SECTION 1

1. INTRODUCTION

At one time, the success of a fishing trip often depended on a fisherman's keen sense of sight, smell and hearing. To that end the value of a high vantage point, the crow's nest, to scan for fish was appreciated. Modern aircraft and satellites have raised mankind's vantage point to a level undreamed of by earlier generations; at the sametime, devices have been developed which have expanded man's perceptions far beyond the limits of the human senses. The combination of these technologies has resulted in the modern science known as remote sensing which may be defined as the acquisition of information about an object or event without being in physical contact. We are just starting to discover some of the ways in which remote sensing can be applied to man's centuries-old quest to harvest food from the sea.

This manual is intended to be an introduction to the field of remote sensing for persons involved in the study, management or utilization of fisheries resources, particularly in developing countries. Although some forms of remote sensing have been in existence for many years and are generally well understood, extraordinary advances have taken place during the past two or three decades, both in the technology and in its application. The sheer volume of literature now available in this field and of technical knowledge needed to understand it makes an introductory manual of this kind essential. It is beyond the scope of this text to attempt a complete or comprehensive description of modern remote sensing or even to document all of the ongoing research programs and their application to the locating and capture of fish. Rather it is intended to provide the reader with a basic understanding of some of the terms, concepts and specific systems used in remote sensing and, through case studies, to illustrate some applications of importance to fisheries personnel.

1.1 Historical Overview

The term “remote sensing” was coined in 1960 by Evelyn Pruitt of the U.S. Office of Naval Research. The history of remote sensing, however, is considerably older. The first aerial photographs were taken from a captive balloon near Paris in 1858. During the next fifty years significant advances were made in camera design and film emulsions. Photographs were taken from such diverse platforms as kites, rockets and even carrier pigeons. The first photograph taken from an airplane was a motion picture shot over Centocelli, Italy, in 1909, in a plane piloted by Wilbur Wright. Most of this early photography provided an oblique rather than a vertical view of the ground. Popular illustrative pictures of a number of large cities and other scenic attractions were also produced using this means. Scientists, however, recognized the potential of aerial photography as a mapping tool and gradually the science of photogrammetry was developed.

It was not until the First World War that aerial photography was acquired and utilized on a large-scale, systematic basis. Cameras were specifically designed for aerial reconnaissance and associated processing facilities were developed to produce thousands of photographs per day. Equally as important as the technological advances was the development of photo interpretation techniques to obtain intelligence information from the images. By observing the deployment of men and material over a period of time, it was possible for strategists to anticipate military manoeuvres. By the end of World War I, there had been substantial improvement in aircraft, cameras and processing equipment and a relatively large number of people had gained experience in different aspects of airphoto acquisition and utilization.

As improved photogrammetric equipment was introduced in the 1920 's and 1930 's, vertical aerial photography became the standard information source for the compilation of topographic maps. Aerial photography was used to a limited extent by geologists, foresters and planners in Europe and North America and by cartographers and geographers for small-scale geographic studies in Africa and South America. Colour film was first developed during this period, although it saw little aerial application until the Second World War. Several areas of scientific research were also initiated which would later form the foundation of modern remote sensing techniques.

World War II provided another catalyst for rapid technological development inthe field of remote sensing. Once again the acquisition of reconnaissance photography for military intelligence was the primary application. Photo interpretation techniques became highly sophisticated. One aspect of value to later coastal zone studies was the use of aerial photography inplanning amphibious assaults. The water penetration capability of aerial film, particularly colour film, made it possible to obtain reliable information on bathymetry and bottom materials when navigational charts were inaccurate or non-existent. The first colour infrared films were also developed during the war for camouflage detection. Large radar networks were erected in the 1940's to provide an early warning system for aircraft detection. Advances in radar technology permitted the development of smaller transmitting and receiving equipment appropriate for airborne use. Included in this class were plan position indicator (PPI) radars which provided an image of the terrain below the aircraft, independent of weather conditions or light availability. The PPI radar, which was used mainly for target detection during nighttime bombing missions or high altitude bombing through cloud cover, also proved useful for coastal navigation.

In the 1950's thermal infrared imaging systems were developed which provided a “heat picture” of objects or terrain. As with radars, thermal infrared systems are not dependent on light availability but, unlike radar, they are unable to “see” through clouds. In the same period, Side Looking Airborne Radar (SLAR) was developed to improve upon the relatively crude imagery produced by PPI radar. Both of these systems were originally designed for military use and did not become available for civilian application for several years.

The launch of SPUTNIK-1 by the U.S.S.R. in 1957 marked the beginning of the “space age”. In 1959 the U.S. satellite EXPLORER-6 transmitted the first image of the earth as seen from space. The world's first meteorological satellite, TIROS-1, was launched in 1960; this was the forerunner of the more advanced weather satellites which are in place today. Manned space flights were particularly important in creating an awareness of the potential for resource mapping and monitoring from space. The first pictures of the earth, taken by hand-held cameras in space, provided an amazing amount of detail of land and water features over a large area. Later missions by the U.S.A. and the U.S.S.R. carried more sophisticated camera and scanning equipment, specifically for the acquisition of imagery for resource evaluation.

