-
Notifications
You must be signed in to change notification settings - Fork 0
/
research.html
286 lines (240 loc) · 29.1 KB
/
research.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
<!DOCTYPE HTML>
<html>
<head>
<title>Tim Lin</title>
<meta charset="utf-8" />
<meta name="viewport" content="width=device-width, initial-scale=1, user-scalable=no" />
<meta name="description" content="" />
<meta name="keywords" content="" />
<link rel="stylesheet" href="assets/css/main.css" />
<!--CSS RWD-->
<style>
.mergbox{
width:100%;border: 0px solid green; margin-left: auto;margin-right: auto;
}
.box12{
border: 0px solid blue;width:30%;display:inline-block;vertical-align: top;
width:62%;
}
.box3{
border: 0px solid yellow;width:35%;display:inline-block;vertical-align: top;
font-size:16px
}
p1:hover {
color:rgb(235, 112, 112);
/* background-color:black; */
transition: 0.3s;
}
div1:hover{
/* box-shadow: 0 0 0 50vmax rgba(0, 0, 0, .5); */
width:64%;
transition: 0.3s;
}
.tooltip {
position: relative;
display: inline-block;
border-bottom: 1px dotted black;
}
.tooltip .tooltiptext {
visibility: hidden;
width: 200px;
background-color: black;
color: #fff;
text-align: center;
border-radius: 5px;
padding: 2px 0;
/* Position the tooltip */
position: absolute;
z-index: 1;
}
.tooltip:hover .tooltiptext {
visibility: visible;
}
@media screen and (max-width:960px){
.box12{width:40%}
.box3{display:unset}
}
</style>
</head>
<body class="is-preload">
<!-- Header -->
<header id="header">
<a class="logo" href="index.html">Home</a>
<a class="logo" href="research.html">Research</a>
<a class="logo" href="publication_talk.html">Publications</a>
<a class="logo" href="https://github.com/jiunting">Github</a>
<a class="logo" href="https://drive.google.com/file/d/1Do5SogcdvddbXexriWdWdPXk1-cQLR_h/view?usp=sharing">CV</a>
<a class="logo" href="database.html">Database</a>
<a class="logo" href="https://sites.google.com/site/jtimweb/">Old Webpage</a>
<a class="logo"></a>
<a class="logo"></a>
<a class="logo"></a>
<nav>
<a href="#menu">Menu</a>
</nav>
</header>
<!-- Nav -->
<nav id="menu">
<ul class="links">
<li><a href="index.html">Home</a></li>
<li><a href="research.html">Research</a></li>
<li><a href="publication_talk.html">Publications</a></li>
<li><a href="https://github.com/jiunting">Github</a></li>
<li><a href="https://drive.google.com/file/d/1Do5SogcdvddbXexriWdWdPXk1-cQLR_h/view?usp=sharing">CV</a></li>
<li><a href="database.html">Database</a></li>
<li><a href="https://sites.google.com/site/jtimweb/">Old Webpage</a></li>
<li><a href="mailto:jiunting@uoregon.edu"> Contact me</a></li>
</ul>
</nav>
<!-- Banner -->
<section id="banner">
<div class="inner">
<!-- <h1 style="font-weight: bold;font-size:20px" >Hello!</h1> -->
<p style="font-weight: bold;font-size:46px; color:white">Research Highlights</p>
</div>
</section>
<!-- Highlights -->
<section class="wrapper">
<div class="inner">
<header class="special">
<p></p>
<div style="font-size: 28px;background-color: rgb(200, 200, 200,0.6); font-weight: bold; text-shadow:0 0 0px rgb(20, 20, 20);">Deep-Learning on EEW</div>
<div style="display: block; text-align:left; margin-inline-start: 0px; margin-inline-end: 0px; margin-block-start: 0.5em; margin-block-end: 2em; font-size: 18px;background-color: rgb(156, 206, 240,0); font-weight: normal; text-shadow:0 0 0px rgb(20, 20, 20);"> <a href="pubs/Lin_MLARGE_2021JB022703.pdf" style="text-decoration: none; color:cornflowerblue"><i class="icon fa-file-pdf-o"></i></a> <a href="https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2021JB022703" style="text-decoration: none;color:dimgray"><i class="icon fa-link"></i></a> Early Warning for Great Earthquakes From Characterization of Crustal Deformation Patterns With Deep Learning (2021). <i style="font-weight: normal">Journal of Geophysical Research-Solid Earth</i> </div>
<div class="mergbox">
