Ecologists, Farmers, Tourists - GIS Support Planning of Red Stone Park, China
( This paper is printed from: Geographic Information Research：Bridging the Atlantic (Craglia, M. and Hellen, C. Eds.)Tayor &Francis. pp. 480-494.1997

ABSTRACT: Landscape planning is considered as a procedure of defense taking place among defenders of various processes. Defending by security patterns (SPs) --strategic portions and positions -- may significantly increase the efficiency of safeguarding the processes of our concern. SPs are defined and identified based on the threshold-type quality of the dynamics of the processes. Alternative change models are proposed based on SPs. Decision making based on SPs is also discussed. The SP approach is illustrated by the case study of the Red Stone National Park, China, in which a defending procedure among ecologists (defenders of ecological processes), tourists (defenders of visual perceptual processes) and farmers (defenders of agricultural conversion processes) is simulated. Integrated with the SP approach, GIS shows great potential for supporting decision making in landscape changes.

INTRODUCTION SECURITY PATTERNS AND SP APPROACH
Landscape planning is considered a procedure of defense involving defenders of various processes. How can we defend the processes of our concern more effectively while maximising opportunities for changes? This paper tries to answer this question using the concept of security patterns (SPs) and demonstrates how GIS can be combined with the SP approach in landscape planning.
By definition, SPs are the spatial patterns composed of strategic portions, positions, critical scales (sizes), numbers, shapes and inter-relationships that are associated with certain thresholds in the non-linear dynamics of processes in the landscapes. SPs have or potentially have a critical significance in safeguarding certain processes, e.g. the process of species dispersal, spread of fire and other disturbances, visual perception and preference, agricultural conversion, etc.
In terms of their significance for the processes of our concern, security landscape components have three basic characteristics:
(i) Initiative, the quality of a portion or position whose occupation is likely to give it the advantage of initiating certain processes;
(ii) Efficiency, the quality of a position or portion whose occupation will give it the advantage of less cost in energy and materials and be much more effective in promoting or controlling certain processes;
(iii) Co-ordination, the quality of a position or portion whose occupation will give it the advantage of effective spatial communication among neighboring elements.
SPs are multi-leveled. Each individual process in the landscape has its own security patterns (Figure 1), and these individual SPs may compete and overlap spatially.
Furthermore, each individual process has SPs at various security levels.
Figure 1 A presumed hierarchy of landscape security patterns
The concept of SPs is based on two assumptions concerning spatial patterns and processes: (a) landscape patterns effect processes, and (b) there are strategic landscapes associated with some thresholds in the dynamics of certain processes.
Numerous observations suggest that the spatial patterns of a landscape influence various ecological processes such as species dispersal and population dynamics (Forman and Godron, 1986; Turner, 1989); human processes such as residential development and demographic dynamics (e.g. Berry and Horton, 1970), and visual perceptual processes (Gibson, 1950; Lynch, 1960).
Not all portions and positions of the landscape are equally important in terms of their influence on individual processes, some are more important than others, and some are strategically critical. Examples of such strategic portions and positions include the inlets and outlets of a basin and breaks in a corridor that have critical values for ecological processes (Forman and Godron, 1986; Merriam, 1984); the conspicuous land marks, narrow defiles, gorges and bridges that have significant visual perceptual effects (Stein and Niederland, 1989; Tuan, 1974); as well as certain places that have a strategic significance for economic processes (Taaffe and Gauthier, 1973).
It is important to note , however, that in some cases various processes in the landscape may be controlled by spatial patterns that are not intuitively obvious nor visually apparent to a human observer. It is assumed that some kinds of thresholds exist in the trajectories of the dynamics of processes. At some points (in terms of number, size, shape and inter-distance of landscape elements), a slight change in landscape property produces sudden changes in the response of the process. Such thresholds have been recognized in urban development (Kozlowski, 1986). Similar to thresholds, other concepts have been proposed that may also be useful in understanding my ideas concerning the strategic landscape and security patterns such as safe minimum standards (SMS) (Bishop, Fullerton, et al, 1974; Ciriacy-Wantrup, 1968), carrying capacity, and ultimate environmental thresholds (UETs) (Kozlowski and Hill, 1993), etc.
