Trends in Forum Ecology & Evolution Investigating the impacts of artificial light via blackouts Arjun Amar 1,*, Chevonne Reynolds2, Robert L. Thomson1, and Davide Dominoni3 Natural experiments provide re- markable opportunities to test the large-scale effects of human activ- ities. Widespread energy black- outs offer such an ‘experiment’ to test the impacts of artificial light at night (ALAN) on wildlife. We use the situation in South Africa, where regular scheduled black- outs are being implemented, to highlight this opportunity. Understanding impacts of ALAN on animal behaviours ALAN originates from man-made light sources such as streetlights and buildings, and it is increasing globally [1]. This sen- sory pollutant has been linked to a range of negative effects on wildlife and the envi- ronment, including the disruption of a range of animal behaviours such as move- ments, timing of activity, feeding behav- iours, and species interactions [2–4]. Exploring the impacts of ALAN on the behaviours of free-ranging animals is chal- lenging. Most research has been correla- tive, examining responses across ALAN gradients [5], but this approach often fails to distinguish ALAN effects from other urban pressures such as noise pollution [4]. Although some ALAN exposure exper- iments have occurred, these have either been small-scale (e.g., adding lights to nest boxes [6,7]), or (by necessity) have 612 Trends in Ecology & Evolution, July 2024, Vol. 39, No. limited replicates or only illuminated small areas [8,9]. We highlight that energy blackouts, where electricity supply fails or is turned off, pro- vides an exciting opportunity to better understand how species respond to the presence and absence of ALAN in a direct manner. Blackouts are increasingly com- mon in many parts of the world, and in South Africa load-shedding (with sched- uled blackout locations and times) has been introduced as a means of managing these blackouts, providing an ideal oppor- tunity for research on the impacts of ALAN. South African load-shedding: what is it and how does it operate? In South Africa, Eskom (the national power utility) is currently unable to fully meet the demand for electricity, resulting in sched- uled blackouts throughout the country. Unlike random outages that occur else- where, these load-shedding episodes are pre-planned for specific zones. Zone size varies with population density but is gen- erally many square kilometres (Figure 1). Eskom provides schedules in advance for these blackouts, which last approxi- mately 2–4 h, occurring once to three times daily based on Eskom’s set ‘stage’. Each stage indicates 1 GW of energy re- duction across the country. For instance, stage 1 cuts 1 GW, while stage 6 cuts 6 GW and is equivalent to ca. 10% of South Africa’s overall grid capacity of 60 GW. The frequency of these blackouts in South Africa has risen in recent years [10]. In 2021, there were 75 days of load- shedding, rising to 205 days in 2022, and in 2023 there were only 30 days without load-shedding. Load-shedding intensity (i.e., stage levels) have also escalated over this periodi. Despite mitigation efforts, fore- casts suggest that these blackouts will persist for many years [10]. Blackouts cause a clear and obvious re- duction in ALAN (Figure 1). The impact of 7 load-shedding on ALAN is quantifiable from remotely sensed data, with night- time radiance in South African cities de- creasing in recent years from a monthly average of 15.6 (nanowatts/steradian/ cm2) in 2014–2018 to a monthly average of 12.6 (nanowatts/sr/cm2) during 2022– 2023 as load-shedding has increased (Figure 2). During a typical South African winter night, load-shedding lasting 2–4 h would cover a loss of ALAN across 17% to 33% of the night. Blackouts and load-shedding in South Africa: opportunities for research We propose that the international research community leverage South Africa’s natural experiment of load-shedding to enhance our global knowledge of the effects of light pollution on animal behaviour. These sched- uled outages across vast urban zones allow for comparative studies on animals’ reac- tions to artificial light (non-load-shed areas) versus its absence (load-shed areas) within the same landscape and timeframe. This scenario provides a global first: an opportu- nity for a large-scale, controlled experiment to investigate ALAN’s impacts on animal behaviours. Load-shedding presents several unique strengths for research. Implemented na- tionwide, this practice provides numerous replicates over a vast, biodiverse area (1.22 million km2). This enables studies of ALAN responses across varied climates. For instance, cities like Durban have a tropical climate, others (like Cape Town) are Mediterranean. Urban areas in South Africa also span a diverse range of biomes and habitats, including forested, grassland and fynbos areas, freshwater habitats, and coastal marine areas, providing opportuni- ties to examine the impacts of ALAN across a suite of different taxa encompassing a broad phylogenetic spectrum. The size of load-shedding zones typically encom- passes the home range of most species, ensuring that for some individuals their http://crossmark.crossref.org/dialog/?doi=10.1016/j.tree.2024.04.006&domain=pdf https://orcid.org/0000-0002-7405-1180 CellPress logo TrendsTrends inin EcologyEcology && EvolutionEvolution Figure 1. Load-shedding zones for areas supplied by the City of Cape Town, illustrating the size of the areas and the manner in which power cuts (lasting 2–4 h) are implemented. Zones in black are being load-shed, and those in yellow are not. The hashed areas have power supplied by a provider other than the City of Cape Town. A zone’s size is based primarily on population density, with more populated areas having smaller zones. Load-shedding causes meaningful and obvious changes in artificial light at night (ALAN) across large areas of a city, as shown by these photos of Cape Town taken a few minutes before, with load-shedding implemented in most of the Cape Town Central Business District (A) and then after load-shedding has ended (B). The red square on themap indicates the general regions seen in the images, which are taken from a YouTube videoiii with permission from Eric Nathan© (2023). Trends in Ecology & Evolution entire habitat goes dark during times of load-shedding. Of course, this natural experiment has its own limitations. The power cuts last only a few hours nightly, making them less suitable for studying longer-term ALAN impacts such as population-level effects. Even dur- ing nights with load-shedding, some ALAN remains, albeit in varying and diminished quantities. Furthermore, increases in the use of alternative energy sources (such as batteries and inverters) could result in ALAN increasing during load-shedding Tre over time, and also argues for some on- site light measurements within study areas. Additionally, because load-shedding has occurred regularly now for several years in South Africa, it could mean that species may have already responded to it, for exam- ple by modifying their distribution or other behaviours; this lack of baseline measure- ments of animal behaviours prior to any load-shedding may thus be an additional limitation. Lastly, although no information currently exists, it is possible that other human activities are also reduced during load-shedding. If so, human activity – and thus human disturbance – may be some- what confounded with load-shedding, cre- ating challenges to disentangle the effect of ALAN from other forms of disturbance. Another important constraint with this sys- tem is the insecurity associated with field work at night in South Africa, especially in urban areas where crime rates are high. However, many of these risks can be circumvented by using remote technology to monitor behaviours, such as deploying GPS tracking devices, camera traps, or re- mote acoustic monitors. Given these advantages and limitations, we suggest that load-shedding is best suited for examining ALAN’s impact on shorter-term behavioural responses, such as movement, foraging patterns and suc- cess, species displays, or species interac- tions. Several of these responses can be explored using GPS tracking devices. Ad- ditionally, variation in load-shedding sched- ules (e.g., 0, 2, or 4 h load-shedding across any one night) could provide for a dose– response analysis, which can expand the types of responses that can be explored, such as examining stress markers or sleep levels under varying degrees of nightly ALAN exposure. Similarly, different areas of cities might be dominated by different light sources (e.g., warm tungsten lamps versus cold, blue-rich LED lights), which may provide opportunities to test for behavioural responses to different light spectra [11]. nds in Ecology & Evolution, July 2024, Vol. 39, No. 7 613 Image of &INS id= CellPress logo TrendsTrends inin EcologyEcology && EvolutionEvolution Figure 2. Declines in average monthly radiance levels (nanowatts/sr/cm2) across Johannesburg, South Africa (black line) – blue line: locally weighted scatterplot smoothing (LOESS) trendline – following an increase in average monthly load-shedding stages in recent years (red bars). Monthly average radiance data were extracted from composite images of night-time data from the Visible Infrared Imaging Radiometer Suite (VIIRS) Day/Night Band (DNB), and data on load-shedding stages is publicly available and provides national scale data on the load-shedding status as they are updated. These load-shedding data were averaged daily and then used to calculate the mean monthly load-shedding stageiv. Trends in Ecology & Evolution Research initiative to exploit this research opportunity Our goal here is to alert the global research community to this unique opportunity. South Africa's situation is one example of well-documented load-shedding, but sev- eral global cities, especially in developing economies, may offer similar opportunities [12]. Ongoing changes in global climate and extreme weather are leading to in- creased energy demand and consequently, more frequent load-shedding events by energy providers. This trend may result in comparable research opportunities in vari- ous countries in the future. While we invite researchers worldwide to utilise the situation in South Africa, we ad- vocate for a ‘global science’ approach, as 614 Trends in Ecology & Evolution, July 2024, Vol. 39, No. outlined by Asase et al. [13], promoting collaboration between researchers from developed and developing nations, rather than the historically exploitative parachute approach. Coetzee et al. [14] have previ- ously highlighted the paucity of studies of ALAN in Africa, and this opportunity also presents a chance to directly address this knowledge gap. A global research fund could facilitate inter- national collaborative research, allowing experts in ALAN research, predominantly from the global north (wheremost research to date has occurred [14]), to collabo- rate with local South African researchers, leveraging their invaluable taxon- and locale-specific knowledge. Such collabo- ration could yield novel ecosystem-level 7 insights into ALAN impacts by enabling studies across different taxa within the same system (e.g., within predator–prey networks). We advocate for the estab- lishment of an ‘ALAN load-shedding ini- tiative’, mirroring efforts initiated to study the ecological effects of COVID-19ii. Declaration of interests No interests are declared. Resources ihttps://loadshed.theoutlier.co.za/ iiwww.bio-logging.net/ iiihttps://youtu.be/BK0zOMFLnQs ivhttps://docs.google.com/spreadsheets/d/1ZpX_ twP8sFBOAU6t–vvh1pWMYSvs60UXINuD5n-K08/ edit#gid=863218371 1FitzPatrick Institute of African Ornithology, Department