Diplomacy Homepage >> Virtual
Diplomacy Publications >> Space
Aid: Current and Potential Uses of Satellite Imagery in UN Humanitarian
SPACE AID: CURRENT AND POTENTIAL USES OF SATELLITE
IMAGERY IN UN HUMANITARIAN ORGANIZATIONS
Einar Bjorgo, Ph.D.
The views expressed here are those of the author
and not necessarily those of the United Nations, the United Nations
High Commissioner for Refugees (UNHCR), or the United States Institute
United Nations (UN) humanitarian organizations are all in need
of relevant geographic information to prepare for and respond to natural
and human-devised, or manmade, disasters. By organizing collected information
in a geographic information system (GIS), UN humanitarian organizations
can more effectively analyze situations and coordinate relief efforts.
These organizations, as listed by the UN Office for Coordination of
Humanitarian Affairs (OCHA), include the UN High Commissioner for Refugees
(UNHCR), World Food Programme (WFP), UN Development Programme (UNDP),
UN Children's Fund (UNICEF), Food and Agriculture Organization (FAO),
World Health Organization (WHO), UN High Commissioner for Human Rights
(UNHCHR), and UN Population Fund
as well as OCHA itself. 1
The U.S. Geological Survey (USGS) has described a GIS as "a computer
system capable of storing, manipulating, and displaying geographically
referenced information, i.e., data identified according to their locations.
Practitioners also regard the total GIS as including operating personnel
and the data that go into the system."
In complex emergencies, a GIS has become an extremely useful tool whereby
organizations store geographic information on, for example, infrastructure,
population distribution, food availability, and climatic conditions, which
is then translated into maneuverable databases, analyzed, and finally
mapped. It is in this context that satellite images have a role. 2
The objective of this paper is to provide an up-to-date overview of the current and potential use of satellite imagery within the UN humanitarian organizations and their operations and to illustrate the wide range of applications of earth observation (EO) data among this community. An in-depth assessment of all activities related to satellite imagery within each agency is beyond the scope of this paper.
Because UN humanitarian organizations provide relief for a wide range of natural and manmade disasters, often in unfamiliar areas and in complex situations involving large groups of affected people, government agencies, nongovernmental organizations (NGOs), military, and other international organizations, the need for accurate and relevant geographic information is fundamental (Bouchardy 1997, 2000; King and Dilley 2001). Depending on their specific mandates, these organizations have different information needs, some of which can be met using various types of imagery acquired by orbiting satellites. EO images range from those with large area coverage and low spatial resolution to those with small area coverage and high resolution (Bjorgo 2000a, 2001) (see figure 1). Satellite images are processed and analyzed according to the specific needs of the organizations and distributed to users in the field and/or decision makers in regional offices or headquarters.
Figure 1: Image coverage and resolution for different types of satellite sensor imagery.
Even though satellite imagery has been available for more than three decades, for the first fifteen years its use was limited to the military and scientific communities. Practical, operational use of this technology for civilian purposes was not particularly successful at that time, and the industry's expectations about what useful information could be delivered to end users were too high. The remote-sensing industry -- including data providers, value-added service providers, and, most important, the end users themselves -- has matured and accordingly makes a more practical assessment of the value of this kind of data to its work.
Data from earth observation satellites are now regularly
used by UN organizations following natural disasters (Committee for Earth
Observation Satellites [CEOS]
FAO is one organization with considerable expertise in this respect. However,
the use of satellite images for manmade disaster relief operations such
as refugee flows is relatively new within the UN humanitarian relief community.
UNHCR was among the first to use this technology for information management
(IM) directly related to refugee assistance in the early 1990s. Many other
UN humanitarian organizations now focus on IM applications that support
their mandates. With more proven processing and analysis methods, extended
spatial and temporal coverage, and improved software for desktop publishing,
several UN humanitarian organizations now include satellite imagery as
a core element in their IM system and use it to support the various phases
of their humanitarian field activities, ranging from preparedness and
climate monitoring to emergencies and rehabilitation. 3
Because there is no UN agency in charge of centralized satellite image purchasing, this overview of the current and potential use of satellite imagery in UN humanitarian organizations is based on communication with and documentation from selected contacts within the organizations. Because of the complexity of the UN humanitarian system, this paper should not be taken as comprehensive but rather as an effort to illustrate the wide range of information needs and related applications of EO data within this community.
Humanitarian relief is a collective term, comprising actions taken to alleviate suffering from natural disasters such as floods, forest fires, hurricanes, earthquakes, and droughts, and manmade disasters such as civil and interstate wars, with their resulting population displacements and refugee flows. With responsibilities related to coordination, refugee protection and assistance, food delivery, development, children's rights, nutrition, agricultural productivity, health, human rights, and population issues, the UN humanitarian organizations have numerous geographic information requirements depending on which sectors with which they are involved. Nevertheless, several common needs prevail. All UN organizations require that geographic information be accurate, objective, timely, and in a suitable, standardized format.
To make efficient use of geographic information collected before or during humanitarian operations, the data should be stored and analyzed in a GIS. This requires that the data be geo-referenced and ideally available in a standardized format, so that data can be swiftly uploaded to the GIS. Satellite images can be delivered with geographic coordinates for referencing when purchased through a data provider. However, the geometric accuracy of the data and their end products vary depending on how the geo-referencing was carried out and on the quality and number of ground control points used. Such geo-referencing is often carried out using field-based readings from global positioning system (GPS) handheld receivers.
A major improvement occurred on May 1, 2000, when the so-called
GPS Selective Availability (related to an intentional degradation of the
precise military GPS signal for public users) was stopped
GPS readings are now ten times more accurate than before -- between three
and ten meters depending on the satellite signal coverage (see figure
2) -- and thus can be used to improve the accuracy of relatively detailed
satellite imagery. However, for the highest resolution of satellite imagery
commercially available (currently one-meter resolution), one still needs
differential GPS for accurate geo-referencing. The point is that satellite
images should be geo-referenced according to the application and level
of accuracy the information managers need. For example, geo-locating damaged
infrastructure in war-torn areas requires less accuracy than does detailed
refugee camp planning. 4
Figure 2: Comparison of accuracy of GPS signal with and without selective availability (SA). The figures show that with the end of SA, the accuracy of the GPS signal improved from approximately fifty meters (left) to five meters (right) for the test location (Erlanger, Kentucky). Source: National Geodetic Survey, National Oceanographic and Atmospheric.
