Article Category: Original study
Data publikacji: 01 paź 2021
Zakres stron: -
Otrzymano: 01 lut 2018
Przyjęty: 10 lip 2021
DOI: https://doi.org/10.2478/jms-2021-0011
Słowa kluczowe
© 2021 Jeffrey C. Cegan et al., published by Sciendo
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Base camps are vital components of overseas military missions and must provide four key functions: (1) force protection; (2) critical infrastructure (e.g. housing, maintenance areas, dining facilities, aid stations, chapels, postal service centre, electric power, water treatment, fuel storage, road networks and athletic fields); (3) training support; and (4) maintenance support (Ezell et al. 2001). Providing these functions in a dynamic environment requires complex planning, development and evaluation of site suitability by a team of subject-matter experts (e.g. a Location Selection Team or Forward Engineering Support Team (FEST)). Although flexible, these teams are usually led by one officer-in-charge, and are comprised of one non-commissioned officer and six highly skilled U.S. Army Corps of Engineers (USACE) civilian technical experts, who work together to determine base camp locations within an area of interest (AoI) (USACE 2009). Each alternative site’s suitability is evaluated based on a variety of factors that can reduce risk and maximise mission effectiveness, and can be used in conjunction with Geographical Information Systems (GIS) data, software and/or capabilities. It is important to include a holistic evaluation of all relevant factors when assessing relative site suitability to address the U.S. Department of Defense (DOD) needs with the resources available. A formal decision-support tool could provide a useful framework for systematically comparing alternatives and ensuring inclusion of criteria that may otherwise be unintentionally overlooked or undervalued during the base camp selection process. This provides decision support for both planning and in-field purposes, as well as a platform for incorporating new information and lessons learned.
Geographic locations for base camp construction are prioritised as described by existing military doctrine (Ezell et al. 2001; Lindberg and Vargesko 2007; Lockridge 2017). The DOD has established guidelines for standardised base camp construction, operation, maintenance and security through such references as the ‘RedBook,’ ‘SandBook’ and Assistive Technology Planner (Lindberg and Vargesko 2007). While these construction guidelines discuss the broad intent of the doctrine, they leave many details to be interpreted without establishing criteria known to minimise mission risks beyond initial construction (Ezell et al. 2001; Parry and Kwarta 2001; Cegan et al. 2017a). This means that battlespace commanders responsible for determining site selection must interpret their site-specific research without specific decision guidance from military doctrine and without an established set of base camp suitability criteria. These limitations could potentially decrease the base-camp resiliency, introducing vulnerability for soldiers and ultimately compromising mission success rates (Parry and Kwarta 2001; Lockridge 2017).
There are several historical examples of base camp failures that highlight the need for a holistic and structured decision framework. For these sites, a formal framework could have influenced the base camp’s initial location and allowed soldiers and commanding officers to have a greater awareness of the risk (or benefits) posed by the camp’s location. Flawed siting locations over the past 15 years have resulted in base camp evacuations due to the detection of chemical, biohazard, nuclear and radiological hazards such as nerve and mustard gas and heavy metal contaminants (Williams 2002). They have also resulted in temporary closures due to unforeseen rain patterns and landslides and other large-scale environmental risks (Fisher 2012). On more than one occurrence, civilian engagement, threat levels, cultural norms, environment, enemy vantage points and topography constraints culminated in base camp closures, increased combat and international outrage (Ellinger et al. 2015; CNN 2009). While these closures and delays exemplify a base camp’s effect on soldier safety, site suitability also impacts military spending during logistics operations such as supply and resupply (Marlott 2003). A tool would help reduce common risks: unnecessary vulnerability to dissatisfied locals, environmental risks like water supply contamination or slope destabilization, over-reliance on single supply routes that could lead to isolation and so on.
