Transcript
BOKU – Institute of Mountain Risk Engineering
Identification of alpine mass movements with i f infrasound d and d seismic i i signals. i l 1 2, Johannes Hübl Andreas Schimmel A d S hi l1,2 J h Hübl1 1 University of Natural Resources and Life Sciences, Vienna, Institute of Mountain Risk Engineering 2
corresponding corresponding author:
[email protected] author:
[email protected]
Introduction Sediment‐related hazards like landslides, debris flows, and debris floods are an increasing threat to people and property due to the last socio‐economic development of mountain areas and the climatic change. Monitoring debris flow torrents is essential for warning issues and gaining more knowledge about the processes. Debris flows and debris floods induce, by the collision of stones and by the friction of the flow to the pattern in the time‐frequency range for the infrasound and seismic signals, permits an channel, waves in the low‐frequency infrasonic spectrum and seismic waves. This fact that debris flows have a characteristically , q y p yp q y g g ,p early detection and identification of these events, even before a surge passes the sensor location. This work aims to create a reliable approach for detecting mass movement processes and identify and quantify them in respect to their process‐type, and –magnitude. There have already been several approaches to identify mass moment processes with infrasound or seismic signals, but this work aims to develop a method which uses a combination of both technologies and plans to define a common set of identification rules for e ent t pe and magnitude event‐type magnit de identification. identification
System Setup E Event t Detection D t ti and d Identification Id tifi ti Method M th d
Infrasound Sensor:
The event detection is based on the method developed at the FFG‐Project "Automatic Detection of Alpin Mass Movements (AMM‐Detection)” (Schimmel et al. 2015). The developed system is build up on a minimum of one seismic and one infrasound sensor which are co‐located and a microcontroller which runs a detection algorithm to detect debris flows and debris floods with high accuracy in real time directly on‐site. The developed detection algorithm analyses the evolution in time of the frequency content from the infrasonic and seismic mass movement signals Therefore different frequency bands are used to analyse the infrasound signal, signals. signal whereby a 5 to 15 Hz band characterises debris flows and a 15 to 30 Hz band is used for debris floods. The frequency bands below (manly dominated by wind) (and above) are used to eliminate false alarms. Further the variance of the amplitudes is used to eliminate artificial false alarms. For the main detection criteria the average amplitudes of the debris flow/ debris flood frequency bands has to exceed a certain limit for a specific time span. To distinguish between different event sizes, sizes two limits are used for the average amplitudes (Level 1 (L1) and Level 2 (L2)). Analyses with this method have shown different evaluation of the average amplitudes at the observed frequency bands for debris flows, debris floods and avalanches. So this approach can serve as a basis for the process identification. Detection principle depicted in a running spectrum of a debris flow Detection principle depicted in a running spectrum of a debris flow
• Chaparral Model 24 Sensitivity 2 V/Pa, frequency range 0,1‐100 Hz • MK‐224 MK 224 Sensitivity 50 mV/Pa, frequency range 3‐200 Hz • Electret Condenser Micophone KECG2742WBL‐25‐L Sensitivity ‐42±3 dB, frequency range ~20‐20000 Hz
Geophon:
• Luminary LM3S8962: 50 MHz ARM Cortex‐M3 Processor 4 ADC‐Channels – 100 Samples/s Data recording on MicroSD‐Card (16 GB: >4 months) Data recording on MicroSD‐Card (16 GB: >4 months) User‐Interface (display, keys), Ethernet (web server remote‐control, e‐mail alarms..)
