO properties

Compilation of molecular and tissue-specific properties

Identification of the organism Stability of the transgenic properties and their phenotypic characteristics Metabolic characterisation Multiple transgenic properties and molecular interaction potential Performance of the GM plant (yield, pest infestation and further agronomic data) Pollen dispersal Necessary and specific management measures Observation during cultivation (farmer questionnaires) Seed purity Development of resistance in target organisms Changes in population size of relevant non-target organisms Changes in mass balances and element cycles Changes in soil mineral nutrient retention capacity

Compilation of the properties of plant stands (field level)

Compilation of population properties of target and non-target organisms Compilation of ecosystems properties

Compilation of properties on the landscape level

Compilation of effects on the regional levels

is used to derive specific sets of hypotheses. For each description of processes on the population, ecosystem and landscape level that can be related to environmental effects of GMO, we develop hypotheses which allow testing effects in a specific situation. In order to verify the hypotheses (to evaluate and enhance quality and completeness of GMO monitoring plans), they are used to identify observation targets, which involves introducing a corresponding check point for the monitoring. These can serve as a checklist for setting up the monitoring plan for a specific GMO with each check point being linked to a cause–effect chain. The derived check points are systematically prioritised according to safeguards, extent or probability of possible effects and may also consider a check point’s development. Prioritisation of checkpoints may also include results from previous monitoring thus reducing priority if no effects have been measured for long time intervals or increasing priority in the opposite case. For instance, a check point of minor priority can be implemented within

Enrichment and decomposition of substances Potential interactions with the abiotic environment Dispersal and persistence of transgene plants (e.g. via pollen and seeds) Dispersal and persistence of transgenes Abundance and population dynamics of hybridisation partners Large-scale development of target and non-target organisms Changes in species composition Effects of changes in agricultural practices Neighbourhood relationships between fields Regional field geometries and distributions Regional changes in crop rotation pattern Regional changes in agricultural structures Long-term changes in species composition of regions

a monitoring plan if suitable environmental monitoring programmes have already been set up or could be put in place with minor effort. 3.4 Assessment and evaluation of monitoring plans Our overall strategy is to integrate information about available data from existing environmental monitoring networks with the evaluation and development of monitoring plans at a very early stage. This constitutes a double strategy to either establish new or to evaluate existing monitoring programmes (Fig. 2): 1. The knowledge-guided approach is used to establish relevant indicators and rank their priority. This part of the strategy summarises the steps from chapter 3.1 to 3.3. 2. A data-guided approach based on availability of monitoring data and existing environmental monitoring networks features information on the status of realisa-

Environ Sci Pollut Res (2010) 17:1479–1490

Fig. 2 Double strategy using cause–effect-hypotheses and existing environmental monitoring systems to derive and prioritise checkpoints for GMO monitoring

tion of specific indicators. To some extent, priority indicators identified under strategy 1 may already be monitored. For already existing environmental monitoring data, it is only necessary to establish causal relations with the cultivation area of GM plants and the distribution of their transgenes in the environment. A further option is to use sampling strategies or sampling sites in existing environmental monitoring networks and supplement these with measurements of indicators relevant for the respective GMO. Utilising already established infrastructure and saving resources is a distinct advantage. Combining both approaches is an effective strategy to establish and evaluate GMO monitoring programmes and is flexible with respect to new plants and traits. The potential to use existing networks to evaluate and further prioritise indicators is a decisive reason to make their information accessible through ISMO. 3.5 Data availability and data flow The monitoring database (Fig. 1) helps to store and provide data from GMO monitoring and other relevant monitoring programmes and landscape information. One key task is to relate GMO cultivation effects to changes in the environment as well as to assess combinatory and long-term effects. The database has to provide access to different types of data including (a) data from environmental monitoring programmes that also contain indicators for potential environmental GMO effects, (b) GMO monitoring data provided by the notifiers, (c) location of GMO cultivation areas, differentiated for plants and traits, (d) background data such as on the environmental situation on eco-regions and biotopes, geo-data on land cover and agricultural statistics, climate data and (e) maps of protected areas, topological information and administrative boundaries. The data access should be facilitated by using interface standards and a standardised meta-data structure which

1485

stores relevant information on environmental monitoring networks as well as on locally stored GMO monitoring data sets. However, on both national levels and within the European Union data sets and data infrastructure schemes are highly heterogeneous, and standardisations for environmental and GMO data coordination are still in preparation (EC 2007; Graef et al. 2008). For an ISMO implementation, this situation makes it mandatory to use meta-data for data characterisation and to implement statistical algorithms to combine and analyse data with different qualities. For details, see the technical report (Reuter et al. 2007). The monitoring database also should provide a set of analysis and evaluation tools. This includes spatial analysis with classical GIS operations (e.g. overlay, buffering and geostatistics) as well as complex spatial analysis such as representativeness and neighbourhood analysis. The basic maps and the results of spatial statistics should be visualised within a GIS environment. ISMO should enable users to download monitoring data for detailed further analysis. 3.6 Implementation of the WebGIS GMO: the case study of Baden-Württemberg (Germany) As outlined in the ISMO concept, a Web-based geoinformation system (WebGIS) application can and should be linked with the monitoring database. Schröder et al. (2006b) implemented such a WebGIS GMO and integrated data from different environmental monitoring networks established in the Federal State of Baden-Württemberg (South-Western Germany) and evaluated their relevance for GMO monitoring and their representativity for the landscape coverage of Baden-Württemberg. The monitoring data and their meta-data had been collected in earlier research projects (Schröder et al. 2003, 2006a). They comprise information on localities and the measurement parameters as well as methods of 43 environmental monitoring networks with about 7,500 observation sites. Moreover, relevant maps describing environmental factors in Baden-Württemberg such as climate, soil texture, land coverage and eco-regions were compiled. Finally, coordinates and attributes of each observation site were stored in the GMO monitoring database. This enabled spatial and logical data queries by a user interface that also allows the download and visualisation of query results with templates and maps (Fig. 3). Geographical and statistical functions allow for analysing monitoring data by various GIS tools (clipping, intersection and buffering) and integrated analysis tools (descriptive statistics, frequency analysis and histogram plots). The user of this WebGIS, which is part of the monitoring database, requires neither additional software nor detailed knowledge on the GIS algorithms. Applying the approaches of Schröder et al. (2003) and Graef et al. (2005a, b), we tested whether the monitoring

1486

Environ Sci Pollut Res (2010) 17:1479–1490

Fig. 3 WebGIS GMO monitoring developed by Schröder et al. (2006b). The Web-GIS facilitates a selection of information according to relevance and representativeness criteria

sites in Baden-Württemberg can be used for GMO monitoring, in particular for general surveillance (EC 2001). In a first step, an analysis of representativeness was performed. It tested whether the sites of the BadenWürttemberg monitoring network cover each ecoregion (Schröder and Schmidt 2001) proportional to its particular area (Schröder et al. 2006a). For instance, if the landscape “Swabian Highlands” covers 20 % of the territory of Baden-Württemberg, then 20 % of the soil monitoring sites (eight of 40) should be located in this ecoregion. Due to the high interference with GMO monitoring methods and indicators, we used observation sites of the nationwide Ecological Area Survey (ÖFS; Middelhoff et al. 2005), which was primarily established to measure changes in floral and faunal biodiversity. By using the GMO monitoring WebGIS, we determined that a sub-sample of the ÖFS sites (n=29) had more (+5%) observation sites in the Swabian Highlands (ecoregion 26) than necessary considering proportionality. Considering the criterion ecoregion representativeness, the ÖFS monitoring sites are considered suitable to detect changes in biodiversity induced by GMO cultivation. For other potential effects on further integration levels (see Table 2) a similar evaluation of available databases would have to be performed. In addition to database operations, the WebGIS as a tool in the GMO monitoring context facilitates an overview and provides information in a form that can be used for integration, regionalisation and decision making. This makes it an important part of the ISMO, generating overviews on

existing environmental monitoring programmes and helping to economise and coordinate GMO monitoring activities. The WebGIS GMO was further refined and, meanwhile, encompasses GMO relevant data from the regional to the European scale (Kleppin et al. 2008).

4 Discussion 4.1 Technical aspects and implementation Key requirements for designing and implementing an ISMO are demonstrated above. In the following we emphasise further implications of implementing this ISMO structure. The technical implementation of the monitoring database is challenging. Since it should store monitoring data and manages interfaces to other environmental information systems, not all information will be physically available on a particular server. This requires remote access routines. The database must also provide tools for statistical and spatial evaluation. These tools can be used to evaluate and enhance monitoring plans and to collect monitoring results. Other important features are the keyword and metainformation search. Much of the information is accessible with no or limited restrictions, in particular (a) meta-data (descriptors to all contained data sets and stored results from monitoring programmes), (b) geo-referenced data from monitoring programmes and basic data including tools for data administration, (c) access pathways to

Environ Sci Pollut Res (2010) 17:1479–1490

externally stored data (from other monitoring programmes that are evaluated for GM effects, (d) meta-data facilitating access to measurement methods, (e) tools for data analysis and visualisation, (f) details of monitoring activities (e.g. indicators, site locations and monitoring plans) and (g) results and reports from GMO monitoring. Restricted access with accompanying security measures is required for some database components (confidential or raw data). The administrative database provides and structures the data relevant to the approval process. Administrative data require a structured access policy and management as some of the information must be kept confidential. Therefore, access restrictions must be in place. Such data may include (a) dossiers of the notifiers, (b) correspondence and information concerning current and finalized notification procedures, (c) dossier examination details, (d) conditions or restrictions for the placing on the market, use or handling of the products and (e) laboratory results and monitoring plans of the notification holders. The three major ISMO components (Fig. 1) should be interconnected, but conceptualised as independent units with regard to their implementation. A major constraint is the necessity to manage contents retrieved from a variety of different sources. Apart from continuously updating the GMO monitoring data, the information system primarily collects basic data that rarely changes (spatial geometries of protected areas, maps, etc.). For existing environmental monitoring programmes, the data can be kept on the respective external servers, with restricted access to confidential or raw data. The up-to-date datasets can be retrieved on demand from external systems. This solution makes it unnecessary to continuously check system data consistency. Full access to each external information system requires using local meta-data referring to them. This necessitates standardised interfaces, data formats and protocols for data exchange, making it useful to fully comply with international standards such as those proposed by the Open GIS Consortium ( Kresse and Fadaie 2004; Uhrich et al. 2008; Vögele et al. 2005). 4.2 Data coordination and harmonisation on the national and international level Data harmonisation is needed for setting common standards for GMO monitoring. Without harmonisation, data from different sources can hardly be managed and analysed in an integrative manner. This includes common methodologies (Fink et al. 2006), spatio-temporal data collection designs (Graef et al. 2005a, b), data and meta-data structures, their exchange formats and common quality criteria (Jones et al. 2006). Harmonisation, however, increases the explanatory power of data and may help extrapolating local monitoring results to larger areas and supports long-term studies.

1487

Collecting and storing GMO monitoring data from different sources thus demands coordination by trained staff using specific tools. For GMO monitoring, different spatial and temporal scales must be considered. The sites should be representative for the area of GMO cultivation because monitoring every exposure site is not feasible. Hence, standardising the monitoring design for specific spatial and temporal scales enhances overall representativeness and explanatory power (Graef et al. 2005a, b; Stein and Ettema 2003). Based on known variances of the measured indicators, the monitoring can be designed to be representative for large areas (Perry et al. 2003), considerably reducing the overall monitoring and analysis efforts. The result aspired is standardised data quality with known error sources and uncertainties as well as a feasible number of replications for statistical tests (Jones et al. 2006). Finally, standardised monitoring data can also be internationally shared for common and aggregated analyses, again reducing overall effort. Monitoring non-target effects of different GM crops may involve different indicator species. Such species may have a specific spatial distribution depending on ecological site requirements, and most methods are specific to selected organisms. Both the selection of indicator species (Andow and Hilbeck 2004) and the methods used to measure them, if harmonised, improve large-scale data comparability (Fink et al. 2006). Coordination and harmonisation of the notifiers’ GMO monitoring data in the Federal States in Germany and the EU is still in the early stages and mostly conceptual (VDI 2006; Graef et al. 2008). A harmonised industry approach exists only for farmer questionnaires that may constitute part of the GMO monitoring on farmland (EuropaBio 2009). The Infrastructure for Spatial Information in the European Community (EC 2007) is designed to also contain standardised GMO monitoring data to some extent. To collect data in a harmonised way, the infrastructure scheme European Environment Information and Observation NETwork (Eionet 2009) has been developed which is also relevant to GMO monitoring. On the global level, this task is considered under the Cartagena Biosafety Protocol (CBD 2009), but concepts for GMO monitoring data coordination and harmonisation are even less advanced than on the EU level. Currently, there is still little awareness about the relevance of GMO monitoring data or why it should be harmonised in the EU and beyond. Several environmental monitoring programmes are in effect at both the national and EU level, which can provide a valuable information source, especially as baseline or reference data for interpreting GMO monitoring results (Graef et al. 2008). However, most national programmes are very heterogeneous with regard to the date types and availability. None of the programmes established for other purposes

1488

would currently satisfy the requirements of GMO monitoring. Programmes on the EU level having a potential applicability to GMO monitoring include (a) the Habitats Directive on the conservation of natural habitats and of wild fauna and flora (EEC 1992), (b) the Birds Directive on the conservation of wild birds (EEC 1979) and (c) the Water Framework Directive (EC 2000). Some would need only minor adaptations to cover aspects of GMO monitoring. According to Graef et al. (2008), the process of data coordination under the Water Framework Directive and the Habitats Directive is a good example of how monitoring activities at the national level could be integrated within a monitoring design covering several or all EU member states.

