Two separate expert working groups undertook programme design exercises, using examples that reflect the range of capacities to develop and undertake monitoring. They proposed processes and mechanisms for developing a monitoring programme that meet the needs of country or region with a) substantial knowledge of potential hazards and programmes for monitoring environmental effects of GM crops and b) limited knowledge of potential hazards and little experience in monitoring environmental effects of GM crops. Two Monitoring Programme Design Templates are presented below in Table 1a and Table 1b.
Table 1a: This example illustrates the systematic development of a programme of goal-setting, monitoring, analysis and assessment that is possible where potential hazards and their consequences are known, and environmental protection standards and policies are effective such that they enable monitoring goals to be refined to address the specific concerns of stakeholders. Some key recommendations following this analysis are given at the foot of the table.
| Programme design elements for regions with substantial prior knowledge | Points to be considered in programme formulation | Case example Monitoring programme for GM potato cultivation in the Netherlands |
| Identify responsible (lead) organization | The organization should have the trust of all stakeholders and should follow a transparent process for requiring, reviewing and using monitoring data | Monitoring was conducted on GM potatoes in the Netherlands. The responsible group was the “Commission on Genetic Modification (COGEM) who advise the Ministry of Housing, Spatial planning and the Environment. |
| Determine general societal concerns | A list of general societal concerns should be developed through broad consultation (e.g. direct and indirect ecological impacts, gene flow, impacts on traditional or protected farming systems) | The primary concerns were the potential for gene flow to non-GM potatoes and the presence and importance of antibiotic resistance as a marker |
| Determine trait specific concerns | Specific concerns related to the crop, the cropping system, and the trait or traits (e.g. persistence of transgene products in soil, cumulative effects) | Two trait-specific concerns were investigated prior to field-scale monitoring: frost tolerance and alkaloid content. Smaller experiments allowed for higher levels of statistical control prior to full field studies. The trait-specific concern in field monitoring was whether volunteers would occur in follow-on crops and if they could be controlled. This could not be fully investigated in smaller experiments. |
| Prioritize concerns | Some data may be of scientific interest, but may not play a significant role in decision-making (e.g. differences between GM and non-GM crops that are less than differences between conventional crop varieties) | The presence of volunteers in following crops was the primary concern. Changes in frost tolerance or alkaloid content were viewed as items of lesser importance in the monitoring studies because it was believed that these parameters had been effectively investigated using small plot studies conducted earlier. |
| Identify information gaps | Closely examine the data that are available to see what “data gaps” exist. Gaps may be filled by using data from the literature, from previous studies, or using modeling. Conceptual models are extremely effective tools for identifying gaps in knowledge or understanding. |
Two information gaps identified: 1) What was the rate of volunteers in typical potato culture, and 2) would the problem be significantly greater with GM potatoes? The first gap was filled using results from previous studies. Data were not available to fill the second gap |
| Clearly define question(s) | It is important to define the questions that need to be answered before a decision can be made. For example, can a hypothesis be tested for a specific crop-trait-geographic scenario, and will the results assist a management or regulatory decision? | The specific question could be defined as “Under commercial cultivation (flowering, harvesting, etc.), is the level of volunteer plants significantly different with GM potato than with typical potatoes?” |
| Determine actions to be taken to answer the question | The process of defining the question can also define the type of study that is most appropriate to answer the question. For example, if the question has to do with the potential variation in response, then multiple sites will be needed. However, if the question concerns variation over time, then multiple growing seasons may be needed. Some questions are better addressed with small-scale or semi-field studies | Smaller studies were used to determine potential changes in frost tolerance and alkaloid content. Then field monitoring using fields undergoing commercial cultivation was deemed appropriate to answer the question about the increased presence of volunteers at the commercial scale |
| Design, conduct, interpret, and communicate appropriate study(ies) | Careful study design is a critical step in obtaining data that are useful in making a regulatory decision. Study design should take into account the appropriate endpoints, robustness of sampling and interpretation techniques and the statistical power of the study |
Prior data were lacking. Sufficient resources were available to design and conduct a relatively large monitoring study using three potato varieties and 200 farmers. Potatoes were grown for one year and harvested. The fields were then rotated into grass (typical agricultural system) and the presence of volunteer potatoes was noted. Appropriate control fields were also included in the monitoring design. |
| Refine conceptual model and integrate data within the regulatory process | Once the study has been conducted and the data have been analyzed, the new information can be integrated within the conceptual model and/or used in crop management or a regulatory decision making process. This is the major test of the rigor and integrity of the previous steps. | The data compiled showed no significant differences between GM and non-GM varieties. When the results from the field monitoring studies were combined with the previous data on frost tolerance and alkaloid content, all of the original concerns were addressed. One variety had excessive flowering, two varieties had only slightly lower yield and one variety had a very low yield. All three had a kanamycin resistant marker (something that became a concern after the study was conducted). |
| Develop a basis for country/global/regional networking and communication | Communicate the results of monitoring so that others might make use of the information. It is important to note, however, that different countries may have different perceptions of risk. At present there are no mechanisms to help in regional communication of GM monitoring results |
