WEBVTT NOTE This file was exported by MacCaption version 7.0.06 to comply with the WebVTT specification dated March 27, 2017. 00:00:09.910 --> 00:00:13.280 align:center line:-1 position:50% size:40% Okay, folks, I can see we just hit one minute after the hour. 00:00:13.280 --> 00:00:15.282 align:center line:-1 position:50% size:37% My name is Matt Thomas, and thank you for tuning in 00:00:15.282 --> 00:00:17.784 align:center line:-1 position:50% size:39% to the USGS Landslide Hazards Seminar. 00:00:17.784 --> 00:00:20.120 align:center line:-1 position:50% size:44% This meeting is hosted by the Landslide Hazards Program, 00:00:20.120 --> 00:00:21.655 align:center line:-1 position:50% size:24% and co-organized with contributions 00:00:21.655 --> 00:00:24.191 align:center line:-1 position:50% size:33% from Stephen Slaughter and Jaime Kostelnik. 00:00:24.191 --> 00:00:25.826 align:center line:-1 position:50% size:39% For those of you that are new to this meeting, 00:00:25.826 --> 00:00:27.160 align:center line:-1 position:50% size:27% you have the ability to submit questions 00:00:27.160 --> 00:00:29.396 align:center line:-1 position:50% size:39% via the chat window or use the Raise Your Hand feature 00:00:29.396 --> 00:00:31.865 align:center line:-1 position:50% size:40% in combination with your microphone or video camera. 00:00:31.865 --> 00:00:33.600 align:center line:-1 position:50% size:42% We're gonna wait until the end of today's presentation 00:00:33.600 --> 00:00:34.534 align:center line:-1 position:50% size:25% to take questions, 00:00:34.534 --> 00:00:36.270 align:center line:-1 position:50% size:33% so in the meantime please just do your best 00:00:36.270 --> 00:00:38.138 align:center line:-1 position:50% size:36% to make sure your microphone is muted 00:00:38.138 --> 00:00:40.774 align:center line:-1 position:50% size:39% and your video camera is off 00:00:40.774 --> 00:00:42.609 align:center line:-1 position:50% size:35% when you're not intending to speak. 00:00:42.609 --> 00:00:44.444 align:center line:-1 position:50% size:33% Today we're joined by Sebastian Uhlemann 00:00:44.444 --> 00:00:47.214 align:center line:-1 position:50% size:34% from Lawrence Berkeley National Laboratory. 00:00:47.214 --> 00:00:49.683 align:center line:-1 position:50% size:30% Sebastian is currently a freelancing scientist 00:00:49.683 --> 00:00:51.385 align:center line:-1 position:50% size:40% based in Potsdam, Germany, 00:00:51.385 --> 00:00:54.187 align:center line:-1 position:50% size:45% and a guest scientist at the Lawrence Berkeley National Lab. 00:00:54.187 --> 00:00:56.290 align:center line:-1 position:50% size:39% He's also a visiting professor at the Amrita Center 00:00:56.290 --> 00:00:59.559 align:center line:-1 position:50% size:34% for Wireless Networks and Applications in India. 00:00:59.559 --> 00:01:02.829 align:center line:-1 position:50% size:32% In 2018, Sebastian obtained a PhD degree 00:01:02.829 --> 00:01:05.465 align:center line:-1 position:50% size:44% in Applied Geophysics from ETH Zürich in Switzerland, 00:01:05.465 --> 00:01:07.868 align:center line:-1 position:50% size:30% where he focused (inaudible) monitoring 00:01:07.868 --> 00:01:09.569 align:center line:-1 position:50% size:40% of moisture-driven processes 00:01:09.569 --> 00:01:11.638 align:center line:-1 position:50% size:32% in natural and engineered slopes. 00:01:11.638 --> 00:01:13.840 align:center line:-1 position:50% size:27% From 2012 to 2018, 00:01:13.840 --> 00:01:16.109 align:center line:-1 position:50% size:40% Sebastian worked at the British Geological Survey 00:01:16.109 --> 00:01:17.844 align:center line:-1 position:50% size:37% as a research geophysicist 00:01:17.844 --> 00:01:19.579 align:center line:-1 position:50% size:31% and next to monitoring landslides, 00:01:19.579 --> 00:01:20.714 align:center line:-1 position:50% size:23% also working on the development 00:01:20.714 --> 00:01:22.716 align:center line:-1 position:50% size:32% of near-surface geophysical techniques 00:01:22.716 --> 00:01:24.384 align:center line:-1 position:50% size:41% for resource and groundwater assessment 00:01:24.384 --> 00:01:27.854 align:center line:-1 position:50% size:37% in England, Europe, Southeast Asia, and Africa. 00:01:27.854 --> 00:01:31.024 align:center line:-1 position:50% size:45% In 2018, he joined Lawrence Berkeley National Lab, 00:01:31.024 --> 00:01:32.326 align:center line:-1 position:50% size:35% first initially as a post-doc 00:01:32.326 --> 00:01:34.227 align:center line:-1 position:50% size:32% and then later as a research scientist. 00:01:34.227 --> 00:01:35.595 align:center line:-1 position:50% size:27% And he's worked on various projects 00:01:35.595 --> 00:01:37.664 align:center line:-1 position:50% size:30% related to the impacts of climate change. 00:01:37.664 --> 00:01:39.032 align:center line:-1 position:50% size:30% Last year, he received 00:01:39.032 --> 00:01:42.502 align:center line:-1 position:50% size:47% an AGU Near-Surface Geophysics Early Career Achievement Award 00:01:42.502 --> 00:01:45.639 align:center line:-1 position:50% size:33% for novel advances in integrated monitoring 00:01:45.639 --> 00:01:47.541 align:center line:-1 position:50% size:40% and modeling of geophysical, hydrological, 00:01:47.541 --> 00:01:50.777 align:center line:-1 position:50% size:42% and geomechanical processes in the subsurface environment. 00:01:50.777 --> 00:01:53.613 align:center line:-1 position:50% size:35% And I also want to congratulate Sebastian 00:01:53.613 --> 00:01:56.616 align:center line:-1 position:50% size:32% for choosing to join us at 10:00 in the evening 00:01:56.616 --> 00:01:58.685 align:center line:-1 position:50% size:27% Germany time for today's seminar. 00:01:58.685 --> 00:02:00.520 align:center line:-1 position:50% size:34% So, Sebastian, I believe that makes you 00:02:00.520 --> 00:02:04.691 align:center line:-1 position:50% size:41% our first international speaker, so thank you very much. 00:02:04.691 --> 00:02:06.226 align:center line:-1 position:50% size:26% (Sebastian laughs) 00:02:06.226 --> 00:02:08.028 align:center line:-1 position:50% size:26% Thanks a lot, Matt. 00:02:08.028 --> 00:02:11.298 align:center line:-1 position:50% size:27% And actually it's 11, 11:00 p.m., so, 00:02:11.298 --> 00:02:14.001 align:center line:-1 position:50% size:36% yeah, it's even a little later, but that's good. 00:02:14.001 --> 00:02:16.603 align:center line:-1 position:50% size:43% So, yeah, thank you very much for the introduction, 00:02:16.603 --> 00:02:22.242 align:center line:-1 position:50% size:41% and also for--just for inviting me to give this seminar today. 00:02:22.242 --> 00:02:24.077 align:center line:-1 position:50% size:37% So, today what I would like to do is, 00:02:24.077 --> 00:02:26.246 align:center line:-1 position:50% size:31% I would like to give you a little overview 00:02:26.246 --> 00:02:27.781 align:center line:-1 position:50% size:30% of the work that we've been doing 00:02:27.781 --> 00:02:31.651 align:center line:-1 position:50% size:36% in the last couple of years or so here at Berkeley Lab 00:02:31.651 --> 00:02:34.521 align:center line:-1 position:50% size:43% with regards to characterization and monitoring of landslides 00:02:34.521 --> 00:02:36.723 align:center line:-1 position:50% size:25% using geophysics, remote sensing, 00:02:36.723 --> 00:02:38.892 align:center line:-1 position:50% size:42% and wireless sensor networks. 00:02:38.892 --> 00:02:40.994 align:center line:-1 position:50% size:39% And I also want to show you a little bit how we-- 00:02:40.994 --> 00:02:43.697 align:center line:-1 position:50% size:37% how we may use that for landslide early warning. 00:02:45.532 --> 00:02:48.168 align:center line:-1 position:50% size:40% And this is basically a project that, of course, you know, 00:02:48.168 --> 00:02:49.970 align:center line:-1 position:50% size:24% has a few people that work on it. 00:02:49.970 --> 00:02:54.107 align:center line:-1 position:50% size:34% So it's mostly myself, but also Sylvain Fiolleau, 00:02:54.107 --> 00:02:56.777 align:center line:-1 position:50% size:27% who is the post-doc working on that; 00:02:56.777 --> 00:02:58.111 align:center line:-1 position:50% size:36% Stijn Wielandt, who's an electric engineer 00:02:58.111 --> 00:03:01.081 align:center line:-1 position:50% size:45% who did most of the development of the sensor network 00:03:01.081 --> 00:03:02.315 align:center line:-1 position:50% size:28% we're going to show; 00:03:02.315 --> 00:03:03.483 align:center line:-1 position:50% size:29% and Baptiste Dafflon, 00:03:03.483 --> 00:03:07.687 align:center line:-1 position:50% size:43% who is also a chief physicist who is working with me on that. 00:03:10.624 --> 00:03:12.626 align:center line:-1 position:50% size:21% So, (inaudible). 00:03:15.796 --> 00:03:17.130 align:center line:-1 position:50% size:35% Yeah, Lawrence Berkeley National Lab 00:03:17.130 --> 00:03:20.367 align:center line:-1 position:50% size:34% is a part of Energy Research Lab, 00:03:20.367 --> 00:03:23.036 align:center line:-1 position:50% size:22% and so the work that we're doing 00:03:23.036 --> 00:03:25.872 align:center line:-1 position:50% size:24% is not necessarily hazard related. 00:03:25.872 --> 00:03:28.875 align:center line:-1 position:50% size:31% So this project really frames this work 00:03:28.875 --> 00:03:31.445 align:center line:-1 position:50% size:36% as actually, you know, not focused on landslides, 00:03:31.445 --> 00:03:35.615 align:center line:-1 position:50% size:41% but is more generally focused on stability of earthworks 00:03:35.615 --> 00:03:38.418 align:center line:-1 position:50% size:35% that either form or protect critical infrastructure. 00:03:38.418 --> 00:03:40.053 align:center line:-1 position:50% size:29% So it's all about critical infrastructure. 00:03:40.053 --> 00:03:43.723 align:center line:-1 position:50% size:44% That's the way how we formed-- how we developed this project. 00:03:43.723 --> 00:03:46.893 align:center line:-1 position:50% size:42% And critical infrastructure itself are the assets, networks, 00:03:46.893 --> 00:03:49.863 align:center line:-1 position:50% size:35% and systems that provide the essential services 00:03:49.863 --> 00:03:51.832 align:center line:-1 position:50% size:29% that underpin society. 00:03:51.832 --> 00:03:53.633 align:center line:-1 position:50% size:35% And these are, you know, 00:03:53.633 --> 00:03:55.802 align:center line:-1 position:50% size:31% the things that are vital to our safety, 00:03:55.802 --> 00:03:57.971 align:center line:-1 position:50% size:36% prosperity, and well-being. 00:03:57.971 --> 00:03:59.940 align:center line:-1 position:50% size:42% But now a lot of infrastructures actually are in poor condition, 00:03:59.940 --> 00:04:01.108 align:center line:-1 position:50% size:32% and you can see this when you drive around, 00:04:01.108 --> 00:04:02.209 align:center line:-1 position:50% size:40% for instance, at the Bay Area, 00:04:02.209 --> 00:04:06.780 align:center line:-1 position:50% size:35% you can see bridges that are pretty (inaudible), 00:04:06.780 --> 00:04:07.881 align:center line:-1 position:50% size:40% you know, poorly maintained. 00:04:07.881 --> 00:04:09.850 align:center line:-1 position:50% size:44% You can see some underpasses that are falling apart. 00:04:09.850 --> 00:04:11.418 align:center line:-1 position:50% size:41% So, really what we need to do 00:04:11.418 --> 00:04:14.154 align:center line:-1 position:50% size:43% is we need to maintain and manage this infrastructure, 00:04:14.154 --> 00:04:15.489 align:center line:-1 position:50% size:34% but this comes at a cost. 00:04:15.489 --> 00:04:17.324 align:center line:-1 position:50% size:45% And here are just some numbers that, for instance, 00:04:17.324 --> 00:04:19.326 align:center line:-1 position:50% size:28% by 2040, in the U.S., 00:04:19.326 --> 00:04:21.628 align:center line:-1 position:50% size:30% we will have to spend $12.4 trillion, 00:04:21.628 --> 00:04:24.397 align:center line:-1 position:50% size:34% and worldwide this will be like $4 trillion, 00:04:24.397 --> 00:04:26.099 align:center line:-1 position:50% size:33% so really a lot of money. 00:04:26.099 --> 00:04:28.235 align:center line:-1 position:50% size:30% And now in the future, this infrastructure 00:04:28.235 --> 00:04:31.738 align:center line:-1 position:50% size:36% is increasingly making use of connected technology 00:04:31.738 --> 00:04:34.141 align:center line:-1 position:50% size:24% to really integrate energy systems, 00:04:34.141 --> 00:04:35.509 align:center line:-1 position:50% size:32% buildings, and industry, 00:04:35.509 --> 00:04:36.843 align:center line:-1 position:50% size:36% and basically to adapt and evolve 00:04:36.843 --> 00:04:39.146 align:center line:-1 position:50% size:41% with the way we work and live 00:04:39.146 --> 00:04:41.448 align:center line:-1 position:50% size:31% and what we do with this infrastructure. 00:04:41.448 --> 00:04:43.216 align:center line:-1 position:50% size:35% And this is also becoming increasingly important 00:04:43.216 --> 00:04:46.219 align:center line:-1 position:50% size:35% for earthen infrastructure, like, for instance, 00:04:46.219 --> 00:04:48.922 align:center line:-1 position:50% size:28% levees or, you know, embankments, 00:04:48.922 --> 00:04:52.492 align:center line:-1 position:50% size:38% and also just the slopes that surround infrastructure. 00:04:52.492 --> 00:04:53.660 align:center line:-1 position:50% size:42% Where we see more and more 00:04:53.660 --> 00:04:55.128 align:center line:-1 position:50% size:36% that we have wireless sensor networks, 00:04:55.128 --> 00:04:57.931 align:center line:-1 position:50% size:40% we have digital twins and other integrated systems 00:04:57.931 --> 00:04:59.766 align:center line:-1 position:50% size:33% and networks that are increasingly being used 00:04:59.766 --> 00:05:01.735 align:center line:-1 position:50% size:31% for landslide mitigation 00:05:01.735 --> 00:05:03.303 align:center line:-1 position:50% size:34% but also for landslide early warnings. 00:05:03.303 --> 00:05:05.939 align:center line:-1 position:50% size:35% So I'm going to show you a little bit of this today, 00:05:05.939 --> 00:05:08.975 align:center line:-1 position:50% size:35% or I'm hoping to show you a little bit of this today. 00:05:08.975 --> 00:05:11.778 align:center line:-1 position:50% size:24% So in this project, 00:05:11.778 --> 00:05:15.182 align:center line:-1 position:50% size:33% we are trying to develop a novel approach 00:05:15.182 --> 00:05:18.318 align:center line:-1 position:50% size:35% to provide a robust and reliable early warning 00:05:18.318 --> 00:05:22.889 align:center line:-1 position:50% size:39% of slope failure by combining various sensor technologies, 00:05:22.889 --> 00:05:25.492 align:center line:-1 position:50% size:39% which range from, you know, 00:05:25.492 --> 00:05:29.462 align:center line:-1 position:50% size:44% low-cost geophysical monitoring but also environmental sensing, 00:05:29.462 --> 00:05:31.298 align:center line:-1 position:50% size:24% and combine this with hydrological 00:05:31.298 --> 00:05:32.432 align:center line:-1 position:50% size:40% and geomechanical modeling 00:05:32.432 --> 00:05:37.103 align:center line:-1 position:50% size:37% to basically try and provide effective safety of slopes 00:05:37.103 --> 00:05:38.738 align:center line:-1 position:50% size:16% in real time. 00:05:38.738 --> 00:05:42.275 align:center line:-1 position:50% size:43% But also by using these models and feeding them with, 00:05:42.275 --> 00:05:45.278 align:center line:-1 position:50% size:42% for instance, weather forecasts or, you know, 00:05:45.278 --> 00:05:47.447 align:center line:-1 position:50% size:33% machine learning-based prediction techniques, 00:05:47.447 --> 00:05:49.749 align:center line:-1 position:50% size:33% we also want to be able to forecast 00:05:49.749 --> 00:05:53.019 align:center line:-1 position:50% size:33% what the effect of safety is doing in the future. 00:05:53.019 --> 00:05:55.956 align:center line:-1 position:50% size:21% And this work really builds on 00:05:55.956 --> 00:05:57.791 align:center line:-1 position:50% size:23% recent advances in various fields, 00:05:57.791 --> 00:05:59.326 align:center line:-1 position:50% size:30% particularly in sensing technologies, 00:05:59.326 --> 00:06:00.160 align:center line:-1 position:50% size:31% where, in recent years, 00:06:00.160 --> 00:06:01.528 align:center line:-1 position:50% size:32% geophysical monitoring has seen 00:06:01.528 --> 00:06:04.064 align:center line:-1 position:50% size:37% a number of developments that enable us 00:06:04.064 --> 00:06:06.600 align:center line:-1 position:50% size:43% to perform long-term monitoring of landslides 00:06:06.600 --> 00:06:07.867 align:center line:-1 position:50% size:30% to provide actionable information 00:06:07.867 --> 00:06:11.271 align:center line:-1 position:50% size:37% like volumetric 3D changes on moisture content 00:06:11.271 --> 00:06:13.373 align:center line:-1 position:50% size:33% that we can record before, during, and after 00:06:13.373 --> 00:06:15.609 align:center line:-1 position:50% size:30% landslide movements. 00:06:15.609 --> 00:06:22.649 align:center line:-1 position:50% size:45% And before I go to the work that we've done at Berkeley Lab, 00:06:22.649 --> 00:06:24.584 align:center line:-1 position:50% size:27% I would like to focus a little bit 00:06:24.584 --> 00:06:27.287 align:center line:-1 position:50% size:32% on these developments that are specifically 00:06:27.287 --> 00:06:28.989 align:center line:-1 position:50% size:33% on low-cost geophysical monitoring, 00:06:28.989 --> 00:06:30.790 align:center line:-1 position:50% size:29% because also, as Matt already said, 00:06:30.790 --> 00:06:31.825 align:center line:-1 position:50% size:42% my background is geophysics, 00:06:31.825 --> 00:06:32.826 align:center line:-1 position:50% size:34% so I know this pretty well 00:06:32.826 --> 00:06:33.760 align:center line:-1 position:50% size:32% and I thought I'd take this opportunity 00:06:33.760 --> 00:06:35.829 align:center line:-1 position:50% size:28% to also show you a little bit about that. 00:06:35.829 --> 00:06:37.697 align:center line:-1 position:50% size:42% Hey, Sebastian, before you make the transition 00:06:37.697 --> 00:06:38.598 align:center line:-1 position:50% size:25% to your next slide, 00:06:38.598 --> 00:06:41.768 align:center line:-1 position:50% size:41% I just recommend turning your video camera off. 00:06:41.768 --> 00:06:46.273 align:center line:-1 position:50% size:28% I will try and do that. 00:06:48.975 --> 00:06:50.043 align:center line:-1 position:50% size:10% Better? 