Waging a war against light pollution to save the world — Machine Learning is all set to help

Author(s): Supriya Ghosh

Originally published on Towards AI.

Technology

Light Pollution is a known phenomenon that is seeking significant attention globally due to its alarming rate of harming all living species including humans. It can prove hazardous if not monitored and controlled effectively.

Only one good sign is that it can be reversed unlike many other forms of pollution if each one of us starts making a difference by contributing towards minimizing it. Hence, in this context, only being aware of this problem is not enough, instead, there is a need to take action.

What is light pollution?

Light pollution is an injudicious or unwanted or excessive use of artificial light in an area that is above an optimal threshold.

Components include :

• Glare — excessive or unnecessary brightness that causes temporary blinding effect or vision perturbance. e.g., High beams from a car coming from the opposite end.

• Sky glow — brightening of the night sky due to scattering of lights.

• Light trespass — light falling into areas where it is not needed.
• Clutter — excessive groupings of light sources which often create confusion, distraction, and can cause accidents. e.g., roads where the streetlights are badly designed.

Hazards to living beings including humans

1. It disrupts the day and night cycle as the night sky gives an illusion of daylight endangering creatures (animals, insects, birds, aquatic species, and plants) by interfering with their pattern of reproduction, nutrition, sleep, safeguarding from predators, etc.

2. Artificial Lights cause baby sea turtles to die as sea turtles live in the ocean but brood at night on the beaches. Babies from such brooded eggs find their way to the sea by detecting the natural light in the sky over the ocean. But in presence of artificial lights, they are driven in opposite direction instead of their way back towards the ocean.

3. Birds migrating at night uses natural lights of the moon and stars. Artificial light causes them to lose their direction and deviate towards the skyscrapers in cities making them collide and die.

4. Artificial lights cause migratory birds to migrate before their intended migratory seasons and make them move either too early or too late. This actually deprives them of the ideal climate needed for nesting, tracking their breeds, and other behaviors. This severely affects their migration patterns which in turn affects the ecosystem.

5. A large number of insects which are food sources for birds and other animals are attracted to artificial lights and get killed while coming in contact with light sources. This affects the food chain in dangerous ways.

6. Too much light emission/bright light wastes energy resources leading to unpleasant economic and environmental effects.

7. Sky Glow caused by uncontrolled outside lighting fades the visibility of constellation of stars and milky way and makes it hard to view many celestial bodies. This can go to such an extent where our coming generations can only see the stars, Milky Way, and galaxies in a planetarium, rather than with the naked eye.

8. Artificial light at night negatively affect human health, increasing the risks for obesity, increased heart rate and palpitations, depression, sleep disorders, diabetes, breast cancer, and many more by disrupting our biological clock.

9. Night exposure to artificial light especially blue light suppresses melatonin production, which is essential for boosting the immune system, and helps in lowering cholesterol, supporting the functioning of the thyroid, pancreas, and many other glands.

Ways to Address the Problem

The only solution is to bring back the natural Dark Sky which existed at the times of our ancestors or at least close to it.

This can be achieved by following one of the following measures –

a. Implementation of certain regulations to limit the use of artificial lighting as well as advocating the right mechanism for its use.

There is much traction seen in this direction and global experts and regulatory bodies are busy setting rules and standards.

b. Measurement and prediction through mathematical and machine learning models.

It is often believed that if something can be measured then it can be monitored and controlled as well. But only measurement is not enough, and prediction using machine learning models is also a necessary part so that it can be monitored and controlled in the present and future both.

Understanding the implementation Steps — The technical aspects

Modeling of Light pollution is fundamentally a radiative transfer problem.

1. Experts and researchers use radiative transfer (energy transfer in the form of electromagnetic radiation) to capture the scattering and the propagation of light pollution in the atmosphere.

2. The radiative transfer mathematical model considers multiple propagation parameters like spectral reflectance, obstacles, atmospheric conditions for computation. Another important parameter considered is the population of the specific area under consideration.

3. All these parameters are utilized for developing machine learning models using required algorithms and techniques.

4. A detailed analysis and hyper tuning of machine learning and mathematical model is carried out to gain desired efficacy.

5. These ML models then conveniently determine if an area has light pollution or not.

6. Only one specific area is targeted at one time.

7. With the use of these developed models, the digital and graphic map of the area capturing sky brightness levels and light pollution levels are created.

8. Once contributions and maps from each specific area are ready, these are summed to get the total contribution to the night-sky brightness and light pollution in the entire region.

9. The results captured are extrapolated i.e., prediction is done for getting future trends showing what the skies will be like in the next few years.

The primary goal is studying, monitoring, and predicting the extent of light pollution in different cities and towns (urban, suburban, and rural areas).

This result completely shows the light pollution prospects in different regions for the coming years so that corrective programs can be employed beforehand to curb it or lessen the impacts.

Conclusion

There is no doubt that light pollution is hazardous and needs to be controlled. Hence it is the right time to act in the present to save the future. And especially with technologies like AI and ML at our disposal, it is high time to leverage these to the core and save the world from this terrible phenomenon.

One more important job to do at our end is spreading awareness as many people still don’t know or don’t understand about light pollution and its negative impacts. We have to ensure that, more and more people understand and come forward to take the necessary steps to preserve the natural night/dark sky.

Thanks for reading !!!

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI