91av

FLOODS, EARTHQUAKES and volcanic eruptions strike with depressing frequency
and usually little warning. Drought and famine develop over months and years,
yet details tend to emerge too late to avert death and suffering. But remote
sensing by satellite can give us data that can warn of imminent disaster
or provide effective emergency relief for remote areas. Depending on their
orbits, satellites can produce images of the same area at intervals ranging
from a few hours to a few weeks. Those images can pinpoint the location,
extent and severity of sudden disasters, but long term monitoring, over
months and years, can document changes that foretell drought and famine.
Analysts can map most of the vegetation and, to some extent, the soils and
rocks that cover the Earth’s surface, as the images change through the year.
The new understanding of the dynamics of the environment that the images
bring can help people in the developing world to identify potential resources
and plan sustainable development.

Remote sensing evolved through military intelligence, starting in the
American Civil War with the use of cameras on balloons to photograph trenches.
The Second World War and the Cold War of the 50s and 60s forced the pace
in research and development, particularly in ways to distinguish between
camouflage and growing vegetation. Plants look green to us because they
absorb red and blue light, and reflect green light strongly, but they reflect
infrared radiation more strongly, giving analysts a clear pattern from which
to pick out living vegetation . The military need for surveillance at night
and in all weathers led to the development of heat detectors and imaging
radar, that can ‘see’ through clouds. Once engineers had developed electronic
techniques to record energy coming from the Earth’s surface, satellites
began to send back an almost continuous flow of information.

Nowadays useful images come from the US Landsat, the French SPOT, a
number of lower resolution weather satellites and, increasingly, from satellites
operated by Japan, India and the Soviet Union. Costs vary widely from much
less than a penny per square kilometre to about 50 pence, depending on the
amount of processing and interpretation required.

The potential of remote sensing for research has led governments to
make military hardware and techniques more widely available. Secrecy now
shrouds only the degree of detail available from military satellites, and
precisely where such spies in the sky are looking at any particular time.
But there are still limits on the sale of images to the public if they reveal
features in which there is a security interest. The success of global information
gathering by the USA, France, the Soviet Union and Japan has spurred several
developing countries, including Brazil, China and Indonesia, to plan their
own remote-sensing satellites. India has taken the first independent step,
and in 1988 launched its own system (IRS-1) that rivals Landsat and SPOT
in the quality of its images.

Meteorologists have tracked hurricanes routinely since the launch of
the first weather satellites in the 1960s, illustrating the way in which
forewarning can reduce the cost and intensity of efforts needed to relieve
disasters of this type. SPOT images, which include information about topography,
help researchers to find the areas most prone to flooding, as well as the
places that offer the safest haven as the floods continue. Geologists know
which regions are prone to earthquakes and eruptions, but precisely where
and when a disaster might strike is still a hit or miss affair. With the
help of remote sensors, researchers can detect the heat emitted by volcanoes
in the months before an eruption, pinpointing those to beware of among the
hundreds that lie dormant. This method has proved itself in the volcanoes
of the Andes.

But how can remote sensing help if disaster strikes unannounced? In
areas for which there are inadequate maps, satellite images can pick out
routes that trucks can take to distribute supplies. A crucial part of effective
relief after, for example, floods, is finding out exactly which areas have
been affected in the first few days.

But despite the possibilities, and the claims of speedy delivery by
those who distribute data from satellites, this sort of information has
rarely if ever helped to direct relief operations immediately after a disaster.
Often, satellites cannot help: the Nile floods of 1988 were hidden by clouds,
the toxic gas release in Cameroon in 1987 had no tangible form. But in many
other cases useful images could have been available, but distributors failed
to live up to their claims. They can deliver on time: Landsat images helped
British troops to find their way in the Falklands War, and within days of
the Chernobyl accident in 1986, Eosat, which distributes Landsat data, released
computer-enhanced images of the damaged reactor. When a military satellite
spots a missile a few centimetres larger than agreed limits, or an alleged
chemical warfare plant, the publicity is rapid and assured.