• Sercel SG‐5 Sensitivity 80 V/m/s, natural frequency 5 Hz • Sensor Nl S Nl SM‐6/H‐A SM 6/H A Sensitivity 28 V/m/s, natural frequency 4,5 Hz
For the process and magnitude identification additional sensors like radar, ultrasonic sensors or video are used to get the required information for categorizing the signal and get a relation of the infrasound and seismic amplitudes to the discharge of the event. event Schimmel, A; Hübl, J. (2015): Automatic detection of debris flows and debris floods based on a combination of infrasound and seismic signals In: Landslides, online first, DOI: 10.1007/s10346‐015‐0640‐z
Example Events Debris flow at Lattenbach on 09.08.2015
Microcontroller:
Debris flood at Dristenau on 08.07.2015
This example shows the seismic and infrasound signal of a debris flow at Lattenbach and a debris flood at the test site Dristenau (Both test sites are located in Tyrol and operated by the Institute of Mountain Risk Engineering). The debris flow at Lattenbach was recorded on 09.08.2015 with a peak discharge of 64 m³/s, a total volume of 16000 m³ and a total duration of 1500 s. The max. infrasound amplitudes are up to 4 Pa and the max. seismic i i amplitudes li d are up to 4∙10 4 10‐33 m/s. / The Th event was detected by the detection algorithm at sec. 149 for level 1 and at 204 s for level 2. So the time between detection and passing g at the sensor site was ca. 50 s. of the of first surge The debris flood at Dristenau occurred on 08.07.2015 with a peak discharge of 3,5 m³/s and a duration of approximately 2400 s. The max. infrasound amplitudes are up to 600 mPa and th max. seismic i i amplitudes lit d are up to t 1∙10 1 10‐33 m/s. / The Th eventt the was detected by the algorithm at sec. 1149 for level 1 and at 1196 s for level 2. The frequency q y distribution in the total spectra p of the infra‐ sound signal reveals the difference between debris flow and debris flood signals. The infrasound signals of the debris flow at Lattenbach has its peak frequencies in the 5 to 15 Hz frequency band whereas for the debris flood event at Dristenau the peak frequencies range occur in the 15 to 30 Hz frequency band. (a) Infrasound time series; (b) Seismogram; (c) Running spectrum of the infrasound signal; (d) Running spectrum of the seismic signal; Signals are represented with a common base of time. Line: Time of first detection at Level 1 (e) Total infrasound spectrum of the events
Results Event Detection
Test Sites
This table shows an overview of the number of events and false alarms in the last three years. The event detections are split in smaller events (level 1 detections (L1), mostly small debris floods) and larger events (level 2 detections (L2)).
A long period of testing with high occurrence of different kind of events is necessary for the development of the identification method. Therefore this system is currently tested at five test sites in Austria, two in Italy and one in Switzerland:
Test Site
Year
Lattenbach
2013 2014 2015 2013 2014 2015 2013 2014 2015 2013 2014 2015 2013 2015 2015 2015 2015 Sum:
Dristenau
Farstrinne
Schüsserbach
Wartschenbach Illgraben Gadira Marderello
Detected events 1 2 3 18 7 12 0 2 1 3 0 2 5 1 6 1 3 67
Detections L1 Detections L2 1 2 0 14 7 9 0 0 0 2 0 2 5 1 2 1 2 48
D t ti Detections L1 L1 Detections L2 not classifiable d t ti detections not detected events False alarms
0 0 3 4 0 3 0 2 1 1 0 0 0 0 4 0 1 19
False alarms 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 1 3 6
not classifiable not detected detections events 1 0 0 0 0 1 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 9 0 0 0 1 0 0 0 1 2 14 4
All larger l events t could ld be b detected d t t d att every test t t site in the testing period and only four smaller events (level 1) could not be detected. Only six false alarms occurred in the last yyears and 14 detections could not be clearly classified as event (nine of them at Wartschenbach due to technical problems).
Lattenbach: Grins, Tyrol; Catchment area 5,3 km² Installation of the detection system since July 2012 Other equipment at the site: 3 ultrasonic sensors, weighing precipitation gauge, seismometer, 2 video cameras, 2D laser scanner, debris flow radar Illgraben: Leuk, Switzerland; Catchment area 9,5 km² Installation of the detection system since July 2015 Other equipment at the site: ultrasonic sensors, weighing precipitation gauge, geophones, 2 video cameras gauge, geophones, 2 video cameras
Schüsserbach: Radmer, Styria; Catchment area 0,7 km² Installation of the detection system since August 2013 Other equipment at the site: radar sensors, geophone, camera, weighing precipitation gauge Wartschenbach: Lienz, East Tyrol; Catchment area 2,6 km² Installation of the detection system since July 2013 Other equipment at the site: radar sensors, 2 weighing precipitation gauges, 2 video cameras Dristenau: Pertisau, Tyrol; Catchment area 9,97 km² Installation of the detection system since June 2013 Other equipment at the site: 2 radar sensors, video g gp p g g camera, weighing precipitation gauge Farstrinne: Umhausen, Tyrol; Catchment area 5,5 km² Installation of the detection system since July 2013 Other equipment at the site: debris flow radar, video camera Gadria: Alliz, Italy; Catchment area 6,3 km² Installation of the detection system since July 2015 Other equipment at the site: 2 ultrasonic sensors, weighing precipitation gauge, geophones, 3 video cameras Marderello: Novalesa, Italy; Catchment area 6,6 km² Installation of the detection system since July 2015 Other equipment at the site: ultrasonic sensor, weighing precipitation gauge, geophones, video camera precipitation gauge, geophones, video camera