<div1 class="box12" style="transition: 0.3s; border: 0px solid red; vertical-align:middle; font-size:22px; font-family: Helvetica, sans-serif; "> <a href="https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2021JB022703" target="_blank" rel="noopener noreferrer"> <img width="100%" src="./images/Exp_architecture.png" title="Click here to view the full article" ></img> </a></div1>
<div class="box3" style="vertical-align:top;"> <p style="font-weight: bold;margin-block-end:0px;font-size:22px;max-width: 100%;margin: 0 0 0 0;">Why EEW System?</p>Large earthquakes and the resulting tsunami are catastrophic. According to the report from Federal Emergency Management Agency and the US Geological Survey, the annual economic loss due to earthquakes in the U.S. amounts to $6.1 billion, and they often cause casualties. Earthquake early warning (EEW) systems aim to provide advanced warning to the public and automated systems, giving them a few seconds to minutes of notice before the onset of strong shaking and tsunamis. </div>
</div>
<div class="mergbox">
<!-- <div class="box12" style="border: 0px solid red; vertical-align:middle; width:62%; font-size:22px; font-family: Helvetica, sans-serif; ">M-LARGE architecture<img width="100%" src="./images/Exp_architecture.png"></img></div> -->
<div class="box3" style="vertical-align:middle"> <p style="font-weight: bold;margin-block-end:0px;font-size:22px;max-width: 100%;margin: 0 0 0 0;">Why Deep-Learning?</p> <p style="font-weight: bold;margin-block-end:0px;font-size:18px;max-width: 100%;margin: 0 0 0 0;">Earthquakes are complex in time and space</p> The existing models estimate earthquake magnitude by the earliest metrics of rupture signal-P waves from the near-field. However, there are three well-known issues in the current systems. <span style="color:rgb(0, 0, 0); font-weight:bold">First</span>, sensitive broadband sensors can be distorted and clipped in the near-field during strong shaking. <span style="color:rgb(0, 0, 0); font-weight:bold">Second</span>, the amplitude spectrum of large earthquakes saturates at higher frequencies, where accelerometers are most sensitive (this is called magnitude saturation). <span style="color:rgb(0, 0, 0); font-weight:bold">Finally</span>, the early signal of rupture may not provide the final information about the rupture (this is called non-determinism). As a result, determining large earthquake sources <span style="color:rgb(255, 60, 0); font-weight:bold">"quickly"</span> and <span style="color:rgb(255, 60, 0); font-weight:bold">"accurately"</span> has been proven to be challenging. This is a trade-off between accuracy & warning time.</div>
<div class="box12" style="border: 0px solid red; vertical-align:middle; width:62%; font-size:22px; font-family: Helvetica, sans-serif; "> <img width="80%" src="./images/EQexample1_1.png" onmouseover="this.src='./images/EQexample1_2.png';" onmouseout="this.src='./images/EQexample1_1.png';"></img> </div>
</div>
<div class="mergbox">
<!-- <div class="box3" style="border: 1px solid blue; vertical-align:middle"> <p style="font-weight: bold;margin-block-end:0px;font-size:22px;max-width: 100%; width:100%">M-LARGE performance: Maule earthquake</p></div> -->
<div style="font-size: 20px;background-color: rgb(156, 206, 240,0); font-weight: bold;">2010 Mw8.8 Maule earthquake: Magnitude, Location, and Source Extent prediction (blue rectangle)</div>
<div class="box12" style="border: 0px solid red; width:80%"> <video width="100%" height="100%" autoplay loop muted playsinline> <source src="./images/movie_Maule.mp4" type="video/mp4">Video Loading... </video> </div>
</div>
<a href="./images/movie_Maule.mp4">Or play the video link HERE</a>
<br> </br>
<hr></hr>
<br> </br>
<!-- The second part -->
<p></p>
<div style="font-size: 28px;background-color: rgb(200, 200, 200,0.6); font-weight: bold; text-shadow:0 0 0px rgb(20, 20, 20);">Earthquake Source Joint Inversion</div>
