A SCOURGE LIKE NO OTHER Cartographer: Markian Kuzmowycz DHP-P207: GIS for International Applications Fall 2018 ASSESSING LANDMINE VULNERABILITY AND 17 December 2018 Projection: Ukraine 2000 3 Degree GK Zone 9 RISK IN EASTERN UKRAINE

BACKGROUND METHODOLOGY RESULTS AND LIMITATIONS

Since the onset of war in Eastern Ukraine in To find the most sensitive areas, separate The results show a ring of high risk around 2014, monitoring groups have struggled to risk and vulnerability assessments were cre- the line demarcating government-controlled maintain a clear picture of the damage ated of the Donetsk and Luhansk Oblasts, vs non-government controlled areas caused by war and the changing vulnerabili- the two regional Ukrainian administrative ties for residents who have not been able to units which comprise an area known collec- (NGCAs). The highest risk areas tend to cor- relocate. tively as Donbas. Raster analysis tech- relate with densely populated areas where niques were applied to four separate vulner- there has been significant recorded military The OSCE and Human Rights Watch have ability variables: population density, proximi- activity since fighting began, particularly in reported the use of cluster munitions and FACTORS RISK ty to schools, proximity to roads, and dis- the important industrial cities of Mariupol, landmines by government forces and the self tance from hospitals and clinics. Euclidean Donetsk, Horlivka, Sloviansk, Luhansk, and -declared, Moscow-backed, Donetsk and Distance was applied to separate transpor- Luhansk People’s Republics. In less than tation, education, and health shapefiles – Kramatorsk. Additionally, areas east of five years, Ukraine has become one of sourced from OpenStreetMap – and the Re- Lisichansk, and northeast of Luhansk show world’s heaviest mined countries. Today the classify tool assigned all four vulnerability high scores, while there is slightly less vul- HALO Trust confirms that Ukraine led the factors scores from 1-5. The Raster Calcula- nerability to the population. world in unexploded ordinance (UXO) acci- tor overlaid these rasters together to create dents for three years running. Over 1,800 a composite score. mine casualties were recorded 2014-2017, Area Pop. at Risk and mine-related injuries represent two- Separately, this process was carried out for Profile (Km2) (Est.) thirds of all child fatalities in the region. risk factors. As the war in Donbas is ongoing and fairly young, comprehensive UXO data This project aims to determine the areas of does not yet exist. However, the HALO Low Risk & Vulnerability 49,351 2,028,908 Eastern Ukraine most at risk by war, includ- Trust, an international demining agency, has ing from both conventional weapons and UX- published Rapid Assessment shapefiles Elevated Risk & Vulnerability 3,558 1,336,074 Os through a geospatial vulnerability assess- which were utilized to determine both con- ment and consider where demining activities firmed and potential minefield locations. To Highest Risk & Vulnerability 726 392,830 may have the greatest societal impact. support this data, data was referenced from the Uppsala Conflict Data Program (UCDP), The analysis estimates that over 392,000 which has geotagged war events since the residents live in these zones classified as beginning of the war. By the estimate of the Highest Risk & Vulnerability, an area cover- Ukrainian government, only 15-20% of UX- 2 Os are landmines; by considering the densi- ing 726 km , while another 1.3 million live in an elevated risk and vulnerability zone cov- ty of more conventional forms of warfare, an 2 estimate can be made of where explosive ering 3,558 km . It is recommended that remnants of war (ERWs) are likely to be. demining activities prioritize these most sen- The Kernel Density tool was applied to this sitive areas to generate the greatest humani- data to create rasters which were also re- tarian and economic impact. The scale of the classified as 1-5 scores indicating density of UXO problem in Ukraine has not yet been risk. A risk composite was generated fully understood, but as international agen- through an identical process as vulnerabili- cies such as the OSCE and the HALO Trust, SOURCES ties. among others, work together to survey the Landmine and Cluster Munition Monitor (2018), OpenStreetMap (2018), Uppsala Conflict territory, special attention should be given to Together, these rasters were overlaid, with Data Program (2018), HALO Trust (2018), EU Civil Protection and Humanitarian Opera- areas where significant mine activity could risks receiving a 2.5x weight to vulnerabili- tions (2017), WorldPop, ESRI, UN Humanitarian Data Exchange, International Committee be expected, and where their clearance of the Red Cross ties; ultimately, there is no greater indicator would be most beneficial. of areas at risk to mines than proximity to weapons activity. This composite raster was This analysis is limited by a lack of publicly reclassified once again, considering the dis- available data from non-government con- tribution of the data, to separate the areas of trolled areas. Furthermore, the displacement high and most extreme risk and vulnerability. of over 1.2 million Ukrainians, and subse- Finally, the Zonal Statistics tool was used to quent lack of census data has made popula- estimate the population and the areas where tion analysis difficult. sensitivity is highest. VULNERABILITY FACTORS

Mine Casualties by Year, 2014-17 900 800 700 600 277 510 500 400 300 309 200 429 275 100 102 120 0 46 2014 2015 2016 2017

Killed Injured