Prototype of a LPWA Network for real-time hydro-meteorological monitoring and flood nowcasting

Auteurs

A. Douinot, A. Dalla Torre, J. Martin, J.-F. Iffly, L. Rapin, C. Meisch, C. Bastian, and L. Pfister

Référence

in International Conference on Ad-Hoc Networks and Wireless (ADHOC-NOW 2019): Ad-Hoc, Mobile, and Wireless Networks, Luxembourg, 1-3 October, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11803 LNCS, p. 566-574, ISBN: 978-303031830-7, 2019

Description

Over the past decade, Luxembourg has experienced an increase in extreme precipitation events during the summer season. As a direct consequence of this evolution, increasingly flashy hydrological responses have been observed in the river network – eventually leading to several flash flood events. In an area that has been historically prone to large scale winter floods, the current flood forecasting and monitoring chain is not adapted to this specific type of events, characterized by small spatial extents (<200 km2), short time scales (<6 h) and high magnitudes. The SIGFOX network deployed in Luxembourg since 2016 provides a unique opportunity to leverage the potential for a Low-Power Wide-Area Network (LPWAN) to implement and operate a hydro-meteorological monitoring system of unprecedented spatial and temporal density. The restricted message size (12 bytes) supported by this technology is suited for the small data sets that are to be transmitted from each monitored hydro-meteorological site. In addition, the inter-connection of the different sensing devices allows for an immediate adaptation of the data acquisition and transmission frequencies – according to rapidly changing hydro-meteorological states. The entire network is able to switch from a sleep and low power consumption mode to a warning mode, with high frequency recordings and transmissions. The initial limit of 140 daily transmissions is eventually compensated for by a systematic compression of the numeric data on the one hand and the switch to cellular transmission during flash flood warning mode. Here we present the prototype implemented in the Ernz Blanche catchment (102 km2), which was successively exposed to flash flood events in 2016 and 2018. The overall set-up consists of 4 raingauges, 4 streamgauges, and 4 soil moisture sensors. The first three months of operation of the prototype monitoring and transmission system on the Ernz Blanche catchment are assessed for validating the data transmission, acquisition, and emergency warning system.

Lien

doi:10.1007/978-3-030-31831-4_40