The purpose of this chapter is to conceive of how a marine fisheries resource GIS might best be set up in terms of its varying potential to help with fisheries management. As was indicated in Chapter 1, marine fisheries is both a subject area which is extremely broad, and an activity which is carried out over a significant proportion of the Earth's surface by a vast number of people and groups. Comparative freedom of access to the marine resource base, coupled with fragmentation of the fishing communities, has meant that in the past coordinated fisheries management strategies have seldom been implemented. During this century there have emerged various attempts at fisheries management, first by biologists and later by economists, sociologists and then environmentalists. But for too long these attempts have been pursued in a more or less mono-disciplinary way. Continued over-exploitation and declining fish stocks have shown this to be the case, and the situation has now been reached whereby, if marine areas are to continue to yield fish and other biomass resources, then better, integrated management strategies are imperative.
In Chapter 1 it was also made clear that a major problem for fisheries was that it was a multi- faceted discipline, operating essentially in a spatially extensive and spatially variable environment. In the following chapters we demonstrated that GIS is a tool which has the necessary requirements and functionality to allow for spatial management across multi- disciplinary subject fields. For any GIS to achieve optimum operative capability then, as was shown in Chapter 6, it is clear that the data which is entered into the system needs to be managed. In this chapter we intend to give suggestions and examples of how this data management can best be achieved.
Before looking at the potential individual database areas into which a marine fisheries GIS might be subdivided, it is relevant to outline some of the general methodological problems which are likely to be encountered across all or some of the database areas. Under this heading we are referring not to specifically GIS based operating or functional problems, but to those considerations which might prove difficult to solve when setting up databases for the various facets of marine fisheries.
The vast majority of entities which are mapped are distributed in 2-D space. Although this mapping may give rise to considerations of scale, symbolism, classification, generalisation, etc, relatively speaking it causes comparatively few problems. Some entities may need mapping, or displaying, in 2.5-D, i.e. where altitudinal differences are shown by various means on a 2-D surface. There are ways of easily accomplishing this either by showing the height dimension (or depth dimensions for the oceans or seas) using contours, hachures and various shadings, or by the use of various “three” dimensional, elevational models. However, the problem for those wishing to map fish biomass distributions is that they nearly all occur in genuine 3-D space (the exceptions are, of course, wholly demersal or bottom living species).
From a practical viewpoint the implication of this is that the fisheries management team, in conjunction with those operating the GIS and those setting up data collection systems, have to make decisions concerning how best to portray species distributions in this 3-D space and how to set up databases which record in three dimensions. Clearly one answer is to simply ignore the third dimension. This would mean that any map drawn would be a spatial representation which ignored the vertical distribution of the species (or of water qualitative parameters). For many mapping instances this may not be a problem, and indeed, in many mapping scenarios it is likely that there is insufficient data available to allow for mapping in any other way. An alternative answer is to map distributions by suitable vertical depth class categories, i.e. with there being as many maps for any one area as there were depth class categories. The actual categories used would vary with the number of class divisions, the purposes of mapping, the depth of the water and likely species or water qualitative distributions. This mapping would involve the procurement of quite detailed distributional data, often via acoustic methods. Data sets would be very large per unit of surface area.
Work is now in progress on sophisticated modelling and mapping methods whereby, for any desired mapping plane along any axis, it is possible to take a “slice” through a representation of any plotted 3-D distribution so as to show the distribution in 2-D across the desired plane. Figure 7.1 shows a hypothetical example of a 3-D “cube” of marine space with a “slice” taken through the latitudinal axis. The slice shows fish biomass density variations given the specific combinations of latitude, longitude and water depth. Although this methodology has yet to be incorporated into present GIS's, it will in the future allow for some quite sophisticated 3-D maps to be drawn, especially in instances at a larger scale whereby particular vertical distributions were comparatively stable over time and space, e.g. perhaps generalised seasonal thermoclinal models. Manley and Tallet (1990) have shown examples of experimental work which shows how optimal visualization is being achieved using 3-D GIS images of factors such as salinity and water temperature variations. There are also other initiatives underway which are aiming to develop genuine 3-D GIS's (Raper, 1989; Mason et al, 1992; O'Conaill et al, 1992; Li and Saxena, 1993; Hack and Sides, 1994; Houlding, 1994), but it may be some time before these are available for use in typical marine fisheries applications. The eventual decision on how to map in a 3-D environment should be made according to the purposes of the mapping and thus of the final data output required. Clearly there is no right or wrong way. The decision will be influenced by the type of source data held, the costs of getting primary data which is valid in 3-D space, the scale or spatial extent of the mapping and perhaps a cost/benefit analysis between alternative ways of mapping.
Figure 7.1 A 3-D Hypothetical Cube of Oceanic Water With a Slice to Show Fish Biomass Variations
The vast majority of maps produced are spatial representations of the distribution of static objects or areas in the terrestrial environment. In this case it is easy to locate the entities in a relatively exact location vis a vis other entities. However, there are increasingly attempts being made to map objects which are non-static (e.g. Miller, 1994). This mapping may involve various types of representation such as flow lines, which show volumetric amounts of movement along certain corridors, e.g. the number of bus movements along various routes per a 24 hour time period, or perhaps mapped biological distributions of various faunal species. Although many of these latter maps can look successful, the map user will need to look with care at exactly what the map is purporting to show. Thus they might be showing the maximum extent of a species, areas where the species have been recorded within a certain time period, areas where the species have ever been recorded or areas where a species most commonly occurs. For the fisheries manager, there is the added mapping complexity in that not only do the species themselves show a high time/space variability, but so does their environment, i.e. the water quality may change, and this is frequently the causal factor for species distributional change.
This introduces the additional and related problem of actually defining boundaries in the marine or coastal environment, i.e. how, in this constantly changing milieu, is it possible to draw clear boundaries around zones or areas having relatively non-homogeneous properties? (e.g. imagine the difficulties in drawing boundaries along a river estuary between saline, brackish and fresh water). Thus marine or coastal entities tend to have fuzzy or transitory boundaries which require methods of representation which, strictly speaking, can take this uncertainty into account. At the present state of GIS development, this is impossible, although progress is being made towards dynamic modelling which should result in measures towards solving these problems.
However, there are several ways of overcoming these various mapping problems. As alluded to above, mapping can show extreme, average or seasonal distributions, and for some types of mapping it might even be necessary to show diurnal distributions. The overall success of mapping which allows for the mobility of both species and their environment will depend on factors such as the importance of the mapping, the accuracy required, the scale of the area being mapped, the frequency of mapping and the relative amount of movement of the species or the water qualitative factor. Again, the fisheries management team will need to discuss exact requirements in terms of the task being undertaken.
We have previously mentioned (section 4.2.1) that all mapping entities may be thought of in terms of points, lines and areas, and the fisheries manager will be required to map examples of each of these in any GIS exercise. Static point features, linear features and all spatially extensive areas will be comparatively easy to map since their position and extent is easily determined. However, mobile point features (and sometimes other features) are not so conducive to normal mapping, and so they are frequently mapped as grouped quantitative variables. According to the purposes of mapping, to the numeric scale of measurement used, and the way in which the data has been gathered, quantitative variables may be mapped in various ways. Figure 7.2 shows three main methods which are commonly used, though there are many variates of these. Again, depending on the purpose of mapping, there is no right or wrong method to use.
Figure 7.2 The Three Main Ways of Mapping Variable Quantities
From a practical viewpoint, in a GIS the fisheries manager would most frequently use either isoline or cellular mapping techniques. Where variables change slowly over a spatial continuum, such as water quality changes, then isoline maps can be reliably used so long as the source data is accurate enough and sufficient point samples are available. Although this method could also be used for mapping perhaps fish densities, it would seldom be advisable because it is difficult to provide any guidance as to the statistical validity of the map. For this reason density mapping by unit cells is the best method. However, in the use of unit cells there is the problem of cell size to be considered. It is impossible to advise precisely on this since there are many factors that need considering, the most obvious of which are the total size of the marine area being mapped, and the density of species sampling data which is either held or planned. We would advise on the establishing of a nested hierarchy of cells to cover the whole marine area. Though this is discussed in Meaden (1994) we show here (Figure 7.3) how this might be represented.
