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 the prio-grid month level for state-based conflict continue to be in northeastern Nigeria, the Anglophone region of Cameroon, the North and South Kivu provinces in DR Congo, Somalia (southern states in particular), Egypt's Sinai, and the northeastern Cabo Delgado Province of Mozambique where an Islamist insurgency emerged at the end of 2017. The risk of violence in Mali and Burkina Faso also remains high, but is more spread out geographically. Most of these regions have been facing violence for years, reflecting that countries' recent conflict history is the strongest predictor of future violence.
Compared to last month, we find the strongest increases in the risk of state-based violence in the northeast of Nigeria given ramped up security operations in the region since April. Interestingly, the risk appears to have increased this month in the southern parts of Borno State, while the risk has decreased in its northern parts. Nonetheless, the risk remains significant across the region. While the cluster of violence in central Mali and northern Burkina Faso has shown a decline in the probability of state-based violence compared to last month, the risk does appear to have increased along the southern border with Niger. The Anglophone areas of Cameroon show a slightly increased risk of state-based violence in September, while by contrast in DR Congo, the risk has again increased significantly in Ituri province given a large-scale government operation against militias there. Continued clashes with Islamist militants in Mozambique futhermore produces a significant elevation in the risk of state-based violence in the Cabo Delgado Province. In Somalia, finally, we find an unclear pattern of increase and decrease in its southern regions and around the country's capital.
The forecasts for non-state conflict and one-sided violence depend on the same factors although with somewhat different implications. For non-state conflict, we most notably predict a worsened cluster of non-state violence in the northeastern DR Congo province of Ituri, given escalated ethnic violence between ethnic Hema and Lenuda, leaving at least 170 dead in Djugu area in June. Compared to last month, Nigeria shows an increased risk of non-state violence in the country's southwest in particular. On 12 July a group alleged to be Fulani herders killed the daughter of a pan-Yoruba leader. With regard to the risk of one-sided violence, we again find a strong increase in the DRC's Ituri province given escalated violence against civilians perpetrated by the ADF and other militias. Nigeria's Borno State furthermore shows a continued rise in the risk of one-sided violence given increased Boko Haram attacks in July. Ethiopia, too, shows a southern cluster of risk as a result of deadly clashes between security forces and ethnic Sidama that protested the delay of a referendum on the creation of a Sidama State. Somalia, finally, faces a heightened risk due to ramped up attacks by Al-Shabaab.
For more information see the monthly forecasts report for September 2019
- Forecasts September 2019, geographical level, esemble model (state-based)
- Forecasts September 2019, geographical level, ensemble model (non-state)
- Forecasts September 2019, geographical level, ensemble model (one-sided)
- Forecasts September 2019, geographical level, changemaps (state-based)
- Forecasts September 2019, geographical level, changemaps (non-state)
- Forecasts September 2019, 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.