In order to assess base camp life cycle risks and maximise mission effectiveness and soldier safety, site suitability criteria must be compared across possible location alternatives. Multicriteria decision analysis (MCDA) provides an efficient methodology to assist siting teams wishing to strategically optimise their base camp location, weighting information in a holistic, balanced, quantitative and geographic approach. MCDA is a branch of operations research and management science that aids alternative selection in complex decision problems by providing a structured framework of criteria, value functions and weights to systematically aggregate individual details in a composite comparison of relative suitability. MCDA frameworks are powerful because they establish a common understanding among stakeholders and guide the process of incorporating subjective decision-maker tradeoffs into a transparent and defensible decision model (Keeney and Raiffa 1976; Linkov and Moberg 2011). Complex decision environments often require consideration of multivariate, conflicting spatial criteria to prioritise a set of alternatives (Malczewski 2006). MCDA tools can be introduced to help solve complex problems in a way that is practical and technically reliable and that can incorporate research, focus groups and surveys for criteria valuation along with spatial mapping or GIS-based considerations to include expanded neighbouring environments in the analysis as well (Huang et al. 2011). Although the military has developed some tools for supporting site selection (e.g. Geoblast, which considers a select set of mission parameters, and the ArmyBaseCamp/JFOB.net knowledge management system, which acts as a repository for documents and best practices), these are discrete tools lacking the ability to integrate cultural data, geographical data and engineering data into one decision-analysis framework (Ezell et al. 2001; Trainor et al. 2008).
Screening a large geographic area to find the most suitable location for a base camp requires consideration of various tradeoffs. MCDA can assist by explicitly capturing various insights and expertise of personnel involved in site selection in order to facilitate discussion and provide a transparent and cohesive process. By incorporating a GIS-based MCDA framework when locating military base camps, commanders will have accessible interfaces to help evaluate vulnerabilities, characterise and integrate multidimensional impacts and improve decision-making to mitigate spatially distributed risks.
Past applications of MCDA frameworks include land use development, transportation studies, site suitability for infrastructure (Hill et al. 2005; Hamilton et al. 2016; Yatsalo et al. 2010), nanotechnology management (Linkov et al. 2013; Bates et al. 2015; Bates et al. 2016), environmental risk management (Yatsalo et al. 2011), and humanitarian assistance and relief (Curran et al. 2014) with growing acceptance and application over time (Cegan et al. 2017b; Kurth et al. 2017). Given the success of these applications in informing site suitability decisions based on divergent criteria, we seek to apply similar MCDA methods with a new set of decision criteria related to increasing soldier safety and mission effectiveness through military base camp siting. To do this, we integrate expert elicitation and military doctrine regarding base camp life-cycles into a set of 10 criteria that can be used across a generic mission type to access and select alternative sites. We propose default weighting for these criteria from surveys with additional military commanders. However, this method must be tailored through scenario-specific weights and value functions. Inclusion of specific GIS tools, as appropriate based on uncertainty present in the area, can also be applied (Malczewski 2006). Finally, we apply these weights in an illustrative example and discuss the future work that would need to be done in order for the military to adopt this type of framework. This research contributes to the ongoing development of a new tool that will aid military teams in efficiently siting base camps and also allow for in-field decision support through the establishment of criteria and weights that are comprehensive, well-defined and easy to reference in situations calling for quick, unbiased and transparent decision-making.
During the criteria development process, the authors consulted with 14 current and former army officers (including 1 Colonel, 1 Lieutenant Colonel and 12 Captains) with experience in base camp siting and operations to discuss factors important for identifying promising sites for future base camps (Cegan et al. 2017a). While these conversations were naturally influenced by the US Army’s more recent experience in the Middle East and the Balkans than past conflicts in South-East Asia, Europe, North Africa, or elsewhere, the interviews were steered to cover a variety of terrain, climatic and engagement conditions broadly relevant for expected future operations. These elicitations parallel the deep army history of After Action Reviews (AARs) by documenting lessons learned to identify areas of concern and develop criteria to focus on for improving site suitability assessment.
In addition to the officer interviews, relevant army literature on base camp siting, including descriptions of current doctrine and suggestions for future improvement, was reviewed to develop our list of key criteria. While often not as specific or prospective as the firsthand descriptions from the interviewees, over a dozen publically available military documents provided useful context and background knowledge related to base-camp siting. A list of these documents is included in Appendix A.
