Assessing changes in nighttime lighting in the aftermath of the Turkey–Syria earthquake using SDGSAT-1 satellite data

On February 6, 2023, a powerful earthquake with a magnitude of 7.8 MW struck southeastern Turkey and northwestern Syria. The epicenter of the earthquake was in Gaziantep, Turkey. The earthquake was followed by a devastating aftershock 9 h later, which registered a magnitude of 7.5 MW and which had an epicenter located 100 km farther north in Kahramanmaras, Turkey. Thousands of subsequent tremors and aftershocks since the initial earthquake have further exacerbated the impacts on lives and infrastructure, resulting in more than 50,000 deaths and damage to hundreds of thousands of buildings in the region. The 2023 Turkey-Syria earthquake occurred near the intersection of the Anatolian, Arabian, and African plates. Turkey lies primarily on the Anatolian plate, which is being squeezed by the Arabian plate in the east and which is moving westward because the Eurasian plate obstructs its northward movement.1Reilinger R.E. McClusky S.C. Oral M.B. et al.Global Positioning System measurements of present-day crustal movements in the Arabia-Africa-Eurasia plate collision zone.J. Geophys. Res. 1997; 102: 9983-9999Crossref Google Scholar Moreover, the squeeze is amplified as the African plate moves beneath the Anatolian plate along the Cyprus and Hellenic Arc offshore.1Reilinger R.E. McClusky S.C. Oral M.B. et al.Global Positioning System measurements of present-day crustal movements in the Arabia-Africa-Eurasia plate collision zone.J. Geophys. Res. 1997; 102: 9983-9999Crossref Google Scholar These geological forces can result in a sudden release of pressure along tectonic boundaries, making Turkey particularly vulnerable to frequent earthquakes. With increasing accessibility to high-resolution remotely sensed images, the damage caused by earthquakes on human inhabitants can be more rapidly and intuitively identified and evaluated more objectively. Changes in nighttime light patterns in the aftermath of an earthquake represent a key indicator that is widely used to assess conditions on the ground that affect the livelihoods of survivors. In this study, we collected publicly released images of nighttime light across Turkey and Syria captured at a spatial resolution of 10 m by the SDGSAT-1 satellite from the International Research Center of Big Data for Sustainable Development Goals (CBAS) on August 22, 2022, and February 13, 2023, to map the pre- and post-disaster conditions, respectively, of the 2023 Turkey-Syria earthquake. Both images encompassed the same 21 Turkish provinces and nine Syrian provinces and were centered on Gaziantep and Kahramanmaras. For both images, we conducted a pixel-wise classification of lighted regions followed by noise removal using BM3D algorithms (Block Matched 3-Dimensional Collaborative filtering denoising algorithms)2Yahya A.A. Tan J. Su B. et al.BM3D image denoising algorithm based on an adaptive filtering.Multimed. Tool. Appl. 2020; 79: 20391-20427Crossref Scopus (21) Google Scholar to map the nighttime light levels in regions affected by the earthquake. We then used a change detection analysis of the classified images from the two dates to identify regions that displayed increased or decreased lighting patterns in the aftermath of the earthquake. Synthesizing data on peak ground acceleration (PGA) from the US Geological Survey, we found that nighttime lighting in Turkey after the earthquake largely vanished from regions that had experienced strong seismic shaking (PGA ≥ 0.5%g), such as the cities of Kirikhan and Antakya in Hatay Province (Figures 1A and 1B). This may have largely been a consequence of the damages caused to power facilities, the cancellation of outdoor entertainment events, and the diversion of power from non-essential lighting. However, we found that areas that had experienced weak seismic impacts (PGA < 0.5%g) such as Kahramanmaras City (Figure 1C), displayed increased levels of nighttime light after the earthquake. In Syria, nighttime lighting area increased substantially in many areas, such as in the cities of Aleppo and Ar-Raqqah (Figures 1D and 1E). This could have been attributed to the construction of temporary infrastructure to support active search and rescue operations in the area. Moreover, nighttime lighting area along urban roads decreased substantially. This indicated damage to transport infrastructure and suggested the presence of obstacles to effective access and mobility in these regions (Figure 1F). Specifically, an about 94% decrease in nighttime lighting area can be observed due to strong seismic shaking (PGA ≥ 0.5%g) (Figure 1G). This highlighted the weak earthquake-resistance capabilities of the power network infrastructure in the region. In contrast, 49% of the decrease in lighting area is observable in areas that witnessed weak seismic shaking (PGA < 0.5%g). Furthermore, in comparison with nighttime light conditions prior to the earthquake, areas that