CHAPTER 2 - THE ACQUISITION OF PRIMARY DATA FOR A MARINE FISHERIES RESOURCE GIS

2.1 The Importance of Data Collection

At the beginning of section 1.2 we highlighted the importance of data to the task of spatial management. It will be clear therefore that a GIS cannot function without data, and that generally the more data there is then the greater versatility that a GIS will have and the greater will be the potential functionality of any GIS. With regard to the implementation of GIS generally, the situation has now been arrived at whereby considerations of data are more important than issues concerning hard or software. This situation exists because, given the commercial advantages to be gained, a great deal of research effort and funding has been devoted to technological improvements in the performance of computer systems per se. For very little capital outlay processors can now perform countless functions at incredible speeds and software is cheaply available that will allow for the pursuance of an almost limitless number of requisite tasks. Conversely, given the infinite forms that data may take then it is clearly a far harder task to expedite its collection. So, whereas as recently as five years ago it was hardware and software considerations that formed the first point of interest in establishing a GIS system, it is now considerations of data acquisition that have come to the fore. Costs concerned with data acquisition are now considerably more than those for purchasing hard or software, e.g. Frank et al (1991) estimate that the ratio of the costs for hardware, software and data respectively, over the lifetime of a GIS, are 1:10:100.

At the outset it seems relevant to introduce a few cautionary words on data and its uses. Initially, in this study we define “data” as purposeful observations which have been recorded and stored; this is different from “information” which we would simply define as organised data. Data itself may usefully be classified as being “primary” or “secondary” data. The former represents more a process than a state in the sense that it is facts which are being gathered, usually in order to be converted into secondary data. Thus secondary data represents available data or information which may be refined and/or organised primary data, and which may be available in a range of formats. Table 2.1 gives a summary of ways in which both primary and secondary data collection processes can be categorised. This chapter and chapter 3 will be elaborating on much of this table, and many of these introductory remarks concern both types of data. Figure 2.1 gives an initial indication of the data “flow” both in its collection stages and as it proceeds through the GIS. Again, Chapter 6 will elaborate on the data processing functions which occur within the GIS box. There are some special problems related to the collection and handling of data that is related to marine environments and resources, and some of these are discussed in section 7.2.

Table2.1    A Summary of Primary and Secondary Data Collection for a Marine Fisheries Resources GIS

Once the gathering of data has commenced, we would strongly advise that a meta database (or “informational database”) is established at the earliest opportunity. This is catalogue or listing, held in digital form, which gives details on datasets found which may be potentially useful, i.e. such information as spatio-temporal coverage, appropriateness, date of data collection, source of the data, format and the quality of the data should be recorded. This is described in more detail in section 4.4. Meta databases are important since the data holdings are soon likely to grow exponentially. We would also advise that every effort be made to ascertain the quality of any data which might be considered for use. Quality would include not only the level of accuracy but also its completeness, its detail (or resolution) and perhaps its level of aggregation. If mistakes are embedded then they will multiply once the GIS becomes operative. For most data relative to a marine resources GIS it will be important to understand the sampling techniques which have been used to compile the data. What are their tolerances, variances, etc. Efforts should be made to standardise the quality and methods of sampling to a unified high level. Copyright issues might need to be checked before any data is used.

Figure 2.1 The Transformation of Data Via Various Stages of the Total GIS Operation

2.2 The Data Needed for a Marine Fishery Resources GIS

Although in Chapter 7 we will be setting out some potential database areas, it will be impossible to decide exactly what data is needed for a marine fishery resources GIS until the potential user has established clear aims and objectives for the system. As Roberts and Ricketts (1990) have noted, with regard to a marine resources information system, “…problems concerning the anticipated use of the GIS and related databases can usefully be minimized with the development of a clear statement regarding its potential relevance and use in given instances…. This clearly places the onus on project managers to objectively and unambiguously define the applications and objectives of the particular project and to determine the scope of the GIS applications.” For many reasons the aims and objectives themselves will vary between users or organisations, but once established it will be possible to work out which areas of data collection it is useful to concentrate upon. Since not all areas of a marine fisheries GIS can be instigated simultaneously, it is likely that certain areas will emerge as being important from an “emergency” or “urgency” viewpoint, whilst other areas will be preferential from a feasibility aspect, i.e. “we can get this going quite easily because we have the data.”

Certain core fisheries management aims and objectives are likely to be almost universal. For instance, concerns such as the maintenance or improvement of the coastal zone will be one clear priority, i.e. since the well-being of estuaries, lagoons, mangroves or coastal shallows are vital to the life cycle of many marine resources. Another core aim or objective, though more difficult to achieve, might be the monitoring of fish yields by species for given marine unit areas. Whilst it is not our function to decide what any aims and objectives should be, we should point out that their establishment is crucial to the success of a GIS approach and that their achievement will inevitably involve discussions and decisions across a spectrum of involved parties.

Having established aims and objectives, such as those suggested, it may be a relatively simple task to list the main types of data required. Thus, in our coastal zone management example, the user would clearly require data (for any specified geographic area) on existing land use, any proposed land use changes, land ownership, transport routes, etc., and he may want additional data on factors such as height or gradient of the land, sea-bed types, existing natural vegetation, locations of jetties or harbours, local population densities, etc. Obviously, for the example of monitoring fish yields, then essential data might be that which is shown in Table 2.2. These brief examples simply illustrate the range of data which might initially be required; after a while the GIS user would be a position to add to these basic data requirement lists.