00:06:50.043 --> 00:06:51.311 align:center line:-1 position:50% size:23% Yup, keep going. 00:06:51.311 --> 00:06:53.146 align:center line:-1 position:50% size:19% Sounds good. 00:06:53.146 --> 00:06:54.547 align:center line:-1 position:50% size:16% Cool, yeah. 00:06:54.547 --> 00:06:57.284 align:center line:-1 position:50% size:29% So, yeah, this is work that I've done before, 00:06:57.284 --> 00:06:58.451 align:center line:-1 position:50% size:27% actually, (inaudible) 00:06:58.451 --> 00:07:00.387 align:center line:-1 position:50% size:40% when I was at the British Geological Survey. 00:07:00.387 --> 00:07:01.955 align:center line:-1 position:50% size:33% And there I was working in a group 00:07:01.955 --> 00:07:04.524 align:center line:-1 position:50% size:31% that was developing geoelectric monitoring, 00:07:04.524 --> 00:07:06.359 align:center line:-1 position:50% size:34% and my task there was basically to develop 00:07:06.359 --> 00:07:07.927 align:center line:-1 position:50% size:41% sound techniques that enable 00:07:07.927 --> 00:07:10.764 align:center line:-1 position:50% size:34% long-term geoelectrical monitoring of landslides. 00:07:10.764 --> 00:07:12.365 align:center line:-1 position:50% size:34% And these developments have been done 00:07:12.365 --> 00:07:15.268 align:center line:-1 position:50% size:32% at the Hollin Hill Landslide Observatory. 00:07:15.268 --> 00:07:17.003 align:center line:-1 position:50% size:42% This is a landslide observatory 00:07:17.003 --> 00:07:19.706 align:center line:-1 position:50% size:38% that's located in North Yorkshire in England, 00:07:19.706 --> 00:07:23.009 align:center line:-1 position:50% size:34% and it's sitting on a slope that sits within rocks 00:07:23.009 --> 00:07:25.879 align:center line:-1 position:50% size:30% of the Lias group of the Lower Jurassic, 00:07:25.879 --> 00:07:28.448 align:center line:-1 position:50% size:24% where we have Whitby Mudstone 00:07:28.448 --> 00:07:30.216 align:center line:-1 position:50% size:38% failing over slate sandstone formations. 00:07:30.216 --> 00:07:33.653 align:center line:-1 position:50% size:42% So, Whitby Mudstone is a very weak, very clay material 00:07:33.653 --> 00:07:36.089 align:center line:-1 position:50% size:36% that fails, then, over this kind of sandstone material 00:07:36.089 --> 00:07:39.292 align:center line:-1 position:50% size:41% that's also not really a nice, competent sandstone, 00:07:39.292 --> 00:07:43.963 align:center line:-1 position:50% size:33% but is more like a weak, silty, sandy material. 00:07:43.963 --> 00:07:47.300 align:center line:-1 position:50% size:40% But still we get, yeah, these failures in these areas. 00:07:47.300 --> 00:07:48.735 align:center line:-1 position:50% size:35% And actually, geologically, 00:07:48.735 --> 00:07:50.937 align:center line:-1 position:50% size:26% this is pretty typical in England. 00:07:50.937 --> 00:07:52.672 align:center line:-1 position:50% size:27% So in a lot of places 00:07:52.672 --> 00:07:54.407 align:center line:-1 position:50% size:33% where Whitby Mudstone is outcropping, 00:07:54.407 --> 00:07:55.942 align:center line:-1 position:50% size:28% that's where you get a lot of landslides. 00:07:55.942 --> 00:07:58.745 align:center line:-1 position:50% size:40% In this map here, we can see the Whitby Mudstone here 00:07:58.745 --> 00:08:00.647 align:center line:-1 position:50% size:41% failing, basically, over the slate sandstone formation 00:08:00.647 --> 00:08:02.982 align:center line:-1 position:50% size:31% in this kind of somewhat lighter gray. 00:08:02.982 --> 00:08:05.719 align:center line:-1 position:50% size:27% The landslide itself is reasonably small 00:08:05.719 --> 00:08:08.855 align:center line:-1 position:50% size:34% and has a slope gradient of only 12 degrees, 00:08:08.855 --> 00:08:11.991 align:center line:-1 position:50% size:35% which already shows you how weak this material is. 00:08:11.991 --> 00:08:14.928 align:center line:-1 position:50% size:37% (inaudible) also shows you how unstable the slope is. 00:08:14.928 --> 00:08:16.763 align:center line:-1 position:50% size:42% So it's a slow-moving landslide 00:08:16.763 --> 00:08:19.432 align:center line:-1 position:50% size:23% with movements that are usually 00:08:19.432 --> 00:08:21.134 align:center line:-1 position:50% size:24% less than a meter per year or so, 00:08:21.134 --> 00:08:22.635 align:center line:-1 position:50% size:43% where translational movements 00:08:22.635 --> 00:08:25.004 align:center line:-1 position:50% size:26% that are happening in this area here 00:08:25.004 --> 00:08:26.873 align:center line:-1 position:50% size:25% are forming these earth lobes, 00:08:26.873 --> 00:08:28.942 align:center line:-1 position:50% size:42% and these translational movements that we have here 00:08:28.942 --> 00:08:30.844 align:center line:-1 position:50% size:33% actually are taking away the support 00:08:30.844 --> 00:08:33.012 align:center line:-1 position:50% size:36% of some secondary slopes on the back. 00:08:33.012 --> 00:08:35.482 align:center line:-1 position:50% size:33% Right then you can see some rotational failures. 00:08:35.482 --> 00:08:37.384 align:center line:-1 position:50% size:35% So this shows you that this landslide system 00:08:37.384 --> 00:08:38.251 align:center line:-1 position:50% size:23% is quite complex, 00:08:38.251 --> 00:08:40.754 align:center line:-1 position:50% size:33% where you have translational movement, 00:08:40.754 --> 00:08:41.955 align:center line:-1 position:50% size:38% you have rotational failures, 00:08:41.955 --> 00:08:44.758 align:center line:-1 position:50% size:31% and you have the formation of lobes. 00:08:44.758 --> 00:08:45.925 align:center line:-1 position:50% size:23% Now, at this site, 00:08:45.925 --> 00:08:48.395 align:center line:-1 position:50% size:28% we used it as kind of a development site, 00:08:48.395 --> 00:08:51.030 align:center line:-1 position:50% size:26% we have a lot of instrumentation. 00:08:51.030 --> 00:08:53.633 align:center line:-1 position:50% size:33% There are GPS markers that we use 00:08:53.633 --> 00:08:57.837 align:center line:-1 position:50% size:35% to monitor the formation every month or two or so. 00:08:57.837 --> 00:09:00.073 align:center line:-1 position:50% size:38% We have turf meters that measure the formation. 00:09:00.073 --> 00:09:03.877 align:center line:-1 position:50% size:42% But we have also piezometers, some moisture sensors, 00:09:03.877 --> 00:09:06.246 align:center line:-1 position:50% size:36% and these type of sensors that are measuring, 00:09:06.246 --> 00:09:08.148 align:center line:-1 position:50% size:32% basically, the hydrology of the system. 00:09:08.148 --> 00:09:10.884 align:center line:-1 position:50% size:28% And then we have, outlined in red here, 00:09:10.884 --> 00:09:13.052 align:center line:-1 position:50% size:30% an electrical resistivity tomography system 00:09:13.052 --> 00:09:16.089 align:center line:-1 position:50% size:37% that was measuring the resistivity of the ground 00:09:16.089 --> 00:09:17.357 align:center line:-1 position:50% size:22% every other day. 00:09:17.357 --> 00:09:19.926 align:center line:-1 position:50% size:30% And this is something that we can then use 00:09:19.926 --> 00:09:22.228 align:center line:-1 position:50% size:38% to actually track moisture movements in 3D, 00:09:22.228 --> 00:09:24.731 align:center line:-1 position:50% size:26% and I will show this in a second. 00:09:24.731 --> 00:09:28.501 align:center line:-1 position:50% size:30% But before that, I thought I'd show you 00:09:28.501 --> 00:09:30.703 align:center line:-1 position:50% size:28% how we characterize the landslide 00:09:30.703 --> 00:09:32.238 align:center line:-1 position:50% size:32% using a different geophysical technique. 00:09:32.238 --> 00:09:36.342 align:center line:-1 position:50% size:45% In this case, P- and S-wave seismic refraction measurements 00:09:36.342 --> 00:09:39.412 align:center line:-1 position:50% size:39% that provide us with the P- and S-wave velocities 00:09:39.412 --> 00:09:41.014 align:center line:-1 position:50% size:25% or P- and S-wave velocity variations 00:09:41.014 --> 00:09:42.348 align:center line:-1 position:50% size:25% in the subsurface. 00:09:42.348 --> 00:09:46.419 align:center line:-1 position:50% size:37% Now, for those of you who are not geophysicists, 00:09:46.419 --> 00:09:47.887 align:center line:-1 position:50% size:29% the way you do these kind of surveys 00:09:47.887 --> 00:09:51.491 align:center line:-1 position:50% size:28% is you take basically a sledgehammer 00:09:51.491 --> 00:09:54.961 align:center line:-1 position:50% size:39% and you hit a metal plate, or in this case a metal prism, 00:09:54.961 --> 00:09:56.663 align:center line:-1 position:50% size:39% and by doing so you're exciting some waves, 00:09:56.663 --> 00:09:58.164 align:center line:-1 position:50% size:27% so they're traveling through the ground. 00:09:58.164 --> 00:09:59.666 align:center line:-1 position:50% size:25% And every time you get a contrast 00:09:59.666 --> 00:10:03.069 align:center line:-1 position:50% size:28% is seismic velocities, at the right angles, 00:10:03.069 --> 00:10:04.671 align:center line:-1 position:50% size:34% you get a refracted wave 00:10:04.671 --> 00:10:06.706 align:center line:-1 position:50% size:33% that's basically traveling along this interface, 00:10:06.706 --> 00:10:08.241 align:center line:-1 position:50% size:26% and while traveling along this interface 00:10:08.241 --> 00:10:10.510 align:center line:-1 position:50% size:25% it's kind of exciting additional waves 00:10:10.510 --> 00:10:13.046 align:center line:-1 position:50% size:31% that you then measure using geophones. 00:10:13.046 --> 00:10:15.515 align:center line:-1 position:50% size:41% In this case, these geophones are shown here. 00:10:15.515 --> 00:10:17.484 align:center line:-1 position:50% size:40% And we use three component geophones 00:10:17.484 --> 00:10:21.221 align:center line:-1 position:50% size:40% because that enables us to not only look at the P-waves, 00:10:21.221 --> 00:10:23.022 align:center line:-1 position:50% size:37% or the pressure wave that's going through the ground, 00:10:23.022 --> 00:10:24.891 align:center line:-1 position:50% size:34% but also the shear wave. 00:10:24.891 --> 00:10:26.826 align:center line:-1 position:50% size:29% And having both is actually pretty nice 00:10:26.826 --> 00:10:29.863 align:center line:-1 position:50% size:37% because the P-wave is sensitive to, for instance, 00:10:29.863 --> 00:10:33.500 align:center line:-1 position:50% size:34% variations in fluid content in the ground, 00:10:33.500 --> 00:10:34.934 align:center line:-1 position:50% size:34% which the S-wave is not. 00:10:34.934 --> 00:10:36.703 align:center line:-1 position:50% size:30% So by having the two, 00:10:36.703 --> 00:10:40.006 align:center line:-1 position:50% size:26% you can also make some assessment 00:10:40.006 --> 00:10:43.443 align:center line:-1 position:50% size:46% of where or where you don't have any moisture in the ground. 00:10:43.443 --> 00:10:44.577 align:center line:-1 position:50% size:42% And we're doing these surveys 00:10:44.577 --> 00:10:46.513 align:center line:-1 position:50% size:30% to try and get a good understanding 00:10:46.513 --> 00:10:49.282 align:center line:-1 position:50% size:42% on where the interface between the Whitby Mudstone 00:10:49.282 --> 00:10:51.351 align:center line:-1 position:50% size:33% and the slate sandstone formation is. 00:10:51.351 --> 00:10:54.287 align:center line:-1 position:50% size:40% Now if you look at the results that are shown here, 00:10:54.287 --> 00:10:56.523 align:center line:-1 position:50% size:35% you can actually see that, you know, 00:10:56.523 --> 00:10:58.024 align:center line:-1 position:50% size:37% we don't see this interface. 00:10:58.024 --> 00:11:00.627 align:center line:-1 position:50% size:43% We were expecting an interface to be somewhere around here, 00:11:00.627 --> 00:11:03.696 align:center line:-1 position:50% size:42% but as you see, there's nothing that you see in these images 00:11:03.696 --> 00:11:06.733 align:center line:-1 position:50% size:35% that indicate any interface of two geological units. 00:11:06.733 --> 00:11:11.404 align:center line:-1 position:50% size:45% What we do see, though, is some high-velocity anomalies, 00:11:11.404 --> 00:11:13.106 align:center line:-1 position:50% size:29% especially down here and up there, 00:11:13.106 --> 00:11:14.741 align:center line:-1 position:50% size:41% which we align to the regional table 00:11:14.741 --> 00:11:16.075 align:center line:-1 position:50% size:31% that you can see here, 00:11:16.075 --> 00:11:18.711 align:center line:-1 position:50% size:38% and the perched drum table that you can see up here. 00:11:18.711 --> 00:11:20.380 align:center line:-1 position:50% size:30% Now, if you look at the S-wave velocities, 00:11:20.380 --> 00:11:21.648 align:center line:-1 position:50% size:28% really, the only thing you can see 00:11:21.648 --> 00:11:23.650 align:center line:-1 position:50% size:43% is kind of a weathering horizon, 00:11:23.650 --> 00:11:25.852 align:center line:-1 position:50% size:34% so the deeper you go the higher the velocities, 00:11:25.852 --> 00:11:28.588 align:center line:-1 position:50% size:40% so the shallower material is just much more weathered 00:11:28.588 --> 00:11:31.291 align:center line:-1 position:50% size:32% than the other material. 00:11:31.291 --> 00:11:32.358 align:center line:-1 position:50% size:37% But as I said, we had both, 00:11:32.358 --> 00:11:34.727 align:center line:-1 position:50% size:45% we had P- and S-wave velocities in this case, 00:11:34.727 --> 00:11:35.628 align:center line:-1 position:50% size:40% and what we can do with that 00:11:35.628 --> 00:11:38.464 align:center line:-1 position:50% size:30% is we can calculate a ratio between them. 00:11:38.464 --> 00:11:40.533 align:center line:-1 position:50% size:35% In this case, we calculate the Poisson's ratio 00:11:40.533 --> 00:11:43.269 align:center line:-1 position:50% size:26% because that's an elastic property 00:11:43.269 --> 00:11:45.204 align:center line:-1 position:50% size:32% that's very well defined. 00:11:45.204 --> 00:11:46.739 align:center line:-1 position:50% size:36% And to keep it very simple, 00:11:46.739 --> 00:11:48.474 align:center line:-1 position:50% size:40% if the Poisson's ratio is small, 00:11:48.474 --> 00:11:49.909 align:center line:-1 position:50% size:33% so in this case if it's red, 00:11:49.909 --> 00:11:52.245 align:center line:-1 position:50% size:30% and this means that the ground is dry. 00:11:52.245 --> 00:11:53.813 align:center line:-1 position:50% size:34% And if the Poisson's ratio is high, 00:11:53.813 --> 00:11:55.381 align:center line:-1 position:50% size:25% so if it goes to 0.5, 00:11:55.381 --> 00:11:58.484 align:center line:-1 position:50% size:37% that means that the ground is fully saturated. 00:11:58.484 --> 00:12:00.720 align:center line:-1 position:50% size:38% And now suddenly we could really see this interface 00:12:00.720 --> 00:12:01.955 align:center line:-1 position:50% size:43% between the Whitby Mudstone, 00:12:01.955 --> 00:12:03.656 align:center line:-1 position:50% size:37% which is the saturated blue material 00:12:03.656 --> 00:12:04.891 align:center line:-1 position:50% size:30% that we can see here, 00:12:04.891 --> 00:12:08.494 align:center line:-1 position:50% size:46% and then underneath that we saw this kind of dry material, 00:12:08.494 --> 00:12:10.296 align:center line:-1 position:50% size:42% which is the slate sandstone formation, 00:12:10.296 --> 00:12:12.165 align:center line:-1 position:50% size:37% which is the formation that is much better drained 00:12:12.165 --> 00:12:14.267 align:center line:-1 position:50% size:39% and has much less moisture, 00:12:14.267 --> 00:12:16.102 align:center line:-1 position:50% size:34% especially in the top part of it. 00:12:16.102 --> 00:12:17.804 align:center line:-1 position:50% size:37% And if we look at this in 3D, 00:12:17.804 --> 00:12:19.906 align:center line:-1 position:50% size:35% we can see that we could really nicely track 00:12:19.906 --> 00:12:22.442 align:center line:-1 position:50% size:38% this interface across space, 00:12:22.442 --> 00:12:24.944 align:center line:-1 position:50% size:31% and we could basically have a good idea 00:12:24.944 --> 00:12:27.280 align:center line:-1 position:50% size:28% on what the lithology at the site looks like. 00:12:28.648 --> 00:12:29.716 align:center line:-1 position:50% size:30% And then what we did 00:12:29.716 --> 00:12:32.785 align:center line:-1 position:50% size:31% is we also did resistivity tomography. 00:12:32.785 --> 00:12:34.420 align:center line:-1 position:50% size:32% And the reason for that 00:12:34.420 --> 00:12:36.522 align:center line:-1 position:50% size:27% is that the resistivity of the ground 00:12:36.522 --> 00:12:38.024 align:center line:-1 position:50% size:28% is actually a very nice indicator 00:12:38.024 --> 00:12:42.729 align:center line:-1 position:50% size:36% of first the lithology but also the water content 00:12:42.729 --> 00:12:45.431 align:center line:-1 position:50% size:31% and how water content may change over time, 00:12:45.431 --> 00:12:47.533 align:center line:-1 position:50% size:35% and also the temperature of the ground. 