Seventeenth Session

Rome, 31 March-4 April 2003

Agri-Environmental Information and Decision Support Tools for Sustainable Development

Item 8 of the Provisional Agenda

Table of Contents

I. Introduction

1. Environmentally related decisions are taken in the agricultural sector every day for the purpose of long-term strategic policy formulation, on-farm management, and for rapid assessment and management of emergency operations. Good decision-making requires reliable data and analytical tools that vary both in scale (in time and space) and in scope (policy to technical) as they are applied to a wide spectrum of issues ranging from food security, to plant protection, to management of natural resources. Over the last decades FAO has developed, applied and disseminated a wide range of information and decision support tools in support of its programme activities.

2. Agricultural decision-making has been undergoing rapid change over the last decade in response to greater emphasis on sustainable agriculture and rural development, the emergence of multi-lateral environmental agreements, and other development priorities such as embodied in the World Food Summit Plan of Action to reduce world hunger. There has also been progress in information technology that has led to availability, analysis and communication of data in ways that had not previously been anticipated. Agri-environmental information and decision support tools for sustainable development therefore need to be re-examined with a view to further developing their potential and improving their uptake and use in Member Nations.

3. This paper discusses ways in which agri-environmental information and decision support tools might become more widely used in the field of agriculture. The topic is covered in three parts: (1) overview of key technical, policy and institutional issues related to agricultural environmental information and decision support tools; (2) constraints and gaps, and (3) areas where FAO seeks COAG’s guidance for further action.

II. Overview of Key Issues

4. A range of issues must be addressed to enhance the use of environmental data in agriculture. Many of the issues were recognised in the Strategic Framework for FAO 2000-2015 (in particular element E) and several initiatives have been taken to tackle conceptual issues, methods, analytical tools, and their application to programme areas such as biodiversity, climate change, disaster prevention and response, land and water, gender, integrated production systems and organic agriculture.

5. The Organization is addressing some issues through two related Priority Areas for Interdisciplinary Action (PAIAs). The PAIA “Definitions, Norms, Methodologies and Quality of Information" focuses on consistency and quality of the basic data, while "Spatial Information Management and Decision Support Tools" facilitates access to harmonised spatial information.

6. Three other recent initiatives offer the prospect for improved agri-environmental data and information access: the modernisation of the FAO database of agricultural statistics (FAOSTAT2); the development of GeoNetwork (an interactive window to maps and related information); and, RANET (Radio and InterNet for the communication of hydro-meteorological and climate related information). Annex I provides an overview of these and related datasets and decision support tools.

7. Three types of issues are covered in the discussion on the use of environmental data in agricultural decision-making:


8. Decision support tools are used to evaluate decision alternatives so that users can assess the effects of different choices and arrive at the “best” decision, given the circumstances. Such tools are one component of the decision-making process. The explicit or implicit comparison of the impact of different decisions is termed “simulation” or “modelling”. The two ends of the “decision chain” (from data collection through analysis to assess the impacts of decisions to decision proper) often belong to different technical areas covering the spectrum of geography, environment, agriculture and socio-economics. This results in some operational implications: decision support tools should keep in view the need for conformity with the decision-making1 process, in particular the requirements of the decision-makers in terms of types of data and spatial resolution (villages, towns, districts, provinces, regions) as well as the dissemination of products to beneficiaries, including farm level.

9. As every piece of information is characterised by its unique time and spatial scale (point or village, district, region, watershed, agro-ecological zone, etc.), accuracy, “age” and accessibility, there is always some degree of incompatibility between data types. Georeferenced data (i.e. data with geographic coordinates attached) can easily be converted to “statistics” (data usually identified only by the name of an administrative unit), while the reverse operation is more difficult since it requires data disaggregation tools (models) and may lead to products of uncertain quality.

10. In general, local decisions are closer to the “raw” data (i.e. direct in situ observations of agriculture and environment) while decisions taken at the national or international level use more synthetic and value-added variables (or indicators). Indicators as well as the methods used to assess the impact of various decisions (i.e. the simulation methods) should all be seen as being components of decision support.


11. Deciding which data to collect and the frequency of collection usually takes place through consultation at the national and international levels. Compromises are made between cost, accuracy, planned use and sampling frequency. There is always a risk that specifications for more costly information products, especially global ones, will be beyond the reach of the least developed countries and may not address their priority needs. FAO tends to be quite sensitive to this issue and normally seeks a balanced approach.

