Editorial Manager(tm) for Journal of Water and Climate Manuscript

Editorial Manager(tm) for Journal of Water and Climate
Manuscript Draft
Manuscript Number:
Title: Participatory Scenario Development Analysis for the Future of Water in Seyhan Basin, Turkey
Article Type: Special Issue Article
Corresponding Author: Dr. Erol Hasan Cakmak, Ph.D.
Corresponding Author's Institution: Middle East Technical University
First Author: Erol Hasan Cakmak, Ph.D.
Order of Authors: Erol Hasan Cakmak, Ph.D.; Hasan Dudu, M.Sc.; Ozan Eruygur, Ph.D.; Metin Ger, Ph.D.;
Sema Onurlu, M.Sc.; Ozlem Tonguc, M.Sc.
Manuscript
Click here to download Manuscript: Cakmak et al._JWCC_scenes_last.doc
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Participatory Scenario Development Analysis for the Future of Water
in Seyhan Basin, Turkey1
Erol H. Cakmak2,3, Hasan Dudu2 , Ozan Eruygur4, Metin Ger5, Sema Onurlu6, Özlem Tonguç2,
Abstract
Stress on water resources of Turkey is expected to increase in near future. This paper
presents the results of a case study in one of the most important basins of Turkey, Seyhan
Basin, developing scenarios in a participatory process with stakeholders in the region. We
employed modified fuzzy cognitive mapping with a new learning algorithm and back-casting
method with STEEP framework. The results obtained from both methods are consistent.
Participants envisioned that water supply, water demand and water use will decline in the
future in response to the increasing impacts of climate change. Improvements in sustainable
water management, irrigation efficiency and water saving technologies will diminish the
severity of scarcity that is expected to occur due to climate change.
Keywords: Dynamic analysis, Fuzzy Cognitive Maps, Seyhan Basin, sustainable water
management, water.
Short Title: Participatory Scenario Development Analysis for the Future of Water in Seyhan
1
The authors gratefully acknowledge financial support for the project Water Scenarios for Europe and
Neighbouring States (SCENES) from the European Commission (FP6 contract 036822).
2
Department of Economics, Middle East Technical University, 06531, Ankara, Turkey.
3
Corresponding author. Tel: +90 312 210 3088, Fax: +90 312 210 7964, E-mail: [email protected]
4
Department of Economics, Gazi University, Ankara, Turkey.
5
Department of Civil Engineering, Istanbul Kultur University, Istanbul, Turkey.
6
Sintek Muhendislik Ltd., Ankara, Turkey.
1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
INTRODUCTION
Water has been designated as a scarce economic resource by international community for at
least two decades. Increase in the irrigated land area in the 20th century has been
considered as the major reason behind the scarcity. The change in the volatility of rainfall
which is attributed to the climate change has also put a significant pressure on water
resources. This has lead to significant shifts in researchers’ focus; analyzing the issues about
irrigation water management and developing better policies and practices in a global scale
have become priorities (Dudu & Sinqobile, 2008).
This paper reports the findings of a qualitative scenario development process implemented
in the Seyhan Basin, Turkey, as a part of SCENES (Water Scenarios for Europe and for
Neighbouring States) project. SCENES uses an integrated approach by combining and
balancing several dimensions of issues related to water to address complex questions about
the future of water resources in Europe, Mediterranean, Caucasus and Ural Mountains.
SCENES adopts an iterative process for scenario enrichment on three levels: the panEuropean scale, the regional scale and basin scale. The enrichment works in both directions
from pan-European to basin and from basin to pan-European iteratively (Kämäri et al.,
2008). This study is a part of the iterative process at the basin level. Seyhan is selected as
one of the pilot areas to develop storylines and to feedback the upper level scenarios
interactively.
Stakeholder workshops are organized in the Seyhan Basin to identify the issues, drivers, and
their interactions, and to envision the future of water in the basin area. Fuzzy cognitive maps
2
(FCM) are formed for the present and the future. Estimated future states of the variables are
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
obtained through dynamic analysis of the FCMs. The stakeholders are also asked to develop
scenarios using STEEP (social, technological, economic, environmental and policy) approach
for self-defined objectives to be attained in 2030.
Next section briefly introduces the pilot area. Then a literature survey on implemented
methodologies follows. Third section describes the process and the workshop settings.
Fourth section presents the findings of the dynamic analysis. The last section is reserved for
conclusions.
OVERVIEW OF WATER IN TURKEY AND SEYHAN BASIN
Turkey has a water potential of 501 km3 of which 274 km3 is lost to evapotranspiration. 69
km3 feeds aquifers and 158 km3 flows to seas and lakes. Surface runoff is 193 km3 of which
98 km3 is usable. 31 km3 is consumed out of this 98 km3. 41 km3 of ground water recharge is
added surface run-off to supply a total of 234 km3 of renewable water potential. 14 km3 of
ground water resources is safe yield and hence total usable net water resources add up to
112 km3. Total consumption sums up to 43 km3 of which 12 km3 supplied by ground water
resources (Cakmak et al., 2008). 75 percent of this consumption belongs to agricultural
sector to irrigate approximately 5.28 million hectares of land (approximately 60 percent of
total economically and technically feasible irrigable area and 23 percent of total cultivated
area) (DSI, 2009). Almost all of the irrigation schemes are managed by farmers either in the
form of Water User Associations (WUA) or as Village Communities (VC). Irrigation water is
priced on per hectare basis in Turkey. Rather than being based on cost recovery criteria,
3
price depends on the type of crop and season of irrigation. The average price of irrigation is
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
around 70 Euro/ha (DSI, 2006a).
Although current stress on water resources of Turkey at the national level is considered to