Of course, the recorded satellite data signals must be accurate for policy decision-making processes. This accuracy is related to technical parameters such as radiometric and spectral accuracy. Most satellite sensors used by UN humanitarian organizations have been tested for many years, and data analysis algorithms are verified. Even so, satellite images consist of information collected from 400 to 36,000 kilometers above the earth's surface. Haze and on-the-ground meteorological conditions such as snow and ice may affect the data sensed from space. That is why image processing, analysis, and interpretation should be carried out by experts and the classification of images verified by "true" field values -- that is to say, if satellite data classify an area as deforested, such information must be verified with direct observation from the field.
One of the benefits of a satellite image is its objectivity. This turns satellite image-derived assertions into valuable baseline information when presenting facts to governments or donors. Because satellites record what actually exists on the ground, nobody can argue that the information has been omitted or changed, as can be argued if potentially biased individuals or companies carry out field surveys in remote areas. An objective view from above thus provides common ground for stating facts and a framework for making decisions.
Of course, as with any information, products derived from raw satellite images can be altered or interpreted with a bias toward the results one wants to achieve. But because the original information is objective and can be verified, digital raw data can be reviewed and reanalyzed should an interpretation be disputed.
It is important to be aware of other factors such as atmospheric effects and the time of image acquisition when interpreting satellite images because clouds may obstruct the view to certain areas or, as in the case of a flood, the image may not have been acquired during its peak period.
A critical information requirement is for timely data. UN humanitarian organizations are involved in a multitude of applications, ranging from long-term operational monitoring to emergency assistance and contingency planning (Bjorgo 1999). Typical requirements for timely delivery are listed in table 1.
REQUIRED DELIVERY TIME
Annual updates of aggregated climatic data
Baseline infrastructure data
One to two weeks
Ten-day averaged land cover data
Monitoring agricultural production
Imagery of disaster-struck areas (e.g. floods, landslides)
One to two days
Table 1: Typical requirements for timely delivery of satellite image-derived
information to UN humanitarian organizations.
In some cases, such as for hurricane warnings, UN humanitarian organizations need data daily or on a near-real-time basis (here defined as less than twenty-four hours). The requirements vary with the applications and mandates of the various organizations. Several factors affect the time it takes to deliver an image to the end user. First, the area of interest must be within the view of a satellite. Depending on which satellite is used, this can vary from several times a day (as for geostationary weather satellites tracking hurricanes) to once in several weeks (as for mapping of food production potential during spring growth using the U.S. Landsat 7 satellite). Second, the data-receiving station and its facilities may not be able to process the image immediately after reception. Depending on the urgency of a specific data order, processing can take from several days to weeks. Third, the data providers, often local resellers, have their own priorities depending on market potential, delivery services offered, and workload. All in all, many factors often lead to serious delays in obtaining useful satellite imagery for humanitarian relief operations. UN humanitarian organizations must be aware that near-real-time data are not always available with satellite-obtained data.
Standard data formats
The Geographic Information Support Team (GIST), an informal
interagency support group
for sharing geographic information between actors in the humanitarian
relief community, has developed a set of core standards for information
sharing (King and Dilley 2001). These standards, following the "structured
humanitarian assistance reporting exchange" (SHARE) approach, encourage
actors to include 5
geographic reference or location information on where data were collected;
time-stamp indicating when data were collected, and where suitable, at what frequency new data are collected;
metadata (information about the data themselves), including information on sources of data, what the data values represent, which standards were used, and how the data were acquired.
One of the main benefits of satellite imagery is that service providers already follow these standards. This makes satellite data particularly suitable to be included in GIS and to be shared with other actors. Hence, there is no need to develop new formats for the relief community.
First, satellite images are normally delivered as raster
data, either already geo-referenced (e.g., in the GeoTIF format) or with
header files (metadata) containing information on image center and corner
satellite data contain detailed information about the exact time the image
was acquired. Because satellite images are no more than snapshots of the
situation on the ground, the time-stamp is crucial, which is why it is
included in satellite data standards. For derived products, such as annual
averages of rainfall in a region, information about when and where the
data were collected, accompanied by a detailed list of all the individual
scenes used in the production of the end product, should be included with
the images. This is important to ensure objectivity and transparency in
the interpretation of the satellite-image-derived end products. Third,
metadata, as described above, appear either as separate files (e.g., header
files) or as part of the image itself. Typical parameters listed in satellite
metafiles are the type of sensor, center and corner coordinates, acquisition
date and time, satellite orbit parameters, file size in pixel width and
height, number of channels for multispectral images, processing level
and data provider, information on missing data, and number of bits/pixels.
The more advanced the data processing and interpretation, the more metadata
are required to "verify" the image's meaning. Metadata are also useful
in a GIS, because users can search for image coverage of special areas
or during particular time intervals. 6
Even though the various types of images and sensors often
have specific formats, most of them can be directly read using off-the-shelf
image-processing software. Such technical image standards -- for example,
GeoTIF, EOSAT Fast Format, ERMapper raster -- are important for rapid
processing and information sharing. With these standards in place, satellite
imagery is an ideal source of information for inclusion in GIS
and for sharing among humanitarian organizations. 7
The legal framework for using satellite imagery to monitor
the Earth is described in one main treaty
the "Treaty on Principles Governing the Activities of States in the Exploration
and Use of Outer Space, Including the Moon and Other Celestial Bodies"
(United Nations, 1967). This treaty is commonly known as the "Outer Space
Treaty." In addition, there is a set of legal principles 8
for the application of remote satellite observations of Earth: The "Principles
Relating to Remote Sensing of the Earth from Outer Space" (United Nations,
1986). According to Article I of the Outer Space Treaty, (United Nations,
1967) "the exploration and use of outer space, including the moon and
other celestial bodies, shall be carried out for the benefit and in the
interests of all countries, irrespective of their degree of economic or
scientific development, and shall be the province of all mankind." Furthermore,
Article 2 in the same treaty reads: "Outer space, including the moon and
other celestial bodies, is not subject to national appropriation by claim
of sovereignty, by means of use or occupation, or by any other means."