Table 1b: Monitoring programme development is a greater challenge in cases where possible hazards are not clearly understood, the stakeholder community is not well defined, the level of protection afforded by environmental protection measures is low, and there is a lack of capacity and resources. The outline below examined the process from the perspective of a monitoring design template: the elements of the programme, points to be considered and the challenges of implementing the various elements in the context of herbicide tolerant lowland rice in Asia are addressed. Some key recommendations following this analysis are given at the foot of this table
| Programme design where there is limited information and experience | Elements for programme formulation | Points to be considered | Hypothetical example: monitoring programme for herbicide tolerant (HT) rice in a developing country in Asia. |
| Develop and state programme goals in consultation with stakeholders in the final outcome (e.g. farmers, stewards of local protected areas etc.) | Identify and engage stakeholders, recognizing that different skills tend to be found in different sectors. Define the ultimate goals of the monitoring programme, expressed in terms that stakeholders value. Develop consensus on precisely stated goals to enable effective monitoring design, and eventual follow up. |
Are the goals clear and simple enough to be addressable? If there are broader concerns, should the programme be nested within a larger process? Does the programme adhere to laws and relevant conventions? Has a fair and equitable selection programme been used to identify relevant stakeholders? |
Goals: To avoid weedy rice becoming more weedy because of gene flow and selection. To maintain the native gene pool of rice. To maintain the livelihoods of Asian farmers |
| Identify barriers to achieving goals | Identify all the practices, and stressors that may compromise the system. Identify the resource affected by each practice or stressor. This will aid the later identification of indicators. Summarize the characteristics of the above in terms of frequency, extent, magnitude, selectivity and variability. |
Competing interests and marketing forces could prevent consensus. Lack of success can result from failure to engage civil society: people with important expertise may be excluded from communication and access to resources. Poor communication between stakeholders limits goals setting, and engagement |
Weedy rice is already widespread in direct seeded areas, less so in transplanted areas. Good management practice is well understood, but not always practiced for various reasons. Marketing forces will influence the adoption of GM rice, and may not acknowledge risks. No obvious technical barrier to effective monitoring |
| Develop a simple, robust, conceptual model for the system based upon stakeholder and expert knowledge | Outline interconnections between system components, the strength and direction of links and the state of the system. Outline the scales at which processes operate and consider how the system ‘works’ with an emphasis on response to practices or stressors. What is acceptable variability and what constitutes a normal pattern? |
Engage all sources of knowledge from farmer, public, private and civil society sector. Need to ensure their participation throughout the programme. |
GM technology, with low adoption of good practice leads to HT gene flow into wild relatives. Herbicide resistance in weedy rice is selected by increased use of herbicides (which can happen with or without gene flow). Weedy rice densities can increase and production consequently decreases |
| Identify possible indicators that are connected to key elements of the conceptual model, and to the concerns of stakeholders | Make measurements that reflect agricultural and ecological processes that are sensitive to change across the range of GM crop release and provide information on the status of unmeasured resources. Temporal and spatial scales must be stated. |
Indicators may work, but must be able to be measured cost-effectively. Need provisions for entry and validation of data received from farmers and other stakeholders. |
Counts of weedy rice m-2 Yield loss. Seeding rate, kg/ha. Frequency of herbicide use. Need to establish sampling regime that may be undertaken by extension services, farmer groups, farm consultants etc. |
| Estimate the status and trends in the indicator, in comparison with control areas, baseline values before crop release or ideally both | Determine the required frequency and intensity of sampling effort to obtain the necessary level of statistical power. A successful outcome depends on a high level of inferential power in the comparisons that will be made |
The choice of reference site or condition is complicated where adoption is rapid or widespread. Reference points and baselines may be hard to identify if GM cropping becomes the norm |
Reference point – non-GM systems (may want to use sentinel plots/farms). Need to report variation in indicator responses, as well as mean values. Important to clearly visualize results and express in terms that have clear meaning to stakeholders |
| Determine trigger values for the selected indicators that lead to management action | Determine appropriate magnitude of effect size for a response, based on an understanding of spatial and temporal variation in response relative to baseline or reference condition. | The trigger value must be connected to an adverse effect on resources of concern to stakeholders. Intensively managed systems tend to become depleted and trigger values must take into account broader goals for sustainability, as well as the status and trends in the indicator in the reference site(s). Placing long -term societal goals for sustainability ahead of short-term, possibly unsustainable goals is a challenge and requires confidence building measures among stakeholders |
To be effective in early warning about serious hazards, triggers are needed that result in a change in farmer behavior in time to reverse adverse impacts. Need to address balance between long- and short-term costs and benefits. May ask farmers to make decisions that are not cost-effective or valued in the short term. |
| Link monitoring results to decision making through clarity, transparency, effective policy development and capacity building | List and evaluate all possible interpretations of indicator values, the likelihood of each being true and the societal values associated with each interpretation. This engages stakeholders and provides guidance in effective decision making | The experts recognized that there were few effective models for this process in the recent history of adopting new technologies in agriculture. Full stakeholder engagement however, is essential for adaptive and effective technology adoption. | Establish chains of multi-way communication that extend from local government to farmer, to researcher, educator, regulator and policy developer |
