case study index | terms of use | help

Case study:
Modelling site suitability for wind farms

Manual

1. Introduction
2. Context
3. GIS Data Management
4. GIS Modelling and Analysis
5. Results & Outcomes

Appendix 1
Appendix 2

4. GIS Modelling and Analysis

We have now got the basic ingredients we need to pull the previously shape files and the previously described Criteria together to carry out the GIS modelling. You will need to construct slightly different data together for the TWO different modelling approaches in this Phase I. Your first step will be to do the INITIAL MODELLING (IC). You should have in place the core datasets that you need and will now carry out the GIS processing to enable you to model all the criteria you need within a GIS project.

For each of the criteria listed below there are some tips and hints on how to start to model them. As already stated the case study is not intended to be a simple step-by-step guide. Second-level GIS students will be expected to have a basic grasp of the software and the data. This is also an assessed piece of work and as such students will be expected to demonstrate a basic grasp of GIS principles and a sound understanding of the data you are working with. This will be apparent in the way in which you put the datasets together and how you develop and model your criteria. This will also test your project management skills and how well you understand the task set before you.

Phase I: Initial and Modified Modelling

INITIAL CRITERIA MODELLING (IC)

You now have gathered 10 criteria together for the modelling. Again we will not be spoon-feeding but suggest to you that the different criteria should all be in a single format - raster - which will make modelling all of the criteria together a simpler process. This should be the case anyway if you have used Map Query and set up the criteria correctly. You should be careful to re-name as you go along especially as you are working with a great deal more than 10 files. This is especially important when using the Map Query and Map Calculator commands as they tend to rename things by just adding a version number to the end.

Hint - you can laboriously multiply each criteria one after the other but there is also an option in Map Calculator of multiplying them all together. This is one advantage of using Map Query when you are actually setting the criteria up. By going into Map Calculator you can click on the first criteria on the list. Double click on this and then double click on the multiply icon ( * ) and then select the other criteria in turn clicking on the multiply button after each one. This will then run all the criteria together and produce a result of your initial modelling.

You should now convert the results of your modelling into a vector format and label it IC.shp. This is then ready for some subsequent modelling which will be carried out in Phase 2

MODIFIED CRITERIA MODELLING (MC)

Having carried out the initial modelling we can see that there are a relatively small number of available sites. It may well be that when you carry out your own project, that no sites are identified at all. You might decide that, within reason, you can relax this initial approach. You can do this in a number of ways, as outlined below, but whatever you decide, you must justify and explain why you made these choices. The coherence and common sense and understanding of the problem will be contained within this explanation and will form an important element of the assessment as the approach you take will also show your knowledge of the GIS processes you have just carried out.

2 main approaches are suggested to creating and running the Modified Criteria (MC);

The first is simply to drop a number of the criteria. By modifying the process you will inevitable come up with a different result. You may choose to drop anything from 1 to 4 criteria but whatever you do must be explained and justified. It is not recommended to drop any more than four criteria and indeed you will have four times the work justifying their demotion.

A second approach and one that may be easier to explain and rationalise is to re-visit the specific values within one or more of the criteria. As an example, if you wished to modify Criteria 2, Slopes, then you might choose to broaden the slope range to pick up a bigger area. This could be done by re-modelling, in the example here, the criteria for slope as now being Greater then 5 degrees and Less than 20 degrees (rather than the >8 and <15 previously used). However this would have to be justified again.

The MC component of the project is an interesting one and we would recommend that you experiment. You may find that by narrowing one criteria and expanding another that you get a similar result. Or it might be that by relaxing the criteria you open up a number of new potential sites. There are no set rules on what results you will get and again, it is the process and the understanding of that process that matters. It is unlikely however that by relaxing 3 or 4 of the criteria, you will identify a smaller number of sites and this should give you more to play with in the final phase of the exercise.