5 Conclusions Introduction and use of GMO in agriculture imply new regulatory tasks. The EU regulations require that state-ofthe-art measures be taken to detect and avoid potential adverse effects. GMO monitoring is one of them. The efficiency and sustainability of the agricultural sector and the surrounding ecosystems is of highest importance and deserves administrative priority. Agriculture operates within a complex ecological context, and an increasing number of GMO have to be surveyed. This calls for collecting existing information using a structure that allows to acquire a coherent overview and supports decision making. Information and knowledge gaps have to be detected and closed. An ISMO is the adequate tool to structure, handle and evaluate GMO-related information and data. Our proposition to structure the content of such an ISMO was developed in an inductive way, transforming requirements into steps to be taken. The advantage of this approach is its flexibility in integrating new information. It helps to document the appropriateness of decisions and provides a basis for updates or corrections if necessary. Furthermore, it increases efficiency by facilitating the re-use of existing approaches in new contexts. Environmental information management is required on different administrative levels, linking the regional, national and transnational level. This aspect can serve as a guideline for the further development of the ISMO concept and application. In other fields like the exchange of weather data, highly developed protocols for data access are already established. Progress needs to be made in structuring and facilitating the free flow of GMO monitoring data to detect potential environmental effects and support decision making. No information systems for monitoring GMO exist today. A combination of national and international ISMOs could utilise existing national environmental networks with the option to produce results applicable to large scales.

Environ Sci Pollut Res (2010) 17:1479–1490

6 Recommendations Currently, there is no legal basis to regulate the coordination and harmonisation of GMO monitoring data in and between EU member states and on a global level. Some regulations have implications for general information or data exchange. These include the creation of national registers for GMO cultivation (Directive 2001/18/EC), for the structure of results from deliberate releases into the environment (Decision 2003/701/EC) and in the context of transboundary movements of GMOs (Regulation No 1946/ 2003). None refer to GMO monitoring data per se. A legal basis is thus required, and there should be a common strategy and time schedule for data coordination and harmonisation. Once legally established, a technical framework will still depend on varying national conditions and take several years for adaptation and implementation. To accelerate this process, national solutions should be advocated in the short term. Ultimately—and in a parallel approach—the goal must be an international GMO monitoring information system, which should include the elaborated subunits and functionalities We emphasise to implement a system with a strictly science-based approach applying hypotheses on ecological organisation levels to design and evaluate monitoring plans and to analyse monitoring data. Acknowledgements Funding by the German Federal Agency for Nature Conservation under Grant No 80464030 is gratefully acknowledged. Further support came from the Federal Ministry for Education and Research under grant No FKZ: 0312637A and 07VPS14A, respectively. Many thanks also to M. Stachowitsch for improving the English.

References AGBIOS Company (2009) GM database, information on GM approved products. Available at: http://www.agbios.com/dbase. php?action=ShowForm Allen TFH, Starr TB (1982) Hierarchy–perspectives for ecological complexity. Chicago University Press, New York Allen TFH, Hoekstra TW (1991) Role of heterogeneity in scaling ecological systems under analysis. In: Kolasa J, Pickett STA (eds) Ecological heterogeneity. Ecological studies no. 86. Springer Verlag, Berlin, pp 47–68 Andow DA, Hilbeck A (2004) Science-based risk assessment for nontarget effects of transgenic crops. Bioscience 54:637–649 Beckie HJ, Warwick SI, Nair H, Seguin-Swartz G (2003) Gene flow in commercial fields of herbicide-resistant canola (Brassica napus). Ecol Appl 13:1276–1294 Breckling B, Reuter H (2006) General surveillance of genetically modified organisms—the importance of expected and unexpected environmental effects. J Consum Protect Food Safety 1(Supplement 1):72–74 CBD (2009) The cartagena protocol on biosafety to the convention on biological diversity. Available at: http://bch.cbd.int/ protocol/

