by Fred Snijders,
Remote Sensing Officer
Environment and Natural Resources Service (SDRN)
This article was first published in Remote Sensing & Geo-Information, nr 2, August 2001. RS&GI is a publication of the Dutch "Beleidscommissie Remote Sensing" (BCRS), Delft, The Netherlands.
One of the goals of the Food and Agriculture Organization (FAO) of the United Nations is the reduction of food insecurity in the world. This was further re-iterated at the World Food Summit held in 1996 where a Plan of Action was adopted aiming at reducing the number of the world's hungry people in half by 2015. While the proper foundation of this goal lies, among others, in the increase of food production and ensuring access to food, there is also a need to monitor the current food supply and demand situation, so that timely interventions can be planned whenever the possibility of famine, starvation and malnutrition exists. With an imminent food crisis, actions need to be taken as early as possible as it takes time to mobilise resources and logistic operations are often hampered by adverse natural or man-made conditions, including war and civil strife. The availability of objective and timely information is, therefore, crucial.
While much emphasis is generally given to estimation of crop or food production in a specific country or region, there is also a need for information on the overall demand for food, access to food, availability of food stocks, and on commercial imports and exports, and in some cases pledges for food aid. In addition, information concerning the food situation in neighbouring regions or countries is required as surpluses can be used as potential source for food aid. Therefore, in the overall context of food security, knowledge of local food production provides only part of the information required to obtain a clear picture of what is available and what is needed and, not only are various inputs required for an adequate picture of the food supply situation, but the information must be up-to-date and available on a regular basis for the operational monitoring of the food supply and demand at all levels.
Within FAO these monitoring activities are co-ordinated by the Global Information and Early Warning System (GIEWS) on Food and Agriculture. Its goal is to provide policy makers and analysts with up-to-date information on all aspects of food supply and demand, including production, trade, prices, etc. and aims at answering questions such as "where are food interventions most needed", "how much food is the world producing" and "what is the impact of civil war on food security".
The FAO GIEWS is not the only group active in this field. There are a number of commercial companies that closely monitor the main grain producing areas and market this information. However, this type of commercially available information often does not cover countries that are considered low income - food deficit, as their local food production is of little relevance to the international grain trade. There are also other governmental and non-commercial groups involved in these activities, most notably the US Famine and Early Warning System Network (FEWSNET), funded by the US Agency for International Development, which focuses on a number of countries in Africa, and the SADC Regional Early Warning System. However, FAO remains the only institution that monitors the food situation on a global scale using objective methodologies and assessments, and does not focus on the needs for such information by any particular government or economic group.
Global monitoring requires a large network of information sources and collaborative activities. One of the pillars of this network is a direct flow of information from the individual countries to FAO. In particular in the developing world, this is often coupled to capacity building activities by FAO, including the establishment or support of national "Early Warning and Food Information Systems" (EWFIS). Furthermore, close contacts are usually kept with NGOs; other monitoring groups, such as the afore-mentioned US-AID FEWS with which there have been many collaborative efforts; with related international programmes, such as the World Food Programme and the UN Office for the Co-ordination of Humanitarian Affairs (OCHA); and with Donors of food assistance.
While a large portion of the relevant data is obtained through these various collaborative efforts, FAO also performs its own crop monitoring, with a particular focus on countries and regions which suffer from a lack of continuous, reliable information on weather and crop conditions. In addition, independent monitoring capabilities can improve the objectivity of the final analysis by FAO. FAO's crop monitoring activities rely heavily on satellite remote sensing and agrometeorological techniques. The remote sensing support to GIEWS is provided by the ARTEMIS system. ARTEMIS was originally developed with the generous support provided by the Government of the Netherlands and is now fully integrated into the FAO Regular Programme. Operational since 1988, it originally focussed on Africa, as the continent had experienced a number of severe droughts during the 1970s and '80s. ARTEMIS was developed to ensure that the information gap that existed during these droughts did not occur in future years. Had ARTEMIS data been available, it could have been used to provide better early warning of the impending droughts and famines.
The main satellite systems used by ARTEMIS at the time of its inception were METEOSAT for rainfall monitoring and NOAA-AVHRR (GAC) for vegetation monitoring. At present, ARTEMIS provides a global coverage and the resolution has been reduced from, in general, 8 to 1 km, slowly moving towards the 500/250-meter scale; a development that has been facilitated by the establishment of many partnerships and involves, among others, NASA, USAID-FEWS, USGS, JMA, NOAA, ESA, EUMETSAT, SPOT-IMAGE, JRC and the European Commission (EC).
Improvements in telecommunications and information networking, in national capacities, and applications of newer technologies such as remote sensing have greatly increased the amount of information available to GIEWS and other early warning systems. This increase in information has also created a need to develop systems suited to managing, analysing and rapidly integrating the data for early warning and crop monitoring. A response to this has been the development of the "GIEWS Workstation" which provides PC based access to many of the GIEWS information sources and allows, for instance, the rapid display of ARTEMIS imagery. Through the Workstation, much of the technological complexity of accessing the data has been hidden from the country analyst through a series of customised menus.
The use of remote sensing imagery in support of crop monitoring has greatly improved over the years. Desktop and laptop personal computers now allow early warning analysts to access and analyse satellite data from their desks or even while on mission. Considerable progress when one compares this to the early start of ARTEMIS when imagery was distributed in the form of Polaroid prints! The geographic coverage (global) and spatial resolution of data available through the ARTEMIS system has greatly increased since 1998 with the incorporation of the SPOT/VEGETATION data, purchased by FAO with the support of the EC.
However, less progress has been made in the quantitative use of the remote sensing imagery. For example, the type of analysis conducted by FAO using NDVI data, one of the two mostly used image types, at continental and regional scale, has not changed much since 1988. While data registration and calibration have become much better, the main type of analysis is still a qualitative assessment of the current vegetation situation, or of the whole growing-season, as compared to previous years or the average by comparing trends in NDVI.
The use of METEOSAT data as single source for rainfall estimates has only been partially successful. Although quantitative estimates of rainfall derived from METEOSAT are becoming more refined, the most popular METEOSAT derived product for early warning and crop monitoring remains the "cold cloud duration" images. An interesting fact is that these images were originally intended to be an intermediate product only, and not suitable for distribution! More recently, METEOSAT data has been merged with data from other sources to improve the quality of the estimates. At the national level, this is often done by interpolating between a large number of observations from meteorological stations using the cold cloud images as a weighted surface guiding the interpolation patterns and the ground observations for the quantification. At a continental scale, only rainfall data through the Global Telecommunication System (GTS) is readily available in real-time, which is often of uncertain quality over many food insecure countries. Recent approaches now combine METEOSAT with GTS data and data from microwave images or sounding instruments, the most popular microwave data being SSM/I.
While improvements are still needed before remote sensing can provide the quantitative data required for crop production estimate models, the remote sensing data remain an essential, yet partial, component for the monitoring of food supply and demand to improve world food security.