Spatial Multi-Criteria Analysis for Humanitarian Mine Action

The positive impact that the Information Management System for Mine Action [IMSMA] has had in Humanitarian Mine Action [HMA] Programs around the world is undeniable. Kosovo, where IMSMA was deployed for first time, ended as one of the most successful HMA programs to date. A comprehensive understanding of the landmine problem is required to effectively manage an HMA program. Without the Information Management tools that IMSMA offers, this understanding would be very limited. Yet, it is arguable that IMSMA cannot be all things to all users, it does not contain all of the tools that an HMA program needs to make decisions about the most effective use of scarce funds and resources. There will always be a need for tools and methods that address specific problems and existing tools that have proven successful in other relief, development or governmental sectors should be at the top of the list for consideration.

 

The Problem 

 A large number of data collection and mapping tools are available within the HMA community. Most of these tools, have been developed specifically for enhancing reporting on different HMA activities. IMSMA is currently operational in more than 50 countries and is the only tool recognized as the international standard for the HMA operations. However, IMSMA lacks the tools necessary to support the planning and prioritization activities of an HMA program The other HMA tools that do support planning and prioritization do not generally have the geographic information system [GIS] functionality necessary for spatial analysis. In the context of this article, GIS does not refer to the ability to simply draw a map of a polygons, points or lines on a map. GIS for our purposes means the capability to evaluate the location and proximity of hazards in relationship to populated places, infrastructure and other critical resources necessary for communities to thrive. These GIS tools seek to answer some of the following questions:
  • How far away is a hazard? 
  • Are hazards close to schools or other important places like roads or markets? 
  • Are there many hazards that are close together? 
  • What are the terrain, vegetation and soil characteristics of specific hazards?

There is no “Super Tool” available, in the HMA community, which can efficiently perform all information capture, processing, reporting and management functions. IMSMA is not expected to, or able, to accomplish all possible information management tasks required or desired by HMA programs. Beyond IMSMA’s business process functionality, HMA planners must address strategic planning, operational planning, and cost estimation for projects, etc. The amount of information that is stored in IMSMA requires significant attention and resources to process and make it useful for planning and prioritizing of clearance activities. Mine Action managers need to be able to consider a wide range of factors to be able identify which HMA activities in which locations will have the greatest positive impact on communities and the economy. The impact of HMA on these communities must be measured not only in terms of the traditional mine action metrics like devices removed or destroyed and decreasing or elimination of future victim counts but must also be measured in terms of other national and local priorities. HMA planners and managers must be able to demonstrate how HMA activity contributes to positive outcomes in things like agriculture, economic development, infrastructure investment and stability to name a few.

 

National Mine Action Coordination Centers [MACC] often suffer from barriers between the Operations and Information Management Departments. The view that the Information Management Department should be a sub-set of the Operations Department illustrates a lack of understanding of information management processes, tools and the role they should play at all levels of HMA programs. There is a need for tools that not only facilitate intra-departmental information management work between HMA planning and operations but also facilitate this work on an intra-sectorial basis between HMA organizations and organizations engaged in other sector activity. Given this need, the author explored other tools that have proven their worth in governmental planning and/or socio-economic analysis fields, and that could be adopted to complement use of IMSMA in support of HMA policy and planning activities [see Figure 1].

 
Complementing IMSMA 
HMA programs depend upon IMSMA to manage data and information. The ability to adapt to the specific nature and needs of each HMA program presents a complex challenge. Information management plays, or should play a major role in prioritizing, planning and tasking operational activities. It is a prerequisite to have information management personnel qualified on not only information management principles, practices and technology, but they must also understand the processes and complexity of HMA programs.
 
SMCE within IMSMA Context

Figure 1: SMCA within IMSMA Context

 

Furthermore, program and operations managers face the challenge of limited resources. To maximize the use of those resources, Information Management Departments must seek to identify and implement the tools, which will allow for a more well-informed decision-making processes concerning strategic and operational planning, prioritizing HMA activities and tasking of HMA assets. Effective information management is key to guiding the efforts of decision makers. The process of analysis for prioritizing clearance requires not only HMA data and knowledge but also input from the wider relief and development arena. We do not undertake HMA as a standalone activity, we do them in support of wider relief and development objectives. This decision-making activity, with access to a large number of diverse high quality data sets, makes a perfect case for the application of Decision Support Systems [DSS].

There are many tools that can serve the DSS role, however, few are able to do so effectively while incorporating spatial data. In addition, analyzing large amounts of data of varying types and importance requires a transparent process of collaborative decision making in order to ensure credibility and user “buy-in”.

The Challenge
Prioritizing HMA clearance is a complex task, in part because there are often numerous stakeholders associated with the process, each with his/her own view as to what constitute critical considerations. The operations department, communities, local and national economic considerations, victim assistance concerns, and many other interests all come into play, and often these groups do not view the problem from the same perspective or with the same value scale. Therefore, a method is required that can weigh and accommodate the requirements of these stakeholders. IMSMA is limited in doing this because the analysis tools that it contains are non-spatial and the score that is generated is per community [this score is calculated in LIS data] and not per hazard area. 

There are also generic GIS tools like ArcGIS available. ArcGIS is a very flexible tool that allows for countless ways in which to conduct spatial analysis. Unfortunately, this flexibility comes with a cost. To effectively use ArcGIS as a DSS tool, users must have a high level of technical GIS skills, require a superior understanding of analytical modeling and must be able to create all or most of the steps in the analytical process from scratch.