Environ Sci Pollut Res (2010) 17:1479–1490 Dively GP, Rose R, Sears MK, Hellmich RL, Stanley-Horn DE, Calvin DD, Russo JM, Anderson PL (2004) Effects on monarch butterfly larvae (Lepidoptera: Danaidae) after continuous exposure to Cry1ab-expressing corn during anthesis. Environ Entomol 33:1116–1125 EC (2000) Directive 2000/60/EC of the European Parliament and of the Council of 23 October 2000 establishing a framework for the community action in the field of water policy EC (2001) Directive 2001/18/EC of the European Parliament and of the Council of 12 March 2001 on the deliberate release into the environment of genetically modified organisms and repealing Council Directive 90/220/EEC EC (2002) Council decision of 3 October 2002 establishing guidance notes supplementing Annex VII to Directive 2001/18/EC of the European Parliament and of the Council on the deliberate release into the environment of genetically modified organisms and repealing. Council Directive 90/220/ EEC:1–10 EC (2003) Commission Decision 2003/701/EC establishing pursuant to Directive 2001/18/EC of the European Parliament and of the Council a format for presenting the results of the deliberate release into the environment of genetically modified higher plants for purposes other than placing on the market EC (2003) Regulation (EC) No 1946/2003 of the European Parliament and of the Council of 15 July 2003 on transboundary movements of genetically modified organisms EC (2007) Directive 2007/2/EC of the European Parliament and of the Council of 14 March 2007 establishing an Infrastructure for Spatial Information in the European Community (INSPIRE) EEC (1979) Council Directive 79/409/EEC of 2 April 1979 on the conservation of wild birds EEC (1992) Council Directive 92/43/EEC of 21 May 1992 on the conservation of natural habitats and of wild fauna and flora EFSA (2004) Guidance document of the Scientific Panel on Genetically Modified Organisms for the risk assessment of genetically modified plants and derived food and feed, the EFSA Journal 99, 1–94 (updated July 2005) Eionet (2009) Environment Information and Observation NETwork. Available at: http://www.eionet.europa.eu EuropaBio (2009) Available at: http://www.europabio.org/positions/ GBE/PP_090204_Farmer_Questionnaire.pdf Fink M, Seitz H, Beismann H (2006) Concepts for general surveillance: VDI proposals. Standardisation and harmonisation in the field of GMO monitoring. J Consum Protect Food Safety 1 (Supplement 1):11–14 Fränzle O, Kappen L, Blume H-P, Dierssen K (eds) (2008) Ecosystem organization of a complex landscape. Ecological Studies 202. Springer, Berlin Graef F (2009) Agro-environmental effects due to altered cultivation practices with genetically modified herbicide-tolerant oilseed rape and implications for monitoring. A review. Agronom Sustainable Dev 29:31–42 Graef F, Züghart W, Hommel B, Heinrich U, Stachow U, Werner A (2005a) Methodological scheme for designing the monitoring of genetically modified crops at the regional scale. Environ Monit Assess 111:1–26 Graef F, Schmidt G, Schröder W, Stachow U (2005b) Determining ecoregions for environmental and GMO monitoring networks. Environ Monit Assess 108:189–203 Graef F, De Schrijver A, Murray B (2008) GMO monitoring data coordination and harmonisation at the EU level - Outcomes of the European Commission Working Group on Guidance Notes supplementing Annex VII of Directive 2001/18/EC. J Consum Protect Food Safety 3(Supplement 2):17–20 Hölker F, Breckling B (2002) Concepts of scales, hierarchies and emergent properties in ecological models. In: Hölker F (ed)