Table 2.2  Essential Data Required for a GIS Which is Being Used for Monitoring Fish Yields

2.3 Collection Methods for Primary Data

In the collection of primary data we are essentially considering the collection of raw data in the field, or more frequently for marine resources, in the water! Some of the data which may be of value to a marine fisheries GIS could be collected in any one of several ways. The actual methods employed, the degree of detail collected and the volume of data collected could depend on a number of factors, e.g.:

a) The time available,
b) the capital outlay planned,
c) the skills and number of personnel involved,
d) the availability and usefulness of any existing data,
e) the size of the area being studied,
f) the equipment available,
g) and the purpose for which it is required.

Data acquired by primary collection techniques could be in many forms, e.g. photographic, numeric, digitally encoded, labelled pictorial, written descriptions, colour coded graphical, and it could be used in several ways. In this section we propose to describe some of the important data collection methods under separate sub-headings which progress from the simple to the sophisticated, and we will discuss both the data collection processes and any equipment involved. Space prohibits anything but a brief descriptive examination of each - those wishing to find further information should consult any of a number of more specialised texts. Within any of the methods described, it would be possible to acquire varying degrees of skills and there are often recognised rules or conventions in collection techniques which are devised so as to ensure degrees of detail, accuracy, consistency, etc. Most methods described require that preparatory work is undertaken so as to achieve the desired level of reliability. This may include being familiar with the use of statistical and sampling techniques i.e. since for many forms of data collection it will be impossible to obtain a complete data coverage.

2.3.1 Manual Methods of Data Collection Using No Equipment

a)  Direct mapping or survey sketching. There will be many instances, especially when dealing with the mapping of very small areas, when it could be necessary to draw maps or plans by hand. For instance, if a new mariculture facility were being planned, perhaps on a coastal flood plain area, then it might be necessary to make the first draughts by hand. This is especially true in areas where no detailed or up to date maps exist. Sketches or maps lend themselves to annotation and other form of labelling. They are inexpensive to produce and they can easily be upgraded such that they could form the basis of later digitising.

b)  Interviewing. Though at first sight this method might not appear to be useful in a spatial context, in practice it is often the only means of obtaining certain sorts of data. Fishermen in any country might need to be asked questions about catches, operating gear, costs, market outlets, etc. Figure 2.2 gives an indication of the types of information gained from interviewing community leaders in coastal sites in Libya (from Meaden and Reynolds, 1994). Fishermen are often familiar with all aspects of their occupation and they may be able to provide insights into local fish habits, climatic factors, water quality, marketing, fishing strategies, etc. It can also be useful to interview people who are not directly involved in marine resources exploitation, e.g. marketing experts, financial or credit experts or various extension officers. There is plenty of information which might only be obtained by means of interviewing, e.g. data on what conditions were like at some time in the past, species habitat preferences, the prevalence of endemic diseases and local transport availability. The disadvantages of interviewing is that the data obtained could be very subjective and it could rarely be converted directly to a mappable or statistical form.

c)  Questionnaires and counts. These are obviously the major way of ascertaining certain types of information. They have advantages over interviewing in that they can provide objective data, much of which could lend itself to mapping. Questionnaires are a very good method of finding out about personal preferences, e.g. types of fish, type of fishing method used, target species, etc. as well all sorts of demand or market preferences. Although they are comparatively cheap to administer, adaptable to sampling techniques and they can provide data which is easily graphed and which can form the basis of written analyses, questionnaires should always be rigorously structured and methodically administered in order to eliminate any forms of bias. Data collected is usually entered onto a previously established form, although increasingly simple electronic data loggers are being used (see section 2.3.3.3).

d)  Form filling. It is standard practice, especially amongst government departments, research departments and larger organisations, to have a range of forms which are required to be completed in order to fulfil supposed data requirements, or to comply with some aspect of the law. Example forms are shown in Figure 2.3. Since a standard form, which is often obligatory, is usually sent out to a large number of respondents, then there is the potential for a large body of useful data to be acquired. Clearly, the potential range of information which could be received is almost limitless and, since the respondents are not usually paid for the information they provide, then this can be an inexpensive form of data gathering.

2.3.2 Methods of Data Collection Using Simple, Non-electronic Equipment

Since there would be an almost limitless list of both methods and equipment we shall only mention a few of the major ones which could prove directly useful to gathering data for a marine fisheries GIS. It should be mentioned that, with the introduction of digital functioning in most data gathering instruments, then non-electronic equipment is rapidly becoming obsolete.

a)  Photography. Although aerial photography is the most useful form of photography for GIS purposes, we shall consider it later along with satellite remote sensing (RS) (see Section 2.2.4.2), i.e. since they are both related RS techniques. Photography is of little direct use to GIS, though it can be very useful as a means of recording information, e.g. recording pictures of fish species caught for later identification, and remotely controlled cameras may be used for various underwater or deepwater survey activities.

Figure 2.2 Fishery Information Sought in a Frame Survey of Communities Along the Libyan Coastline (From Meaden and Reynolds,1994)

Figure 2.3 Examples of Standardised Data Collection Sheets Used on a Baitfish Project in Tonga (from King, 1995)

b)  Site surveying or measurement. In the absence of data specific to small sites, which are perhaps being developed for mariculture or port activities, then surveying and measurement techniques will be necessary for planning purposes. There are an array of acknowledged measuring devices which can be used with little experience. For survey purposes, when levelling and accurate measurements are required, then there are a number of instruments such as clinometers, theodolites and abney levels which can be used, given sufficient knowledge.

c)  Meteorological data gathering. It could well be important to a marine fisheries GIS to gather meteorological data, especially that pertaining to particular micro sites for which there is no other data, or perhaps to particular conditions at some time and place relating to a specific activity or event. Further details on the importance of meteorological data can be obtained from Cushing (1982). There is a range of fairly simple measuring equipment available such as rain gauges, thermometers, barometers and anemometers. The use of these instruments is either self- explanatory or they can easily be referenced.

