Using the environmental light field method for measuring biologically relevant light characteristics at the household scale

The increasing prevalence of artificial light at night (ALAN) alters the natural light cycle and has been found to have harmful effects on both human and environmental health. Given its importance as a disease vector, we use Anopheles mosquitoes as a case study. Anopheles mosquitoes, several of which act as malaria vectors, and compare how light characteristics (details of the method followed below) influence key mosquito behaviours differ among regions and houses. We characterise the artificial light environment inside and immediately outside houses in rural settlements (South Africa, Mozambique, and Uganda) as the lighting present in these areas will likely play a role in affecting people’s risk of being infected by vector-borne diseases. 

Methods
Light measurements

The full methods for the environmental light field (ELF) method and analysis are described in Nillson & Smolka (2021), Quantifying biologically essential aspects of environmental light. Journal of The Royal Society Interface 18:rsif.2021.0184, 20210184.  

A Nikon D850 camera body and a Sigma 8mm 180° fish eye lens were used in this study. We measured the light characteristics in houses in rural settlements in South Africa, Mozambique, and Uganda. Houses in Venda were measured from May to July 2023 in Mozambique, Mamfene, and Uganda.

Each house is classified as an 'environment' in the ELF method. To characterise each environment, images were taken of four 'scenes' within that environment. Scenes were captured by rotating the camera 90° in a clockwise direction between scenes in order to ensure a 360° view of the environment was captured, with overlap between the scenes. Each scene was bracketed, capturing three photographs of each scene, separated by three exposure evaluation settings (3EV). Bracketing is used to increase the dynamic range of the images in the analysis process. In total, 20 houses were measured (Table S2), both inside and outside. Therefore, a total of 480 photographs were taken (20 houses, 2 locations per house, 4 scenes with 3 photographs per scene [2024*3]).  

The camera was placed on a tripod at 1 metre above the ground and positioned so that it was level in both the vertical and horizontal planes, using a spirit level mounted on the camera. The camera was placed in the centre of the room for inside measurements, and no further than 2 m outside the houses for outside measurements. The ELF method characterises the entire light environment as detected by the animal eye, and captures a range of potential variation at the different elevation angles in the vertical field of view. Therefore, replication within an environment is not necessary.    

Data analysis       

To characterise the light environment, the photographs taken were processed in Matlab (The MathWorks Inc., 2022) using the ELF open-source software (Available at: https://github.com/sciencedjinn/elf/archive/refs/heads/master.zip). The ELF analysis software used remaps all images from a circle into an equirectangular format, calculates the image statistics, the mean image of the 360◦ environment, average image statistics, and creates the final plots describing the light characteristics of the environment. All the image statistics calculated include (for each spectral channel, i.e., white, red, green, and blue wavelengths) the mean radiance, radiance standard deviation, median radiance, 2.5th, 25th, 75t,h and 97th percentiles and minimum and maximum radiances for 3◦ elevation angle bins within the 180◦  field of view (Nilsson and Smolka, 2021).     

The light environment for each house is described at that specific position for a living organism in a nuanced and meaningful manner (e.g., for a range of biologically relevant wavelengths), reflecting the way animals observe their environment. Radiances are described as photon flux per nm (photons m -2s -1nm -1) for white light and the separate channels (Nilsson and Smolka, 2021). Radiance values are log10 transformed, allowing for comparison over all natural intensities, producing values in a unit called 'lit' (Nilsson and Smolka, 2021).   

Note: The picture_start_number in the data file corresponds to the first image of the bracketed sequence for each scene within an environment/house. The three bracketed images of each scene are considered one file in analysis.