00:12:47.533 --> 00:12:49.636 align:center line:-1 position:50% size:44% So, especially the water content is something 00:12:49.636 --> 00:12:51.871 align:center line:-1 position:50% size:34% that for landslide studies is super interesting 00:12:51.871 --> 00:12:53.506 align:center line:-1 position:50% size:34% because if you can track how water is moving 00:12:53.506 --> 00:12:55.608 align:center line:-1 position:50% size:40% in 3D through your landslide, 00:12:55.608 --> 00:12:57.477 align:center line:-1 position:50% size:27% that tells you a lot on what's going on. 00:12:57.477 --> 00:12:58.578 align:center line:-1 position:50% size:29% And the way we do these measurements 00:12:58.578 --> 00:13:01.514 align:center line:-1 position:50% size:43% is we basically have electrodes in the ground, 00:13:01.514 --> 00:13:03.650 align:center line:-1 position:50% size:29% and we--in this case, 00:13:03.650 --> 00:13:08.287 align:center line:-1 position:50% size:36% we have 196 electrodes distributed across the site, 00:13:08.287 --> 00:13:11.958 align:center line:-1 position:50% size:45% and then we do measurements between four of them every time, 00:13:11.958 --> 00:13:14.494 align:center line:-1 position:50% size:26% we just swap them so that we can try 00:13:14.494 --> 00:13:17.196 align:center line:-1 position:50% size:32% and sample the ground as densely as possible, 00:13:17.196 --> 00:13:20.600 align:center line:-1 position:50% size:33% as kind of was indicated in this schematic here 00:13:20.600 --> 00:13:23.102 align:center line:-1 position:50% size:41% where depending on how you choose these four electrodes, 00:13:23.102 --> 00:13:25.571 align:center line:-1 position:50% size:32% you sample the ground at different locations. 00:13:25.571 --> 00:13:28.875 align:center line:-1 position:50% size:26% And then you need to do an inversion 00:13:28.875 --> 00:13:32.078 align:center line:-1 position:50% size:34% to basically recover what the resistivity distribution 00:13:32.078 --> 00:13:33.946 align:center line:-1 position:50% size:22% in 3D looks like. 00:13:33.946 --> 00:13:35.515 align:center line:-1 position:50% size:20% Now, as I said, 00:13:35.515 --> 00:13:37.583 align:center line:-1 position:50% size:34% the resistivity depends on the lithology, 00:13:37.583 --> 00:13:39.118 align:center line:-1 position:50% size:28% and we can see this very nicely here 00:13:39.118 --> 00:13:41.254 align:center line:-1 position:50% size:41% where we can see very nicely 00:13:41.254 --> 00:13:42.855 align:center line:-1 position:50% size:44% where we have the Whitby Mudstone formation, 00:13:42.855 --> 00:13:44.590 align:center line:-1 position:50% size:42% because that's where we have, basically, 00:13:44.590 --> 00:13:45.725 align:center line:-1 position:50% size:30% these low resistivities, 00:13:45.725 --> 00:13:47.493 align:center line:-1 position:50% size:43% and that's mainly because the Whitby Mudstone formation 00:13:47.493 --> 00:13:49.262 align:center line:-1 position:50% size:37% is very high in clay content, 00:13:49.262 --> 00:13:51.297 align:center line:-1 position:50% size:25% has a lot of clay, has a lot of water, 00:13:51.297 --> 00:13:52.999 align:center line:-1 position:50% size:28% and so the resistivity is very low. 00:13:52.999 --> 00:13:55.001 align:center line:-1 position:50% size:42% And underneath that, we have the slate sandstone formation, 00:13:55.001 --> 00:13:56.469 align:center line:-1 position:50% size:27% which has less clay, less water, 00:13:56.469 --> 00:13:59.639 align:center line:-1 position:50% size:37% and has a higher resistivity. 00:13:59.639 --> 00:14:02.208 align:center line:-1 position:50% size:43% Now the nice thing about these resistivity measurements 00:14:02.208 --> 00:14:04.410 align:center line:-1 position:50% size:35% are that if we repeat them over time, 00:14:04.410 --> 00:14:07.480 align:center line:-1 position:50% size:42% we can assume that the lithology's not going to change 00:14:07.480 --> 00:14:09.482 align:center line:-1 position:50% size:30% but that the only thing that's going to change 00:14:09.482 --> 00:14:11.918 align:center line:-1 position:50% size:29% is the water content and the temperature. 00:14:11.918 --> 00:14:14.253 align:center line:-1 position:50% size:45% Now the temperature we actually measured at the site, 00:14:14.253 --> 00:14:17.690 align:center line:-1 position:50% size:38% and so we know how the temperature varies 00:14:17.690 --> 00:14:19.892 align:center line:-1 position:50% size:19% across space, 00:14:19.892 --> 00:14:21.360 align:center line:-1 position:50% size:37% well, how it varies laterally, 00:14:21.360 --> 00:14:23.663 align:center line:-1 position:50% size:29% but also how it varies with depth. 00:14:23.663 --> 00:14:25.264 align:center line:-1 position:50% size:36% So we can correct for that. 00:14:25.264 --> 00:14:26.432 align:center line:-1 position:50% size:28% Now, the other thing that we can do 00:14:26.432 --> 00:14:29.168 align:center line:-1 position:50% size:32% is we can take samples from the site 00:14:29.168 --> 00:14:31.804 align:center line:-1 position:50% size:42% and analyze them with regards to their moisture content. 00:14:31.804 --> 00:14:33.873 align:center line:-1 position:50% size:44% So we did this, we took samples from the Whitby Mudstone, 00:14:33.873 --> 00:14:35.007 align:center line:-1 position:50% size:42% the slate sandstone formation. 00:14:35.007 --> 00:14:36.642 align:center line:-1 position:50% size:28% We took them to lab and measured 00:14:36.642 --> 00:14:39.912 align:center line:-1 position:50% size:41% how the resistivity is changing at different moisture contents. 00:14:39.912 --> 00:14:42.982 align:center line:-1 position:50% size:38% And having this basically allows us to create a model 00:14:42.982 --> 00:14:45.084 align:center line:-1 position:50% size:34% that we can then use to translate the resistivity 00:14:45.084 --> 00:14:48.988 align:center line:-1 position:50% size:32% that we measure into a moisture content. 00:14:48.988 --> 00:14:50.490 align:center line:-1 position:50% size:42% And now we can see this here, 00:14:50.490 --> 00:14:52.391 align:center line:-1 position:50% size:44% and we can see again that we have very high moisture content 00:14:52.391 --> 00:14:53.926 align:center line:-1 position:50% size:33% in the Whitby Mudstone formation, 00:14:53.926 --> 00:14:55.194 align:center line:-1 position:50% size:41% we have low moisture content 00:14:55.194 --> 00:14:56.796 align:center line:-1 position:50% size:29% in the slate sandstone formation. 00:14:56.796 --> 00:14:59.365 align:center line:-1 position:50% size:28% But we can also see a lot of detail where, 00:14:59.365 --> 00:15:01.934 align:center line:-1 position:50% size:34% for instance, here we can see an area 00:15:01.934 --> 00:15:05.171 align:center line:-1 position:50% size:26% of slightly elevated moisture content, 00:15:05.171 --> 00:15:06.139 align:center line:-1 position:50% size:35% which is basically an area 00:15:06.139 --> 00:15:07.440 align:center line:-1 position:50% size:27% where every time we went to the field 00:15:07.440 --> 00:15:09.509 align:center line:-1 position:50% size:39% we can see--or we could see some water coming out 00:15:09.509 --> 00:15:12.945 align:center line:-1 position:50% size:37% of just the head of the lobe, 00:15:12.945 --> 00:15:16.749 align:center line:-1 position:50% size:39% and so we saw a little spring that was there. 00:15:16.749 --> 00:15:19.752 align:center line:-1 position:50% size:37% So in this video that I'm going to show now, 00:15:19.752 --> 00:15:22.088 align:center line:-1 position:50% size:37% I'm basically going to show a time series 00:15:22.088 --> 00:15:24.423 align:center line:-1 position:50% size:39% across the monitoring period 00:15:24.423 --> 00:15:26.526 align:center line:-1 position:50% size:31% that I analyzed a couple of years ago. 00:15:26.526 --> 00:15:27.593 align:center line:-1 position:50% size:31% So first we see, again, 00:15:27.593 --> 00:15:29.896 align:center line:-1 position:50% size:36% how the Whitby Mudstone formation 00:15:29.896 --> 00:15:32.632 align:center line:-1 position:50% size:28% in through here is kind of riding over 00:15:32.632 --> 00:15:34.033 align:center line:-1 position:50% size:41% the slate sandstone formation in red. 00:15:34.033 --> 00:15:35.902 align:center line:-1 position:50% size:36% We can see how the lobes are forming. 00:15:35.902 --> 00:15:37.503 align:center line:-1 position:50% size:34% And now in the following, 00:15:37.503 --> 00:15:39.005 align:center line:-1 position:50% size:32% what I'm going to show is just the changes 00:15:39.005 --> 00:15:40.673 align:center line:-1 position:50% size:26% in moisture content that we calculated, 00:15:40.673 --> 00:15:42.208 align:center line:-1 position:50% size:32% just to give you an idea 00:15:42.208 --> 00:15:46.179 align:center line:-1 position:50% size:38% of how we can track moisture movements in 3D. 00:15:46.179 --> 00:15:47.313 align:center line:-1 position:50% size:30% And basically we start in winter, 00:15:47.313 --> 00:15:49.549 align:center line:-1 position:50% size:38% where we could see high moisture content here. 00:15:49.549 --> 00:15:51.684 align:center line:-1 position:50% size:30% We go then further into a summer month, 00:15:51.684 --> 00:15:56.422 align:center line:-1 position:50% size:41% where we can now see how the water is basically draining 00:15:56.422 --> 00:15:59.158 align:center line:-1 position:50% size:23% in this lower part of the slope. 00:15:59.158 --> 00:16:00.993 align:center line:-1 position:50% size:43% And basically this water that is draining here in summer 00:16:00.993 --> 00:16:03.029 align:center line:-1 position:50% size:40% is what accumulates in winter up here. 00:16:03.029 --> 00:16:05.498 align:center line:-1 position:50% size:37% And we can see again now in summer, 00:16:05.498 --> 00:16:07.967 align:center line:-1 position:50% size:29% these patterns that will appear here. 00:16:07.967 --> 00:16:11.504 align:center line:-1 position:50% size:40% So that's something that is happening every year. 00:16:11.504 --> 00:16:14.006 align:center line:-1 position:50% size:31% So now we are getting into the year of 2012, 00:16:14.006 --> 00:16:17.877 align:center line:-1 position:50% size:35% which is when this slope reactivated this landslide. 00:16:17.877 --> 00:16:19.245 align:center line:-1 position:50% size:27% And this was a year where we had 00:16:19.245 --> 00:16:20.613 align:center line:-1 position:50% size:31% an exceptional amount of rainfall. 00:16:20.613 --> 00:16:22.248 align:center line:-1 position:50% size:32% So already the summer is now much wetter 00:16:22.248 --> 00:16:23.749 align:center line:-1 position:50% size:29% than the year before, 00:16:23.749 --> 00:16:24.817 align:center line:-1 position:50% size:39% but now if you look at winter, 00:16:24.817 --> 00:16:26.385 align:center line:-1 position:50% size:32% it was really, really wet. 00:16:26.385 --> 00:16:28.754 align:center line:-1 position:50% size:42% And that's where we had some rotational failures in the back, 00:16:28.754 --> 00:16:32.258 align:center line:-1 position:50% size:37% but we also had a massive translational movement 00:16:32.258 --> 00:16:35.895 align:center line:-1 position:50% size:25% of these lobes that formed there. 00:16:35.895 --> 00:16:37.563 align:center line:-1 position:50% size:24% So, unfortunately, I went back, 00:16:37.563 --> 00:16:41.534 align:center line:-1 position:50% size:33% but here I have just a few of those plots 00:16:41.534 --> 00:16:43.302 align:center line:-1 position:50% size:25% to show you again the drainage. 00:16:43.302 --> 00:16:45.805 align:center line:-1 position:50% size:37% So we have this water that accumulates in winter here 00:16:45.805 --> 00:16:46.806 align:center line:-1 position:50% size:28% and then drains in the summer here, 00:16:46.806 --> 00:16:49.041 align:center line:-1 position:50% size:38% so that's something that we can see every year. 00:16:49.041 --> 00:16:50.409 align:center line:-1 position:50% size:42% But what was more interesting 00:16:50.409 --> 00:16:52.778 align:center line:-1 position:50% size:36% was actually what we saw during December 2012 00:16:52.778 --> 00:16:54.313 align:center line:-1 position:50% size:42% when we had this reactivation. 00:16:54.313 --> 00:16:56.649 align:center line:-1 position:50% size:32% Because what we were expecting to see 00:16:56.649 --> 00:16:59.452 align:center line:-1 position:50% size:31% was a lot of movement on this side, 00:16:59.452 --> 00:17:00.653 align:center line:-1 position:50% size:27% so one of the lobes, 00:17:00.653 --> 00:17:04.590 align:center line:-1 position:50% size:38% because when the landslide failed before in 2009, 00:17:04.590 --> 00:17:07.927 align:center line:-1 position:50% size:32% that's where we saw most of the movement. 00:17:07.927 --> 00:17:11.731 align:center line:-1 position:50% size:37% Now in 2012, actually, this lobe didn't move at all. 00:17:11.731 --> 00:17:13.999 align:center line:-1 position:50% size:28% It basically remained completely stable. 00:17:13.999 --> 00:17:16.535 align:center line:-1 position:50% size:27% And what we found is that we had 00:17:16.535 --> 00:17:20.640 align:center line:-1 position:50% size:32% this interesting pattern of moisture movement, 00:17:20.640 --> 00:17:22.842 align:center line:-1 position:50% size:41% which indicates to us that we had some preferential flow 00:17:22.842 --> 00:17:28.114 align:center line:-1 position:50% size:46% where we had some relatively big fractures in the surface 00:17:28.114 --> 00:17:31.083 align:center line:-1 position:50% size:42% that allowed water to infiltrate relatively deep into the ground, 00:17:31.083 --> 00:17:33.085 align:center line:-1 position:50% size:29% and then underneath this lobe here 00:17:33.085 --> 00:17:34.487 align:center line:-1 position:50% size:40% you have some aeolian sand, 00:17:34.487 --> 00:17:38.090 align:center line:-1 position:50% size:36% so there were really nice preferential flow pathways 00:17:38.090 --> 00:17:40.026 align:center line:-1 position:50% size:25% and basically took the water away 00:17:40.026 --> 00:17:41.727 align:center line:-1 position:50% size:29% from the slip surface, 00:17:41.727 --> 00:17:43.229 align:center line:-1 position:50% size:38% produced the pore pressure of the slip surface, 00:17:43.229 --> 00:17:45.031 align:center line:-1 position:50% size:30% and actually stabilized the slope. 00:17:45.031 --> 00:17:46.165 align:center line:-1 position:50% size:28% So this is something we didn't have 00:17:46.165 --> 00:17:47.833 align:center line:-1 position:50% size:28% on the other side, on the eastern lobe, 00:17:47.833 --> 00:17:52.004 align:center line:-1 position:50% size:44% where we see this accumulation of moisture on the top here. 00:17:52.004 --> 00:17:54.540 align:center line:-1 position:50% size:45% And this is basically a site where we got about three meters 00:17:54.540 --> 00:17:58.778 align:center line:-1 position:50% size:39% of movement in just less than two months. 00:17:58.778 --> 00:18:00.112 align:center line:-1 position:50% size:35% So there was a lot of movement on this side, 00:18:00.112 --> 00:18:03.015 align:center line:-1 position:50% size:44% which was basically triggered by this additional load of moisture 00:18:03.015 --> 00:18:06.152 align:center line:-1 position:50% size:38% that we managed or that we imaged up there. 00:18:06.152 --> 00:18:08.421 align:center line:-1 position:50% size:31% So this just shows you that geophysics 00:18:08.421 --> 00:18:09.855 align:center line:-1 position:50% size:23% can really reveal some processes 00:18:09.855 --> 00:18:11.157 align:center line:-1 position:50% size:30% that control landslide movements, 00:18:11.157 --> 00:18:13.993 align:center line:-1 position:50% size:33% and can do this in 3D or even 4D in this case. 00:18:15.294 --> 00:18:17.096 align:center line:-1 position:50% size:41% So then we were also looking how we could use this 00:18:17.096 --> 00:18:21.200 align:center line:-1 position:50% size:34% potentially as, you know, 00:18:21.200 --> 00:18:23.269 align:center line:-1 position:50% size:37% for early warning purposes, let's say. 00:18:23.269 --> 00:18:25.972 align:center line:-1 position:50% size:32% And for this, I extracted the moisture content 00:18:25.972 --> 00:18:28.808 align:center line:-1 position:50% size:27% at the top part of this eastern lobe. 00:18:28.808 --> 00:18:30.876 align:center line:-1 position:50% size:45% And what we can see is basically some seasonal effects. 00:18:30.876 --> 00:18:33.112 align:center line:-1 position:50% size:44% We can see that we have low moisture content in summer, 00:18:33.112 --> 00:18:35.081 align:center line:-1 position:50% size:42% we have high moisture content in winter. 00:18:35.081 --> 00:18:37.616 align:center line:-1 position:50% size:42% And in general over the period that we analyzed, 00:18:37.616 --> 00:18:41.654 align:center line:-1 position:50% size:41% the moisture content basically increased over time. 00:18:41.654 --> 00:18:46.692 align:center line:-1 position:50% size:43% This also agrees with the water levels that we measured, 00:18:46.692 --> 00:18:51.163 align:center line:-1 position:50% size:35% which also show some sort of seasonal behavior, 00:18:51.163 --> 00:18:54.433 align:center line:-1 position:50% size:40% but also clearly show this rise over the couple of years 00:18:54.433 --> 00:18:57.403 align:center line:-1 position:50% size:31% or few years that we monitored this. 00:18:57.403 --> 00:18:58.471 align:center line:-1 position:50% size:28% On the bottom here, 00:18:58.471 --> 00:19:01.907 align:center line:-1 position:50% size:35% what we see is a displacement record. 