You can now use the Map Calculator to multiply your modified criteria together. You may find it helpful as you go through to re-label the revised version of any one criterion as Criteria5a, where you have updated the previous layer, Criteria5. Make sure you use the revised version when modelling all the criteria together. The resulting shape file from this modified process can be labelled MC.shp.

In this example we have taken the step of changing the slope specifications (Criteria 2) as listed above and also extended the distance used in Criteria 6 to 1.5 kilometres. We have also excluded woodland (Criteria 10) as there is not many large areas of woodland near our area. This means that the MC approach uses Criteria 1,3,4,5,7,8,9 and new Criteria 2a and 6a.

You will now have finished Phase I of the project and we will now take our identified sites from both the IC and MC approaches forward to the final section of the project.

4.1 Final Technical Modelling - Phase II

The initial modelling phase having been carried through, we now need to carry out a further modelling, a Phase II in effect which will model some additional specific requirements for a wind farm. We will also introduce a secondary environmental impact element by looking at visual effects and population affected to further clarify and complete the siting process.

You will now have now identified a number of sites from our combined IC and MC modelling. There are a further series of criteria we need to incorporate in Phase II.

Area (Criteria 11)
We have identified a number of sites. Obviously not all of these will be big enough to house a wind farm, while other sites will be far too large for the energy output required. We therefore need to identify (in the sample case study) those sites from both IC.shp and MC.shp are larger than 25,000 square metres and smaller than 150,000 square metres (or if you prefer between 2.5 and 15 hectares). It is likely that the minimum size will be more relevant than the maximum size. Also these limits may change depending on where you are working so if there are problems with the data, then a possible second option would be to pick the 4 largest sites found under both approaches.

This process will require us to convert the result file containing the sites from a raster to a vector format (if this has not been already done). We will now have a set of vector polygons but how do we know the area of those polygons? To find this out we need to run an ArcView script (called calcarea - included with the pack) to estimate the area of the sites and then run a query to identify those within the specified range.
Again there is a presumption that students will have some knowledge of scripting but as an aide memoire, the key stages are to go to the project window, click on the script icon and then click on new. A script window appears and you should run the Script > Load text file command and choose calcarea. You then compile the script by clicking on the tick icon. You then click on the view window and then go back to the script window and run the script by clicking on the running man. We then have the data we need in the table file associated with the sites to do a minimum and maximum size query.

Environmental Impact
A further consideration, in part incorporated in the Environmental criteria already used, is the visual impact of the wind farm. Wind farms with turbines of up to 70 metres in height will be visible from a long way away. It is important therefore, to minimise this visual impact. There is a free extension, Viewshed which is specifically designed to identify areas from which specific elements in the landscape can be seen. When you load it up from the file >extensions area you will see two new icons just above the View window which look like an hourglass and a pair of binoculars.

This extension can be run from the mid points of the sites identified as being of the right size (generally by using the highest point on the site as a marker and using that height as the starting point for the viewshed modelling). You will need to use the original merged height file as your base. After clicking on the Binoculars Icon, you will be asked for some options. It is recommended that you set the tower height to 70m (the average size of a modern turbine), the offset height to 5m , the field of vision to 359 and make sure the far distance is at least 5000m. You then point and click in the centre of the site, holding the mouse button down. Draw a line out in any direction, preferably south-west, let go of the mouse button, and the visibility extension will run. This creates a polygon which shows the area from which the tower can be seen, its viewshed in effect. The polygons created (one for each of the 4 sites in both IC and MC) will need to be saved as shape files for the next stage of the modelling.

Population affected
While it is one thing to see the areas from which the wind farm can be modelled, it would also be useful to identify the populations affected within this area. We have attached a ward map with basic population information for the whole of the south east (Wardseast.shp). By taking each of the sites in turn it should be possible to cookie-cut using the Geo-Processing extension and identify from the table the populations which fall within the viewshed areas for each site. You will have to run a vector query on the viewshed file to find all those areas where Gridcode = 1 and convert this to a shapefile. You can then run this through the View > GeoProcessing > Intersect two themes option choosing the selected site theme and Wardseast theme. The resultant intersect file (you should name it within the GeoProcessing wizard) will have a Pop Column which you can select and then get a sum of the population affected by using the Field > Statistics option. This will be one of the outputs to your final table.

