IN 1972, just after the launch of the first fully functional remote sensing satellite, Landsat 1, the world鈥檚 leading satellite experts gathered in Alice Springs to celebrate the dawn of a new era. They talked about how they could turn remote sensing into a viable business.
Twenty-three years later, they are still puzzling. Although the power and performance of the satellites has improved dramatically, and remote sensing has become commonplace, it is still not a commercial proposition. And according to Ken McCracken 鈥� a joint winner of the 1995 Australia Prize for his contributions to remote sensing while at the CSIRO, and widely regarded as the father of the field in Australia 鈥� it may never become so. The reason he gives is that satellite imagery helps society in a general way rather than giving a direct boost to the profits of any one individual or company. 鈥淚ts value is enormously diffuse,鈥� says McCracken. 鈥淪atellite sensing tells you where to look 鈥� whether it be for blue-green algae or a particular rock formation. Then you can bring in much more focused devices, which are carried on aeroplanes. Airborne remote sensing is used only when someone needs it, and the majority of customers pay upfront.鈥�
Remote sensing from satellites has made a profound difference to meteorology and our ability to monitor the environment. It has underpinned a new understanding of geology. It is even having an impact on the way farmers husband their land. Satellite images are routinely used by mining companies, local councils, water managers and researchers. But none of these users pay what it costs to provide the data, and remote sensing is heavily subsidised by governments worldwide.
Advertisement
In Australia, for example, a government body 鈥� the Australian Centre for Remote Sensing (ACRES) 鈥� is responsible for obtaining and disseminating almost all the satellite data used across the country. Each year, ACRES outlays about A$10 million to gather information, but it recoups only about A$2 million from sales.
Government departments and instrumentalities make up about 70 per cent of the total Australian market for remote sensing products. The mining industry buys almost all the rest, although a small percentage is used in agriculture.
ACRES dominates the Australian marketplace. The only other important suppliers are the European Space Agency (ESA) and the French company Spot Image of Toulouse, which both provide small mounts of data direct to a few Australian customers.
Beyond that, the Australian remote sensing industry comprises about 20 consulting companies, which buy images from ACRES, enhance them, and pass them on. Most have a staff of fewer than five, and a small client base. There is also a fledging instrument manufacturing sector. In recent years, Australian companies have built equipment for ground stations (鈥淒own to earth鈥�), supplied instrument parts for use on ESA sateellites, and developed a working model of a device for meterologists that measures barometric pressure across the full extent of the Earth鈥檚 atmosphere.
But why do they bother? Would Australia not be better off if it shut down the industry and put the money it spends on remote sensing into general government revenue? It is not hard to find people in the finance ministries in Canberra who argue that way.
Perhaps surprisingly, remote sensing in Australia is relatively healthy compared with the rest of the world. The costs are minimal because the country does not have to find billions of dollars to launch and maintain satellites. It only spends a relatively small amount each year on space instrumentation.
The benefits for such a large and sparsely populated country as Australia are potentially enormous, especially considering that two of its major industries 鈥� mining and agriculture 鈥� rely heavily on the analysis and management of land.
This growing reliance on remote sensing is reflected in the healthy position of ACRES itself. 鈥淩evenue has been growing at about 17 per cent a year over the past four years,鈥� says Dennis Puniard, the director of marketing for ACRES. 鈥淎gainst the background of a major recession, growth rates above this have been rare in almost any business field.鈥�
Remote sensing has become part of Australia鈥檚 infrastructure 鈥� just like roads, railways and hospitals. Satellite images, for instance, have been at least partially responsible for every significant mining discovery in the past 20 years. While no mining company will ever admit that remote sensing was responsible for any particular find, rumours abound, says McCracken. And the proof is in the amount of money that such companies pour into remote sensing. 鈥淣o exploration team goes out into the field without studying Landsat images first,鈥� he says.
Satellite sensing has also underpinned much of our new understanding of the Australian environment. At CSIRO鈥檚 Centre for Arid Zone Research in Alice Springs, Geoff Pickup leads a team of researchers which is trying to understand what is happening to the fragile lands in the outback. Distances there are so vast, the population so sparse, that collecting even the simplest information is difficult.
鈥淔irst, we wanted to know the distribution of plants and animals. Then we needed a picture of the stresses they experience 鈥� such as those from introduced weeds, farm animals and drought,鈥� says Pickup. The only practical way to tackle such problems was to use satellite images. 鈥淔or example, we were able to use the patterns of change in vegetation derived from satellite images to produce accurate estimates of the impact of cattle grazing,鈥� he says.
Or consider a case much closer to home for many of Australians. The effluent produced by the five million people in the greater Sydney area is now pumped several kilometres offshore in giant underground pipes. Satellite images have been used by the New South Wales Environment Protection Authority to monitor the behaviour of this plume of pollution as it is dispersed on the current.