The officer interviews and army literature reviews resulted in identification of an initial list of potential factors relevant for base camp siting. However, benefits of a limited yet thorough development of criteria have been shown to improve both the transparency of the results and the consistency of weighting (Weber et al. 1988; Marttunen et al. 2017), and combined with the need for the capability for quick (in-field) decision-making, a large list of criteria is not ideal. Therefore, this initial list was then pared down to criteria judged to have been emphasised most consistently across the sources consulted, resulting in a total of 10 criteria. For these final criteria, summary definitions were developed that are clear and succinct, meeting the exhaustive, consistent, non-redundant, comprehensible and measurable requirements (Adiat et al. 2012) for use in MCDA analysis. Importantly, many of these criteria consider the holistic nature of the environment, drawing not only on site-specific (e.g. bounded) but also larger and broader environment factors, such as surrounding land use and access to the site, which incorporate the vital consideration that a basecamp is not independent of its larger setting and neighbours.
After the criteria were finalised, in-person surveys were conducted to develop default weights, which are used in an illustrative example in Section 3.3, which may provide a starting point for further scenario customisation in the decision-support tool, and can likewise be incorporated with larger spatial analysis programmes given available data and time through GIS software for further understanding of more specific neighbouring or direct threat. Surveys were conducted with 14 different officers attending the 85th Military Operations Research Symposium at West Point, NY on June 19–22, 2017. Due to the limited time available for in-person discussion and elicitation and the illustrative nature of the default weights, a simple and readily understandable approach of direct ranking and weighting was selected over alternate weight-elicitation techniques like swing weight elicitation (Winterfeldt and Edwards 1986). The officers were first asked to rank the criteria in importance from 1 to 10 and then use a four-star rating system for identifying relative criteria importance. The ranking exercise helped participants organise their broad thoughts before giving each criteria a specific rating of importance. Four stars were used for the rating system because research on surveys for valuation show that choice consistency is maximised when approximately four alternatives are given (Cook et al. 2007). The completed surveys were used to calculate the default weights for the MCDA model by averaging the star ratings across survey respondents normalised by the total number of stars indicated to derive percent weight.
The MCDA framework for base camp site suitability follows the method of Dodgson et al. (2000) and applies the following equation to the developed criteria and weights, where
Sensitivity analysis was conducted to further examine the impact of criteria weighting on final alternative scores. The sensitivity analysis compares the default weights to equal weighting, rank-reciprocal weighting with the officer-supplied tradeoffs, and an ‘opposite weight’ scenario where the officers’ rank order of criteria importance is entirely reversed prior to calculating rank-reciprocal weights (i.e. the most important criterion becomes the least important, the second most important criterion becomes the second least important etc.). This extreme scenario helps to show how much of the total score depends on the details of the officer’s subjective weighting versus overwhelming differences in the site data. The results of this sensitivity analysis are shown in Appendix B.
Rank-reciprocal weights are calculated by normalising the reciprocals of each criterion’s rank order of weight importance (recall that the officers were asked both to rank the criteria in terms of their importance for site assessment and then to directly provide numerical weights using a scoring system). The rank-reciprocal weight,
The most important criteria for site selection are drawn from the expert interviews and review of relevant army literature. These are divided between criteria related to social and demographic data, special analysis of military communications capabilities available between that site and other existing facilities, civil infrastructure and terrain and environmental data (Table 1). Each of these types of information will be supported by different spatial data sources.
Key criteria for base camp siting, categorised by data type
| Population | Social/demographic |
| Enemy/threat | |
| Interference/signal | Capability analysis |
| Roads | Civil infrastructure |
| Local utilities | |
| Slope | Terrain/environment |
| Soils | |
| Land cover | |
| Aquifer access | |
| Land use |
These criteria are relevant for screening-level assessments of potential alternatives within an AoI, leaving many case-specific considerations to be further evaluated by the siting team outside of the MCDA framework before final site selection. The AoI includes both the potential site and the potential immediate surroundings due to its intentional scope. Smaller terrain features such as ditches, hilltops and gullies will also need to be considered for specific facility placement within any chosen site. Note that a score of zero indicates poor performance in that area rather than a ‘no build zone’ (which is specially designated by the military outside of this process); any fully unfeasible alternatives should be screened out before starting the MCDA analysis.
