Mapping Access to Electricity and Urban Night Lights: Leveraging the Massive Repository of Astronaut Photography of the Earth

Abstract

The challenge

Nighttime light (NTL) images from the International Space Station (ISS) have been used by very few researchers without exploring their full potential. Potential application fields have a large range, including the monitoring of energy consumption, light pollution, urban extent, socio-economic and population modelling, and electrification. The low spatial resolution of satellite-based observations, e.g., 750m for Suomi-NPP VIIRS-DNB has restricted researchers from exploring the full potential of NTL images for within-urban scales. In addition to the coarse spatial resolution, NPP was designed primarily for meteorological, and the recent shift to LED lights is not well captured because VIIRS-DNB is blind to blue wavelengths. Since the 1990s, NASA astronauts and later also ESA, JAXA, ROSCOSMOS and CSA-ASC, have been acquiring NTL images with DSLR cameras. However, images are not systematically calibrated, georeferenced and accessible.

Methodology

Urban monitoring with nighttime remote sensing is very different with daytime remote sensing. Commonly built-up areas have traces of NTL, however, many informal areas in the Global South are not (well) connected to electricity (Figure 1b). In ISS images, NTLs are often linear features along the main transport axes (Figure 1a). Acquired ISS NTL images come with various geometric distortions and require georeferencing (high order transformation) and radiometric correction for quantitative, temporal and comparative applications. The Citizen Science program Cities at Night tagged, located and georeferenced them to make NTL images accessible. Leveraging this repository of ISS NTL, we show the potential of ISS to support diverse EO application fields. For several application fields, we explore the required spatial resolution (defined by the focal length of the DSLR camera and the position of the ISS). For example, for monitoring the energy transition and the reduction of light pollution (e.g., for European cities), high spatial resolution is required to identify individual pollution sources. High spatial resolution nightlights are also needed for mapping access to the electricity grid (e.g., in African cities), to map the heterogenous intra urban patterns of NTL. We also explore whether high radiometric resolution and RGB images allow determining sources of lighting (e.g., the shift to LED lights), comparing images of different years for the same city.

Results

Related to lighting sources, since the LED lights have started to replace the old discharge lights, the radiometry values that VIIRS-DNB provides (DMSP-OLS did not have any radiometric calibration) can be misleading as they largely exclude light emitted in the blue spectrum. Therefore, VIIRS-DNB images might indicate a light reduction, which is not the case when compared to ISS images. Related to electrification, high spatial resolution (e.g., ISS images acquired with a focal length of 400 mm), is required to detect small neighbourhoods not connected with the electricity grid, while the presence or absence of street lights can be concluded with a focal length of 180 mm (slightly lower spatial resolution). Such information is essential for infrastructure development and the upgrading of informal settlements or planning to connect services (e.g., health, education) to the electricity grid. Related to light pollution, very high-resolution images allow the mapping of sources of light pollution and the modelling of dark corridors to support applications in the field of biodiversity. For example, the massive reduction of insects is related to the increase in light pollution in urban, suburban and peri-urban areas. Outlook for the future For the development of a high-resolution NTL satellite-based sensor system characteristics are not well understood and are challenging to define. Generally, there is limited scientific knowledge across application fields of suitable sensor requirements. The repository of ISS NTL images, which is extremely rich but very under-utilised by researchers, is an optimal testbed to develop applications and an understanding of requirements. Such requirements include defining spatial, spectral, radiometric and temporal base requirements for the application fields. We are presently exploring, collecting and expanding different applications and documenting their requirements toward defining sensor requirements for a high-resolution nighttime light mission.