d)  Trawl surveys. Most marine fisheries personnel will have some familiarity with the range of trawl survey methods which are used as a major source of information on species distributions and abundance. Since the processes involved in fishery resources surveys are extremely varied and sometimes quite complex, it is beyond the scope of this study to elaborate on these, i.e. more than to say that survey vessels will need to be equipped with an array of nets (usually purse seines and otter, beam or midwater trawls) and other items for the passive or active collection of samples. It is also essential that, in order to obtain valid data, the trawl survey itself is designed in a way which ensures the capture of statistically valid information. An example of the type of data gathered in a trawl survey is shown in Figure 2.4. Note that it has been prepared so that all the data received can be easily and accurately entered into the correct columns on a digital database. We recommend as further reading on various survey techniques Smith and Richardson (1977), Doubleday and Rivards (1981), Grosslein and Laurec (1982), Gulland (1983), Fogarty (1985), Hilborn and Walters (1992), Stromme (1992) and Gunderson (1993).

2.3.3 Methods of Data Collection Using Simple Electronic Equipment

In this section we will consider “simple electronic equipment” to be any item which can be used by one person working alone and which is reasonably portable, i.e. even though some of the equipment itself might be technologically fairly complex. Since one range of equipment, Global Positioning Systems (GPS), is quite new, and since it offers a huge potential for the gathering of marine data, we will look at this in more detail than the others.

As implied in Section 2.3.2 above, over the past decade there has been a considerable move towards the manufacture of data collection equipment which incorporates some elementary electronics. This has been of great benefit to the data collector for a number of reasons. Thus, items are invariably far lighter to transport, they can achieve greater degrees of accuracy, their prices are usually far lower than previous mechanical equipment, they are often easier to use and data collection can be greatly speeded up. Under this heading the equipment can be subdivided into two categories, i.e. (i) that which simply produces electronic read-outs and (ii) that which is able to store and download data for further use.

2.3.3.1 Electronic Read-Out Equipment

These devices can nearly all be classified as automatic measurement devices and electronic distance measuring instruments. They constitute a huge range of devices which are purposefully designed to give readings on any parameter for which they have been calibrated. A perusal through a science equipment catalogue reveals digital pH meters, digital thermometers, humidity/temperature meters, ultrasonic volumetric readers, digital light meters, water flow meters, electronic pressure gauges, various weighing devices, etc. Obviously the use of any such device would assume that a systematic survey scheme had been planned and that at least some standard data recording form was available.

Figure 2.4 Example of a Data Collection Form Used for Trawl Surveys (from Gunderson, 1993)

Since so many of the measurements which are made in collecting primary data will form the basis of further work, it has not taken equipment manufacturers long to realise the potential for making devices which can not only do the measurements but which can store them, either on floppy disks or via some fixed internal storage system, i.e. enabling the data to be captured in a digital format for export to another software package. Particular fisheries related examples include a “salinity, temperature, depth recorder” and a “marine temperature profiling system”. In the latter a sophisticated temperature probe is lowered into any marine waters, and temperature records are relayed back, at accuracies of 0.1 C, for disk storage and display. Temperature profiles can be built up, or the information can be used directly so as to concentrate fishing effort into areas where certain species are known to have temperature preferences.

2.3.3.2 Multi-Media Devices

A range of data input methodologies has recently emerged which may be generically referred to as “multimedia”. Thus it is now possible to input to a GIS data from oblique aerial photographs, ordinary photographs, film or video images, graphics, animations, cartoons and even sound. Obviously, to be useful, these media must be converted to a digital format and, to be classed as a GIS, then some sort of geo-referencing is necessary. An example of a multiple media device would be the Still Video Camera. These small cameras offer full colour computer imaging by capturing the image on a small 2 inch re-useable floppy disk. Data can then be transferred to a host computer in a variety of formats for later use in perhaps a GIS or desk-top publishing programmes. Most multimedia devices are beyond the remit of this paper but the interested reader can obtain useful summaries from Parsons (1992) or Cassettari (1993).

2.3.3.3 Data Loggers

There are now a large number of specialist devices which are designed for use in the field and are therefore very portable. They may be semi-automatic, hand held, battery operated instruments which run simple software which has been specifically designed for a particular data recording task, e.g. a hand held survey laser which records azimuth range, heights, inclinations and co-ordinates for later input to other software. Alternatively, they may be automatic devices which are placed in some location where data needs to be constantly and regularly gathered, e.g. weather variables, air pollution monitoring or river water quantity or quality variables. Data entry into hand held devices may be via keyboard input or by pen plotting. Curtis and Bowler (1994) provide a basic description of how the UK Ordnance Survey are now starting to update most of their larger scale topographic mapping data via the use of pen plotters which are being used in the field on notepad computers (Figure 2.5). Intelligent data loggers are available which, for example, can switch between sampling strategies or which can do certain mathematical transformations, etc. The data obtained by loggers, which is typically numeric values or attribute information, is usually stored internally for later transference to another programme, though some systems may display readouts directly to the computer screen. There are now specialist software systems into which trawl survey data can be logged and stored for a range of analyses (Stromme, 1992). A summary of some digital data collection devices is given in Price (1992).

2.3.3.4 Global Positioning Systems

These are devices which allow the user to very accurately establish his location on the Earth's surface. Their ability to do this relies on 21 military (NAVSTAR) satellites which have been developed and launched by the U.S. Department of Defense, and which became fully operational in 1994. These satellites are in predictable orbits and they carry atomic clocks which allow them to transmit highly accurate radio signals. As illustrated in Figure 2.6, the GPS user's position is determined by using a portable receiver which receives and compares the signals from at least three satellites. The receiver is able to perform elementary geometric calculations in order to establish the user's location (at the intersection of the three cones), which is usually given in latitude and longitude co-ordinates.

