The maps above present forecasts at fine-grained sub-national geographical locations for each of the three outcomes, focusing primarily on where events happen. We refer to this level as pgm.
The densest risk clusters at pgm level for state-based conflict at s=3 continue to be found in north-eastern Nigeria, the Anglophone region of Cameroon, the Ituri and Kivu provinces of DRC, southern Somalia, the Niger delta and Sinai in Egypt, around Tripoli in Libya, the Cabo Delgado province of Mozambique, and in the border areas between central Mali, northern Burkina Faso, and south-western Niger.
Compared to our April forecasts at s=3, the most pronounced changes in the risk assessment at the PRIO-GRID level also mostly align with the high-risk clusters. We continue to see both local increases and decreases across Central Mali and northern Burkina Faso, illustrating the local dynamics of the conflicts.
Only increases are expected in western Niger, where the conflict between the government and IS on the one hand, and the government and Boko Haram on the other, continue to shed lives. Of particular note is also an attack on a military post in the Diffa region mid-March, after which the Nigerien government claimed to have killed 50 Boko Haram members (although the attack was later claimed by ISWAP). Its affect on the country's risk profile is illustrated by the bright red grid cells on the south-eastern border to Nigeria.
While north-eastern Nigeria continues to be plagued by both IS and Boko Haram, we see a geographic differentiation in the risk assessment for July 2020. We also see that the UCDP candidate data set has picked up new conflict events in eastern Niger state, where an ambush killed nearly 30 security officials on March 22nd. In Katsina state, in turn, the indicated conflict events concerned the killing of 24 bandits by troops attached to Operation Hadarin Daji a few days earlier. Part of the same operation, two bandits were also neutralised in Zamfara state on March 18th.
Compared to our forecasts of last month, we find grid cells with both increased and decreased risks this month also in the Anglophone region of Cameroon, and the Ituri and Kivu provinces of DRC, and in Somalia's southern-most regions (with decreasing trends notable in and around Kismayo and Mogadishu). The grid cells of increased risk in southern DRC come as a result of clashes between Kata Katanga militiamen and the army in three different towns on March 28th.
The forecasts for non-state conflict and one-sided violence depend on the same factors although with somewhat different implications. Nigeria remains a hotbed for non-state conflict, the highest risks of which continue to be found in the southern and central regions, underpinned by continued inter-communal violence of various forms, coupled with criminality and cultist violence. Other dense risk clusters include the Ituri and Kivu provinces of DRC. The Horn of Africa remains a larger general risk cluster, whereas Libya offers an interesting geographic distribution with higher risks predicted in individual and often separated PRIO-GRID cells.
Compared to our forecasts at s=3 last month, there are thus few notable changes in the risk assessment for non-state violence. Most prominent is a general decrease across southern Nigeria given the time lapsed since the last UCDP-coded conflict events there, albeit the risk levels remain relatively high.
For one-sided violence, Mali and Burkina Faso, north-eastern Nigeria, the Anglophone regions of Cameroon, and the Kivu provinces of the Democratic Republic of Congo continue to feature the strongest risk clusters. Compared to last month's forecasts at s=3, most visible is an increase in Burkina Faso's Nord region, in which over 40 people were killed by a vigilante self-defence militia in targeted attacks against three Fulani villages in early March. A number of extrajudicial killings were also reported in the Sahel region in March, coupled with the killing of 19 Fulani civilians in the Soum province on 2 March by suspected Koglweogo.
For more information see the monthly forecasts report for May 2020
- Forecasts May 2020, geographical level, ensemble model (state-based)
- Forecasts May 2020, geographical level, ensemble model (non-state)
- Forecasts May 2020, geographical level, ensemble model (one-sided)
- Forecasts May 2020, geographical level, changemaps (state-based)
- Forecasts May 2020, geographical level, changemaps (non-state)
- Forecasts May 2020, geographical level, changemaps (one-sided)
Please cite: Hegre, Håvard, Marie Allansson, Matthias Basedau, Michael Colaresi, Mihai Croicu, Hanne Fjelde, Frederick Hoyles, Lisa Hultman, Stina Högbladh, Naima Mouhleb, Sayeed Auwn Muhammad, Desiree Nilsson, Håvard Mokleiv Nygård, Gudlaug Olafsdottir, Kristina Petrova, David Randahl, Espen Geelmuyden Rød, Nina von Uexkull, Jonas Vestby (2019) ‘ViEWS: A political violence early-warning system’, Journal of Peace Research, 56(2), pp. 155–174. doi: 10.1177/0022343319823860.