be moderate, it is expected to increase significantly in the near future with the increasing
competition for water for the industrial and domestic use (Cakmak et al., 2008; Alcamo et al.
2007). Furthermore, current irrigation management policies of Turkey are “far from forming
an integrated framework for effective management of water resources” (Cakmak et al.,
2008, p.15). The need to develop a vision about the issues related to the state of water
resources in Turkey is urgent if policy makers intend to take the measures necessary to avoid
the possible negative impact of increasing stress on water resources.
Seyhan basin is located in the eastern Mediterranean. Seyhan River, which is formed by
confluence of Zamantı and Göksu Rivers, drains the Çukurova plain and discharges to
Mediterranean Sea. The basin consists of 20,450 km2 of land and an average water flow of
8.01 km3 (DSI, 2007). Total irrigated area is about 271 thousand hectares which is around 45
percent of total cultivated area in the region (DSI, 2006a; TURKSTAT, 2009). Irrigation ratio is
quite high compared to the national average of 23 percent. Almost all irrigation schemes are
managed by Water User Associations and only 15 percent of irrigable area is used as rainfed.
Çukurova plain is among the most important agricultural production areas of Turkey. Seyhan
Basin’s share in total harvested area is about 3 percent (Table 1). On the other hand, 4
percent of total agricultural production value is obtained in Seyhan Basin. The share of
4
Seyhan Basin in production rises as high as 11 percent for oil seeds and 7 percent for cereals.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Most distinguishing characteristic of Seyhan Basin is reflected in yield numbers. The average
yield of Turkish agricultural production is around 70 percent of yields in Seyhan Basin. This
ratio is as low as 50 percent (indicating the fact that yields in Seyhan Basin is two times the
overall average) for pulses. Seyhan basin is also important in vegetable and fruit production
for which share of Seyhan basin in national production is 5 percent.
Table 1. Some agricultural indicators about Seyhan Basin, 2008
Cultivated
area (%)
Harvested
area (%)
Production
(%)
Total Crops
3.21
3.39
3.88
Ratio of
yields
(Turkey/Seyhan)
0.70
Pulses
1.36
1.58
1.57
0.50
Industrial Crops
6.51
6.52
1.78
0.72
Cereals
2.83
3.01
6.66
0.72
Oil Seeds
7.71
7.73
10.74
0.93
Feed Crop
0.75
0.78
1.02
0.74
Tuber Crop
3.20
3.21
4.42
0.64
Vegetables
N.A.
N.A.
4.90
N.A.
Fruits
1.84
N.A.
4.91
0.44
Source: TURKSTAT (2009)
Regional population growth rates and ratio of urban to rural population are depicted in
Tables 2 and 3. The figures show that Seyhan Basin is highly urbanized. The ratio of urban to
rural population is significantly higher than the national ratio, with the rural population
consistently declining, while the urban population increasing. Since growth or total
population in the region is also positive, it can be concluded that in-migration to the basin is
also an important factor in the demographic dynamics of the region.
Table 2. Yearly Average population growth rates
1975
Seyhan Basin
Total
Urban
Rural
1.82
2.03
1.37
Total
1.68
5
Turkey
Urban
1.88
Rural
1.48
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
1980
1990
2000
2008
1.82
1.41
-1.16
1.33
1.81
1.51
1.13
1.50
1.83
-1.25
-1.37
-1.59
1.61
-1.26
1.52
1.24
1.75
-1.28
1.66
1.47
1.47
-1.23
1.11
-1.49
Source: TURKSTAT (2009) Website
Note: The averages are calculated in a compound basis
Table 3. Ratio of urban population to rural Population
1970
1975
1980
1990
2008
Seyhan Basin
1.03
1.32
1.31
2.31
6.71
Turkey
0.62
0.72
0.78
0.75
2.99
Source: TURKSTAT (2009) Website
There are 6 dams in the basin. Four of these dams are used for irrigation. Their water
holding capacity add up to 4500 hm3 with irrigation capacity of about 350,330 ha (Table 4).
Although the irrigated area steadily increases in the basin, its share in the total cultivated
area does not change much (Table 5).
Table 4. Dams in the Seyhan Basin
Dam
Setup
Year
Berke
1999
Çatalan
1997
Kesiksuyu
1971
Kozan
1972
Nergizlik
1995
Seyhan
1956
Arıklıkaş
1998
Aslantaş
1984
Kalecik
1985
TOTAL
Normal Irrigation
Volume
Area
hm3
(ha)
427
2,126
53
8,760
170
10,220
22
2,326
1,200
174,000
1,872
285
1,150
149,849
33
4,890
7,053
350,330
Source: DSI (2009) Website
6
Power
(MW)
Annual
Production
(GWh)
510
1,672
169
596
59
350
138
569
876
3,187
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Table 5. Irrigated land in the Seyhan Basin, 2001-2008 (ha)
Irrigated
Land
2001
2002
2003
2004
2005
2006
2007
2008
250,279
258,405
264,381
306,224
279,457
279,113
267,332
270,965
Total
Share of
Cultivated Irrigated
Land
Land
581,459
43.04
576,388
44.83
612,645
43.15
595,996
51.38
594,363
47.02
562,195
49.65
554,265
48.23
555,427
48.79
Source: DSI (2004, 2005, 2006b, 2007b, 2008), TURKSTAT (2009).
PARTICIPATORY SETTING, MAIN ISSUES AND PRESENT STATE
Careful selection of stakeholders was necessary to obtain a scenario that will both help the
domestic policy makers and provide feedbacks to the pan-European scenario building
activities for the SCENES project. The participants were from diverse societal groups.
Representatives from central and local public institutions related to water and agriculture,
environmental and farmers’ NGOs, irrigation associations and local university participated in
the workshop.
The stakeholder workshop was conducted both in forum and group settings depending on
the issue that was covered. There were also cases where participants were asked to convey
their personal views. The personal and group forms were collected, processed and the
results were immediately shared with the participants. Homogenous groups were formed to
save time in reaching consensus, but the results obtained from the groups were finalized in a
forum discussion. The groups formed by the stakeholders were:
7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
(1)
Training and Academic Personnel
(2)
Technocrats and Bureaucrats
(3)
Farmers and Irrigation Associations
(4)
Non-governmental organizations (NGOs)
Despite the fact that many of the participants were not used to this kind of meetings and
there were a lot of work to do, the participation remained high throughout the workshop.
