Decision-support modelling report
Introduction
The model was made to understand the different factors that can be influential in the supply of contaminants to the study area of Seyhan dam reservoir in Turkey. The Seyhan dam is located in the Adana basin and three rivers has its inflows to the dam. The dam is surrounded by three different land use types: urban, agricultural and mining. The city of Adana (1.717 millions people) is located adjacent to the dam in the south.
Three different scenarios is posed to further investigate the most influential parameters for the area and possible negative or positive interaction between them.
Analytical model
A system analysis was made to find out how the different parameters interact with each other in the system. The results of the influence matrix are shown in the cause effect diagram figure 2 .
The groundwater flow is the most dominant parameter in the system because it is the parameter that has the most effect on the others. However groundwater is also strongly affected by the other parameters. Agricultural area is a quite dominant parameter that is not much affected by the other parameters. River flow is a quite active parameter, mining and urban area are moderately active. Particle size (suspended in river water) is quite active but strongly affected by the other parameters.
Figure 1. Map of Adana basin and Seyhan dam. (Google maps 2017)
Graph above showing the strong and crucial influence parameters relationship to each other. The strongest interactions is from groundwater flow to river flow to particle size. The Urban area parameter is the only strong influence that has no other parameter that strongly affects it. Groundwater flow is the parameter which is the most active both affecting and being affected. Particle size is the member just being affected in the system.
Synthetical model
Multi-criteria evaluation
A Multi criteria evaluation method was made to evaluate the effects of the different parameters on the supply of contaminants to the Seyhan dam reservoir. The different parameters influence was weighted against each other for this specific environment.
The different scenarios was chosen by using the analytical model results. First we varied the most active parameters and then adapted the others to that scenario to get a more realistical result.
Scenario 1
Very low flow in the Seyhan due to flow control in a dam upstream (catalan baraji dam).
Seyan river is the largest river contributing to Seyhan reservoir.
PARAMETERS:
Mining area: small
Agricultural area: big
Urban area: big
River flow: very low
Particle size: clay
GW flow: low
Scenario 2
More agricultural area due to land use change.
PARAMETERS:
Mining area: small
Agricultural area: big
Urban area: big
River flow: high
Particle size: clay
GW flow: medium
Scenario 3
Dryer climate affecting groundwater flow and less agricultural area due to land use change.
PARAMETERS:
Mining area: small
Agricultural area: small
Urban area: big
River flow: medium
Particle size: clay sand
GW flow: very low
Graphs of utility values
Graphs was then made to illustrate the influence that the different parameters have on the supply of contaminants, figure 4. The utility values can be noticed on the y- axis. This is then used for the different scenarios proposed to get a value to multiply with the weight that the parameter has in the system. This gives the total utility.
Total utility table of scenario 1-3.
Discussion
The results of the MCE applied for the different scenarios shows that scenario 2 means the highest supply of contaminants to the dam reservoir and scenario 3 the lowest with almost ⅓ .
The result is mainly influenced by the parameters agricultural area and suspended particle size because the total utility value of them differs most in scenario 2 and 3.
The relative low values of total utility for scenario 1 is due to the very low value for river flow compared with the two other scenarios.
To choose parameters is difficult because there is a large amount of parameters affecting the system. Therefore some simplifications has been made, for example groundwater flow is affected by different parameters like hydraulic gradient, aquifer grain size, etc. In order to simplify only groundwater flow has been used in the model and not the main parameters affecting it.
Another problem we encountered in our specification of the parameters was particle size. First we interpreted it as both size of suspended particles in the river and reservoir and also size of particles in the surrounding soil. When we were to judge its utility curve it was discovered and posed a problem since size effects in opposite directions. We solved it by concentrating on the suspended particles and let the other factor merge into the groundwater flow parameter.
The main limitation of the modelling is that it requires a lot knowledge of the site in order to make realistic assumptions about how the parameters interact in the system. And some of the needed information is almost impossible to obtain.
Further the utility values graphs are extremely difficult to estimate for some parameters, for example how the agriculture area affects supply of contaminants to the river and therefore the graphs are greatly simplified as linear functions and that is a source of error.
The parameter that caused much confusion in the calculation of Total Utility was again particle size. How did the river flow effect our particle size in a way that was related to supply of contaminants? A low river flow does supply mostly fine fractions but lesser quantities. In relation to a high river flow with more energy and thus still lots of small particles but also a larger range in grain sizes. Would addition of another parameter such as sediment supply or erosion give us a more balanced system. Would it counteract the high influence of particle size?
Assignment 3_ Decision- Support modelling_ Miguel and Amie