=Paper=
{{Paper
|id=None
|storemode=property
|title=Determining Ecotones by Decision Support Systems
|pdfUrl=https://ceur-ws.org/Vol-706/papersg03.pdf
|volume=Vol-706
|dblpUrl=https://dblp.org/rec/conf/dateso/PechanecBC11
}}
==Determining Ecotones by Decision Support Systems==
https://ceur-ws.org/Vol-706/papersg03.pdf
Determining Ecotones by Decision Support
Determining ecotones by decision support systems
Systems
Pechanec Vilém, Brus Jan and Caha Jan
Vilém Pechanec, Jan Brus, and Jan Caha
Faculty of Science, Palacky University in Olomouc, Czech Republic
Department ofvilem.pechanec@upol.cz,
Geoinformatics, Faculty of Science, Palacký
jan.brus@upol.cz, University in Olomouc
jan.caha@klikni.cz
Tř. Svobody 26, 771 46 Olomouc, Czech Republic
vilem.pechanec@upol.cz, jan.brus@upol.cz, jan.caha@klikni.cz
Abstract The investigation into ecotones enabled a better understanding of the
causal relationships between certain landscape elements, landscape utilisation
categories and ecotones. By studying ecotones, we wanted to expand
understanding of patterns having an influence on the landscape condition,
structure, functions, landscape elements and their relationships. The fuzzy
approach was applied for better understanding and modelling of ecotones. This
methodology was afterwards built in expert system. The landscape assessment
process based upon an expert system allows for a multidisciplinary view of a
landscape. The landscape is evaluated from several perspectives: the ecological
stability, soil erosion, retention capacity and economic landscape calculations
by designed expert system. These results from expert system were taken into
consideration during mapping ecotones. Ecotones may serve as one of the
distinctive indicators of the impact humans have on the landscape. It is
therefore necessary to develop more sophisticated methods of their studying.
1 Introduction
Ecotones are significant regions of landscape heterogeneity that contain elements,
patterns, and processes existing and operating at varying spatial scales [19], [14],
[24]. Ecotones play a pivotal role in landscape. This can be seen from several
viewpoints — environmental, biological, economic, historic, aesthetic, etc. Their
most crucial task is ecological. They represent specific ecosystems in a landscape,
corridors for the migration of animals or distribution of plants. They also contribute to
soil protection against erosion, etc. During their existence ecotones go through
developmental stages which depend primarily on the dynamics of factors in the
surrounding environments. Some ecotones are temporally stable some may migrate or
mutate. Spatial relationships within the surrounding environments are of great
importance since ecotones are the zones where communities, populations of plants
and animals meet, where they compete, create some tension and blend. In nature
ecotones often function as corridors or barriers enabling or restricting the flow of
mass or energy. The temporal and spatial stability of ecotones contributes to the
ecological value of the ecotone community and to the greater value of a landscape.
V. Snášel, J. Pokorný, K. Richta (Eds.): Dateso 2011, pp. 206–215, ISBN 978-80-248-2391-1.
2 Pechanec Vilém, Brus Determining
Jan and CahaEcotones
Jan by Decision Support Systems 207
We can classify ecotones by various aspects, e.g. by structure. It is difficult to
perfectly simulate the real world, due to its complexity. It has been argued that
uncertainty information is a vital component in the use of spatial data for decision
support [11], [1]. We are forced to model uncertainties which suggest fuzzy logic to
problem solving. Many techniques have been developed for communicating
uncertainty in data and models for specific visualization applications, such as remote
sensing, land allocation, [3], [16], [1]. Development in the field of Geographical
Information Systems (GIS) facilitates the integrating of flexible decision support
functionalities to perform multicriteria evaluations to solve allocation or location
problems, to perform suitability analysis, to integrate different criteria for options
choice and group decisions [13], [18]. For such a specific and complex field as
ecotones there is compulsory to use unique technique. Fuzzy approach can be applied
to help to better mapping and modelling, due to non-sharp nature of ecotones, also
their uncertainties. Fuzzy theory is practically implemented through decision-making
processes in expert system. Build in expert knowledge helps to coping with vague
concepts such as imprecise and uncertain values of attributes of spatial entities of
ecotones. This research was mainly focused on testing possibilities of using Spatial
Decision Support Systems (SDSS) for mapping ecotones.
