<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE article PUBLIC "-//NLM//DTD Journal Publishing with OASIS Tables v3.0 20080202//EN" "journalpub-oasis3.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:oasis="http://docs.oasis-open.org/ns/oasis-exchange/table" xml:lang="en" dtd-version="3.0">
  <front>
    <journal-meta><journal-id journal-id-type="publisher">NPG</journal-id><journal-title-group>
    <journal-title>Nonlinear Processes in Geophysics</journal-title>
    <abbrev-journal-title abbrev-type="publisher">NPG</abbrev-journal-title><abbrev-journal-title abbrev-type="nlm-ta">Nonlin. Processes Geophys.</abbrev-journal-title>
  </journal-title-group><issn pub-type="epub">1607-7946</issn><publisher>
    <publisher-name>Copernicus Publications</publisher-name>
    <publisher-loc>Göttingen, Germany</publisher-loc>
  </publisher></journal-meta>
    <article-meta>
      <article-id pub-id-type="doi">10.5194/npg-26-61-2019</article-id><title-group><article-title>Inverting Rayleigh surface wave velocities for crustal thickness in eastern
Tibet and the western Yangtze craton based on deep learning neural networks</article-title><alt-title>Inverting Rayleigh surface wave velocities</alt-title>
      </title-group><?xmltex \runningtitle{Inverting Rayleigh surface wave velocities}?><?xmltex \runningauthor{X. Cheng et al.}?>
      <contrib-group>
        <contrib contrib-type="author" corresp="yes" rid="aff1">
          <name><surname>Cheng</surname><given-names>Xianqiong</given-names></name>
          <email>chxq@cdut.edu.cn</email>
        </contrib>
        <contrib contrib-type="author" corresp="no" rid="aff2">
          <name><surname>Liu</surname><given-names>Qihe</given-names></name>
          
        </contrib>
        <contrib contrib-type="author" corresp="no" rid="aff1">
          <name><surname>Li</surname><given-names>Pingping</given-names></name>
          
        </contrib>
        <contrib contrib-type="author" corresp="no" rid="aff1">
          <name><surname>Liu</surname><given-names>Yuan</given-names></name>
          
        </contrib>
        <aff id="aff1"><label>1</label><institution>College of Geophysics, Chengdu University of Technology, Chengdu,
P.R. China</institution>
        </aff>
        <aff id="aff2"><label>2</label><institution>School of Information and Software Engineering, University of
Electronic Science and Technology of China,<?xmltex \hack{\break}?> Chengdu, P.R. China</institution>
        </aff>
      </contrib-group>
      <author-notes><corresp id="corr1">Xianqiong Cheng (chxq@cdut.edu.cn)</corresp></author-notes><pub-date><day>17</day><month>April</month><year>2019</year></pub-date>
      
      <volume>26</volume>
      <issue>2</issue>
      <fpage>61</fpage><lpage>71</lpage>
      <history>
        <date date-type="received"><day>6</day><month>February</month><year>2018</year></date>
           <date date-type="rev-request"><day>9</day><month>March</month><year>2018</year></date>
           <date date-type="rev-recd"><day>19</day><month>March</month><year>2018</year></date>
           <date date-type="accepted"><day>2</day><month>April</month><year>2019</year></date>
      </history>
      <permissions>
        <copyright-statement>Copyright: © 2019 </copyright-statement>
        <copyright-year>2019</copyright-year>
      <license license-type="open-access"><license-p>This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this licence, visit <ext-link ext-link-type="uri" xlink:href="https://creativecommons.org/licenses/by/4.0/">https://creativecommons.org/licenses/by/4.0/</ext-link></license-p></license></permissions><self-uri xlink:href="https://npg.copernicus.org/articles/.html">This article is available from https://npg.copernicus.org/articles/.html</self-uri><self-uri xlink:href="https://npg.copernicus.org/articles/.pdf">The full text article is available as a PDF file from https://npg.copernicus.org/articles/.pdf</self-uri>
      <abstract><title>Abstract</title>
    <p id="d1e115">Crustal thickness is an important factor affecting lithospheric structure and
deep geodynamics. In this paper, a deep learning neural network based on a
stacked sparse auto-encoder is proposed for the inversion of crustal
thickness in eastern Tibet and the western Yangtze craton. First, with the
phase velocity of the Rayleigh surface wave as input and the theoretical
crustal thickness as output, 12 deep-sSAE neural networks are constructed,
which are trained by 380 000 and tested by 120 000 theoretical models. We
then invert the observed phase velocities through these 12 neural networks.
According to the test error and misfit of other crustal thickness models, the
optimal crustal thickness model is selected as the crustal thickness of the
study area. Compared with other ways to detect crustal thickness such as
seismic wave reflection and receiver function, we adopt a new way for
inversion of earth model parameters, and realize that a deep learning neural
network based on data driven with the highly non-linear mapping ability can
be widely used by geophysicists, and our result has good agreement with
high-resolution crustal thickness models. Compared with other methods, our
experimental results based on a deep learning neural network and a new
Rayleigh wave phase velocity model reveal some details: there is a
northward-dipping Moho gradient zone in the Qiangtang block and a relatively
shallow north-west–south-east oriented crust at the Songpan–Ganzi block.
Crustal thickness around Xi'an and the Ordos basin is shallow, about 35 km.
The change in crustal thickness in the Sichuan–Yunnan block is sharp, where
crustal thickness is 60 km north-west and 35 km south-east. We conclude
that the deep learning neural network is a promising, efficient, and
believable geophysical inversion tool.</p>
  </abstract>
    </article-meta>
  </front>
<body>
      

