Site Map - Home : Methods and Techniques : Mapping and Decision Diagram
The Mapping and Decision (MAD) Diagram is a hand drawn diagram showing how spatial data is utilized throughout the project. It shows how source data layers are processed using selected geoprocessing functions to produce derivative layers, and how those layers can be used to produce additional layers and so forth.
The Mapping and Decision Diagram shown above (on the right) shows the relation between the Steinitz Framework for Geodesign and a stylized representation of a typical model diagram, similar to those produced by ModelBuilder, Esri's graphical environment for doing geoprocessing.
The following 19 slides describe the various components of the MAD diagram, concluding with an example of how the MAD diagram was used on a real project.
1 The Geodesign Framework, previously referred to as the Landscape Change Model, proposed by Carl Steinitz, consists of six models defining the nature of landscape (more broadly defined in this presentation as our geo-scape … the planet’s life zone), and how that landscape (geo-scape) might be changed.
The first three models comprise the assessment process, the second three models comprise the intervention process.
The assessment portion of the framework includes:
- Representation models (data) that answer the question, “How should the geo-scape be described?”
- Process models (information) that answer the question, “How does the geo-scape operate?”
- Evaluation models (knowledge) that answer the question, “Is the geo-scape working well?”
The intervention portion of the framework includes:
- Change models (data) that answer the question, “How might the geo-scape be altered?”
- Impact models (information) that answer the question, “What differences might the changes cause?”
- Decision models (knowledge) that answer the question, “Should the geo-scape be changed?”
The Geodesign Framework provides an excellent conceptual model for working with all aspects of assessment (of the existing conditions) and intervention (of the proposed changes). It does not, however, suggest how these models should be constructed or processed.
2 This modeling diagram, which is very similar to a ModelBuilder diagram, shows how spatial information is modeled to produce derivative information. In this case, the diagram shows how Elevation, Soils, Vegetation and Rainfall data is modeled (evaluated and combined) to produce maps showing Slope (a derivation of Elevation), Erosion Potential (a derivative of Slope, Soils and Vegetation) and Erosion Hazard (a derivative of Erosion Potential and Rainfall).
The rectangles in the diagram represent layers of spatial data. The circles represent a geographic function (process) used to convert the data in the input layer to the data in the output layer. Notice how the output from one function can be the input to another function.
One of the unfortunate limitations of the current edition of Esri’s ModelBuilder is that it represents the map layers (geo-spatial data) as circles and the processing functions a rectangles, which is just the opposite of what is shown above. Thus, making it virtually impossible to render graphic representations of the map layers as shown in the above diagram.
3 The Geodesign Framework (shown vertically in the first slide) is displayed horizontally across the top of the MAD diagram (shown above), where it serves as a reference to the flow of spatial information represented in the main portion of the MAD diagram.
Project data maps (layers of spatial information) are analyzed (using various geoprocessing techniques) to produce derivative maps, which are in turn used to assess, evaluate, or better understand relevant issues.
These issue maps are combined or otherwise analyzed (often using some type of group-base assessment process) to create a set of evaluation maps depicting the suitability for, or sensitivity to, a particular land use (or set of land uses).
These evaluation maps then serve as the context for developing alternative land use plans (or land use management strategies).
The alternative plans/strategies are then assessed to determine their respective impacts on the existing landscape (geo-scape) including physical, biological, social, economic, etc. conditions.
The decision makers can then use these assessments to decide if a proposed alternative is acceptable, to what degree, and when/if the alternative should be implemented.
The following slides describe this process.
4 The MAD diagram is predicated on an understanding of the various issues, principally spatial issues, effecting or otherwise influencing the development of the project.
Most issues can be measured. Spatial issues are issues that can be measured and mapped, that is, where the measured representation of the issue’s intensity can be mapped, such as seismic risk, environmental sensitivity, crime, agricultural potential, and so forth.
Having a clear understanding of the various issues (issues of concern) provides the basis for determining what data is need for the project.
5 Project data (spatial data) is data that is relevant and ready to use, as opposed to source data that may, or may not, be ready to use or even relevant. One of the pit falls of any project is naively assuming that the data needed by the project is on hand and in a compatible format.