00:19:01.907 --> 00:19:04.710 align:center line:-1 position:50% size:37% So first we see that we had a little bit of creep, 00:19:04.710 --> 00:19:06.145 align:center line:-1 position:50% size:43% there was not much movement, 00:19:06.145 --> 00:19:09.515 align:center line:-1 position:50% size:36% but then as soon as we hit the winter of 2012, 2013, 00:19:09.515 --> 00:19:12.118 align:center line:-1 position:50% size:38% we see really that the slope started to accelerate 00:19:12.118 --> 00:19:13.552 align:center line:-1 position:50% size:38% and had a lot of movement. 00:19:13.552 --> 00:19:17.023 align:center line:-1 position:50% size:44% Now if we relate this to the measured moisture content, 00:19:17.023 --> 00:19:20.226 align:center line:-1 position:50% size:31% we can actually derive a threshold. 00:19:20.226 --> 00:19:21.594 align:center line:-1 position:50% size:17% In this case, 00:19:21.594 --> 00:19:24.630 align:center line:-1 position:50% size:45% we found that if we have more than 48% moisture content 00:19:24.630 --> 00:19:25.698 align:center line:-1 position:50% size:31% in this part of the lobe, 00:19:25.698 --> 00:19:26.832 align:center line:-1 position:50% size:25% and so just above the eastern lobe, 00:19:26.832 --> 00:19:28.601 align:center line:-1 position:50% size:30% that's where we have to expect movements. 00:19:28.601 --> 00:19:31.537 align:center line:-1 position:50% size:35% And now if I would extend this plot into the future, 00:19:31.537 --> 00:19:32.371 align:center line:-1 position:50% size:28% and actually we have all the data 00:19:32.371 --> 00:19:34.106 align:center line:-1 position:50% size:28% because the system is still monitoring, 00:19:34.106 --> 00:19:36.075 align:center line:-1 position:50% size:37% every time we exceed 48% 00:19:36.075 --> 00:19:38.110 align:center line:-1 position:50% size:35% is actually really where we see movement, 00:19:38.110 --> 00:19:42.214 align:center line:-1 position:50% size:42% so this threshold proved better than the last few years. 00:19:42.214 --> 00:19:45.117 align:center line:-1 position:50% size:39% So this was just to show you what you can do 00:19:45.117 --> 00:19:49.055 align:center line:-1 position:50% size:39% with geophysical monitoring, let's say. 00:19:49.055 --> 00:19:50.356 align:center line:-1 position:50% size:37% The problem that you have with that 00:19:50.356 --> 00:19:52.558 align:center line:-1 position:50% size:37% is that it's very site specific. 00:19:52.558 --> 00:19:53.893 align:center line:-1 position:50% size:27% So that's something that you can do 00:19:53.893 --> 00:19:56.228 align:center line:-1 position:50% size:38% on a place where you know you have a problem, right? 00:19:56.228 --> 00:19:58.631 align:center line:-1 position:50% size:40% Where the problem is severe 00:19:58.631 --> 00:20:00.132 align:center line:-1 position:50% size:31% and you want to understand it better. 00:20:00.132 --> 00:20:03.169 align:center line:-1 position:50% size:40% That's actually also how the British Geological Survey 00:20:03.169 --> 00:20:05.404 align:center line:-1 position:50% size:36% is using these techniques. 00:20:05.404 --> 00:20:06.872 align:center line:-1 position:50% size:34% They're working together quite a lot 00:20:06.872 --> 00:20:08.107 align:center line:-1 position:50% size:24% with, for instance, Network Rail, 00:20:08.107 --> 00:20:11.410 align:center line:-1 position:50% size:39% which is the railway provider in the UK, 00:20:11.410 --> 00:20:16.348 align:center line:-1 position:50% size:42% which often asks them to install these kind of systems 00:20:16.348 --> 00:20:19.051 align:center line:-1 position:50% size:43% at areas where they really have big problems and instabilities, 00:20:19.051 --> 00:20:20.519 align:center line:-1 position:50% size:35% but they really don't know what's triggering 00:20:20.519 --> 00:20:22.054 align:center line:-1 position:50% size:28% the sort of instability, for instance, 00:20:22.054 --> 00:20:24.223 align:center line:-1 position:50% size:33% and so they're using this as a measure 00:20:24.223 --> 00:20:26.358 align:center line:-1 position:50% size:40% to really image what's going on in the ground 00:20:26.358 --> 00:20:29.061 align:center line:-1 position:50% size:37% to then design some mitigation measures. 00:20:29.061 --> 00:20:31.397 align:center line:-1 position:50% size:39% But now what we want to do is we wanted to bring this 00:20:31.397 --> 00:20:33.365 align:center line:-1 position:50% size:22% a little bit further and upscale it, 00:20:33.365 --> 00:20:35.101 align:center line:-1 position:50% size:27% and that, of course, is not possible 00:20:35.101 --> 00:20:36.769 align:center line:-1 position:50% size:43% by just using geophysics alone. 00:20:36.769 --> 00:20:38.137 align:center line:-1 position:50% size:26% You also need to combine it, then, 00:20:38.137 --> 00:20:39.839 align:center line:-1 position:50% size:36% with some environmental monitoring, 00:20:39.839 --> 00:20:41.140 align:center line:-1 position:50% size:31% some remote sensing, 00:20:41.140 --> 00:20:42.808 align:center line:-1 position:50% size:32% and then, as I already mentioned, 00:20:42.808 --> 00:20:45.244 align:center line:-1 position:50% size:40% combining it with hydrological and geomechanical modeling 00:20:45.244 --> 00:20:47.680 align:center line:-1 position:50% size:24% to get to basically a prediction 00:20:47.680 --> 00:20:50.316 align:center line:-1 position:50% size:24% of factor of safety for an entire site. 00:20:50.316 --> 00:20:53.719 align:center line:-1 position:50% size:39% And that's something that we try to do at Berkeley Lab, 00:20:53.719 --> 00:20:57.156 align:center line:-1 position:50% size:31% because Berkeley Lab is located in Berkeley, 00:20:57.156 --> 00:20:58.691 align:center line:-1 position:50% size:43% in the San Francisco Bay Area, 00:20:58.691 --> 00:21:00.292 align:center line:-1 position:50% size:33% where the Berkeley Hills and the Oakland Hills 00:21:00.292 --> 00:21:02.995 align:center line:-1 position:50% size:28% really have a history of landslides, 00:21:02.995 --> 00:21:05.931 align:center line:-1 position:50% size:33% there are many deep-seated landslides, 00:21:05.931 --> 00:21:08.834 align:center line:-1 position:50% size:33% which, obviously, shown in this map here. 00:21:08.834 --> 00:21:10.469 align:center line:-1 position:50% size:36% Most of them are dormant, 00:21:10.469 --> 00:21:13.405 align:center line:-1 position:50% size:40% but there are also quite a few active landslides. 00:21:13.405 --> 00:21:15.341 align:center line:-1 position:50% size:33% But next to these deep-seated landslides, 00:21:15.341 --> 00:21:19.712 align:center line:-1 position:50% size:29% you also have a lot of shallow landslides. 00:21:19.712 --> 00:21:23.549 align:center line:-1 position:50% size:40% And that's something that we were mostly interested in. 00:21:23.549 --> 00:21:28.554 align:center line:-1 position:50% size:34% And what we focused on is basically my office, 00:21:28.554 --> 00:21:29.755 align:center line:-1 position:50% size:31% so Lawrence Berkeley National Lab, 00:21:29.755 --> 00:21:31.790 align:center line:-1 position:50% size:37% or the campus of Lawrence Berkeley National Lab, 00:21:31.790 --> 00:21:35.194 align:center line:-1 position:50% size:43% because, as you can see in this other map here on the right, 00:21:35.194 --> 00:21:36.929 align:center line:-1 position:50% size:33% also there we have a lot of landslides. 00:21:36.929 --> 00:21:39.131 align:center line:-1 position:50% size:34% So you have a lot of (inaudible) landslides, 00:21:39.131 --> 00:21:43.202 align:center line:-1 position:50% size:36% so, for instance, out here you have an old landslide, 00:21:43.202 --> 00:21:44.370 align:center line:-1 position:50% size:29% you have some here. 00:21:44.370 --> 00:21:45.404 align:center line:-1 position:50% size:44% Most of them actually happened 00:21:45.404 --> 00:21:46.739 align:center line:-1 position:50% size:33% during the development of the site, 00:21:46.739 --> 00:21:48.974 align:center line:-1 position:50% size:27% where they did a lot of excavations, 00:21:48.974 --> 00:21:50.943 align:center line:-1 position:50% size:27% changed the slopes quite significantly, 00:21:50.943 --> 00:21:53.612 align:center line:-1 position:50% size:25% and then you get a lot of landslides. 00:21:53.612 --> 00:21:55.447 align:center line:-1 position:50% size:31% But also more recently you get quite a-- 00:21:55.447 --> 00:21:56.682 align:center line:-1 position:50% size:42% you get quite a few landslides. 00:21:56.682 --> 00:22:00.386 align:center line:-1 position:50% size:35% So one of the active ones that's still creeping 00:22:00.386 --> 00:22:02.922 align:center line:-1 position:50% size:38% is Chicken Creek Landslide that you have here. 00:22:02.922 --> 00:22:04.757 align:center line:-1 position:50% size:26% You have one that's called LS-40, 00:22:04.757 --> 00:22:07.860 align:center line:-1 position:50% size:29% so likely calling-- or being called LS-40 00:22:07.860 --> 00:22:09.495 align:center line:-1 position:50% size:41% because landslide number 40, 00:22:09.495 --> 00:22:10.729 align:center line:-1 position:50% size:36% which is affecting a bridge here, 00:22:10.729 --> 00:22:12.464 align:center line:-1 position:50% size:29% so I will focus on this in a little bit. 00:22:12.464 --> 00:22:13.332 align:center line:-1 position:50% size:24% But basically what you can see 00:22:13.332 --> 00:22:15.267 align:center line:-1 position:50% size:40% is there are a lot of landslides at this site, 00:22:15.267 --> 00:22:16.936 align:center line:-1 position:50% size:23% but also in terms of landslide risk, 00:22:16.936 --> 00:22:18.370 align:center line:-1 position:50% size:28% we have a very high landslide risk 00:22:18.370 --> 00:22:22.174 align:center line:-1 position:50% size:37% because we also have a lot of elements at risk, right. 00:22:22.174 --> 00:22:25.544 align:center line:-1 position:50% size:39% So we have unique scientific facilities in this area, 00:22:25.544 --> 00:22:26.912 align:center line:-1 position:50% size:25% and not only here, 00:22:26.912 --> 00:22:30.649 align:center line:-1 position:50% size:37% also downslope just in the U.C. Berkeley campus. 00:22:30.649 --> 00:22:33.452 align:center line:-1 position:50% size:30% So, really, we use this as our field laboratory 00:22:33.452 --> 00:22:35.554 align:center line:-1 position:50% size:35% and try to develop some of these techniques 00:22:35.554 --> 00:22:38.257 align:center line:-1 position:50% size:40% of combining hydrological and geomechanical modeling 00:22:38.257 --> 00:22:41.694 align:center line:-1 position:50% size:42% with remote sensing and so on at this site. 00:22:41.694 --> 00:22:44.597 align:center line:-1 position:50% size:30% So, first what we tried to do here 00:22:44.597 --> 00:22:49.268 align:center line:-1 position:50% size:42% is really we first aimed to assess the landslide hazard 00:22:49.268 --> 00:22:50.636 align:center line:-1 position:50% size:38% through some geostatistical 00:22:50.636 --> 00:22:52.671 align:center line:-1 position:50% size:36% and hydro-geomechanical modeling, 00:22:52.671 --> 00:22:56.442 align:center line:-1 position:50% size:40% which we did here using the Python package Landlab. 00:22:56.442 --> 00:23:00.813 align:center line:-1 position:50% size:33% Landlab basically takes a digital elevation model 00:23:00.813 --> 00:23:02.648 align:center line:-1 position:50% size:19% of the area, so, you know, 00:23:02.648 --> 00:23:04.149 align:center line:-1 position:50% size:28% DEM, which we had a very good one, 00:23:04.149 --> 00:23:07.219 align:center line:-1 position:50% size:43% because there's high resolution lidar data on the area. 00:23:07.219 --> 00:23:11.857 align:center line:-1 position:50% size:42% And then you take distributions of various input parameters 00:23:11.857 --> 00:23:14.627 align:center line:-1 position:50% size:31% to basically calculate a landslide probability. 00:23:14.627 --> 00:23:18.664 align:center line:-1 position:50% size:40% And before we ran this model with the-- 00:23:18.664 --> 00:23:22.067 align:center line:-1 position:50% size:43% basically distributions that we knew that we had to expect, 00:23:22.067 --> 00:23:25.537 align:center line:-1 position:50% size:35% we ran just some sensitivity analysis. 00:23:25.537 --> 00:23:27.172 align:center line:-1 position:50% size:33% And that's something that's shown down here, 00:23:27.172 --> 00:23:30.309 align:center line:-1 position:50% size:40% where we basically estimated how much influence 00:23:30.309 --> 00:23:34.847 align:center line:-1 position:50% size:40% each variable has on the outcome of the model. 00:23:34.847 --> 00:23:37.583 align:center line:-1 position:50% size:31% And what we found is that really the most-- 00:23:37.583 --> 00:23:41.453 align:center line:-1 position:50% size:37% the most important variable is the slope angle, 00:23:41.453 --> 00:23:44.923 align:center line:-1 position:50% size:40% followed by the soil thickness and the soil cohesion. 00:23:44.923 --> 00:23:47.092 align:center line:-1 position:50% size:27% And less important are the soil density, 00:23:47.092 --> 00:23:48.827 align:center line:-1 position:50% size:25% the friction angle, the transmissivity, 00:23:48.827 --> 00:23:49.828 align:center line:-1 position:50% size:33% and the water recharge. 00:23:49.828 --> 00:23:53.198 align:center line:-1 position:50% size:39% So, basically water recharge just basically defining 00:23:53.198 --> 00:23:55.901 align:center line:-1 position:50% size:34% your precipitation pattern over the year. 00:23:55.901 --> 00:23:57.403 align:center line:-1 position:50% size:44% But so basically we had to know the slope angle 00:23:57.403 --> 00:24:00.272 align:center line:-1 position:50% size:30% and the soil thickness and the soil cohesion. 00:24:00.272 --> 00:24:01.173 align:center line:-1 position:50% size:30% Now, the slope angle, as I said, 00:24:01.173 --> 00:24:02.641 align:center line:-1 position:50% size:38% is something that we can get pretty easily 00:24:02.641 --> 00:24:04.543 align:center line:-1 position:50% size:35% because we have high-resolution lidar data, 00:24:04.543 --> 00:24:06.345 align:center line:-1 position:50% size:38% so we didn't worry too much about that. 00:24:06.345 --> 00:24:08.881 align:center line:-1 position:50% size:42% But we had to find ways of estimating the soil thickness 00:24:08.881 --> 00:24:11.050 align:center line:-1 position:50% size:24% and the cohesion across the site. 00:24:11.050 --> 00:24:12.851 align:center line:-1 position:50% size:29% So the soil thickness, 00:24:12.851 --> 00:24:17.690 align:center line:-1 position:50% size:27% we estimated basically combining 00:24:17.690 --> 00:24:20.225 align:center line:-1 position:50% size:34% former information that we had from the site 00:24:20.225 --> 00:24:22.227 align:center line:-1 position:50% size:47% and then some geophysical measurements. 00:24:22.227 --> 00:24:25.164 align:center line:-1 position:50% size:41% And in this case, ambient seismic noise measurements. 00:24:25.164 --> 00:24:26.598 align:center line:-1 position:50% size:32% And we did this because the boreholes, 00:24:26.598 --> 00:24:28.367 align:center line:-1 position:50% size:38% that are shown in blue here, 00:24:28.367 --> 00:24:32.971 align:center line:-1 position:50% size:41% are in locations that are not necessarily particularly useful 00:24:32.971 --> 00:24:37.343 align:center line:-1 position:50% size:35% to make an assessment of what the soil thickness 00:24:37.343 --> 00:24:38.844 align:center line:-1 position:50% size:23% of the slopes in this area area, 00:24:38.844 --> 00:24:42.314 align:center line:-1 position:50% size:46% because most of these boreholes are actually at points 00:24:42.314 --> 00:24:44.350 align:center line:-1 position:50% size:37% where they needed to build a new building, 00:24:44.350 --> 00:24:47.619 align:center line:-1 position:50% size:32% or--it was that, actually. 00:24:47.619 --> 00:24:48.520 align:center line:-1 position:50% size:31% They're mostly at sites 00:24:48.520 --> 00:24:50.456 align:center line:-1 position:50% size:32% where they had to build new buildings, 00:24:50.456 --> 00:24:53.258 align:center line:-1 position:50% size:30% so they are just trying to estimate 00:24:53.258 --> 00:24:56.462 align:center line:-1 position:50% size:38% what the conditions are in already developed areas. 00:24:56.462 --> 00:24:58.230 align:center line:-1 position:50% size:27% So they were very little boreholes 00:24:58.230 --> 00:25:00.699 align:center line:-1 position:50% size:37% that were actually telling us what the soil thickness is. 00:25:00.699 --> 00:25:03.902 align:center line:-1 position:50% size:44% And so we added these ambient seismic noise measurements, 00:25:03.902 --> 00:25:05.104 align:center line:-1 position:50% size:37% which basically just means 00:25:05.104 --> 00:25:08.974 align:center line:-1 position:50% size:41% you take a three component geophone, 00:25:08.974 --> 00:25:10.509 align:center line:-1 position:50% size:31% you put it out in a field, 00:25:10.509 --> 00:25:13.479 align:center line:-1 position:50% size:36% you record for half an hour or an hour so, 00:25:13.479 --> 00:25:16.915 align:center line:-1 position:50% size:31% and then you calculate the spectral ratio 00:25:16.915 --> 00:25:18.951 align:center line:-1 position:50% size:30% of the horizontal to vertical component 00:25:18.951 --> 00:25:21.086 align:center line:-1 position:50% size:27% of the signals that you measured. 00:25:21.086 --> 00:25:23.489 align:center line:-1 position:50% size:37% And what you get from that is basically just 00:25:23.489 --> 00:25:26.358 align:center line:-1 position:50% size:34% the resonance frequency of the ground. 00:25:26.358 --> 00:25:29.061 align:center line:-1 position:50% size:35% And you can see this in these plots down here. 00:25:29.061 --> 00:25:31.930 align:center line:-1 position:50% size:41% And this resonance frequency 00:25:31.930 --> 00:25:33.565 align:center line:-1 position:50% size:38% gives you an indication of what the soil thickness is, 00:25:33.565 --> 00:25:35.768 align:center line:-1 position:50% size:40% because usually there's resonance frequency, 00:25:35.768 --> 00:25:38.637 align:center line:-1 position:50% size:29% it just depends on the soil thickness. 00:25:38.637 --> 00:25:41.940 align:center line:-1 position:50% size:40% And the higher the frequency the thinner the soil cover, 00:25:41.940 --> 00:25:45.010 align:center line:-1 position:50% size:38% and the lower the frequency the thicker the soil cover. 00:25:45.010 --> 00:25:48.414 align:center line:-1 position:50% size:35% Now to really translate these peaks in frequency 00:25:48.414 --> 00:25:49.381 align:center line:-1 position:50% size:32% into your soil thickness, 00:25:49.381 --> 00:25:52.117 align:center line:-1 position:50% size:38% you need to know the shear velocity of the ground. 00:25:52.117 --> 00:25:54.853 align:center line:-1 position:50% size:38% And we were lucky because there are quite a number 00:25:54.853 --> 00:25:57.122 align:center line:-1 position:50% size:44% of shear velocity measurements across this site, 00:25:57.122 --> 00:25:58.223 align:center line:-1 position:50% size:38% so we had a very good idea 00:25:58.223 --> 00:26:00.759 align:center line:-1 position:50% size:38% on what the expected shear across these are, 00:26:00.759 --> 00:26:03.228 align:center line:-1 position:50% size:30% so we could actually translate these peaks 00:26:03.228 --> 00:26:05.030 align:center line:-1 position:50% size:28% into a soil thickness, 00:26:05.030 --> 00:26:06.865 align:center line:-1 position:50% size:40% add this to the borehole measurements, 00:26:06.865 --> 00:26:08.967 align:center line:-1 position:50% size:30% and then interpolate it across the site 00:26:08.967 --> 00:26:10.002 align:center line:-1 position:50% size:39% to get a good understanding 00:26:10.002 --> 00:26:12.704 align:center line:-1 position:50% size:38% of what the soil thickness is across the site. 00:26:12.704 --> 00:26:14.840 align:center line:-1 position:50% size:43% And then the second parameter that we had to know 00:26:14.840 --> 00:26:17.976 align:center line:-1 position:50% size:28% is the total cohesion, 00:26:17.976 --> 00:26:20.646 align:center line:-1 position:50% size:27% which is the sum of the soil cohesion 00:26:20.646 --> 00:26:22.347 align:center line:-1 position:50% size:31% and the root cohesion. 00:26:22.347 --> 00:26:25.284 align:center line:-1 position:50% size:39% So the soil cohesion is something that we could get 00:26:25.284 --> 00:26:26.785 align:center line:-1 position:50% size:29% from all of these geotechnical drillings 00:26:26.785 --> 00:26:28.954 align:center line:-1 position:50% size:38% that we had across the site. 00:26:28.954 --> 00:26:31.757 align:center line:-1 position:50% size:28% But what we needed to better understand 00:26:31.757 --> 00:26:33.625 align:center line:-1 position:50% size:28% is the root cohesion. 00:26:33.625 --> 00:26:38.063 align:center line:-1 position:50% size:40% And basically what we did is we used some planar data, 00:26:38.063 --> 00:26:42.100 align:center line:-1 position:50% size:33% so some high-resolution aerial remote sen-- 00:26:42.100 --> 00:26:44.236 align:center line:-1 position:50% size:42% or optical remote sensing data, 00:26:44.236 --> 00:26:47.406 align:center line:-1 position:50% size:34% has a resolution of about three meters in this area. 00:26:47.406 --> 00:26:50.809 align:center line:-1 position:50% size:44% And we took this at various times across the year 00:26:50.809 --> 00:26:52.878 align:center line:-1 position:50% size:36% and then trained a machine learning model 00:26:52.878 --> 00:26:57.716 align:center line:-1 position:50% size:43% to predict which vegetation type we have across the site. 00:26:57.716 --> 00:27:01.353 align:center line:-1 position:50% size:31% And we made it also somewhat simple 00:27:01.353 --> 00:27:03.755 align:center line:-1 position:50% size:31% in that we only distinguished between 00:27:03.755 --> 00:27:05.657 align:center line:-1 position:50% size:33% what is high or what's tall vegetation, 00:27:05.657 --> 00:27:07.226 align:center line:-1 position:50% size:31% what is low vegetation, 00:27:07.226 --> 00:27:09.394 align:center line:-1 position:50% size:40% and where do we really have vegetation, 00:27:09.394 --> 00:27:12.498 align:center line:-1 position:50% size:36% so where do we only have bare soil across. 00:27:12.498 --> 00:27:18.337 align:center line:-1 position:50% size:41% And then by employing a simple root density function, 00:27:18.337 --> 00:27:21.073 align:center line:-1 position:50% size:37% we could actually calculate what the root cohesion is 00:27:21.073 --> 00:27:24.776 align:center line:-1 position:50% size:34% with that depending on each vegetation type. 00:27:24.776 --> 00:27:27.713 align:center line:-1 position:50% size:43% So that's then bringing us the total cohesion that we need 00:27:27.713 --> 00:27:30.249 align:center line:-1 position:50% size:42% for the model, and then the other parameters 00:27:30.249 --> 00:27:33.752 align:center line:-1 position:50% size:40% on which the model is not that sensitive anyways, 00:27:33.752 --> 00:27:35.721 align:center line:-1 position:50% size:28% like the friction angle or the soil density, 00:27:35.721 --> 00:27:37.689 align:center line:-1 position:50% size:33% is just other parameters that we took 00:27:37.689 --> 00:27:39.925 align:center line:-1 position:50% size:39% from the geotechnical drilling and from soil testing 00:27:39.925 --> 00:27:42.928 align:center line:-1 position:50% size:44% that had been performed at site. 00:27:42.928 --> 00:27:45.063 align:center line:-1 position:50% size:38% And then we ran this model, 00:27:45.063 --> 00:27:48.166 align:center line:-1 position:50% size:36% and basically the outcome is shown here. 00:27:48.166 --> 00:27:52.037 align:center line:-1 position:50% size:37% So what we get from it is just probability of failure, 00:27:52.037 --> 00:27:54.773 align:center line:-1 position:50% size:35% that's coming from a Monte-Carlo simulation 00:27:54.773 --> 00:27:56.875 align:center line:-1 position:50% size:34% of an infinite slope model 00:27:56.875 --> 00:27:58.877 align:center line:-1 position:50% size:28% that is implemented at a certain Landlab, 00:27:58.877 --> 00:28:02.814 align:center line:-1 position:50% size:43% and then we can see areas that have high landslide probability, 00:28:02.814 --> 00:28:05.183 align:center line:-1 position:50% size:41% particularly in the eastern part of the site, 00:28:05.183 --> 00:28:06.785 align:center line:-1 position:50% size:20% which is also (unintelligible). 00:28:06.785 --> 00:28:08.287 align:center line:-1 position:50% size:36% I was actually sitting here, 00:28:08.287 --> 00:28:12.024 align:center line:-1 position:50% size:37% so just below this very high landslide probability, 00:28:12.024 --> 00:28:16.061 align:center line:-1 position:50% size:42% but also actually sitting on one of the old putty landslides, 00:28:16.061 --> 00:28:17.896 align:center line:-1 position:50% size:34% and we have high landslide probability 00:28:17.896 --> 00:28:20.832 align:center line:-1 position:50% size:41% in this area, and we also have high landslide probability 00:28:20.832 --> 00:28:22.000 align:center line:-1 position:50% size:16% in this area. 00:28:22.000 --> 00:28:23.602 align:center line:-1 position:50% size:39% So now we know, you know, 00:28:23.602 --> 00:28:26.572 align:center line:-1 position:50% size:34% where we have kind of high landslide probability 00:28:26.572 --> 00:28:28.040 align:center line:-1 position:50% size:27% across the site, so that's something 00:28:28.040 --> 00:28:31.176 align:center line:-1 position:50% size:26% that we'll later use to then also decide 00:28:31.176 --> 00:28:33.579 align:center line:-1 position:50% size:27% where we put our sensor network. 00:28:33.579 --> 00:28:36.014 align:center line:-1 position:50% size:43% The other thing that I already mentioned is that we-- 00:28:36.014 --> 00:28:37.849 align:center line:-1 position:50% size:34% to calculate this landslide probability, 00:28:37.849 --> 00:28:41.019 align:center line:-1 position:50% size:36% we also included the vegetation distribution. 00:28:41.019 --> 00:28:43.589 align:center line:-1 position:50% size:35% That's actually pretty neat because it also means 00:28:43.589 --> 00:28:46.625 align:center line:-1 position:50% size:41% that if the vegetation changes, we could recalculate 00:28:46.625 --> 00:28:49.161 align:center line:-1 position:50% size:25% what the landslide probability is. 00:28:49.161 --> 00:28:52.764 align:center line:-1 position:50% size:45% And in this case, where you have active vegetation management 00:28:52.764 --> 00:28:55.434 align:center line:-1 position:50% size:33% but also wildfires, that's actually very interesting. 00:28:55.434 --> 00:28:57.202 align:center line:-1 position:50% size:23% And we had one of these cases. 00:28:57.202 --> 00:29:00.706 align:center line:-1 position:50% size:29% In this case, one of the areas had 00:29:00.706 --> 00:29:02.774 align:center line:-1 position:50% size:26% eucalyptus trees that were pretty tall 00:29:02.774 --> 00:29:04.676 align:center line:-1 position:50% size:38% and were actually pretty close to a power line, 00:29:04.676 --> 00:29:07.646 align:center line:-1 position:50% size:37% and so PG&E had decided that they had to go 00:29:07.646 --> 00:29:09.948 align:center line:-1 position:50% size:26% because they were a wildfire hazard, 00:29:09.948 --> 00:29:12.951 align:center line:-1 position:50% size:40% and so we were actually able to recalculate-- 00:29:12.951 --> 00:29:15.153 align:center line:-1 position:50% size:39% or to calculate what the landslide probability was 00:29:15.153 --> 00:29:18.857 align:center line:-1 position:50% size:39% before and after this removal of the eucalyptus trees, 00:29:18.857 --> 00:29:21.793 align:center line:-1 position:50% size:46% and we can see the change here, and what we found is that, 00:29:21.793 --> 00:29:23.629 align:center line:-1 position:50% size:24% in areas where we had thick soil, 00:29:23.629 --> 00:29:28.266 align:center line:-1 position:50% size:40% so where the soil was thicker than roughly three meters, 00:29:28.266 --> 00:29:30.202 align:center line:-1 position:50% size:37% we saw hardly any change. 00:29:30.202 --> 00:29:31.970 align:center line:-1 position:50% size:38% So here, that's shown here, 00:29:31.970 --> 00:29:35.140 align:center line:-1 position:50% size:41% where we see, yeah, basically not much change at all. 00:29:35.140 --> 00:29:38.276 align:center line:-1 position:50% size:38% And then, for the other area 00:29:38.276 --> 00:29:41.246 align:center line:-1 position:50% size:26% where the soil was relatively thin, 00:29:41.246 --> 00:29:43.682 align:center line:-1 position:50% size:31% the removal of the eucalyptus trees 00:29:43.682 --> 00:29:45.951 align:center line:-1 position:50% size:33% actually really increased the landslide probability, 00:29:45.951 --> 00:29:47.753 align:center line:-1 position:50% size:31% and we talked with site management, 00:29:47.753 --> 00:29:49.087 align:center line:-1 position:50% size:27% and they're actually fully aware of that, 00:29:49.087 --> 00:29:53.992 align:center line:-1 position:50% size:43% and they are now implementing a vegetation plan 00:29:53.992 --> 00:29:58.797 align:center line:-1 position:50% size:40% where they're actually putting some native oak trees 00:29:58.797 --> 00:30:01.800 align:center line:-1 position:50% size:41% and they're putting some small shrubs and bushes here 00:30:01.800 --> 00:30:04.903 align:center line:-1 position:50% size:22% to basically help the vegetation-- 00:30:04.903 --> 00:30:10.575 align:center line:-1 position:50% size:33% or to use the vegetation to stabilize the slope. 00:30:10.575 --> 00:30:14.046 align:center line:-1 position:50% size:30% Now I would like to go to the next slide, but... 00:30:14.046 --> 00:30:16.715 align:center line:-1 position:50% size:27% Yeah, there you go, it's just a bit slow. 00:30:16.715 --> 00:30:19.151 align:center line:-1 position:50% size:25% So the other thing that we did is, 00:30:19.151 --> 00:30:22.754 align:center line:-1 position:50% size:35% we looked at another site. 00:30:22.754 --> 00:30:24.823 align:center line:-1 position:50% size:33% As I said, we're kind of interested, 00:30:24.823 --> 00:30:26.224 align:center line:-1 position:50% size:34% you know, it's nice to get 00:30:26.224 --> 00:30:29.795 align:center line:-1 position:50% size:37% the landslide probability distribution across the site, 00:30:29.795 --> 00:30:33.298 align:center line:-1 position:50% size:34% but we're also interested in better understanding 00:30:33.298 --> 00:30:34.966 align:center line:-1 position:50% size:27% landslide dynamics, and in particular, 00:30:34.966 --> 00:30:36.702 align:center line:-1 position:50% size:31% the landslide dynamics of a landslide, 00:30:36.702 --> 00:30:39.104 align:center line:-1 position:50% size:23% this Landslide 40 that's out here, 00:30:39.104 --> 00:30:41.573 align:center line:-1 position:50% size:24% which is affecting this bridge. 00:30:41.573 --> 00:30:44.609 align:center line:-1 position:50% size:30% And so at this site, we did various things, 00:30:44.609 --> 00:30:46.378 align:center line:-1 position:50% size:37% a range of geophysical techniques, 00:30:46.378 --> 00:30:48.513 align:center line:-1 position:50% size:26% and I will talk about two of them. 00:30:48.513 --> 00:30:50.282 align:center line:-1 position:50% size:30% So, first, I'm going to show you 00:30:50.282 --> 00:30:51.983 align:center line:-1 position:50% size:39% some of the characterization that we did 00:30:51.983 --> 00:30:54.953 align:center line:-1 position:50% size:37% on a relatively small scale, but basically we measured, 00:30:54.953 --> 00:30:58.290 align:center line:-1 position:50% size:40% again, the P- and S-wave velocities. 00:30:58.290 --> 00:30:59.791 align:center line:-1 position:50% size:42% So we have P-wave velocities. 00:30:59.791 --> 00:31:01.126 align:center line:-1 position:50% size:25% This is actually S-wave velocities, 00:31:01.126 --> 00:31:02.527 align:center line:-1 position:50% size:27% even though it says P-wave velocities, 00:31:02.527 --> 00:31:03.995 align:center line:-1 position:50% size:35% but it's S-wave velocities, 00:31:03.995 --> 00:31:07.766 align:center line:-1 position:50% size:42% and we can see how they vary across the site, 00:31:07.766 --> 00:31:09.801 align:center line:-1 position:50% size:30% which basically shows that we just have 00:31:09.801 --> 00:31:13.338 align:center line:-1 position:50% size:41% a pretty slow velocity material 00:31:13.338 --> 00:31:16.608 align:center line:-1 position:50% size:39% that's overlying more higher velocity ground, 00:31:16.608 --> 00:31:19.845 align:center line:-1 position:50% size:43% so pretty weak material on top of more competent rock. 00:31:19.845 --> 00:31:22.748 align:center line:-1 position:50% size:45% And basically what we have, too, here is basically the interface 00:31:22.748 --> 00:31:25.784 align:center line:-1 position:50% size:43% between the Moraga Formation and the Orinda Formation, 00:31:25.784 --> 00:31:27.486 align:center line:-1 position:50% size:31% and in this case, the Moraga Formation 00:31:27.486 --> 00:31:29.788 align:center line:-1 position:50% size:41% is somewhat more permeable than the Orinda Formation. 00:31:29.788 --> 00:31:31.957 align:center line:-1 position:50% size:21% And if we look across the site, 00:31:31.957 --> 00:31:33.992 align:center line:-1 position:50% size:29% almost at every place where we are 00:31:33.992 --> 00:31:35.627 align:center line:-1 position:50% size:42% at this interface between these two formations, 00:31:35.627 --> 00:31:37.362 align:center line:-1 position:50% size:34% that's also where we see some landslides, 00:31:37.362 --> 00:31:39.231 align:center line:-1 position:50% size:28% which is probably just because we get, 00:31:39.231 --> 00:31:40.866 align:center line:-1 position:50% size:41% you know, water that's coming 00:31:40.866 --> 00:31:42.834 align:center line:-1 position:50% size:38% from the Moraga Formation and the Orinda Formation, 00:31:42.834 --> 00:31:45.137 align:center line:-1 position:50% size:37% it just cannot go anywhere, and we get rise 00:31:45.137 --> 00:31:49.474 align:center line:-1 position:50% size:39% in pore pressures that's then driving some deformation. 00:31:49.474 --> 00:31:51.877 align:center line:-1 position:50% size:34% So as I said, we have this bridge here 00:31:51.877 --> 00:31:55.380 align:center line:-1 position:50% size:31% which is actually being heavily affected 00:31:55.380 --> 00:31:56.915 align:center line:-1 position:50% size:36% by this creeping landslide. 00:31:56.915 --> 00:31:58.583 align:center line:-1 position:50% size:29% So the landslide itself is moving 00:31:58.583 --> 00:32:00.986 align:center line:-1 position:50% size:43% by about ten millimeters a year, so not a lot, 00:32:00.986 --> 00:32:04.156 align:center line:-1 position:50% size:36% but it's actually having a big effect on this bridge. 00:32:04.156 --> 00:32:06.057 align:center line:-1 position:50% size:29% And at Berkeley Lab, 00:32:06.057 --> 00:32:07.826 align:center line:-1 position:50% size:37% we use advanced monitoring instrumentation, 00:32:07.826 --> 00:32:11.129 align:center line:-1 position:50% size:33% so one of them is a ruler that is just measuring 00:32:11.129 --> 00:32:12.597 align:center line:-1 position:50% size:29% how much this bridge is deforming. 00:32:12.597 --> 00:32:15.367 align:center line:-1 position:50% size:41% So we're just looking at one of the pillars of the bridge here, 00:32:15.367 --> 00:32:17.235 align:center line:-1 position:50% size:35% and basically you have this ruler that's measuring 00:32:17.235 --> 00:32:18.470 align:center line:-1 position:50% size:34% how much it's deforming. 00:32:18.470 --> 00:32:21.439 align:center line:-1 position:50% size:32% The only downside of such instrumentation 00:32:21.439 --> 00:32:23.508 align:center line:-1 position:50% size:40% is that, every now and again, it falls off 00:32:23.508 --> 00:32:24.976 align:center line:-1 position:50% size:40% and then it has to be reglued, 00:32:24.976 --> 00:32:26.978 align:center line:-1 position:50% size:28% so really it's actually of not much use. 00:32:26.978 --> 00:32:30.348 align:center line:-1 position:50% size:39% So, but anyway, I mean, they're monitoring the bridge 00:32:30.348 --> 00:32:31.917 align:center line:-1 position:50% size:30% and actually right now 00:32:31.917 --> 00:32:33.552 align:center line:-1 position:50% size:35% are completely rebuilding the bridge. 00:32:33.552 --> 00:32:38.190 align:center line:-1 position:50% size:38% So, yeah, having a good bit of construction there 00:32:38.190 --> 00:32:39.958 align:center line:-1 position:50% size:21% at the moment. 00:32:39.958 --> 00:32:42.594 align:center line:-1 position:50% size:37% Basically what we did here is we installed 00:32:42.594 --> 00:32:45.463 align:center line:-1 position:50% size:36% a range of instrumentation along a short transect 00:32:45.463 --> 00:32:47.232 align:center line:-1 position:50% size:26% that's shown here, so we have, again, 00:32:47.232 --> 00:32:48.400 align:center line:-1 position:50% size:39% some resistivity tomography, 00:32:48.400 --> 00:32:50.268 align:center line:-1 position:50% size:21% but what I want to focus on first 00:32:50.268 --> 00:32:53.371 align:center line:-1 position:50% size:40% is actually some ambient seismic noise measurements 00:32:53.371 --> 00:32:59.978 align:center line:-1 position:50% size:42% that we did in October of 2021. 00:32:59.978 --> 00:33:04.182 align:center line:-1 position:50% size:39% And at that time, we got about 150 millimeters of rain 00:33:04.182 --> 00:33:05.717 align:center line:-1 position:50% size:22% within 24 hours, 00:33:05.717 --> 00:33:09.120 align:center line:-1 position:50% size:27% so pretty significant rainfall event 00:33:09.120 --> 00:33:12.090 align:center line:-1 position:50% size:34% that actually caused some of the deformation. 00:33:12.090 --> 00:33:15.961 align:center line:-1 position:50% size:36% And so, what we see here is basically the results 00:33:15.961 --> 00:33:17.729 align:center line:-1 position:50% size:41% of these ambient seismic noise measurements. 00:33:17.729 --> 00:33:20.232 align:center line:-1 position:50% size:29% So what you do there is you just record 00:33:20.232 --> 00:33:24.135 align:center line:-1 position:50% size:40% the seismic wavefield as it's happening around you. 00:33:24.135 --> 00:33:26.938 align:center line:-1 position:50% size:29% So it will be the noise from cars, 00:33:26.938 --> 00:33:30.108 align:center line:-1 position:50% size:45% it will be the noise of the trees moving around in the wind, 00:33:30.108 --> 00:33:31.977 align:center line:-1 position:50% size:32% it will be the noise of the waves in the bay. 00:33:31.977 --> 00:33:34.379 align:center line:-1 position:50% size:38% So everything that is kind of going on around you 00:33:34.379 --> 00:33:36.381 align:center line:-1 position:50% size:41% is what we're measuring here. 00:33:36.381 --> 00:33:39.985 align:center line:-1 position:50% size:33% And then what we do is, we correlate 00:33:39.985 --> 00:33:43.655 align:center line:-1 position:50% size:34% the signal of each station to itself. 00:33:43.655 --> 00:33:47.459 align:center line:-1 position:50% size:47% So we correlate basically what's happening 48 hours before 00:33:47.459 --> 00:33:48.894 align:center line:-1 position:50% size:32% to what we're recording right now, 00:33:48.894 --> 00:33:50.161 align:center line:-1 position:50% size:45% and then what you would expect, 00:33:50.161 --> 00:33:51.496 align:center line:-1 position:50% size:29% if nothing is changing the ground, 00:33:51.496 --> 00:33:54.366 align:center line:-1 position:50% size:35% is that you get a very high correlation coefficient 00:33:54.366 --> 00:33:56.001 align:center line:-1 position:50% size:35% because basically your signals don't change 00:33:56.001 --> 00:33:57.402 align:center line:-1 position:50% size:32% if nothing in the ground changes, 00:33:57.402 --> 00:34:01.072 align:center line:-1 position:50% size:49% because the noise you can assume all the time is pretty constant. 00:34:01.072 --> 00:34:02.540 align:center line:-1 position:50% size:32% But that's not the case, 00:34:02.540 --> 00:34:04.643 align:center line:-1 position:50% size:35% and that has been shown in other papers, too, 00:34:04.643 --> 00:34:06.011 align:center line:-1 position:50% size:40% that when you get landslides, 00:34:06.011 --> 00:34:09.481 align:center line:-1 position:50% size:43% you actually can get, you know, some responses in the ground 00:34:09.481 --> 00:34:11.683 align:center line:-1 position:50% size:41% that sometimes happen hours before actual movement, 00:34:11.683 --> 00:34:13.418 align:center line:-1 position:50% size:26% and that's also what we saw here. 00:34:13.418 --> 00:34:15.387 align:center line:-1 position:50% size:31% So, first, before the rain started, 00:34:15.387 --> 00:34:17.322 align:center line:-1 position:50% size:29% we get this high correlation coefficient 00:34:17.322 --> 00:34:18.857 align:center line:-1 position:50% size:25% which is shown in red colors here, 00:34:18.857 --> 00:34:20.592 align:center line:-1 position:50% size:24% but then as soon as the rain starts, 00:34:20.592 --> 00:34:24.129 align:center line:-1 position:50% size:32% we get these lower correlation coefficients, 00:34:24.129 --> 00:34:25.397 align:center line:-1 position:50% size:23% and in particular, at the beginning, 00:34:25.397 --> 00:34:26.965 align:center line:-1 position:50% size:35% we get them at the higher frequencies, 00:34:26.965 --> 00:34:29.267 align:center line:-1 position:50% size:39% which means that, especially in the shallow soft surface, 00:34:29.267 --> 00:34:31.069 align:center line:-1 position:50% size:28% things are changing. 00:34:31.069 --> 00:34:36.641 align:center line:-1 position:50% size:41% And then, later on, we see that we lose the signal 00:34:36.641 --> 00:34:38.877 align:center line:-1 position:50% size:29% throughout almost all the frequencies, 00:34:38.877 --> 00:34:42.314 align:center line:-1 position:50% size:39% and we basically, yeah, so our correlation coefficient 00:34:42.314 --> 00:34:44.616 align:center line:-1 position:50% size:28% almost goes to zero, and we saw-- 00:34:44.616 --> 00:34:47.285 align:center line:-1 position:50% size:41% this shows us that we get, you know, a disturbed ground 00:34:47.285 --> 00:34:49.821 align:center line:-1 position:50% size:35% not only at the surface but also at deeper layers. 00:34:49.821 --> 00:34:52.757 align:center line:-1 position:50% size:36% And here in this case, this happens around here, 00:34:52.757 --> 00:34:55.527 align:center line:-1 position:50% size:32% and then six hours later is when we see 00:34:55.527 --> 00:34:58.897 align:center line:-1 position:50% size:38% some significant movement throughout the soil count. 00:34:58.897 --> 00:35:02.934 align:center line:-1 position:50% size:38% So this plot here shows you this placement 00:35:02.934 --> 00:35:06.204 align:center line:-1 position:50% size:36% from basically the surface down to 1.8 meters depth, 00:35:06.204 --> 00:35:08.540 align:center line:-1 position:50% size:20% and I will talk a little bit more 00:35:08.540 --> 00:35:10.942 align:center line:-1 position:50% size:44% about these deformation probes in one of the following slides 00:35:10.942 --> 00:35:13.111 align:center line:-1 position:50% size:41% because they're pretty unique. 00:35:13.111 --> 00:35:14.913 align:center line:-1 position:50% size:30% But this is basically for the bottom sensor, 00:35:14.913 --> 00:35:16.915 align:center line:-1 position:50% size:36% then we had a sensor that was further to the top, 00:35:16.915 --> 00:35:19.551 align:center line:-1 position:50% size:29% and we can see a similar thing again. 00:35:19.551 --> 00:35:22.454 align:center line:-1 position:50% size:26% So we see the loss in the signal 00:35:22.454 --> 00:35:23.888 align:center line:-1 position:50% size:44% throughout the frequency range, 00:35:23.888 --> 00:35:26.124 align:center line:-1 position:50% size:38% and then, about three hours later in this case, 00:35:26.124 --> 00:35:27.959 align:center line:-1 position:50% size:39% is where things start moving. 00:35:27.959 --> 00:35:30.128 align:center line:-1 position:50% size:45% And in this case, as you can see from these plots, 00:35:30.128 --> 00:35:31.329 align:center line:-1 position:50% size:41% this is really small movement. 00:35:31.329 --> 00:35:33.631 align:center line:-1 position:50% size:37% It's only like ten millimeters of movement, 00:35:33.631 --> 00:35:36.634 align:center line:-1 position:50% size:45% so that's not a lot, but we see significant movement 00:35:36.634 --> 00:35:39.838 align:center line:-1 position:50% size:41% that we can resolve into these probes really nicely. 00:35:39.838 --> 00:35:42.107 align:center line:-1 position:50% size:41% And the interesting thing here, too, is that, you know, 00:35:42.107 --> 00:35:44.409 align:center line:-1 position:50% size:40% we get first movement at the bottom then at the top, 00:35:44.409 --> 00:35:47.145 align:center line:-1 position:50% size:48% so basically what this means is we get some movement down here 00:35:47.145 --> 00:35:48.480 align:center line:-1 position:50% size:22% that's releasing some stressors, 00:35:48.480 --> 00:35:51.916 align:center line:-1 position:50% size:36% and then we get some movement up there. 00:35:51.916 --> 00:35:53.618 align:center line:-1 position:50% size:36% We also, as I said, we had 00:35:53.618 --> 00:35:55.286 align:center line:-1 position:50% size:30% an electrical resistivity tomography line 00:35:55.286 --> 00:35:58.523 align:center line:-1 position:50% size:45% installed at the site, and we were basically using that 00:35:58.523 --> 00:36:02.761 align:center line:-1 position:50% size:45% to better understand how storm-- how these winter storms 00:36:02.761 --> 00:36:04.729 align:center line:-1 position:50% size:25% cause variations in road conditions 00:36:04.729 --> 00:36:07.565 align:center line:-1 position:50% size:39% and how this changes, then, the slope stability. 00:36:07.565 --> 00:36:11.136 align:center line:-1 position:50% size:42% So we monitored the electrical resistivity of the ground 00:36:11.136 --> 00:36:13.805 align:center line:-1 position:50% size:34% using electrical resistivity tomography, 00:36:13.805 --> 00:36:17.308 align:center line:-1 position:50% size:30% and the tomograms are shown again here 00:36:17.308 --> 00:36:19.044 align:center line:-1 position:50% size:41% where we can see, you know, 00:36:19.044 --> 00:36:21.913 align:center line:-1 position:50% size:33% the conductive putty landslide deposits, 00:36:21.913 --> 00:36:24.182 align:center line:-1 position:50% size:36% probably highly saturated. 00:36:24.182 --> 00:36:25.917 align:center line:-1 position:50% size:44% Above here, we have the Moraga Formation, 00:36:25.917 --> 00:36:28.820 align:center line:-1 position:50% size:43% and below here, we have the Orinda Formation. 00:36:28.820 --> 00:36:31.423 align:center line:-1 position:50% size:30% Now, usually, what we would expect 00:36:31.423 --> 00:36:35.760 align:center line:-1 position:50% size:41% when it's raining is that the resistivity would decrease, 00:36:35.760 --> 00:36:38.897 align:center line:-1 position:50% size:35% usually because we get an increase in saturation. 00:36:38.897 --> 00:36:43.635 align:center line:-1 position:50% size:39% In this case, we actually saw an increased resistivity, 00:36:43.635 --> 00:36:47.072 align:center line:-1 position:50% size:47% which is mostly because we have, 00:36:47.072 --> 00:36:50.341 align:center line:-1 position:50% size:43% in some really high conductivity pore water, 00:36:50.341 --> 00:36:52.410 align:center line:-1 position:50% size:42% that then during the fall season and winter 00:36:52.410 --> 00:36:55.246 align:center line:-1 position:50% size:39% where you have a lot of rain, it's washed out and replaced 00:36:55.246 --> 00:37:00.285 align:center line:-1 position:50% size:41% by more resistive rainwater that's coming into the system. 00:37:00.285 --> 00:37:02.387 align:center line:-1 position:50% size:41% So in this case, we can track the groundwater 00:37:02.387 --> 00:37:05.223 align:center line:-1 position:50% size:37% just by looking at these negative changes, 00:37:05.223 --> 00:37:07.692 align:center line:-1 position:50% size:30% and when we do that, we can actually 00:37:07.692 --> 00:37:09.627 align:center line:-1 position:50% size:42% really nicely extract where the groundwater table is 00:37:09.627 --> 00:37:11.996 align:center line:-1 position:50% size:38% and we can put this into simple infinite slope models 00:37:11.996 --> 00:37:13.631 align:center line:-1 position:50% size:29% and we can calculate how this changes 00:37:13.631 --> 00:37:16.000 align:center line:-1 position:50% size:26% the factor of safety of these landslides 00:37:16.000 --> 00:37:18.169 align:center line:-1 position:50% size:22% or, in this case, of these slopes. 00:37:18.169 --> 00:37:20.805 align:center line:-1 position:50% size:30% And so in this case, what we found is that, 00:37:20.805 --> 00:37:23.608 align:center line:-1 position:50% size:43% beginning or before the storms, 00:37:23.608 --> 00:37:26.144 align:center line:-1 position:50% size:34% we had a factor of safety of about 1.42, 00:37:26.144 --> 00:37:29.214 align:center line:-1 position:50% size:45% and then after the storms, when the groundwater table rose 00:37:29.214 --> 00:37:30.715 align:center line:-1 position:50% size:32% a little bit, by not much, 00:37:30.715 --> 00:37:32.784 align:center line:-1 position:50% size:34% just by a half meter or so in places, 00:37:32.784 --> 00:37:37.589 align:center line:-1 position:50% size:42% we see that the factor of safety goes down to 1.38. 00:37:37.589 --> 00:37:41.059 align:center line:-1 position:50% size:31% We also had a student working on developing 00:37:41.059 --> 00:37:42.994 align:center line:-1 position:50% size:34% some automation of this, 00:37:42.994 --> 00:37:45.063 align:center line:-1 position:50% size:33% so we used unsupervised clustering 00:37:45.063 --> 00:37:46.531 align:center line:-1 position:50% size:31% to extract the groundwater table, 00:37:46.531 --> 00:37:50.235 align:center line:-1 position:50% size:37% and we could use this then to really get 00:37:50.235 --> 00:37:54.973 align:center line:-1 position:50% size:38% this updated factor of safety in almost real time, 00:37:54.973 --> 00:37:56.374 align:center line:-1 position:50% size:27% so that's something that you can do 00:37:56.374 --> 00:38:00.378 align:center line:-1 position:50% size:35% with these kinds of monitoring techniques. 00:38:00.378 --> 00:38:03.181 align:center line:-1 position:50% size:35% So, for monitoring, again, we can understand 00:38:03.181 --> 00:38:06.651 align:center line:-1 position:50% size:44% how the soil moisture conditions indeed play a critical role 00:38:06.651 --> 00:38:09.420 align:center line:-1 position:50% size:31% in landslide processes at Berkeley Lab, 00:38:09.420 --> 00:38:12.223 align:center line:-1 position:50% size:38% and as I already mentioned, this is really great 00:38:12.223 --> 00:38:14.792 align:center line:-1 position:50% size:37% because you can get a very good understanding 00:38:14.792 --> 00:38:19.097 align:center line:-1 position:50% size:38% of how things are behaving, but as I've also shown, 00:38:19.097 --> 00:38:20.732 align:center line:-1 position:50% size:40% we only focused on this side. 00:38:20.732 --> 00:38:22.667 align:center line:-1 position:50% size:34% It doesn't tell us anything about what's happening 00:38:22.667 --> 00:38:25.937 align:center line:-1 position:50% size:38% anywhere in the other areas of the site. 00:38:25.937 --> 00:38:27.472 align:center line:-1 position:50% size:39% So what we decided to do is 00:38:27.472 --> 00:38:29.140 align:center line:-1 position:50% size:37% we decided to install a wireless sensor network, 00:38:29.140 --> 00:38:31.609 align:center line:-1 position:50% size:42% and you can see the sensor network that we installed here. 00:38:31.609 --> 00:38:33.378 align:center line:-1 position:50% size:26% So it's quite dense. 00:38:33.378 --> 00:38:34.746 align:center line:-1 position:50% size:35% We have a lot of sensors. 00:38:34.746 --> 00:38:36.614 align:center line:-1 position:50% size:29% We actually-- we have 18 locations 00:38:36.614 --> 00:38:40.552 align:center line:-1 position:50% size:44% where we measure soil moisture and suction 00:38:40.552 --> 00:38:41.719 align:center line:-1 position:50% size:34% at three different depths, 00:38:41.719 --> 00:38:45.290 align:center line:-1 position:50% size:36% at 10, 50, and 30 centimeters depth. 00:38:45.290 --> 00:38:49.127 align:center line:-1 position:50% size:48% And then, in addition, we have some deformation probes 00:38:49.127 --> 00:38:53.031 align:center line:-1 position:50% size:24% that I will explain on the next slide. 00:38:53.031 --> 00:38:55.567 align:center line:-1 position:50% size:32% We installed them at 60 different locations 00:38:55.567 --> 00:38:58.036 align:center line:-1 position:50% size:37% to get a good estimation of where things are moving 00:38:58.036 --> 00:39:00.505 align:center line:-1 position:50% size:39% and where things are stable. 00:39:00.505 --> 00:39:03.975 align:center line:-1 position:50% size:37% So these probes, they're actually pretty neat. 00:39:03.975 --> 00:39:07.445 align:center line:-1 position:50% size:41% We basically built them based on some developments 00:39:07.445 --> 00:39:08.913 align:center line:-1 position:50% size:22% at Berkeley Lab 00:39:08.913 --> 00:39:12.083 align:center line:-1 position:50% size:43% that were focused on temperature measurements. 00:39:12.083 --> 00:39:14.953 align:center line:-1 position:50% size:39% So we do quite a lot of research also in the Arctic 00:39:14.953 --> 00:39:17.956 align:center line:-1 position:50% size:46% and in mountainous environments where it's very interesting 00:39:17.956 --> 00:39:21.326 align:center line:-1 position:50% size:37% to see what soil temperatures are 00:39:21.326 --> 00:39:23.261 align:center line:-1 position:50% size:37% but also what, for instance, snow temperatures are 00:39:23.261 --> 00:39:25.363 align:center line:-1 position:50% size:39% because if you monitor them over time, 00:39:25.363 --> 00:39:28.600 align:center line:-1 position:50% size:33% you can estimate soil hydraulic properties 00:39:28.600 --> 00:39:31.336 align:center line:-1 position:50% size:35% but you can also estimate snow properties. 00:39:31.336 --> 00:39:33.271 align:center line:-1 position:50% size:43% Now, what we did in this project 00:39:33.271 --> 00:39:36.641 align:center line:-1 position:50% size:45% is we also added three-component accelerometers 00:39:36.641 --> 00:39:39.177 align:center line:-1 position:50% size:38% to this probe, and basically what we have 00:39:39.177 --> 00:39:42.847 align:center line:-1 position:50% size:36% is now an alternating array of temperature sensors 00:39:42.847 --> 00:39:47.452 align:center line:-1 position:50% size:47% and then basically a three-component accelerometer, 00:39:47.452 --> 00:39:49.454 align:center line:-1 position:50% size:31% and we have this every five centimeters. 00:39:49.454 --> 00:39:53.558 align:center line:-1 position:50% size:41% So, every five centimeters, we have a temperature probe, 00:39:53.558 --> 00:39:55.660 align:center line:-1 position:50% size:36% then an accelerometer, then a temperature probe. 00:39:55.660 --> 00:39:58.196 align:center line:-1 position:50% size:31% So basically, we have every ten centimeters, 00:39:58.196 --> 00:40:00.231 align:center line:-1 position:50% size:32% then an accelerometer. 00:40:00.231 --> 00:40:02.500 align:center line:-1 position:50% size:31% And these are all built on these flexible PCBs 00:40:02.500 --> 00:40:04.202 align:center line:-1 position:50% size:26% that I'll show in the middle here. 00:40:04.202 --> 00:40:06.971 align:center line:-1 position:50% size:39% And basically, they allow you to really bend the probe 00:40:06.971 --> 00:40:08.373 align:center line:-1 position:50% size:26% without it breaking, 00:40:08.373 --> 00:40:13.611 align:center line:-1 position:50% size:39% and we get pretty nice deformation data out of that. 00:40:13.611 --> 00:40:16.981 align:center line:-1 position:50% size:33% So, so far, we have built a modular 00:40:16.981 --> 00:40:19.284 align:center line:-1 position:50% size:36% so that we can have them, you know, any length 00:40:19.284 --> 00:40:22.153 align:center line:-1 position:50% size:30% between, in our case, 00:40:22.153 --> 00:40:24.956 align:center line:-1 position:50% size:29% 20 centimeters and 1.8 meter length, 00:40:24.956 --> 00:40:27.525 align:center line:-1 position:50% size:32% and basically we limit it to 1.8 meter length 00:40:27.525 --> 00:40:29.360 align:center line:-1 position:50% size:29% because the way we install them is just 00:40:29.360 --> 00:40:32.263 align:center line:-1 position:50% size:34% we take a long drill bit, we drill a hole in the soil, 00:40:32.263 --> 00:40:33.631 align:center line:-1 position:50% size:31% then just slide them in. 00:40:33.631 --> 00:40:36.067 align:center line:-1 position:50% size:39% This makes installation super easy and super quick, 00:40:36.067 --> 00:40:37.669 align:center line:-1 position:50% size:37% but if we would make them longer, 00:40:37.669 --> 00:40:40.104 align:center line:-1 position:50% size:40% it would just make installation much more difficult. 00:40:40.104 --> 00:40:44.175 align:center line:-1 position:50% size:31% And also, you know, as at many other sites, 00:40:44.175 --> 00:40:45.943 align:center line:-1 position:50% size:41% at Berkeley Lab, you need a penetration permit 00:40:45.943 --> 00:40:47.278 align:center line:-1 position:50% size:27% to do a lot of things, 00:40:47.278 --> 00:40:50.448 align:center line:-1 position:50% size:43% so the bigger you make your excavation or penetration, 00:40:50.448 --> 00:40:52.684 align:center line:-1 position:50% size:28% the more paperwork you have to file, 00:40:52.684 --> 00:40:55.720 align:center line:-1 position:50% size:36% so this was also one way of making it relatively easy 00:40:55.720 --> 00:40:58.089 align:center line:-1 position:50% size:29% for us to install these. 00:40:58.089 --> 00:41:00.692 align:center line:-1 position:50% size:36% We didn't only install them in the field, 00:41:00.692 --> 00:41:03.194 align:center line:-1 position:50% size:39% we also tested them in a lab, and what we found is that 00:41:03.194 --> 00:41:06.097 align:center line:-1 position:50% size:33% we have an accuracy of about two millimeters 00:41:06.097 --> 00:41:07.598 align:center line:-1 position:50% size:27% per meter of probe, 00:41:07.598 --> 00:41:09.867 align:center line:-1 position:50% size:33% but also we have submillimeter sensitivity, 00:41:09.867 --> 00:41:14.972 align:center line:-1 position:50% size:40% and we can see this in this plot in the middle here 00:41:14.972 --> 00:41:19.677 align:center line:-1 position:50% size:40% where I just, you know, very slightly moved the probe 00:41:19.677 --> 00:41:24.449 align:center line:-1 position:50% size:38% and looked at the change in basically the sat location. 00:41:24.449 --> 00:41:26.084 align:center line:-1 position:50% size:27% In this case, this was really little, 00:41:26.084 --> 00:41:29.120 align:center line:-1 position:50% size:36% so you have less than a millimeter of movement, 00:41:29.120 --> 00:41:32.323 align:center line:-1 position:50% size:40% but even though you may not be able to really resolve 00:41:32.323 --> 00:41:34.659 align:center line:-1 position:50% size:34% how much it's moving, you really get every time 00:41:34.659 --> 00:41:36.060 align:center line:-1 position:50% size:26% when the probe is actually moving, 00:41:36.060 --> 00:41:40.531 align:center line:-1 position:50% size:36% so you have submillimeter sensitivity in this case. 00:41:40.531 --> 00:41:44.068 align:center line:-1 position:50% size:36% And then, the system itself has a logger, of course, 00:41:44.068 --> 00:41:46.938 align:center line:-1 position:50% size:40% which is also very low power, so the entire system, 00:41:46.938 --> 00:41:49.741 align:center line:-1 position:50% size:40% including logger, operates off two AA batteries, 00:41:49.741 --> 00:41:53.311 align:center line:-1 position:50% size:31% and if we don't turn on data transmission, 00:41:53.311 --> 00:41:55.847 align:center line:-1 position:50% size:41% the system should be running, you know, 00:41:55.847 --> 00:41:58.750 align:center line:-1 position:50% size:32% at five-minute sampling for about three years 00:41:58.750 --> 00:42:01.552 align:center line:-1 position:50% size:36% on these two AA batteries. 00:42:01.552 --> 00:42:04.555 align:center line:-1 position:50% size:47% If we enable LoRa communication that we have on it, 00:42:04.555 --> 00:42:08.893 align:center line:-1 position:50% size:42% so long-range communication, this battery life decreases, 00:42:08.893 --> 00:42:11.262 align:center line:-1 position:50% size:35% but because we came up 00:42:11.262 --> 00:42:13.464 align:center line:-1 position:50% size:41% with some good data compression techniques, 00:42:13.464 --> 00:42:15.433 align:center line:-1 position:50% size:38% and also we are not using LoRaWAN 00:42:15.433 --> 00:42:18.269 align:center line:-1 position:50% size:45% but we are using our own communication protocol, 00:42:18.269 --> 00:42:21.305 align:center line:-1 position:50% size:35% still we have a battery life of about 1.5 years 00:42:21.305 --> 00:42:24.308 align:center line:-1 position:50% size:44% off two AA batteries even though they're transmitting. 00:42:24.308 --> 00:42:28.513 align:center line:-1 position:50% size:44% So, we deployed quite a number of them at Berkeley Lab, 00:42:28.513 --> 00:42:31.115 align:center line:-1 position:50% size:37% but we also deployed them in the Arctic 00:42:31.115 --> 00:42:34.685 align:center line:-1 position:50% size:40% because it's a good example and also it's a good test case 00:42:34.685 --> 00:42:37.321 align:center line:-1 position:50% size:35% because, in the Arctic, what you would expect is 00:42:37.321 --> 00:42:39.791 align:center line:-1 position:50% size:46% where you have frozen ground, that's where nothing should move 00:42:39.791 --> 00:42:42.527 align:center line:-1 position:50% size:35% because it's pretty-- well, it's frozen into place, 00:42:42.527 --> 00:42:44.028 align:center line:-1 position:50% size:28% and where you have unfrozen ground 00:42:44.028 --> 00:42:46.230 align:center line:-1 position:50% size:47% is where it should move, and these are just some examples 00:42:46.230 --> 00:42:48.032 align:center line:-1 position:50% size:35% of this here where we can see, really, 00:42:48.032 --> 00:42:49.901 align:center line:-1 position:50% size:35% that in the frozen ground, nothing moves, 00:42:49.901 --> 00:42:51.769 align:center line:-1 position:50% size:33% but as soon as we're on the unfrozen ground, 00:42:51.769 --> 00:42:54.572 align:center line:-1 position:50% size:32% we get movements of about ten millimeters 00:42:54.572 --> 00:42:56.874 align:center line:-1 position:50% size:42% in the period of about five days in this case. 00:42:56.874 --> 00:42:59.544 align:center line:-1 position:50% size:40% So, really, these probes work very nicely 00:42:59.544 --> 00:43:04.148 align:center line:-1 position:50% size:36% to get your soil information (unintelligible). 00:43:04.148 --> 00:43:06.184 align:center line:-1 position:50% size:42% So the data from these deformation probes 00:43:06.184 --> 00:43:08.453 align:center line:-1 position:50% size:40% and the soil moisture sensors is then transmitted 00:43:08.453 --> 00:43:10.588 align:center line:-1 position:50% size:36% via LoRa to a base station 00:43:10.588 --> 00:43:14.158 align:center line:-1 position:50% size:30% which receives and decodes the data 00:43:14.158 --> 00:43:17.862 align:center line:-1 position:50% size:29% and then passes it to an Influx database 00:43:17.862 --> 00:43:19.464 align:center line:-1 position:50% size:47% which is an open-source database 00:43:19.464 --> 00:43:21.866 align:center line:-1 position:50% size:39% that has been specifically designed to work 00:43:21.866 --> 00:43:24.735 align:center line:-1 position:50% size:30% with time series data, so really handy for us. 00:43:24.735 --> 00:43:27.839 align:center line:-1 position:50% size:33% And, yeah, we have this in the field 00:43:27.839 --> 00:43:31.642 align:center line:-1 position:50% size:34% and it's operational now since February last year, 00:43:31.642 --> 00:43:33.744 align:center line:-1 position:50% size:25% and so we have quite a lot of data, 00:43:33.744 --> 00:43:36.113 align:center line:-1 position:50% size:28% also because before we actually installed 00:43:36.113 --> 00:43:38.783 align:center line:-1 position:50% size:44% all the LoRa connectivity, some probes were already there 00:43:38.783 --> 00:43:40.918 align:center line:-1 position:50% size:28% since October 2019. 00:43:40.918 --> 00:43:42.920 align:center line:-1 position:50% size:29% So quite a lot of data that we have, 00:43:42.920 --> 00:43:47.692 align:center line:-1 position:50% size:41% and all of this data is now being fed into Grafana, 00:43:47.692 --> 00:43:50.094 align:center line:-1 position:50% size:34% which is an open-source browser application 00:43:50.094 --> 00:43:53.064 align:center line:-1 position:50% size:32% that we're using to visualize all the data, 00:43:53.064 --> 00:43:56.434 align:center line:-1 position:50% size:36% but unfortunately, this needs a user account, 00:43:56.434 --> 00:43:58.703 align:center line:-1 position:50% size:35% so it's not a public service right now, 00:43:58.703 --> 00:44:01.239 align:center line:-1 position:50% size:31% but we're also working on a public data portal 00:44:01.239 --> 00:44:02.974 align:center line:-1 position:50% size:23% so that everyone can get access 00:44:02.974 --> 00:44:04.675 align:center line:-1 position:50% size:28% to the data that we're acquiring. 00:44:04.675 --> 00:44:08.779 align:center line:-1 position:50% size:21% And, you know, just to reiterate 00:44:08.779 --> 00:44:11.749 align:center line:-1 position:50% size:33% that, you know, changes in soil moisture 00:44:11.749 --> 00:44:13.718 align:center line:-1 position:50% size:42% and pore pressure really make a difference at this site, 00:44:13.718 --> 00:44:16.487 align:center line:-1 position:50% size:44% so here we see some examples of the data that we acquired. 00:44:16.487 --> 00:44:18.723 align:center line:-1 position:50% size:31% So on the top, we see just some data 00:44:18.723 --> 00:44:21.859 align:center line:-1 position:50% size:35% from the suction sensors, we see water potential 00:44:21.859 --> 00:44:24.595 align:center line:-1 position:50% size:36% which shows the response after one rainfall event 00:44:24.595 --> 00:44:27.965 align:center line:-1 position:50% size:35% that we had in April 2022, so last year 00:44:27.965 --> 00:44:31.002 align:center line:-1 position:50% size:28% where we had a rise in pore pressure 00:44:31.002 --> 00:44:33.070 align:center line:-1 position:50% size:42% that then caused the displacement, in this case, 00:44:33.070 --> 00:44:34.472 align:center line:-1 position:50% size:30% of just one centimeter. 00:44:34.472 --> 00:44:38.276 align:center line:-1 position:50% size:30% So, clearly, you know, a little bit of change 00:44:38.276 --> 00:44:42.246 align:center line:-1 position:50% size:43% in the hydrology in this case causes quite a bit of movement 00:44:42.246 --> 00:44:45.149 align:center line:-1 position:50% size:38% of at least the very shallows of surface. 00:44:45.149 --> 00:44:47.585 align:center line:-1 position:50% size:29% And so, just because you may have heard 00:44:47.585 --> 00:44:49.987 align:center line:-1 position:50% size:41% that we had pretty bad storms in the Bay Area 00:44:49.987 --> 00:44:53.958 align:center line:-1 position:50% size:35% or, you know, almost all of Western United States 00:44:53.958 --> 00:44:56.260 align:center line:-1 position:50% size:29% in the last few weeks and months, 00:44:56.260 --> 00:44:58.729 align:center line:-1 position:50% size:42% here's some of the data that we acquired during this period. 00:44:58.729 --> 00:45:00.998 align:center line:-1 position:50% size:26% So we're starting in December 2022, 00:45:00.998 --> 00:45:02.667 align:center line:-1 position:50% size:32% and basically what we're seeing here 00:45:02.667 --> 00:45:04.502 align:center line:-1 position:50% size:25% is the interpolated moisture content, 00:45:04.502 --> 00:45:07.939 align:center line:-1 position:50% size:43% and the moisture sensors are shown here in these diamonds, 00:45:07.939 --> 00:45:10.608 align:center line:-1 position:50% size:35% and then we can also see the recorded deformation 00:45:10.608 --> 00:45:14.912 align:center line:-1 position:50% size:25% that is shown in these dots here 00:45:14.912 --> 00:45:18.816 align:center line:-1 position:50% size:47% where the more red they become, the more movement they've seen. 00:45:18.816 --> 00:45:20.284 align:center line:-1 position:50% size:31% Basically, we start in September, 00:45:20.284 --> 00:45:21.786 align:center line:-1 position:50% size:43% so there's not much movement, 00:45:21.786 --> 00:45:24.288 align:center line:-1 position:50% size:43% there's not much moisture content in the ground. 00:45:24.288 --> 00:45:26.991 align:center line:-1 position:50% size:34% Most of it is actually less than five percent. 00:45:26.991 --> 00:45:29.627 align:center line:-1 position:50% size:28% We get into October, still no change. 00:45:29.627 --> 00:45:31.529 align:center line:-1 position:50% size:24% We see a little bit of creep here, 00:45:31.529 --> 00:45:37.301 align:center line:-1 position:50% size:42% which is just some of the creep in the eastern part of the area 00:45:37.301 --> 00:45:40.571 align:center line:-1 position:50% size:31% where we just have a constant bit of creep 00:45:40.571 --> 00:45:43.608 align:center line:-1 position:50% size:31% and it's also the old landslide sites. 00:45:43.608 --> 00:45:46.677 align:center line:-1 position:50% size:38% Then if we go to November, 00:45:46.677 --> 00:45:48.846 align:center line:-1 position:50% size:38% that's where we already got a few rains. 00:45:48.846 --> 00:45:50.715 align:center line:-1 position:50% size:45% We see that the moisture content is slightly increasing, 00:45:50.715 --> 00:45:53.351 align:center line:-1 position:50% size:46% but still there's not much, there's also not much movement. 00:45:53.351 --> 00:45:55.152 align:center line:-1 position:50% size:39% Now if we go into December, 00:45:55.152 --> 00:45:58.689 align:center line:-1 position:50% size:37% that's when we had quite a number of rainfall events, 00:45:58.689 --> 00:46:01.325 align:center line:-1 position:50% size:40% and we see that, you know, the moisture content is rising. 00:46:01.325 --> 00:46:04.428 align:center line:-1 position:50% size:34% We see now some more deformation, 00:46:04.428 --> 00:46:06.097 align:center line:-1 position:50% size:31% especially in this area. 00:46:06.097 --> 00:46:08.065 align:center line:-1 position:50% size:40% And now if we go to January, 00:46:08.065 --> 00:46:11.202 align:center line:-1 position:50% size:36% we really see how the site has completely saturated. 00:46:11.202 --> 00:46:12.870 align:center line:-1 position:50% size:31% Everything is basically fully saturated. 