<sec id="Ch1.S1" sec-type="intro">
  <label>1</label><title>Introduction</title>
      <p id="d1e127">Eastern Tibet and the western Yangtze craton are one of the key areas for
understanding the collision process between the Indo-European plate and are
an important area for understanding the collision and contact relationship
between the Qinghai–Tibet Plateau and the Yangtze craton. In the field of
geosciences, because of the strong seismic activity, the nature of the two
blocks is different, especially the special topography. The altitude of the
two blocks suddenly rises from about 500 m in eastern Tibet to 5000 m in
the western Yangtze craton. Many studies focus on understanding the crust and
upper mantle structure in this region; there have especially been heated
debates on crustal thickness. The discontinuity between the crust and the
mantle is called Moho discontinuity, which varies greatly on a small scale
and is an important factor in geodynamics, including crustal evolution,
tectonic activity, gravity correction of the crustal effect, seismic
tomography, and geothermal models. Many studies focus on obtaining the depth
of Moho discontinuity called crustal thickness by various data and different
methods.</p>
      <p id="d1e130">Usually, the crustal thickness can be inverted by many kinds of data, such as
inversion of deep seismic sounding profiles on the Chinese mainland for
crustal thickness (Zeng et al., 1995), inversion of satellite gravity data
for global<?pagebreak page62?> crustal and lithospheric thickness (Fang, 1999), inversion of Bouguer gravity and topography data
to calculate the crustal thickness of China and its surrounding areas (Huang
et al., 2006; Guo et al., 2012), and an inversion receiver function used to
calculate the crustal thickness and Poisson's ratio of the Chinese mainland
(Chen et al., 2010; Zhu et al., 2012; Xu et al., 2007). In particular, one of
the newest models' crust 1.0 at <inline-formula><mml:math id="M1" display="inline"><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup><mml:mo>×</mml:mo><mml:mn mathvariant="normal">1</mml:mn><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:mrow></mml:math></inline-formula> (Laske et al.,
2013; Stolket al., 2013) is based on refraction and reflection seismology as
well as receiver function studies. In addition, these data related to crustal
thickness mentioned above, and crust thickness has significant effects on
fundamental mode surface waves (Meier et al., 2007; Grad and
Tiira, 2009). Dispersion characteristicz
of surface waves provide a powerful tool to research the structure of the
crust and upper mantle (Legendre et al., 2015). So far phase and group
velocity measurements of fundamental mode surface waves have most commonly
been used to constrain shear-velocity structure in the crust and upper mantle
on a global scale (Zhou et al., 2006; Shapiro and Ritzwoller, 2002) or on a
regional scale (Zhang et al., 2011; Yi et al., 2008); also, the newly
developed ambient noise surface wave tomography has been used to constrain
shear-velocity structure (Sun et al., 2010; Yao et al., 2006; Zheng et al.,
2008; Zhou et al., 2012), with a few works to invert fundamental mode surface
wave data for global or regional crustal thickness and to present a global or
regional crustal thickness model (Devile et al., 1999; Meier et al., 2007;
Das and Nolet 2001; Lebedev et al., 2013). Although the measurement period
and method of group velocity and phase velocity are different, the detection
depth and measurement error are also different, and phase velocity is more
sensitive to deeper structure, so it is easier to infer deep structure from
phase velocity measurements. We use phase velocity as input to infer the
crustal thickness.</p>
      <p id="d1e153">There are several inversion methods to get crustal thickness which can be
broadly classified into two classes: (1) the model-driven method and (2) the
data-driven method. For the model-driven method, the researchers mainly
consider the physical relation between earth parameter space and data space
to calculate an inversion function. Most model-driven methods deal with the
inversion of crustal thickness as a linear problem. More importantly, their
results largely depend on the initial earth model. Compared with model-driven
method, another fully non-linear data-driven method is called a neural
network, which is used to obtain crustal thickness (Devile et al., 1999;
Meier et al., 2007). Data-driven, highly non-linear mapping neural networks
are widely used in geophysical inversion methods, which use actual seismic
logging data and their attributes to predict the earth's parameters. Compared
with model-driven inversion, data-driven inversion does not need to consider
the physical relationship between the parameters of the earth model and data
space and can map and predict arbitrary non-linear relationship quickly and
accurately. Neural networks can be very useful in situations where the
forward relation is known, but the inverse mapping is unknown or difficult to
establish by more conventional analytical or numerical methods (de Wit et
al., 2013). So the aim of neural network inversion is to find the mapping
from a set of training data. Neural networks have been widely used in
different geophysical applications, well summarized by van der Baan and
Jutten (2000) such as in electrical impedance tomography (Lampinen, and
Vehtari, 2001) and in seismic processing including trace editing, travel time
picking, horizon tracking, and velocity analysis. Devilee et al. (1999) were
the first to use a neural network to invert surface wave velocities for
Eurasian crustal thickness in a fully non-linear and probabilistic manner.
Meier et al. (2007) further developed the methods of Devilee et al. (1999),
and then inverted surface wave data for global crustal thickness on a
<inline-formula><mml:math id="M2" display="inline"><mml:mrow><mml:mn mathvariant="normal">2</mml:mn><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup><mml:mo>×</mml:mo><mml:mn mathvariant="normal">2</mml:mn><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:mrow></mml:math></inline-formula> grid globally using a neural network.</p>
      <p id="d1e176">As seismology points out that there are many factors affecting phase
velocity, inverting phase velocity for discontinuities within the earth forms
a non-linear inverse problem (Meier et al., 2007). Because of strong
non-linear relations between crust thickness and surface wave dispersion, we
cannot treat it with a linear inverse problem as Montagner and Jobert (1988)
stated. Although a shallow neural network with a lower number of hidden
layers can present a non-linear inverse function, it maybe cannot learn or
approximate the true inverse function well when the true inverse function is
too complicated. In contrast, a deep learning neural network can overcome
this defect since it has powerful representation abilities and can discover
intricate structures in large data sets, because it makes use of the
back-propagation algorithm to indicate how a machine should change its
internal parameters that are used to compute the representation in each layer
from the representation in the previous layer (LeCun et al., 2015).</p>
      <p id="d1e180">In this paper, in view of the advantages and characteristics of a deep
learning neural network, a new fast inverse method based on a data-driven,
deep stacked sparse auto-encoders (sSAE) neural network is introduced to
solve the non-linear geophysical inverse problems. We focus on deep learning
neural networks to solve the non-linear inverse problem, and then apply them
to retrieve the crustal thickness for eastern Tibet and the western Yangtze
craton from the newest and high-resolution phase velocity maps. Based on
normal mode theory we compute phase velocities for the sampled radially
symmetric earth models to generate 500 000 theoretical models. First, the
theoretical phase velocity of a Rayleigh surface wave under random noise is
used as input to enhance the robustness of the neural network, and the
corresponding theoretical crustal thickness is used as output. Twelve deep
neural networks are constructed and trained by 380 000 and tested by
120 000 synthetic models. We then invert the observed phase velocities
through these 12 neural networks. According to test errors and misfits with
other crustal thickness models, the optimal crustal thickness model is
selected as the crustal thickness of the study area.</p>
      <?pagebreak page63?><p id="d1e183">To the best of our knowledge, we are the first to introduce deep learning
neural networks to learn and invert crustal thickness, and our results show
that crustal thickness is strongly non-linear with respect to phase velocity.
The merits of our methods include the following: first, since deep learning
neural networks can represent complex functions, it is possible to learn the
crustal thickness inverse function precisely. Using a deep learning neural
network, we can learn the relationship between surface wave phase velocity
and model parameters on the basis of large data (i.e. 500 000 theoretical
models in this study) and relax the a priori constraints on model parameters
(the crustal thickness is limited to 20–100 km). Secondly, inverse mapping
based on neural networks is of high efficiency because new observations can
be inverted instantaneously once well-trained deep learning neural networks
with multiple hidden layers are constructed. Thirdly, we can invert any
combination of model parameters without resampling model space (we will
invert crustal thickness and shear wave velocity simultaneously in future
work). Last but not least, the results show that when the number of hidden
layers reaches six and the test error is about 4.5e-6, the change in the number of neurons in
each layer has little effect on the test error, which indicates that the deep
learning neural network has strong robustness to the neural network structure
with appropriate layers. In the following, we will first briefly introduce
the deep learning neural network.</p>
</sec>
<sec id="Ch1.S2">
  <label>2</label><title>Deep learning neural networks</title>
      <p id="d1e194">In geophysics the real inverse function is usually a very complicated one
between data space and model space. The traditional linear inverse methods
that treat the real inverse function as a linear one can resolve the linear
relation problems. However, they depend on physical relationships between the
parameter space and initial earth model. A neural network has its origins in
attempts to find mathematical representations of information processing in
biological systems (Bishop, 1995). The deeper strength of artificial neural
networks (ANNs) is that more capabilities learn to infer complex, non-linear,
underlying relationships without any a priori knowledge of the model (Bengio,
2009). A shallow neural network has gained in popularity in geophysics in the
last decade and has been applied successfully to a variety of problems such
as well-log, interpretation of seismic data, or geophysical inversion.
Although a shallow neural network can present a non-linear inverse function,
it can only learn the relatively simple inverse function. In contrast, many
research results indicate that a deep learning neural network has a powerful
representation ability and can apply big geophysical observable data to learn
and approximate the complicated inverse function well (Lecun et al., 2015;
Bengio et al., 2006; Liu et al., 2015).</p>
      <p id="d1e197"><?xmltex \hack{\newpage}?>Based on the analysis above, we design a deep learning neural network to
obtain crustal thickness for eastern Tibet and the western Yangtze craton.
Compared with shallow neural networks, a deep learning neural network allows
computational models that are composed of multiple processing layers to learn
representations of data with multiple levels of abstraction and can learn
complex functions. The essence of deep learning is building an artificial
neural network with deep structures to simulate the analysis and
interpretation process of the human brain for data such as image, speech, or
text. However, many research results suggest that gradient-based training of
a deep neural network gets stuck in apparent local minima, which leads to
poor results in practice (Bengio, 2009). Fortunately, the greedy layer-wise
training algorithm proposed by Hinton and Salakhutdinov (2006) overcomes the optimization difficulty of deep
networks effectively. The training processing of deep neural networks is
divided into two steps. First, unsupervised learning methods are employed to
pre-train each layer's parameters with the output of the previous layer as
input, giving rise to initialize parameter values. After that, the
gradient-based method is used to finely tune the whole neural network
parameter values with respect to a supervised learning criterion as usual.
The advantage of the unsupervised pre-training method at each layer can help
guide the parameters of that layer towards better regions in parameter space
(Bengio, 2009). There are multiple types of deep learning neural networks,
such as convolutional neural networks, deep belief nets, and sSAE. sSAE works
very well in learning useful high-level features for better representation of
input raw data. Since the sSAE learning algorithm can automatically learn
even better feature representations than the hand-engineered ones, sSAE is
used widely in many domains such as computer vision, audio processing, and
natural language processing (Hinton and Salakhutdinov, 2006; Deng et al., 2013). Similar to these problems, we need to
extract earth feature representation from dispersion of surface waves. Here
we introduce sparse auto-encoders briefly, and a detailed description of the
network training method is given by Liu et al. (2015).</p>
      <?pagebreak page64?><p id="d1e201">The structure of sSAE is stacked by sparse auto-encoders to extract abstract
features. A typical sparse auto-encoder (SAE) can be seen as a neural network
with three layers, as shown in Fig. 1, including one input layer, one hidden
layer, and one output layer. The input vector and the output vector are
denoted by <inline-formula><mml:math id="M3" display="inline"><mml:mi mathvariant="bold-italic">v</mml:mi></mml:math></inline-formula> and <inline-formula><mml:math id="M4" display="inline"><mml:mover accent="true"><mml:mi mathvariant="bold-italic">v</mml:mi><mml:mo mathvariant="normal" stretchy="false">^</mml:mo></mml:mover></mml:math></inline-formula>, respectively. The matrix <inline-formula><mml:math id="M5" display="inline"><mml:mi mathvariant="bold">W</mml:mi></mml:math></inline-formula> is associated with the
connection between the input layer and the hidden layer. Similarly, the
matrix <inline-formula><mml:math id="M6" display="inline"><mml:mover accent="true"><mml:mi mathvariant="bold">W</mml:mi><mml:mo stretchy="false" mathvariant="normal">^</mml:mo></mml:mover></mml:math></inline-formula> connects the hidden layer to the output layer. The
vectors <inline-formula><mml:math id="M7" display="inline"><mml:mi mathvariant="bold-italic">b</mml:mi></mml:math></inline-formula> and <inline-formula><mml:math id="M8" display="inline"><mml:mover accent="true"><mml:mi mathvariant="bold-italic">b</mml:mi><mml:mo mathvariant="normal" stretchy="false">^</mml:mo></mml:mover></mml:math></inline-formula> are the bias vectors associated with
the units in the hidden layer and the output layer, respectively. The SAE is
trained to encode the input vector <inline-formula><mml:math id="M9" display="inline"><mml:mi mathvariant="bold-italic">v</mml:mi></mml:math></inline-formula> into some representation so that
the input can be reconstructed from that representation. Let <inline-formula><mml:math id="M10" display="inline"><mml:mrow><mml:mi>f</mml:mi><mml:mo>(</mml:mo><mml:mi>x</mml:mi><mml:mo>)</mml:mo></mml:mrow></mml:math></inline-formula> denote
the activation function, and the activation vector of the hidden layer then
is calculated (with an encoder) as