Figure 2.5 Portable Pen Plotter for the Field Recording of Survey Data

GPS can be used in two ways, i.e. autonomously or differentially. Autonomous GPS is where a single GPS receiver collects inputs from the satellites only. In this case accuracy may not be very great, i.e. typically from dozens to hundreds of metres. Inaccuracy is caused by atmospheric and ionospheric distortions, satellite clock or orbital errors and by intentional degradation imposed by the US Department of Defense (known as “Selective Availability”) - imposed to ensure that 95% of GPS positions are only within 100 metres of their true position.

Differential GPS largely negates the effects of selective availability. It relies upon two GPS receivers (a base or reference station and a rover) that collect data simultaneously. The base receiver occupies a known co-ordinate and, for real time positioning, the rover GPS receiver can move anywhere within transmitting distance of the base receiver. The base receiver calculates correction values associated with the satellite measurements. These corrections can either be transmitted (for use in real time) or they can be downloaded and applied later. With “top of the range” GPS's, accuracies to within a few centimetres can be obtained using differential GPS, though typically it is 0.3 to 2.0 metres. Accuracy rates deteriorate with distance between the base station and the roving GPS at about 1 cm per kilometre. GPS requires a direct line of sight between the receiver and each satellite accessed. At the present time differential GPS accuracy can only be obtained at sea in those areas where there are shore based differential stations, e.g. most of the north west European coastline (Figure 2.7), but with the launch of five Inmarsat-3 navigational satellites in 1996, sub-metre accuracy will then be possible worldwide (and the service will be free of charge).

Figure 2.6 The Functioning of a Global Positioning System

Differential GPS is mostly being used for mapping and to accurately track moving vehicles or vessels. Recent advances in GPS technology means that it is now possible to utilize GPS to help map several thousand objects in a day. The potential that this has for a marine GIS is huge. The use of maps on land has always enabled positions to be quite accurately determined, but of course at sea there are few guides to exact location, especially when out of sight of land. Now it is possible for fishing or survey vessels to accurately determine where trawl hauls were made, where underwater obstacles occur, where water quality samples were taken, etc. Commercial marine equipment companies are already offering base receiver facilities for ships fishing in most of the north west European waters. Governments in many countries, including Japan, New Zealand, Australia, Chile, Canada, the USA and most EC countries, in order to monitor fishing activity, are either making it compulsory for vessels of certain types to carry satellite positioning equipment, or they are experimenting with the implementation of various GPS schemes.

The latest GPS equipment also includes software which can allow for the capture of any attribute or feature data, plus its GPS given position, so as to form a field mapping system. The data obtained can then be directly exported to most GIS packages. In marine use this would allow, for instance, for a survey vessel to continuously monitor water quality along any transect whilst recording the exact location at a rate of twice per second. The data could then be mapped in near real time. It is also possible to build a customised database complete with menus and data dictionaries, so that the GPS can function and record data in any required structure. Ridgeway (1994, personal communication) reports the development of an on-board “automatic position recorder” (APR) which has been specifically designed to both receive GPS information on a trawler's location (at 30 minute intervals) and to record the trawlers catch data, i.e. at the end of a fishing trip the vessel would have a complete record on disk of catch against location. The APR was seen as an alternative way of monitoring vessel location, i.e. from the use of a satellite communication link to a position transponder on-board the fishing vessel. The European Commission are undertaking trials of the two fisheries remote surveillance systems during 1995, with a view to introducing a surveillance system on all EC vessels from 1996. Further information on GPS can be obtained from Cross (1991), Leick (1990) and Gilbert (1994), and the Geodetical Info Magazine has a complete GPS special edition (Vol.7, No.3, November, 1993).

Figure 2.7 The Marine Areas of Northwest Europe Presently Covered by Differential GPS

2.3.4 Methods of Data Collection Using Complex Electronic Equipment

In this section we shall limit our discussion to three major sources of data collection which would be of value to a marine resources GIS, i.e. (i) acoustic SONAR devices, (ii) satellite remote sensing and (iii) aerial photography. Although aerial photography does not necessarily use complex electronic equipment, we have included it here since there is now a rapid move towards the use of video air photography. We should also note that there are many sensors similar to those being carried in satellites, which can and are being used from aircraft platforms. The other two methods are complex in the sense that they require a good deal of expertise and are very expensive to install and/or to operate. The potential user should also be warned that the use of any of these more complex systems means that very large volumes of data will need to be handled and stored, and that advanced skills are required in processing the data acquired.

As an adjunct to the above, it is relevant to note that most of the larger fishing vessels in the developed world, and the majority of research vessels, now have the potential to be equipped with very advanced electronic instrumentation systems. The systems are primarily designed for navigation and for “fish-finding”, and they might consist of a variety of components such as computer processing units, GPS, data loggers, navigational and plotting aids, digital charts, communication systems, and acoustic SONAR equipment. There is some indication that this degree of technological innovation is proving to be seen as extremely complex by those who are having to manage these systems. If as seems likely, GIS capability is added to these systems in the near future, then this can only make the situation even more complex. It now seems essential that moves are made towards systems integration, i.e. so that one “package” is capable of providing an intelligent on-board navigation and fish-finding capability. Failure to do this may soon make it very difficult for fishermen to provide management with requisite data and information.