Principle XII of the "Principles Relating to Remote Sensing of the Earth
from Outer Space" allows states to receive copies of imagery acquired
over their territory: "As soon as the primary data and the processed data
concerning the territory under its jurisdiction are produced, the sensed
State shall have access to them on a non-discriminatory basis and on reasonable
cost terms." However, it is important to be aware that according to international
law, no state can deny acquisition of satellite imagery of its territory. 9
UN humanitarian organizations are facing many of the same limitations of satellite imagery as other users (Bjorgo 2000b). Limitations can be categorized as physical or institutional. Among the physical limitations are cloud cover, repeat frequency for imaging the same area on the ground, and image resolution (the amount of detail in an image). Institutional limitations include delivery time of images, cost of data, copyright, and "shutter control."
Clouds: Most commercial earth observation satellites cannot photograph areas covered by clouds. This presents a particular problem in the tropics and temperate zones, where most humanitarian relief efforts take place. Synthetic aperture radar (SAR) sensors operate at wavelengths not impeded by cloud cover or the lack of illumination (radar satellites can also acquire imagery during night passes). SAR technology, however, requires more processing and specialist interpretation than do satellites that operate in the visual and near-visual part of the electromagnetic spectrum, such as Landsat, SPOT, and Ikonos. This does not rule out the use of SAR, especially in areas known to exhibit heavy cloud cover. For flood monitoring, for example, SAR imagery is often the only type of spatial information available during heavy rainfalls.
Repeat frequency: Humanitarian organizations frequently operate in highly dynamic situations, which require constant information updates. Satellites have to revisit and re-image areas periodically (referred to as "repeat frequency"). Currently, repeat frequency varies from one to more than twenty days for various types of non-geostationary (i.e., polar orbiting) satellites. Because most systems require cloud-free conditions, it can take much more than one revisit cycle before a sensor can acquire another image of the same area. As more satellites populate space, however, images can be ordered from different satellite systems. The problem with this approach is that comparing data is more difficult because different cameras from slightly different orbits and incidence angles produce slightly different images. Fortunately, more sensor cameras can now be programmed to adjust their viewing angle while in orbit in order to photograph the same area at more frequent intervals. This is the case, for example, with the SPOT and Ikonos satellites.
Image resolution: Now that one-meter resolution is
commercially available, the detail in satellite images is beginning to
meet the needs of UN organizations for humanitarian operations. Although
more detailed images are always preferable, we are not likely to see such
a dramatic jump in the technology as the recent hundredfold improvement
in spatial resolution that the ten-meter to one-meter resolution represents.
Today's most advanced restricted military satellites, thought to have
a resolution at ten centimeters or less according to the Federation of
American Scientists, 10
are the only satellites that offer images of such minute detail. Most
UN activities, such as monitoring land cover changes, do not require such
high resolutions. 11
Delivery time: For standard operational applications, such as regularly updated land cover and various climatic data, delivery times are routinely met. Image processing and delivery are automated, and institutional delays caused by processing individual custom orders are usually avoided. For regular, nonemergency requests, there are Internet services that deliver certain types of archived imagery -- for example, Landsat 7 -- within one to five days, a considerable improvement over the former lag time of several weeks. The problem comes when humanitarian organizations need near-real-time data for emergency applications. Although near-real-time meteorological data are available, they provide fewer details than are needed for damage assessments. When it comes to detailed single-scene imagery, no data provider has yet dependably delivered this quality of detail in a twenty-four-hour cycle, because, for example, of obstructing cloud cover and limited processing capabilities.
Unless there are ongoing monitoring programs, such as FAO's ten-day updates of global vegetation cover using SPOT one-kilometer data, new images are ordered on a one-off custom basis. This is a slow process because of institutional procedures on both the customer side -- related to budget and the time-consuming processing of purchase orders -- and the distribution side -- handling new incoming orders, locating the imagery in the database, and subsequent processing. It is therefore crucial that established procedures between customers and data are in place before emergency image requests arise. Nor should these procedures depend on personal contacts, whose absence at any point could obstruct and slow the processing of an emergency request.
Cost: UN humanitarian organizations generally assume that satellite imagery is expensive. With their limited budgets for purchasing satellite data, they tend to opt for the less expensive solutions. Cost, however, is relative. It depends on the type of imagery purchased and the urgency of the need compared with the cost of acquiring, processing, and analyzing data by other cheaper, less accurate, and probably slower means. The cost of data acquisition may range from no cost (e.g., Defense Meteorological Satellite Program [DMSP] data used by World Health Organization [WHO] for nighttime light mapping) to more than $2,000 (e.g., Ikonos single one-meter resolution of a scene). In addition to the cost of acquiring data, processing and analyzing data is expensive. Indeed, processing and analyzing data is often more expensive than buying raw data and should be included in the budget when planning imagery purchases, if no in-house experience is available.
Comparing the relatively high cost of acquiring, processing, and analyzing satellite data with other lower-cost information usage, such as field surveys or aerial photographs, depends on a number of other contextual factors. Copies of existing aerial photos from local institutions, which are available at low cost, may meet an organization's information requirements. On the other hand, it can be time-consuming, and thus expensive, to collect raw data over a large area by contracting for a field survey. The cost of renting a plane with pilot, fuel, equipment, and so forth for non-archived aerial photography can quickly exceed that of satellite imagery. When up-to-date information on isolated or no-access areas is needed, satellite imagery may be the only practical source. In addition, unlike satellite imagery, which is delivered in a geo-coded format, aerial images are not, and geo-coding adds considerably to the processing cost.
On a positive note, the price policy of Landsat 7 data is encouraging. Today's $600 per image is affordable for most UN organizations. Moreover, the quality and lower cost of Landsat 7 data could force other imagery providers, such as SPOT, which operates similar satellites, to reduce their prices. The resulting benefit to the humanitarian assistance community, which could then more easily afford to use satellite imagery in its work, is obvious.