              <disp-formula id="Ch1.E1" content-type="numbered"><label>1</label><mml:math id="M11" display="block"><mml:mrow><mml:mstyle displaystyle="true" class="stylechange"/><mml:mi>z</mml:mi><mml:mo>=</mml:mo><mml:mi>f</mml:mi><mml:mo>(</mml:mo><mml:mi mathvariant="bold">W</mml:mi><mml:mi mathvariant="bold-italic">v</mml:mi><mml:mo>+</mml:mo><mml:mi mathvariant="bold-italic">b</mml:mi><mml:mo>)</mml:mo><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

        where <inline-formula><mml:math id="M12" display="inline"><mml:mi>z</mml:mi></mml:math></inline-formula> is the encoding result and some representation for the input
<inline-formula><mml:math id="M13" display="inline"><mml:mi mathvariant="bold-italic">v</mml:mi></mml:math></inline-formula>. The representation <inline-formula><mml:math id="M14" display="inline"><mml:mi>z</mml:mi></mml:math></inline-formula> or code is then mapped back (with a decoder
into a construction <inline-formula><mml:math id="M15" display="inline"><mml:mover accent="true"><mml:mi mathvariant="bold-italic">v</mml:mi><mml:mo mathvariant="normal" stretchy="false">^</mml:mo></mml:mover></mml:math></inline-formula> of the same shape as <inline-formula><mml:math id="M16" display="inline"><mml:mi mathvariant="bold-italic">v</mml:mi></mml:math></inline-formula>. The
mapping happens through a similar transformation, e.g.
          <disp-formula id="Ch1.E2" content-type="numbered"><label>2</label><mml:math id="M17" display="block"><mml:mrow><mml:mover accent="true"><mml:mi mathvariant="bold-italic">v</mml:mi><mml:mo mathvariant="normal" stretchy="false">^</mml:mo></mml:mover><mml:mo>=</mml:mo><mml:mi>f</mml:mi><mml:mo>(</mml:mo><mml:mover accent="true"><mml:mi mathvariant="bold">W</mml:mi><mml:mo stretchy="false" mathvariant="normal">^</mml:mo></mml:mover><mml:mi>z</mml:mi><mml:mo>+</mml:mo><mml:mover accent="true"><mml:mi mathvariant="bold-italic">b</mml:mi><mml:mo mathvariant="normal" stretchy="false">^</mml:mo></mml:mover><mml:mo>)</mml:mo><mml:mo>.</mml:mo></mml:mrow></mml:math></disp-formula>
        SAE is an unsupervised learning algorithm which sets the target values to be
equal to the inputs and constrain outputs of the hidden layer which are near
to zero and most hidden layers are inactive; the cost function is expressed
as

              <disp-formula specific-use="align" content-type="numbered"><mml:math id="M18" display="block"><mml:mtable displaystyle="true"><mml:mtr><mml:mtd><mml:mrow><mml:mstyle class="stylechange" displaystyle="true"/><mml:msub><mml:mi>J</mml:mi><mml:mi mathvariant="normal">sparse</mml:mi></mml:msub><mml:mfenced open="(" close=")"><mml:mrow><mml:mi mathvariant="bold">W</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="bold-italic">b</mml:mi></mml:mrow></mml:mfenced></mml:mrow></mml:mtd><mml:mtd><mml:mrow><mml:mstyle displaystyle="true" class="stylechange"/><mml:mo>=</mml:mo><mml:mi>J</mml:mi><mml:mfenced close=")" open="("><mml:mrow><mml:mi mathvariant="bold">W</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="bold-italic">b</mml:mi></mml:mrow></mml:mfenced><mml:mo>+</mml:mo><mml:mi mathvariant="italic">β</mml:mi><mml:munderover><mml:mo movablelimits="false">∑</mml:mo><mml:mrow><mml:mi>j</mml:mi><mml:mo>=</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow><mml:mrow><mml:msub><mml:mi>S</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:munderover><mml:mi mathvariant="italic">ρ</mml:mi><mml:mi>log⁡</mml:mi><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mi mathvariant="italic">ρ</mml:mi><mml:mrow><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">ρ</mml:mi><mml:mo stretchy="false" mathvariant="normal">^</mml:mo></mml:mover><mml:mi>j</mml:mi></mml:msub></mml:mrow></mml:mfrac></mml:mstyle></mml:mrow></mml:mtd></mml:mtr><mml:mlabeledtr id="Ch1.E3"><mml:mtd><mml:mtext>3</mml:mtext></mml:mtd><mml:mtd><mml:mstyle class="stylechange" displaystyle="true"/></mml:mtd><mml:mtd><mml:mrow><mml:mstyle displaystyle="true" class="stylechange"/><mml:mo>+</mml:mo><mml:mo>(</mml:mo><mml:mn mathvariant="normal">1</mml:mn><mml:mo>-</mml:mo><mml:mi mathvariant="italic">ρ</mml:mi><mml:mo>)</mml:mo><mml:mi>log⁡</mml:mi><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>-</mml:mo><mml:mi mathvariant="italic">ρ</mml:mi></mml:mrow><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>-</mml:mo><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">ρ</mml:mi><mml:mo mathvariant="normal" stretchy="false">^</mml:mo></mml:mover><mml:mi>j</mml:mi></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>.</mml:mo></mml:mrow></mml:mtd></mml:mlabeledtr></mml:mtable></mml:math></disp-formula>

          Here <inline-formula><mml:math id="M19" display="inline"><mml:mrow><mml:mi>J</mml:mi><mml:mfenced close=")" open="("><mml:mrow><mml:mi mathvariant="bold">W</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="bold-italic">b</mml:mi></mml:mrow></mml:mfenced></mml:mrow></mml:math></inline-formula> is the cost function without a
sparsity constraint, <inline-formula><mml:math id="M20" display="inline"><mml:mi mathvariant="italic">β</mml:mi></mml:math></inline-formula> controls the weight of the sparsity penalty
term, <inline-formula><mml:math id="M21" display="inline"><mml:mrow><mml:msub><mml:mi>S</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> is the number of neurons in the hidden layer, and the index
<inline-formula><mml:math id="M22" display="inline"><mml:mi>j</mml:mi></mml:math></inline-formula> sums over the hidden units in our network. <inline-formula><mml:math id="M23" display="inline"><mml:mrow><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">ρ</mml:mi><mml:mo stretchy="false" mathvariant="normal">^</mml:mo></mml:mover><mml:mi>j</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> is the
average activation of hidden unit <inline-formula><mml:math id="M24" display="inline"><mml:mi>j</mml:mi></mml:math></inline-formula>, and <inline-formula><mml:math id="M25" display="inline"><mml:mi mathvariant="italic">ρ</mml:mi></mml:math></inline-formula> is a sparsity parameter,
typically a small value close to zero.</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F1"><?xmltex \currentcnt{1}?><label>Figure 1</label><caption><p id="d1e573">An auto-encoder with one hidden layer (Liu et al., 2015).</p></caption>
        <?xmltex \igopts{width=170.716535pt}?><graphic xlink:href="https://npg.copernicus.org/articles/26/61/2019/npg-26-61-2019-f01.png"/>

      </fig>

      <p id="d1e582">Further, sSAE are neural networks consisting of multiple layers of SAE in
which SAE are stacked to form a deep neural network by feeding the
representation of the SAE found on the layer below as input to the current
layer. Using unsupervised pre-training methods, each layer is trained as sSAE
by minimizing the error in reconstructing its input, which is the output code
of the previous layer. After all layers are pre-trained, we add a logistic
regression layer on top of the network, and then train the entire network by
minimizing prediction error as we would train a traditional neural network.
For example, an sSAE with two hidden layers is shown in Fig. 2. This sSAE is
composed of two SAEs. The first SAE consists of the input layer and the first
hidden layer, and the representation or code of the input <inline-formula><mml:math id="M26" display="inline"><mml:mi mathvariant="bold-italic">v</mml:mi></mml:math></inline-formula> is
<inline-formula><mml:math id="M27" display="inline"><mml:mrow><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>=</mml:mo><mml:mi>f</mml:mi><mml:mo>(</mml:mo><mml:msub><mml:mi mathvariant="bold">W</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mi mathvariant="bold-italic">v</mml:mi><mml:mo>+</mml:mo><mml:msub><mml:mi mathvariant="bold-italic">b</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>)</mml:mo></mml:mrow></mml:math></inline-formula>. The second SAE comprises two
hidden layers, and the code of <inline-formula><mml:math id="M28" display="inline"><mml:mrow><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> is
<inline-formula><mml:math id="M29" display="inline"><mml:mrow><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mo>=</mml:mo><mml:mi>f</mml:mi><mml:mo>(</mml:mo><mml:msub><mml:mi mathvariant="bold">W</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mi mathvariant="bold-italic">b</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mo>)</mml:mo></mml:mrow></mml:math></inline-formula>. Each SAE is added to a
decoder layer as shown in Fig. 1, and we can then employ unsupervised
pre-training methods to train each SAE by expression (1). Finally, the matrix
<inline-formula><mml:math id="M30" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold">W</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M31" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold">W</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> and bias vector <inline-formula><mml:math id="M32" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">b</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> and
<inline-formula><mml:math id="M33" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">b</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> are
initialized. We then apply supervised fine-tuning methods to train an entire
network. Since our aim is to calculate crustal thickness and this is a
regression problem, we attach a layer connected fully with the last layer of
the encoder part (the matrix <inline-formula><mml:math id="M34" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold">W</mml:mi><mml:mi>s</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>). After that, we train this
network as done in a traditional neural network.</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F2"><?xmltex \currentcnt{2}?><label>Figure 2</label><caption><p id="d1e730">Stacked sparse auto-encoder with two hidden layers.</p></caption>
        <?xmltex \igopts{width=170.716535pt}?><graphic xlink:href="https://npg.copernicus.org/articles/26/61/2019/npg-26-61-2019-f02.png"/>