experienced a weak seismic impact displayed a 70% increase in nighttime lighting; this represented a 6% increase in nighttime lighting over areas that had experienced strong seismic shaking. Knowledge of these changes in nighttime lighting across different disaster areas would be crucial for optimizing rescue efforts as well as efforts to re-establish essential power supplies needed by survivors. We also noted that changes in the pattern of nighttime light varied significantly across different provinces (Figure 1H). Heavy damage inflicted by the earthquake caused the provinces of Hatay, Sanliurfa, and Kilis to display a 70% decrease in nighttime lighting area following the disaster. In these three provinces, the proportion of areas that had increased nighttime lighting after the earthquake was lower than that of areas that had decreased nighttime lighting. The proportion of areas with nighttime lighting increased least in the province of Kilis, where it had increased by only 28.8%. Apart from the three heavily damaged provinces discussed above, we found that in all other provinces, the proportion of the area where nighttime lighting had been created exceeded that where nighttime lights had vanished. This finding reflected the extensive efforts by local governments to secure and improve the livelihoods of survivors during the post-disaster relief operations. The provinces of Malatya, Kayseri, and Sivas experienced less damage from seismic shaking. In these provinces, the decrease in the proportion of total area displaying nighttime lighting was marginal (<10%). Rather, the same provinces displayed a 58% increase in the total proportion of lighted areas at nighttime in comparison with conditions prior to the earthquake. It is worth noting that the Syrian provinces of Aleppo and Ar-Raqqah established more areas with nighttime lighting. For instance, the area of nighttime lighting in Ar-Raqqah expanded by 138.38% after the earthquake. Images of nighttime light are generated based on the spectral responses of different radiative attributes of multiple ground objects. Such images have been used for detecting and mapping the human footprint over land surfaces, as well as for estimating the intensity of an activity. The method is based on the level of illumination detected, a measure that has been found to correlate strongly with economic activities.3Gibson J. Olivia S. Boe-Gibson G. Night lights in economics: sources and uses 1.J. Econ. Surv. 2020; 34: 955-980Crossref Scopus (36) Google Scholar Recent studies have been conducted using nighttime light images from the National Polar-Orbiting Partnership’s Visible Infrared Imaging Radiometer Suite (NPP-VIIRS) and the Operational Linescan System (OLS) from the Defense Meteorological Satellite Program (DMSP).4Zheng Q. Weng Q. Wang K. Developing a new cross-sensor calibration model for DMSP-OLS and Suomi-NPP VIIRS night-light imageries.ISPRS J. Photogrammetry Remote Sens. 2019; 153: 36-47Crossref Scopus (94) Google Scholar Such images have a coarser spatial resolution of 15 arcsec (about 500 m) and 30 arcsec (about 1 km), respectively; however, these are often inadequate for detecting detailed changes to lighted areas in the aftermath of emergency situations. The public release of nighttime light images captured at 10 m spatial resolution by the SDGSAT-1 satellite make it possible to detect and estimate the status of a disaster recovery effort as well as to estimate the ground capabilities at a finer scale. In the aftermath of an emergency, areas that continue to receive power (e.g., road networks and building blocks) can therefore be identified directly from SDGSAT-1 images. This will promote the continuous monitoring of disaster recovery and rescue capability at multiple scales, from the fine scale of a residential block to the national level. We therefore contend that the SDGSAT-1 data represent an important and novel data resource for enhancing the monitoring of disaster impacts as well as post-disaster rescue and rehabilitation. Changes in nighttime lighting patterns were evident in the SDGSAT-1 data. The exact causes of these changes could be further explored through the aid of remotely sensed images captured at a high spatial resolution in daytime, for instance, building damage,5Naddaf M. Turkey–Syria earthquake reveals building danger.Nature. 2023; 614: 398-399Crossref PubMed Scopus (6) Google Scholar road disruption, sheds constructed for rescue, and migration of disaster victims. To enhance emergency responses, a more comprehensive approach for assessing the damages caused by a disaster and for evaluating the success of a post-disaster recovery effort can likely be achieved by integrating SDGSAT-1 data on nighttime lighting with remotely sensed images captured in the daytime. This work was supported by the Strategic Priority Research Program of the Chinese Academy of Sciences (XDA19030101) and the National Key R&D Program of China (2022YFC3800701). The authors declare no competing interests.