It is likely that as the use of GIS for marine fishery resources analyses increases, then acoustic SONAR, RS and videography will become the major source of data. However, because these methods have been extensively discussed elsewhere, we shall only briefly describe each so that a recognition of their potential is fully appreciated. Since the acquisition of data using satellite RS techniques could not practically be made directly by those operating a marine fisheries GIS, in Chapter 3 we suggest actual sources for this data. We shall only be considering satellites as used for their data gathering capabilities and not in their communication or transmission capacity. For readers who are interested in obtaining further information on remote sensing, we recommend Curran (1985), Lo (1986), Butler et al (1988), Lantieri (1988), Drury (1990), Cracknell and Hayes (1991), Faust et al (1991), Foody and Curran (1994) and Petrie (1994), and for further information on various underwater acoustic systems we recommend Burczynski (1979), Johannesson and Mitson (1983), MacLennan and Simmonds (1992), Simmonds et al (1992), Foote and Stefansson (1993) and Gunderson (1993). For further information on aerial photography see Cassettari (1993) or Petrie (1994).

2.3.4.1 Acoustic SONAR Systems

The use of underwater acoustic or echo-sounding devices has two main purposes as far a marine fisheries GIS is concerned, i.e. to gather data for underwater mapping and to locate fish or other underwater objects. The location of fish is usually for commercial exploitation reasons but echo sounding is also used extensively in biomass survey work. In carrying out acoustic detection techniques certain pieces of equipment are essential:

 (i) A Transducer - this instrument converts an electrical signal to mechanical vibrations (pulses or pressure waves) which physically move the adjacent water particles (Figure 2.8).

(ii) A platform - this carries the transducer and other equipment over the data collection area.

(iii) An echo sounder - this transmits and receives electronic signals to control the pulsing of the transducer and reception of the echoes.

(iv) A computer plus the relevant software systems to store transmitted data for later application to a GIS (or to other software analysis programmes).

(v) The necessary hardware to capture and view the transmitted echoes on, i.e. usually a Visual Display Unit (VDU).

Other equipment might be necessary such as fish netting gear to verify the species and the age or size distribution of observed fish.

Figure 2.8 Propagation of an Acoustic Wave Produced by Vibration of the Transducer (after Gunderson, 1993)

In carrying out survey data collection, the transducer is generally carried or towed across the study area, i.e. suspended under the platform (Figure 2.9). This area will have been previously selected and a trawl or survey path will have been designed, usually in zig-zags or in parallel lines (see below). Obviously the extent of the survey will be dependent upon available time, finance, area of interest and whether or not anything has been located. Whilst the survey is in progress the transducer should be in continuous operation. The transducer pulsing occurs at intervals of typically one “ping” per second, though to avoid confusions with echoes being returned from the sea bottom, pulse repetition rates in fact vary with water depth. It normally uses a frequency of 10–12 kHz with a beam width of 10–20 degrees and a bandwidth of 500Hz, though importantly all of these configuration values can be adjusted or optimised for specific data collection tasks, e.g. for the detection of bottom living species it would be important to have a narrow beam width and a short pulse length. When the sound pulse strikes an object it causes an echo which rebounds to the transducer. Given that the transducer's beam angle, ping rate, the speed at which the pulse travels and the speed of the towing vessel are known, then it is possible to detect where an object is relative to the transducer. Acoustic devices can operate to depths of 1200 metres and may either sweep continuously through 360 degrees or they may sweep at selectable speeds, resolutions and in selectable sectors. Figure 2.9 shows that there are variations in transducer configurations, i.e. ship mounted or net mounted, and Figure 2.10 illustrates how an acoustic scanner can vary the area being sensed, i.e. (1) a headrope mounted scanner can look up to 800 metres in front of the trawl; (2) in downlook scanning mode quantities of fish entering the net can be viewed and the trawl opening can be measured and; (3) in net profile mode the correct net opening geometry can be maintained.

Figure 2.9 Basic Variations in Transducer Configurations

Echoes received are collected as marks, in black and white or in colour, along a strip of moving paper (or on a VDU screen or even speakers that convert the echo signal into audible sounds). The paper is calibrated so that the marks are drawn relative to the distance from the object - this produces an echogram, i.e. a visual image of what has been detected by the transducer. Clearly, this information can be collected and stored digitally for use in a computer. There are now a number of software packages which can convert the data into a variety of useful output forms (see Chapter 3).

Figure 2.10 Different Acoustic Transducer Sensing Modes

Different characteristics of various aquatic species such as their size, shape, distribution, schooling habits, etc., means that the echogram will exhibit different marking patterns. An experienced user of the equipment may be able to interpret the markings so as to calculate fish densities (though not yet the species of an individual fish), and distribution patterns within the survey area. Often, for various reasons, this is impossible so target identification is then necessary using net trawls or purse seines, etc. The measurements obtained can be used as interpolation points which can form the basis of describing the total fish quantity and distribution over the entire survey area.

Whilst what we have described briefly here shows how fish numbers can be calculated, it should be noted that the compilation of bathymetric maps (Figure 2.11), or 3-D bathymetric images (Figure 2.12), using similar acoustic technology are equally feasible. Here, of course, the survey should only need to be carried out once, and there are no complications caused by the size of the object being detected or by its mobility. Additionally, what are called “ground discrimination units” are now available. These are echo sounders which are specifically designed to capture the data necessary to allow for differentiation between sea bottom types. Useful summaries on the application of acoustic methods to ocean mapping can be obtained from Talukdar and Tyce (1990), Miller (1991), Mills and Perry (1992) or Somers (1992).

As well as basic echo sounding devices to determine the presence of fish or the depth of the sea, new instruments are now being marketed which are described as “integrated information systems”. This means that the one set of instrumentation can provide multi-window displays showing fish detection, catching operations, and water depth and they can function as a navigation aid. Other acoustic instruments have been developed solely for navigational use. It should also be mentioned that there a number of variations on the basic acoustic systems, e.g. there are multi-beam swath echo sounders, split beam transducers, sidescanning systems and combined sidescan/swath bathymetry systems (Figure 2.13). Potentials users should do some research before investing in any of this expensive technology.