      </fig>

</sec>
<sec id="Ch1.S3">
  <label>3</label><title>Inverting surface wave data for crustal thickness</title>
      <p id="d1e747">As Meier et al. (2007) demonstrated, the neural network approach for solving
inverse problems is best summarized by three major steps as shown in Fig. 3:
(1) forward problem. In this stage we proceed by randomly sampling the model
space and solve the forward problem for all visited models based on seismic
wave normal mode theory. (2) Designing a neural network structure. In this
stage taking phase velocities with random noise as input and theoretical
crustal thickness as output, we train the deep learning neural networks and
the optimized neural network is obtained. (3) Inverse problem. Based on
trained networks, we invert crustal thickness from observed phase velocities.</p>
      <p id="d1e750">In what follows we introduce how to train sSAE deep learning neural networks
to model surface wave dispersion based on synthetic seismograms, and then
invert dispersion curves based on the trained networks. Finally we compare
our crustal model with other crustal thickness models, and discuss the
geodynamic implications of our model.</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F3"><?xmltex \currentcnt{3}?><label>Figure 3</label><caption><p id="d1e755">Crustal thickness inversion based on an sSAE neural network composed
of two parts: forward problem and inverse problem.</p></caption>
        <?xmltex \igopts{width=241.848425pt}?><graphic xlink:href="https://npg.copernicus.org/articles/26/61/2019/npg-26-61-2019-f03.png"/>

      </fig>

<?pagebreak page65?><sec id="Ch1.S3.SS1">
  <label>3.1</label><title>Data preparation</title>
      <p id="d1e772">We closely follow the model parameterization methodology outlined in de Wit
et al. (2014), which is based on the Preliminary Reference Earth Model (PREM,
Dziewonski and Anderson, 1981) and is parameterized on a discrete set of 185
grid points used by the Mineos package (Masters et al., 2014). In addition,
these models we have got show no correlations between physical parameters
such as velocity, density, anisotropic parameter, and attenuation
profiles. As with the model parameterization method mentioned above, we
generate 500 000 synthetic models based on 1-D reference model PREM, which
are randomly drawn from the prior model distribution; also, prior ranges for
the various parameters in our model are given in Tables A.2–A.4. of de Wit
et al. (2014). We use the Mineos package to compute phase velocity for
fundamental mode Rayleigh waves for all 500 000 synthetic 1-D earth models.
As for observation data used in the stage of inversion below, phase
velocities are more sensitive to the deep structure than group velocity.
Based on Rayleigh wave phase velocity from ambient noise (Xie et al., 2013)
shown in Fig. 4 averaged from 10 to 35 mHz, we take these as input for our
neural networks.</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F4"><?xmltex \currentcnt{4}?><label>Figure 4</label><caption><p id="d1e777">Average phase velocity of the western Yangtze craton (Xie et al.,
2013) from 10 to 35 mHz. The black lines in the figure show structure lines.
The blue lines show boundaries of sedimentary basins. The red dots show
seismic events in this region from 1975 to 2015, and the size of a dot
demonstrates the size in magnitude from <inline-formula><mml:math id="M35" display="inline"><mml:mrow><mml:msub><mml:mi>M</mml:mi><mml:mi mathvariant="normal">s</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> 6.0 to <inline-formula><mml:math id="M36" display="inline"><mml:mrow><mml:msub><mml:mi>M</mml:mi><mml:mi mathvariant="normal">s</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> 8.0. The yellow and purple
stars demonstrate the Wenchuan and Lushan earthquakes respectively. These are
the same as Figs. 4, 6, and 7.</p></caption>
          <?xmltex \igopts{width=236.157874pt}?><graphic xlink:href="https://npg.copernicus.org/articles/26/61/2019/npg-26-61-2019-f04.png"/>

        </fig>

</sec>
<sec id="Ch1.S3.SS2">
  <label>3.2</label><title>Training an sSAE deep learning neural network</title>
      <p id="d1e816">It is well known that the artificial neural network can approximate any
non-linear function to solve the non-linear inverse problem by using a
corresponding set of input–output pairs. These examples are presented to a
network in a so-called training process, during which the free parameters of
a network are modified to approximate the function of interests (de Wit et
al., 2014). Here adopting an sSAE deep learning neural network, with detailed
methods presented in Sect. 2 above, we pre-train the neural network, taking
the theoretical phase velocity of Rayleigh waves with random noise as inputs
and the theoretical crustal thickness as outputs to attain the initial
weights and bias for a neural network. And then we take the theoretical phase
velocity of Rayleigh waves with random noise as input and crustal thickness
as output to fine-tune the neural network as done in a traditional neural
network.</p>
      <p id="d1e819">How to find a satisfactory structure of a neural network is a difficult
problem because neural network training is sensitive to the random
initialization of the network parameters. Therefore, de Wit et al. (2014)
pointed out that it is common practice to train several neural networks with
different initializations and subsequently choose the network which performs
best on a given synthetic test data set, and the network which performed best
on the test set is used to draw inferences from the observed data (de Wit et
al., 2014). After trying many times, we find the proportion of the training
data set to test one is <inline-formula><mml:math id="M37" display="inline"><mml:mrow><mml:mn mathvariant="normal">3</mml:mn><mml:mo>:</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:math></inline-formula>, which is reasonable (Fig. 5). The final<?pagebreak page66?> test
error depends on not only the number of input neurons, hidden layers, and
intermediate neurons, but also on the number of trainings, batch sizes, and
other optional parameters. In addition, the type of activation function,
learning rate, zero masked fraction, and non-sparsity penalty value will
affect final test errors. We give 12 cases and their corresponding test
errors in Table 1.</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F5"><?xmltex \currentcnt{5}?><label>Figure 5</label><caption><p id="d1e836">The relationship between proportions of training data sets to test
data sets and test errors.</p></caption>
          <?xmltex \igopts{width=199.169291pt}?><graphic xlink:href="https://npg.copernicus.org/articles/26/61/2019/npg-26-61-2019-f05.png"/>

        </fig>

</sec>
<sec id="Ch1.S3.SS3">
  <label>3.3</label><title>Inverting crust thickness</title>
      <p id="d1e853">Based on all our 12 neural networks, we invert Rayleigh phase velocities
(10–35.0 mHz) to attain 12 crustal thickness models for eastern Tibet and
the western Yangtze craton. Considering not only the test errors of sSAE
networks but also misfits and correlation coefficients of our 12 models with
crustal thickness models from other studies, we choose the network structure
indicated by <inline-formula><mml:math id="M38" display="inline"><mml:msup><mml:mi/><mml:mo>*</mml:mo></mml:msup></mml:math></inline-formula> in Table 1. We find the best fit crustal thickness model
from sSAE (Fig. 6). We compare our model with the crustal thickness model
from the receiver function (Zhu et al., 2012), and the other two global
crustal thickness models, CRUST2.0 from Bassin et al. (2000) based on
refraction and reflection seismics as well as receiver function studies and
the CUB2 model from Shapiro and Ritzwoller (2002) (Fig. 7), who inverted a
similar data set for crustal thickness using a Monte Carlo approach. The
correlation coefficients and scatter plots of our model versus ZJS, our model
versus CRUST2.0, and our model versus CUB2 (Fig. 8) indicate overall
agreement between the three models. However, the agreements of our model with
CUB2 and CRUST2.0 are better than with ZJS, since model ZJS attained from Zhu
et al. (2012) has relatively sparse stations with poor data coverage and
lower resolution. In addition, taking the Monte Carlo method
(Hansen et al., 2013) using four
processors only for 1000 iterations, it takes 3 weeks to invert the
Xie et al. (2013) data set to the
crust thickness of the same region. As Shapiro and
Ritzwoller (2002) pointed out, the major
disadvantage of this method is computational expense. Maybe the result is
high resolution after many more iterations using the Monte Carlo method.
However, when we make use of sSAE, using only one processor, a six-layer
network for 380 000 training samples and 120 000 test samples for network
training takes 5 h, while a well-trained neural network inversion takes only
a few seconds to complete.</p>