Developing a rough-cut, or first draft, of a MAD diagram can help determine what data is need and to what degree it is relevant. Giving consideration to availability, extent, scale, taxonomy, and cost can provide the project team with an assessment of its value and what might be required to convert that source data into useful project data.
Once this is done, the project team should create a spatial database for the project. This database should provide an organized structure for storing and managing source data, project data, and all derivative data, as well as all planning/design alternatives and their subsequent evaluations.
6 Assessment models are geoprocessing models that assess or otherwise analyze the geo-spatial conditions with respect to a particular issue, such as the erosion hazard model shown in Slide 2.
Assessment models should be developed for each of the predominant issues effecting, or potentially effecting, the creation of a proposed plan or design.
Issues and their respective assessment models can pertain to most anything, including the geography of the study area, the surrounding environment, broader social concerns, community expectations, political agendas, and economics.
The results of each of the assessment models should be evaluated to assure the quality of their representations.
7 The final assessment models are used to produce a series of issue-based assessment maps showing the relative degree of favorable and unfavorable conditions associated with each issue.
At this point in the process, each of the condition assessment maps pertain to a single issue. These maps will be combined in the next two steps to produce a series of composite suitability and vulnerability maps, one set for each land use or land related activities associated with the (yet to be) proposed plan or design.
8 The issue maps are combined or otherwise analyzed (often using some type of group-base assessment process) to create a set of composite maps depicting the suitability (or vulnerability) of the landscape to a particular land use (or set of land uses).
The Delphi Process, developed by the Rand Corporation, is often used for making these multidisciplinary assessments. The technique, coupled with value scales for assessing goodness and the expected degree of influence gives individuals in a group the ability to make their individual assessments and then compare those assessments to those of the group as a whole.
This comparison initiates discussion within the group which often leads to a change in the assigned values and influences. This process is repeated until the group has reached stability. The resulting values and influences are then used to specify the input parameters to some type of combinatorial process, such as weighted overlay, to produce the composite maps.
9 The composite maps resulting from the combinatorial process, or suitability maps as they are sometimes called, show the spatial distribution of the project team’s measures of suitability (established during the group-based decision process) for each of the primary land uses, or land use management strategies, under consideration.
Multiple suitability maps, showing each of the specified measures of suitability (development suitability, development capacity, environmental sensitivity, etc.) deemed important by the project team, are typically prepared for each of the primary land uses, or land management strategies, under consideration. These maps serve as the platform, or background, for assessing the preferred locations for each of the primary uses.
10 The composite suitability maps, which show suitable locations for each of the predominate land uses, can be referenced by the design team as they develop their land use plans.
One can imagine sketching a proposed plan on a transparent sheet of paper (real or digital) that has been laid over one of the composite suitability maps; for example, one showing the preferred locations for commercial development. The designer could then visually reference that map as he/she sketches in the location of possible commercial areas. The process can then be repeated for each of the other land uses. This process does not produce a single solution but rather serves as a spatial guidance system leading to the development of many solutions.
This process, called geo-sketching, can be used to create a series of preferred locations for each of the predominate land uses, considering one land use at a time. That is, without giving any consideration to the other land uses. It can also be used (which is frequently the case) to sketch out complete plans.
While geo-sketching may be used to generate complete plans, the results are often sub-optimum, leaving the best alternatives unexplored. A much better approach would be to use geo-sketching to determine the preferred locational alternatives for each of the predominate land uses, and then use some form of morphological analysis to generate the best composite plans.
11 Each of the alternative land use plans should be fully delineated, showing an overview the plan, with selected (critical) areas of the plan rendered with greater detail, 3D views or 3D movies designed to give the viewer a sense of place, and performance dashboards showing the key performance indicators.
The value associated with this effort is two-fold: it provides the planning team, their sponsors, review agencies, and stakeholders with an informed and consistent view of each of the plans, it also forces the planning team to develop a deeper appreciation of the pros and cons associated with each of the plans.
12 The alternative plans/strategies are assessed to determine their positive and negative impacts on various aspects of the environment (physical, biological, social, economic, etc.). These assessments should provide an estimate of both the degree and probability of the near-term, mid-term and long-term impacts.
The decision makers can then use these assessments to decide if a proposed alternative is (or is not) acceptable, the degree to which it is (or is not) acceptable, and whether or not it should be implemented.