Therefore, the challenge is:
To complement HMA’s use of IMSMA for planning, prioritizing and executing clearance activities by incorporating as many relevant spatial data sets and stakeholder priorities as possible while keeping the process transparent
 
Meeting the Challenge
In concert with IMSMA, Spatial Multi-Criteria Analysis [SMCA] provides an effective set of tools for addressing this challenge. One of the most important benefits provided by SMCA is the ability to include both spatial and non-spatial data in the analysis.

SMCA is a DSS tool, and as such does not make a decision on behalf of the user. Rather, it aids the user in making well-informed decisions. The problem being analyzed is structured in a way that allows addressing goals that are common to HMA programs. For example, the objective of “reducing the direct threat that Explosive Remnants of War [ERW] poses to a community” can be measured with any number of criteria. The user must have data for both the sub-objective and the criteria and a firm understanding of how the criteria and sub-objective are related.

For illustration purpose, we can take an example as follows [see Figure 2 & 3]. Let’s assume that we only have three criteria on “sub-objective 1”

 
 
SMCE Decision Tree
Figure 2: SMCA Decision Tree
 
  1. Distance of hazard from community 
  2. Number of people living in nearest community 
  3. Size of hazard area

Under this assumption, besides hazard data of course, we need to have data on community. To be more specific, we need: 

  1. The boundaries of settlements for the first criterion
  2. The number of people in each community for second criterion
  3. The boundaries of the hazard areas for the third criterion 
 
Close-up of SMCE Tree

Figure 3: Close-up of SMCA Tree

 
Figure 2 gives a snapshot of how the problem can be structured in SMCA. There is a main goal or objective and a number of sub-objectives, in this case four. In order to quantify the role of each criterion within the sub-objectives, a score and weight for each must be identified and defined. Based on how each criterion in the problem tree is defined and structured that is how it will affect clearance priority. Though, not all criteria are measured in same unit, like distance, area, blockages, victims or money, it is required to transform their values into comparable units. This is done by standardizing each criterions value from the raw into the range of between 0 and 1. The SMCA tool is equipped with a number of functions which will help on this process.

However, not all criteria within a sub-objective are equally important, and each are assigned a weight accordingly, for example by raking the above-mentioned criteria in order from highest to lowest importance. Say, distance from hazard is most important criterion, thus we list it on the top, and then the number of affected people, and at the bottom would be size of hazard. There are cases where one or more criteria may have equal importance, SMCA allows for this possibility as well. This method is called Ranking Order, and there are other methods such as Pair-Wise Comparison and Direct Input. The total weight of all criteria within the sub-objective must equal to the 100%.

 
Process of SMCE Analysis
Figure 4: Process of SMCA Analysis
 
The next step in the process requires that we next look at weighting of the four sub-objectives In a similar manner to the criterion in each sub-objective may or may not be of equal importance within our problem tree, and their weights will be defined accordingly. Throughout the process of weighting criterion and sub-objectives, each MAC department, and other stakeholders within the HMA community will provide documented input, delivering transparency and a sense of ownership and responsibility toward the outputs.

The analysis and the outcomes from SMCA are heavily dependent upon spatial data, providing greater benefits than tabular data analysis alone.

Hazard Area with four different scores [48, 44, 40 and 36]

Figure 5: Hazard Area with four different scores [48, 44, 40 and 36] 
 
In the SMCA spatial analysis environment, the hazard area is not necessarily treated as a single feature, but may be divided into a collection of cells [pixels] containing common scores. Based on user input of standardizing and allocating weights to the criteria and the weighting of sub-objectives, SMCA calculates the composite index for each cell. This may result with hazard areas that have more than one score, meaning different part of hazard area have different score of priority for clearance. This phenomenon is due to criterions that are of continuous spatial data format. The distance layers, elevation model, slope and alike, where each cell has different value, hence different composite index. The size of hazard area also may affect the score. In Figure 5, we show a large hazard that is broken into several portions. Portions of the hazard that are closer to community get a higher score. As the distance increases, the portions that are farther away from the community get a lower score. Unfortunately, in the real world often we face lack of data for a number of different factors. Decision makers need to plan with a degree of certainty and understand how other factors may impact the operational plan. To address these no-data issues and factors in SMCA, we simulate different scenarios and test the overall calculated score [composite index] through a process called sensitivity analysis [see Figure 4]. Sensitivity analysis is used to see how vulnerable the results are to changes small changes in weights.
 
Sensitivity Analysis – Alternatives 
Figure 6: Sensitivity Analysis – Alternatives
 
One way of executing this is to change the weights of sub-objectives [see Figure 6]. Naturally, if we change something in the model, then the result should change as well, however the change of output values should be comparable to the changes we do in input values. If this is not the case, and small changes in weight produce large changes in the outputs then our model of the problem tree needs to be revisited. The same is true in the opposite case, if changes in the weights produce no changes in the outputs then the factors that were changed are irrelevant to goal trying to 
be achieved.

This process provides additional feedback to the decision makers on the expectation and performance, and to foresee the potential problems so that they can 
be addressed in shortest time possible.

Conclusion
IMSMA is a superior tool for HMA information management, and is nearly unique across the relief and development continuum. As such, it is a rare example of an entire sector adhering to the same basic set of information management business processes and structures. However as we stated earlier there are no super tools available that can be all things to all people. The HMA community will continue to need specialized DSS tools to facilitate the complex process of planning and prioritizing to make the best the use of the available human, technical and financial resources. The integration of SMCA into information management processes improves the utility of IMSMA and improves the quality of information available to decision makers, both internally and externally.

Comments welcome, of course.

 
About these ads

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s