12. The costs of decision-making derive from data acquisition, analysis and the dissemination of results.

13. Many data “owners” at the national and international levels tend to charge a fee for data access. National data policies vary enormously. Some countries make all of their data freely available through easily accessible means, including the Internet. Others impose high levels of confidentiality even on common and non-sensitive data such as rainfall. Often, restrictive policies of data access are used to recover at least part of the investment made to collect the data, rather than for confidentiality. In some cases, tight data access policies are used to discourage casual and unqualified data requests. There is a special need for data to be easily accessible in work with a humanitarian component, such as food security and emergency operations, as well as for scientific uses. It is FAO policy to make its official data widely available and easily accessible.


14. At the international level, global environmental data are usually collected and pre-processed by countries owning the technology (primarily satellites). Beyond the policy aspects mentioned above, data dissemination to all stakeholders requires the coordination of international databases, their maintenance, and the definition of access protocols, in addition to covering the operating costs. FAO’s leadership of the Global Terrestrial Observing System (GTOS) is an example of a programme that seeks to build global databases related to climate change and to biodiversity based on common measurement methodologies and data access protocols.

15. Various approaches are used to share data or replicate datasets on a regional basis. FAO has played a key role in using and building awareness of the usefulness of satellite data and tools for food security, crop monitoring and forecasting and early warning. As a member of the Committee of Earth Observing Satellites (CEOS) FAO plays an important role in conveying to these organizations the specifications and processing requirements for using their data in agricultural development activities.

16. Particular difficulties are encountered when data are collected, analysed and disseminated in countries with poor communication infrastructures. For example, Response Farming2, a technique that allows a farmer to adjust practices to the prevailing environmental conditions, can have limited effectiveness when data transmission facilities are weak.

III. Constraints and Gaps

17. The use of environmental data for improved agricultural decision-making is constrained in five key areas.


18. Data of good quality can be “unavailable” in Member Nations for a variety of reasons: they were never collected, they cannot be located easily, their format and presentation are inadequate, their cost is too high, their volume is too large for the processing capacity available locally, or the national or international owner of the data imposes restrictions to their distribution.

19. Furthermore, data must be of sufficient quality to be put to the intended use, particularly if they are to be combined from official or unofficial sources.


20. Different types and sources of data must be coherent. For instance, agricultural production data must be compatible with climate and soil conditions, with available land and the level of technology; similarly, there are links between population, food requirements and water use. Data coherence is an essential factor of the integration of environmental and natural resource considerations into overall planning at the national and international levels.

21. To be used easily, data must have some degree of inter-compatibility or harmonisation, an issue which includes definitions, formats (tables, maps, graphs) and media. Furthermore, the scale and level of detail of the data must be appropriate to the decision-making needs. National applications often require much more detailed and accurate data than do regional or global studies.

22. Data harmonization is particularly relevant for decision-making that involves bringing together data sets from a variety of providers and involving diverse technical, institutional and policy actors. At the national level, data and tools should be harmonized across government levels and sectors if they are to enter the same decision-making process.

23. In FAO, beyond the statistical data in FAOSTAT, more work is needed to standardise environmental, agricultural and socio-economic data. Tools such as GeoNetwork are being used to standardize both metadata and data collection, and to overcome differences in data presentation and format. This will allow for more integrated use of data and information from different sources. Additional sustained efforts are needed in order to achieve greater compatibility across data sets.

24. Metadata, information about the “who, what, where and when” which lies behind data and how it was collected, is an essential tool for ensuring that data is used properly and leads to good decision-making. National and international lists of data sources, including information about their formats, limitations, and the way to access them are now gradually becoming available, as are those of the International Standards Organization (ISO) that establish a common baseline for metadata information to be registered.


25. Some constraints are associated with technology for data collection and transmission, and for data storage and processing. Usually the difficulties are associated with low levels of technological development, particularly the communications infrastructure.

26. A barrier in many developing countries is keeping up with the rapid development of new technologies for data collection, processing and analysis. In particular, data from recent earth resources satellites are not easily accessible if continuous and sometimes costly changes are not made to national data acquisition and processing infrastructures. Data processing and analytical techniques are also evolving rapidly in areas like geographic information systems, satellite data processing and modelling.

27. FAO has developed good technical capacities and services to overcome these difficulties but often lacks the resources to develop the full potential of information for the benefit of Member Nations. Dedicated efforts are needed, especially in the lower income countries, to improve their access to and use of available technologies to obtain and handle environmental information for improved decision-making at local, sub-national, national, regional and global levels.