The information carried by satellite images can continue to improve
relief operations for months and even years during a drought. Food and shelter
is commonly available only to those fit enough to make their way to feeding
stations near major towns; the overcrowding that results poses fresh problems
of sanitation and overgrazing by the stock which people bring with them.
If aid workers could send relief food and farming supplies to the small
settlements where they are needed, they could ease this pressure. Conventional
sources of information do not always help: the people who most need aid
are semi-nomadic, their settlements do not appear on maps, and there are
no roads that trucks could use. The sharp resolution of Landsat and SPOT
images, acquired routinely before disaster strikes, could help aid agencies
to locate temporary villages and plot the best routes to reach them across
otherwise unmapped country.

More importantly, satellite images can help to give advance warning
of famine. Weather satellites cover the whole of the planet each day, so
that analysts can monitor the two most important indicators of famine –
rainfall and vegetation. Chris Tucker of NASA and scientists at the UN Food
and Agriculture Organisation (FAO) have found that the infrared radiation
emitted by clouds because of their temperature is a means of assessing rainfall.
The colder the top of a cloud system, the greater the chance that rain is
forming. But this does not guarantee that rain is actually falling. One
of the realities of remote sensing is that observations from afar have to
be matched to reality on the ground. In this case meteorologists must calibrate
the satellite data with measurements of rainfall over a wide area. They
convert these analyses to estimates of actual rainfall, month by month,
across entire continents. The estimates that the FAO now provides for Africa
are sketchy, but at least provide a guide where none was possible before.

Of all the materials on the Earth’s surface, vegetation is the easiest
to spot on satellite images. Data from remote sensors, suitably analysed,
allows researchers to estimate the density and health of the ground cover
at any time of year; they can even show the effects of infestation by pests
and diseases. And because they show the extent of these problems, the images
help make the most of resources needed to stamp them out. Calibrating the
images is once again a matter of matching observations in the field to the
satellite data. Teams at NASA and the FAO now routinely provide monthly
maps of vegetation for Africa and South America based on data from weather
satellites. Jelle Hjelkma of FAO has devised means of using such images
to identify the breeding ground of locusts and to predict when and where
swarms will break out.

Researchers also estimate the likelihood of famine in different areas
by matching sparse records of crop yields to satellite images through the
years. Pam Kennedy of the University of Reading has used seasonal weather
satellite data to chart the variations in grass cover across Tunisia. But
remote sensing is still mainly used in agriculture for fine tuning of production
and even checking the honesty of farmers seeking subsidies for their crops
in the US and Europe.

The major problem in areas prone to drought is not that there is no
rain, but that the rain generally falls in a downpour; most of the water
flows away or evaporates before farmers can use it on their crops. Only
a small proportion soaks in through the surface to form groundwater. More
is lost if the ground slopes steeply, and if the soil is compacted. Plants,
however, tend to open up the soil around their roots, and slow down runoff,
allowing more water to soak into the soil. Most water used in dry areas
comes from wells that tap groundwater, but they often run out during a drought
and as the local population grows. Three possible strategies can improve
the situation: digging or drilling new wells, storing ephemeral stream water
at the surface, and increasing the amount of rainfall that seeps into soil.
Surface water is prone to pollution in hot climates. Floods soon fill reservoirs
with boulders and silt. Capturing water in soil is effective, simple and
cheap. It entails slowing the flow at the surface, to give water more time
to enter the soil. Satellite images show the courses of streams, the steepness
of slopes and the distribution of different types of soil and vegetation
cover; mapping the catchment area and estimating the flow of each major
stream during rains become simple and speedy tasks.

The next step is to locate areas where the torrents slow down naturally
after storms. Cheap engineering methods, such as placing boulders in stream
beds at ‘thresholds’, where the current usually slows down anyway, make
water form ponds. It will then seep into soil, provided that the surface
of the ground has been broken up by ploughing. Shallow trenches filled with
porous sand above the thresholds make the capture of water even more efficient.
It took me 15 days to evaluate the best places for this type of water conservation
across 120,000 square kilometres of Red Sea Province in the Sudan, using
cheap photographs taken from the space shuttle, at a scale of 1:250,000.
These images cost only Pounds sterling 200.