Copyright: With the exception of Landsat 7, all purchased satellite imagery comes with varying degrees of copyright restrictions. This is a serious problem because useful original raw data cannot be shared among UN humanitarian organizations or their implementing partners. Although their derived products (e.g., classifications of forest cover) can be shared, the raw data are restricted to the purchasing organization, sometimes even to a single department within an organization. This restriction limits the use of satellite imagery by financially strapped organizations. As the need for imagery among humanitarian organizations increases, the need to overcome restrictions against sharing becomes more pressing. A possible solution is to transact a memorandum of understanding (MOU) between UN humanitarian organizations and satellite image providers. The MOU would include negotiated conditions and cost for sharing critical humanitarian data. By sharing images, the various UN users would contribute financially toward one image but receive several copies for the MOU group ("pool-sharing" of a single image). Image-sharing does not necessarily mean less income for the commercial data providers. Sales of imagery may actually increase because more users will pass the threshold for investing in imagery; hence, a win-win situation should be possible.
Shutter control: Shutter control is government-ordered
restriction of the distribution of satellite imagery. According to the
license granted to commercial U.S. satellite operators, distribution of
imagery can be restricted "during periods when national security or international
obligations and/or U.S. foreign policies may be compromised -- as defined
by the Secretary of Defense or the Secretary of State."
Except during the Gulf War, when the public was denied access to several
types of updated imagery over Iraq and Kuwait, the U.S. government has
never enforced shutter control. With more non-U.S. satellites now in space,
UN humanitarian organizations can use non-U.S. imagery and avoid shutter
control on satellite companies operating under U.S. licenses if necessary.
(Of course, satellites operated under other nations' licenses may also
be restricted, as was the case with the French SPOT system during the
Gulf War.) Although not strictly considered "shutter control," imagery
access can also be restricted through purchasing of exclusive image rights,
thus denying other potential users timely access to information. This
was the case with Ikonos imagery acquired over Afghanistan in the aftermath
of the September 11 attacks in the United States. (Lindsey Butts, Space
Imaging, personal communication). Compared with the cost of data and copyright
restrictions, shutter control is normally not a major limitation for humanitarian
organizations -- at least for the time being. More in-depth discussion
on shutter control can be found in Secrets for Sale (Dehqanzada and Florini,
Depending on their respective mandates and particular needs (see
table 2), UN humanitarian organizations use a wide range of imaging
satellites. The range of satellite sensors and applications used by
these organizations is exemplified below (figures 3 - 8). Several UN
organizations have extensive experience in using satellite remote sensing
as a critical tool in fulfilling their mandates. FAO, UNHCR, and to
some extent UNDP are the most experienced users of satellite imagery,
although they use various image types. The examples below do not represent
a comprehensive list of the applications of satellite imagery within
these organizations but rather illustrate the wide range of earth observation
satellite data used by UN humanitarian organizations.
Use of satellite imagery to support the protection of and assistance to the world's refugees started with UNHCR environmental assessments in areas close to refugee camps (Bouchardy 1995). From this application, UNHCR and its Geographic Information and Mapping Unit (GIMU) have moved to applying a wide range of satellite imagery on an operational basis. These applications include support of emergency assistance, repatriation and rehabilitation, and contingency planning. Although UNHCR has some experience using radar imagery (Radarsat and ERS-1 and 2), the main sensors used are SPOT, Landsat, IRS-1D, and Ikonos. UNHCR has also used archived Russian data (KVR-1000). The image in figure 3 illustrates the detailed level of information commercially available from very high-resolution satellite data. A refugee camp in Nepal was imaged at one-meter resolution by Space Imaging's Ikonos satellite. Note the level of detail: individual shelters, narrow streets and paths, and small rivers. Such detailed images are useful for camp planning and detailed environmental assessments.
Figure 3: This Ikonos image of Beldangi refugee camp in Nepal was acquired and analyzed by Satellus Metria of Sweden as part of the ENVIREF project, of which UNHCR is a partner. Copyright: Space Imaging.
UN Humanitarian Organization
Geographic Information Need
Expertise in satellite imagery
In-house and/or outsourced
Basic data layers, mechanism to share GIS data in order to facilitate coordination.
Supports purchase and coordination of satellite imagery, but no in-house experience per se.
Spatial information on areas of refugee locations for contingency planning emergency assistance, environmental monitoring, repatriation, and rehabilitation to assist field operations.
Extensive expertise since 1995. Has used data from most types of satellite sensors, including SPOT, Landsat, IRS-1D, KVR-1000, Ikonos, Radarsat, ERS 1 and 2.
In-house and outsourced
Land cover and rainfall data for crop monitoring, food availability, and early warning assessments.
Some experience with imagery applicable for crop monitoring using Meteosat and NOAA Advanced Very High Resolution Radiometer (AVHRR).
In-house, some outsourced
Wide range of environmental, population, health, logistics, and economic data to assist development programs to combat poverty.
Expertise using data from a wide rang of radar and optical satellites, such as Radarsat, Lansat, and SPOT.
Spatial data to complement children's programs in areas such as education, health, and living conditions.
Limited experience in satellite imagery. Involved in use of imagery from Radarsat, Landsat, and IRS 1C on a few occasions.
Spatial data with focus on temporal changes of vegeation cover, and use, and related parameters for early warning, crop monitoring, and food availability assessments.
Extensive expertise using data from optical satellites, such as Landsat, SPOT Vegetation, SPOT, IRS 1C/D, NOAA AVHRR.
In-house and outsourced
Distribution of population and health services, type and extent of diseases, climatic data to plan and monitor programs.
Limited experience in satellite imagery. Some experience with data from DMSP Operational Line Scanner (OLS) and other optical sensors.
In-house and outsourced
Information on population distribution and living conditions in regions and countries to strengthen field presence and work toward compliance with international human rights laws.
In general, no use of satellite imagery.
National, regional, and global population data and related parameters, such as health, to assist developing countries in finding solutions to population problems.
In general, no use of satellite imagery.
Table 2: Information needs and expertise in satellite imagery of UN humanitarian agencies.
The World Food Programme (WFP) has experience with different types of EO satellites. It is the only agency of those described here that uses a geostationary satellite series. The Meteosat meteorological satellites are used for weather forecasting and for assessing climatic conditions, such as rainfall data, important for predicting potential droughts. WFP also uses NOAA Advanced Very High Resolution Radiometer (AVHRR) data to monitor land cover and to provide early warning about poor crop production. The use of satellite imagery within WFP, however, is ad hoc and not mainstreamed in the decision-making process.