2 Theoretical background
2.1 Fuzzy theory
Sometimes crisp information isn’t the best approximation of reality. This is the main
reason for use of fuzzy set theory and fuzzy logic in GIS and decision making
process. This point of view is shared by many authors [6], [8], [12], [22] and [23].
The best overview of use of fuzzy sets in modelling geographical elements and
phenomena are presented in [6] and [8]. Concepts of fuzzy set theory and fuzzy logic
were firstly presented by L. A. Zadeh in 1968 [25] in his article "Fuzzy sets".
Compared to classical set theory where elements are evaluated like belonging or not
belonging to the set, fuzzy set theory uses membership function to assess each
element's membership value. Membership value is the number from interval [0, 1]
where 0 means that element is not included in set. Value 1 means that element is fully
included in the set, any other value from given interval means that element is a fuzzy
member of the given set. The higher of the membership value predicate how much the
element belongs to this set Fig.1. Within fuzzy set theory, vague and imprecise
numeric values can be represented by fuzzy subsets on the basic numeric domains.
There are several ways how can we define membership values for elements, for finite
sets we can clearly define membership value for each element and for infinite sets
(i.e. real numbers) we can define a function, named membership function, which
defines membership value for each element of the set.
208 Vilém Pechanec, Jan Brus, Jan Caha ecotones by decision support systems
Determining 3
Fig. 1 Boolean versus Fuzzy sets [11]
Unlike from the crisp set theory where element either belongs to set A or B with
fuzzy sets it is possible for an element to belong to both sets A, B to each with
different membership value. Fuzzy logic is derived from those assumptions. Fuzzy
logic is a multi-valued logic that deals with reasoning where classic binary sets are
not appropriate.
3 Technical background
3.1 SDSS
In recent years, the GIS is increasingly understood as a means used to support
decision-making and recognized as the basis for the Spatial Decision Support Systems
(SDSS). SDSS are a specific kind of information system. There is no unambiguous
and generally accepted definition because forms of technology have not been profiled
yet. However, the majority of authors agree that it is a spatial expansion of Decision
Support Systems (DSS), or rather the integration of GIS and DSS. SDDS are
computer information systems that provide support for problems difficult to formulate
and structure. They are usually considered when it is impossible to use a fully
automated system. SDSS are closely related to knowledge-based and expert systems
whose creation was possible due to artificial intelligence. SDSS also provide detailed
displays resulting in reduced decision time and enabling a better grasp of spatial
problems due to better visualization of the problem to be solved [16].
In relation to the previous paragraph, we can set apart special category of SDSS
which are expert systems. They can stand alone, but in combination with GIS they
integrate into a very powerful tool. We must emphasize here that the possibility of
easy handling without deeper knowledge of processes and algorithms may lead to a
completely incorrect interpretation of results, and thus, erroneous decisions.
3.2 Expert systems
Expert systems are computer programs able to simulate actions of an expert in a
4 Pechanec Vilém, Brus Determining
Jan and CahaEcotones
Jan by Decision Support Systems 209
particular field when solving complicated tasks . Other authors [20, 5] describe expert
system as software or combination of software and hardware, which can complete the
exercise of specific complex tasks. These tasks can also be solved by a human expert,
but require significant expertise in the solution. Expert systems provide a powerful
tool for solving many problems that often cannot be solved by other, more traditional
methods. The usage of expert system has proved to be crucial in the process of
decision support and problem solving [8]; therefore, their usage has spread into many
sectors. They are considered a sub-category of knowledge-bases systems. They are
based on symbolic representation of knowledge and its implementation in an
inference mechanism. Experts in the given field present the source of knowledge and
procedures. These systems are able to justify solution procedures. They are used
primarily for tasks difficult to structure and algorithmize, e.g. problems with
recognition of situations, diagnosis of status, construction, planning, monitoring of
status, corrections, management and decision-making. However, experience and
intuition have to be part of the solution. Certain authors [10] look at the expert analyst
required to operate the system as posing a barrier to decision makers who must
translate the problem into a form that can be understood by experts who, in turn must
translate their understanding of the problem into a form that can be evaluated and
solved [21].