00:46:12.870 --> 00:46:14.772 align:center line:-1 position:50% size:42% The variation that we see here is just changes 00:46:14.772 --> 00:46:17.041 align:center line:-1 position:50% size:39% in the porosity of the soil which basically doesn't allow 00:46:17.041 --> 00:46:20.144 align:center line:-1 position:50% size:38% for higher moisture content, but we can also see 00:46:20.144 --> 00:46:23.848 align:center line:-1 position:50% size:36% very high movement rates up to seven centimeters. 00:46:23.848 --> 00:46:26.617 align:center line:-1 position:50% size:32% And give it a bit of time. 00:46:26.617 --> 00:46:28.819 align:center line:-1 position:50% size:35% Now we had a few weeks without much rain, 00:46:28.819 --> 00:46:30.888 align:center line:-1 position:50% size:34% actually moisture content decreased again, 00:46:30.888 --> 00:46:34.125 align:center line:-1 position:50% size:39% but, yeah, things had moved by then already, 00:46:34.125 --> 00:46:36.360 align:center line:-1 position:50% size:31% and we see again that, in some areas, 00:46:36.360 --> 00:46:39.964 align:center line:-1 position:50% size:42% we see higher movement rates on the top of the slope 00:46:39.964 --> 00:46:42.466 align:center line:-1 position:50% size:43% and then lower movement rates at the bottom, 00:46:42.466 --> 00:46:44.802 align:center line:-1 position:50% size:34% and we can also see this in, for instance, 00:46:44.802 --> 00:46:47.371 align:center line:-1 position:50% size:37% the western part of the site. 00:46:47.371 --> 00:46:51.042 align:center line:-1 position:50% size:34% And then if you look at the Grafana interface, 00:46:51.042 --> 00:46:52.343 align:center line:-1 position:50% size:32% that's what it looks like. 00:46:52.343 --> 00:46:55.146 align:center line:-1 position:50% size:42% And we see, here, we just plot the pore pressure, 00:46:55.146 --> 00:46:56.547 align:center line:-1 position:50% size:34% and basically, red is that, 00:46:56.547 --> 00:46:59.784 align:center line:-1 position:50% size:34% and almost all of the site is still in that condition 00:46:59.784 --> 00:47:02.386 align:center line:-1 position:50% size:41% where we have very high pore pressure across the site, 00:47:02.386 --> 00:47:04.955 align:center line:-1 position:50% size:42% and then, here we also see the moisture content recorded, 00:47:04.955 --> 00:47:08.059 align:center line:-1 position:50% size:43% and especially in January, we see how every rainfall event 00:47:08.059 --> 00:47:10.628 align:center line:-1 position:50% size:28% really leads to a rise in pore pressure, 00:47:10.628 --> 00:47:13.197 align:center line:-1 position:50% size:40% which basically is just the height that the rain brings 00:47:13.197 --> 00:47:18.269 align:center line:-1 position:50% size:41% just gives rise to the moisture content again. 00:47:18.269 --> 00:47:21.839 align:center line:-1 position:50% size:38% And just to kind of finish up, 00:47:21.839 --> 00:47:24.141 align:center line:-1 position:50% size:39% I want to show you a little bit of the work that we're doing 00:47:24.141 --> 00:47:26.577 align:center line:-1 position:50% size:45% with regards to machine learning and prediction. 00:47:26.577 --> 00:47:29.280 align:center line:-1 position:50% size:43% So, what we're trying to do here 00:47:29.280 --> 00:47:31.816 align:center line:-1 position:50% size:33% is basically we try to use machine learning 00:47:31.816 --> 00:47:34.685 align:center line:-1 position:50% size:31% to predict what's going to happen in the future 00:47:34.685 --> 00:47:38.022 align:center line:-1 position:50% size:40% to basically make this sensor network adaptive, 00:47:38.022 --> 00:47:43.728 align:center line:-1 position:50% size:34% so to basically save data and basically record 00:47:43.728 --> 00:47:46.964 align:center line:-1 position:50% size:36% in small sampling intervals when nothing's happening 00:47:46.964 --> 00:47:49.166 align:center line:-1 position:50% size:37% but then increase the sampling automatically 00:47:49.166 --> 00:47:51.035 align:center line:-1 position:50% size:35% when we expect changes to happen. 00:47:51.035 --> 00:47:52.470 align:center line:-1 position:50% size:34% And then the other thing, of course, 00:47:52.470 --> 00:47:54.438 align:center line:-1 position:50% size:29% is trying to predict further into the future 00:47:54.438 --> 00:47:56.640 align:center line:-1 position:50% size:36% to provide some more early warning. 00:47:56.640 --> 00:47:58.843 align:center line:-1 position:50% size:34% That's what we did here. 00:47:58.843 --> 00:48:02.179 align:center line:-1 position:50% size:42% So we tested two approaches, 00:48:02.179 --> 00:48:03.981 align:center line:-1 position:50% size:24% first predicting the next one hour 00:48:03.981 --> 00:48:05.649 align:center line:-1 position:50% size:44% based on the previous 24 hours, 00:48:05.649 --> 00:48:08.085 align:center line:-1 position:50% size:31% and second we predict the next 12 hours 00:48:08.085 --> 00:48:11.722 align:center line:-1 position:50% size:37% based on the prior 14 days of data. 00:48:11.722 --> 00:48:13.424 align:center line:-1 position:50% size:33% And here in these plots, 00:48:13.424 --> 00:48:16.060 align:center line:-1 position:50% size:30% basically, blue is the input data, 00:48:16.060 --> 00:48:21.332 align:center line:-1 position:50% size:38% green is the true data, and yellow is the prediction. 00:48:21.332 --> 00:48:25.803 align:center line:-1 position:50% size:34% And what we use for it is a one-year time series 00:48:25.803 --> 00:48:28.139 align:center line:-1 position:50% size:32% of actual and antecedent rainfall, 00:48:28.139 --> 00:48:30.007 align:center line:-1 position:50% size:28% we use temperature, relative humidity, 00:48:30.007 --> 00:48:33.677 align:center line:-1 position:50% size:34% and barometric pressure just as the, you know, 00:48:33.677 --> 00:48:36.847 align:center line:-1 position:50% size:42% kind of parameters that we use to train the model, 00:48:36.847 --> 00:48:38.849 align:center line:-1 position:50% size:32% and then we can see that actually the results 00:48:38.849 --> 00:48:41.185 align:center line:-1 position:50% size:40% are pretty good for predicting just one hour, 00:48:41.185 --> 00:48:45.089 align:center line:-1 position:50% size:39% so we're usually pretty close to the actual value, 00:48:45.089 --> 00:48:47.158 align:center line:-1 position:50% size:29% so we have very high prediction accuracy, 00:48:47.158 --> 00:48:49.827 align:center line:-1 position:50% size:34% but then if we look at longer-term forecasts, 00:48:49.827 --> 00:48:53.130 align:center line:-1 position:50% size:30% in this case, 12 hours, it's good if we have 00:48:53.130 --> 00:48:55.766 align:center line:-1 position:50% size:32% just gradation changes, as for instance, here. 00:48:55.766 --> 00:48:58.202 align:center line:-1 position:50% size:34% So we see that, yeah, prediction is pretty good, 00:48:58.202 --> 00:49:00.938 align:center line:-1 position:50% size:36% but it's pretty poor if we have sudden events. 00:49:00.938 --> 00:49:03.941 align:center line:-1 position:50% size:41% So where we have sudden increases in moisture content, 00:49:03.941 --> 00:49:06.177 align:center line:-1 position:50% size:45% it really doesn't work, and that's mostly because, here, 00:49:06.177 --> 00:49:07.778 align:center line:-1 position:50% size:28% we just trained on one year of data, 00:49:07.778 --> 00:49:10.648 align:center line:-1 position:50% size:44% and one year of data where we only had three storm events, 00:49:10.648 --> 00:49:12.616 align:center line:-1 position:50% size:31% so really there was not much data to train, 00:49:12.616 --> 00:49:15.085 align:center line:-1 position:50% size:35% especially for these kinds of sudden events. 00:49:15.085 --> 00:49:17.755 align:center line:-1 position:50% size:42% So, we think here that we have sufficient accuracy 00:49:17.755 --> 00:49:20.658 align:center line:-1 position:50% size:42% for adaptive sensing strategies but really we need more data 00:49:20.658 --> 00:49:23.227 align:center line:-1 position:50% size:36% to predict more accurately in the future, 00:49:23.227 --> 00:49:26.597 align:center line:-1 position:50% size:31% and that's also what recent papers all show 00:49:26.597 --> 00:49:29.567 align:center line:-1 position:50% size:31% that if you want to do a long-term prediction, 00:49:29.567 --> 00:49:31.936 align:center line:-1 position:50% size:30% usually it doesn't work that well 00:49:31.936 --> 00:49:33.804 align:center line:-1 position:50% size:30% and the accuracy isn't particularly good, 00:49:33.804 --> 00:49:35.406 align:center line:-1 position:50% size:26% and that's probably because you need 00:49:35.406 --> 00:49:37.308 align:center line:-1 position:50% size:32% quite a long time series to actually cover 00:49:37.308 --> 00:49:38.909 align:center line:-1 position:50% size:30% these extreme events that then cause, 00:49:38.909 --> 00:49:40.678 align:center line:-1 position:50% size:30% especially, landslides. 00:49:40.678 --> 00:49:42.913 align:center line:-1 position:50% size:36% And just before I finish up, 00:49:42.913 --> 00:49:45.249 align:center line:-1 position:50% size:44% I want to show you also some of the data from the Arctic 00:49:45.249 --> 00:49:47.885 align:center line:-1 position:50% size:33% just because, you know, in Berkeley, 00:49:47.885 --> 00:49:51.322 align:center line:-1 position:50% size:36% we saw seven centimeters of movement, so not a lot. 00:49:51.322 --> 00:49:55.092 align:center line:-1 position:50% size:31% Here we actually see-- well, in this case, 00:49:55.092 --> 00:49:58.462 align:center line:-1 position:50% size:41% it's only showing until five centimeters of movement, 00:49:58.462 --> 00:50:00.464 align:center line:-1 position:50% size:29% but we can see from these plots here 00:50:00.464 --> 00:50:03.200 align:center line:-1 position:50% size:44% that we actually see up to 30 or 40 centimeters of movement. 00:50:03.200 --> 00:50:05.569 align:center line:-1 position:50% size:40% So here, really, we have a lot of movement in the Arctic, 00:50:05.569 --> 00:50:07.738 align:center line:-1 position:50% size:40% and what I thought is very nice to show you here 00:50:07.738 --> 00:50:11.408 align:center line:-1 position:50% size:32% is also why having co-located temperature 00:50:11.408 --> 00:50:13.878 align:center line:-1 position:50% size:44% and deformation measurements make a lot of sense, 00:50:13.878 --> 00:50:15.980 align:center line:-1 position:50% size:34% because here, of course, we are in an area 00:50:15.980 --> 00:50:18.782 align:center line:-1 position:50% size:45% where we have permafrost, where we have unfrozen ground, 00:50:18.782 --> 00:50:21.185 align:center line:-1 position:50% size:41% and as I've already said earlier, what we would expect 00:50:21.185 --> 00:50:23.487 align:center line:-1 position:50% size:35% is that nothing's changing in the frozen ground 00:50:23.487 --> 00:50:26.390 align:center line:-1 position:50% size:46% but that we see mostly movement in the unfrozen ground, 00:50:26.390 --> 00:50:28.525 align:center line:-1 position:50% size:30% but it's something that you would expect 00:50:28.525 --> 00:50:31.428 align:center line:-1 position:50% size:44% but is very difficult to show except if you have these probes 00:50:31.428 --> 00:50:35.332 align:center line:-1 position:50% size:40% because what we see now is basically that we have 00:50:35.332 --> 00:50:38.469 align:center line:-1 position:50% size:44% most of the movement occurring exactly at this interface 00:50:38.469 --> 00:50:40.337 align:center line:-1 position:50% size:30% between the unfrozen and frozen ground, 00:50:40.337 --> 00:50:42.273 align:center line:-1 position:50% size:28% which is shown here just by the change 00:50:42.273 --> 00:50:44.341 align:center line:-1 position:50% size:29% in the angle of the accelerometer, 00:50:44.341 --> 00:50:48.445 align:center line:-1 position:50% size:44% so that's the gradient of the measured accelerometer rating, 00:50:48.445 --> 00:50:53.083 align:center line:-1 position:50% size:40% and it shows really nicely how basically the slip surface 00:50:53.083 --> 00:50:55.119 align:center line:-1 position:50% size:25% follows very nicely this interface 00:50:55.119 --> 00:50:57.588 align:center line:-1 position:50% size:35% between the frozen and the unfrozen ground. 00:50:57.588 --> 00:50:58.923 align:center line:-1 position:50% size:41% And this brings me to the end. 00:50:58.923 --> 00:51:00.591 align:center line:-1 position:50% size:30% So what I was hoping to show here 00:51:00.591 --> 00:51:02.693 align:center line:-1 position:50% size:33% is basically an approach that we're working on 00:51:02.693 --> 00:51:05.396 align:center line:-1 position:50% size:38% that tries to integrate various sensing techniques, 00:51:05.396 --> 00:51:07.231 align:center line:-1 position:50% size:44% so we integrate remote sensing, 00:51:07.231 --> 00:51:09.300 align:center line:-1 position:50% size:40% geophysical and environmental modeling, 00:51:09.300 --> 00:51:11.535 align:center line:-1 position:50% size:40% with hydrological and geomechanical modeling 00:51:11.535 --> 00:51:13.938 align:center line:-1 position:50% size:41% to get to a factor of safety assessment. 00:51:13.938 --> 00:51:16.473 align:center line:-1 position:50% size:39% Of course, this still needs development. 00:51:16.473 --> 00:51:20.044 align:center line:-1 position:50% size:37% So far, what I've shown is mostly a characterization 00:51:20.044 --> 00:51:23.480 align:center line:-1 position:50% size:39% of the site using a combined hydro-geomechanical model 00:51:23.480 --> 00:51:26.116 align:center line:-1 position:50% size:41% and some very simple models 00:51:26.116 --> 00:51:29.286 align:center line:-1 position:50% size:43% where we integrate geophysics with environmental monitoring 00:51:29.286 --> 00:51:31.555 align:center line:-1 position:50% size:30% to, for instance, track the groundwater table 00:51:31.555 --> 00:51:33.524 align:center line:-1 position:50% size:36% and see how this changes the factor of safety. 00:51:33.524 --> 00:51:35.125 align:center line:-1 position:50% size:23% But really where we want to go to 00:51:35.125 --> 00:51:36.627 align:center line:-1 position:50% size:30% is take all of this data, 00:51:36.627 --> 00:51:39.496 align:center line:-1 position:50% size:37% take the 3D distribution of the different parameters, 00:51:39.496 --> 00:51:41.799 align:center line:-1 position:50% size:34% and get really a nice 3D representation 00:51:41.799 --> 00:51:44.001 align:center line:-1 position:50% size:30% of the factor of safety. 00:51:44.001 --> 00:51:47.671 align:center line:-1 position:50% size:42% What we're still working on in terms of the sensor network 00:51:47.671 --> 00:51:49.740 align:center line:-1 position:50% size:31% and so on is that it's a public data portal 00:51:49.740 --> 00:51:51.308 align:center line:-1 position:50% size:25% so that everybody can access it, 00:51:51.308 --> 00:51:54.712 align:center line:-1 position:50% size:47% and we're also still working on some machine learning algorithms 00:51:54.712 --> 00:51:59.817 align:center line:-1 position:50% size:39% that maybe could avoid doing this kind of framework 00:51:59.817 --> 00:52:01.418 align:center line:-1 position:50% size:30% where you first do all the measurements 00:52:01.418 --> 00:52:03.787 align:center line:-1 position:50% size:38% then run the models, where maybe we could just, 00:52:03.787 --> 00:52:05.656 align:center line:-1 position:50% size:37% you know, run the models beforehand 00:52:05.656 --> 00:52:09.793 align:center line:-1 position:50% size:34% on a set of scenarios and then use this to train 00:52:09.793 --> 00:52:12.062 align:center line:-1 position:50% size:36% a machine learning model to actually then assess 00:52:12.062 --> 00:52:14.465 align:center line:-1 position:50% size:36% what our field data means basically in real time 00:52:14.465 --> 00:52:17.334 align:center line:-1 position:50% size:30% without having a need to run these models. 00:52:17.334 --> 00:52:19.470 align:center line:-1 position:50% size:35% And one thing that's nice, you know, 00:52:19.470 --> 00:52:21.905 align:center line:-1 position:50% size:38% all this work at first was just a research project, 00:52:21.905 --> 00:52:23.841 align:center line:-1 position:50% size:25% but actually now it's being taken up 00:52:23.841 --> 00:52:26.076 align:center line:-1 position:50% size:33% by the site management of Berkeley Lab 00:52:26.076 --> 00:52:30.414 align:center line:-1 position:50% size:35% which is taking our data and is actually, you know, 00:52:30.414 --> 00:52:31.915 align:center line:-1 position:50% size:31% is asking us what to do 00:52:31.915 --> 00:52:36.720 align:center line:-1 position:50% size:37% and how to try and mitigate the landslide hazard 00:52:36.720 --> 00:52:41.392 align:center line:-1 position:50% size:45% and asking us where do we have a very high hazard, 00:52:41.392 --> 00:52:43.127 align:center line:-1 position:50% size:39% where is the hazard smaller, 00:52:43.127 --> 00:52:45.963 align:center line:-1 position:50% size:36% what was now the impact of the storms that we saw, 00:52:45.963 --> 00:52:47.965 align:center line:-1 position:50% size:35% and so on, so really we're starting to integrate 00:52:47.965 --> 00:52:51.535 align:center line:-1 position:50% size:39% the data that we're gathering into the site response 00:52:51.535 --> 00:52:53.170 align:center line:-1 position:50% size:15% at this site. 00:52:53.170 --> 00:52:55.339 align:center line:-1 position:50% size:27% And, yeah, that's all I want to talk about. 00:52:55.339 --> 00:52:57.107 align:center line:-1 position:50% size:33% So thank you very much for your attention, 00:52:57.107 --> 00:52:59.309 align:center line:-1 position:50% size:42% and if you have any questions, I'd be very happy 00:52:59.309 --> 00:53:01.111 align:center line:-1 position:50% size:22% to answer them, and if not now, 00:53:01.111 --> 00:53:03.981 align:center line:-1 position:50% size:32% just send me an email and I'll get back to you.