Validation of an integrated satellite-data-driven response to an effusive crisis: the April–May 2018 eruption of Piton de la Fournaise

Andrew Harris,Lucia Gurioli,Magdalena Chevrel,Nicolas Villeneuve,Andrea Di Muro,Jean‐Luc Froger,Diego Coppola,Simon Thivet,Aline Peltier,Alexis Hrysiewicz,Philippe Kowalski,Massimiliano Favalli,Michael Ramsey

doi:10.4401/ag-7972

Abstract

Satellite-based surveillance of volcanic hot spots and plumes can be coupled with modeling to allow ensemble-based approaches to crisis response. We complete benchmark tests on an effusive crisis response protocol aimed at delivering product for use in tracking lava flows. The response involves integration of four models: MIROVA for discharge rate (TADR), the ASTER urgent response protocol for delivery of high-spatial resolution satellite data, DOWNFLOW for flow path projections, and PyFLOWGO for flow run-out. We test the protocol using the data feed available during Piton de la Fournaise’s April–May 2018 eruption, with product being delivered to the Observatoire du Piton de la Fournaise via Google Drive. The response was initialized by an alert at 19:50Z on 27 April 2018. Initially DOWNFLOW-FLOWGO were run using TADRs typical of Piton de la Fournaise, and revealed that flow at >120 m 3 /s could reach the island belt road. The first TADR (10–20 m 3 /s) was available at 09:55Z on 28 April, and gave flow run-outs of 1180–2510 m. The latency between satellite overpass and TADR provision was 105 minutes, with the model result being posted 15 minutes later. An InSAR image pair was completed two hours after the eruption began, and gave a flow length of 1.8 km; validating the run-out projection. Thereafter, run-outs were updated with each new TADR, and checked against flow lengths reported from InSAR and ASTER mapping. In all, 35 TADRs and four InSAR image pairs were processed during the 35-day-long eruption, and 11 ASTER images were delivered.

Highlights

Throughout the 1990’s and 2000’s methods were developed to extract lava flow discharge rates from 1 km spatial resolution satellite data collected by satellite sensors operating in the thermal infrared [e.g., Harris et al, 1997; 2007; Harris and Bologa, 2009; Coppola et al, 2010]
The MIROVA “watch” began at 20h30 (16h30 UTC) on 27 April, with an ASTER urgent response protocol (URP) being triggered at 04h25 (00h25 UTC) on 29 April (Appendix C)
On receipt of the Bulletin, DOWNFLOWGO was loaded with the most recent DEM of Piton de la Fournaise, this being the 5-m DEM generated from LiDAR data in 2010 modified by adding the largest flow fields in the area that are the October 2010 and the August 2015 using the InSAR-based thickness maps

Summary

Introduction

Throughout the 1990’s and 2000’s methods were developed to extract lava flow discharge rates from 1 km spatial resolution satellite data collected by satellite sensors operating in the thermal infrared [e.g., Harris et al, 1997; 2007; Harris and Bologa, 2009; Coppola et al, 2010]. FLOWGO has been initialized for and tested for lava channels at Piton de la Fournaise by Harris et al [2016] and Rhéty et al [2017], and - to allow improved model initialization, iteration and application - has been rewritten and rebuilt in Python as PyFLOWGO [Chevrel et al, 2018]. It is this version of FLOWGO that we use here

Methods

Results

Discussion

Conclusion