Mapping Crimea's Blackouts from Space
Mapping Crimea's Blackouts from Space
In the summer of 2026, Crimea experienced a series of major disruptions to its electricity supply. Official information on the actual scale of electricity consumption and outages was limited, making it difficult to understand how widespread these events were from reports on the ground alone.
Together with Meduza, Vertical52 approached the problem from a different perspective: we looked at Crimea at night from space.
Using satellite observations of nighttime lights, we reconstructed how artificial illumination changed across the peninsula between April and the end of August 2026. The analysis revealed three distinct waves of dramatic light loss, in late June, late July and late August, separated by periods of partial recovery. During the strongest events, nighttime illumination across much of Crimea dropped to only around 10 to 15 percent of its normal level.
Measuring an electricity crisis through nighttime lights
Cities, roads, industrial sites and residential areas produce a characteristic pattern of artificial light that can be observed from orbit after sunset. When electricity supply is disrupted on a large scale, this pattern changes.
For the analysis, we used daily nighttime-light observations from NASA's VIIRS Black Marble dataset. Rather than interpreting individual satellite images visually, we transformed the observations into a time series that allowed us to compare the same locations over several months.
The core of the methodology consisted of four steps. First, we divided the peninsula into municipalities rather than treating Crimea as a single area, since a large city and a predominantly rural district naturally have very different levels of nighttime illumination.
Second, we established a normal level of light for each municipality, using observations from April, before the major summer decline, as an individual reference baseline. Each later observation could then be compared with what was normal for that particular area rather than with an arbitrary peninsula-wide brightness value.
Third, we converted brightness into a relative index, with the baseline for each municipality defined as 100. A value of 50 means nighttime illumination had fallen to approximately half of its reference level; a value of 10 means only around one tenth remained. This allowed very different parts of Crimea to be compared on the same scale.
Fourth, we tracked the signal through time and across the peninsula, combining the municipal measurements into a peninsula-wide time series and mapping the changes geographically, which allowed us to distinguish an isolated local reduction in light from an event affecting most of Crimea at the same time.
The result was not simply a sequence of darker satellite images, but a quantitative record of how nighttime illumination collapsed and recovered across different parts of the peninsula.
What the satellites revealed
The first major decline appeared in late June. Nighttime illumination dropped sharply across Crimea, falling well below its April reference level. This coincided with the beginning of a period in which attacks on electrical substations intensified.
The lights subsequently began to return, but the recovery was incomplete. During the second half of July, the peninsula remained substantially darker than during the baseline period.
Then came a second and much deeper collapse. Around 29 July and the first days of August, nighttime lights fell across large parts of Crimea almost simultaneously. At the lowest point, the peninsula-wide indicator reached roughly 10 percent of the normal level. In several municipalities, the difference was so extreme that areas normally covered by a visible network of settlements and infrastructure appeared almost dark from orbit.
This was followed once again by a gradual recovery rather than an immediate return to normal conditions.
Finally, on 26 August, the satellite record showed a third major drop in nighttime illumination. Taken together, the observations point to three separate waves: late June, late July and late August.
The scale of these changes was the most important result of the analysis. What initially looked like a single disruption followed by recovery instead developed into a repeating pattern: a rapid collapse in nighttime illumination, a period of partial restoration, and then another collapse.
Connecting the satellite signal with events on the ground
Nighttime-light data alone cannot explain why a region becomes darker. For the joint investigation, Meduza therefore compared the satellite-derived timeline with independently geolocated reports of Ukrainian attacks in Crimea, including strikes documented by the Geoconfirmed project.
The timing showed a clear relationship between periods of concentrated attacks on electrical substations and the strongest reductions detected in the nighttime-light data. In particular, Meduza identified the 29 July wave as involving at least 15 geolocated strikes on substations; the satellite observations immediately afterwards showed one of the deepest reductions in nighttime illumination during the entire study period.
The broader pattern was also significant. Illumination partially recovered when attacks shifted away from substations, before falling again during later waves of disruption. This does not by itself prove a specific military strategy, but it provides an independent, observation-based record of the consequences visible on the ground.
What nighttime lights can, and cannot, tell us
Satellite-measured nighttime illumination should not be interpreted as a direct measurement of electricity consumption. Electricity is used for far more than lighting, and a 50 percent reduction in visible nighttime light does not necessarily mean a 50 percent reduction in total electricity use. What the satellite data provide is a proxy: an independent indicator of how strongly the functioning of illuminated settlements and infrastructure has changed.
For this reason, the analysis is most useful for identifying the timing, geographic extent and relative severity of large disruptions. In Crimea, that signal was unusually strong. Across several months of observations, the peninsula did not simply become gradually darker. Instead, satellite data recorded three abrupt and geographically widespread collapses in nighttime illumination, with the most severe episodes reducing visible light to approximately 10 to 15 percent of normal conditions.