Figure 4: Rainfall estimates for northeast Ethiopia carried
out by WFP
To fulfill its mandate to combat poverty in hardship regions, the UN Development Programme (UNDP) has a wide range of geographic information needs. It uses satellite imagery in a variety of ways to meet those needs. Using commercially available EO data, as shown here from the UNDP Somalia GIS lab, UNDP can assess "informal settlements" around Nairobi, Kenya. According to satellite image spectral analysis, what is here seen in white represents informal settlements; black represents urban areas; and the two shades of gray represent vegetation (darker gray for dense vegetation, lighter gray for less dense vegetation). Other UNDP uses of satellite imagery include radar imagery for flood assessments and optical sensors for land-use analyses and population estimates of major towns.
Figure 5: Assessment of informal settlements (seen in white) carried out by UNDP Somalia GIS lab.
Although the UN Children's Fund (UNICEF) does not use EO data regularly, it has used it on occasion to support its mission. The example shown here is a Radarsat image of the main towns affected by the flooding of the Juba River in Somalia in 1997. UNICEF has also used IRS and Landsat imagery of areas exposed to earthquakes. Such assessments, together with demographic data, are used to plan special relief and assistance programs directed toward children in affected areas.
Figure 6: Flood extent as imaged by the cloud-penetrating Radarsat satellite.
As one of the main users of satellite imagery within the UN system, the Food and Agriculture Organization (FAO) has considerable knowledge and expertise in various types of EO sensors, such as Landsat, SPOT Vegetation, SPOT, and IRS. FAO uses EO data on a regular, operational near-real-time basis for early-warning purposes. Geographic information obtained using satellite imagery includes cloud-cover duration for surrogate rainfall estimates, vegetation monitoring for food security assessments, land-cover analyses, deforestation, and change in cultivation patterns. FAO has also developed its own satellite image processing and analysis tool for rapid assessments of the core geographic information parameters useful to support its mandate, such as land-cover classification. The image shown in figure 7 is a typical normalized differential vegetation index (NDVI) analysis of the Horn of Africa's vegetation cover. Satellite-derived information constitutes a core part of FAO's decision-making procedures.
Figure 7: Normalized differential vegetation index (NDVI) analysis of the Horn of Africa's vegetation cover using NOAA AVHRR.
The use of satellite imagery is relatively limited within the World Health Organization (WHO). Working as a project partner, WHO has used EO technology to indicate climatic conditions and to map areas vulnerable to malaria in Africa. As seen in the example shown here, WHO used EO data innovatively -- DMSP operational line scan (OLS) data measuring area of illumination on a global scale -- to show populations at health risk. WHO has a well-developed GIS, which can also incorporate remote-sensing imagery.
Figure 8: Global nighttime lights (in white) for income prediction using DMSP OLS data. The inset is a close-up of the Nile River delta and parts of the Middle East.
Overview of Systems
Table 3 provides an overview of the various EO satellite systems used by UN humanitarian organizations. Systems are separated into two groups depending on their typical applications. Group A includes satellite imagery suitable for more detailed, local analyses and emergency situations -- Landsat 7, SPOT, IRS-1C/D, KVR-1000, Ikonos, Radarsat, and ERS-1/2. Group B is more suitable for regional coverage at less detailed resolution and longer term monitoring -- DMSP OLS, NOAA AVHRR, and Meteosat.
Within Group A, the most widely used satellite systems are Landsat 7 and Landsat 4 - 5 (four user organizations) and SPOT and IRS-1C/D (three user organizations). For cloud-penetrating images, two organizations have used Radarsat.
Based on an overall comparison of system quality, range of applications, cost of data, and copyright restrictions, Landsat 7 offers the best range of features as an EO satellite for the UN humanitarian organizations involved in emergencies and for relatively detailed assessments. The cost of data from Landsat 7 is an affordable $600 per image. It covers a large area (185 km x 185 km), and with its eight channels, is suitable for a range of proven classification techniques. The fifteen-meter spatial resolution in panchromatic (black and white) mode is comparable to that of SPOT (ten meters), and archived data can be delivered over the Internet in one to five days. Finally, this is the only satellite in Group A that has no copyright restrictions. Raw data and derived products from Landsat 7 can be shared freely. This feature is important during UN humanitarian operations, which often require interagency information management and sharing of imagery.
If a user organization requires very high levels of detail, then Ikonos (one-meter resolution) is suitable for emergency-type assessments. The recently launched EROS-1A satellite will compete with the more costly Ikonos among humanitarian users, because its 1.8-meter resolution provides very detailed imagery and the cost of data is significantly less ($1,500 compared with $3,500 for a similar imaged area). EarthWatch's QuickBird satellite, with its 0.61-meter resolution, was launched on October 18, 2001, and is undergoing in-orbit testing in preparation for commercial operations. This satellite has the potential to be a good data source when detailed information is required. Information on prices for QuickBird scenes was not available when this article went to press.
Among Group B, NOAA AVHRR is the system most used (two UN organizations). NOAA AVHRR data are used at the operational level more regularly than data from the systems in Group A. NOAA AVHRR is an integrated part of early warning and decision-making processes in FAO.
NOAA AVHRR is a good up-to-date source for information on global vegetation cover and areas of potential food shortages. Meteosat is also useful for assessing cloud cover and therefore rainfall information. Because these data are available in near real-time and at no or low cost, it is not surprising that FAO and WFP use these satellites.
In addition to the systems listed in table 3, the geostationary Meteosat series used by WFP provides imagery of half the Earth -- showing the Earth as a two-dimensional disk -- at a frequency of six hours to provide near-real-time weather information. The spatial resolution for assessing cloud cover is 150 kilometers. Meteosat data are generally available for free. Derived products are used for weather forecasts and to assess average rainfall. Additional satellites of potential use for UN humanitarian organizations include the recently launched EROS-1A (1.8-meter resolution), QuickBird (0.61-meter resolution) and SPOT 5 (2.5-meter resolution, launch planned in 2002) satellites.
AGENCY WITH EXPERIENCE
SPATIAL PIXEL RESOLUTION (M)
REVISIT FREQUENCY (DAYS)
DELIVERY TIME (DAYS)
Landsat 7 ETM+
UNHCR, UNICEF, FAO, UNDP
600 for all channels
Landsat 4-5 TM
UNHCR, UNICEF, FAO, UNDP
2,500 for all channels
UNHCR, FAO, UNDP
26, but shorter frequency possible
170 for all channels
UNHCR, UNICEF, FAO
24 (12 for C/D couple)
0.5 (two satellites)
NOTE: IR = Infrared, MS = Multispectral, Pan = Panchromatic, SWIR = Short Wave Infrared, VIS =
Landsat 7 ETM+
Proven record; multiple applications; matches previous Landsat historic data records for change assessments; inexpensive data; no copyright restrictions for sharing data.