It was necessary to identify the main issues/drivers to start the FCM process. Stakeholders
were first given a list of predetermined issues identified by the researchers during their
prior visits to the basin area. The list was formed by a combination of social, economic and
environmental issues that may be crucial in forming water scenarios (Table 6).
Table 6. List of Predetermined Issues
Issues
01
Rate of recycled waste water
02
Impact of Increasing Urbanization
03
Maintenance, Repair and Overhaul
04
Environmental Consciousness
05
Drainage Problem
06
Internal Migration
07
Impacts of Climate Change
08
Employment
09
Decrease in Forestry
10
Impact of Industrial Production
11
Water Supply
12
Water Delivery Losses
13
Support for Publications on Water Use
14
15
16
17
18
19
20
21
22
23
24
25
Water Demand
Use of Water-Saving Methods
Irrigation Infrastructure
Irrigation Water Pollution
Irrigation Water Use
Price of Irrigation water
Irrigation Efficiency
Sustainable Water Management
Agricultural Support Policies
Increase in Agricultural Output
Salinity
Ground-Water Use
However the participants were not restricted with the predetermined list. They were
encouraged to contemplate and add any missing issues. Eventually, the participants came up
with a list of 32 variables with the context of the issues precisely defined. It was encouraging
to note that there were not many divergence views on the issues. Almost all participants
8
agreed on the relevance of predetermined issues and the ones that were added by other
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
participants.
Participants were then divided into four groups to decide on the most important issues. Final
list of top 15 issues were decided by the ratings provided by the stakeholders (Table 7). The
list seems to reflect the sectoral orientation of the participants. Most of them are closely
related to irrigation and agriculture. Given the fact that more than 80 percent of the water
used in irrigation, this was not a surprising result. The environmental issues survived to the
extent that they were related to agriculture.
Table 7. Final List of Variables/Drivers
Variables/Drivers
D01
Impact of Increasing Urbanization
D09
Irrigation Water Use
D02
Water Supply
D10
Irrigation Efficiency
D03
Water Demand
D11
Water Pollution
D04
Irrigation Water Price
D12
Use of Water-Saving Methods
D05
Agricultural Support Policies
D13
Sustainable Water Management
D06
Impacts of Climate Change
D14
Soil Degradation
D07
Water Delivery Losses
D15
Use of Ground-Water
D08
Irrigation Infrastructure
Source: Workshop results
Following the establishment of final list of variables, it was necessary to establish the
present state of the selected variables. The participants were asked to assign weights to
each variable on a scale from one to five individually and the outcome was discussed in a
group setting. The scaling was as follows: 1 for Null, 2 for Very-Small, 3 for Small, 4 for Big,
and 5 for Very-Big. This corresponds to a modified spider gram exercise. The assigned values
represent the relative position of the designated variable with respect to its desired level
(Figure 1).
9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Figure 1. Overall Ratings of the Issues in the Present State
Source: Workshop results
The averages of the input from four groups for each variable were further discussed in forum
setting. Educational and Academic personnel were generally more optimistic while the
farmers sketched a rather pessimist view about the state of the variables. However
divergence was not significant among the groups. The main issues in which divergence of
views was relatively high were water demand (D03), water use (D09) and use of water saving
technologies (D12). Farmers thought that first two (D03 and D09) were sufficient while NGO
representatives believed that the last one (D12) was far from being satisfactory. Agricultural
support (D05) and use of water-saving technologies (D12) and sustainable water
management (D13) were considered to be far less satisfactory compared to their desired
level.
10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
METHODOLOGY
Fuzzy Cognitive Maps, introduced by Kosko (1986), are used in many fields varying from
ecological modeling to information systems; to model different things varying from
ecological systems to producing intelligent decision making engines.
Examples of fuzzy cognitive modeling in the literature are (but not limited to): The study of
Siraj et al. (2004) where FCMs and fuzzy-rule bases are used to create a decision engine that
detects intrusions to a computer system. Sharif and Irani (2006) use fuzzy cognitive mapping
and morphological analysis to formulate a conceptual model of decision making behavior
within the information systems evaluation task. Özesmi and Özesmi (2004) utilize FCM to
create ecological models with both expert and local people’s knowledge. Similarly, there are
studies in the literature where FCM is used to model political developments (Taber, 1991),
electrical circuits (Styblinski and Meyer, 1988), and virtual worlds (Dickerson and Kosko,
1994).
Recently, FCMs are widely used to facilitate public participation by modeling local and expert
knowledge. Studies that utilize FCM in such a fashion include: Özesmi (2006), where FCM is
used to map social and economic conditions of local people before their resettlement due to
a dam construction. Kastens and Newig (2008) study active involvement of regional
stakeholders in North-west Germany in effective implementation of the WFD. Mouratiadou
and Moran (2007) use FCM to model stakeholder and public perceptions on water related
issues in a river basin in Greece, thereby promoting the involvement of stakeholders and
11
public for successful implementation of WFD. Özesmi and Özesmi (2003) use FCM to create a
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
participatory management plan for a lake ecosystem.
It is necessary to normalize the state vector to ensure convergence in FCM iterations. In this
way we also avoid problems related to the scaling, and interpret the negative numbers
obtained at the end of iterations more easily. Normalization is done according to the
formula:
si 
si 
s
i i
n