WARNING - when you look at the intersected file, check carefully for duplicates which are bound to appear. This will affect the total population figure. There are two ways of managing this problem.
1) The first approach is to use the GeoProcessing wizard. Select the top option Dissolve feature based on attribute. Choose Name as the attribute to dissolve and when requested by the wizard, select 'pop by average' as the field/option to be included in the output file. This will average the population values for all the duplicate records. When finished identify the total population of the affected area from the table (select column name pop and choose field > statistics)
2) The second approach is to put the table in edit mode, select the duplicate rows (shift click) and delete them (delete records). When finished identify the total population of the affected area from the table (select column name pop and choose field > statistics)

Final Choice of Site
Just when you though it was all over, the final stage of the whole process is to take a measure of the detailed wind speeds at the chosen sites to clarify and compare their suitability for a wind farm. This will essentially involve identifying the minimum and maximum wind speeds (generally found at the lowest and highest points) of each site as well as estimating the average wind speed and comparing these to a table of the population affected. In this way it should be possible to identify maximum potential power input combined with minimum impact and in effect, identify (and justify) our optimal site from the 8 we have identified under both MC and IC.

You will hopefully still have a copy of the Wind Speed surface layer you have created and you can overlay this with the vector sites files (IC.shp and MC.shp). Make the vector layer (IC or MC) active and choose one of the sites by using the select feature icon and drawing a box around it. Now make the raster wind speed surface layer active. Choose the drop down command Grid Analyst > Extract grid theme using polygon. Choose the next fault and then choose the chosen vector layer i.e. IC.shp as the polygon layer to be used. Choose the remaining default and you will be presented with a new raster grid file drawn around the chosen site. Make that layer active and then choose the command Grid Analyst > Calculates statistics of a grid theme. From this you can glean the mean, minimum and maximum wind speeds within that site.

Below is a pre-prepared table with those values as well as a data on the population affected. You will be expected to come up with something similar when you carry out the project yourself. Looking at the respective results, you might decide that for the IC modelling, the fourth site 4 (IC), has the highest average wind speed. This would suggest maximum power output (though this is tempered by the size of the site - perhaps another measurement you could work with) along with the smallest visual impact on local populations which makes it the best site. For the MC modelling the sites are bigger and higher and have higher wind speeds. Those close to town 1 (MC) and 4 (MC) have high scores for population affected, whereas 2 (MC) and 3 (MC) are a better balance. Again it would be a close choice with perhaps site 3 (MC) having the edge due to slightly higher wind speeds. Interestingly both of the chosen sites from each approach, IC and MC are very close together suggesting that across the board these sets of sites near Lewes will be the optimum location with 3 (MC) being the best location across the board.

Table 1a. Results for IC Modelling of Wind Farm Sites
Site Size (ha) Min Speed Max Speed Ave. SpeedSite Centre Pop Affected
1 (IC) 3.85 6.15 6.29 6.24 63,105
2 (IC) 10.68 5.74 6.12 6.01 59,011
3 (IC) 3.82 5.86 6.03 5.96 60,765
4 (IC) 4.37 6.29 6.52 6.34 24,420

Table 1b. Results for MC Modelling of Wind Farm Sites
Site Size (ha) Min Speed Max Speed Ave. SpeedSite Centre Pop Affected
1 (MC) 13.22 6.75 7.14 6.86 129,680
2 (MC) 11.35 6.29 6.61 6.42 24,420
3 (MC) 8.23 6.22 6.86 6.56 26,136
4 (MC) 11.11 6.66 6.89 6.74 115,320

Last updated: December 4, 2003