While the manned missions were successful in demonstrating the value of space imagery, they were generally of short duration and did not provide uniform global coverage. These limitations were largely overcome by the development of earth resource satellites such as the U.S. LANDSAT series. Operating in a lower orbit than the meteorological satellites, LANDSAT, and later similar systems, provided greater spatial detail, although less frequent temporal coverage. Although the sensors were chosen primarily for land-based applications, they have proved to be useful for several coastal and marine studies.

In addition to the operational satellites now in place, there have been a number of experimental systems launched which have further demonstrated the value of monitoring the marine environment from space. Prominent among these have been the NIMBUS-7 satellite which carried the Coastal Zone Colour Scanner (CZCS) and the short-lived SEASAT satellite which carried a number of specialized marine sensors. In recent years several nations in addition to the U.S.S.R. and the U.S.A. have launched their own remote sensing satellites.

It should be emphasized that aircraft continue to play a major role in remote sensing because of their flexibility in terms of altitude, scheduling and sensor complement. Depending upon the user's information requirements and available resources, aircraft suitable for remote sensing studies can range from single-engine light aircraft to high-altitude, multi-engine jets. Aircraft are often used as platforms to test sensor designs before the sensors are used onboard satellites.

The rapid advances in computer technology have affected all aspects of remote sensing. Of particular interest to users are the digital processing techniques developed for data analysis. Programs are available for geometric correction, image enhancement and quantitative data extraction. The speed of digital computers allows a single operator to perform data analysis for large geographic areas in a relatively short period of time.

While there continue to be advances in data collection and analysis techniques, the major challenge for remote sensing today is to provide usable information in a timely fashion to those who need it. This requires an infrastructure to access and distribute the data and a user community who can apply it to their programs. Meteorological remote sensing is perhaps the only area of application which can truly be considered operational in this sense, at least on a global basis. The Food and Agriculture Organization (FAO) of the United Nations provides a framework for international cooperation in remote sensing and, through its research programs, publications and training courses, is actively encouraging more widespread and effective use of this technology.

1.2 Basic Terms and Concepts

Before we proceed further, some of the terms and concepts which are used throughout this manual will be explained. Although they will be discussed in greater detail in subsequent sections, they are introduced briefly at this point to facilitate anunderstanding of the text.

Remote sensing may be defined as the acquisition of information about an object or event on the basis of measurements taken at some distance from it. In practice the term is normally used to describe the collection and analysis of data made by instruments carried in or above the earth's atmosphere.

A sensor is a device which detects and measures a physical parameter, such as radiation, and converts it into a form which can be stored or transmitted. In other words, it is the device which “sees” the objects or terrain towards which it is pointed. While devices which sense gravity, magnetic fields or sound waves can properly be classified as remote sensors, many authors restrict their use of the term remote sensing to describe measurements of electromagnetic radiation. That convention will be followed in this manual although a brief section is included on underwater acoustic devices such as sonars and echo sounders because of their importance to the fishing industry.

Electromagnetic radiation (EMR) is a type of energy which appears in such forms as X-rays, visible light, microwaves and radio waves. While these forms of EMR may initially seem to be separate phenomena, they are in fact part of a continuous spectrum. This can be understood best by considering how a prism separates white light into different colours; each colour represents a different wavelength of light. Visible light is the only segment of EMR which human vision can detect.

A given sensor can detect EMR only over a limited range of wavelengths: this range is referred to as a spectral band. The width of the spectral band, i.e. the extent of the limited range of wavelengths detected, is referred to as spectral resolution. Some sensors are comprised of a number of detectors, each of which is sensitive to a different spectral band. These are called multispectral or multiband sensors. By our looking at the earth in two or more bands simultaneously, it is possible to discriminate a wider range of features. The combination of typical responses coming from a specific target seen by a sensor in various spectral bands is called the spectral signature of that target.

Sensors may be classified according to a number of different criteria. For example, there are imaging and non-imaging sensors. As their name implies, imaging sensors produce a two-dimensional “picture” while non- imaging sensors produce point measurements or profiles. Sensors are also described as being either active or passive. Active sensors transmit radiation to “illuminate” the surface and to receive and measure the amount of radiation which is reflected back. Passive sensors, in contrast, measure naturally produced radiation which is either reflected solar energy or emitted terrestrial energy.

In order to provide a view of the earth's surface a sensor must be mounted on a platform which is simply the device or vehicle from which the sensor operates. Although stationary platforms, either attached or tethered to the ground, are sometimes used for specialized applications, aircraft and satellites are the most commonly used platforms for remote sensing. A general rule is that the higher the altitude of the platform, the larger the area that can be“seen” by the sensor; however, the ability to discriminate small objects will be reduced.

The level of spatial detail which can be observed or recorded by a sensor is referred to as its spatial resolution. For a given sensor/platform system, spatial resolution is usually described in terms of the smallest unit area which can be distinguished from its neighbours. In an imaging sensor system, the individual elements which make up the image are called pixels, a term derived from “picture elements”. The area on the earth's surface represented by a pixel normally corresponds to the spatial resolution of the sensor, i.e. the ground resolution cell size.