1489 Scales, hierarchies and emergent properties in ecological models. Peter Lang Verlag, Frankfurt, pp 7–27 Jones MB, Schildhauer MP, Reichman OJ, Bowers S (2006) The new bioinformatics: integrating ecological data from the gene to the biosphere. Ann Rev Ecol Evol Syst 37:519–544 JRC (2009) GMOinfo: placing on the market of GMOs as or in products. Available at: http://gmoinfo.jrc.ec.europa.eu/ gmc_browse.aspx?DossClass=0 Kleppin L, Aden C, Schmidt G, Schröder W (2008) Monitoring of genetically modified maize cultivation by means of WebGIS and Google Maps. In: Car A, Griesebner G, Strobl J (eds) Geospatial crossroads @ GI_Forum ’08: Proceedings of the Geoinformatics Forum Salzburg. Wichmann, Heidelberg, pp 170–179 Knispel AL, McLachlan SM, Van Acker RC, Friesen LF (2008) Gene flow and multiple herbicide resistance in escaped canola populations. Weed Sci 56:72–80 Kresse W, Fadaie K (2004) ISO standards for geographic information. Springer, Berlin Heidelberg Lang A, Vojtech E (2006) The effects of pollen consumption of transgenic Bt maize on the common swallowtail, Papilio machaon L. (Lepidoptera, Papilionidae). Basic Appl Ecol 7:296–306 Losey JE, Rayor LS, Carter ME (1999) Transgenic pollen harm monarch larvae. Nature 399:214 Middelhoff U, Hildebrandt J, Breckling B (2005) Die Ökologische Flächenstichprobe als Instrument eines GVO-Monitorings. BfNSkripten 172. 95p. Available at: http://www.bfn.de/fileadmin/ MDB/documents/service/skript172.pdf Perry JN, Rothery P, Clark SJ, Heard MS, Hawes C (2003) Design, analysis and statistical power of the farm-scale evaluations of genetically modified herbicide-tolerant crops. J Appl Ecol 40:17– 31 Reuter H, Breckling B, Middelhoff U, Verhoeven H, Filser J, Batz T, Bonn G, Weis M, Schmidt G, Schröder W, Pappenberger R, Kreft I, Bachhuber I, Carstens A, Winter G, Rösing J (2007) ISMO— Informationssystem für ein Monitoring gentechnisch veränderter Organismen. BfN, Bonn, BfN Skripten 125 Schröder W, Schmidt G (2001) Defining ecoregions as framework for the assessment of ecological monitoring networks in Germany by means of GIS and classification and regression trees (CART). Gate EHS 2001:1–9 Schröder W, Schmidt G, Pesch R (2003) Harmonization of environmental monitoring. Tools for examination of methodical comparability and spatial representativity. In: Gate to EHS 2001, July 2003, pp. 1–13 Schröder W, Schmidt G, Zipperle J (2006a) Geodaten, Messdaten und Analyseabläufe zur Messflächenauswahl bei unterschiedlichen Skalen. Überlegungen zum GVO-Messnetz in BadenWürttemberg. In: BfN-Skripten 189, S. 29–41 Schröder W, Weis M, Schmidt G (2006b) Flächenbezogene Aspekte des Informationssystems GVO-Monitoring. Grafische Konzeption, Entwicklung und Erprobung raumbezogener Dokumentations- und Analysemethoden.—Abschlussbericht des Teilvorhabens 3 in dem Verbundprojekt ´Informationssystem für ein Monitoring gentechnisch veränderter Organismen (ISMO)´. FKZ 804 67 030 im Umweltforschungsplan des Bundesministers für Umwelt, Naturschutz und Reaktorsicherheit. Im Auftrag des Bundesamtes für Naturschutz, Bonn Snow AA, Andow DA, Gepts P, Hallerman EM, Power A, Tiedje JM, Wolfenbarger LL (2005) Genetically engineered organisms and the environment: current status and recommendations. Ecol Appl 15:377–404 Stein A, Ettema C (2003) An overview of spatial sampling procedures and experimental design of spatial studies for ecosystem comparisons. Agric Ecosyst Environ 94:31–47 Uhrich S, Kruse F, Klenke M, Giffei C, Peters S (2008) The information technology of the German Environmental Informa-

1490 tion Portal PortalU® as tool for the European Level? EnviroInfo 2008, 22th International Conference on Informatics for Environmental Protection, Lueneburg VDI (2006) VDI Guidelines: monitoring the ecological effects of genetically modified organisms. Genetically modified plants. Basic principles and strategies. VDI 4330(Part 1):1– 16 Vögele T, Kruse F, Klenke M (2005) Flexible und modulare Systemarchitektur für gein 2.0., Koordinierungsstelle UDK/

Environ Sci Pollut Res (2010) 17:1479–1490 GEIN Hannover. Available at: http://www.udk-gein.de/publikat/ 2005/udb/UDB05_Voegele.ppt Züghart W, Breckling B (2003) Konzeptionelle Entwicklung eines Monitoring von Umweltwirkungen transgener Kulturpflanzen. UBA-Texte 50/03, Berlin. Available at: http://www.bfn.de/ fileadmin/MDB/documents/UBA-Texte_50-03-Teil-1_Konzept_ transgen_Kulturpfl.pdf http://www.bfn.de/fileadmin/MDB/ documents/UBA-Texte_50-03-Teil-2_Konzept_transgen_ Kulturpfl.pdf