 si
 i  si  ni




2
(1)
where si is the ith element of the state vector S before normalization while si is the
normalized value of si and n is the total number of elements of S .
We have also rescaled the elements of the transition matrix to 0 – 1 interval according to
formula
Aij 
Aij
max  Aij 
(2)
where Aij is the element of the transition matrix A in the ith column and jth row. Aij is the
normalized value f Aij .
Dynamic analysis of FCM is done by multiplying the normalized state vector with the
rescaled transition matrix iteratively until convergence is reached for all variables. To ensure
convergence we have weighted the state vector at each step. The state vector at iteration t
is calculated by
12
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

1
St  

 max abs min si,t S1t 1

  
 , max  s 
St 1
i ,t 1

 S  A for t  2...T
 t 1


where St is the state vector at step t of iterations, si,t S1t 1 is the ith element of the normalized
state vector St1 at iteration t  1 . St is normalized at each step according to equation (1).
A follows from equation (2).
Convergence to undesired steady-states is highly possible in the dynamic analysis since FCM
is a nonlinear system outcome and its development FCM relies heavily on human experience
and knowledge. In order to direct a system to a desired steady state several methods are
used. Learning algorithms are one of the methods used for this purpose. A learning
algorithm basically determines the weights and outlines the convergence for a neural
network to reach a desired steady state via local search techniques. When applied to the
case of FCM, this process corresponds to updating the strengths of causal links
(Papageorgiou et al., 2004).
There are various types of learning algorithms proposed in the literature. Papageorgiou et al.
(2004) list these algorithms as: Differential Hebbian Learning (DHL) (proposed by Kosko, an
unsupervised learning without any mathematical formulation), Adaptive Random FCMs
(given initial state this algorithm adapts weights so that FCM converges to a desired steady
state), Nonlinear Hebbian Learning (adapts the magnitude of non-zero weights only), Particle
Swarm Optimization, and Active Hebbian Learning algorithm (an algorithm with
mathematical formulation).
13
(3)
Our approach on introducing the learning mechanism was to focus on 7 key variables: Water
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
supply (D02), Water demand (D03), impacts of climate change (D06), Irrigation water use
(D09), use of water-saving technologies (D12), sustainable water development (D13) and
irrigation efficiency (D10). These variables are selected according to the number and
magnitude of their links with each other. In this sense, we used a modified Nonlinear
Hebbian Learning Algorithm. For this purpose we have compared the 7-element subsets of
the combinations of 15 variables in terms of number of their “inner” links and strength of
these links. Since it is obvious that impacts of climate change (D06) and sustainable water
management (D13) have quite “strong” links, instead of inspecting all combinations we have
compared the 7-element subsets of 15 drivers which consists of these two variables.
Suppose li , j denotes the magnitude of the link from driver i to driver j and let ki , j be
defined as

