The Geography of Armed Civil Conflict

Led by Halvard Buhaug
Jan 2002 - Dec 2005

Main theme

The purpose of my doctoral project is to improve the study of civil war by including certain geographic and military aspects in a systematic, quantitative fashion. I aim to answer questions like these: To what extent are geographic factors like topography, natural resources, climate and conflict location key determinants of course and outcome of internal conflicts? How do geographic elements interact with military attributes of the warring parties?


The end of the Cold War removed the most important source of interstate rivalry and the attention of peace research has increasingly turned to civil conflicts, often of complicated nature, involving multiple parties. The works by Collier (1999) and Collier & Hoeffler (1998, 2001) are among the leading studies on causes and consequences of civil war. Influenced by economic theories and terminology, Collier & Hoeffler distinguish between ‘greed’- and ‘grievance’-motivated rebellion. The goal of the first group is mainly individual economic gain by controlling and exploiting valuable regions, typically producing primary commodities (diamonds in Angola and Sierra Leone, cocaine in Colombia). On the other hand, grievance-motivated rebellions are triggered by ethnic or religious hatreds, inequality, political exclusion or historical vengeance, and the ultimate goal here is either secession or improved civil conditions (Bosnia, Chechnya). Economic status and distribution, political system and representation, ethnic and religious fractionalisation, and population distribution are central exogenous variables in these studies.

Most empirical studies within the quantitative conflict research tradition have failed to acknowledge the relevance of military and geographic attributes of conflicts. How different combatant actors are trained and equipped, and how the terrain facilitate and influence the course of (civil) war, as well as the outcome, are questions that - to my knowledge - never have been addressed in a proper, systematic manner. And until most recently, the rather few studies that actually do discuss geography explicitly, do so within the limited context of geographic proximity and distance.

Geography and Conflict

Knowledge about the potential impact of geography on conflicts is probably as old as the art of war. And geographers, as well as theorists of international relations have for long claimed climate, topography and location to be important determinants of state behaviour (Sprout 1963). The empirical conflict studies’ definition of geography has been less inclusive. In an assessment of previous work on geography and war, Diehl (1991) discusses three theoretical frameworks of particular influence: Sprout & Sprout (1965) with the notion of ‘environmental possibilism’, Boulding (1962) with the loss-of-strength gradient and Starr (1978) with the concepts of opportunity and willingness. All three emphasise physical distance as the crucial geographic factor affecting the risk of conflict.

The theoretical link between geographic proximity and conflict has resulted in the development of the Correlates of War project’s contiguity dataset (Gochman 1991) as well as Gleditsch & Ward’s (2001) minimum distance dataset. Bremer's (1992) study, finding geographic proximity to both facilitate conflict and being a source of conflict, is a well-regarded empirical study, applying the geographic term in this manner. But while the pioneering works by the Sprouts, Boulding and Starr still are considered highly influential - and later empirical studies have indeed revealed some convincing findings - they interpret geography merely as a concept of contiguity and distance. Accordingly, the geography concept acts more as a proxy of interstate interaction opportunities than measuring the impact of physical, geographic attributes of conflicting countries. Moreover, international borders and interstate distance as analytical concepts are mainly relevant when dealing with international conflicts.

To the extent that geography and conflict have been linked in other and more fashionable ways, they have been subject to either of two approaches. The first one deals with micro-level analyses of battlefield effectiveness, typically from a military point of view. In this respect, issues like weapon and soldier performance in varying topographic and climatic conditions, and how to exploit geographic advantages (ranging from hills and weather to tidal water) are central concerns, sometimes illustrated by certain well-selected historical battles (see Collins 1998). On the other hand, we have system-level discussions of geopolitics and structures of the (post-)Cold War. Here, spheres of ideological influence and strategies of nuclear deterrence are central themes (cf. Pepper & Jenkins (1985) for a discussion of Cold-War geopolitics). Neither of these approaches is suitable for a cross-national study for the entire post-WW II period; the first is inappropriately detailed, the latter allows little variation between cases.

But perhaps we are witnessing a renewed attention to the influence of geographic attributes of conflict zones. In a theoretical paper on the microfoundations of rebellion, Gates (2001) identifies three factors determining military success and shaping rebel recruitment: geography, ideology and ethnicity. His central themes are how geography interacts with ethnicity and ideology, and how this interaction affects the distance between rebels and government, and how distance enables a rebel group to expand. Among the first empirical studies to include variables on terrain, Fearon & Laitin (1999) construct a rough dummy variable on mountains for the Eastern European countries. Collier & Hoeffler (2001) include indicators of both forests and mountainous terrain as causal variables explaining civil wars - their geographic variables measures proportion of total land area covered by forest and mountain, respectively. Collier & Hoeffler's findings weakly support the expected positive mountain-war correlation, whereas forests fail to generate significant effects. Herbst (2000) and Collier & Hoeffler also control for geographic dispersion of the population within each country.

