During the night of 24 April 2026, Russian drones attacked residential areas of Odesa, including buildings in the Moldavanka neighbourhood. According to information from the State Emergency Service reported by Suspilne, a three-storey residential building and three two-storey buildings sustained significant damage. Two of the two-storey buildings were almost completely destroyed. A fire broke out in the three-storey building, damaging apartments and structural components.
The main environmental footprint is associated with the fire, the destruction of older residential buildings and the generation of a substantial volume of construction waste. Materials that may have burned included wooden floors and roofing components, furniture, polymer-based interior finishes, cable insulation, household appliances and thermal insulation.
Incomplete combustion of these materials can release CO₂, carbon monoxide, smoke, soot, ash, volatile organic compounds and fine particulate matter. The blast wave and structural collapse also generated dust, fragments of brick, glass and metal, and other potentially contaminated waste.
For the scenario assessment, the total affected building area was estimated at 1,200–4,000 m², while the possible burned area was estimated at 150–800 m². Open sources provide no official information on the precise duration of the fire or the mass of burned materials, so the assessment uses a scenario with an unknown duration or burning lasting up to 12 hours. The lower bound reflects localised burning within part of one building and the minimum quantity of materials involved. The upper bound represents an expanded fire and damage scenario within the same residential location.
The tree figure relates exclusively to notional compensation of the CO₂ footprint, assuming that the trees are planted and maintained for at least 10 years. Trees do not compensate for soot, toxic combustion products, construction dust, soil or water contamination, or hazardous waste.
Browse all preliminary environmental damage assessments published by ECO AI.