The criteria are defined as follows:
Fourteen officers across the army were surveyed to determine default criteria weights (Table 2). The results, ranging from 7.5% to 12.6%, show no clear categorical trend related to the data types outlined in Table 1.
Average ranking and importance order comparison and calculated weighting
| Roads | 12.6 |
| Threat/enemy | 12.0 |
| Slope | 11.7 |
| Soils | 11.5 |
| Population | 10.1 |
| Local utilities | 9.8 |
| Land cover | 8.9 |
| Aquifer access | 8.1 |
| Interference/signal | 7.8 |
| Land use | 7.5 |
To demonstrate the MCDA site selection framework, seven hypothetical site alternatives within an AoI are evaluated in an illustrative example. Here, normalised value scores (
Alternative regions for suitable base camp sites based on criteria scores
| Site 1 | 0.6 | 0.8 | 0.2 | 0.8 | 0.7 | 0.3 | 0.7 | 0.3 | 0.8 | 0.5 |
| Site 2 | 0.9 | 0.2 | 0.0 | 0.3 | 0.5 | 0.2 | 0.3 | 0.9 | 0.4 | 0.3 |
| Site 3 | 0.9 | 0.8 | 0.3 | 0.7 | 0.8 | 0.9 | 0.3 | 0.5 | 0.0 | 0.3 |
| Site 4 | 0.7 | 0.3 | 0.1 | 0.9 | 0.3 | 0.8 | 0.6 | 0.4 | 0.3 | 0.8 |
| Site 5 | 0.9 | 0.8 | 0.6 | 0.9 | 0.7 | 0.7 | 0.7 | 0.5 | 0.7 | 0.9 |
| Site 6 | 0.4 | 0.1 | 0.5 | 0.1 | 0.4 | 0.2 | 0.8 | 0.0 | 0.3 | 0.2 |
| Site 7 | 0.9 | 0.9 | 0.9 | 0.6 | 0.6 | 0.7 | 0.8 | 0.9 | 0.9 | 0.5 |
Table 4 shows the sum product of the alternative scores per criterion (Table 3) and the criterion weights (Table 2). The resulting total MCDA score (
Ranking of alternative scores
| Site 7 | 0.78 |
| Site 5 | 0.75 |
| Site 3 | 0.58 |
| Site 1 | 0.57 |
| Site 4 | 0.52 |
| Site 2 | 0.40 |
| Site 6 | 0.30 |
In addition to total score, the MCDA results allow one to drill-down and see why each alternative received its corresponding total score (Figure 1). That Site 7 ranked first overall is not surprising since it also has the highest values for the top three weighted criteria: roads, threat/enemy and slope. Site 2 scores as high as Site 7 in suitability with respect to roads, but poor performance in other areas still leads to a low total score. Several sites outperform Site 7 in suitability with respect to population, but fail to generate superior scores in enough other areas to rank first. Site 5 ranks second after Site 7, with a somewhat different score profile, and might be preferable overall for base camp construction depending on the more detailed analyses performed next by the site selection team. These top two sites would be the recommended for further consideration and in-depth analysis.
Fig. 1
MCDA factor distribution by alternative site. MCDA, multicriteria decision analysis.
This breakdown is also significant in showing that in-field decisions that must be made quickly (perhaps for temporary basecamp siting) can prioritise the more heavily weighted criteria in decision-making. Although in an ideal situation all criteria should be leveraged for a final score, given limited or uncertain information and time, the higher weighted criteria can be used in the field, especially given the greater likelihood of physical visibility (roads and slope) or knowledge (threat/enemy) that military personnel would have on hand.
To represent these findings in a manner that can be used by a potential user interface in a selection tool, we employed a colour-coding system (Figure 2). Red shaded regions represent site locations having the lowest level of suitability (≤0.50), yellow shaded areas represent alternative regions that pose a medium suitability (0.50−0.75), and green shaded areas represent regions believed to be best suited (≥0.75) for a base camp in the MCDA screening analysis. While the diagram of results for the illustrative example has each alternative with an equal area, alternatives will likely vary in size in real settings. This clear visualisation enables more precise in-field selection of base camps by giving clear geographically ranked results.