It is thus reasonable to assume that:
(1) landscape patterns associated with these critical thresholds or constraints are likely to be strategically critical in controlling or promoting certain processes;
(2) landscape design and management following these strategically critical patterns can more effectively safeguard or control the processes.
Therefore, it is worthwhile to identify and apply SPs in landscape planning. The following two aspects of exploration become the major focus of this paper:
(1) How can we define and identify SPs and what are they?
(2) How can we apply SPs in landscape planning to achieve a less detrimental landscape, while at the same time, maximally making changes acceptable to decision makers and/or developers?
These two aspects of inquiry compose an approach to landscape planning which I call the SP approach, or the approach of security patterns. It is an approach to defending various processes of our concern, aiming at a good balance of acceptable changes and a securer landscape through identifying and applying security patterns (SPs). The SP approach tries to establish 'stop signs' in the procedures of decision making for various landscape changes, and to safeguard the security of the processes at critical points. In a certain sense, defining SPs is a strategy of spatial defense, an operational weapon of negotiation aimed at a less harmful change by controlling critical points, or 'frontiers'. Defense by these SPs is expected to be more effective in safeguarding the landscape processes of our concern. GIS has great potential when combined with the SP approach in landscape planning and decision making (see Yu, 1995c for more detailed discussion on the SP concept).
A case study of the Red Stone National Park in south China, is used to illustrate the SP approach. This case is selected since it dramatically represents a defensible procedure of landscape change among defenders of three interacting, and often competing, processes in landscapes, including ecological, visual and agricultural conversion processes .

DEFENDING THE SECURITY OF PROCESSES IN RED STONE NATIONAL PARK: A CASE STUDY
Red Stone National Park is 313 square kilometers in size (Figure 2). The dominant regional natural vegetation is composed of sub-tropical
Figure 2 The landscape of the Red Stone National Park in South China
evergreen forests which have been seriously destroyed at the peripheral area with some isolated remnant patches scattered in the remote areas. The landscape is made up of hundreds of heavily eroded rocky hills, square with flat top and steep slopes. This unique land form is the primary factor affecting the distribution of soil, vegetation, wildlife habitats, visual quality and agriculture. The remnant biological islands are extremely valuable in terms of biodiversity conservation and landscape restoration. The visual quality is extraordinary . It is one of the major tourist attractions in south China. The fertile soil and sub-tropical climate make this land one of the most productive agricultural areas. About twenty thousand farmers live in seventy villages scattered in the small alluvial planes in this hilly landscape. The problems this national park now faces are typical of other national protected areas, namely, the conflicts between development , ecological and visual conservation. Landscape planning in this park is a defensible procedure dramatically taking place among defenders of various processes. As a result, this case study a an illustrative example for the SP approach.
Three processes are concerned in this case study: ecological, visual perceptual, and agricultural. The objectives in this case study are to explore an effective way of defending various landscape processes in this national park by identifying and applying SPs, and to demonstrate how GIS can be integrated into the defensible procedure of landscape change and decision making.
Security patterns in the Red Stone National Park
Ecological SPs: Ecologists' Defensive Frontiers
Ecological processes concerned in this case are species dispersal and maintenance. Three groups of species are targeted: medium-sized mammals (Cervidae and Viverridae families), pheasants (Phasianidae family) and amphibians (Cryptobranchidae and Ranidae families). These species are native to this region and have an endangered status. Ecological SPs are identified by analyzing accessibility surface that represent the potential coverage by the species of our concern.