Fig 7.3 A Possible Means of Dividing a Marine Area for Mapping by Using a Nested Hierarchy of Cells
The choice of square cells relates to their usefulness in nesting, to the fact that simple algorithms exist which can provide a measure of the statistical significance of any survey data obtained, and to the fact that other raster type data is gathered in this way, e.g. RS data. Additionally, data held in cell units is structured in a format which will be advantageous to any raster based GIS software, and the geo-referencing of this data will be simple. However, for some aspects of a potential marine fisheries GIS, square cells may not be the optimum shape. For instance, if the concern is with modelling (simulating) the movement of marine organisms from a given starting point then, although any one square cell has eight surrounding cells, it is far more likely that the movement will be into one of the four adjoining side cells than into one of the four adjoining corner cells. Since this would clearly be unfair, then hexagonal shaped cells have been sometimes used, being a shape which also tessellates (Taylor and Ullman, 1993). Obviously a suitable geo-referencing system would need to be devised for the use of hexagonal tiling.
To obtain data on species existence and likely species quantities, various data sampling methods have been traditionally used. Although the data obtained is obviously useful, it is clear that in the vast 3-D space of the open sea, then there will be the likelihood of a huge margin for error, i.e. since the survey can only gather samples from a very small fraction of the total space. The margin for error will potentially be less for demersal species, who mainly occupy 2.5-D space (2.5-D indicates a surface link), than for pelagic species who may occupy much of any 3-D space. Differences in the behaviour of certain stocks and their rate of mobility is also important, e.g. the sampling of stocks which exhibit a high shoaling density will call for different sampling strategies than those where members of a stock are ubiquitously distributed, and this sampling logic also applies to species which are more migratory than others. The implication of this is that, when it comes to any mapping of the species surveyed, the data used has the potential to contain a large margin of error, i.e. the statistical variance in grouped sample data is likely to be very high. In many cases therefore, confidence in the validity of the data will not be high. The extent to which this is true will depend on the accuracy and frequency of surveys, plus the survey methods used, though it might also depend upon the dispositional habits of the particular species being mapped. Meaden (1995) has suggested that, where relevant, GIS output can incorporate maps which give an indication of variance. Thus, Figure 7.4 shows an example whereby hypothetical estimations of fish biomass and of variance in the survey data is shown simultaneously.
From the GIS viewpoint, this lack of confidence in the data and the high likelihood of error, is a cause for concern. This lies mainly in the fact that, if the data is to be used as the basis for any manipulations, then errors are going to be multiplied. So, for instance, if the fisheries manager is hoping to maintain a record of fish stocks in a particular marine zone, via the use of GIS, then manipulations would need to be made which involved adding recruitment estimates to existing surveyed stocks and then subtracting the total of fish catches and/or fish mortalities and migration. Notice that data on recruitment, mortality and migration might also be subject to large statistical variance. So, given the present state of fisheries surveys and GIS functionality, then it must be recognised that output data (map or tabular) should in many cases be treated with some caution. However, this is an area where there are likely to be rapid future improvements, especially with the widespread introduction of acoustic methods of stock assessment, and with the likely advent of realistic models to show marine species quantity and distribution both temporally and in 3-D space. In fact, we would contend that the use of GIS is likely to help provide evidence that, for certain species at least, their spatial disposition is likely to follow certain quite well defined patterns. Readers interested in the problem of spatial variance associated with fish stock distributions should consult Smith (1990), Foote and Stefansson (1993), Gunderson (1993) or Petitgas (1993).
It has already been indicated that, since most marine stocks are to a greater or lesser degree, mobile then any map showing their distribution is going to have a poor temporal resolution, i.e. the map is rapidly going to become dated, and it could only be “accurate” if it was purporting to show the approximate species distribution for a very particular time period. Other factors relative to marine fisheries and resources are also going to be variable over time, e.g. the distribution of fishing effort, the extent of market zones, water quality or quotas per unit area. The temporal scale used in GIS mapping has to be a function of many variables most of which will relate to the purposes of the mapping, the temporally variable data available and probably a perceived cost/benefit decision by the fisheries manager.
Figure 7.4 Simulated GIS Output to Show the Incorporation of Variance Estimates Into a Fish Biomass Survey Map
Another important decision for which there are no hard and fast rules is that concerning the spatial scale(s) of any mapping. The scale decided upon will vary as a function of the purposes of the management task, the detail required, the input efforts available, the relative mobility of a species or the area that it occupies, and the extent of the marine area to be mapped. Stoms (1994) provides an excellent example, relating to the terrestrial species richness of vertebrates in southern Idaho, of how the scale used for data collection and mapping can influence the validity and usefulness of the final map (Figure 7.5). Here it can be seen that, with the use of 890 hectare cells a reasonably accurate perception of the distribution of vertebrate species can be obtained (level 1), but by the time cell size is increased to 57,000 hectares (level 4) then the information as portrayed for this part of Idaho is virtually useless. To briefly illustrate scale variability in the marine situation, we could envisage that maps showing legal or exclusive economic zones (EEZs) could be compiled at quite a small scale (perhaps 1:5 000 000, but depending on the region being mapped) whereas, at the other extreme, maps showing the possible configuration of mariculture facilities at a particular location could be at a scale of perhaps 1:2 000. Figure 7.6 provides a good indication of the scale variations which might need consideration.
Figure 7.5 Species Richness Maps Showing Absolute Richness of Vertebrates in Part of Southern Idaho (USA) at Five Sampling Unit Sizes (from Stoms, 1994)
Figure 7.6 Typical Time and Space Scales Associated with Plants (P), Zooplankton (Z) and Pelagic Fish (F)(from Rothschild,1986)
A further way in which the mapping scale chosen can be an important decision relates to the fact that, to instigate changes of scale in a GIS, can be very difficult due to the problem of “generalisation”. Briefly, this means that once mapping entities are drawn at a given scale, then it is difficult to change this scale without the map either becoming too cluttered or too sparse, i.e. simply because the degree of detail varies at each scale. Thus the zooming in or out from a given map can only be realistically accomplished over a relatively small range of scale variations.
It is recommended that before a database is set up, then some thought should be given to the likely mapping scale to be used. It also seems sensible that, for any given fisheries management area, then the marine space should have been divided up into a hierarchy of nested cells. In other words surveying or mapping of any marine related variable, could take place at any chosen scale, and if a consistent hierarchy has been used, then comparisons or overlay functions can routinely be executed using established cellular areas. With regard to cell size or the scale of mapping then certainly before the final mapping output for any specific purpose was required, some experimentation should have been carried out.
Since the subject area of marine fisheries is so wide and since it has been shown that a GIS offers the potential to perform a wide variation of management related functions, then of prime concern is how best to subdivide the subject into recognisable "database areas". Obviously there is no right or wrong way of doing this. Caddy and Garcia (1986) identified eight main areas to which fisheries mapping could be applied (Table 7.1), and undoubtedly there could be others. The way which is chosen will clearly depend on a variety of factors such as those listed in Table 7.2.
|*||Mapping of local environments and production systems.|
|*||For designing statistical data collection surveys.|
|*||For planning scientific resources surveys.|
|*||For preparing inventories of resources.|
|*||For the elaboration of management plans.|
|*||Mapping fishing effort distribution and fishing grounds.|
|*||Mapping in support of international agreements.|
|*||For remote sensing and subsequent mapping.|
|*||The scale of the marine area which was being managed.|
|*||The resources which were available to fund the GIS.|
|*||The degree of urgency behind the need for fisheries management.|
|*||The purposes to which the GIS were to be allocated.|
|*||Any individual or departmental perceptions as to fisheries management needs or priorities.|
|*||The adoption methods chosen to implement the GIS.|
|*||The ways in which the subject field were already being subdivided-perhaps for other purposes.|
|*||The type and quantity of relevant data which already existed.|
Our approach to this problem of dividing up the subject will be from the perspective of implementing a national or regional fisheries management policy, i.e. the approach will be practical rather than academic or theoretical. There will thus be a presumption that data from the whole subject field, plus several linked but perhaps more marginal fields, might need to be included. What we are not doing therefore is approaching the subject from the viewpoint of perhaps an individual fisherman, fishing company, local fishing community or any sub-sector of the activity, or from the viewpoint of an academic institution interested solely in theoretical modelling or simulation. Obviously, any of these sectors could still find a marine fisheries GIS to be of use, but it would not encompass the same functional requirements or emphases.