Finally, a couple of ongoing projects accentuate the need for more precise data on conflict location. Braithwaite (2001), studying temporal and spatial conflict clustering, improves the MID data by coding latitude and longitude co-ordinates for the centre of every dispute, and performs a visual assessment through the aid of GIS. And testing the theory of Gates (2001), the ongoing work by Buhaug & Gates (2001) on the geography of armed civil conflicts includes variables on location and scope of battle-zones, distance from the conflict zone to the capital, and dummy variables on the presence of valuable natural resources.

As for the military element of conflicts, not a single empirical, cross-national analysis has been conducted to my knowledge. This is probably much due to the problem of gathering precise, reliable data on military strength of the combatants - a task that is particularly difficult regarding multiparty conflicts in the distant past.


A quick review of the works on geography and conflict reveals that much is left undone. Few relevant studies that include geographic variables extend their operationalisation beyond measures of contiguity and borders. And hardly any work has been done on exploring the impact of military capability on the cause, course and outcome of civil or international wars. Therein lies a huge challenge.

This project will rely substantially on the ongoing work by Buhaug & Gates (2002) particularly when it comes to the geographic data sample. The new dataset generated by PRIO and the Department of Peace and Conflict Research at Uppsala University (Gleditsch et al. 2001) will serve as the conflict database, to which a number of geographic and military variables will be added (some of which have already been tested in Buhaug & Gates 2002). I believe the PRIO/Uppsala conflict dataset is better suited than the Correlates of War database as it has a considerably lower threshold for civil (and interstate) conflicts and therefore a larger set of conflicts.

Until now, there has not existed any precise data on the location of armed conflicts. Thus, any civil war has been defined as stretching throughout the entire country - a procedure with obvious flaws if we consider limited conflicts in rural areas like Chechnya, Kashmir or Tibet. But with quantitative data on more exact geographic location of the battle-zones, we will be able to perform considerable more comprehensive analyses of the pattern of civil wars. Controlling for total area of the conflicts, attributes of the corresponding terrain, and distance to the opposing forces and to administrative centres are only a few of the possibilities that the geographic data enable.

Geographic and military variables of particular relevance are

  • Conflict location and scope
  • Forest
  • Mountain
  • Rivers, lakes
  • Resources
  • Population dispersion and density
  • Size and military capability of main actors

In work performed previously under a cross-disciplinary project at NTNU and co-sponsored by PRIO’s civil war project, I have determined for each conflict a centre-point and a distance (radius) variable approximating the total extent of the conflict area. This dataset is used in Buhaug & Gates (2002). As few conflicts are likely to have a precisely circular shape, effort will be put into refining the conflict location variable, either by constructing conflict polygons made out of 4-6 specific points or by defining entire regions/administrative units as affected by conflict. The latter method is far more advanced (but not necessarily more representative for all conflicts) and requires the use of GIS tools.

The dataset on mountains and forests in Collier & Hoeffler (2001), giving proportional cover of each country, is usable, but a major improvement would be to have data on mountain and forest cover reflecting the terrain where the conflicts actually occurs (and relative to the 'mean' proportion of mountain/forest in the country). Whether gathering such data is feasible within the temporal frame of this project remains to be decided.

The variables on valuable natural resources in the Buhaug & Gates (2002) study need serious refinement. Resources also have a temporal dimension that must be included, as well as some measure/gradient of lootability. Le Billion's (2001) typology of location (proximate/distant) and concentration (point/diffuse) might be a good starting-point.

Gathering reliable data on military capability of the combatants is most likely an enormously challenging and time-consuming task, and time will show whether such data is available for any conflict bar the most recent ones.

I also plan to make use of the work by Kathryn Furlong and Nils Petter Gleditsch on measuring the length of national borders. Most studies agree that land contiguity affects the probability of interstate wars, though the effect varies substantially across different regime types (Buhaug 2001). A longer land boundary should have a greater potential for negative (as well as positive) interaction. The length of land boundaries may also affect the spread of internal conflicts to external parties, so this factor is also relevant for understanding the course and diffusion civil wars.


The empirical analyses in this project will be of mostly quantitative, cross-national character. But intrastate conflicts where international forces have intervened will be of particular interest, and additional attention will be given to such cases. Hopefully, the combination of geographic and military factors may produce plausible explanations as to why (and how) some UN operations are deemed successful when others fail.


Boulding, K., 1962. Conflict and Defense. New York: Harper & Row.

