Climate Change and Conflict in East Africa: Mediation, Land, and Water in the Karamoja-Turkana Corridor
When the Land Runs Dry, Who Decides Who Stays?
A new Mercy Corps report from Uganda and Kenya's most climate-stressed borderlands argues that how communities negotiate over water and pasture may matter more than the rain itself.
In the borderlands where northeastern Uganda meets northwestern Kenya, the rains no longer arrive the way they used to. And when they do come, people fight over what's left.
The Karamoja–Turkana corridor has long been one of Africa's hardest places to govern — a semi-arid stretch of rangelands and dry riverbeds where pastoralist communities have moved with their cattle for generations, following seasonal water and pasture across a border that colonial mapmakers drew through the middle of a single ecological zone. Now, a new learning report from Mercy Corps' Climate Change Leaders Advancing for Peace (CCLAP) program is adding empirical weight to what field practitioners have long suspected: climate change isn't just stressing livelihoods here. It is restructuring the conditions under which violence either happens or doesn't.
The program, funded by the Austrian Development Agency and implemented across Uganda's Kotido and Kaabong districts and Kenya's Turkana West, is one of the most carefully documented efforts to integrate climate adaptation and conflict prevention into a single programmatic framework. Its findings have implications well beyond this corridor.
Climate Change Doesn't Pull the Trigger — But It Loads the Gun
One of the clearest contributions of the CCLAP report is its precision about causality. A persistent temptation in policy discourse is to draw a straight line from drought to violence — to say that climate change causes conflict. The evidence doesn't support that, and CCLAP is careful to say so. What climate change does, the report argues, is stress the systems through which communities normally manage competition: governance structures, trust between groups, access to shared resources like water points and grazing corridors.
It's not the heat that starts the fight. It's the breakdown of the arrangements that kept the peace when resources were manageable.
In Turkana West, the drought season — historically spanning January to May — now extends through December, sharply reducing livestock productivity and pushing communities into competition over water points and pastureland that previously sustained multiple groups through careful seasonal movement. The CCLAP assessments documented this shift systematically across both sides of the border, confirming that the timing and intensity of climate disruption is now outpacing traditional adaptive strategies.
Mediation as Infrastructure, Not Intervention
The program's most practical innovation is treating community mediation not as an emergency response to violence, but as standing governance infrastructure — the equivalent of roads or water systems, something that needs to be built and maintained before the crisis, not deployed during it.
CCLAP facilitated Natural Resource Sharing Agreements (NRSAs) between communities across the Uganda-Kenya border, including a cross-border forum that formalized access arrangements for the Turkana-Karamoja grazing corridors. Critically, these agreements went beyond just specifying who can graze where. They also ensured access to shared social services — health centers and markets — that lie across what are otherwise hard ethnic and administrative boundaries.
"Community-led mediation — not guns and rifles — became the mechanism for resolving livestock disputes."
— Finding from Mercy Corps' EKISIL program, the forerunner to CCLAP
The report is also honest about where these agreements hit their limits. The same cross-border forum that resolved corridor access did not produce a settlement for a more localized conflict between the Ik and Turkana communities in the Nawontos green belt. The lesson: regional-scale agreements need to be nested within sub-catchment and community-specific processes that can address particular grievances that corridor deals leave untouched. Governance architecture needs to match the geography of the conflict, not just the geography of the program.
The Inclusion Problem Is a Function Problem, Not a Values Problem
The report makes a pointed argument about inclusion that goes beyond standard programmatic language. It is not enough, CCLAP argues, to ensure that women and youth are "represented" in community governance and mediation processes. Without decision-making power — not just a seat at the table — these groups cannot contribute what they actually know, and the agreements produced tend to reflect the priorities of whoever already holds power.
This is a functional argument, not merely an ethical one. Women in these communities are frequently the primary managers of water and firewood collection, traveling longer distances as resources contract. They have direct, granular knowledge of where resource pressure is intensifying and where early warning signs are appearing. Excluding that knowledge from governance makes the governance worse.
The gendered dimensions of climate stress documented by CCLAP are severe. Resource scarcity correlates with increased domestic violence as men facing livelihood pressure displace frustration onto households. Girls face elevated school dropout rates when family survival takes priority over education during climate shocks. Women exposed to longer-distance resource collection face heightened risk of gender-based violence. These are not separate issues from the conflict prevention agenda — they are the same issue.
Why Standard Metrics Miss What's Actually Working
Perhaps the most important methodological contribution in the CCLAP findings concerns how we measure whether this kind of programming works. The dominant evaluation logic for peacebuilding programs counts incidents of violence. The assumption is that if a program is succeeding, the numbers should drop.
CCLAP's evidence suggests this is a lagging indicator that misses what happens first. Before violent incidents decline, something else changes: the quality of governance improves, trust between communities increases, local dispute resolution capacity strengthens, and agreements become more durable. These are the mechanisms through which violence is eventually prevented — and they are invisible to incident-counting frameworks during the window in which most donor evaluations are conducted.
For funders, this has real implications. Conventional Results Frameworks that gate continued investment on incident reduction within 24-to-36-month project cycles may systematically defund exactly the interventions that are working. The field needs better ways to measure governance quality, institutional trust, and mediation capacity as leading indicators — not consolation prizes for programs that haven't yet produced the violence-reduction numbers donors are looking for.
How This Fits the Wider Picture
CCLAP's approach aligns with guidance from the UN Department of Political and Peacebuilding Affairs (DPPA), which has made climate-informed mediation a formal priority. The UN framework specifically advocates for "climate-proofing" peace agreements — designing them to remain durable as environmental conditions worsen, not just as they stand today. The Nawontos case above illustrates exactly why this matters: an agreement calibrated to current resource availability may not hold as precipitation patterns continue to shift.
There is also a broader field alignment with what environmental peacebuilding researchers call the cooperation hypothesis — the observation that shared environmental stress doesn't inevitably produce conflict. Under the right institutional conditions, shared scarcity creates a powerful incentive for dialogue, collective management, and the rebuilding of cross-community trust. The CCLAP program is, in effect, a sustained attempt to build those institutional conditions before the scarcity tips into violence.
What's Still Unresolved
The report doesn't resolve everything, and it's worth naming the gaps. The weapons overhang is significant — an influx of small arms from South Sudan has reduced the capacity of civilian governance mechanisms to hold, even when communities want them to. Cross-border mediation agreements can be ignored or undermined by armed actors. CCLAP's current design doesn't have a documented answer to this.
Oil exploration in Turkana County adds a new extractive-economy layer that pastoral resource-sharing frameworks were not designed to handle. And like most donor-funded programs of this type, CCLAP faces the sustainability question: what happens when the Austrian Development Agency funding ends? The logic of the program requires these governance mechanisms to be institutionalized within government planning frameworks — Uganda's national development plans, Kenya's County Integrated Development Plans — to survive the project cycle. The path to that institutionalization is not yet clear.