Meaden (1993) identifies seven database areas which must be of broad relevance to any marine fisheries GIS, and these can usefully form the basis of the subdivisions used in this chapter. However, it is important to note that the seven areas chosen only represent a starting point. Once data from primary or secondary data collection initiatives had started to accrue, then it would soon become apparent that further subdivisions of the database areas would be necessary to meet the real management needs of the time. This need would be reinforced if or when it was found that links to other relevant data sets could be obtained, i.e. with an implication that not all of the datasets would need to be held by the fisheries management team. In fact it is likely that, as the marine GIS became more complex, only a minority of the data needs would be held "in- house", with the majority being accessed via various WAN's. The database areas identified in this section are thus simply a means of classifying data into a range of spatio-related ideas, and within each there would be a huge range of potential subjects. GIS will allow for the functionality of cross border analyses between the seven main database areas identified.
Within each of the database areas, our approach is (i) to look at the range of factors that the area includes in terms of major topics or subjects, (ii) to exemplify the type of non-GIS published spatially related work or research being undertaken, (iii) to review the types of management problems and perspectives, and, remembering that potential data sources have been outlined in Chapters 2 and 3, (iv) to suggest ways in which any specific data could be mapped, modelled or analysed. Our examples of spatially related fisheries work have been chosen for their variety of mapping techniques, and in many cases the original captions have been retained as a means of explanation. The main aim is to be suggestive rather than speculative. It is important that the ideas given on the uses of GIS are both realistic and practicable. It is also important to realise that, within and between each of the seven suggested database areas, GIS technology will allow for not only descriptive analyses of current or past situations, but also for prescriptive formulations based on variable simulations and/or modelling. Given these aims, and the breadth of the GIS potential, then only a small range of possibilities can be introduced here.
Water qualitative parameters on which data might be collected include temperature, salinity, DO levels, water colour/algae content, turbidity, sea currents, density, etc. As well as these marine qualitative parameters, for the success of fisheries management and resource sustainability, it is important to have an indication of habitat types and their spatial disposition. This would include acquiring data on various ecosystems such as coral reefs, sea grass beds, estuarine areas, upwellings, mangroves, sandbanks, etc. It would also include having a detailed indication of oceanographic factors such as sea bottom types, the location of tidal and other fronts plus major and minor gyres, prevailing sea currents, turbidity and bathymetry. Eventually it might be desirable to have more general climatological and oceanographic data concerning perhaps climatic variables, tidal information, wave height, etc. The main sources of data for any of these parameters is shown in Table 7.3.
|*||Marine trawl surveys.|
|*||Marine acoustic surveys.|
|*||Hydrographic charts (paper and digital).|
|*||Topographic mapping (paper and digital).|
|*||Specialist thematic maps or atlases.|
|*||Remotely sensed data - especially data from these sensors:|
|Advanced Very High Resolution Radiometer (AVHRR), Coastal Zone Colour Scanner (CZCS), Radar altimeter data, Sea-viewing Wide-Field-Of-View Sensor (SeaWIFS), High-resolution Multifrequency Microwave Radiometer (HMMR), Ocean Colour Monitor (OCM). Synthetic Aperture Radar (SAR).|
Spatially variable water qualitative data readings may already form the basis of isoline or other maps, and new data being gathered would form the basis of additional databases from which new maps could be compiled. Thus, it is important to note that there already exists a vast amount of marine survey data consisting of geo-referenced point measurements and which, through the use of GIS, could form the basis of interpolated mapped surfaces. GIS manipulative functionality will allow for the easy production of, for instance, spatial correlation or time series analyses between and within any of the mapped variables. RS and acoustic techniques are allowing for huge datasets to be built up on sea bottom typology, and again GIS will enable the mapping of this, as well as the construction of digital terrain models (DTM) which might allow for the visualisation of perhaps relationships between depth and various habitat types. Although there is abundant evidence of the importance of water qualitative factors to fish growth and recruitment, e.g. Rothschild (1986), Gulland (1988), Mann (1993) and Cushing (1995), few detailed longer term spatially based analyses have previously been possible.
Our first example of how GIS could have been effectively utilised concerns the type of modelling study carried out by Werner et al (1993). As Figure 7.7 shows, the authors were attempting to simulate, by the use of passive particles, the fate of the early life stages of cod or haddock which spawn on the Georges Bank in the North West Atlantic. Time series GIS can easily be implemented by the use of temporally interpolated data, and the GIS could usefully have shown various indicators of the degree of relationship between the passive particles and the current speed and direction. Any type of modelling which involves diffusion processes can be comparatively easily simulated. So, for instance, Thomson et al (1992) proposed a conceptual model to show the effect of the North Pacific currents on sockeye salmon (Oncorhynchus nerka) migration routes and their likely landfall location (Figure 7.8) and, as inferred in their paper, this could have been expedited by the use of GIS.
Figure 7.7 Horizontal Disposition of Particles at 30 Metre Depth on Georges Bank, N.W. Atlantic Ocean, Compared With the Main Currents at the Same Depth (from Werner et al, 1993)
Figure 7.9 shows data published by D'Amours (1993) on temperature and oxygen levels for August, 1991 in the Gulf of St. Lawrence. In his study D'Amours sought to show the relationship between the distribution of cod, water temperature and oxygen levels in the Gulf, and for this GIS could have been applied to advantage to achieve mapped interpolated surfaces for use in any type of spatial auto-correlation analysis. Traganza et al (1983) have used satellite data to measure water temperatures in the upwelling off California. From associated shipboard observations they have fitted inferred nitrate readings to the temperature isotherms (Figure 7.10). Obviously GIS based techniques would have allowed for the expansion of this work so as to either incorporate other water qualitative variables, or to have established relationships between water quality and measurements of fishery resources distribution and biomass. Recently, both Mann (1993) and Cushing (1995) have emphasised that it is vital to make spatial analyses between various factors relating to climate, physical oceanography, marine food chains and fish stocks - the use of GIS would greatly enhance our capacity to do this. Finally, examples of gathering data for bathymetric mapping, via the use of acoustic surveys, are given in Mills and Perry (1992) and Somers (1992). The storage of this digitised depth data into a GIS, will allow for its immediate functional use.
Figure 7.8 Simulated Movements of Sockeye Salmon in the North Pacific According to Different Start Locations (A = June; B = July; C = August) (from Thomson et al, 1992)
Here the essential concern is with the mapping and analyses of quantitative or qualitative estimates of marine biomass distributions or densities. Separate databases can be established for any particular geographic area, and/or at any particular time(s), covering criteria such as individual fish species, sea mammals, planktons, cephalopods, shell fish, crustaceans, etc. Obviously these could be further subdivided into genera, subspecies, etc. Most previous mapping of species distribution has simply utilised some sort of quantitative location symbol, such as proportional circles or various classes of shading, to illustrate species density or distributional variations. With the use of specialist GIS algorithms it would be possible to refine distribution maps so that they showed both species density variations and statistical variance (see Section 7.2.4). As also inferred in section 7.2 above, the frequency of mapping species distributions will be a function of both the temporal variability of the distributions and the purposes to which the maps are being put plus, of course, the availability of new data. Table 7.4 gives an indication of the likely sources of data for maps showing biomass distributions.
Figure 7.9 (A) Bottom Temperatures and (B) Bottom Oxygen Levels (Mg/I) in the Northern Gulf of St. Lawrence, Canada for August-September, 1991 (from D'Amours, 1993)
|*||Marine trawl surveys|
|*||Marine acoustic surveys|
|*||Specialist maps or atlases|
|*||Government and other marine research institutes|
|*||Commercial catch records|
|*||Remotely sensed data (satellite or airborne)|
|*||Specialist fisheries or biologic databases|
The usual reason for doing resources surveys or analyses is to gather quantitative data which may be of help in estimating biomass and establishing not only yield potential, but changes in stock numbers or indications of the biological productivity of the water. Not only would GIS be able to provide additional graphical output on this type of information, but it would also allow for more refined analyses such as time series correlations, inter species correlative analyses, contiguity analysis (as an aid to establishing patterns of adjacency) and analyses such as the nearest neighbour index which allows for objective indications of the degree of clustering or dispersal of or between marine species. Clearly all these analytical functions can be carried out between, for instance, fishing catch/effort indicators and biomass quantitative factors.