Long (16-day) revisit interval; obstructed by clouds.
Landsat 4-5 TM
Proven record; multiple applications; historic data record available; matches previous Landsat data for change assessments.
Relatively expensive data; copyright restrictions for sharing data; long (16-day) revisit interval; obstructed by clouds.
Proven record; multiple applications; in-orbit programming possible; historic data record available.
Relatively expensive data; copyright restrictions for sharing data; obstructed by clouds.
Global daily coverage; provides aggregated 10-day average of global vegetation cover.
Relatively short history of data (available since 1998); copyright restrictions for sharing data; obstructed by clouds.
Relatively detailed imagery.
Relatively expensive data; copyright restrictions for sharing data; obstructed by clouds.
Detailed imagery; historic data available.
Original data not in digital form (photographic film); relatively long delivery time and expensive data; obstructed by clouds.
Very detailed imagery; in-orbit programming possible.
Relatively expensive data; copyright restrictions for sharing data; obstructed by clouds.
Not affected by cloud cover and poor illumination conditions; relatively detailed imagery available; wide scenes possible.
Relatively expensive data; copyright restrictions for sharing data; radar imagery can be difficult to interpret; limited classification possibilities.
Independent on cloud cover and illumination conditions.
Relatively expensive data; copyright restrictions for sharing data; radar imagery can be difficult to interpret; limited classification possibilities.
Data available since 1978 at low/no cost.
Data available since 1978; near real-time delivery; NDVI vegetation index highly robust; data available at low/no cost.
Coarse resolution; obstructed by clouds.
Technological developments in the EO satellite industry, such as more detailed imagery and online distribution mechanisms, are proceeding very rapidly. For the UN humanitarian organizations to benefit from these developments and be able to cooperate with other actors in the humanitarian relief community, such as NGOs and the military, they must follow developments and test new types of sensors and products as they become available. Although most UN humanitarian organizations already know what kind of information they need to support their mandates, they should continuously explore more accurate, more timely, and more cost-effective ways to do so. Developing closer links to satellite data providers as well as participating in interagency forums -- such as the UN Office for Outer Space Affairs (UNOOSA) Scientific and Technical Sub-Committee of the UN Committee on Peaceful Uses of Outer Space, the UN Geographic Information Working Group (UNGIWG), and the GIST -- would facilitate this effort.
Need for policy developments
The trend in the remote-sensing industry is toward higher
resolution imagery: NOAA has granted a 0.5-meter resolution commercial
license to Space Imaging,
and EarthWatch launched its 0.61-meter resolution satellite in October
2001. 16 Accordingly,
the United Nations should expedite developing policies and procedures
for handling what some actors may argue to be sensitive information. The
current international legal framework clearly allows for purchase of very
high-resolution satellite imagery over any country. 17
However, some states or actors may object to being imaged at this level
of detail, citing national security interests. Although this is a potential
concern, it should by no means limit the use of such techniques to derive
information that will meet the needs of UN humanitarian organizations.
The images themselves do not necessarily need to be shared with a wide
audience, as long as the derived information gets to the right decision
Commercial and military sources of data
Both the recent UN General Assembly's "Report of the Panel on United Nations Peace Operations" (A/55/305 - S/2000/809, also referred to as the "Brahimi report") and "Strengthening of the Coordination of Emergency Humanitarian Assistance of the United Nations: Report of the Secretary-General" (A/55/82 - S/2000/61) acknowledge the importance of using modern technologies, including GIS and related tools such as satellite imagery, during complex and humanitarian emergencies. Military sources of imagery are often preferred to commercial sources because they offer greater detail and in some cases the images have already been interpreted and are ready to use, as in Kosovo (Bouchardy 2000). Although military data are important for many humanitarian operations, it is not a given that the United Nations will receive military imagery of certain "hot spots." Moreover, commercial satellite images should also be used for two important reasons: first, to ensure information transparency among all the UN organizations that will be working together during an operation; second, to build stronger, more self-sufficient UN organizations that use satellite-derived information when needed. In every case, UN organizations should have routine knowledge about what imagery is available commercially and from the military; the respective image details (see figure 1); the potential limitations on distribution (e.g., copyright and classification); and the delivery times. Mechanisms for sharing commercially available imagery or information derived from it among UN humanitarian organizations should already be in place before an emergency occurs.
Lower-priced data yield higher
The new pricing policy of Landsat data -- as described earlier
-- has already produced some interesting results. Figure 9 shows the U.S.
Geological Survey (USGS)
numbers of individual scenes sold for the three various types of Landsat
data 19 for fiscal years
1999 and 2000. The total number of Landsat 7 scenes sold in FY 2000 was
more than twice the total number of scenes sold for all its predecessors
(Landsat 1 - 5). 20
Figure 9: U.S. Geological Survey (USGS) number of individual scenes sold for the three types of Landsat data for FY 1999 and 2000. Sales figures from USGS.
Figure 10: Sales of Landsat data in millions of U.S. dollars for FY 1999 and 2000. Sales figures from USGS.
With respect to both the number of individual scenes sold and actual income, these numbers show the drastic increase in sales of Landsat 7 data compared with sales of Landsat 1-5 data. USGS's FY 2000 sales income from Landsat 7 was more than twice that from the other Landsat series. Of course, this figure does not take into consideration the increased cost of processing more individual scenes. User-friendly Internet search engines have recently lessened the workload for the data providers in this respect. Customers can now specify the exact scene identifier that corresponds to their image of choice, a task previously carried out manually by the data providers. The sales results shown in figures 9 and 10 can be attributed to the improved quality of Landsat 7, with its fifteen-meter resolution panchromatic (black and white) channel, low cost ($600 per scene for international coverage), and, most important, free sharing of data (no copyright). When compared with previous Landsat satellites, with their thirty-meter resolution, high cost ($2,500 per scene, recently dropped from $4,500 per scene), and strict copyright limitations, it is no wonder that Landsat 7 is popular. The success of Landsat 7 proves that the users are willing to invest in satellite imagery, but not at any cost.