0 if li , j  0
ki , j  
for i  and j 

1 if li , j  0
where  is the set of drivers.
Then we define
N  t    i , jC li , j t  1,,1287
t
and
M  t    i , jC ki , j t  1,,1287
t
where Ct is the t th 7-element subsets of combinations of 15 drivers which include the
impacts of climate change (D06) and sustainable water management (D13). Then we
ordered all subsets in a descending order with respect to their
14
Nt
scores and applied the
Mt
FCM framework described above to the ones that are in the top 250. In this case, the subset
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
which consists of the elements “Water supply (D02), Water demand (D03), impacts of
climate change (D06), Irrigation water use (D09), use of water-saving technologies (D12),
sustainable water development (D13) and irrigation efficiency (D10)” gives the best result.
Our criterion for the best result is stability of all variables and maximum possible
The subset mentioned above has the second largest
Nt
score.
Mt
Nt
score and all variables in the set
Mt
become stable at the end of the iterations. However to reach stability, we introduced 0.25 to
diagonal elements of transition matrix. This implies self-maintenance for all variables. That is
to say, level of a variable in the current term increases its level in the next term. Considering
the nature of the issues under investigation, this is not a poor assumption. Furthermore,
better results are obtained when a link from “impact of climate change (D06)” to
“sustainable water management (D13)”. This link implies that as the effects of climate
change increases water authorities will put more emphasis on sustainable water
management.
RESULTS AND ANALYSIS
The participants were asked to create two desired future state vectors for 2015 and for 2030
in preparation for FCM analysis. Every group assigned desirable values for each of the 15
variables for the two future dates. Average values for each variable were calculated,
discussed and finalized in forum setting. The future state vectors are used for two purposes.
First, they are compared with the results of FCM analysis to check the consistency of the
15
results. Secondly, they formed the basis in the scenario building phase. The results are given
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
in Table 8 below, and the output from group work can be found in Appendix Figure A3.
Table 8. Current and Future States of Variables
TODAY
2015
2030
TODAY
2015
2030
Impact of
Increasing
Urbanization
Water Supply
4.00
2.53
1.75
Irrigation Water Use
4.25
4.00
4.18
4.00
4.33
4.65
Irrigation Efficiency
3.00
4.45
4.85
Water Demand
3.75
3.75
4.20
Water Pollution
3.75
1.93
1.08
Irrigation
Water Price
Agricultural
Support Policies
Impacts of
Climate Change
Water
Delivery Losses
Irrigation Infrastructure
3.00
3.08
3.13
2.00
3.75
4.43
2.50
3.95
4.55
2.75
4.07
4.88
4.00
2.13
1.58
Use of Water-Saving
Methods
Sustainable
Water Management
Soil Degradation
4.00
1.83
0.98
4.50
1.65
1.28
Use of Ground-Water
3.00
2.28
1.63
3.00
4.30
4.95
Source: Workshop results
A declining trend in the impact of increasing urbanization is observed. This shows that the
participants expect urbanization effects to be weaker in the future, which is consistent with
the observations about the demographic dynamics of the region. Water supply, water
demand and water price increases in time, suggesting a possible shift of agricultural
production to the water intensive crops which create higher value added. However,
irrigation water use does not change significantly while use of ground water sources declines
considerably. This is consistent with decline in water pollution and soil degradation as well as
increase irrigation efficiency, water delivery losses, use of water saving methods.
Improvements in irrigation infrastructure, agricultural subsidies and sustainable water
management are the key drivers for these enhancements.
16
The participants filled the matrices individually towards the construction of the FCM. They
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
defined weights and direction of relationships among selected issues. Average values were
calculated for each entry and the results are presented to the stakeholders. Afterwards, the
participants were again divided into four groups, where they were asked to discuss the
matrices further. The final FCM was established in a forum setting by asking directions of the
links and their final weights. The final outcome depended heavily to the outcomes of the
group work. As a result, it was possible to obtain one FCM for the pilot area and, also four
FCMs from each group. The final FCM is presented in Figure 2.
Figure 2. Final FCM
17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
No facilitators are assigned to the groups, but in any case participants seemed to handle
FCM framework well. Since the groups were homogenous, cognitive and social learning was
limited, but this was a necessary sacrifice due to the time constraint. Designated
relationships among the issues in the transition matrix are presented in Table 9. A variable in
the first column of the Table affects a variable in the second column. The size of effect is
given in the last column. A negative number in last column implies a negative relationship
between the variables. That is, as the level of the variable in the first column of the table
increases, the level of the variable in the second column of the table decreases, and vice
versa.
Most of the relationships in the table are as expected. However, some of them were