<div style="display: block; text-align:left; margin-inline-start: 0px; margin-inline-end: 0px; margin-block-start: 0.5em; margin-block-end: 2em; font-size: 18px;background-color: rgb(156, 206, 240,0); font-weight: normal; text-shadow:0 0 0px rgb(20, 20, 20);"> <a href="pubs/1-s2.0-S0012821X20302971-main.pdf" style="text-decoration: none; color:cornflowerblue"><i class="icon fa-file-pdf-o"></i></a> <a href="https://www.sciencedirect.com/science/article/pii/S0012821X20302971?via%3Dihub" style="text-decoration: none;color:dimgray"><i class="icon fa-link"></i></a> Overlapping regions of coseismic and transient slow slip on the Hawaiian décollement (2020). <i style="font-weight: normal">Earth and Planetary Science Letters</i> </div>
<div class="mergbox">
<div class="box3" style="vertical-align:middle"> <p style="font-weight: bold;margin-block-end:0px;font-size:22px;max-width: 100%;margin: 0 0 0 0;">Why Joint Inversion?</p> For most of the time we can not see the rupture processes during earthquakes. However, the emitted seismic waves carry the fingerprint of the source. From how large is the slip? how fast the fault has moved? what is the fault geometry? to what is the slip direction and how long is the rupture duration? Inversion is an <span style="color:rgb(255, 60, 0); font-weight:bold">"reverse engineering"</span>, and the goal is to find the source that best produces the observed data. With joint inversion model, we find a model that reconcile different types of data simultaneously, providing a better resolution of the earthquake source process. The figure shows an example of joint (a) GPS coseismic displacement, (c) strong motion accelememoter, and (d) tsunami tide gauges data to reveal the detailed rupture process of the 2018 Hawaii Big Island earthquake [<a href="https://earthquake.usgs.gov/earthquakes/eventpage/hv70116556/executive" style="text-decoration: none;color:rgb(0,40,255)">US Geological Survey</a>].</p></div>
<div1 class="box12" style="transition: 0.3s; border: 0px solid red; vertical-align:middle; font-size:22px; font-family: Helvetica, sans-serif; "> <a href="https://www.sciencedirect.com/science/article/pii/S0012821X20302971" target="_blank" rel="noopener noreferrer"> <img width="80%" src="./images/FigS1.png" title="Click here to view the full article"></img> </a></div1>
<!-- <div1 class="box12" style="border: 0px solid red; width:55%;vertical-align:middle"><img width="80%" src="./images/FigS1.png"></img></div1> -->
</div>
<div class="mergbox">
<div class="box12" style="border: 0px solid red; width:55%;vertical-align:middle"><img width="100%" src="./images/research_2_2.png"></img></div>
<div class="box3" style="vertical-align:middle"> <p style="font-weight: bold;margin-block-end:0px;font-size:22px;max-width: 100%;margin: 0 0 0 0;">What Did We Find?</p> In this study, we jointly invert the rupture history of the Mw7.1 2018 May 4 Hawaii earthquake by GPS, strong motion accelerometer, and additionally, calibrated tsunami data with a novel empirical tsunami approach. We explore the model space with more than 2000 inversion models. The result show that the earthquake (fast-slip) ruptured through an adjacent region that regularly hosts slow-slip. Additionally, we perform numerical modeling to simulate fast and slow slip behavior that supports our observations. This suggests that fault limit that usually constrained by coseismic slip is likely not a limit for a big event. As a result, earthquake hazard study should include the adjacent SSE regions for better hazard assessment. Check out our publication <div class="tooltip" style="color:rgb(0, 40, 255); font-size:bold" > HERE. <a class="tooltiptext" href="https://www.sciencedirect.com/science/article/pii/S0012821X20302971?via%3Dihub" target="_blank" rel="noopener noreferrer">Overlapping regions of coseismic and transient slow slip on the Hawaiian décollement</a> </div> </p></div>
</div>
<br> </br>
<hr></hr>
<br> </br>
<!-- The third part -->
<p></p>
<div style="font-size: 28px;background-color: rgb(200, 200, 200,0.6); font-weight: bold; text-shadow:0 0 0px rgb(20, 20, 20);">Signals Detection by Cross-correlation and Machine Learning</div>