28. The most serious constraints are in the area of human resources. Many developing countries lack the personnel to collect, analyze and manage data from national and international sources for use in decision support. This is particularly true where fast evolving scientific methods or technologies are involved. Training and refresher training are required, as well as initiatives to raise the awareness of decision-makers about the potential of new tools and strengthening capacity in data analysis.


29. The least developed countries are rarely fully involved in data acquisition, processing, storage, and exchange and could benefit from formal and informal networks, particularly in cases where data about their country are used in developing global datasets related to biodiversity, climate change or desertification.

30. Global, regional and local data can be highly complementary but more effort is needed in this area. While international data sources, especially from satellites, can provide spatially detailed data, many of them can be converted into meaningful data for decisions at the national and international levels only if they are combined with local ground observations on soil, climate, weather, vegetation etc. This is clearly an area where national and international partnerships are of mutual benefit. FAO has a two-fold role to play: looking after technical aspects such as access to data and data compatibility, and in ensuring, as an honest broker, that the partnerships are fair and equitable to all participants.

31. Many large datasets are best maintained centrally to reduce maintenance and development costs. Access to and use of these datasets can be achieved only through bi- and multi-lateral partnerships.

IV. Areas Where FAO Seeks COAG Guidance

32. The above discussion highlights the significant role environmental information systems and decision support tools play in FAO’s programme of work. The Organization has a long standing presence in this field and actively contributes to its further development and application in the agriculture sector. It is expected that FAO will continue developing tools to analyse environmental impacts, to assess natural resource conditions contributing to food insecurity and poverty, and to assist in agricultural production planning and management decision-making.

33. In the light of new and emerging data requirements and technological tools that allow more rapid and complete data analysis and presentation, numerous countries are requesting assistance, in particular as regards their reporting requirements for multi-lateral environmental agreements. Current challenges call for improving the awareness of data types and technologies, and for closer collaboration both within FAO and at the inter-agency level, to ensure compatibility of data formats and the coherence of data across different sectors.

34. Some data requirements arise internally from FAO regular programme activities to promote sustainable food security through initiatives such as the Food Insecurity and Vulnerability Information and Mapping System (FIVIMS), the Global Information and Early Warning System on Food and Agriculture (GIEWS), the internet-based Geographic Mapping initiative (GeoNetwork), land and water databases and statistics, and use of digital radio and internet for communication (RANET). Other data requirements arise from external sources such as multi-lateral environmental agreements, the Millennium Development Goals, the Millennium Ecosystem Assessment, disaster and emergency response, and sustainable development in general.

35. The guidance of COAG is sought on the following issues that have implications for future FAO activities in this field:

Annex: FAO data sets and decision support tools

i. This annex describes FAO datasets and decision support tools. Special attention is given to three initiatives: the “new” FAOSTAT (under datasets) as well as GeoNetwork and RANET (under Decision Support Tools).


ii. The global nature of many FAO data sets is one of the assets of the Organization. The FAO web site lists more than 1000 references to FAO data sets of various kinds, including some lists of databases. Many qualify as agricultural environmental decision support tools both at the global scale (most data sets) and for regional work (SADC, Horn of Africa, Bangladesh, Pacific Animal Health Information System, etc.) Some databases were assembled from a variety of external sources and therefore may pose complex ownership problems. The main datasets fall into three categories.

iii. Geo-referenced databases:

iv. Geo-referenced data are processed by Geographic Information Systems (GIS). The technology for such analysis has advanced rapidly during the past five years and it is now possible to have extremely powerful GIS systems run on desktop personal computers. As such, GIS software is a fundamental decision support tool for environment and agriculture analysis. An inventory of geo-referenced data sets was prepared under the GIS Policy Group in 1998 and is now being updated in the ambit of GeoNetwork.

v. Mostly non-geo-referenced:3

vi. Reference material and metadatabases. This covers essentially reference information of a textual nature, including bibliographic references, information on institutions and legislation, utilities (conversion factors), list of experts, etc.

vii. It is to be noted that the three categories above are often difficult to link for purposes of inter-comparison. More attention is needed to the development of tools for converting data to common standards, and aggregating and disaggregating it. Systematic registration of metadata can go far in enabling users to understand the strengths and limitations of data sets, whether spatial or tabular.


viii. Decision support tools mostly take the form of computer software. A number of tools are available for the agriculture sector and its numerous interactions with the environment. Many of them are easily accessible on the FAO website, including:

ix. GIS software is an important decision support tool for agri-environmental analysis. GIS applications are used across all of the FAO technical departments, in areas such as the Forest Resources Assessment (FRA), in fisheries resources (FIGIS), climate and soil resources. FAO has developed numerous applications, including WinDisp, KIMS, and GeoNetwork, which make extensive use of GIS technology and are exploited both within and outside of FAO.

