Oil spills has a significant detrimental impact on the environment, and remote sensing via satellite imagery has an increasingly important role in oil spill detection and response.

On May 29, 2020, a fuel storage reservoir failed at a power plant outside Russia’s northernmost city of Norilsk. This caused leaking of more than ten thousand tonnes of oil contaminating the Ambarnaya river. The river flows into lake Pyasino and its surrounding subsoil.

A satellite image captured over Norilsk on May 31, 2020, shows an overview of the locations of the oil spill and the factory oil depot that caused the environmental disaster.

The Promimagery team collected the latest satellite imagery from May 30, 2020 up to June 13, 2020. This series of satellite images focused on the northern part of the Ambaryana river just around 5 kilometers from the lake Pyasino shows a concentration of leaked oil.

Loading...

Time series of images captured from May 30, 2020, to June 13, 2020, shows images of the blood-red color of Ambarnaya river in Norilsk Russia due to an oil spill happened on May 29, 2020

A satellite image captured last June 13, 2020, also shows multiple makeshift structures and oil control booms used to control and contain the spread of oil and avoid it further reaching Lake Pyasino.

Satellite imagery captured June 13, 2020, over the Amabarnaya river shows the location of oil control booms and makeshift structures with an estimated size of 400 square meters used for containing the further spread of the oil spill.

Applications of remote sensing in oil spill detection

Instead of visually inspecting the extent of the oil spill from satellite imagery, the Proimagery team used remote sensing techniques to scientifically quantify and pinpoint the extent of the spill on the Ambaryana river. Using object-based image analysis, the Proimagery team processes multiple satellite images captured over the surrounding area to determine and pinpoint affected regions.

Loading...

Classified oil spill extent derived from satellite images using object based image classification