Figure 2.11 Typical Bathymetric Output Achieved from an Acoustic Bottom Survey

Having outlined the technological functioning of acoustic surveying, it is relevant to briefly mention some important considerations on survey design for estimating species numbers. Factors we consider here can be applied to both acoustic trawl surveys and to ordinary surveys for fish, planktons and various water qualitative factors. Since the acoustic data gathered will only represent a sample of the number of fish present, then if this sample is to be used to make estimates of a total population, it is important that the survey transects occupy times or locations such that the information finally produced will have statistical reliability, i.e. it will be unbiased. Track-line patterns used during a survey could take several forms as illustrated in Figure 2.14. As an example of bias in a survey, then if the survey were to follow the zig-zag pattern it is clear that there is a higher sampling intensity near the turns compared to other parts of the track- line. There are several ways of ensuring that bias is minimised e.g. see Williamson (1982), Francis (1985), Simmonds et al (1992) and Hansson (1993). In an acoustic survey of anchovies off South Africa, Jolly and Hampton (1990) suggested that discrete transects placed quasi- randomly within previously defined stratum areas would give unbiased results (Figure 2.15), though Francis (1985) suggested that simply placing the transects uniformly in parallel lines would still give valid results. Whichever locational method is adopted, it is clear that consideration would also need to be given to factors such as the length of the transects, the time employed in terms of tow length, clock time and time of year, standardisation of all procedures used, etc.

Figure 2.12 3-D Bathymetric Image Derived from Acoustic Data

Figure 2.13 The Main Components of a Sidescan SONAR System (after Robinson et al, 1995)

Figure 2.14 Possible Transect Patterns for Use in Trawl Surveys

2.3.4.2 Satellite Remote Sensing Systems

Like acoustic surveys, aerial remote sensing (RS) is concerned with the collection of data by sensing devices which are not in contact with the object being sensed. Various technological developments had become integrated to the extent that satellite observations from space first became feasible in the 1960's. Since that time several series of satellites have been launched, first by the Americans and Russians, and more recently by France, a European consortium (the European Space Agency - ESA), plus India and Japan. Generally, the equipment used has become progressively more sophisticated enabling a greater range of images to be captured in a more detailed spatial resolution. This in turn has led to a greater utility of RS as a viable data gathering medium. Figure 2.16 shows the main satellite RS systems which might provide directly useful data for a marine fishery resources GIS. From this it can be seen that the main marine parameters which can be monitored relate to water temperatures, wave height and direction, bathymetry, ocean currents and water colour, though indications can be estimated for sediment concentrations, phytoplankton standing stock, turbidity patterns, current speed and direction and light attenuation coefficients. Figure 2.17 gives a simple illustration of how useful RS could be to fisheries resources management. These satellite images show the plankton densities for one time period in the upwellings off the coast of (a) Peru and (b) West Africa. Red and yellow equals dense blooms, darker blues are areas having almost no plankton and black is the land. It is interesting to contrast the coastal areas in Peru where the upwelling occurs with those having no upwelling and thus no plankton. Recent work has also shown that, in cases where water quality is good, then it is possible to use RS to differentiate between certain sea bottom types (Luczkovich et al, 1993). For additional information on how satellite RS techniques have been used to aid marine fisheries see Hock (1986), European Space Agency (1987), Butler et al (1988), Myers and Hick (1990), Allan (1992), Cusido et al (1992), Simpson (1992) and Prasad and Haedrich (1993).

Figure 2.15 Transect Plan for an Acoustic Survey Off South Africa Using Quasi- Randomly Generated Track-Lines (after Jolly and Hampton, 1990)

Figure 2.16 Current and Planned Satellite Systems in Support of Marine Meterology and Oceanography - 1992 to 2010 (From UNESCO, 1992)

For its functionality satellite RS relies upon several pieces of hardware:

(a) A satellite which is in space either in a geo-stationary orbit or it is earth orbiting. This is the platform for:

(b) Various sensors, which gather certain wavelengths of electromagnetic radiation (EMR) given off from the earth. The variable amounts of radiation detected must be relayed back to:

(c) Ground control centres which control the RS activity and produce RS output data. The relaying of data from the satellite is achieved by:

(d) Data storage and transmitting devices on board the satellite.

Figure 2.17 Processed Satellite Imagery Showing Plankton Densities Off Peru (left) and West Africa (right)

So RS relies on the fact that different facets of the earth's land or water surfaces transmit or reflect different amounts of energy, i.e. they have so called spectral signatures as shown in Figure 2.18. The sensors detecting EMR may be either push broom (framing), in which case an entire image is captured instantaneously (as in still photography), or scanning in which case the sensor sweeps across a scene in a series of parallel lines collecting data, via numerous detectors, to build up an image (Figure 2.19). Each detector records a value, using 8-bit coding to give a possible range of 256 values corresponding to emitted radiation, for each square on the earth's surface, and the size of these individual picture elements (pixels) denotes the sensor's resolution. The sensor's resolution is a function of the height of the satellite, the focal length of the lens, the wavelength of the radiation and other characteristics of the sensor itself. The use of different spectral bands produces different EMR values. There are so-called passive sensors, which record reflected or emitted EMR, or active sensors which illuminate an object with their own radiation source and then record the echo, e.g. the underwater acoustic devices as discussed above would be active sensors.

