The short life of the volcanic island New Lateʻiki (Tonga) analyzed by multi-sensor remote sensing data

The study of volcanic island formation and lifespan is crucial for understanding volcanic processes and their impact on the environment. Submarine volcanic eruptions can create new islands, which may be short-lived or long-lasting. The longevity of these islands depends heavily on the composition of the eruptive materials. Long-lasting islands provide valuable opportunities to study ecological succession on virgin land, while ephemeral islands highlight the dynamic nature of volcanic processes and the rapid erosional forces of the ocean. This paper focuses on the October 2019 eruption at Late'iki Volcano, Tonga, which produced a new island, New Late'iki. Located on the Tonga Volcanic Arc, Late'iki has a history of eruptions, resulting in both ephemeral and relatively long-lived islands. This makes it an ideal location to investigate the factors controlling the lifespan of volcanic islands. The remote location of Late'iki and the often challenging conditions for in-situ observation make satellite-based earth observation a particularly powerful tool for monitoring such events. The study uses a combination of thermal (MODIS, VIIRS), optical (OLI, MSI), and SAR (Sentinel-1, TerraSAR-X) data to comprehensively analyze the eruption and the subsequent evolution of New Late'iki.

Several volcanic islands have formed in recent decades, each with varying lifespans. Surtsey, formed off Iceland in 1963, is a notable example of a long-lasting volcanic island. Other examples include Nishinoshima and Niijima (Japan) and Zubair (Yemen), which all persisted after their formation. In contrast, several islands formed by surtseyan eruptions have been short-lived, lasting only months or years. These include Kuwae (Vanuatu), Fukutoku-Okanoba (Japan), Kavachi (Solomon Islands), Metis Shoal and Home Reef (Tonga). Previous eruptions at Late'iki, formerly known as Metis Shoal, have produced islands with varying lifespans. Eruptions between 1851 and 1894, 1967-1968, and 1979 resulted in islands that quickly eroded, whereas the 1995 eruption created an island that lasted for over 25 years. The persistence of newly formed volcanic islands is relevant beyond purely scientific interest, also playing a role in territorial claims under the UN Convention on the Law of the Sea.

The study utilized a multi-sensor approach employing both high-spatial resolution and high-temporal resolution satellite data. High-resolution multispectral data from Sentinel-2 (24 datasets between September 5, 2019 and February 17, 2020) and Landsat-8 were used for detailed analysis of the island's morphology and evolution. This included visual inspection of various band combinations (Sentinel-2: 12/4/2; Landsat-8: 7/4/2) to identify features like water vapor plumes and changes in surface reflectance. The Normalized Hotspot Index (NHI) algorithm was applied to Sentinel-2 and Landsat-8 data to identify and quantify thermal anomalies. High-frequency, lower-resolution data from MODIS and VIIRS provided continuous monitoring of volcanic activity, especially in terms of volcanic radiant power (VRP) detection. MODIS bands 1/4/3 and VIIRS bands 11/M4/M3 were used for visual inspection, while thermal bands were used for hotspot detection. SAR data from Sentinel-1 (15 dual-pol images) and TerraSAR-X (one Spotlight and one High-Resolution Spotlight) provided complementary information, particularly regarding surface changes and volcanic activity regardless of cloud cover. Polarimetric SAR processing was applied to Sentinel-1 data, including the Wishart unsupervised classification, to analyze the backscatter characteristics of different surface types. Finally, seismic data from various global catalogs were consulted to determine if any seismic unrest coincided with the volcanic activity, though given the distance to the nearest station, seismic detection was not expected to be conclusive.

The analysis of multi-sensor data revealed a comprehensive picture of the Late'iki eruption and the evolution of New Lateʻiki. Sentinel-2 imagery showed a complete change in the area around the older Late'iki island beginning October 10, 2019, with the emergence of New Lateʻiki on October 15. The island reached a maximum area of approximately 21,000 m² on October 30, 2019, before rapidly eroding at a rate of approximately 464 m²/day. The erosion slowed after November 19, and by December 14, the island was completely submerged. The NHI analysis of Landsat-8 and Sentinel-2 data identified thermal anomalies on October 16 and 20, corroborating the eruption. MODIS and VIIRS detected volcanic activity from October 13-23, with a peak VRP of 49 MW. Sentinel-1 SAR data showed increased backscatter around the older island, starting October 17, consistent with the eruption and the emergence of New Lateʻiki. The Wishart classification of Sentinel-1 data effectively distinguished the area of strong backscatter from the surrounding ocean. No volcanic ash was observed, only a steam plume, in line with aviation advisory reports. Analysis of SO2 data did not show intensified emissions, but discoloration of the seawater suggested gas emissions continued for at least a year after the eruption. The area of the older island, significantly smaller before the 2019 eruption (≈400 m²), eroded at a much slower rate (≈8.4 m²/day) over the preceding 25 years. The much more rapid erosion of New Lateʻiki (464 m²/day) suggests it was composed of easily erodible pyroclastic debris, unlike the hardened lava of the 1995 island.

The results confirm the rapid and significant changes occurring at Late'iki during the October 2019 eruption. The stark contrast in erosion rates between the 1995 island and New Lateʻiki underscores the influence of eruptive material composition on island longevity. The effectiveness of multi-sensor satellite data in monitoring this remote eruption highlights its importance for volcanic monitoring, particularly in regions lacking extensive ground-based networks. The lack of significant seismic signals, despite the eruption, emphasizes the value of satellite monitoring in these circumstances. While the study successfully observed the eruption and erosion, potential limitations include the absence of nighttime acquisitions and the effects of solar irradiation on thermal anomaly measurements. Furthermore, the relatively sparse seismic data limit a thorough examination of the eruption's seismic precursors.

The study demonstrates the power of multi-sensor remote sensing in monitoring remote volcanic eruptions and island formation. The rapid erosion of New Lateʻiki contrasts sharply with the longevity of the 1995 island, revealing the crucial role of eruptive material composition in determining island lifespan. Future research could focus on refining thermal anomaly detection algorithms, incorporating additional data sources (e.g., gas measurements), and exploring the link between eruptive processes and island morphology.

The study's limitations include the lack of nighttime satellite imagery to mitigate the influence of solar irradiation on thermal measurements, and the sparse seismic data, hindering a thorough assessment of seismic precursors to the eruption. The absence of comprehensive gas data also restricts conclusions about the volcanic emissions' magnitude and composition.