Data from sensors may be stored in analog or digital formats. In an analog system variations in the strength of the original input signal (e.g., the brightness variation in an image) are represented by continuous variations in some other medium such as voltage or film density. A digital representation, in contrast, divides the original signal into discrete ranges, each of which is assigned a numerical value. The range of the original signal as represented by a single numerical value is termed the radiometric resolution of the sensor system. Digitally recorded data, unlike analog data, can be processed easily by computers and can be copied repeatedly without negatively affecting the original or copied data. For human interpretation, however, an analog display such as a photograph or television picture is more useful. With appropriate equipment, it is possible to convert data from one format to the other.

A final concept which should be mentioned is the timeliness of remotely sensed information. The term real time is used to describe data that is available for display or analysis at the same time and rate at which it is acquired. Most commonly, there is some delay between the time the sensor “observes” the surface and the time the data is available for use. If this delay is short, for example, a few hours, the data is said to be near real time. When the data has been collected considerably in advance of being analyzed, it is referred to as historical or archival data. Timeliness is a particularly important consideration for fisheries applications because of the dynamic nature of marine resources and ocean processes.

1.3 Fisheries Applications

The discussion has focussed thus far on the technology of remote sensing without reference to its application to fisheries management. The sections which immediately follow describe remote sensing principles, systems and analysis techniques in greater detail. Before we proceed with this examination of remote sensing, it is appropriate to outline the types of fisheries-related studies to which remote sensing may be applied. It should be emphasized that a number of the applications described are in the research stage and are not presently operational.

Although direct detection of fish stocks would appear to be the most obvious goal for remote sensing, it is in fact the most difficult to achieve. Visual fish spotting from aircraft is used successfully for locating a number of pelagic species such as anchovy, swordfish, menhaden and tuna. In this case, a trained observer is the “sensor” and direct radio communication is maintained with vessels in the area. If a camera is also carried onboard the aircraft, photographs can be taken for subsequent stock assessment. Different species can be distinguished on the basis of their colour, behaviour and schooling patterns. Table 1.1 lists a number of species which are directly observable from low-level aircraft. Fish spotting is limited by the range of the aircraft and is only feasible when the probability of fish detection is reasonably high and the economic return derived fromthecatch justifies the expense of aerial surveillance.

A modified type of fish spotting makes use of the phenomenon of bioluminescence which is the emission of light by certain types of plankton when they are disturbed by the movement of fish. This phenomenon has been recognized by fishermen for centuries and is regularly used to locate fish when bioluminescent organisms are abundant. Sensitive low- light level television (LLLTV) systems equipped with image intensifier tubes can be used to amplify the relatively small amount of biologically produced light. Information derived from aircraft-mounted LLLTV systems can be used to direct vessels towards schools of fish. Attempts also have been made to image bioluminescence from an orbiting satellite while scanning the night side of the earth.

While the direct detection of fish is not always feasible, their indirect detection may be possible by observation of sea surface phenomena associated with species distribution. This may simply involve mapping the distribution of fishing activities within a given area. Changes in ocean colour from blue to green may also serve as an indicator of increasing plankton abundance. The green colour is associated with the presence of chlorophyll, the light retaining pigment of phytoplankton. While ocean colour has long been used locally by fishermen to locate fish species, aircraft and satellite imagery can record colour variations over a much wider area in a more precise manner. Techniques have been developed to quantity biological productivity on the basis of chlorophyll distribution and abundance.

Water temperature is another important factor in determining species distribution and thermal sensors can be used to produce maps of the sea surface temperature (SST). Such mapping can be used to identify cold water upwelling of nutrient-rich water and to locate boundary areas between warm and cold waters where certain species are know to congregate.

In addition to resource detection, remote sensing can be valuable in characterizing the marine and costal environments. This may involve such activities as updating navigational charts with coastline and bathymetric data; mapping the distribution and types of coastal wetlands; identifying marine plants and sediment types in the intertidal zone and in shallow waters; and monitoring the condition of coral reefs. While the above applications are related to relatively static or slowly changing conditions, remote sensing can also be used to observe more dynamic phenomena on a regular, repetitive basis. Examples in this category include turbidity patterns (due to both organic and inorganic materials), currents, freshwater and saltwater mixing, and wind and wave regions. Long term monitoring of these phenomena can provide a better understanding of the physical environment which supports biological activity and establishes a baseline against which divergent or unusual events can be measured.

Improved weather forecasting, aided in part by remote sensing, can mean greater safety for fishermen at sea. Pollution from coastal or offshore sources which can negatively affect fishing grounds can be monitored by remote sensing. The intensity and type of fishing activity also can be remotely sensed. This information can be used to determine the rate of resource exploitation and to assist in the enforcement of fishing regulations.

The examples cited above illustrate some of the remote sensing applications which may be of interest to fisheries personnel. It must be stressed, however, that remote sensing can seldom be used in isolation; it must be integrated with other sources of information. The sections which follow explain how remotely sensed data is acquired, processed and analyzed and they demonstrate, through a series of case studies, how it is currently being utilized to facilitate fisheries exploitation and management.