Figure 7.10 Inferred Nitrate Readings from Temperature Isotherms off the Southern California Coast, USA (from Traganza et al, 1983)
As well as the mapping of biomass density or distributions, this database area must concern itself with the study of those natural and anthropogenic biomass processes which have a strong spatial element. Examples of these would be:
(a) Patchiness - the propensity for both phyto- and zooplankton to aggregate in non-homogeneous patches. The schooling of fish may also be considered to be a kind of patchiness.
(b) Blooms - the huge growth of plankton which occurs mostly in the early spring following increased insolation. This may also refer to other types of rapid reproduction such as the development of toxic algal blooms.
(c) Migrations- many species incorporate migrations into their life cycles. These may be diurnal feeding migrations or annual or periodic spawning migrations.
(d) Displacement- this refers to the change in species dominance which might occur over time in a given area. An example would be the displacement of anchovy by sardines in the Peruvian Current of the eastern Pacific, or the displacement of herring by gadoids in the North Sea.
(e) Enhancement- referring to the selective stocking of marine areas with organisms which have been reared in captivity.
(f) Extinctions- there have been a number of recorded occasions when, for various reasons, species have been found to no longer occur in an area. Clearly it is in any fisheries manager interest to establish the causes for this.
(g) Recruitment - the number of recruits joining the adult stock each year is notoriously variable. More knowledge on the dynamics of this process is of paramount interest to most fisheries managers.
Explanations and understandings of these processes are essential to the long term viability of many fisheries, as well as to the well being of marine ecosystems. Since the processes incorporate temporo-spatial change at various scales, then it is likely that the functionality of GIS's will prove the only way of successfully modelling, simulating or recording the variety of factors underlying any single process.
Studies which exemplify the mapping of marine biomass resources take a large number of forms. They may start from fairly basic accounts which seek only to show a specific species distribution and/or density. They may then attain sophistication by perhaps showing a density or distribution relative to some former temporal period or show a sequence of temporal changes for the species. Spatial analyses might then go on to portray relationships, either between a species distribution and some water qualitative or environmental parameter(s), or between the distribution of one species and one or several others. At a most sophisticated level GIS could be used in the simulation or modelling of factors concerning a species spatial potential or any intra- or inter-species relationships. In the following paragraphs examples of some of these types of potential GIS mapping are given.
Examples of works which show, in a selected variety of ways, relatively straight-forward biomass density and distribution, and for which GIS could have been usefully employed, include Carpentier et al (1989), Isaev and Seliverstov (1991), Lago de Lanzos et al (1993) and Armstrong and Briggs (1993). Some mapped distributions may be extremely basic in that they simply allocate numerical quantities to cells. This is the method used by both Carpentier et al and Isaev and Seliverstov to show respectively the distribution of cod in the English Channel (Figure 7.11) and blue whiting to the west and north of the British Isles (Figure 7.12). Lago de Lanzos plots the density distribution of mackerel eggs in the southern Bay of Biscay in 1990 (Figure 7.13), using two forms of mapping representation. The use of GIS in either case would have allowed for a variety of mapping types, as well as many statistical manipulations and analyses either between cells or between similar maps constructed for successive time periods. Likewise, the Armstrong and Briggs data (Figure 7.14) could have formed the basis of temporal spatial distributional change analyses had a GIS been used.
Studies which show the relationship between a biomass distribution and a water qualitative or environmental factors include Iversen et al (1993), who showed the relationship between anchovy (hatched areas) and sea surface temperatures in the Yellow Sea (Figure 7.15) and Brunetti and Ivanovic (1992), whose study related the occurrence of the early life stages of the squid (Illex argentinus) to water temperature, plus the positions of the Brazil Current and the continental shelf (Figure 7.16). The capacity of GIS to function optimumly in a mode which involves overlaying various classes of data, makes GIS an excellent medium to analyse the types of relationships shown in these examples. The benefit of this type of analytical ability would also have been useful in the situation as exemplified in the study by Shuntov et al (1990) who related fish biomass to the abundance of macroplankton in the Sea of Okhotsk (Figure 7.17).
Besides being of use in the plotting of various types of biomass distribution, GIS functionality can be applied to statistical and resource modelling situations. Thus Simard et al (1992) investigated the possibilities of using various statistical procedures to deal with estimation problems in the spatial autocorrelation of shrimp samples in the Gulf of St. Lawrence. Since they had very detailed data, then, given an appropriately sophisticated GIS package, this complex type of kriging work would prove relatively straight-forward-indeed the maps they produced might well have been produced using a simple GIS (Figure 7.18). Crawford and Fox (1992), using echo sounding data and a graphics program called SURFER, showed a number of ways in which fish biomass could be portrayed- Figure 7.19 shows one example. This work could readily have been accomplished by most GIS packages. There have also been a large number of marine fisheries simulations which have attempted to show how dispersal patterns occur (or might occur) over time for various marine species. Figure 7.8 (in section 7.3.1) showed work carried out by Thomson et al (1992), and MacCall (1990) showed the basis of a diffusion model to simulate the transport of anchovy eggs, from three starting places of the coast of southern California, under the influence of both Ekman and geostrophic flow patterns (Figure 7.20). The use of time series GIS programs would allow this type of modelling under an almost infinite number of variable conditions.
Figure 7.11 Distribution of Cod in the Eastern English Channel — October, 1988 (from Carpentier et al, 1989)
The management and regulation of fisheries is desirable for a number of disparate reasons. Although the obvious reasons are related to the direct necessity to guarantee that fish stocks are sustainable, i.e. so as to ensure the maintenance of the fishing industry, the back-up industrial sector and to ensure the continuance of supply, there are also less obvious reasons. Thus management is necessary, in an economic sense, to ensure a fair distribution of the resource and to help improve the efficiency of the industry. From a social sense it may be desirable in order to maintain the continuance of fishing communities, or for the protection of public health. And from the environmental viewpoint, there are a number of important management perspectives, e.g. in order to enhance marine systems biodiversity it is essential that monitoring, protection and possibly improvement tasks are undertaken, and it is also important to realise that a total ecosystems approach to fisheries sustainability is now being seen as desirable (Alexander, 1993). Management is additionally necessary in the sense that fisheries are one of several economic and social activities which may be competing for marine space. Further details on spatial aspects concerning fisheries management and regulation can be found in Arnason (1991), Morgan (1991), Waters (1991), Hinds (1992), Symes (1992), Pearse and Walters (1992), Green and Stockdale (1993), Holden (1994), and the journal “Marine Policy” is very informative. Management policies emanate from various hierarchical levels. At an international scale fisheries policies may be activated by institutions such as the United Nations or the European Community. At other levels there will be national, regional and even local sources of management decision- indeed there are moves towards encouraging fisheries management to operate at the local community level (McGoodwin, 1990, Siar et al, 1992, Dyer and McGoodwin, 1994). Many of the regulations are formulated by multilateral regional fisheries bodies, examples of which are shown in Table 7.5. At whichever level the policies come from, there will be a need to set out rules, to monitor them and, if necessary, to adjudicate or arbitrate as to their compliance. Different hierarchical levels will usually be entrusted with making management policies covering differing spatial scales, and of course it will be necessary to make agreements on a bilateral or multilateral basis among and between the various levels. The original sources of data for a GIS which outlines management and regulatory areas or zoning will usually be from legal documentation or bulletins produced by the various authorities concerned, although a are shown in Table 7.5. At whichever level the policies come from, there will be a need to set out rules, to monitor them and, if necessary, to adjudicate or arbitrate as to their compliance. Different hierarchical levels will usually be entrusted with making management policies covering differing spatial scales, and of course it will be necessary to make agreements on a bilateral or multilateral basis among and between the various levels. The original sources of data for a GIS which outlines management and regulatory areas or zoning will usually be from legal documentation or bulletins produced by the various authorities concerned, although a miscellaneous variety of secondary sources often provide the necessary information, including most of the regional fisheries bodies and maritime atlases.