x. WinDisp is a package jointly developed by FAO (ESCG/Global Information and Early Warning System and the Environment and Natural Resources Service SDRN), the United States Agency for International Development (USAID), the Southern African Development Community (SADC), the US Forest Service (USFS) and the US Geological Survey (USGS). WinDisp was originally meant to be a simple tool to analyse satellite imagery for early warning purposes. It gradually evolved into a multipurpose tool for many food security early warning and crop monitoring and forecasting applications.

xi. The Key Indicators Mapping System (KIMS) was developed by the Information Systems Service (AFIS) within the World Agriculture Information Centre (WAICENT) framework as a contribution to FIVIMS. KIMS is a user-friendly software, accessible to non-GIS specialists and developed with the specific purpose of collecting, mapping and disseminating food insecurity and vulnerability indicators. KIMS allows for the collection, analysis and visual display of data at different levels of aggregation, and can assist the monitoring and surveillance of the food and nutrition situation over time. The Key Indicators Data System (KIDS) is the Web version of KIMS and is employed in the realisation of several web-based applications for data visualisation that require dynamic mapping functionality.

xii. GeoNetwork aims at providing a common “window” to work interactively with FAO’s vast wealth of map and related information, making the earth’s geography a starting point for finding, retrieving and using information. This includes population density, infrastructure, administrative boundaries, land cover/use, soils, crop zones, water, marine fisheries, inland aquaculture and forest resources, livestock distribution, nutrition profiles and early warning information. GeoNetwork is an integral part of the spatial information infrastructure being developed by SDRN, explicitly as an aid to decision-making for sustainable development among FAO Member Nations and other stakeholders such as UN Agencies, intergovernmental organizations and NGO’s. GeoNetwork is using an ISO standard for its metadata (ISO-19115 DIS) and an internationally established Open GIS Consortium (OGC) standard for its actual spatial databases.

xiii. RANET (the Radio & InterNET for the Communication of Hydro-Meteorological and Climate Related Information: project was designed specifically to address the digital divide between North and South and make climate and weather related information more accessible to national and provincial institutions and rural communities. It was conceived and started by the African Centre of Meteorological Applications for Development (ACMAD), with valuable support from partners in the United States (University of Oklahoma, the National Oceanic and Atmospheric Administration (NOAA), USAID), and is currently supported by a wide range of national and international institutions from all over the world. FAO is actively cooperating in RANET with its near real time Advanced Real-Time Environmental Monitoring Information System (ARTEMIS) and agroclimatological databases, including the Desert Locust forecast bulletins.

xiv. In cooperation with its partners, RANET is now able to make observations, forecasts, and bulletins more readily available to hydro-meteorological and extension services in Africa. The information is broadcast over the AfriStar satellite as a digital radio signal on the ALC (Africa Learning Channel) data broadcast. The satellite services are made available by the WorldSpace Foundation. With an appropriate radio receiver, this information can be transmitted to a computer on which it can be seen as text, illustrations and pictures. In this way the limitations of Internet availability are no longer a barrier to critical information access.


1 "Data integration and its role in the development of better agriculture and food information systems” by L. Kabat, L. Naiken and P. Narain paper prepared for Invited Paper Session 15 - Integrating Agriculture and Food Statistics: National and International Perspectives, of the International Conference on Establishment Surveys – II organised jointly by USDA, Statistics Canada and American Statistical Association at Buffalo, NY, USA - June 2000.

2 Response Farming was originally developed by Ian Stewart through the WHARF Foundation (World Hunger Alleviation through Response Farming) at the University of California - Davis. Some environmental variables are measured and determined at the village or farm level, for instance rainfall and soil type. The elementary version of RF, guides farmer decisions by comparing the environmental variable with reference values listed in decision tables. Typical examples include the definition of planting dates and rates of fertiliser application. The decision tables have been centrally prepared taking into account historical information on local climate effect on crop yields, local soil types and local crops, crop prices. More complex forms of response farming take advantage of modern data transmission technology to collect and disseminate local, village-level data and real-time analysis of local data using simulation models. The analyses may also involve processing of satellite imagery. Thus, the final cost of advice to farmers is best expressed as cost-per-farmer.

3 Some data sets can be said to be “semi-geo-referenced”, for instance the rice databases of AGP where varieties are associated with agroecological zones.