Mark Howden, a consultant who worked on the project, says conventional sampling techniques would not have been given a wide enough view of Sydney鈥檚 coastal waters. 鈥淪uch a view was essential to understanding the external influences on the waters in the immediate vicinity of the outfalls,鈥� he says.
Weather forecasting in Australia is also now highly dependent on remote sensing techniques (鈥淲inds of change鈥�). Because of the huge distances and the small, dispersed population, obtaining data using ground-based sensors is difficult and expensive. This is true both for inland Australia and the Southern Ocean, both of which greatly influence the country鈥檚 weather. 快猫短视频s at the Bureau of Meteorology in Melbourne now claim to be on a par with the world鈥檚 best interpreters of satellite images.
Essentially, weather-data is obtained from two major sources: the Japanese Geo-stationary Meteorological Satellite (GMS-4) and the polar orbiting satellites of the US National Oceanic and Atmospheric Administration (NOAA). GMS-4 is parked directly over the equator just north of Australia and an image is received from this satellite every hour. The NOAA satellites orbit once every 12 hours and so pass over at the same time every day. These satellites are much lower than GMS-4, and give a much more detailed picture of what is happening. They also carry atmospheric sounders, which enable the meteorologists to obtain three-dimensional views of activity in the atmosphere.
By and large, weather forecasting in Australia now consists of plugging satellite data into a computer model of the atmosphere and running the model forward in time. The model contains data from the previous 24 hours, and is constantly updated as new data comes in.
Remote sensing has been taken up on a wide scale in Australia. A recent survey by the industry magazine GIS User estimated that remote sensing was used in at least 250 government departments and instrumentalities across Australia. It is also used by almost every division of CSIRO.
Indeed, the CSIRO has such a commitment to remote sensing that it has an office of Space Sciences and Applications in Canberra to co-ordinate its effort.
Commercially, a great deal of interest is now focused on using remote sensing in agriculture. 鈥淚f the market were to consist only of 10 per cent of the landowners with more than 500 hectares,鈥� says Brian Button from Agricultural Reconnaissance Technologies, 鈥渋t would take my company 25 years to service them all if they bought just one image a year.鈥�
But in spite of a plethora of farm management software, and the urgings of a flock of computer salesmen, farmers are notoriously loath to use computers. David Hodgkinson from the New South Wales Farmers鈥� Federation says straw polls indicate only 10 per cent of farmers use a personal computer to run their farms.
Norm Campbell at the Leeuwin Centre for Earth Sensing Technologies in Perth runs training courses for agricultural officers in the Department of Agriculture. He says it will be five years before the technology really takes off in rural areas. 鈥淭here are problems of distribution that need to be solved, as well as questions of how to present information in the most meaningful way.鈥� But more fundamental than this is spreading awareness of remote sensing, and convincing farmers that it can increase productivity.
As long as single satellite images are priced at around A$10 000 it will be difficult to convince farmers to use them. However, by using standardised procedures, suppliers like ACRES are able to offer farmers images of their farms for only a few hundred dollars. At the same time, the cost of the computer hardware necessary to run and interpret remote sensing images is dropping rapidly.
There is further cause for optimism. Technologies are now under development which will make remote sensing even more useful. For example, hyper-spectral imaging, in which sensors can make extremely fine discriminations between different substances, should help both geologists and environmental scientists. And improved resolution offers the potential for even more profound changes.
At present, the most powerful sensors in general use can discriminate between two points 10 metres apart in black and white and 30 metres in colour. Since the end of the Cold War in the early 1990s, both America and Russia have begun to contemplate putting 鈥渟py quality鈥� imagery, with a resolution of one metre or so, on the market. At that level of resolution, space-based remote sensing could offer comparable products to aerial photography for a fraction of the price.
And there is a large potential market for it 鈥� local government. City and shire councils are all about land use 鈥� planning, roads and parks for example. Ipswich Council in southeast Queensland has already installed a geographic information system based on 1-metre resolution aerial photographs of the entire council area. It was an expensive exercise, but council workers now have maps which show everything in the city, from power poles to trees to backyard swimming pools. This is a great boon to the efficiency of almost every council activity from providing up-to-date property plans to locating manhole covers.
While most of the next generation of space-based, high-resolution sensors will be built and operated in America and Europe, one Australian company is attempting to enter the market. Spatial Information Services Australia (SISA) in Adelaide, a joint venture of several Australian information technology companies, plans to buy imagery from overseas supplies, enhance it, and make it available to customers in Australia and perhaps Southeast Asia.