Figure 7.12 Distribution of Blue Whiting to the West and North of the British Isles (from Isaev and Seliverstov, 1991)
Figure 7.13 Dot Map Showing Total Mackerel Egg Abundance in a 1990 Survey, and choropleth Map Showing Daily Mackerel Egg Production in the Same Period - in the Southern Bay of Biscay (from Lago de Lanzos et al, 1993)
Figure 7.14 Distribution of the 1991 Year Class of Haddock in the Irish Sea for September, 1991 and 1992 (from Armstrong and Briggs, 1993)
Figure 7.15 Mean Surface Temperatures and Distribution of Anchovy in the East China Sea During March (from Iversen et al, 1993)
Figure 7.16 Water Qualitative Parameters (A) and the Abundance of Juvenile Squid (B) off the North Coast of Argentina in Summer, 1989 (from Brunetti and Ivanovic, 1992)
Figure 7.17 Distribution of Macroplankton Biomass Compared With Fish Catches in the Summer of 1998 in the Sea of Okhotsk (from Shuntov et al, 1990)
Figure 7.18 Maps to Show Shrimp Biomass Contours of 5, 500, 1000 and 1500 kg/km² in the Gulf of St. Lawrence, Canada in 1988. (+ = sampling point < 1000 kg/km²; * = > 100 kg/km²; o = non-sampled areas)
|*||International Council for the Exploration of the Seas (ICES)|
|*||North-East Atlantic Fisheries Commission (NEAFC)|
|*||North Atlantic Salmon Conservation Organization (NASCO)|
|*||Northwest Atlantic Fisheries Organization (NAFO)|
|*||General Fisheries Council for the Mediterranean (GFCM)|
|*||Fishery Committee for the Eastern Central Atlantic (CECAF)|
|*||Western Central Atlantic Fishery Commission (WECAFC)|
|*||Regional Fisheries Advisory Commission for the Southwest Atlantic (CARPAS)|
|*||International Commission for the Conservation of Atlantic Tunas (ICCAT)|
|*||International Commission for the South East Atlantic Fisheries (ICSEAF)|
Figure 7.19 3-D Plot of Water Column Biomass Along a 3.5 km Track Using Data from an Acoustic Survey (from Crawford and Fox, 1992)
It is important to point out that the tools for management, allocation and regulation may be either directly or indirectly spatially related. Directly spatially related regulation is where all or part of a fisheries area is partitioned in some way for management purposes. Thus, for instance, King (1995) shows how an ecosystems approach to management may set aside various “Marine Protection Areas” in which access and exploitation are controlled (Table 7.6). Clearly, it would be a simple task to map these regulation zones. Regulation may be indirectly spatially related in the sense that the regulation itself has nothing directly to do with specific areas, e.g. limiting the efficiency of types of fishing gear, but the regulation or management policy could well be assigned to specific spatial areas.
|* Preservation.||No access.|
|* Wilderness.||Access allowed, but no exploitation.|
|* Recreational.||Regulated amateur fishing.|
|* Traditional fishing.||Subsistence fishing by people living on coast.|
|* Scientific.||Authorised research purposes.|
|* Experimental.||Fishing controlled at different levels to assess effects of exploitation.|
Figure 7.20 Cell Diagram for a Transport Simulation of Anchovy Eggs in the Pacific Ocean off Southern California, USA (from MacCall, 1990)
The actual division of marine space for management or regulation can take several forms. Initially, under a natural classification, marine areas often form obvious units such as “the Mediterranean Sea”, the “Black Sea” or the “Persian Gulf”. Then there has recently been a move towards recognising marine areas in terms of “large marine ecosystems” (LME's), i.e. areas of the sea which have large scale unified hydrographic regimes and trophically related populations of marine organisms. Figure 7.21 shows the LME's which have currently been defined. Very recently the concept of the marine catchment basin (MCB) has been proposed (Caddy, 1993). This refers to the idea that an important way of examining marine areas, especially closed or semi-enclosed seas, is via an appreciation that a river basin is inextricably linked with the marine area into which its waters flow. Thus, for instance, the problems which are so obvious in the Black Sea stem in large part from inputs which are received there from inflowing rivers such as the Danube, the Don and the Volga.
Figure 7.21 World Map of Large Marine Ecosystems (from Alexander, 1993)
For a long time now, most of the major marine areas have been officially sub-divided into fisheries zones by some of the main fisheries regulatory bodies. Thus for instance, there are official ICES zones which cover much of the north east Atlantic area, and there are NAFO regulatory areas in the north west Atlantic (Figure 7.22). Other forms of spatial division, as manifest from a legal viewpoint, are the 200 mile Exclusive Economic Zone (EEZ) and the Exclusive Fishery Zone, and these might form the mapping boundary for any national fisheries GIS. Figure 7.23 gives examples of these zones as they apply in the south west Pacific area. Within any of these units of marine space, for both management and regulation purposes, it may be necessary to further subdivide the area, possibly using a hierarchical nested cell structure, i.e. a more refined version of the “ICES rectangles” which are presently used in the North Sea.
Figure 7.22 Example of Northwest Atlantic Fisheries Organization Divisions and Subdivisions
Once the unit areas for management have been established, then the areas form the basic units into which variables of regulation are plotted. There are a large number of ways in which regulation is administered which we cannot go into detail here. However, we should note that the type of regulatory variables which we are conceiving of include catch regulations (which might include factors such as net mesh size, length of drift net, season of fishing, etc,) plus fishing rights, quotas and total allowable catches, closures, plus indices of fish availability such as maximum sustainable yield or even the amount of restocking per cell which had taken place. The different ways of recording and mapping these regulatory variables would need to be considered before any GIS output could be obtained. It should be mentioned here that fishing fleet tracking using satellite systems is now seen by many governments as the most practical way of maintaining checks on the operations of fishing vessels, and this “spy-in-the-sky” system is likely to be increasingly utilised, especially since the monitoring of fishing for migratory and straddling stocks has now been agreed upon. If these system store their records, as they have the capability to do, then it will form a valuable GIS data resource for mapping the logistics of specific vessel activity.
Figure 7.23 200 Mile Exclusive Economic Zones in the South Western Pacific
An initial example of how GIS might be employed in delimiting access rights is given by Bidi (1993) with reference to the Gulf of Guinea coastal states in West Africa. He suggests that the twelve coastal nations from Mauritania in the west to Nigeria in the east, will need to have a degree of “flexible co-operation” in their approach to managing the limited fish stocks of the area. In the neighbouring area to the north, i.e. between Guinea Bissau and southern Morocco, management of the prolific fish stocks urgently needs to be strengthened because of the almost total lack of control on access rights (Goffinet, 1992). In this area, the rich upwelling of the Canaries currents ensures a high natural productivity and thus the fisheries potential is huge. However, the nearshore coastal states are both thinly populated and comparatively impoverished. Although a regional management organisation has been set up, the Committee for East Central Atlantic Fisheries (CECAF), it has no enforcement capabilities. The area has been divided up into the divisions as shown in Figure 7.24, with the coastal states being responsible for specific areas. These divisions will form the basis of data gathering units to which GIS functionality should be able to usefully contribute. The granting of territorial use rights for small scale fisherfolk in almost any country, under community based management schemes, is seen as a major way of enhancing the livelihood of the people involved (Ruddle, 1987). The examples in this paragraph all involve rights of access and the right to determine levels of fishing effort and catches with the necessary legal backing. The management of such schemes would usefully employ GIS techniques at micro or macro scales, so allowing for GIS learning to occur in conditions where data gathering was unlikely to be too overwhelming.
Figure 7.24 Committee for East Central Atlantic Fisheries (CECAF) Divisions off the North West African Coast (from Goffinet, 1992)
In the small and heavily overfished North Sea, Symes (1992, p.336) recognises that “effective central control is essential”, and that this must incorporate the need to vary the system of regulation to suit ever changing conditions of stock availability and consequent Total Allowable Catches (TAC's). Similarly, in the Peruvian and Chilean upwelling coastal waters, which have witnessed huge stock fluctuations from both natural and human intervention reasons, Caviedes and Fik (1993) advocate strict resource management and restrictions on fish capture. In a final heavily fished area, that off the eastern seaboard of China, we illustrate (Figure 7.25) how the marine area has been zoned such that certain fisheries related activities are allocated to specified zones. Each zone is managed under a complex set of regulations as laid down under the 1986 Chinese Fishery Law, and as explained by Wang and Zhan (1992). In all of these studies, which relate to heavily fished areas, GIS has the functionality to address a range of aspects of management.
