Mariupol evacuation-window building damage lab

Compares one fixed 2021 baseline with imagery nearest each evacuation period, calculates visual change inside OpenStreetMap building footprints, fills only high-change building footprints in red and prints the detected damaged-building percentage directly on the after image. Intact buildings remain unmarked, and no circles are drawn. The result is a research screening layer—not verified damage.

TEST AOI · 47.0975, 37.5448 · Z17
Loading maps and archive releases…
Important: exact images for every requested date may not exist. The page selects the nearest locally changed archive release and shows its underlying acquisition date when Esri metadata provides it. A release outside the requested window is marked as a proxy.
BASELINE · TARGET 2021-02-24
Metadata pending
AFTER · 5–9 MAR 2022
Metadata pending

UNOSAT Official Damage Assessment

United Nations Satellite Centre (UNITAR/UNOSAT) verified damage points overlaid on post-damage satellite imagery. Dots update based on selected evacuation window.

UNOSAT DAMAGE OVERLAY
Loading UNOSAT damage points…

UNOSAT Assessment

of structures with visible damage

Nighttime Lights — Simulated from VIIRS Black Marble

Satellite imagery rendered as nighttime view with simulated city lights. Light intensity calibrated to real VIIRS Day/Night Band measurements. Bright pixels in the satellite image indicate buildings and infrastructure — their glow is reduced according to measured light loss during the siege.

BASELINE · Pre-conflict · 20 Feb 2022city illuminated
Simulated nighttime — 100% light output
DURING SIEGE
Loading…

Light Intensity Comparison

estimated light reduction over Mariupol
Light intensity calibrated to VIIRS DNB pixel brightness measured within Mariupol city bounds. Values relative to 20 Feb 2022 pre-invasion baseline.
NASA VIIRS Day/Night Band · Suomi NPP
Satellite base: Esri Wayback World Imagery

Data-Driven Evacuation Analysis

Evacuation routes from the most damaged and darkest zones of Mariupol to Zaporizhzhia (agreed safe destination). Origins are data-driven: cells with highest damage + lowest light survival + most buildings. Routes thread through least-damaged cells via Dijkstra pathfinding on a 250 m grid. Hover over a zone for details.

EVACUATION ROUTE ANALYSIS
Loading…

Evacuation Analysis

Population: OSM building centroids (45,544)
Damage: UNOSAT CE20220223UKR · Lights: VIIRS Black Marble