Better ways to find water

A clean, dependable supply of drinking water would improve the lives
of at least a third of the world’s population. Sources at or close to the
surface are usually polluted. The best supply comes from groundwater tapped
by wells, yet finding such a supply in a desert requires slow and costly
surveys by teams of hydrologists and geologists on the ground. Remote sensing
contributes to this search by targeting places where the essential exploration
on the ground and construction of wells stand the best chances of success.
Between 1985 and 1987, UNICEF carried out a $3 million programme to improve
water supply in the Red Sea Hills of the Sudan without geological and remote
sensing input. Around 80 per cent of their wells failed. In the same terrain,
the Eritrean Relief Association, using remote sensing and geological information,
had a failure rate of less than 30 per cent in the same period.

There are two kinds of target when looking for groundwater; places where
thick loose sediments have been able to soak up some of the surface water,
and those where water comes from the bedrock itself. In the first case,
images can distinguish soft sediments from bedrock by their usually brighter
appearance. They can also highlight areas of natural vegetation that are
a common sign of water below the surface. Desert plants, with their deep
root systems, tap this source of water. In one year, a square kilometre
of dense desert scrub uses enough water through transpiration to supply
a village of 1000 people.

Image interpretation to find water supplies from bedrock depends partly
on distinguishing relatively porous sandstones and limestones, in which
cracks opened out by solution allow the free passage of water, from tighter
and drier crystalline rocks, such as granite, using the way they reflect
radiation . Reconnaissance of crystalline rocks, which are less likely to
carry water, centres on identifying features underground that help channel
the subsurface flow of water within these otherwise impermeable rocks. Among
these are zones of shattering made by earthquakes along ancient faults,
which open up the rocks to water. These fracture zones sometimes show up
as linear depressions where the weakened rock has been etched out by erosion.
In contrast, sheets of igneous rock can form natural dams beneath the surface,
because they are formed completely of interlocking crystals with few pores
to let water seep through. The igneous sheets, which resist erosion by wind
and water, form distinctive linear ridges where they reach the surface of
the ground.

But even when analysts have found places that look likely to hold groundwater,
they have no guarantee of a successful water supply. For each potential
well, hydrologists must find out how quickly the water that people use can
be replaced from aquifers in the surrounding drainage basin. A successful
well depends on matching clues from the satellite images with the experience
of local people in finding and using water, and with observations on the
ground. The eventual choice of sites for wells must take account of local
needs and practices – a deep well sunk by mechanised drilling and served
with a complex pump may soon fall into disuse, while a shallow one dug by
local people and in which the water is brought to the surface by traditional
means will be better cared for.

One of the great advantages of remote sensing is that the images contain
information on all aspects of the surface environment. Their use for one
purpose, such as water exploration, inevitably reveals clues for others.
In using Landsat Thematic Mapper images from 1985 to improve the accuracy
of siting wells in the Eritrean Highlands, Seife Berhe of the Open University
and I soon became aware of their great potential for geological mapping
and exploration for minerals. Moreover, by comparing them with similar Landsat
pictures from 1972, we were able to begin documenting the loss of the fragile
vegetation on these highlands, watered by Red Sea mists, that has resulted
from people having to move into the area during the protracted war of the
past years. A month’s work generated information for three crucial aspects
of Eritrea’s recovery from war and famine – likely sites for new water supplies,
suitable soils and terrain for new agricultural development and signs of
metal mineralisation that could increase the country’s income.