of Biological Sciences, University of Cape Town, South Africa 2School of Animal, Plant and Environmental Sciences, University of Witwatersrand, Johannesburg, South Africa 3School of Biodiversity, One Health and Veterinary medicine, University of Glasgow, UK *Correspondence: arjun.amar@uct.ac.za (A. Amar). https://doi.org/10.1016/j.tree.2024.04.006 © 2024 Elsevier Ltd. All rights reserved. References 1. Hölker, F. et al. (2010) Light pollution as a biodiversity threat. Trends Ecol. Evol. 25, 681–682 2. Sanders, D. et al. (2021) A meta-analysis of biological impacts of artificial light at night. Nat. Ecol. Evol. 5, 74–81 3. Knop, E. et al. (2017) Artificial light at night as a new threat to pollination. Nature 548, 206–209 4. Gaston, K.J. et al. (2021) Pervasiveness of biological impacts of artificial light at night. Integr. Comp. Biol. 61, 1098–1110 5. Senzaki,M. et al. (2020) Sensory pollutants alter bird phenology and fitness across a continent. Nature 587, 605–609 6. McGlade, C.L.O. et al. (2023) Experimental light at night explains differences in activity onset between urban and forest great tits. Biol. Lett. 19, 20230194 7. Raap, T. et al. (2016) Early life exposure to artificial light at night affects the physiological condition: an experimental study on the ecophysiology of free-living nestling song- birds. Environ. Pollut. 218, 909–914 8. Holzhauer, S.I.J. et al. (2015) Out of the dark: establishing a large-scale field experiment to assess the effects of artificial light at night on species and food webs. Sustainability 7, 15593–15616 9. Spoelstra, K. et al. (2015) Experimental illumination of natural habitat: an experimental set-up to assess the direct and indi- rect ecological consequences of artificial light of different spectral composition. Philos. Trans. R. Soc. Lond. Ser. B Biol. Sci. 370, 20140129 10. Lawson, K. (2022) Electricity outages and residential fires: evidence from Cape Town, South Africa. S. Afr. J. Econ. 90, 469–485 11. Gaston, K.J. et al. (2015) The biological impacts of artificial light at night: the research challenge. Philos. Trans. R. Soc. Lond. Ser. B Biol. Sci. 370, 20140133 https://loadshed.theoutlier.co.za/ http://www.bio-logging.net/ https://youtu.be/BK0zOMFLnQs https://docs.google.com/spreadsheets/d/1ZpX_twP8sFBOAU6t--Vvh1pWMYSvs60UXINuD5n-K08/edit#gid=863218371 https://docs.google.com/spreadsheets/d/1ZpX_twP8sFBOAU6t--Vvh1pWMYSvs60UXINuD5n-K08/edit#gid=863218371 https://docs.google.com/spreadsheets/d/1ZpX_twP8sFBOAU6t--Vvh1pWMYSvs60UXINuD5n-K08/edit#gid=863218371 https://doi.org/10.1016/j.tree.2024.04.006 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0005 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0005 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0010 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0010 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0015 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0015 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0020 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0020 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0025 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0025 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0030 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0030 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0030 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0035 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0035 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0035 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0035 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0040 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0040 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0040 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0040 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0045 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0045 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0045 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0045 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0045 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0050 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0050 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0050 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0055 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0055 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0055 Image of &INS id= CellPress logo Trends in Ecology & Evolution 12. Nduhuura, P. et al. (2020) Mapping and spatial analysis of electricity load shedding experiences: a case study of com- munities in Accra, Ghana. Energies 13, 4280 13. Asase, A. et al. (2022) Replacing ‘parachute science’ with ‘global science’ in ecology and conservation biology. Conserv. Sci. Pract. 4, e517 Tre 14. Coetzee, B.W.T. et al. (2023) The impacts of artificial light at night in Africa: prospects for a research agenda. S. Afr. J. Sci. 119, 1–7 nds in Ecology & Evolution, July 2024, Vol. 39, No. 7 615 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0060 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0060 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0060 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0065 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0065 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0065 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0070 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0070 http://refhub.elsevier.com/S0169-5347(24)00088-0/rf0070 CellPress logo Investigating the impacts of artificial light via blackouts Understanding impacts of ALAN on animal behaviours South African load-shedding: what is it and how does it operate? Blackouts and load-shedding in South Africa: opportunities for research Research initiative to exploit this research opportunity Resources References