Figure 7.25 Fishery Zones in Part of the East China Sea (from Wang and Zhan, 1992)
This section covers “effort” and “catches” since, although they are obviously describing and measuring two distinct variables, there are many ways in which they are both strongly related and inextricably linked. It is also important to mention here that in some ways it is difficult to separate the previous section (7.3.3) on management and regulation from this database area. This is because nowadays a large proportion of the world's fisheries have their effort and catches managed by the various regulatory bodies which have been referred to, and there is plenty of evidence that recent regulatory measures have had an impact on fisheries effort and catches, e.g. Bergin and Howard (1992) and Symes (1992). However, from a GIS organisational viewpoint, it is easier to separate management from effort and catch factors. Thus our consideration of management was mostly from the viewpoint of establishing zonations for different regulatory purposes, whereas our considerations relating to effort and catches relates to the measurement of these and to the ability of assigning measurements independently to any spatial zones.
Any evaluation of fishing effort encompasses a variety of techniques and considerations. The common forms of measurement are by indices such as those shown in Table 7.7. At a different level it is also possible to measure effort in terms of the regional or national investment in the fisheries industrial structure and infrastructure. So it can be conceived that there must be varying returns from the fishing activity which are related in some way to the total amount of capital invested in the industry. At one scale investment may be in large scale infrastructure such as fishing harbours, freezing plants, repair yards, etc. At a very different scale, effort may be evaluated in the form of the numbers of small fishing craft which are used by local sea-shore communities. At either scale, the amount and distribution of fisheries investment should form part of the basic data which would be collected for GIS purposes. It is important to note that so- called “technology creep” is occurring whereby inputs of effort are becoming more effective due to the adoption on fishing vessels of technological aids. This often makes it difficult to carry out time series analyses or inter-regional comparisons on fishing effort. Effort is increasingly being subject to monitoring and regulation, and this is being effected by means such as licence limitations, log books and satellite monitoring of vessel activity. Hilborn and Walters (1992) and Anon (1993) both give very good descriptions of the modern practices and problems connected with evaluating and monitoring fishing effort, the former covering all aspects and the latter as it relates to European Community countries.
|*||Weight of fish caught per hook per hour.|
|*||Number of lobsters caught per trap per day.|
|*||Weight of fish caught per hour of trawling.|
|*||Number of hours a vessel is at sea.|
|*||Proportion of hooks on a long-line which have caught fish.|
|*||The working cost of maintaining a vessel at sea per time unit.|
|*||Catch per unit of a vessel's horsepower.|
|*||Catch per man hour fished.|
The monitoring of catches must be one of the core tasks in any fisheries management program. There are a number of obvious reasons for this which are mostly concerned with gaining information on the longer term trends in the welfare of the stock. For GIS use, and for other purposes, it is useful if the catch data per vessel can be disaggregated as much as possible, i.e. by species, by place of capture, time of capture, method of capture, etc. For research purposes additional data is also necessary such as condition, age and size of individuals in the stock. It is also normal to record catch data (or fish production data) by species or product types (chilled, fresh, frozen), or by commercial group sizes, at the ports of landing. If a wide range of categories of catch data is made, then there is the potential to present information in a useful way; unfortunately this has not always been the case in the fishing industry because fishery operations are carried out in a wide range of different communities which themselves are often scattered around long coastlines, and which utilise a variety of recording methods. Additionally, it has often proved too complex and expensive to even justify setting up the necessary data gathering mechanisms.
The acquisition of data or information on fishing effort and catches is often difficult to achieve in a form which is useful to a GIS since, as inferred above, it is often of dubious accuracy or is incomplete, it will seldom be disaggregated to a desirable degree and it will often only be available in very mixed formats, i.e. making time series, or inter-area, analyses impossible. Data which is available is usually held by the appropriate government fisheries department for a region or country. Sometimes individual ports may have local statistics, as may various fishery co-operatives. Marine research institutes might have data available, though this is more likely to refer to stocks or survey catches rather than to commercial catches. It is important to note that sometimes it may prove possible to use proxy data to estimate likely catches, especially where fishing is carried out at a subsistence level. In attempting to secure catch data it should not be forgotten that many catches may be made by licenced (or unlicensed) foreign vessels. In some fisheries by-catches may need to be considered, or estimates might need to be made for discards at sea.
It is not necessary to give many specific examples of the mapping of effort or catch since these will be like many of the other maps shown, i.e. spatially distributed proportional circles or shaded choropleth cells. Meaden and Kemp (in press) show a programme which will be specially adapted to a commercial GIS package, and which is conceived so as to portray either catch or effort data which has been derived from GPS locational output. Here, the intention is to show the trawl path for each individual haul, and to show the catch data from this haul. Aggregations of the catch data can then be made which can be allocated to either cells or to other unit areas of management (Figure 7.26). It will clearly then be possible to use the GIS capabilities to ascertain any spatial relationships which might exist between catch or effort and any of the other database variables, e.g. the GIS can be asked “What is the relationship between unit of effort and sea bottom type in area x?”.
Regarding point effort capitation (the amount of fisheries investment made at any one place), Symes (1992) produces two maps (Figure 7.27) which show UK demersal fish landings for 1975 and 1986, i.e. before and after the imposition of the 200 mile EEZ. GIS could usefully be employed in any analyses of effort variation, e.g. in the Symes case it would have been simple to produce a further map which showed the difference between the two maps (or the net effect of the 200 mile EEZ imposition). At a world scale, using FAO statistical data, Juda (1991) analyses changes in fish catch distribution between different temporal periods. His lengthy study produces 14 tables of figures but no maps (e.g. Table 7.8). The use of GIS would have greatly enhanced the visual interpretation of this data, and again it would have allowed for a large number of more detailed spatial and temporal analyses. Detailed maps showing the spatial distribution of catch per unit of effort (CPUE) have rarely been produced, though Miyabe and Bayliff (1990) produce a map on this topic for bigeye tuna in the eastern tropical Pacific (Figure 7.28). This map was probably produced using a commercial computer mapping package. GIS functionality would be employed in working out why the spatial variations occurred, i.e. since such maps should show that theoretically there ought to be little variation in CPUE over space.
Figure 7.26 Hypothetical GIS Output to Show Fishery Catches (or Effort) by Trawl Hauls for Specific Areas
Figure 7.27 UK Demersal Fish Landings by Ports for 1975 and 1986 (from Symes, 1992)
Although this sector of the industry might not seem very relevant to a marine fishery resources GIS, it needs to be included since the efficient disposal of fisheries products is becoming of increasing importance and since fishery products are increasingly being disposed of (sold) at the international scale. Any considerations of markets in this section will be confined to coastal (point of unloading) markets, i.e. it will not consider any form of retail or end user markets, (except in the sense that in many subsistence economies the products will usually be disposed of directly to end users living in beach or coastal communities). We will also not be including here the transference of fish at sea from perhaps catching to processing vessels. Finally, we will not be considering fishery product marketing per se, i.e. the ways and means of best getting the products to the customers.
Markets, from the fisherman's point of view, represent the coastal unloading point. The form that the market takes will vary, as may the form in which the fishery product is unloaded. So the fish, or other form of commercial marine biomass, may be in a fresh state, may be chilled or frozen, or it may have been processed in some way. At the unloading point the product may be stored, in a suitable environment, usually for later shipment, or it may be processed and packaged into a number of forms or products. Some record of the quantity of fish which is handled in the various ways is usually kept, often by buyers from the company doing the processing or the marketing, but also by the port authorities or by fishermen co-operatives. It may be expected that, since we are dealing with a natural “organic” product, then there will be large fluctuations in the market routine. This might include seasonal variations in supply, irregular changes in product requirements, variations in demand, etc., and the market for fish products may also be strongly related to competition from alternative meat protein sources. So markets for fish and fishery products are highly volatile both from a temporal and spatial perspective. This is why the mapping of markets can be very important.