3.3 Ecosystem Management Decision Support (EMDS)
EMDS - is a product which if interconnected with ArcGIS provides a comprehensive
SDSS product. The supplement comes from the Pacific Northwest Research Station
U.S. Forests Service. According to its authors it integrates logical formalism justified
on the basis of the knowledge base in the GIS environment. It provides support for
decisions on evaluation and assessment of landscape from an ecological point of
view. The EMDS decision-making pattern is based on a knowledge base that uses
fuzzy logic, network architecture and object-based approach. The basic architecture of
objects of EMDS knowledge bases enables an increase in development of dependent
complicated knowledge data. Up-to-date modern research methods dispose of
mathematical models characterised by very specific mathematical dependencies
between the status of monitored objects and the processes influencing them. Fuzzy
logic tools significantly enhance the ability to work with incomplete or vague
information. The proposed network architecture of EMDS knowledge bases allows
evaluation of the influence of the missing information and has the ability to come to
conclusions with incomplete information. The current version of EMDS is the 4.1
version that ensures compatibility with ArcGIS 9.x and includes a newly designed
hotlink browser tool, which speeds up work and provides graphical representation of
the knowledge base for landscape features chosen from topics intended for analysis.
Internet presentation of maps is secured via ArcIMS.
210 Vilém Pechanec, Jan Brus, Jan Caha ecotones by decision support systems
Determining 5
4 Methodology
Many researches were done in the field of determining ecotones [9, 15].
Comprehensive study based on representing transitional zones and determining their
borders was also carried out [2]. Based on upper text, fuzzy theory brings apparatuses
how to deal with specific fuzzy sets. In the case, that we accepted the representation
of ecotones by fuzzy sets, whole ecotone same as its transitional borders can be
modelled.
The research on ecotones was carried out in the catchment area of the stream
Trkmanka in South Moravia. The boundaries of the region are formed by the
boundaries of the Trkmanka catchment area as taken from the hydrological map. The
Trkmanka catchment area comprises the regions of Břeclav, Hodonín, and Vyškov in
the south-east Moravia. This narrow territory drops from the north-east to the south-
west. Its area covers approximately 380 km2. The Trkmanka catchment area lies in
the Carpathian part of the Czech Republic. It consists of the flysh belt of the outer
part of the Western Carpathians and the Vienna Basin. Its prevailing parts are formed
by sedimentary fill. Lowest parts have a flat alluvial relief and belong to the vale
Dolnomoravský úval. Three modelling areas were Kobylí, Ždánice and Rakvice.
The landscape assessment process based upon an expert system allows a
multidisciplinary view of a landscape. The landscape was assessed from four
viewpoints according to the selected indicators. Monitored ecological indicators
mainly include an ecological stability coefficient (1st indicator) showing an
ecosystem’s ability to compensate for changes caused by external factors in order to
keep its natural properties and functions. This closely relates to the erosion hazard (2 nd
indicator) and landscape storage capacity (3rd indicator). A modern tool for nature
conservation is the economic assessment of ecosystems (4 th indicator) and their non-
productive functions. It enables us to compare ecological values and economic profits
in the same terms and hence provides for better reasoning in decision-making
processes. In the assessment the ecological value of nature is always taken into
consideration.