<div style="display: block; text-align:left; margin-inline-start: 0px; margin-inline-end: 0px; margin-block-start: 0.5em; margin-block-end: 0em; font-size: 18px;background-color: rgb(156, 206, 240,0); font-weight: normal; text-shadow:0 0 0px rgb(20, 20, 20);"> <a href="pubs/JGR Solid Earth - 2020 - Yu - S Coda and Rayleigh Waves From a Decade of Repeating Earthquakes Reveal Discordant Temporal.pdf" style="text-decoration: none; color:cornflowerblue"><i class="icon fa-file-pdf-o"></i></a> <a href="https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2020JB019794" style="text-decoration: none;color:dimgray"><i class="icon fa-link"></i></a> S Coda and Rayleigh Waves From a Decade of Repeating Earthquakes Reveal Discordant Temporal Velocity Changes Since the 2004 Sumatra Earthquake (2020). <i style="font-weight: normal">Journal of Geophysical Research-Solid Earth</i> </div>
<div style="display: block; text-align:left; margin-inline-start: 0px; margin-inline-end: 0px; margin-block-start: 0.5em; margin-block-end: 2em; font-size: 18px;background-color: rgb(156, 206, 240,0); font-weight: normal; text-shadow:0 0 0px rgb(20, 20, 20);"> <a href="pubs/2021JB022896.pdf" style="text-decoration: none; color:cornflowerblue"><i class="icon fa-file-pdf-o"></i></a> <a href="https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2021JB022896" style="text-decoration: none;color:dimgray"><i class="icon fa-link"></i></a> Rayleigh-Love discrepancy highlights temporal changes in near-surface radial anisotropy after the 2004 Great Sumatra Earthquake (2021). <i style="font-weight: normal">Journal of Geophysical Research-Solid Earth</i> </div>
<div class="mergbox">
<div1 class="box12" style="transition: 0.3s; border: 0px solid red; vertical-align:middle; font-size:22px; font-family: Helvetica, sans-serif; "> <a href="#" target="_blank" rel="noopener noreferrer"> <img width="90%" src="./images/research_3_1.png" title="Click here to view the full article"></img> </a></div1>
<!-- <div1 class="box12" style="border: 0px solid red; vertical-align:middle; width:62%; font-size:22px; font-family: Helvetica, sans-serif; "> <img width="100%" src="./images/research_3_1.png"></img></div1> -->
<div class="box3" style="vertical-align:middle"> <p style="font-weight: bold;margin-block-end:0px;font-size:22px;max-width: 100%;margin: 0 0 0 0;">LFE Detection and Why?</p>Low frequency earthquakes (LFEs) are tiny earthquakes (with magnitude usually <M2) emitting seismic waves with relatively less high frequency contents. Non-volcanic LFEs can occurr at the down dip subduction zone boundary, shallow portion of subduction megathrust, and within the seismogenic zone. The process of LFE is usually associated with fluid interaction or partial melting at a particular pressure-temperature condition. Thus it is an important indicator to identify subduction zone extent, and to understand geological and geophysical properties of the subduction zone environment. The reason why a LFE detection algorithm is necessary is that unlike regular earthquakes, where waveform signals are obvious, LFEs signals are usually unclear and hidden in noisy time-series. The figure shows example of individual LFE data (gray lines), before stacking of many of such data (red line), clear earthquake signals are barely revealed. </div>
</div>
<!-- The 3-2 part -->
<div class="mergbox">
<div class="box3" style="vertical-align:top"> <p style="font-weight: bold;margin-block-end:0px;font-size:22px;max-width: 100%;margin: 0 0 0 0;">Traditional: Template Matching</p> Despite without apparent <span style="color:rgb(255, 60, 0); font-weight:bold">"waveforms"</span> for each individual event, the most common method to detect LFEs is template matching. In this method a LFE template is pre-defined, new events can be detected when cross-correlating the template and noisy continuous timeseries passes a certain threshold i.e. when the waveform is identical to the template, the correlation coefficient=1. To guarantee the detection is reliable, template matching has to consider the correlation coefficient of multiple stations. The searching procedure can be time consuming and the space complexity is expensive as the network and template grows