Figure 7.28 Spatial Distribution of Catch Per Unit of Effort for Bigeye Tuna in the eastern Pacific
For GIS purposes, market information obtained by the various port or marketing authorities, will need to be stored in a database system, from where it can be extracted for mapping. Since markets will represent coastal points, then mapping is inevitably via the use of proportional circles (as shown in Figure 7.27). Circles can represent either simple totals of fish landed at a port, or they can be progressively refined so as to show volumes of individual products or fish species. Indeed they can be mapped as proportional “pie charts” which give even more information. Where GIS can potentially be very useful is in its capacity to instantly map out both spatial and temporal changes in the various facets of the market, e.g. long term trends or fluxes in can be identified and analysed. It can also be of use in seeking to establish, for instance, under-utilised markets. Here it would a case of matching fish sale and/or consumption levels with a map showing population density and distribution, in order to identify spatial differences and perhaps unexploited markets. GIS has the potential to graphically manage the disposal of fish from catching vessels to on-shore processing plants. This is very important since both vessels and plants have physical capacities which need to be finely managed. Finally, GIS can be of use in the analysis of spatial differences in species or product preferences for any specific market area.
|Top 5 states||50.12%||46.23%|
|Top 10 states||63.70%||62.88%|
|Top 15 states||73.39%||75.12%|
|Top 20 states||80.61%||82.36%|
This need not be major sector of most marine fisheries related GIS's, though its inclusion is considered important for a number of reasons. Some countries now have either a large proportion of their total fisheries output of some species produced using culturing techniques, or the actual volume of cultured fisheries products may be high. Also, in some regions, such as the Atlantic seaboard of Canada, normal marine fish production has been severely hit and the government is turning to increased mariculture potential as one means of providing local employment and income opportunities for former fishing communities. Similarly, fish culturing has been seen as a major way of boosting the economies of some isolated coastal regions in developed countries, and of providing both incomes and protein in coastal settlements in some developing countries (Loayza and Sprague, 1992).
It is also important to include mariculture here since the proportion of fish produced in this way is increasing faster than it is from traditional marine sources (at about 15% per annum), and this trend is likely to continue. The products of mariculture will therefore be increasingly competing in fisheries markets, and the activity itself will be competing with traditional fisheries for (and in) certain aquatic environments, near coast unprotected waters as mariculture zones offshore as well as e.g. fjords, estuaries and mangroves, and this will not be without a certain impact. For instance, Black and Truscott (1994) provide a detailed account of how the Provincial government in British Columbia, Canada, will only issue licences for mariculture after the applications have been vetted for their technical feasibility in relation to the biophysical capability of the environment. All likely mariculture sites here have been monitored and mapped in a “suitability hierarchy”.
For two other reasons it is important to consider mariculture and fisheries together. The first is that both usually are under the same administration and the second is that there can be both spatial and market competition between the two.
The types of mariculture that we are considering are mainly those which is carried out in floating sea cages, and the rearing of various marine shellfish under different production systems. Cage production is usually for the intensive rearing of high value species such as salmonids, sea bass and bream. We are ignoring land based culture systems of marine species, such as that used for rearing eggs and fry to be placed in the cages, and the rearing of shrimp in brackish water coastal ponds, not because they are not important, but because these systems can best be described as being synonymous with fresh water aquaculture in terms of their input requirements, and this topic (with regard to GIS) has been recently reviewed by Beveridge and Ross, 1991 and by Meaden and Kapetsky, 1991. Also, the types of consideration regarding shrimp ponds are looked at in section 7.3.7 covering coastal zone management.
It is difficult to give explicit examples of studies where GIS might have been used, i.e. rather than traditional forms of spatial description and analyses, because there are very few mariculture studies which have sought to introduce a spatial element. Exceptions have been that of Ibrekk et al (1993), who did a nationwide search of the Norwegian coastline to assess the suitability for aquaculture, and who reported that GIS would indeed have been a useful tool to use, and the study by Silvert (1994) in which he outlined a computer based simulation model to help with aquaculture site selection. There are however, a number of aquacultural studies which have utilised various GIS, or remote sensing, techniques and these have been given as case studies in Meaden and Kapetsky (1991) and by Kapetsky and Travaglia (1995). It is interesting to note here that the World Bank has recently said, with regard to implementing small scale aquaculture:
“The appropriate use of Geographic Information Systems and remote sensing technologies in close cooperation with FAO would be highly desirable.” (Loayza and Sprague, 1992; p56)
In this section we have therefore considered it useful to simply describe a number of ways in which GIS could be usefully employed as an aid to mariculture development. Kapetsky and Travaglia (1995) provide a useful list of additional ways in which GIS could be of use to mariculture.
For mariculture to be successful, it is essential that the sea cages, or the artificial shellfish beds, are located correctly. Here correct location can be considered both from a macro and micro viewpoint, i.e. a good location perhaps somewhere along hundreds of kilometres of coastline, or a good location within a fjord, estuary, bay or lagoon. To establish optimum locations it is first necessary to identify the relevant production criterion for the particular mariculture activity. Table 7.9 shows production criteria established by Cordell and Nolte (1988) which might control the mariculture of oysters. It is then necessary to draw up maps for each of production criterion, showing the spatial disposition of them in a way which can be interpreted as being from “ideal” to “impossible”. So, for instance, the depth of water for the mooring of sea cages is an important mariculture production criteria. A bathymetry map can be obtained or drawn up for any area under consideration. Areas of the sea lying at certain submarine contour depths can then be delimited as being, for instance, “ideal”, or “good”, or “fair”, or “poor”, or perhaps “impossible”. Once all the maps have been drawn up in this way, then a GIS is an invaluable tool which allows the various mapped criteria to be overlayed (or superimposed) so as to collate all the data and to allow optimum locations to be established. Various studies using this type of methodology for aquaculture location have been attempted, e.g. Mooneyhan (1985), Meaden (1987), Kapetsky (1989), Kapetsky et al (1988), Kapetsky et al (1990), Ali et al (1991) and Kapetsky (1994).
|a)||Proximity of water and shoreline.|
|b)||Water depth at low tide.|
|c)||Protection from excessive wave action.|
|d)||Existence of stream mouths.|
|e)||Proximity to potential or actual landslide areas.|
|f)||Electric power availability.|
|j)||Proximity to markets.|
|k)||Proximity to labour.|
|n)||Water temperature - summer and winter.|
|o)||Salinity - summer and winter.|
A second way in which GIS could usefully be used for mariculture purposes, is in monitoring the effects that mariculture might have on the local environment. A number of studies have shown that cage production has the potential for degrading the environment in various ways (Fernandez-Pato, 1989; Beveridge and Ross, 1991; Pollnac, 1992; Pillay, 1992; Earll et al, 1992; Cook and Black, 1993). Thus Pillay (1992), in a very detailed analysis, shows the following environmental impacts (Table 7.10).
|*||Conflict with other users - for land and water.|
|*||Sedimentation and the obstruction of water flows.|
|*||Effluent discharges - mostly waste feed and faeces.|
|*||Hypernutrification and eutrophication of water.|
|*||Introduction of exotic species.|
|*||Transmission of diseases.|
|*||Loss of local biodiversity.|
|*||Encouragement of predators.|
|*||Danger of hybridization and reduced genetic diversity.|
|*||Creation of mono-scenery.|
The use of GIS in terrestrial environmental planning is already well established, and its adoption for monitoring and managing mariculture could follow similar methodologies. Thus, for instance, one of the major problems for cage culture is the excessive build up of undesirable detritus underneath cages. The use of GIS would allow for spatial records of this to be kept, for nearby alternative mooring sites to be located (probably on a rotating basis), and for a study of the relationships between detrital build up and other water or benthic parameters.
A further way in which GIS could usefully be applied is in comparability studies. It is in everyone's interest to have records which compare the success of different enterprises operating either under different conditions or in different locations. This is especially true for activities which have production criteria which are highly variable in the spatial domain. So, if some mariculture facilities are being more successful than others, then it is likely to be for reasons relating to variable production criteria inputs. GIS is an excellent tool to reveal this. This factor is likely to be increasingly important in the near future, i.e. with the pressures on the availability of sheltered marine space. Then there will be a move towards the greater use of deep ocean moored cages. Sites for these will need to be carefully chosen and entrepreneurs will wish to monitor their financial viability with some care.
Finally, GIS is presently seen as an aid to the development of mariculture but will be increasingly used as a tool for its management.
Before we investigate parameters relating to the use of GIS for managing the coastal zone, it is important to gain a definition, or at least guidelines, on what is meant by the coastal zone. Several authors have noted the difficulty of precisely defining this (Carter, 1988; Sorensen and McCreary, 1990; Jefferies-Harris, 1992; Kam, 1992), though some facets are obvious. Firstly, there must be a land/sea interface along which there will almost certainly be some degree of marine influences on the land and terrestrial influences on the sea. Thus we would expect to include an element of both land and water into our database parameters. The width of this zone must be variable, but it will usually accord to the strengths of these two way influences. Secondly, the coastal zone must have at least 2.5-D in the sense that coastal waters will have depth, and the shoreline must have some height. Thirdly, there will be both natural and completely artificial shorelines and hinterlands, the extent of which usually accords with a measure of human population density. Carter (1988) and Clark (1992) give illustrations of the different ways in which the coastal zone has been interpreted by various authorities. From a practical viewpoint, the actual definition of the coastal zone which is finally decided upon will be a reflection of the specific purpose to which the marine GIS is being used.