<?xmltex \floatpos{t}?><table-wrap id="Ch1.T1" specific-use="star"><?xmltex \currentcnt{1}?><label>Table 1</label><caption><p id="d1e868">Deep learning neural network structures in this article.</p></caption><oasis:table frame="topbot"><oasis:tgroup cols="12">
     <oasis:colspec colnum="1" colname="col1" align="left"/>
     <oasis:colspec colnum="2" colname="col2" align="left"/>
     <oasis:colspec colnum="3" colname="col3" align="right"/>
     <oasis:colspec colnum="4" colname="col4" align="right"/>
     <oasis:colspec colnum="5" colname="col5" align="right"/>
     <oasis:colspec colnum="6" colname="col6" align="right"/>
     <oasis:colspec colnum="7" colname="col7" align="right"/>
     <oasis:colspec colnum="8" colname="col8" align="right" colsep="1"/>
     <oasis:colspec colnum="9" colname="col9" align="right"/>
     <oasis:colspec colnum="10" colname="col10" align="right" colsep="1"/>
     <oasis:colspec colnum="11" colname="col11" align="right"/>
     <oasis:colspec colnum="12" colname="col12" align="right"/>
     <oasis:thead>
       <oasis:row>
         <oasis:entry colname="col1">sSAE structure</oasis:entry>
         <oasis:entry rowsep="1" namest="col2" nameend="col5" align="center">Parameters </oasis:entry>
         <oasis:entry colname="col6">Error <inline-formula><mml:math id="M39" display="inline"><mml:mrow><mml:mo>×</mml:mo><mml:msup><mml:mn mathvariant="normal">10</mml:mn><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">6</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula></oasis:entry>
         <oasis:entry rowsep="1" namest="col7" nameend="col8" align="center" colsep="1">CUB2 </oasis:entry>
         <oasis:entry rowsep="1" namest="col9" nameend="col10" align="center" colsep="1">CRUST2.0 </oasis:entry>
         <oasis:entry rowsep="1" namest="col11" nameend="col12" align="center">ZJS </oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">Layers</oasis:entry>
         <oasis:entry colname="col3"><inline-formula><mml:math id="M40" display="inline"><mml:mi>D</mml:mi></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col4"><inline-formula><mml:math id="M41" display="inline"><mml:mi>E</mml:mi></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col5"><inline-formula><mml:math id="M42" display="inline"><mml:mi>F</mml:mi></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col6"/>
         <oasis:entry colname="col7"><inline-formula><mml:math id="M43" display="inline"><mml:mi>G</mml:mi></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col8"><inline-formula><mml:math id="M44" display="inline"><mml:mi>H</mml:mi></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col9"><inline-formula><mml:math id="M45" display="inline"><mml:mi>G</mml:mi></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col10"><inline-formula><mml:math id="M46" display="inline"><mml:mi>H</mml:mi></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col11"><inline-formula><mml:math id="M47" display="inline"><mml:mi>G</mml:mi></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col12"><inline-formula><mml:math id="M48" display="inline"><mml:mi>H</mml:mi></mml:math></inline-formula></oasis:entry>
       </oasis:row>
     </oasis:thead>
     <oasis:tbody>
       <oasis:row>
         <oasis:entry colname="col1">[21 50 10 1]</oasis:entry>
         <oasis:entry colname="col2">Layer 1</oasis:entry>
         <oasis:entry colname="col3">0.3</oasis:entry>
         <oasis:entry colname="col4">10</oasis:entry>
         <oasis:entry colname="col5">1e4</oasis:entry>
         <oasis:entry colname="col6">262</oasis:entry>
         <oasis:entry colname="col7">7.29</oasis:entry>
         <oasis:entry colname="col8">0.79</oasis:entry>
         <oasis:entry colname="col9">7.68</oasis:entry>
         <oasis:entry colname="col10">0.80</oasis:entry>
         <oasis:entry colname="col11">9.12</oasis:entry>
         <oasis:entry colname="col12">0.71</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">[21 50 10 1]</oasis:entry>
         <oasis:entry colname="col2">Layer 1</oasis:entry>
         <oasis:entry colname="col3">0.3</oasis:entry>
         <oasis:entry colname="col4">10</oasis:entry>
         <oasis:entry colname="col5">1e3</oasis:entry>
         <oasis:entry colname="col6">79.5</oasis:entry>
         <oasis:entry colname="col7">7.52</oasis:entry>
         <oasis:entry colname="col8">0.77</oasis:entry>
         <oasis:entry colname="col9">8.00</oasis:entry>
         <oasis:entry colname="col10">0.76</oasis:entry>
         <oasis:entry colname="col11">8.42</oasis:entry>
         <oasis:entry colname="col12">0.73</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">[21 50 10 1]</oasis:entry>
         <oasis:entry colname="col2">Layer 1</oasis:entry>
         <oasis:entry colname="col3">0.3</oasis:entry>
         <oasis:entry colname="col4">10</oasis:entry>
         <oasis:entry colname="col5">1e2</oasis:entry>
         <oasis:entry colname="col6">27.83</oasis:entry>
         <oasis:entry colname="col7">7.29</oasis:entry>
         <oasis:entry colname="col8">0.78</oasis:entry>
         <oasis:entry colname="col9">7.32</oasis:entry>
         <oasis:entry colname="col10">0.79</oasis:entry>
         <oasis:entry colname="col11">7.98</oasis:entry>
         <oasis:entry colname="col12">0.72</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">[21 50 10 1]</oasis:entry>
         <oasis:entry colname="col2">Layer 1</oasis:entry>
         <oasis:entry colname="col3">0.3</oasis:entry>
         <oasis:entry colname="col4">100</oasis:entry>
         <oasis:entry colname="col5">1e3</oasis:entry>
         <oasis:entry colname="col6">28.83</oasis:entry>
         <oasis:entry colname="col7">7.44</oasis:entry>
         <oasis:entry colname="col8">0.78</oasis:entry>
         <oasis:entry colname="col9">7.13</oasis:entry>
         <oasis:entry colname="col10">0.80</oasis:entry>
         <oasis:entry colname="col11">7.89</oasis:entry>
         <oasis:entry colname="col12">0.71</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">[21 50 10 1]</oasis:entry>
         <oasis:entry colname="col2">Layer 1</oasis:entry>
         <oasis:entry colname="col3">0.3</oasis:entry>
         <oasis:entry colname="col4">100</oasis:entry>
         <oasis:entry colname="col5">1e2</oasis:entry>
         <oasis:entry colname="col6">11.29</oasis:entry>
         <oasis:entry colname="col7">7.34</oasis:entry>
         <oasis:entry colname="col8">0.79</oasis:entry>
         <oasis:entry colname="col9">6.61</oasis:entry>
         <oasis:entry colname="col10">0.82</oasis:entry>
         <oasis:entry colname="col11">7.79</oasis:entry>
         <oasis:entry colname="col12">0.68</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">[21 50 10 1]</oasis:entry>
         <oasis:entry colname="col2">Layer 1</oasis:entry>
         <oasis:entry colname="col3">0.01</oasis:entry>
         <oasis:entry colname="col4">100</oasis:entry>
         <oasis:entry colname="col5">1e2</oasis:entry>
         <oasis:entry colname="col6">11.28</oasis:entry>
         <oasis:entry colname="col7">7.33</oasis:entry>
         <oasis:entry colname="col8">0.79</oasis:entry>
         <oasis:entry colname="col9">6.61</oasis:entry>
         <oasis:entry colname="col10">0.81</oasis:entry>
         <oasis:entry colname="col11">7.79</oasis:entry>
         <oasis:entry colname="col12">0.68</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">[21 10 2 1]</oasis:entry>
         <oasis:entry colname="col2">Layer 1</oasis:entry>
         <oasis:entry colname="col3">0.01</oasis:entry>
         <oasis:entry colname="col4">100</oasis:entry>
         <oasis:entry colname="col5">1e2</oasis:entry>
         <oasis:entry colname="col6">15.73</oasis:entry>
         <oasis:entry colname="col7">7.08</oasis:entry>
         <oasis:entry colname="col8">0.79</oasis:entry>
         <oasis:entry colname="col9">6.67</oasis:entry>
         <oasis:entry colname="col10">0.82</oasis:entry>
         <oasis:entry colname="col11">7.91</oasis:entry>
         <oasis:entry colname="col12">0.68</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">[21 100 50 20 1]</oasis:entry>
         <oasis:entry colname="col2">Layer 1</oasis:entry>
         <oasis:entry colname="col3">0.5</oasis:entry>
         <oasis:entry colname="col4">100</oasis:entry>
         <oasis:entry colname="col5">1e2</oasis:entry>
         <oasis:entry colname="col6">8.35</oasis:entry>
         <oasis:entry colname="col7">7.37</oasis:entry>
         <oasis:entry colname="col8">0.79</oasis:entry>
         <oasis:entry colname="col9">6.64</oasis:entry>
         <oasis:entry colname="col10">0.82</oasis:entry>
         <oasis:entry colname="col11">7.53</oasis:entry>
         <oasis:entry colname="col12">0.68</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">[21 200 50 20 10 1]</oasis:entry>
         <oasis:entry colname="col2">Layer 1</oasis:entry>
         <oasis:entry colname="col3">0.5</oasis:entry>
         <oasis:entry colname="col4">100</oasis:entry>
         <oasis:entry colname="col5">1e2</oasis:entry>
         <oasis:entry colname="col6">7.62</oasis:entry>
         <oasis:entry colname="col7">7.32</oasis:entry>
         <oasis:entry colname="col8">0.79</oasis:entry>
         <oasis:entry colname="col9">6.69</oasis:entry>
         <oasis:entry colname="col10">0.81</oasis:entry>
         <oasis:entry colname="col11">7.59</oasis:entry>
         <oasis:entry colname="col12">0.68</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">[21 200 100 50 20 10 5 1]<inline-formula><mml:math id="M49" display="inline"><mml:msup><mml:mi/><mml:mo>*</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">Layer 1</oasis:entry>
         <oasis:entry colname="col3">0.5</oasis:entry>
         <oasis:entry colname="col4">100</oasis:entry>
         <oasis:entry colname="col5">1e2</oasis:entry>
         <oasis:entry colname="col6">7.22</oasis:entry>
         <oasis:entry colname="col7">6.75</oasis:entry>
         <oasis:entry colname="col8">0.80</oasis:entry>
         <oasis:entry colname="col9">6.70</oasis:entry>
         <oasis:entry colname="col10">0.82</oasis:entry>
         <oasis:entry colname="col11">8.00</oasis:entry>
         <oasis:entry colname="col12">0.69</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">[21 200 100 50 20 10 5 1]</oasis:entry>
         <oasis:entry colname="col2">Layer 1</oasis:entry>
         <oasis:entry colname="col3">0.5</oasis:entry>
         <oasis:entry colname="col4">100</oasis:entry>
         <oasis:entry colname="col5">50</oasis:entry>
         <oasis:entry colname="col6">4.58</oasis:entry>
         <oasis:entry colname="col7">7.79</oasis:entry>
         <oasis:entry colname="col8">0.79</oasis:entry>
         <oasis:entry colname="col9">8.45</oasis:entry>
         <oasis:entry colname="col10">0.84</oasis:entry>
         <oasis:entry colname="col11">10.7</oasis:entry>
         <oasis:entry colname="col12">0.65</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">[21 50 40 30 20 10 5 1]</oasis:entry>
         <oasis:entry colname="col2">Layer 1</oasis:entry>
         <oasis:entry colname="col3">0.5</oasis:entry>
         <oasis:entry colname="col4">100</oasis:entry>
         <oasis:entry colname="col5">50</oasis:entry>
         <oasis:entry colname="col6">6.04</oasis:entry>
         <oasis:entry colname="col7">7.62</oasis:entry>
         <oasis:entry colname="col8">0.78</oasis:entry>
         <oasis:entry colname="col9">8.35</oasis:entry>
         <oasis:entry colname="col10">0.83</oasis:entry>
         <oasis:entry colname="col11">10.3</oasis:entry>
         <oasis:entry colname="col12">0.66</oasis:entry>
       </oasis:row>
     </oasis:tbody>
   </oasis:tgroup></oasis:table></table-wrap>