Fig. 2
Colour-coded ranking of alternative sites within an AoI. AoI, area of interest.
A sensitivity analysis compared the final rank order of alternative scores using the calculated weights with other potential interpretations using equal weights, rank-reciprocal weights and rank-inverse weights. The results (Tables B1 and B2 in Appendix B) show that the final rank order of alternative sites does not vary between equal weighting and rank-reciprocal weighting and varies only slightly between default weighting and equal weighting. In the opposite-weight scenario, the results are still largely similar, with two alternatives not changing final rank order, four changing rank order by one place, and one changing rank order by two places.
The results from expert elicitation show that, at default, the three most heavily weighted criteria are roads, threat/enemy and slope, with weights of 12.6%, 12.0% and 11.7% respectively. It makes sense that the roads criterion, which relates to vital factors such as safety, logistics/life cycle and constructability, is heavily weighted. As soldier safety and mission effectiveness are of upmost importance and hinge on threat/enemy, this criterion is also valued highly. Slope is at the intersection of logistics/life cycle and safety, because it is a consideration for friendly air resupply and hostile ambush scenarios, among other factors. Decision-makers in the field can leverage prior knowledge of the prioritisation/ranking of the criteria to best manage resources and time as necessary or under uncertainty.
The least valued criteria, interference/signal and land use, are not as cross-cutting across base camp requirements such as constructability, logistics/life cycle and soldier safety. Although further elicitations must be conducted to fine-tune the weights for specific implementations, these results can serve as default guidelines for a base camp site suitability in a decision-support tool. A visual colour-coding ranked output within an AoI gives a valuable starting point for basecamp exploration, and such a portable tool could be used in the field to fine-tune the results as additional data becomes available.
The illustrative example in this paper explores site locations for a hypothetical base camp tasked with a broad mission. For base camps established for specific purposes (e.g. humanitarian assistance, COIN operations, enemy engagement etc.), tailored weights need to be applied that best fit the needs of the mission. Optionally, weight templates that provide a range of more specific default weights appropriate for a handful of common scenarios can anticipate these different needs and provide a non-unique but tailored starting point for case-specific use.
Due to the complexity and case-specific nature of multifaceted base camp objectives, defining the value functions used to derive criteria values (
Existing military doctrine and literature show that there are many factors to consider in base camp site selection, and that taking a comprehensive approach to quantitatively organising and streamlining criteria valuation is necessary. MCDA can streamline the base-camp selection process. Through careful development of the 10 top criteria and expert elicitation across a range of military personnel with base camp experience, criteria and default weights were generated that can broadly aid the process. With case-specific value functions, these will provide an explicit framework for determining site suitability of military base camps. Once vetted by the military approval process, these results are intended to be incorporated into the Engineer Site Identification for the Tactical Environment (ENSITE) tool for determining suitability of potential base camp sites, currently under development by the USACE Environmental Research and Development Center (ERDC).
This paper highlights current work on an MCDA military base camp site suitability process and tool. Further work remains to be done on developing the value functions and perhaps developing more specific weighing templates for some common mission and environmental scenarios. Value functions are important because they normalise raw data in a way that allows for otherwise incomparable information to be transparently included in a comprehensive evaluation. Value function identification is necessary for any MCDA framework to be robust (Malczewski and Rinner 2015).
Due to economic and time constraints, many uncertainties will remain in most pragmatic MCDA evaluations. One approach for prioritising further information gathering to reduce uncertainty is Value of Information (VoI) analysis. Although all the criteria are vital when evaluating the suitability of a base camp, some criteria may present a larger need for measurability over others (e.g. interference/signal may not be mapped by the military data sets across a geographic AoI). In the context of minimising risk of mission failure, some uncertainties naturally matter more than others. Similarly, investments in information gathering to resolve uncertainties (e.g. investments to improve road or aquifer datasets) can differ widely in costs. VoI analysis would combine these factors to prioritise resolving the most valuable uncertainties, and only those that are likely to lead to substantially different outcomes in terms of recommended base camp alternatives (Keisler et al. 2014). Future work could also incorporate VoI analysis into base camp site-suitability assessments.