Accessibility surfaces are developed using a minimum cumulative resistance(MCR) model (Knaapen, Scheffer and Harms, 1992; Yu, 1995b), this model conceives the dynamics of species dispersal as a function of sources, distance and intermediate landscapes. Native habitats of the target species are taken as sources of dispersal. Intermediate landscapes are evaluated for their resistance to the dispersal of species, and the dynamics of the dispersal process is simulated based on the cumulative resistance to the dispersal of a certain species. Comparative resistance values are assigned to various landscape attributes. Various factors such as cover, slope, elevation and aspect may contribute to the resistance value of each cell of the landscape. The probability of successful access to a cell by a species can be expressed as:
i = n
Accessibility = f Min _(Di * Ri)
i = 1
where f is some unknown but monotonically decreasing function. Di and Ri respectively represent the distances (number of cells) and resistance when a species travels across landscape type i. While f is some unknown function, the sum of the weighted distance (Di* Ri) , or the 'cumulative resistance' can be taken as an indication of relative accessibility of the cell to the species through one possible route. There are numerous routes from the sources to the cell, and the routes with the lowest cumulative resistance, namely minimal cumulative resistance (MCR) can be used as the relative measurement of the accessibility of this cell from the sources (habitats of target species).
Resistance classification (Ri ) is based on individual cells of 25 by 25 meters in size. An interactive interface of the GIS model using ARC/INFO is developed to allow the processing of more precise data. Land use and land cover are the major factors contributing to the resistance of the landscape. In our case, it is reasonable to assume the more similar a cell to the natural habitats the less the resistance to the target species of our concern. Eight major land use and land cover categories are observed and they are closely associated with the degree of naturalness or the intensity of human disturbances; these categories range from developed areas to agricultural fields, grass lands, shrubs, coniferous forests, mixed forests, the remnant subtropical forests and water. From developed area to the remnant sub-tropical forests, the degree of human disturbance increases, in this case it is assumed that the resistance to the dispersal of native target species increases accordingly. The developed areas (including roads, housing, tourist service center) have the highest resistance to all target species (assigned a value 10) and the natural remnant forests the lowest resistance to the target species (assigned value 0) . Water bodies are assigned a high resistance value to the medium-sized mammals but have a moderate resistance to pheasants and low resistance to amphibians (here the quality of water is considered).
The topographical factors including elevation and slope also contribute to the resistance to some species. For the medium-sized mammals in this case, gentle slope is considered to have less resistance than a steep slope, the extremely steep slope becoming a barrier to movement. For the pheasants, these topographical factors are not important. Amphibians are sensitive to the hydrological situation, which, in this case is associated with the elevation because of the unique geological formation of this area.
Based on the resistance map, an accessibility surface can be developed using the function discussed above. The resultant accessibility surface resembles a topographic surface that is made up of equal-valued MCR contours. Following Warntz's model of surface interpretation (Warntz, 1966), it 'dips' at the sources, has 'peaks' that are least accessible to target species, has 'courses' with lower MCR value and run from 'pits ' to 'pits', and has 'ridges' with higher MCR values and run from 'peak' to 'peak'. On each of the 'courses' or 'ridges' there is one 'pale' or 'pass'. From the MCR surface, one can also recognize the potential cliffs where values increase or decrease dramatically, and the potential flat planes where the species can spread quickly over the landscape.
The accessibility surface, therefore, reveals the potential patterns of coverage by the target species and the strategic values of landscape in terms of species dispersal and maintenance (Yu, 1995a-b). Based on the features of the accessibility surfaces, four structural components can be identified: buffer zones, inter-source linkages, radiating routes and strategic points. These four components, specified by certain quantitative and qualitative parameters, together with the identified sources (native habitats) compose a security pattern (Figure 3). Changes in these components, quantitatively or qualitatively, will dramatically affect the security of the targeted processes.
Figure 3 A schematic picture showing a typical ecological SP
Among others, three series of ecological SPs are identified respectively at high, moderate and low security levels for different groups of species, e.g. Figure 4-6 for the medium-sized mammals in this case. They could be combined into corresponding overall ecological SPs. These ecological SPs can be used by ecologists as defensive frontiers for the defense of the ecological processes at various security levels in the process of landscape planning and change.
Figure 4 An accessibility surface for the medium-sized mammals and an ecological SP at a less secure level. The SP is composed of sources, strategic points and shortest inter-source linkages

Figure 5 An ecological SP for the medium-sized mammals at a highly secure level. The SP is composed of sources, strategic points, all possible inter-source linkages, big buffer zones and some radiating routes

Figure 6 An ecological SP for the medium-sized mammals at a moderately secure level and the impact of tourist development. The SP is composed of sources, strategic points, some inter-source linkages and some buffer zones.
Visual SPs: Tourists' Defensive Frontiers
The visual SPs are defined on the basis of critical landscape interpreted by visual sensitivity surfaces which are a combination of landscape visibility and preference evaluation. The calculation and mapping of landscape visibility were carried out using function that most GIS packages contain. GIS mapping of landscape preference is relatively more complicated. Firstly, 572 individual from China and USA were interviewed as to their preference evaluation for various landscapes in the case study area (for detailed discussion see Yu, 1995c). Secondly, factor analysis and regression analysis were used to build the preference models. These preference models show the contributions of various spatial information to the visual quality of the landscape. This spatial information consists of landscape elements including water, rocks, vegetation, tourist service buildings, fields, weather conditions, and spatial dimensions associated with the position of viewers including foreground, mid ground and background. Both types of spatial information are classified, mapped and analyzed using GIS. Finally, GIS was used to develop the landscape preference map based on the landscape preference model and the spatial information.
The visual security patterns (SPs) are defined in association with various security levels. Using the histograms of visibility and preference distribution patterns, some thresholds can be identified and used for the identification of visual security levels (Yu, 1995b). Three levels of SPs are identified: low, medium and high. These visual SPs could be used as defensive frontiers by the defenders of visual perceptual processes and tourism during the procedure of spatial bartering and bargaining.
Agricultural SPs: Farmers' Defensive Frontiers
Local farmers have depended on their land for hundreds of years. Population growth requires more land to be converted into agricultural fields. In a certain sense, agricultural conversion in this case study is an issue of survival of the local people in this area. The potential security levels farmers want to achieve are normally determined by socio-economic analysis on local, regional and even national scales. This case study addresses the issue of agricultural SPs based on investigation of the landscape at the local scale and focuses on the issue as to where land conversion should pause or accelerate in terms of efficiency of productivity and impact on other processes in the landscapes.
The procedure of identifying agricultural SPs and ecological SPs is similar. Agricultural conversion is considered a process of disturbance with the seventy villages in the study area taken as the source of the spread of the disturbance. The intermediate landscapes are evaluated according to their resistance or cost of agricultural conversion. A convertibility surface (or the potential of conversion) is developed based on the cumulative resistance of intermediate landscapes. Based on this convertibility surface, agricultural SPs are identified at some thresholds or strategic values. Various agricultural SPs were identified corresponding to different security levels: low, medium and high. They could be used as defensive frontiers by defenders of the process of agricultural conversion in the spatial bargaining and bartering of landscape change and decision making.
Alternative Changes Based On SPs
Various alternative change models can be developed based on SPs. To illustrate this case four change alternatives are discussed as examples:
Differentiation Of Management Concentrations Based On SPs
At the highest level of the management hierarchy is the differentiation between landscape conservation, development (mainly agricultural conversion) and an integration of both (Figure 7). The general management differentiation can be further sub-categorized, when the environmental concern is specified into ecological and visual aspects. A weighing system has to be used when combining SPs of various ecological processes into an overall ecological SP. In this case, it is assumed all of the individual component SPs have the same weight (= 1).
Figure 7 Management differentiation based on two general SPs of environmental conservation and agricultural conversion
Strengthening Landscape Infrastructures Based On SPs
Taking ecological processes as examples, landscape ecological infrastructures can be strengthened by consolidating the SP components . These consolidations include, but are not limited to, the following aspects of improvement in quantity and quality:
(1) Increasing the buffer zones and making the intermediate landscapes more hospitable to native species, (2) having alternative linkages, widening the linkages and improving the connectivity of the linkages, (3) widening the radiating routes with native plants, (4) introducing native patches at the strategic points and expanding their dimensions.