bigger, but it is currently the most robust method and can be applied to various applications e.g. repeating earthquakes. Check out our publication <div class="tooltip" style="color:rgb(0, 40, 255); font-size:bold" > HERE <a class="tooltiptext" href="https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2020JB019794" target="_blank" rel="noopener noreferrer">S Coda and Rayleigh Waves From a Decade of Repeating Earthquakes Reveal Discordant Temporal Velocity Changes Since the 2004 Sumatra Earthquake</a> </div> for application of repeating earthquake searching, coda and Rayleigh waves analysis, and the code for template matching method (click the figure).</div>
<div1 class="box12" style="transition: 0.3s; border: 0px solid red; vertical-align:middle; font-size:22px; font-family: Helvetica, sans-serif; "> <a href="https://github.com/jiunting/RepEQ" target="_blank" rel="noopener noreferrer"> <img width="90%" src="./images/research_3_2.png" title="Click here to view the source code of template matching"></img> </a></div1>
</div>
<!-- The 3-3 part -->
<div class="mergbox">
<div class="box12" style="border: 0px solid red; vertical-align:middle; width:62%; font-size:22px; font-family: Helvetica, sans-serif; "> <img width="90%" src="./images/research_3_3_1.png" onmouseover="this.src='./images/research_3_3_2.png';" onmouseout="this.src='./images/research_3_3_1.png';"></img> </div>
<!-- <div1 class="box12" style="border: 0px solid red; vertical-align:middle; width:62%; font-size:22px; font-family: Helvetica, sans-serif; "> <img width="100%" src="./images/research_3_1.png"></img></div1> -->
<div class="box3" style="vertical-align:top"> <p style="display: block; margin-inline-start: 0px; margin-inline-end: 0px; margin-block-start: 1.5em; margin-block-end: 0em; font-weight: bold;margin-block-end:0px;font-size:22px;max-width: 100%;margin: 0 0 0 0;">New: Deep-Learning</p>In this study, we follow a U-shaped CNN based model first introduced in <a href="https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2021GL093157" target="_blank" rel="noopener noreferrer">Thomas et al. (2021)</a> to identidy LFE waveforms from continuous data in individual stations. The deep-learning model is more computationally efficient than the template matching method once the model has trained. Furthermore, it is more flexable since the detection does not rely on a fixed template, which the arrival times and template waveforms are pre-defined. According to our test it can identify the LFE waveforms from noisy timeseries data with an accuracy of 90%+ and detect new LFEs that are not previously identified. The figure shows the (normalized) number of daily LFE detections at each station. It detects more than 10 millions of arrivals over the 14 stations for over 10 years of recordings. The manuscript is in prep. please check back later for any updates.</div>
</div>
<br></br>
<hr></hr>
<br></br>
<!-- The 4th part -->
<p></p>
<div style="font-size: 28px;background-color: rgb(200, 200, 200,0.6); font-weight: bold; text-shadow:0 0 0px rgb(20, 20, 20);">Earthquake Rupture Simulations with HPC</div>
<div style="display: block; text-align:left; margin-inline-start: 0px; margin-inline-end: 0px; margin-block-start: 0.5em; margin-block-end: 2em; font-size: 18px;background-color: rgb(156, 206, 240,0); font-weight: normal; text-shadow:0 0 0px rgb(20, 20, 20);"> <a href="https://zenodo.org/record/5015610" style="text-decoration: none;color:dimgray"><i class="icon fa-database"></i></a> Chilean Subduction Zone rupture scenarios and waveform data (2020). <i style="font-weight: normal">Zenodo Dataset</i> </div>
<div class="mergbox">