Figure 2.18 Spectral Signatures of Various Natural Earth Features

Satellites may be classified as being either “geostationary” or “polar orbiting” (Figure 2.20). The geo-stationary satellites circle the earth at a height of 35 900 kms usually above the equator. At this height their speed of travel can be easily matched to the speed of the earth's rotation. These satellites, though having a poor spatial resolution of between 2 and 5 kms, are useful for telecommunications, for transmitting real time views of the earth's weather and for monitoring certain environmental factors. Geo-stationary satellites include the GOES and Meteosat series plus India's INSAT and the Japanese HIMAWARI.

Most orbiting satellites are in near polar orbits at heights varying from 270 to 1600 kms, and they complete a revolution of the earth in 95 to 115 minutes, i.e. 12 to 16 revolutions per day (see Figure 2.23). By far the most productive satellites, from the point of view of total data available, have been the U.S.A. Landsat series, using their Thematic Mapper and Multispectral Scanner sensors (Figure 2.21), and the two French SPOT satellites. Both systems have been used for “mapping” vegetation, geological features, soil types, various linear features such as larger rivers and certain water parameters. The SPOT satellite is capable of producing 3-D images through its ability for “off-nadir” viewing at angles of up to 27 degrees (Figure 2.22). There is now a huge amount of imagery available from the Russian Kosmos series of satellites. Most of this imagery has been obtained by short life (11 to 31 days) satellites which have on board framing camera sensors - these take high resolution photographs which are physically returned to earth for all processing. Of special importance to any marine GIS is the European Space Agency's (ESA) ERS-1 satellite, i.e. since its sensors have been designed with a view to obtaining important data relative to seas and oceans including characterisation of ice fields, spectral analysis of surface waves, mapping of surface wind fields and the measurement of sea surface temperatures. Figure 2.23 shows the polar orbital tracks for ERS-1 during one 24 hour period plus the data gained on wind speed (in miles per second). In approximately 80 hours wind patterns for the whole global marine area can be mapped. It is then possible to derive a mean wind fields for any area (Figure 2.24). Data for the compilation of this is based on recording changes in the backscatter of the sea surface every 3.763 seconds. Figure 2.25 shows the data gathering characteristics of some ERS-1 sensors. Thus the main sensor aboard is the Active Microwave Instrument (AMI) which can operate in image mode for capturing certain data (see Figure 2.25 (A)), or wave mode (B).

Figure 2.19 Geostationary and Polar Orbiting Satellite Positioning

The electronic images which have been captured by the sensors are either transmitted as a stream of binary numbers directly to the ground receiving stations or are stored in on-board recorders for later transmission. At this stage the data is in a pre-processed state. To make the data useful for GIS purposes there are an array of image analysis processes which can be performed. The most important of these are reviewed in Chapter 4.

There are a wide range of future satellite missions presently being planned, most of which were shown in Figure 2.16. ESA has three projects at present, i.e. ERS-2, ENVISAT and METOP which will provide a continuous environmentally biased data-stream until 2009. The US plans to launch many satellites over a similar period, including Landsat-7 and further satellites in the Tiros/NOAA series, and France, India, Japan and Russia will continue with their polar and/or geo-stationary platforms. A new trend in the US is that for the first time private companies will be launching their own satellites with an obvious view to making profits from data sales, and at least one US sensor, to be launched by Lockheed in 1996, is planned to have one metre resolution which will give the GIS user a viable alternative to aerial photography. More ground receiving stations worldwide will also be commissioned which will aid the consistency of image supply. A further important furture trend is for the building of lightweight satellites (750 kilos) for placing in low orbit. This will have the effect of considerably reducing costs, especially those of launching.

It is also of interest to mention here that several private companies have recently negotiated commercial deals with private fishing vessels, to supply them directly with up-to-date (near real time) transmitted satellite data, which are made up by combining images showing water colour conditions and sea water surface temperatures. These charts can prove very useful in suggesting where pelagic concentrations are likely to occur. The images are transmitted via the Inmarsat communications satellite and can be displayed on a personal computer.

Before completing this brief look at remote sensing it is important to mention a few of the limitations which are inherent in this technology. We do this because in many senses RS has not achieved the potential which many expected of it. Thus, although the progress in RS technology and applications has indeed been vast, and GIS is proving to be a valid medium for propagation of RS data, the data provided by this methodology is far from easy to use. Some of the main limitations include:

 (i) Cost. Table 3.4 gives some indication of recent RS data costs, and Figure 2.26 gives a comparison of SPOT imagery costs with aerial survey costs as a means of updating 1:10 000 digital base mapping in New Brunswick, Canada (from Rapatz et al, 1990). These costs show that if the one SPOT image can be used for updating up to 80 map sheets then costs are very competitive, though the authors caution on what the mapping priorities might be. Thus, although in terms of cost per square kilometre of surface imaged the prices might be good value, each image only represents one moment in time and several images may be needed to cover a fairly small oceanic or coastal area. There are also large image processing costs to consider.

Figure 2.20 Basic Principles of Scanning and Push Broom Satellite Remote Sensing

(ii) Lack of images. Because cloud cover duration varies greatly from area to area, the number of images available for nearly contiguous areas can vary by a factor of five. So equatorial regions, west facing coastal areas in the mid to high latitudes and areas having seasonal rainfall may have a paucity of data. Also, data is scarce for polar areas and areas beyond the receiving range of some satellite transmitters.

(iii) Ground truthing. Since there is no way of being certain that the EMR value recorded for a pixel represents a particular land use or water condition, then it is essential that the imagery is verified by on-ground verification. This is also inconvenient in that it needs to be done at the same time as the satellite pass occurs, i.e. since pixel values can change from day to day.