      <?xmltex \floatpos{t}?><fig id="Ch1.F6"><?xmltex \currentcnt{6}?><label>Figure 6</label><caption><p id="d1e1533">Crustal thickness of the western Yangtze craton from this paper.</p></caption>
          <?xmltex \igopts{width=236.157874pt}?><graphic xlink:href="https://npg.copernicus.org/articles/26/61/2019/npg-26-61-2019-f06.png"/>

        </fig>

      <p id="d1e1543">In this article, we fixed the following three parameters in every situation:
<inline-formula><mml:math id="M50" display="inline"><mml:mi>A</mml:mi></mml:math></inline-formula> – type of activation function
(sigmoid); <inline-formula><mml:math id="M51" display="inline"><mml:mi>B</mml:mi></mml:math></inline-formula> – learning rate (1); <inline-formula><mml:math id="M52" display="inline"><mml:mi>C</mml:mi></mml:math></inline-formula> – zero masked fraction (0.5).
Various parameters: <inline-formula><mml:math id="M53" display="inline"><mml:mi>D</mml:mi></mml:math></inline-formula> – non-sparsity penalty, which is zero except for
layer 1 in every sASE structure; <inline-formula><mml:math id="M54" display="inline"><mml:mi>E</mml:mi></mml:math></inline-formula> – number of epochs; <inline-formula><mml:math id="M55" display="inline"><mml:mi>F</mml:mi></mml:math></inline-formula> – size of
batch. <inline-formula><mml:math id="M56" display="inline"><mml:mi>G</mml:mi></mml:math></inline-formula> – RMS misfit of our result with other model; <inline-formula><mml:math id="M57" display="inline"><mml:mi>H</mml:mi></mml:math></inline-formula> – correlation
coefficient of our result with another model. <inline-formula><mml:math id="M58" display="inline"><mml:msup><mml:mi/><mml:mo>*</mml:mo></mml:msup></mml:math></inline-formula> – selected sSAE neural
network structure.</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F7"><?xmltex \currentcnt{7}?><label>Figure 7</label><caption><p id="d1e1614">Crustal thickness of model CUB2 from Shapiro and Ritzwoller (2002).
Note: the colour scale is not exactly the same as Fig. 6.</p></caption>
          <?xmltex \igopts{width=236.157874pt}?><graphic xlink:href="https://npg.copernicus.org/articles/26/61/2019/npg-26-61-2019-f07.png"/>

        </fig>

      <?xmltex \floatpos{t}?><fig id="Ch1.F8" specific-use="star"><?xmltex \currentcnt{8}?><label>Figure 8</label><caption><p id="d1e1625"><bold>(a–c)</bold> Scatter plots of our model versus ZJS, our model
versus CRUST2.0, and our model versus CUB2.</p></caption>
          <?xmltex \igopts{width=426.791339pt}?><graphic xlink:href="https://npg.copernicus.org/articles/26/61/2019/npg-26-61-2019-f08.png"/>

        </fig>

</sec>
</sec>
<sec id="Ch1.S4">
  <label>4</label><title>Discussion</title>
      <p id="d1e1645">On the one hand, our results show that deep learning neural networks can
effectively invert crustal thickness because they have the ability to
represent complex inverse functions.</p>
      <p id="d1e1648">A deep neural network can offer improvement over a shallow neural network as
shown in Table 1. Test errors of a deep learning neural network may be
affected by the number of<?pagebreak page67?> hidden layers in networks. The test error of a deep
learning neural network may be affected by the number of hidden layers in the
network, which indicates that the more hidden layers, the smaller the test
error. When the number of hidden layers in networks increases from three to
six, we can attain from Table 1 that the test error decreases from <inline-formula><mml:math id="M59" display="inline"><mml:mrow><mml:mn mathvariant="normal">2.6</mml:mn><mml:mo>×</mml:mo><mml:msup><mml:mn mathvariant="normal">10</mml:mn><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">4</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> to <inline-formula><mml:math id="M60" display="inline"><mml:mrow><mml:mn mathvariant="normal">6.0</mml:mn><mml:mo>×</mml:mo><mml:msup><mml:mn mathvariant="normal">10</mml:mn><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">6</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>. In addition, the robustness of deep
learning neural networks is strong. When the number of hidden layers in a
network reaches six, the change in the number of neurons in each layer has
little influence on test errors, about <inline-formula><mml:math id="M61" display="inline"><mml:mrow><mml:mn mathvariant="normal">5.5</mml:mn><mml:mo>×</mml:mo><mml:msup><mml:mn mathvariant="normal">10</mml:mn><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">6</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>.</p>
      <p id="d1e1705">In addition, we conclude that different training parameters have different
effects on training results. We think that the size of a batch is more
important than epochs, as shown in Table 1. The size of a batch decreases
from <inline-formula><mml:math id="M62" display="inline"><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>×</mml:mo><mml:msup><mml:mn mathvariant="normal">10</mml:mn><mml:mn mathvariant="normal">4</mml:mn></mml:msup></mml:mrow></mml:math></inline-formula> to <inline-formula><mml:math id="M63" display="inline"><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>×</mml:mo><mml:msup><mml:mn mathvariant="normal">10</mml:mn><mml:mn mathvariant="normal">3</mml:mn></mml:msup></mml:mrow></mml:math></inline-formula> and test errors decrease from <inline-formula><mml:math id="M64" display="inline"><mml:mrow><mml:mn mathvariant="normal">2.6</mml:mn><mml:mo>×</mml:mo><mml:msup><mml:mn mathvariant="normal">10</mml:mn><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">4</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> to <inline-formula><mml:math id="M65" display="inline"><mml:mrow><mml:mn mathvariant="normal">7.9</mml:mn><mml:mo>×</mml:mo><mml:msup><mml:mn mathvariant="normal">10</mml:mn><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">5</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>; however, epochs increase from 10 to
100, and corresponding test errors change a little. The neural network
structure indicated by <inline-formula><mml:math id="M66" display="inline"><mml:msup><mml:mi/><mml:mo>*</mml:mo></mml:msup></mml:math></inline-formula> in Table 1 reveals misfits of our model with
models CUB2, CRUST2.0, and ZJS are relatively low at 6.75, 6.70, and 8.0, and
corresponding correlation coefficients are relatively high at 0.8, 0.82, and
0.69 respectively; however, the test error is <inline-formula><mml:math id="M67" display="inline"><mml:mrow><mml:mn mathvariant="normal">7.22</mml:mn><mml:mo>×</mml:mo><mml:msup><mml:mn mathvariant="normal">10</mml:mn><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">6</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> and is
not minimum. This tells us that the test error may not be the only criterion
determining which neural network is best, because over-fitting may lead to
small test errors.</p>
      <p id="d1e1802">Compared with the works of Meier et al. (2007), in order to enhance
robustness of neural networks, random noise is added to synthetic phase
velocity as input in training progress. However, we have not considered the
uncertainty of crustal thickness, which should be revealed by the deep
mixture density network in a probabilistic manner in our future work.</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F9"><?xmltex \currentcnt{9}?><label>Figure 9</label><caption><p id="d1e1808">The relative error of the average phase velocity between the phase
velocity calculated by our model and the phase velocity observed by Xie et al. (2013).</p></caption>
        <?xmltex \igopts{width=236.157874pt}?><graphic xlink:href="https://npg.copernicus.org/articles/26/61/2019/npg-26-61-2019-f09.png"/>