<div class="box3" style="vertical-align:middle"> <p style="font-weight: bold;margin-block-end:0px;font-size:22px;max-width: 100%;margin: 0 0 0 0;">Why Numerical Simulation?</p> Large earthquakes are rare, but their physics are not-- they are the same as what we frequently observed in moderate and small size events. This is why and how earthquake inversion work i.e. seismologist uses Green's functions, which are small earthquakes, to <span style="color:rgb(255, 60, 0); font-weight:bold">"synthesize"</span> large earthquake. Simulations can be considered as real earthquakes, and with many of simulations, we can gain insights from them. Example of the applications including probabilistic tsunami hazard assessment (PTHA), earthquake hazard assessment, and most recently, machine-learning models e.g. identifying focal mechanisms, phase associations, P and S waves detection etc. in earthquake sciences. </div>
<div class="box12" style="border: 0px solid red; width:60%"><img width="100%" src="./images/research_4_1.png"></img></div>
</div>
<!-- The 4-2 part -->
<div class="mergbox">
<div class="box12" style="border: 0px solid red; vertical-align:middle; width:62%; font-size:22px; font-family: Helvetica, sans-serif; "> <img width="90%" src="./images/research_4_2.png" onmouseover="this.src='./images/research_4_2.png';" onmouseout="this.src='./images/research_4_2.png';"></img> </div>
<!-- <div1 class="box12" style="border: 0px solid red; vertical-align:middle; width:62%; font-size:22px; font-family: Helvetica, sans-serif; "> <img width="100%" src="./images/research_3_1.png"></img></div1> -->
<div class="box3" style="vertical-align:middle"> <p style="font-weight: bold;margin-block-end:0px;font-size:22px;max-width: 100%;margin: 0 0 0 0;">How Do We Simulate?</p> We apply rupture kinematics and HPC facility- <span style="font-weight: bold;font-size:16px; "><a href="https://hpcf.uoregon.edu/content/talapas" target="_blank" rel="noopener noreferrer">Talapas</a> </span>, to simulate large Chilean earthquakes. Details of the kinematic rupture modeling can be found in <a href="https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2016JB013314" target="_blank" rel="noopener noreferrer">Melgar et al. (2016)</a>. In this study we simulate thousands of rupture scenarios and ground motion waveforms using Talapas. These earthquakes follow the real global earthquake's source statistic including source duration, fault length, fault width, rupture speed etc. We also apply data augmentation with realistic GNSS noise model and station incompleteness for the synthetic waveforms. The dataset is publicly released and can be downloaded through Zenodo <div class="tooltip" style="color:rgb(0, 40, 255); font-size:bold"> HERE. <a class="tooltiptext" href="https://zenodo.org/record/5015610" target="_blank" rel="noopener noreferrer">Cascadia Subduction Zone Rupture Simulations and Synthetic Seismic Waveforms</a> </div> </div>
</div>
<br> </br>
<hr></hr>
<br> </br>
<!-- The 5th part -->
<p></p>
<div style="font-size: 28px;background-color: rgb(200, 200, 200,0.6); font-weight: bold; text-shadow:0 0 0px rgb(20, 20, 20);">Timeseries Model</div>
<div style="display: block; text-align:left; margin-inline-start: 0px; margin-inline-end: 0px; margin-block-start: 0.5em; margin-block-end: 2em; font-size: 18px;background-color: rgb(156, 206, 240,0); font-weight: normal; text-shadow:0 0 0px rgb(20, 20, 20);"> <a href="pubs/" style="text-decoration: none; color:cornflowerblue"><i class="icon fa-file-pdf-o"></i></a> <a href="https://academic.oup.com/gji/article/219/2/1148/5543903?login=true" style="text-decoration: none;color:dimgray"><i class="icon fa-link"></i></a> Quick determination of earthquake source parameters from GPS measurements: a study of suitability for Taiwan (2019). <i style="font-weight: normal">Geophysical Journal International</i> </div>
<div class="mergbox" style="vertical-align:middle">
<div1 class="box12" style="transition: 0.3s; border: 0px solid red; vertical-align:middle; font-size:22px; font-family: Helvetica, sans-serif; "> <a href="https://sites.google.com/site/jtimweb/gps-analysis/gps-model" target="_blank" rel="noopener noreferrer"> <img width="90%" src="./images/research_5.png" title="Click to see further detail"></img> </a><p style="font-size: 18px;">Click to see further detail</p> </div1>
<!-- <video width="320" height="240" controls> <source src="./images/movie_Maule.mp4" type="video/mp4"> </video> -->