13 The results of the impact assessment should be presented in a manner similar to that used to present the proposed land use plans, giving equitable consideration to the depth of the analysis and the intensity of the graphics used to present the results of that analysis.
Given that most impacts occur over time, special consideration should be given to the use of change models (showing growth, dispersion, etc.) and how those models might be used to present time-dependent data.
User-centric dashboards, showing the key performance indications (KPIs) for each of the alternatives, should be developed for each of the decision groups (the planning team, the project’s sponsors, review agencies and stakeholders). Each of these groups will have different interests and different levels of interest. As such, they will need to be informed within the context of those interests.
14 The decision makers can use the various impact assessments to help decide if a proposed plan is (or is not) acceptable, the degree to which it is (or is not) acceptable, and whether or not it should be implemented.
If a plan is found to be acceptable and is ultimately implemented, then the changes instantiated in accordance with that plan become part of the project data used in subsequent planning studies. That is, they become part of the new existing conditions.
If none of the plans are acceptable then it is “back to the drawing board” which could mean a re-do of all or any part of the project.
15 Most land use plans include some sort of implementation strategy describing which portions of the plan get implemented over time. These strategies are just that, they are strategies, or guidelines, as opposed to detail plans.
Given the fact that the future is not 100% predictable, it is easy to see why the implementation of any plan usually requires some modification of that plan as it is being implemented. Adaptive management embraces this idea, and in so doing provides a means for managing these insitu modifications.
16 This diagram provides a conceptual representation of the ILARIS (Intrinsic Landscape Aesthetic Resource Information System) Model, developed by Jones and Jones, Architects and Landscape Architects, in Seattle, Washington. It was developed as a prelude to the development of a formal MAD diagram (shown in the next slide).
The ILARIS Model was used by Jones and Jones and the Trust for Public Land to assess the intrinsic scenic quality of the landscape in the near-shore areas of the Puget Sound.
17 This slide shows a portion of the actual MAD diagram used to develop the ILARIS Model.
Post-It notes, which are easy to move around, were used to represent segments of the model. Pencil lines, drawn on the background sheet, were used to connect the various elements (data layers and process nodes) of the model. Concept sketches, process diagrams, and related calculations were done on the fly as the diagram was being created.
It took approximately six man-weeks to develop the representational strategies for modeling the intrinsic aesthetic qualities of the landscape. It then took another two man-weeks to delineate and refine the MAD diagram.
Drawing a MAD diagram is easy. Thinking about what to draw is a bit more difficult.
18 ILARIS is an ArcGIS 9 ModelBuilder model developed for the purpose of identifying intrinsic landscape resources and measuring the magnitude of the aesthetic expression of those resources (in this case, for the near-shore areas in and around the Puget Sound).
The ILARIS model consists of approximately 50 data layers, 40 ModelBuilder sub-models, and three Python scripts. The models and scripts are organized into seven ArcGIS toolsets.
ILARIS provides a comprehensive decision-support framework, allowing the presence, quantity, and visual properties of different intrinsic landscape forms to be captured, and the significance of those forms to be identified. The model is used to assess the accumulative effect of the aesthetic view-shed and the relative uniqueness of Puget Sound’s landscapes.
19 The final model as well as its final output, a map showing the composite intrinsic landscape significance for the near-shore areas in the Puget Sound, can been seen in the Winter 2005 issue of ArcNews (Vol. 27 No. 4).
The ILARIS model won an Honor Award for Research from the American Society of Landscape Architects (2006).
See related resources:
- Understand the utility of going MAD
- View the PDF edition of this presentation
- Download the PowerPoint slides
Allen, David, Getting to Know ArcGIS: ModelBuilder, Esri Press, 2011. Available at Amazon.
Esri Staff, "Listening to the Land, The Role of GIS in Protecting the Intrinsic Landscape", ArcNews, Esri, Winter 2005. Content available at ArcNews Online.
Jones, Grant, Jones and Jones ILARIS: The Puget Sound Plan, Source Books in Landscape Architecture, Princeton Architectural Press, 2007. Available at Amazon.
Steinitz, Carl, A Framework for Geodesign: Changing Geography by Design, Esri Press, 2012. Available at Amazon.