Los Angeles Streetlight Retrofit Analysis
The goal of this investigation is to determine the effectiveness of a streetlight retrofitting project in reducing light pollution as measured by the VIIRS sensor aboard the Suomi NPP satellite. It is a collaboration between Darryl Bishop, Christopher Fisher, and Erick N Fox.
Los Angeles Bureau of Street Lighting
HPS vs led streetlights
The city of Los Angeles embarked on a streetlight retrofitting project in 2012 aimed primarily at reducing energy costs by replacing traditional yellow-orange high pressure sodium (HPS) lights with modern LED lights. A secondary goal of this project is to reduce light pollution through the design of the LED streetlights which only direct their light downwards as opposed to the exposed globes on traditional "cobra-head" HPS lights which allow light to escape sideways and upwards into the atmosphere. Notice in the above comparison image that the HPS lights on the left exhibit glare even though the photograph was taken from above the level of the lights. The LED lamp heads however do not emit light above the horizontal, as shown in the right hand image, therefore greatly reducing their light pollution.
viirs sensor: Day/Night Band
The Visible Infrared Imaging Radiometer Suite (VIIRS) instrument comprises 22 narrowband channels from 0.4-12 μm. As a subset of VIIRS, the Day/Night Band (DNB) in the 0.5-0.9 μm range is able to detect light up to 10 million times fainter than reflected sunlight, or 10 to 100 times fainter than even reflected moonlight.
CHANGE DETECTION
The above change detection map for the city boundaries of Los Angeles illustrates that most areas saw a decrease in DNB-measured light (blue areas), while some isolated areas exhibited an increase (red). The bluest areas show a 1% decrease in measured light, and the reddest areas a 1% increase. The mean change was -2.88 nW∙cm²∙sr¹ from 2013 to 2014. This map was generated in ENVI using only data from cloudless and moonless nights.
STATISTICS
A two-tailed t-test at p≤0.01 was conducted on the data to determine if the measured difference was sufficient to reject the null hypothesis that there was no change in measured radiance. Based on the fact that complete satellite coverage of the study area existed for both data sets, the change was shown to be significant with a t-value of 21.731, far above the threshold value of 2.581.
Conclusion
This investigation serves as a proof of concept that the current generation of satellite sensors are sufficiently sensitive to measure changes in nighttime light pollution. It also shows that even the replacement of hundreds of thousands of streetlights only dropped the measured light pollution by a small amount. Other sources such as vehicle lights, commercial signs and lighting, and security lights also contribute to overall light pollution. Governments and other entities which wish to reduce light pollution should look to the DNB data to quantify the results of their efforts going forward.