Whole process of evaluating landscape stands as primary phase in determining of
ecotones. Results computed by expert system are primary data which were used as
initial data in the ecotones determined by botanist. Such an assessment can be
performed with common accessible data, including the land use, biotype mapping of
the Czech Republic which was processed by methodology introduced by NATURA
2000, pedoecological unit (soil-ecological unit in Czech terminology, used for land
appraisal), forest topology and contour lines. A layer of “soil ecological units for soil
rating” was nationally special. Soil ecological unit for soil rating of the agricultural
parcels is a five-digit numerical code of the main soil and climatic conditions that
affect the productive capacity of agricultural land and its economic valuation. Forest
typology is the classification system consisting of differentiation in the management
of the forest lands. This is a nationwide database of permanent environmental
conditions. This database standardizes the potential natural vegetation in relatively
homogeneous territorial units in forests. It is necessary for economic planning of
6 Pechanec Vilém, Brus Determining
Jan and CahaEcotones
Jan by Decision Support Systems 211
forest management. Described layers are finally set up into specified structure and
evaluated in system Assessment in Ecosystem Management Decision Support
(EMDS). In many real cases, the available data are crisp, precise, but they hold
uncertain on their reliability for several reasons: either because the agencies that are
the source of the data cannot be entirely trusted, or because one knows that the means
of acquisition are not enough sophisticated and generate systematic errors; not least
because data are a result of a subjective analysis, such as surveyed data. Uncertainty
on precise or imprecise data can be represented by associating a degree of confidence
or credibility, or reliability with them [4].
5 Results
Three methodologies based were used to determine the indicator of landscape
stability coefficient. The data was valid for the year 2007 and the analysis did not
cover the whole study area, but only parts. We obtained quite varying values for
particular land categories from the category of above-average land use with distinct
disturbances in the natural structures to the category of natural landscape or landscape
close to it. Another indicator of landscape stability is long-term soil loss. It was
determined using a Universal Soil Loss Equation (USLE) to assess the danger of
water erosion to agricultural land (Fig. 4). It gives the potential amount of soil which
could be removed due to water sheet erosion. It however does not include its deposit
at the site or into the lower layers underneath. This parameter is directly linked to
direct runoff. The Runoff Curve Number Method (CN) was applied to determine the
value (Fig. 3). All model areas have the greatest representation of values for the zero
runoff. The greatest runoff values were those for urban areas and roads. In addition,
the modified Hessen biotope assessment method was applied to that part of the study
area if the required data was available. The price of one segment means the average
costs of increasing the land value by one ecological degree for one square meter of
land. In the view of mathematics, ecotones are areas where the particular points can,
with a certain probability, be considered as belonging to one ecological stratum or,
with another probability, to another one.
Assessment of landscape by expert system based on EMDS system is composed of
several parts. Mainly from algorithmized decision schema for appraisal landscape
segments (network in NetWeaver for EMDS), simultaneously is formed and filled
knowledge base about landscape. NetWeaver knowledge bases use an object-based
approach, which makes them very modular; therefore, they are easily created and
maintained. Moreover, the system enables interactive tuning in an arbitrary stage of
the creation of the knowledge base. This significantly speeds up the development
process. Fuzzy logic provides calculation methods which do not require directive
expression. NetWeaver is completely object-oriented system. This means that
networks and data links are programming objects that represent or substitute objects
or notions of the real world (Reynolds, 1998). Implementation of this knowledge
embodied in interconnection between an appraisal network and input entry about
212 Vilém Pechanec, Jan Brus, Jan Caha ecotones by decision support systems
Determining 7
landscape with datalinks in network. Optimized data model with series of
reclassification tables has to be done for connection with entry data. Assessment of
landscape segments by an appraisal network is achieved finally in EMDS.
Fig. 3. An appraisal grid for the Runoff Curve Number Method (CN method)
Fig. 4. The long-term average soil loss of agricultural land and forestland in a model area of
Ždánice
The highest ecological stability is calculated in modeled area Ždánice. The lowest
endangered area by water erosion calculated by USLE (Universal Soil Loss Equation)
is in modeling area of Kobylí. At suggested precipitation 12 mm is a level of straight
runoff equated to 0. Appraisal of landscape by the modified Hessen method is also
significant due to the evaluation of landscape. All this information is taken into
consideration for the final evaluation of possible occurrence of ecotones. Predicted
areas were faced with data of field survey and aerial photography.