<div class="box3" style="vertical-align:top"> <p style="font-weight: bold;margin-block-end:0px;font-size:22px;max-width: 100%; margin: 0 0 0 0;">Function-based regression</p> Raw GPS time series contains information of long-term plate motion, seasonal variation, earthquake deformation, and noise. By knowing these we build time series models to fit observations <img src="./images/GPSmodel_equation.png" width="100%"></img>Information such as inter-seismic, co-seismic, and post-seismic deformation can be extracted from the regression model. Our published work can be found <div class="tooltip" style="color:rgb(0, 40, 255); font-size:bold"> HERE. <a class="tooltiptext" href="https://academic.oup.com/gji/article/219/2/1148/5543903?login=true" target="_blank" rel="noopener noreferrer">Quick determination of earthquake source parameters from GPS measurements: a study of suitability for Taiwan</a> </div></div>
</div>
<br></br>
<hr></hr>
<br></br>
</header>
<!-- CTA -->
<section id="cta" class="wrapper" style= "background-color:black; background-image:linear-gradient(rgb(0, 0, 0,0), rgba(0, 0, 0, 0)), url(./images/bg1.png); height:400px " >
<div class="inner">
<p style="display: block;margin-block-start: 1em; margin-block-end: 0em;font-weight: bold; font-family:Arial, Helvetica, sans-serif; font-size:26px; text-shadow:0 0 20px rgb(20, 20, 20);">More coming soon</p>
</div>
</section>
</div>
</section>
<!-- Footer -->
<footer id="footer">
<div class="inner">
<div class="content">
<section>
<h3 style="margin: 0 0 0.5rem 0">Tim Lin</h3>
<h4 style="margin: 0 0 0.5rem 0">Earth and Computational Scientist</h4>
<p>Processed time series data with 10+ years of experience facilitating modeling and machine learning solutions to improve accuracy and efficiency. Designed NASA project on using deep learning time series forecasting model with significant earthquake warning improvement. Developed and optimized data mining signal detection tools to achieve large-scale data analysis with high performance computing. I like to think about what we can learn from data.</p>
</section>
<section>
<h4>Award & Grant Links</h4>
<ul class="alt">
<li><div class="tooltip"><a href="https://science.nasa.gov/researchers/sara/grant-stats/nasa-earth-and-space-science-fellowship-nessf-selections">NASA Earth & Space Science Fellow</a> <a class="tooltiptext" target="_blank" rel="noopener noreferrer">Click: NESSF 2018 Awards and Information -> Selections -> Earth Science -> then search the name: Jiun-Ting Lin</a> </div> </li>
<li><div class="tooltip"><a href="https://www.agu.org/Learn-and-Develop/Learn/Student-Competitions/OSPA/OSPA-Winners">Outstanding Student Presentation Award</a> <a class="tooltiptext" target="_blank" rel="noopener noreferrer">Click: 2019 Winners -> then search the name: Jiun-Ting Lin</a> </div> </li>
<li><div class="tooltip"><a href="https://www.gst.org.tw/cht/about_award.php?serial=7">Dr. Juan V.C. Best Paper Award</a> <a class="tooltiptext" target="_blank" rel="noopener noreferrer"> Search the 2017 Winners</a> </div> </li>
<li><a href="https://earthsciences.uoregon.edu/">Research Recognition Award @ UO</a></li>
</ul>
</section>
<section>
<h4>Social Media</h4>
<ul class="plain">
<!-- <li><a href="#"><i class="icon fa-twitter"> </i>Twitter</a></li> -->
<li><a href="https://www.facebook.com/supuky/"><i class="icon fa-facebook"> </i>Facebook</a></li>
<li><a href="https://www.linkedin.com/in/jiuntimlin/"><i class="icon fa-linkedin"> </i>LinkedIn</a></li>
<li><a href="https://github.com/jiunting"><i class="icon fa-github"> </i>Github</a></li>
<li><a href="https://sites.google.com/site/jtimweb/"><i class="icon fa-google"> </i>Google Site</a></li>
</ul>
</section>
</div>
<div class="copyright">
© 2021 Jiun-Ting Lin
</div>
</div>
</footer>
<!-- Scripts -->
<script src="assets/js/jquery.min.js"></script>
<script src="assets/js/browser.min.js"></script>
<script src="assets/js/breakpoints.min.js"></script>
<script src="assets/js/util.js"></script>
<script src="assets/js/main.js"></script>
</body>
</html>