By these quantitative and qualitative improvements of the components of the ecological SPs, the security of the landscape for the ecological processes can be improved proportionately. This gradual procedure, however, after reaching a certain threshold, will dramatically increase the security of the landscape and another security level for ecological processes will be achieved, e.g. from Figure 4 to Figure 5.
Modifying Introduced Change Models Or Trade-Off Sps And Exercising Spatial Bartering Based On SPs
Figure 6 shows a proposed tourist development plan and its potential impact on the ecological SP for the dispersal of medium-sized mammals. As a result, the remaining native habitat at the upper-right corner and the immediate buffer zone will be destroyed by the expanding construction of tourist facilities and three ecological corridors will be negatively affected by the tour line.
One solution to reduce the negative impact of tourist development is to modify the plan of tourist development based on ecological SPs. This solution suggests moving the tourist center to the edge of the park and imposing a special management policy on the tour sections across the ecological corridors.
It is, however, possible that any modifications of the tourist development plan may not be acceptable to the developers or local officials. In this case, the defender of ecological processes should consider the solution of spatial bartering to trade-off components of SPs for their overall consolidation. This solution of spatial bartering may include, but is not limited to, the following tactics (Figure 8).
Figure 8 Possible spatial bartering tactics based on ecological SPs
Figure 9 shows an example of how the tourist development plan can be adapted by trading-off some components of ecological SPs, but not security levels. It is suggested that the ecologist may abandon the native habitat at the upper-right corner of the map, but restore a native patch at the middle left and add a corridor to connect the two existing patches that will potentially be isolated because of the interruption of the corridors by the introduced tour line.
Figure 9 Tourist-development plan and spatial bartering on an ecological SP for the medium-sized mammals as a moderately secure level
It should be noted that all the gains in Figure 8 are based on the identified ecological SPs and the accessibility surfaces that represent the accessibility of the landscape.
SP Approach Integrated with GIS in Support of Decision Making for Securer Landscape Changes
Those identified SPs at various security levels are the basis of strategies of spatial defense, or spatial bartering among defenders of ecological, visual and agricultural conversion processes, represented by ecologists, tourists and farmers .
Two situations have been simulated when using SPs in support of decision making:
situation one: negotiations for landscape changes within currently defined security levels. This is an optimistic result of the SP approach. Any of the four change models proposed in the last section may be acceptable.
situation two: negotiations for landscape changes beyond current security levels.
Situation one can be taken as a special case of situation two where further solutions will be explored at other security levels. In situation two, one or more defenders has to give up his current requirement for the security of the concerned processes, and retreat to a lower level of security for further defense. Figure 10 shows the negotiation or gaming procedure and the strategies that reflect the decision making process among the defenders of the three processes, each of the defenders has SPs of three security levels in mind. This procedure of redefining security levels may repeat until solutions can be found.

Figure 10 Strategies for negotiation and GIS-support decision making based on SPs
As is shown in this case study, planners do not provide the optimum solutions or even any solutions at all to solve some projected problems. Instead, they are neutral consultants and moderators providing alternative strategies for each of the defenders in their defending various process in the landscape. These strategies are comparatively more efficient in achieving corresponding utility goals. The final solutions are the acceptable results achieved through the negotiation among all defenders of the processes. GIS has great potential both in the defining of those spatial strategies based on the identification of SPs, and in the defensive procedure of negotiation among defenders of various processes.

CONCLUSION
The general conclusion of this research is that landscape security patterns (SPs) could be very useful in landscape planning aimed at safeguarding various processes. Security patterns can be used as impact models guiding the modification or improvement of proposed change plans, as constraint criteria controlling the maximization course of individual processes, as blue prints for the improvement of landscape structure, and as a basic reference frame for the procedure of spatial bartering.