8 Pechanec Vilém, Brus Determining
Jan and CahaEcotones
Jan by Decision Support Systems 213
6 Conclusion
The main idea of this work was practically to test possibilities of SDSS capability in
GIS software for mapping remarkable landscape structure. Role of ecotones in the
landscape is important and ecotones are formed in a special part of the landscape.
They can be distinguished mainly by field work which was proofed by botanists,
because of their blur border and many ecological factors which affect them. It was
extremely difficult to model all conditions, which come into the process of
developing ecotones. Only datasets which build up the expert system were not
efficient to map ecotones in the study area. In the other hand, in our model was not
decided to calculate with this amount of factors. It was mainly considered as a pilot
and testing application for help with mapping ecotones. As the one of the crucial
problems were considered uncertainties in datasets and also mapping measure of GIS
analysis. This measure was derived from the measure of less detailed map layer.
Uncertainty of entering data is mainly based on the form how the agencies collect
them. This data have to be analysed and tested for accuracy in the field, due to their
error. The second problem was based on problematic expression of ecotones in fuzzy
theory in GIS. Basic ideas of use of fuzzy set theory and fuzzy logic in geography and
geoinformatics can be found in many sources including [6], [8], [12] and [17]. Almost
no of those sources provide some kind of workflow how to deal with imprecise data
in GIS. As suggested in [6] that this fact may be caused by lack of tools for work with
fuzzy sets and fuzzy logic in the widespread GIS. The problem probably lays deeply
in the structure of current GIS. As most of today’s software GIS highly rely on
mathematics and informatics which structure is extremely connected to both Boolean
logic and crisp set theory. Cause of this is probably fact that alternative theory to crisp
set theory and Boolean logic exists only since 1960’s but both other are much older
and thus more rooted in mathematics and informatics. EMDS was chosen as final tool
for developing a appraisal grid. EMDS including NetVeawer was considered as a
very powerful tool with large field of application. Ability of implementation of fuzzy
sets into expert systems is in this case inestimable. Final results of this research stand
only like partially data, which can be used as base point for future analysis of
ecotones in the landscape. Future visions stand on developing more sophisticated
expert system, which will better utilize fuzzy theory and will implemented more
variables into the appraisal grid for assessment the landscape. This approach of
mapping by expert system will be subsequently applied in a recent project such as
monitoring sub-urban areas.
214 Vilém Pechanec, Jan Brus, Jan Caha ecotones by decision support systems
Determining 9
Acknowledgments
This work was supported by the Czech Science Foundation grant GA205/09/1159.
The intelligent system for interactive support of thematic map design.
References
1. Aerts, J. C. J. H., K. C. Clarke, et al. (2003) Research Papers - Testing Popular
Visualization Techniques for Representing Model Uncertainty. Cartography and
geographic information science. 30: 249 (14 pages). 5
2. Arnot, C., Fischer, P. (2007) Mapping the Ecotone with Fuzzy Sets. In Morris, A.,
Kokhan, S. (eds.) (2007): Geographic Uncertainty in Environmental Security, p.
19-32.
3. Aspinall, R.J., Pearson D.M. (1995) Describing and managing uncertainty of
categorical maps in GIS. In: Fisher P (ed) Innovations in GIS 2. Taylor & Francis,
London, pp 71–83
4. Bordogna, G. et al. (2007) A flexible decision support approach to model ill-
defined knowledge in GIS. In Morris, A., Kokhan, S. (eds.) (2007): Geographic
Uncertainty in Environmental Security, p. 19-32.
5. Boss, R. W. (1991) “What Is an Expert System?” ERIC Digest. ERIC Clearing
House on Information Resources, Syracuse, NY, 1-3.
6. Burrough, P., McDonnell, A., Rachael A. (1998) Principles of Geographical
Information Systems. Second Edition: Oxford University Press, 1998. 356 p.