      </fig>

      <p id="d1e1817">On the other hand, we can attain the crustal thickness and resultant
geodynamic implications in a research region from our result. We find
relatively good agreement of our result (Fig. 6) with CUB2 (Fig. 7),
CRUST2.0, (Fig. 8) and other recent studies (Qian et al., 2018; Xu et al.,
2016; Wang et al., 2015). All these models indicate that crustal thickness is
deeper to the west of the Longmen Mountains than to the east of the Longmen
Mountains, which all showed an approximately 45 km thick crust below the
Sichuan basin, thickening beneath the Longmen Mountains and the high Tibetan
Plateau to about 60–80 km. Since we make use of a regional<?pagebreak page68?> high-resolution
phase velocity model, our result reveals more details. In order to prove that
these anomalies are persistent, only accidents, or inversion artifacts, we
verify two aspects: one is that because our model gives a point estimate
without uncertainty information, we propose the relative error between the
phase velocity calculated based on our model and the phase velocity observed
in the same region (Xie et al., 2013) to verify the uncertainty and
non-uniqueness of the inversion results (Figs. 9 and 10). The relative error
calculation formula is shown in Eq. (4).
          <disp-formula id="Ch1.E4" content-type="numbered"><label>4</label><mml:math id="M68" display="block"><mml:mrow><mml:mi mathvariant="normal">RE</mml:mi><mml:mo>=</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:mi mathvariant="normal">|</mml:mi><mml:msub><mml:mi mathvariant="normal">Pha</mml:mi><mml:mi mathvariant="normal">cal</mml:mi></mml:msub><mml:mo>-</mml:mo><mml:msub><mml:mi mathvariant="normal">Pha</mml:mi><mml:mi mathvariant="normal">obs</mml:mi></mml:msub><mml:mi mathvariant="normal">|</mml:mi></mml:mrow><mml:mrow><mml:msub><mml:mi mathvariant="normal">Pha</mml:mi><mml:mi mathvariant="normal">obs</mml:mi></mml:msub></mml:mrow></mml:mfrac></mml:mstyle></mml:mrow></mml:math></disp-formula>
        RE – relative error of average phase velocity; Pha<inline-formula><mml:math id="M69" display="inline"><mml:msub><mml:mi/><mml:mi mathvariant="normal">cal</mml:mi></mml:msub></mml:math></inline-formula> –
calculated average phase velocity based on our model; Pha<inline-formula><mml:math id="M70" display="inline"><mml:msub><mml:mi/><mml:mi mathvariant="normal">obs</mml:mi></mml:msub></mml:math></inline-formula> –
observed average phase velocity from Xie et al. (2013).</p>
      <p id="d1e1873">The relative error between east and west is significantly different (Fig. 9).
The reason is that in our research area, as shown in Fig. 1b of Xie et
al. (2013), the stations in the east are much denser than in the west, and
the Rayleigh wave measurements has a higher resolution in the east.
Therefore, compared with the west, the training data in the east are more
dense, resulting in higher prediction accuracy of the neural network in the
east and lower prediction accuracy in the west. Using the same parameters but
for a traditional shallow neural network with three layers (one input layer,
one intermediate layer, and one output layer), we train this shallow neural
network and obtain a relative error with the observed average phase velocity.
The histogram error of the relative error of the shallow network is larger
than that of our deep learning network (Fig. 10). We can conclude that our
model is more reliable in the eastern part of the Longmen Mountains,
especially in Chengdu, the Qinling–Dabei fold belt, the Xi'an and Ordos
basins, and the Sichuan–Yunnan block, and so the crust thickness anomaly in
these areas is worth explaining. Another reference is to the results of other
studies conducted by C. Y. Wang et al. (2010) in the same area, who attained the crustal thickness estimated by the
<inline-formula><mml:math id="M71" display="inline"><mml:mi>H</mml:mi></mml:math></inline-formula>–<inline-formula><mml:math id="M72" display="inline"><mml:mi>k</mml:mi></mml:math></inline-formula> stacking method based on the broadband tele-seismic data recorded at
132 seismic stations in the Longmen Mountains and adjacent regions
(26–35<inline-formula><mml:math id="M73" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> N, 98–109<inline-formula><mml:math id="M74" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> E) (Fig. 11).</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F10"><?xmltex \currentcnt{10}?><label>Figure 10</label><caption><p id="d1e1910">The histogram of the relative error of the shallow neural network
(SNN) and deep neural network (DNN).</p></caption>
        <?xmltex \igopts{width=213.395669pt}?><graphic xlink:href="https://npg.copernicus.org/articles/26/61/2019/npg-26-61-2019-f10.png"/>

      </fig>

      <?xmltex \floatpos{t}?><fig id="Ch1.F11"><?xmltex \currentcnt{11}?><label>Figure 11</label><caption><p id="d1e1921">Contour map of crustal thickness estimated by <inline-formula><mml:math id="M75" display="inline"><mml:mi>h</mml:mi></mml:math></inline-formula>–<inline-formula><mml:math id="M76" display="inline"><mml:mi>k</mml:mi></mml:math></inline-formula> stacking
analysis based on the receiver function (C. Y. Wang et al., 2010). Small black
solid circles indicate the seismic station used by C. Y. Wang et al. (2010).</p></caption>
        <?xmltex \igopts{width=236.157874pt}?><graphic xlink:href="https://npg.copernicus.org/articles/26/61/2019/npg-26-61-2019-f11.png"/>

      </fig>

      <p id="d1e1945">Our result reveals similar details to C. Y. Wang et al. (2010): the crustal
thickness of the eastern Tibetan Plateau is complex and varies greatly. The
average crust thickness is about above 60 km, especially about 70–75 km at
the Qiangtang block, under which there is a northward-dipping Moho gradient
zone. There is a relatively shallow crust at the Songpa–Ganzi block and is
characteristic of a decrease in northwest–southeast orientation. Our model
still shows some<?pagebreak page69?> changes in the thickness of the crust in this area. For
example, the thickness of the crust around Chengdu is relatively thin,
especially in the north-eastern part of Chengdu, about 50 km thick under the
Qinlin–Dabei fold belt, and the thickness of the crust in the north-east to
the Sichuan basin is about 45–48 km. In addition, crustal thickness around
Xi'an and the Ordos basin is about 35 km. By contrast, a change in the
crustal thickness in the Sichuan–Yunnan block varies greatly, with a
thickness of about 60 km in the north-west and about 35 km in the
south-east. From a geological viewpoint, eastern Tibet and the western
Yangtze craton have a very complex structure and tectonics, where several
tectonic blocks, including the Yangtze Platform, the Songpan–Ganzi Fold
System, the Qiangtang block, and the Indochina block, interact with each
other. It is a site of important processes associated with the India–Asia
collision and abutment against the stable Yangtze Platform, including strong
compressional deformation with crust shortening and thickening; the plateau
surface has been elevated to 4–5 km, and the Tibetan crust has doubled in
thickness since the collision (Chen and Wilson, 1996; Flesch et al., 2005; Z. Wang et al., 2010). East–west crustal
extension and strong earthquakes often occur on the active faults inside and
on the edge of the plateau and are the most active seismic areas within the
mainland. Based on the analysis of the distribution of the epicentres during
1970–2015, the results show that a large earthquake occurred in the brittle
upper crust of the Longmenshan fault zone in Sichuan and Yunnan, and the
crustal thickness changed sharply by about 10 km. The <inline-formula><mml:math id="M77" display="inline"><mml:mrow><mml:msub><mml:mi>M</mml:mi><mml:mi mathvariant="normal">s</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> 8.0
Wenchuan earthquake in 2008 and the <inline-formula><mml:math id="M78" display="inline"><mml:mrow><mml:msub><mml:mi>M</mml:mi><mml:mi mathvariant="normal">s</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> 7.0 Lushan earthquake in
2013 were caused by the reactions associated with the Songpan–Ganzi Fold
System and the Qiangtang block obliquely colliding with the Yangtze Platform.
The reason may be that main fault cut off the Moho discontinuity where
materials exchange between crust and mantle, and accumulated press triggers a
series of earthquakes frequently.</p>
</sec>
<sec id="Ch1.S5" sec-type="conclusions">
  <label>5</label><title>Conclusion and remarks</title>
      <p id="d1e1979">Making use of an sSAE deep learning network, we present a crustal thickness
map of eastern Tibet and the western Yangtze craton (Fig. 7). The data sets
consist of phase velocities of Rayleigh waves from Xie et al. (2013) at
discrete frequencies of 10.0, 12.5, 15.0, 17.5, 20.0, 22.5, 25.0, 27.5, 30.0,
32.5, and 35.0 mHz. We conclude the following.
<list list-type="order"><list-item>
      <p id="d1e1984">Neural network structure is essential for inversion results, determined
by the following parameters: the number of hidden layers, the number of
neurons per layer, number of epoch, batch size, type of activation function,
learning rate, and non-sparsity penalty. We find that the number of
parameters in hidden layers and the size of a batch are crucial for training
neural networks. After many tests, the number of hidden layers was set to 6;
the number of neurons was 200, 100, 50, 20, 10, and 5. The number of periods
and the batch size were both set to 100, the activation function was sigmoid,
and the learning rate was 1; higher resolution and a more reliable crustal
thickness model were attained.</p></list-item><list-item>
      <p id="d1e1988">By training the deep network with six hidden layers and the traditional
shallow network with only one hidden layer, the relative error between the
phase velocity predicted by the deep network and the observed average phase
velocity is smaller than that between the phase velocity predicted by the
traditional shallow network and<?pagebreak page70?> observed average phase velocity, indicating
that the inversion result based on the deep network is more reliable than
that of the traditional shallow network.</p></list-item><list-item>
      <p id="d1e1992">Using only one processor, a six-layer sSAE network for 380 000 training
samples and 120 000 test samples for network training takes 5 h, while a
well-trained neural network inversion takes only a few seconds to complete.
To complete the same inversion task, it takes 3 weeks to use the Monte Carlo
method for four processors. We demonstrate that sSAE deep network inversion
is more efficient than Monte Carlo inversion.</p></list-item><list-item>
      <p id="d1e1996">Comparing our model with current knowledge about crustal structure as
represented by ZJS, CRUST2.0, and CUB2, the overall agreement with these
three models is very good, and agreement is generally better with CUB2 and
CRUST2.0, which are attained from relatively dense stations with rich data
coverage and higher resolution. Our result reveals more details in Chengdu,
the Qinling–Dabei fold belt, the Xi'an and Ordos basins, and the
Sichuan–Yunnan block.</p></list-item></list></p>
</sec>