unexpected. For example water supply increases water while water demand decreases
water supply. It is possible that participants assumed that availability of more water will
encourage the cultivation of more water intensive crops resulting in an increase in water
demand. Participants also believe that as demand increases irrigation water usage will
decline. This recalls an explanation by the competition between farmers such that if demand
by all farmers increases amount of water available for any farmer will decline.
Table 9. Relationship between Issues Used to Form Transformation Matrix
Affecting Issue
Affected Issue
Impact of Increasing Urbanization
Impact of Increasing Urbanization
Impact of Increasing Urbanization
Water Supply
Impacts of Climate Change
Water Demand
Soil Degradation
Water Pollution
Water Demand
Water Supply
2
3
3
1
-3.5
Water Demand
Water Supply
-1.5
18
Size of effect
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Water Demand
Irrigation Water Use
-2
Water Demand
Water Demand
Use of Ground-Water
Irrigation Water Price
2
1.5
Irrigation Water Price
Water Delivery Losses
Use of Water-Saving Methods
Irrigation Water Price
Agricultural Support Policies
Agricultural Support Policies
Water Demand
Water Demand
Water Demand
Irrigation Water Use
Irrigation Water Use
Use of Water-Saving Methods
Impacts of Climate Change
Water Delivery Losses
Water Delivery Losses
Irrigation Infrastructure
Irrigation Infrastructure
Irrigation Infrastructure
Irrigation Water Use
Irrigation Water Use
Use of Water-Saving Methods
Irrigation Efficiency
Water Pollution
Soil Degradation
Use of Water-Saving Methods
Sustainable Water Management
Sustainable Water Management
Use of Ground-Water
Irrigation Water Use
Use of Water-Saving Methods
Irrigation Infrastructure
Water Delivery Losses
Irrigation Efficiency
Irrigation Water Use
Irrigation Efficiency
Water Demand
Irrigation Efficiency
Sustainable Water Management
Soil Degradation
Water Pollution
Sustainable Water Management
Use of Water-Saving Methods
Soil Degradation
Soil Degradation
Use of Water-Saving Methods
Impacts of Climate Change
-3
2
-2
-2
3
4
-3
1.5
1
-3
3
3
3
3
4
1.5
2
3
3
3
-2.5
-1
3
Source: Workshop results
The outcome of dynamic analysis is given in Figure 3. All variables are stabilized in the
interval of 1 and -1 which shows that the system gives comparable and meaningful results.
Dynamic analysis shows that, at the end of the iterations, only irrigation efficiency and
sustainable water management maintain their levels in the state vector. All other variables
are expected to be in a lower state compared to their initial state.
19
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Figure 3. Result of the FCM Analysis
Source: Authors’ calculations from the results of the workshop
Accordingly, impact of urbanization will be declining in the future. This is consistent with the
declining urbanization rates observed in long term time-series. Water supply and water
demand will be declining resulting in a decline in irrigation water use. Relatively higher
impact of climate change is probably the underlying dynamic for declining water supply and
demand. The decline in supply is compensated with better sustainable water management
practices which will not lose importance in the future. This is supported with the relatively
high levels of irrigation efficiency, lower soil degradation and lesser user of underground
water as well as declining water delivery losses. Decline in the irrigation infrastructure calls
for more investment.
20
As mentioned in the methodology section, we introduced learning mechanism for the
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
limited subset of the key variables to obtain a better picture of the FCM with 15 variables.
The final results are depicted in Figure 4. The series labeled as “2009” shows the current
state vector. “2030/Estimated” shows the results of FCM analysis while “2030/Desired”
shows the values of the future state vector that is obtained directly from the participants in
scenario building session. The results with learning mechanism amplify the picture obtained
with the full set of variables. Participants expect a decline in water supply, water demand
and consequently water use. Improvements in the states of efficiency, water saving
technologies and sustainable water management will be the major dynamics underlying the
changes in demand, supply and use. Hence the ultimate effects of climate change will be
lessening.
Figure 4 Results of Dynamic FCM Analysis with Learning
21
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
After the construction and presentation of the final FCM, the stakeholders were requested
to write scenarios for 2030 in group setting. They were asked to define milestones along the
way, and deciding on activities related to these milestones which take place within the
framework of STEEP (Social, Technological, Economic, Ecological, and Political) factors. The
consolidation of the scenarios were accomplished after the workshop.
While groups had differences in their desired future statements, the milestones defined by
different groups had some common elements. These common milestones can be used to
form a general timeline, with milestones grouped under four main headings: required
infrastructure, water related policies, technology and legislations. Consolidated story lines
are presented in Table 10.
The consolidated timeline starts with new projects on water infrastructure, and