ISBN 0198233655.
7. Crossland, M. D., Wynne, B. E. and Perkins, W. C. (1995) Spatial Decision
Support Systems: An overview of technology and a test of efficacy, Decision
Support Systems, 14, 3, 219-235.
8. Dragićevic, S. (2005) Multi-Dimensional Interpolations with Fuzzy Sets. In:
Petry, F.E.; Robinson, V.B.; Cobb, M.A. (eds.) Fuzzy Modeling with Spatial
Information for Geographic Problems. Berlin: Springer, ISBN 3-540-23713-5.
9. Fortin M. J., Olson R. J., Ferson S., Iverson L., Hunsaker C., Edwards G., Levice
D., Butera K. and Klemas V. (2000) Issues related to the detection of boundaries.
Landscape Ecology 15: 453–466.
10. Goodchild, M. F., (1992) Geographical information science. International Journal
of Geographical Information Systems, 6(1), 31–47.
11. Hunter, G.J., and Goodchild, M.F. (1995) A new model for handling vector data
uncertainty in geographic information systems. Proceedings, URISA, San
Antonio, Texas, pp. 410-419.
12. Hwang, S., Thil, J.C. (2005) Modelling Localities with Fuzzy Sets and GIS In:
Petry, F.E.; Robinson, V.B.; Cobb, M.A. (eds.) Fuzzy Modeling with Spatial
Information for Geographic Problems. Berlin: Springer, ISBN 3-540-23713-5.
13. Jankowski, P., Nyerges, T. (2001) Geographic Information Systems for Group
Decision Making, Taylor & Francis: London. March, 2001. equal contribution;
eight chapters synthesizing research 1995-2000.
14. Ji, M. (2002) Fuzzy modelling of African ecoregions and ecotones using AVHRR
NDVI temporal imagery. Geocarto International 17(1): 21-30.
10 Pechanec Vilém, BrusDetermining Ecotones
Jan and Caha Jan by Decision Support Systems 215
15. Kent, M., Gill, W. J., Weaver, R. E., and Armitage, R. P. (1997) Landscape and
plant community boundaries in biogeography, Progress in Physical Geography 21:
315-353.
16. Leitner M, Buttenfield BP (2000) Guidelines for display of attribute certainty.
Cartography and Geographic Information Sciences 27:3–14.
17. Li, Z., Zhou, Q., Kainz, W. (eds.) (2004) Advances in Spatial Analysis and
Decision Making: Proceedings of the ISPRS Workshop on Spatial Analysis and
Decision Making. Lisse: Swets & Zeitlinger, ISBN 90-5809-652-1.
18. Malczewski, J. (1999) GIS and Multicriteria Decision Analysis, John Wiley and
Sons, 392 pp., New York, NY.
19. Nellis, M.D., Briggs, J.M. (1989) The effect of spatial scale on Konza landscape
classification using textural analysis. Landscape Ecology 2(2): 93-100.
20. O'Looney, J. (2000) Beyond maps: GIS and decision making in local government.
Redlands, Calif: ESRI Press.
21. Popper, M., Keleman, J. (1998) Expertné systémy. Bratislava. Alfa.
22. Verstraete, J., Hallez, A. and De Tre, G. (2007) Fuzzy regions: theory and
applications. Geographic Uncertainty in Environmental Security, pp. 1-17.
Springer, Dordrecht
23. Verstraete, J., Hallez, A. and De Tre, G. (2007) Fuzzy regions: theory and
applications In Morris, A., Kokhan, S. (eds.) (2007) Geographic Uncertainty in
Environmental Security, p. 19-32.
24. White, J.D., Running, S.W., Ryan, K.C. and Key. C.C. (2002) Fuzzy logic merger
of spectral and ecological information for improved montane forest mapping.
Geocarto International 17(1): 59-66.
25. Zadeh, L.A. (1968) ‘Probability Measures of Fuzzy Events’, Journal of
Mathematical Analysis and Applications, 23, 421-427.