The SP approach makes policy and management practice specify and concentrate on certain areas which can increase the efficiency of the decision making procedure. Various landscape change alternatives are explored based on SPs within a certain security level for a certain process. Further change alternatives are developed at a lower security level only when none of the alternatives at a higher security level is acceptable to decision makers or defenders of various processes. SPs are 'stop signs' in the decision-making course that reduce the risk of the irreversibility of decision making and reduce the possibilities of catastrophes in landscape changes.
GIS play an important role in simulating various processes, identifying security patterns, evaluating impact and developing landscape change models based on SPs. GIS shows its great potentials when combined with the SP approach in supporting landscape decision making.

ACKNOWLEDGE: Thanks are due to Carl Steinitz, Stephen Ervin and Richard T. T. Forman at Harvard University for their advice and support of this research, Hugh Keegan and other staffs at the ESRI (Environment Systems Research Institute) for their support in the application of ARC/INFO GIS, and Erin Crowley for her editing of the manuscript.

REFERENCES
Berry, B. J. L., and Horton, F. E. (Ed.), 1970, Geographic Perspectives on Urban Systems (1st ed.). Englewood Cliffs, New Jersey: Prentice-Hall, Inc.
Bishop, A. B., Fullerton, H. H., Crawford, A. B., Chambers, M. D., and McKee, M., 1974, Carrying Capacity In Regional Environmental Management Washington, D.C.: Office of Research and Development, U.S Environmental Protection Agency.
Ciriacy-Wantrup, S. V., 1968, Resource Conservation: Economics and Policies (3rd ed.) Berkeley, Los Angeles: University of California Division of Agricultural Science.
Forman, R. T. T., and Godron, M., 1986, Landscape Ecology, New York: John Wiley.
Gibson, J. J., 1950, The Perception of the Visual World, Boston: Houghton Mifflin.
Knaapen, J. P., Scheffer, M., and Harms, B.,1992. Estimating habitat isolation in landscape planning, Landscape and Urban Plann., 23, 1-16.
Kozlowski, J., 1986, Threshold Approach in Urban, Regional and Environmental Planning: Theory and Practice, St. Lucia, Queensland, Australia: University of Queensland Press.
Kozlowski, J., and Hill, G., 1993, Towards Planning for Sustainable Development: A Guide for the Ultimate Environmental Threshold (UET) Method, Vermont, USA: Avebury,, Ashgate Publishing Company.
Lynch, K., 1960, Image of the City, MIT Press. Cognition and Environment: Functioning in an Uncertain World. New York: Praeger.
Merriam, G., 1984, Connectivity: a fundamental characteristic of landscape pattern, In J. Brandt and P. Agger (Ed.), The First International Seminar on Methodology in Landscape Ecological Research and Planning, 1, pp. 5-15, Roskilde, Denmark: Roskilde Universitetsfolag GeoRuc.
Stein, H. F., and Niederland, W. G. (Ed.), 1989, Maps from the Mind: Readings in Psychology. Norman and London: University of Oklahoma Press.
Taaffe, E. J., and Gauthier, H. L., 1973, Geography of Transportation, Englewood Cliffs, New Jersey: Prentice Hall.
Tuan, Y.-F., 1974, Topophilia, Englewood Cliffs, New Jersey: Prentice-Hall, Inc.
Turner, M. G.,1989. Landscape ecology: the effect of pattern on processes., Annual Review of Ecology and Systematics, 20, 171-197.
Warntz, W., 1966, The topology of a social-economic terrain and spatial flows, In M. D. Thomas (Ed.), Papers of The Regional Science Association , pp. 47-61, Philadelphia: University of Washington.
Yu, K., 1995a. Cultural variations in landscape preference: comparisons among Chinese sub-groups and Western design experts, Landscape and Urban Planning, 32, 107-126.
Yu, K., 1995b, Ecological security patterns of landscapes: concept, method and a case, In Fung, T. and Lin, H (Eds.) The proceedings for the International Symposium of Geoinformatics'95, Hong Kong:, The Chinese University of Hong Kong.
Yu, K., 1995c, 'Security Patterns in Landscape Planning: With a Case In South China', doctoral thesis, Harvard University.