maintenance on the existing infrastructure as the budget allocated to maintenance
payments increases. This is followed by the establishment of a water high commission
(WHC), leading to creation of several water policies and inspection of existing water facilities
in the basin. Completion of Yedigoze Dam construction increases energy production in the
basin. Meanwhile studies (mainly on land quality classification and sustainable land use in
the basin) supported by the WHC start the transition to water saving technologies, and
adoption of renewable energy sources. In the following years, we see more farms using
closed irrigation system, and doing non-polluting and organic agriculture. A significant
improvement is observed in water pricing: prices reflect real costs. In 2030, all irrigable lands
in the Seyhan River Basin are irrigated with water saving methods, land quality is better and
there are more green lands. Overall, the region is more prosperous in 2030.
22
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Common to all groups is the need for new and water saving infrastructure for water delivery
and storage. The need for new and renewable sources of energy emphasizes sustainability
concerns of the stakeholders. The establishment of WHC may be linked to stakeholders’
desire to use water resources efficiently. Another benefit of WHC is elimination of the lack of
studies on land quality classification which is expected to contribute to efficient use of water
resources.
The focus of the scenarios is on improvements in technology and legislation to achieve
sustainable water and land management in Seyhan. The group of education and academic
personnel emphasized the importance of irrigation technology, water storage, renewable
energy and legislations/studies on land. The Technocrats and Bureaucrats group focused on
social and environmental issues and indicated the importance of projects aiming to reduce
the pollution and higher social welfare. Farmers, on the other hand, put a special emphasis
on improvement and maintenance of water supply and development of agri-based
industries. Meanwhile the NGO representatives have developed their scenario around
increasing regional prosperity by proper urbanization, water saving technologies, research
and development, social awareness and government support.
23
Table 10. Story lines in consolidated time line
2008
Milestones
*new projects on water
infrastructure and
maintenance start
2010
2015
*establishment *Yedigöze Dam
of water high
completed
commission
*legislations on
sustainable land
use
Social
* increase in
rural welfare;
increase in
cultural and
educational
level
2020
*start of
transition to
water saving
irrigation
technologies
*increase in
social welfare
2025
*storage and
distribution
infrastructure
completion
Non-polluting and organic agriculture on RB level
*completion of
potable and
recycled water
systems
*Imamoglu
irrigation facility
opens
*development of *use of drip
infrastructure for irrigation and
providing energy sprinkler irrigation
requirements of
closed irrigation
systems
24
*Irrigation of all
irrigable lands in
the RB
*EU
Membership
finalized
*improvement *decrease in
in income
diseases
distribution
*development
of "social
fairness" among
producers
*improvement
in construction
technology and
materials
2030
Transition to closed irrigation system
*development
of renewable
energy sources
*studies on land quality
classification
Technological
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
*increase in
*decrease in
human life-span internal
migration
*use of water
saving
technologies
*use of
environmentfriendly
agricultural
technologies
*improvement *regional planning;
in water storage technocity;
technologies
improvement in
transportation
technology
Environmental
Economic
2008
Policy
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
2010
2015
2020
2025
2030
*increase in
energy
production
*increase in
urban
transformation,
increase in
income,
industrial
supports
*improvement in
socio-economic
indicators (e,g,
income
employment)
*development
and
improvement of
agro-based
industry
*attractiveness in
economic activities
and increase in
economic vitality;
increase in income
and employment
level
*decrease in
*decrease in
irrigation water land aridity
loss
*flood control; *decrease in
negative
water pollution
impacts on
endemic species
*regulation o water
distribution
between ASO and
Imamoglu water
facilities *weed
control; newly
acquired areas
*prevention of *decrease in
land
land and water
degradation and pollution
salinization
*improvement in
land quality
*increase in green
lands
*increase in the budget
*establishment *legislations on
allocated to maintenance of water policies making land use
payments
and inspection in accordance to
of existing water land quality
facilities in the compulsory,
basin
prevention of
land partition,
increase in
investments
*implementation *government
*R&D and
of required legal supports for use of publications
regulations for
modern technology
sustainable water
management
25
*irrigation
*budget allocation
payments
for development of
reflecting its real water resources
cost
*proper
urbanization;
urban transition
policies; land use
policies
To sum up, the common ideas stemming from the scenarios can be summarized as the achievement
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
of sustainable water and land management and prevention of pollution in water resources.
Important issues that are put forward consists of:

water supply being fully utilized for human well-being,

implementation and completion of adjustment policies for possible changes in climate,

proper urbanization

irrigation of all irrigable lands via sufficient infrastructure

balanced distribution of water among different sectors
These results are consistent with our findings in FCM framework. Both approches suggest that
increasing water efficiency and water saving through more investment in irrigation infrastructure will
decrease the water demand and this will compensate the decline in the water supply due to climate
change. Ultimately, the impact of climate change will be decreasing and quality of water and land
resources will be improving.
CONCLUSION
In this paper, we present the findings of a qualitative scenario development process on water related
issues in the Seyhan Basin, Turkey. The main outputs of the participatory process were the
determination of the issues, drivers, and their interaction using fuzzy cognitive maps. The desired
state towards 2030 were also obtained by the by participants with the STEEP (social, technological,
economic, and environmental and policy) approach together with the necessary milestones.
26
In the current state, agricultural support, use of water-saving technologies and sustainable water
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
management are considered to be far less satisfactory compared to their desired level. We used a
modified dynamic FCM framework for analysis as well as a qualitative scenario development.
Dynamic analysis shows that, at the end of the iterations, only irrigation efficiency and sustainable
water management maintain their levels in the state vector. All other variables are expected to be in
a lower state compared to their initial state.
Final results suggest that participants expect a decline in water supply, water demand and
consequently water use. Improvements in the states of efficiency, water saving technologies and
sustainable water management will be the main dynamics underlying the changes in demand, supply
and use. Hence the ultimate effect of climate change will be lower. The common ideas in the
scenarios can be summarized as the achievement of sustainable water and land management and
prevention of pollution in water resources. Important issues that are put forward consists of water
supply being fully utilized for human well-being, implementation and completion of adjustment
policies for possible changes in climate, proper urbanization, irrigation of all irrigable lands via proper
infrastructure, balanced distribution of water among different sectors. Both methods suggest that
increasing water efficiency and water saving through more investment in irrigation infrastructure will
decrease the water demand and this will compensate the decline in the water supply due to climate
change. Ultimately, the impact of climate change will be declining and quality of water and land
resources will be improving.
27
References
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Alcamo, J., Flörke, M., Marker, M. 2007 Future long-term changes in global water resources driven by
socio-economic and climatic changes, Hydrological Sciences–Journal–des Sciences Hydrologiques,
52(2), 247-275.
Cakmak, E.H., Dudu, H., Saracoglu, S., Diao, X., Roe, T.L., Tsur, Y. 2008 Macro-Micro Feedback Links of
Irrigation Water Management in Turkey, World Bank Policy Research Working Paper Series, No. WPS
4781.
Dickerson, J.A. and Kosko B. 1994 Virtual worlds as Fuzzy Cognitive Maps. Presence, 3, 173-189.
DSI. 2004 DSI’ce inşa edilerek işletmeye açılan sulama ve kurutma tesisleri 2004 yılı mahsul sayımı
sonuçları [2004 Crop Census Results for irrigation and draining facilities built by the General
Directorate of State Hydraulic Works], General Directorate of State Hydraulic Works Statistics
Department, Electronic copy obtained from www.dsi.gov.tr on 12/05/2009
DSI. 2005 DSI’ce inşa edilerek işletmeye açılan sulama ve kurutma tesisleri 2004 yılı mahsul sayımı
sonuçları [2004 Crop Census Results for irrigation and draining facilities built by the General
Directorate of State Hydraulic Works], General Directorate of State Hydraulic Works Statistics
Department, Electronic copy obtained from www.dsi.gov.tr on 12/05/2009
DSI. 2006a Devredilen Sulama Tesisleri İzleme ve Değerlendirme Raporu: 1999-2006 [Monitoring and
Assesment Report for Transferred Irrigation Facilities: 1999-2006], Digital data obtained from DSI
İşletme ve Bakım Dairesi: Ankara
DSI. 2006b DSI’ce inşa edilerek işletmeye açılan sulama ve kurutma tesisleri 2004 yılı mahsul sayımı
sonuçları [2004 Crop Census Results for irrigation and draining facilities built by the General
Directorate of State Hydraulic Works], General Directorate of State Hydraulic Works Statistics
Department, Electronic copy obtained from www.dsi.gov.tr on 12/05/2009
DSI. 2007a Annual Activity Report of the General Directorate of State Hydraulic Works: 2006, Ankara:
DSI.
DSI. 2007b DSI’ce inşa edilerek işletmeye açılan sulama ve kurutma tesisleri 2004 yılı mahsul sayımı
sonuçları [2004 Crop Census Results for irrigation and draining facilities built by the General
Directorate of State Hydraulic Works], General Directorate of State Hydraulic Works Statistics
Department, Electronic copy obtained from www.dsi.gov.tr on 12/05/2009
28
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
DSI. 2008 DSI’ce inşa edilerek işletmeye açılan sulama ve kurutma tesisleri 2004 yılı mahsul sayımı
sonuçları [2004 Crop Census Results for irrigation and draining facilities built by the General
Directorate of State Hydraulic Works], General Directorate of State Hydraulic Works Statistics
Department, Electronic copy obtained from www.dsi.gov.tr on 12/05/2009.
Dudu, H., Chumi, S. 2008 Economics of Irrigation Water Management: A Literature Survey with Focus
on Partial and General Equilibrium Models, World Bank Policy Research Working Paper, No. WPS
4556.
Fujihara Y., Tanaka, K., Watanabe, T., Nagano, T., Kojiri, T. 2008 Assessing the impacts of climate
change on the water resources of the Seyhan River Basin in Turkey: Use of dynamically downscaled
data for hydrologic simulations, Journal of Hydrology, 353(1-2), 33-48.
Kastens, B., and Newig J. 2008 Will participation foster the successful implementation of the WFD?
The case of agricultural groundwater protection in North-west Germany. Local Environment, 13 (1),
27 – 41.
J. Kämäri, J. Alcamo and I. Bärlund et al. 2008 Envisioning the future of water in Europe—the SCENES
project, E-Water, 1–28.
Kosko, B. 1986 Fuzzy Cognitive Maps. International Journal of Machine Studies, 1, 65 – 75.
Mouratiadou, I., and Moran D. 2007 Public Participation in the Water Framework Directive: an
Application of Fuzzy Cognitive Mapping in the Pinios River Basin, Greece. Ecological Economics, 62(1),
66-76.
Özesmi, U. and Özesmi S. 2003 A Participatory Approach to Ecosystem Conservation: Fuzzy Cognitive
Maps and Stakeholder Group Analysis in Uluabat Lake, Turkey. Environmental Management, 31(4),
518-531.
Özesmi, U. and Özesmi S. 2004 Ecological Models Based on People’s Knowledge: A Multi-step Fuzzy
Cognitive Mapping Approach. Ecological Modelling, 17(1-2), 43-64.
Özesmi, U. 2006 Fuzzy Cognitive Maps of local people impacted by dam construction: their demands
regarding resettlement. Preprint, arXiv:q-bio/0601032.
29
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Papageorgiou, E.I., C.D. Stylios, and Groumpos P.P. 2004 Active Hebbian learning algorithm to train
fuzzy cognitive maps. International Journal of Approximate Reasoning, 37, 219-249.
Sharif, A.M., and Irani Z. 2006 Applying a fuzzy-morphological approach to complexity within
management decision making. Management Decision, 41(7), 930-961.
Siraj, A., S. Bridges, and Vaughn R.B. 2004 Decision Making for Network Health Assessment in an
Intelligent Intrusion Detection System Architecture, International Journal of Information Technology
& Decision Making (IJITDM), 3(2), 281-306.
Styblinski, M.A. and Meyer B.D. 1988 Fuzzy Cognitive Maps, Signal Flow Graphs, and Qualitative
Circuit Analysis. In: Preceedings of the 2nd IEEE International Conference on Neural Networks (ICNN87), San Diego, CA, 549-556.
Taber, W.R. 1991 Knowledge Processing with Fuzzy Cognitive Maps. Expert Systems Application, 2,
83-87.
TURKSTAT (Turkish Statistics Institute), 2009, Veritabanları [Databases], www.tuik.gov.tr, accessed
on 12/05/2009.
30