(iv) Spatial resolution. The best resolution generally available from satellite RS is 10 metres, although much of the Russian satellite photographic data is down to 2m resolution. Whilst this range of resolution is suitable for many purposes, it is clear that smaller objects cannot be detected. The technology exists to get better resolution. When the political, cost and legal barriers to its acquisition are overcome, then RS should prove much more useful.

Figure 2.21 The Configuration of Landsats 4 and 5

(v) Long term planning and payload uncertainty. Most longer term projects using RS data require an assured temporal sequence of images, partly in order to justify the necessary hard and software costs. To some extent RS suffers from “planning blight” in that most satellite launches have suffered from uncertainty in scheduling and payload. Delays, such as the three year hold-up on ERS-1, have been mostly caused by lack of funding assurances and technological problems.

(vi) There are many other minor problems such as the vast data storage capacity needed, obtaining the necessary skills for RS imagery interpretation, plus the actual management and upkeep of the data.

Notwithstanding these problems, we envisage that the longer term future for RS is very healthy, especially if the GIS and RS technologies continue to become better integrated. The following list illustrates some reasons for our optimism:

(a)  There are a large number of planned and committed satellite launches. Some of these are joint satellite ventures between different countries.

(b)  RS is being utilised in an increasing number of fields.

(c)  It is increasingly easy to be able to integrate RS imagery into GIS programs.

(d)  There is some evidence that the technology behind the defence uses of RS will be increasingly diverted to civilian use.

Figure 2.22 Stereoscopic Viewing Capabilities of the SPOT Satellites

(e)  RS is undoubtedly likely to play an increasing role in the move towards environmental awareness.

(f)  Rapid advances in processing hardware means that the vast data increases provided by RS will be relatively easily handled.

(g)  New types of sensor platforms are being developed which will eventually ensure a wider variety of sensors can function and that they can be serviced in orbit.

(h)  A Number of new countries are entering the satellite RS field, e.g. Canada, China, Brazil and Indonesia - they will require some return on their expensive investments.

(i)  The larger number of radar (active) sensors now airborne or planned will ensure better data coverage from areas having cloud cover problems.

(j)  Future imagery capture is likely to be programmable to specific user requirements.

Figure 2.23 Orbital Sequence and Colour Coded Chart of Ocean Surface Wind Speeds Collected on 18th June, 1992 by the ERS-1 Satellite

Figure 2.24 Composite Wind Field Map of the Western Pacific Compiled from Wind Scatterometer Data Collected by the ERS-1 Satellite

(k)  RS is the only means by which an instant synoptic view can be obtained of areas varying from a 60 × 60 km cell to a whole earth disk image.

Figure 2.25 Technical Characteristics of the ERS-1 Active Microwave Instrument Operating in (A) Image Mode and (B) Wave Mode

(l)  The capability that RS has to undertake repetitive coverage so as to achieve temporal monitoring of any area or circumstance.

Figure 2.26 Cost Comparisons for Map Revision in Canada Using Three Different Data Sources

2.3.4.3 Aerial Photography

Since the middle of World War 1, photography from aircraft has been used as a method of obtaining spatial information, and since then there have been many advances which have ensured that nowadays aerial imagery is of a very high standard. Flight programmes for mapping purposes must be flown within the parameters as shown in Figure 2.27, i.e. such that large overlaps are obtained. The overlaps are necessary in order that stereoscopic (3-D) viewing can take place. Until the mid 1980's most aerial photography was accomplished in black and white using panchromatic or infrared film, but with the development of high resolution colour film and cheaper processing techniques then colour output became the norm. Recently however, to avoid the need to convert an aerial photograph into a digital form for input to a GIS, digital cameras have been developed, i.e. which capture images directly as arrays of coded pixels. These cameras can be adjusted to record different spectral bands within the visible and infrared spectrum. Aerial photography as an aid to GIS has advantages over the use of mapping in the sense that more detail can be obtained on existing land use or on sea-based activities, and it may be preferable to satellite RS because of the much higher resolution obtainable and the fact that specific flight programmes can easily be arranged. Imagery is typically available at scales between 1:2 000 and 1:25 000.

The extraction of valuable information from aerial photography depends on the process of photogrammetry. It will be clear that when any in-flight photograph is made there could be many sources of image distortion (Figure 2.28). To correct for these a stereo plotter is used. This device allows an operator to view two overlapping photographs, taken from different positions, in order to form a three dimensional image. The operator can then draw in any desired outlines in their correct location. Using digital photogrammetric techniques all lines and objects in the drawing can be numerically encoded and stored in a computer database. In many areas of the world, original topographic mapping is being done via photogrammetric techniques. It will be appreciated that an accurate orthophoto, i.e. one that has been corrected to remove the distortions, will form an excellent basis upon which to ensure the accuracy of mapping. The interpretation of aerial photographs is a highly skilled task, one that is explained in detail in Star and Estes (1990).

Figure 2.27 Image Overlap Sequence for Attaining Stereoscopic Aerial Photography

Figure 2.28 Some Sources of Error in Aerial Photography

With the likely increased use of aerial surveillance methods for monitoring fishery activities, then a particular aerial data collection system which will be worthy of future consideration is the use of airborne videography, i.e. the use of a simple video recording camera from an aircraft with its potential for the digital processing of the electronic signal (Figure 2.29). Debusschere et al (1992) and Mausel et al (1992) provide further details on this methodology. A similar digital remote sensing data capture method from aircraft is reported by Borstad et al (1992). They have used successfully a Compact Airborne Spectrographic Imager (CASI) to directly capture the spectral signature of schools of Pacific herring (Clupea harengus pallasi) in shallow waters off the coast of Canada.

Figure 2.29 The Main Components of a Videographic Image Collection System (from Robinson et al, 1995)