New initiatives for improved use of EO data
The UNHCR GIMU, in coordination with OCHA and in the framework
of the GIST, has recently implemented a new initiative for increased use
of Landsat 7 - derived information for the benefit of the humanitarian
community. This work includes developing and updating a web site
in cooperation with the University of Georgia (U.S.) and its Information
Technology Outreach Services (ITOS), dedicated to the free sharing of
Landsat 7 satellite imagery as part of a broader data repository. The
web site stores a significant amount of useful Landsat 7 imagery, which
will be regularly updated for the benefit of the global humanitarian relief
community. The site is expected to contribute more quality geographic
information and, accordingly, to increase the use of EO technology within
the humanitarian community. 21
Another recent development is the U.S. National Aeronautics
and Space Administration's (NASA's) release of free Landsat 7 data for
the whole African continent in cooperation with the U.S. EarthSat Corporation.
Although these data are not at the original spatial resolution and do
not allow for multichannel classifications, this initiative is welcomed,
and, more important, it indicates that government agencies see a benefit
in making imagery available to the public at low cost. 22
Furthermore, the European Space Agency (ESA), the French Centre National d'Etudes Spatiales (CNES), and the Canadian Space Agency (CSA) have signed a "charter on cooperation to achieve the coordinated use of space facilities in the event of natural or technological disasters." If a natural or technological disaster strikes and the affected government requests aid, ESA, CNES, and CSA can, through this charter, activate their significant technological resources to provide satellite imagery and derived products to assist in the relief operation. This charter has the potential to become a very useful framework for using satellite imagery in humanitarian relief, although it is currently designed for bilateral implementation and focuses on natural and technological disasters only.
UN humanitarian organizations should strengthen their associations
with space agencies and satellite data providers in order to improve geographic
data collection during emergencies. A potential forum for presenting specific
requirements and areas of collaboration is the Committee on Earth Observation
Satellites, of which the main space agencies and related organizations
are members or associated members.
Conclusions and Recommendations
With more satellite sensors in orbit, which will bring down prices and promote less constrictive copyrights, and with enhanced computer power and more user-friendly analysis software, EO data are positioned to become a routine part of operational decision making to support humanitarian activities. UN organizations should unify their combined efforts to include and maximize the benefits of this tool for humanitarian assistance.
Based on this review, the following specific sensors are
best suited to meet the diversity of geographic information needs of the
UN humanitarian organizations:
Landsat 7 for its quality, low cost, and liberal copyright
Ikonos, and similar very high-resolution satellite sensors, for their detailed imaging capabilities and in-orbit programming for shorter revisit-cycle intervals
SPOT for its in-orbit programming for shorter revisit-cycle intervals
Radarsat for providing imagery over cloud-covered regions
NOAA AVHRR and SPOT Vegetation for their proven record and near real-time delivery of early warning information
These recommendations do not exclude the use of other sensors. In fact, users of specific applications should monitor the development of satellite imagery sensors to ensure maximum benefit of EO data for their particular purposes. The recent financial success of the Landsat 7 program suggests that other data providers have set their prices too high and should lower their price per image in order to increase their overall sales income. In the meantime, because UN humanitarian organizations frequently use satellite imagery, they could act as a single group in offering commercial satellite data providers an agreement in the form of an MOU to purchase images as an umbrella organization. Such an MOU could mean significant reductions in the price of imagery for the UN. It would also open the door to data sharing among these organizations instead of relying on the not-so-quick turnaround of a central satellite data procurer.
Finally, to fully integrate this technology as a standard operating tool among even the most experienced UN organizations, decision makers within the United Nations need to understand its benefits and limitations. The standardized nature of remotely sensed images facilitates a streamlined transfer of information from one UN agency to another. Sharing the product of this technology can therefore act as a catalyst in a much-needed interagency collaboration process.
Rapid technological advancements in the field of satellite remote sensing require that the actors be aware of sensitivities associated with this type of information. Accordingly, as often as possible, organizations should take advantage of the possibility of using civilian sources of data, which allows transparency among the organizations and the public at large. Technological tools such as satellite imagery, analysis software, and Internet distribution -- that is, virtual information mechanisms -- support transparent and efficient humanitarian responses. The UN humanitarian organizations need to ensure that these tools are used to their fullest advantage.
The author wishes to thank the following persons for their valuable input to this paper: Jean-Yves Bouchardy (UNHCR), Pablo Recalde and Rhonda Davis (OCHA), Andrew Nadeau and John Latham (FAO), Dennis King and Zeynep Gungor (UNICEF), Steve Ebener (WHO), Giorgio Sartori (UNDP), Jeffrey Marzilli (WFP), and Brett Lien (USGS).
Finally, I am most grateful for the support provided by the United States Institute of Peace and its Virtual Diplomacy Initiative, especially Sheryl Brown, Margarita Studemeister, and Suzanne Wopperer.
Bjorgo, Einar. "Aiding Refugee Relief."
Imaging Notes 14 (1999): 20 - 21.
Bjorgo, Einar. "Using Very High Spatial Resolution Multispectral Satellite Sensor Imagery to Monitor Refugee Camps."
International Journal of Remote Sensing 21 (3) (2000a): 611 - 616.
Bjorgo, Einar. "Refugee Camp Mapping Using Very High Spatial Resolution Satellite Sensor Images."
Geocarto International 15 (2000b): 77 - 86.
Bjorgo, Einar. "Supporting Humanitarian Relief Operations." In
Commercial Observation Satellites: At the Leading Edge of Global Transparency, edited by John C. Baker, Kevin M. O'Connell, and Ray A. Williamson, pp. 403 - 427. Santa Monica, Calif.: RAND and American Society for Photogrammetry and Remote Sensing (ASPRS), 2001.
Bouchardy, Jean-Yves. "Winning the Peace in Kosovo."
GEO-Europe 9 (July 2000): 36 - 41.
Bouchardy, Jean-Yves. "Development of a GIS System in UNHCR for Environmental, Emergency, Logistics and Planning Purposes." Environment Unit Report 1-95, Division of Programmes and Operational Support (DPOS) UNHCR, Geneva, Switzerland, 1995.
Bouchardy, Jean-Yves. "L'information géographique au sein du HCR."