SISA鈥檚 project director, John Douglas, says his company has just completed a survey to establish whether or not a market for high-resolution data existed. He refused to give details but added: 鈥淚鈥檓 still here, so you can draw your own conclusions. There are no significant technical issues to be overcome.鈥� The company鈥檚 intentions represent a clear change in attitudes on the part of the private sector. It is one that echoes what was first mooted with enthusiasm way back in 1972 and then abandoned. We may not have a sense of d茅j脿 vu after all.
Down to earth
OUR knowledge of what goes on in the ocean that lies between Australia and Antarctica will be vastly improved when the Tasmaninan Earth Resources Satellite Station (TERSS) becomes fully operational in June. One of its main functions will be to download data from radar satellites, but it is also capable of taking data from any of the other satellites around.
Hailed as an example of the next generation of ground stations for receiving satellite data, TERSS cost less than A$2.5 million to build, about a third of the price of a conventional station. It can receive and store massive amounts of data at high speed 鈥� 13 million bytes a second 鈥� and will primarily download data from radar satellites.
TERSS is almost fully automatic and only requires the attention of one person for a few hours a week to change its data storage tapes. Its scheduling can be controlled from anywhere in the world via the Internet.
By comparison, a conventional ground station requires at least two operators to be on hand whenever it is receiving data. And whereas ground stations traditionally record data on giant tape decks called instrumentation recorders, TERSS uses a hard disc array.
How to draw a blank
ANTARCTICA is a difficult place to map. Most of the modern maps of this huge continent are based on aerial photography, but this technique suffers from one major drawback. It is not easy to photograph a landmass that measures about 14 million square kilometres, is predominantly white and is often shrouded in cloud. In particular, the shape and height of topographical features are hard to distinguish.
But a team of Australian cartographers claims to have created the first accurate maps of the Australian Antarctic Territory (AAT), which covers 5.89 million square kilometres or 42 per cent of the Antarctic鈥檚 landmass. Rupert Summerson from the Austalian Antarctic Division in Hobart, Laurie Oliver from the Australian Centre for Remote Sensing in Canberra, and Noel Ward from the Australian Survey and Land Information Group, also in Canberra, joined forces for the project. They used wide-angle photographs from Russian satellites combined with radar readings from ERS-1, a European radar satellite.
In many countries, space imaging is usually conducted from long-lasting, electronic sensors costing billions of dollars. Russia, however, used film cameras on board cheap, short-duration satellites. One of them, Kate-200, had regularly taken photographs of the AAT.
Film gives a much better resolution than electronic scanners because film emulsion has many more pixels than even the most sophisticated scanner. More importantly, each image captured by the Kate-200 camera covered an area of 250 square kilometres, increasing the chances of providing at least one recognisable Antarctic feature. The film had also been shot with considerable overlap so that 60 per cent of the one image was repeated on the next. This means it is possible to make up stereo pairs from which height information can be derived, if the images contain enough surface detail. The film images were digitised in Moscow so that they could be used by the computerised cartographic system in Australia.
As a result maps have been produced that are accurate to 30 metres or less. Those completed so far include the coastline, the areas around the Australian bases of Mawson, Davis and Casey, and most of the mountainous area of the AAT.
But there are still many blank spaces left on the maps where the whiteness of the landscape has defeated the satellitte camera. These are being filled in using radar.
Winds of change
SATELLITE images 鈥� like the onee at right from the Japanese Geostationary Meterological Satellite GMS-4 鈥� have revolutionised the art of weather forecasting in Australia. For the first time, meterologists have been able to obtain a regular coverage of climatic conditions over the whole of Australia and surrounding seas rather than relying on sporadic information about temperature and wind collected by aircraft weather balloons and ships.
The image depicts wind patterns. This one was taken at 5 am Greenwich Mean Time on 25 February this year 鈥� about 36 hours after Cyclone Bobby hit the coastal town of Onslow, about 1500 kilometres north of Perth. The cyclone, as it travels south along the coast of Western Australia, is represented by the swirl of red arrows.
The yellow, blue and red arrows show wind speed, direction and strength recorded at three heights in the clouds above Australia 鈥� up to 3 kilometres for the yellow arrows, between 4 and 7 kilometres for blue arrows, and between 8 and 13 kilometres for the red arrows.
Data from the GMS-4 satellite is sent every hour to antennas on top of the Bureau of Meterology building in Melbourne and fed into a computer model. Four times a day the satellite also sends an extra image on the half-hour, so that a total of 28 images are received every day. To calculate wind speed and direction, the computer tracks movements over time of about 2000 clouds.
At the moment, this technique depends on cloud cover 鈥� no clouds means no images of wind. To overcome this, GMS-5, which will be launched this year, will provide images of water vapour which can be used by meteorologists to measure winds in cloudless areas.