T04. Database for GIS-based Suitability Mapping

Query tool for the visualisation and assessment of different components of GIS-based multi-criteria decision analysis used in the selection of suitable sites for MAR implementation.

Start using now!

To assess the tendencies and preferences applied by the scientific community for MAR site selection using GIS based Multi-Criteria Decision Analysis (GIS-MCDA) a review of available literature was conducted. It resulted in a database with 66 scientific documents from more 18 countries. Based on the database a query tool was developed. This tool will help to use the database in a structured way by querying in order to summarize, analyze and explore the obtained results.

The original database can be downloaded here. A simplified version was uploaded for the query tool.

The GIS-MCDA MAR site selection database

The database was constructed to store and retrieve information related to the methodology used by different authors that applied GIS-MCDA for MAR site selection. Some of the most relevant questions that were formulated and that can be answered by using the database are the following:

• Which MAR type is GIS-MCDA mostly applied for?
• Which water sources were regarded for the studies?
• Which criteria are used the most for GIS-MCDA MAR sites selection? Are there differences regarding the MAR type?
• How often is constraint mapping used for GIS-MCDA MAR site selection? Which thresholds are applied?
• Which weighting assignment methods are used?
• Which decision rules are used the most?
• How often is sensitivity analysis included in GIS-MCDA MAR site selection?

The database was constructed based on the retrieved criteria. This means that each reviewed document has the same number of entries as the number of criteria selected by each author. The following table present all attributes of the database and their meaning (see Table 1):

Table 1. List of database attributes for query tool

The database query tool

In order to query the information contained in the database, a pivot table application was created. Pivot tables are useful tools to query information in a flexible and intuitive way. Users can drag, drop, and filter attributes content to create tables and obtain statistical results.  Documents can be filtered to search for literature that should be further.

The pivot table implemented was developed by Nicolas Kruchten in an open-source third-party project named react-pivottable under MIT-License (https://github.com/nicolaskruchten/pivottable).

The tool interface

The INOWAS query tool works by dragging and dropping attributes from the database into the column or row fields. The user can filter the content of each attribute, can specify how the information should be displayed and add conditional or numerical operator to the pivot table. The following image presents the main objects of the query tool interface (Figure 1):

The user can specify how information should be displayed by the pivot table with the options Table, Table Heatmap, Table Col Heatmap, Table Row Heatmap, and Exportable TSV (Figure 2). The last option is useful to export filtered information.

Conditional and mathematical operator can be included to obtain specific results. Count, Count Unique Values and List Unique Values are the most commonly used operators for lingustic attributes (see Figure 2). Further operators that are specifically useful when working with numeral values are Sum,  Average, Median, Minimum, Maximum and Sample Variance.

Filters can be applied to attribute content. By clicking on the attributes, a list of the information that it contains will be displayed. The user can select which information should be displayed and counted in the pivot tables. When filters are active, the attribute name will appear in italic characters.

Query examples

The INOWAS query tool can be used in different ways because of its operational flexibility. Some examples will be provided based on the main questions that launched this project. These examples are prepared to get the user started on the usability of the tool but it should be kept in mind that there are many other options to visualize the same results or to query information.

EXAMPLE 1. How many and which documents are used for well, shaft and borehole (WSB) recharge analysis?

Querying the documents that are based on a specific MAR type and counting them can be achieved by following these steps:

1. Filter the attribute Well, shaft and borehole recharge to Yes.
2. Drag and drop Document reference attributes to the row field.
3. Select the operator called Count Unique Values. The database was constructed based on the number of criteria for each document. So, in this example the reference “Amineh, 2017” would be counted nine times, as nine criteria were used. Using count unique values will count only one value per document and not the number of criteria used for each document. Under the operator field it needs to be specified which attribute should be regarded for counting. In this case ID needs to be selected.

The results indicate that nine documents were used with well, shaft and borehole recharge as MAR type. The pivot table also displays the reference to search for it (Figure 3).

EXAMPLE 2. Which are the most used criteria for GIS-MCDA MAR site selection? Which are the most used criteria for GIS-MCDA In-channel modification site selection?

If the user what to know which are the most used criteria for GIS-MCDA MAR site selection and which are the most used for In-channel modifications, the following steps could be applied:

1. Drag and drop the Criteria modified attribute to the row field.
2. Display the pivot table as Table Heatmap.
3. Use the Count mathematical operator.
4. Click on the arrow next to the Operator field to organize counted values from highest to lowest value.

The results allow the user to see that slope was the criterion used the most, selected in 57 documents for MAR site selection. Land use/ land cover was used the second most often, followed by geology/lithology (Figure 4).

To find out which are the most used criteria for in-channel modifications site selection, the database needs to be filtered by documents that use in-channel modifications for their analysis. Select as a filter only the In-channel modifications with the Boolean value Yes. The results show that slope was the most used criterion, followed by land use/land cover and geomorphology (Figure 5).

EXAMPLE 3. Which criteria are ranked the highest for GIS-MCDA MAR site selection?

For this example, we want to know which criteria are preferred as the most important for GIS-MCDA MAR site selection. The concept of importance is based on the criteria with the highest rank assigned. For building this query, the following steps can be followed:

1. Drag and drop the Criteria importance rank attribute to the row field.
2. Filter the Criteria importance rank attribute by only showing 1 and 2 (most important criteria).
3. Display the pivot table as Table Heatmap.
4. Use the Count mathematical operator and sort from highest to lowest value.
5. Drag and drop the Criteria modified attribute to the row field underneath the Criteria importance rank attribute.

Using this procedure, the user can observe which criteria were assigned the highest importance rank the most. Results show that slope was the highest ranked criterion, followed geomorphology, land use/land cover and hydrological properties of soils (see Figure 6). Filtering Criteria importance rank to another number will show the listing for the respective criterion rank.

The visualized results are first ordered by the criterion rank. A better way to display the results would be to change the order of attributes in the pivot table row field. Shifting Criteria modified above Criteria importance rank will sort the results first by criteria category (Figure 7). This visualization makes it easier to look for a specific criterion and see if and how often it was used as highest ranked (Rank 1 or 2).

References of analyzed studies