The Landsat Thematic Mapper, which records particular spectral information
invisible to the human eye, not only discriminates between common types
of rock, but also picks out the unusual and valuable rocks. It shows up
subtle geological structures, such as folds and faults, that eased the passage
of the fluids that concentrate important metals and hydrocarbons. As well
as showing features that are invisible to the geologist in the field, satellite
images cover vast areas. Although geologists will always need to do fieldwork,
remote sensing can speed up mapping and exploration. Every multinational
oil and mining company has invested heavily in the hardware and data that
give this improved efficiency. Major companies such as British Petroleum
have invested around $1 million on equipment. They spend up to $1 million
per year on images and have specialist teams of between 10 and 25 people.

The bulk of new exploration targets are in developing countries; the
multinationals often know far more about the mineral potential of these
areas than the governments and people themselves. If development in the
Third World is to be controlled by its people, they need the same or better
information about their potential wealth.

* * *

The true and false colours of the Earth

REMOTE SENSING uses the spectrum of radiation between visible light
and microwaves, which has wavelengths ranging from micrometres to centimetres.
Gases in the atmosphere absorb radiation in some parts of the spectrum,
stopping it from reaching the surface of the Earth; the radiation that does
penetrate through ‘windows’ in the atmosphere interests earth scientists
and oceanographers most.

On land, different materials reflect visible light and infrared radiation
in varying proportions because of their molecular structures. At the simplest
level, substances have different colours; they reflect or absorb visible
wavelengths differently – giving light or dark colours and shades in between.
The colours of rocks and soils come mainly from iron compounds, formed during
weathering. To us, rocks and soils rich in iron appear in various shades
of red, brown and yellow; common iron compounds absorb the shorter blue
and green wavelengths leaving mainly red to be reflected into our eyes.
Venturing into the infrared, as satellites now do routinely, geologists
can exploit particularly strong absorption at wavelengths around 0.9 micrometres
(see graph). This happens to varying degrees for different iron minerals.

Although water and hydroxyl ions (OH^–) in minerals produce little effect
in the visible spectrum, at wavelengths between 1.0 and 2.3 micrometres
in the infrared they absorb energy in a number of narrow bands. The most
important of these, between 2.1 and 2.3 micrometres, comes from clay minerals,
which form by breakdown of rock-forming minerals during weathering (see
graph). Thus satellite images can show how strongly rocks reflect radiation
overall, their iron content and where they contain minerals with hydroxyl
ions. This gives rocks and soils distinctive signatures that promise to
revolutionise the hunt for metals and fuels.

But the odd way in which living vegetation reflects energy has the most
widespread use in remote sensing. We all know that most leaves are green.
This is because the pigment chlorophyll, which photosyn thesises sunlight
into energy for the plant, absorbs more red and blue light than it does
green.

But chlorophyll is also very sensitive to temperature. Most things warm
up as they absorb solar radiation in the visible and short-wave infrared.
To stay cool, land plants have evolved a cell structure that reflects almost
all the infrared component away. On the spectrum of reflectance for a leaf,
this is represented by a sharp increase at wavelengths longer than red.
The infrared signal can be combined with that in red and green light making
the red, green, and blue phosphor spots in a television monitor glow in
proportion to the strength of these signals. The resulting image is said
to show ‘false colour’.

The high infrared reflectance of plants shows up on such an image as
shades of red, which increase in intensity where there is more vegetation.
The spectrum of radiation reflected also helps researchers work out what
plants are growing.

Remote sensing no longer depends on radiation from the Sun. The Earth’s
surface can be illuminated artificially by lasers and, more usefully, by
radar. Images of the radar energy reflected by the surface differ from all
others since their brightness depends on how rough the surface is, rather
than molecular properties of the materials that compose it.

In arid areas, radar can penetrate a few metres beneath the surface
revealing details that are otherwise completely hidden, such as buried river
channels and undulations of the surface of the rock underneath sand. In
some cases this can lead directly to sources of water.

Remote sensing frees our understanding of the environment from the narrow
limits of human vision. Researchers combine data from many parts of the
spectrum to produce colourful and discriminating pictures that are customised
to the materials on which information is needed.

Stephen Drury lectures in Earth Sciences at the Open University, and
is the author of A Guide to Remote Sensing (Oxford University Press).