Braithwaite, A., 2001. The Geography of Militarized Interstate Dispute (MID) Onsets, 1816-1992, paper presented at the annual North American Meeting of the Peace Science Society, Atlanta, GA, October 26-28

Bremer, S.A., 1992. “Dangerous Dyads: Conditions Affecting the Likelihood of Interstate War, 1816-1965”, Journal of Conflict Resolution 36(2): 309-341.

Buhaug, H., 2001. “Assessing the Domestic-Foreign Conflict Nexus - The Discriminating Impact of Political Systems”, hovedoppgave (MA thesis) in Political Science, NTNU.

Buhaug, H. & S. Gates, 2002. ‘The Geography of Civil War’, in forthcoming edition of Journal of peace Research.

Collier, P., 1999. “On the Economic Consequences of Civil War”, Oxford Economic Papers 51: 168-183.

Collier, P. & A. Hoeffler, 1998. “On the Economic Causes of Civil War”, Oxford Economic Papers 50: 563-573.

- 2001. “Greed and Grievance in Civil War”, Working Paper, World Bank.

Collins, J.M., 1998. Military Geography for Professionals and the Public. Washington: Brassey’s.

Diehl, P.F., 1991. “Geography and War: A Review and Assessment of the Empirical Literature”, International Interactions 17(1): 11-27.

Fearon, J.D., 2001. Why Do Some Civil Wars Last so Much Longer than Others?, paper presented at the World Bank conference, Irvine, CA, May 18-20.

Fearon, J.D. & D.D. Laitin, 1999. Weak States, Rough Terrain, and Large-Scale Ethnic Violence Since 1945, paper presented at the annual meeting of the American Political Science Association, Atlanta, GA, September 2-5.

Gates, S. 2001. “Recruitment and Allegiance: The Microfoundations of Rebellion”, Journal of Peace Research, in press.

Gerrard, A.J.W., 2000. “What is a Mountain?”, Mimeo, DECRG, World Bank.

Gleditsch, K.S., & M.D. Ward, 2001. “Measuring Space: A Minimum Distance Database and Applications to International Studies”, Journal of Peace Research, in press.

Gledtisch, N.P., H. Strand, M. Eriksson, M. Sollenberg & P. Wallensteen, 2001. Armed Conflict 1946-2000: A New Dataset, paper presented at the PRIO/Uppsala/World Bank conference on ‘Identifying Wars: Systematic Conflict research and Its Utility in Conflict Resolution and Prevention’, Uppsala, June 8-9, under journal review.

Gochman, C.S., 1991. “Interstate Metrics: Conceptualizing, Operationalizing, and Measuring the Geographic Proximity of States since the Congress of Vienna, International Interactions 17(1): 93-112.

Herbst, J., 2000. States and Power in Africa. Princeton: Princeton University Press.

Kliot, N. & S. Waterman, ed. 1991. The Political Geography of Conflict and Peace. London: Belhaven.

Le Billion, P., 2001. “The Political Economy of War: Natural Resources and Armed Conflicts”, Political Geography 20: 561-584.

Murray, W., 1999. “Some Thoughts on War and Geography”, Journal of Strategic Studies 22(2-3): 201-217.

Pepper, D. & A. Jenkins, ed. 1985. The Geography of Peace and War. Oxford: Basil Blackwell.

Sambanis, N., 2000. “Partition as a Solution to Ethnic War”, World Politics 52(4): 437-483.

Small. M. & J.D. Singer, 1982. Resort to Arms: International and Civil Wars, 1816-1980. Beverly Hills, CA: Sage.

Starr, H., 1978. “Opportunity and Willingness as Ordering Concepts in the Study of War”, International Interactions 4(4): 363-387.

Sprout, H., 1963. “Geopolitical Hypotheses in Technological Perspective”, World Politics 15: 187-212.

Sprout, H. & M. Sprout, 1965. The Ecological Perspective on Human Affairs. Princeton: Princeton University Press.



Peer-reviewed Journal Article

Buhaug, Halvard; & Jan Ketil Rød (2006) Local Determinants of African Civil Wars, 1970-2001, Political Geography 25(3): 315–335.
Buhaug, Halvard; & Päivi Lujala (2005) Accounting for Scale: Measuring Geography in Quantitative Studies of Civil War, Political Geography 24(4): 399–418.
Buhaug, Halvard (2005) Dangerous Dyads Revisited: Democracies May Not Be that Peaceful after All, Conflict Management and Peace Science 22(2): 95–111.
Buhaug, Halvard & Scott Gates (2002) The Geography of Civil War, Journal of Peace Research 39(4): 417–433.

Conference Paper

Gleditsch, Kristian Skrede 2005 The Origin of Conflict Clusters: Contagion or Bad Neighborhoods?, presented at the ECPR general conference, , 8–10 September.