As with the database area of mariculture, at first sight there may appear to be little direct link between a marine fisheries GIS and the database area concerned with coastal zones. However, there are several reasons why this subject area needs some notification. The vast majority of marine species spend at least part of their life cycle in the shallow coastal shelf waters. This period may be for spawning purposes, it may be in occupying nursery grounds or, more frequently, it is for general feeding purposes. So inshore waters play a significant role in species developmental, and as such their existing ecosystems well-being is of profound importance. Given that there were no anthropogenic impacts on the coastal region, then these coastal inshore ecosystems would be able to maintain a self regulating equilibrium i.e. where perhaps long term changes were slowly occurring but where short term variations would be substantially in balance. Although this equilibrium is undoubtedly still able to be maintained, i.e. to a greater degree in perhaps 50% of the worlds coastal waters; in the other 50% varying degrees of degradation are taking place. The critical coastal zone systems which are threatened, and which play a significant role in marine species ecology are listed in Table 7.11.
|*||Submerged seagrass meadows and kelp beds.|
|*||Lagoons and embayments.|
|*||Intertidal marshland and mudflats.|
|*||Beaches and wave cut platforms|
Degradation is a direct result mostly of human activity along the coastal zone. There are a wide range of possible “negative” human activities, some of which are shown in Table 7.12. Obviously there will be enormous spatial and temporal variations in the extent and degree of any of these activities. Nevertheless, as human populations increase, then the impacts become progressively worse, and coastal waters are now quite rapidly becoming less suitable habitats for most species. One factor which has exacerbated this deteriorating situation is the fact that so many of the human activities shown have taken place, for management and jurisdictional purposes, under separate governmental sources of authority. This has meant that action to alleviate problems has been diffuse, slow and difficult to organise. de Freese (1991) gives an excellent summary of the problems in getting coastal zone management programs underway. However, now that the environmental and ecologic situation is so bad in many areas, joint remedial action is starting to materialise, as is the desirability of managing the coastal zone as a complete integrated ecosystem and/or unit of management (Clark, 1992). It is in both the government's and fisheries manager's interest to investigate any causal reasons and links in the interactions which take place between the various sources of coastal zone pressure.
|*||Exploitation of offshore oil and gas deposits.|
|*||Extraction of sand and gravel aggregates.|
|*||Deforestation of coastal timber resources.|
|*||Water based recreation activities.|
|*||Marine disposal of various wastes and sewage effluents.|
|*||Growth in the desire for coastal urban residence.|
|*||Sedimentation from various land based activities.|
|*||Port activities and vessels using coastal waters.|
|*||Development of mariculture activities.|
|*||Establishment of coastal based industry and infrastructure.|
|*||Destruction of coral reefs.|
|*||Drainage of coastal wetlands.|
|*||Intensification of land-based recreational activities.|
|*||Construction of coastal protection schemes.|
|*||The damming of major rivers.|
|*||Land filling to provide more real estate.|
|*||The dredging of navigation channels.|
As well as the fact that the ecosystems quality of coastal waters are increasingly influenced by inshore and onshore human activities, there is another reason why this is an important area for study. There has been a great deal of publicity over the last decade concerning global warming and subsequent sea level rises (see Woodworth, 1993 for a summary of this). Although there has been some debate about definitive causes and effects of this, the results of sea level rise will have profound effects on coastal regions and on fisheries potential. Bigford (1991) gives a useful summary of the fisheries related effects, noting likely changes in habitats, species distributions and fishery yields. It is also important to mention briefly that the coastal zone is particularly threatened by both man-made disasters, such as oil-tankers running aground, and by natural hazards such as hurricanes, tsunamis and storm surges. Because of this, many governments and other authorities have seen that there is an urgent need to manage coastal developments, and worldwide there are many initiatives, e.g. see Crawford (1992), Grip (1992), Halliday and Smith (1992), Suarez de Vivero (1992).
Any database which is established on facets concerned with the coastal zone will inevitably be enormously complex in its range of data holdings. The reason for this relates not only to the fact that the sea/land interface includes an enormous number of varied natural environments, but also because the range of human activities here is large, and this has given rise to a multiplicity of land use types. Another reason for complexity is the necessity of collecting data at a variety of temporal and spatial scales. Thus, not only will factors such as actual land use variety and its rate of spatial variability be varied, but so will the scale of the processes which are occurring. For instance, some coastal areas may be geomorphologically static and change will be slow, whereas in other areas very rapid rates of erosion and deposition are occurring over micro or macro areas. Obviously human generated processes, e.g. marina developments, mangrove clearance and beach resort development, can lead to very rapid spatial changes, which again will occur at a variety of scales. Despite these complexities, there have been attempts at defining suitable database areas for the collection of coastal related information and data. Thus, Figure 7.29 gives an example, taken from Riddell (1992), which gives a strong suggestion as to the types of data which need to be procured. In view of the data complexities, we cannot realistically attempt to suggest data sources on coastal zones and their management, except to mention that many of the sources suggested in Chapter 3 may be relevant, and additional useful information is given by Bartlett (1994).
Unlike the other potential database areas for a marine fisheries GIS, i.e. those discussed in sections 7.3 to 7.8, there has already been a significant amount of coastal zone mapping undertaken using GIS techniques and facilities. We can give a brief resume of some of the work here and some more detailed accounts are presented as case studies in Chapter 9. We can also draw attention to the recent publication by Bartlett (1994) which specifically outlines the impact that GIS has had on coastal zone studies.
Figure 7.29 Structure of a Coastal Zone Management GIS Giving Ideas Of Database Areas (from Riddell, 1992)
In an interesting study which looks generally at the use of various mapping techniques that have been utilised for the purposes of coastal planning, Cendrero (1989) gives a range of examples from different parts of the world. Most of them are fairly complex choropleth maps showing land use detail, and they are accompanied by keys which typically show land use classifications or morphological features. An interesting early example of a “GIS” based map is shown in Figure 7.30. Here a number of spatially variable criteria have been combined, and then weighted, in order to produce a “map” showing the suitability of a coastal area in north east Spain to be designated as a natural park. Class 5 on the map is most suitable; 1 is least suitable. A simple programme was written so that the necessary calculations could be made, and the output from the programme was then “mapped” directly to the VDU, from where hardcopy could be printed.
Figure 7.30 “Map” Showing the Suitability of Parts of North East Spain to be Designated as a Natural Park (from Cendrero, 1989)
There has been tremendous developmental pressure on some of the remaining saltmarsh islands along the east coast of the USA. Preservation of the marshland, and other aquatic ecosystems, here requires detailed and current information on all those factors shown in Figure 7.31. Welch et al (1992) integrated the use of remote sensing, GPS and GIS techniques to produce accurate digital databases on these parameters. Using a series of RS images, plus other hardcopy data, they were able to use GIS functionality to accurately plot land use changes on one particular island, and to measure the rate of logging, the rate of wildlife habitat loss and to measure changes in the size of the island.
Figure 7.31 Sample Layers used in the Sapelo Island Integrated Resource Database (from Welch et al, 1992)
There has been a large amount of coastal zone mapping and GIS work done by a group of Dutch and German workers covering various parameters associated with the Wadden Sea, which lies along the northern coasts of Holland and western Germany, e.g. Dijkema (1991), Schauser et al (1992), Liebig (1994a). Much of this work has involved setting up a large and complex database - the Wadden Sea Information System (WATIS), which operates in connection with another database - the central Wadden Sea Data Base (WADABA). Much of the output from these databases has been in the form of complex modelling of factors such as wave height predictions, sediment transfer routes, current trajectories plus various tidal models. To visually display what is happening along the coast here, ARC/INFO GIS has been used, and one example of the type of GIS output obtained is shown in Figure 7.32. Other examples of this work are shown in Figures 5.9 and 9.6. Although these Figures exemplify water based parameters, other work has been connected with studying human based pressures on the fragile coastal ecologies, plus work on the biological sensitivity of mudflat areas to human intrusion.
Figure 7.32 Currents and Bathymetric Contours in Part of the German Wadden Sea Coastal Area (from Liebig, 1994b)