Bulletin du Comité Français de Cartographie 151-152 (1997): 55 - 62.
Dehqanzada, Yahya A., and Florini, Ann M. Secrets for
Sale - - How Commercial Satellite Iimagery Will Change the World. Washington, D.C.: Carnegie Endowment for International Peace, 2000.
King, Dennis and Maxx Dilley. "How to Share Information in a Complex Emergency,"
Humanitarian Affairs Review, Summer 2001, pp. 10 - 13.
United Nations. "Treaty on Principles Governing the Activities of States in the Exploration and Use of Outer Space, Including the Moon and Other Celestial Bodies" a/res/222(xxi) January 27, 1967.
United Nations. "Principles Relating to Remote Sensing of the Earth from Outer Space" a/res/41/65 December 3, 1986.
List of Internet Addresses (urls) for further information on specific earth observation satellites described in this document, as of March 2002.
Landsat 4-5: http://www.earth.nasa.gov/history/landsat/
SPOT HR: http://www.spotimage.com/home/system/introsat/welcome.htm
SPOT Vegetation: http://www.spot-vegetation.com/
DMSP OLS: http://www.ngdc.noaa.gov/dmsp/descriptions/doc_ols.html
NOAA AVHRR: http://noaasis.noaa.gov/NOAASIS/ml/avhrr.html
SPOT 5: http://www.spotimage.com/home/system/future/spot5/welcome.htm
1. As described by OCHA on the ReliefWeb site: www.reliefweb.int/contacts/dirhomepage.html.
3. Committee on Earth Observation Satellites (CEOS). The Use of Earth Observing Satellites for Hazard Support. A report of the CEOS Disaster Management Support Group. See http://disaster.ceos.org/2000Ceos/progress/overview.html.
4. White House press release, 1 May 2000: Statement by the president on the decision to stop degrading global positioning accuracy. See www.ngs.noaa.gov/FGCS/info/sans_SA/docs/statement.html.
5. The GIST consists of OCHA, UNHCR, FAO, WFP, UNICEF, World Bank, U.S. Agency for International Development Office for Foreign Disaster Assistance, and European Union European Commission Humanitarian Aid Office/Joint Research Centre.
6. Satellite images for purely illustrative purposes can also be delivered without geographic coordinates, but this should be avoided for practical use of the data within UN humanitarian organizations.
7. For more details on the development of standards for GIS applications, visit the OpenGIS Consortium Web site, www.opengis.org.
10. Remember that a satellite image with original pixel sizes representing an area on the ground of 10 m by 10 m (100 m2) has 100 times less detailed spatial information than that of an image with a 1 m by 1 m (1 m2) pixel size.
11. More information on the Federation of American Scientists and its Public Eye initiative can be found at www.fas.org/eye/.
12. U.S. Department of Commerce, National Oceanic and Atmospheric Administration, Application to Operate a Commercial Land Observation System, Section B, part 1. Washington, D.C, 1994.
13. In addition to the humanitarian organizations listed here, both the UN Environment Programme (UNEP) and the World Meteorological Organization (WMO) have extensive experience in the use of satellite imagery, for which some applications would be considered humanitarian. However, UNEP and WMO are not described here because they are not primarily humanitarian organizations, that is, they are not listed as such by OCHA (see footnote 1).
14. The scene size for cost/scene is computed for a quadratic scene with the same dimensions as the scene width listed in column 3 (unless otherwise indicated).
15. Radarsat: Only ranges are listed (minimum to maximum swath width and corresponding price and resolution). More options are available (see Radarsat link in appendix A).
16. See http://www.spaceimaging.com/newsroom/releases/2001/halfmeter_license.htm.
17. See http://www.digitalglobe.com/?goto=news
18. Note that the U.S. data providers of very high-resolution satellite imagery are not allowed to sell detailed imagery displaying Israeli territory as part of the license granted to them by U.S. authorities. This is an internal U.S. limitation on data availability, more strict than the international framework, and does not apply to non-U.S. systems. See U.S. Congressional Record, 1996 Conference Report on H.R. 3230, National Defense Authorization Act for FY 1997: Sec. 1064. Prohibition on Collection and Release of Detailed Satellite Imagery Relating To Israel, as quoted below:
19. USGS is the main supplier of Landsat satellite data, and one can assume that other suppliers will have similar results.
20. Landsat 1-3: MultiSpectral Scanner (MSS), Landsat 4-5: Thematic Mapper (TM), Landsat 7: Enhanced Thematic Mapper Plus (ETM+).
22. See http://zulu.ssc.nasa.gov/mrsid/
for more information.
About the Report
Since 1998 the United States Institute of Peace's Virtual Diplomacy Initiative has examined the role of remote sensing and geographic information systems in preventing, managing and resolving international conflict through its Open Eyes project (/oc/vd/oe99/oe99.html).
Also, the Institute's Grant Program has devoted attention to such innovative tools for conflict management as remote sensing. When he was student at Norway's Nansen Environmental and Remote Sensing Center, Einar Bjorgo won a grant from the Institute to develop and evaluate products and procedures for enhancing the utility of high-resolution satellite images in humanitarian relief operations around the world. Today, he writes from the vantage of a practitioner of both. The Virtual Diplomacy Initiative is delighted to publish Bjorgo's overview of the current and potential uses of satellite imagery within the UN humanitarian organizations. The report illustrates the wide range of applications of earth observation data among this community.
About the Author
Einar Bjorgo, Ph.D., works in the Geographic Information and Mapping Unit (GIMU) of the Population and Geographic Data Section (PGDS) of UNHCR, which provides worldwide operational support to UNHCR's relief operations. Bjorgo is involved in several international projects developing practical applications of space-based technologies for humanitarian relief and georgraphic information management and has published on the use of commerical remote sensing technologies in humanitarian relief operations.
Virtual Diplomacy Initiative Homepage | About the Initiative | What's New | Activities | Grants
Publications | Electronic Discussion Group | Contact Virtual Diplomacy
Institute Home | Education & Training | Grants & Fellowships | Policy Research | Library & Links
Publications | News & Media | About Us | Events | Resources | Jobs | Contact Us Site Map
United States Institute of Peace -- 1200 17th Street NW -- Washington, DC 20036
(202) 457-1700 (phone) -- (202) 429-6063 (fax)