      
      </body>
    <back><notes notes-type="dataavailability"><title>Data availability</title>

      <p id="d1e2003">Data and results of calculations are available by e-mail request.</p>
  </notes><notes notes-type="authorcontribution"><title>Author contributions</title>

      <p id="d1e2009">XC carried out the main investigation and formal analyses of this work and
wrote the initial draft. QL put forward the inversion method and designed the structure of the neural network.
PL performed the realizing inversion. YL revised the written work and contributed to the final text.</p>
  </notes><notes notes-type="competinginterests"><title>Competing interests</title>

      <p id="d1e2015">The authors declare that they have no conflict of
interest.</p>
  </notes><ack><title>Acknowledgements</title><p id="d1e2021">Our work was supported by the National Natural Science Foundation of China
(grant nos. 41774095 and 91755215). The authors are grateful to
Xie Jiayi for providing the phase velocity maps, Shapiro Nikolai for
making model CUB2 available, Gabi Laske for making model CRUST1.0 available,
and Zhu Jieshou for making model ZJS available.</p></ack><notes notes-type="reviewstatement"><title>Review statement</title>

      <p id="d1e2026">This paper was edited by Richard Gloaguen and reviewed by
Harry Matchette-Downes and three anonymous referees.</p>
  </notes><ref-list>
    <title>References</title>

      <ref id="bib1.bib1"><label>1</label><mixed-citation>
Bassin, C., Laske, G., and  Masters, G.: The current limits of resolution for
surface wave tomography in north America, EOS T. Am. Geophys. Un., 81,
F897, 2000.</mixed-citation></ref>
      <ref id="bib1.bib2"><label>2</label><mixed-citation>
Bengio, Y.: Learning deep architectures for AI, Foundations and trends in
Machine Learning, 2, 1–127, 2009.</mixed-citation></ref>
      <ref id="bib1.bib3"><label>3</label><mixed-citation>
Bengio, Y., Lamblin, P., Popovici, D., and Larochelle, H.: Greedy layer-wise
training of deep networks, in: Proceedings of the Advances in Neural
Information Processing Systems, Vancouver, BC, Canada, 153–160, 2006.</mixed-citation></ref>
      <ref id="bib1.bib4"><label>4</label><mixed-citation>
Bishop, C. M.: Neural Networks for Pattern Recognition, Oxford University
Press, Oxford, UK, 1995.</mixed-citation></ref>
      <ref id="bib1.bib5"><label>5</label><mixed-citation>
Chen, S. and Wilson,  C. J. L.: Emplacement of the Longmen Shan Thrust – Nappe
Belt along the eastern margin of the Tibetan Plateau, J. Struct. Geol., 18,
413–430, 1996.</mixed-citation></ref>
      <ref id="bib1.bib6"><label>6</label><mixed-citation>Chen, Y., Niu, F., and Liu, R.: Crustal structure beneath China from receiver function
analysis, J. Geophys. Res., 115, B03307, <ext-link xlink:href="https://doi.org/10.1029/2009JB006386" ext-link-type="DOI">10.1029/2009JB006386</ext-link>, 2010.</mixed-citation></ref>
      <ref id="bib1.bib7"><label>7</label><mixed-citation>
Das, T. and Nolet, G.: Crustal thickness estimation using high frequency
Rayleigh waves, Geophys. Res. Lett., 123, 169–184, 2001.</mixed-citation></ref>
      <ref id="bib1.bib8"><label>8</label><mixed-citation>
de Wit, R. W., Valentine, A. P., and Trampert, J.: Bayesian inference of
Earth's radial seismic structure from body-wave travel times using neural
networks, Geophys. J. Int., 195, 408–422, 2013.</mixed-citation></ref>
      <ref id="bib1.bib9"><label>9</label><mixed-citation>de Wit, R. W. L., Käufl, P. J., Valentine, A. P., and Trampert, J.: Bayesian inversion of free oscillations for Earth's
radial (an)elastic structure, Phys. Earth Planet. In.,
237, 1–17, <ext-link xlink:href="https://doi.org/10.1016/j.pepi.2014.09.004" ext-link-type="DOI">10.1016/j.pepi.2014.09.004</ext-link>, 2014.</mixed-citation></ref>
      <ref id="bib1.bib10"><label>10</label><mixed-citation>
Deng, J., Zhang, Z., Marchi, E., and Schuller, B: Sparse Autoencoder-Based
Feature Transfer Learning for Speech Emotion Recognition, in: Proceedings of
the 5th Biannu Humaine Association Conference on Affective Computing and
Intelligent Interaction, Geneva, Switzerland, 511–516, 2013.</mixed-citation></ref>
      <ref id="bib1.bib11"><label>11</label><mixed-citation>
Devilee, R., Curtis, A., and Roy-Chowdhury, K.: An efficient, probabilistic
neural network approach to solving inverse problems: inverting surface wave
velocities for Eurasian crustal thickness, J. Geophys. Res., 104,
28841–28857, 1999.</mixed-citation></ref>
      <ref id="bib1.bib12"><label>12</label><mixed-citation>
Dziewonski, A. M. and  Anderson, D. L.: Preliminary reference Earth model, Phys. Earth Planet. In., 25, 297–356, 1981.</mixed-citation></ref>
      <ref id="bib1.bib13"><label>13</label><mixed-citation>
Fang, J.: Global crustal and lithosphere thickness inverted by using satellite
gravity data, Crustal deformation and Earthquake, 19, 26–31, 1999 (in Chinese).</mixed-citation></ref>
      <ref id="bib1.bib14"><label>14</label><mixed-citation>
Flesch, L. M.,  Holt, W. E.,  Silver, P. G.,  Stephenson, M.,  Wang, C. Y., and
Chan, W. W.: Constraining the extent of crust – mantle coupling in central Asia using
GPS, geologic, and shear wave splitting data, Earth Planet. Sc. Lett., 238,
248–268, 2005.</mixed-citation></ref>
      <ref id="bib1.bib15"><label>15</label><mixed-citation>
Grad, M. and Tiira, T.: The moho depth map of the European Plate, Geophys. J.
Int., 176, 279–292, 2009.</mixed-citation></ref>
      <ref id="bib1.bib16"><label>16</label><mixed-citation>
Guo, L. H., Meng, X. H., Shi, L., and Chen, Z. X.: Preferential filtering method and its application to
Bouguer gravity anomaly of Chinese continent, Chinese J. Geophys.,  55, 4078–4088, 2012 (in
Chinese).</mixed-citation></ref>
      <ref id="bib1.bib17"><label>17</label><mixed-citation>
Hansen, T. M., Cordua, K. S., Looms, M. C., and Mosegaard, K.: SIPPI: A
Matlab toolbox for sampling the solution to inverse problems with complex
prior information: Part 1 – Methodology, Comput. Geosci., 52, 470–480,
2013.</mixed-citation></ref>
      <ref id="bib1.bib18"><label>18</label><mixed-citation>
Hinton, G. E. and Salakhutdinov, R. R.: Reducing the dimensionality of data
with neural networks, Science, 313,  504–507, 2006.</mixed-citation></ref>
      <ref id="bib1.bib19"><label>19</label><mixed-citation>
Huang, J. P., Fu, R. S., Xu, P. ,Huang, J. H., and Zheng, Y.: Inversion of gravity and topography data
for the crust thickness of China and its adjacency, Acta Seismologica
Sinica, 28, 250–258, 2006.</mixed-citation></ref>
      <ref id="bib1.bib20"><label>20</label><mixed-citation>
Lampinen, J. and Vehtari, A.: Bayesian approach for neural<?pagebreak page71?> networks-review
and case studies, Neural Networks, 14, 257–274,2001.</mixed-citation></ref>
      <ref id="bib1.bib21"><label>21</label><mixed-citation>
Laske, G., Masters, G., Ma, Z., and Pasyanos, M.: Update on CRUST1.0 – A
1-degree Global Model of Earth's Crust, Geophys. Res. Abstracts, 15, Abstract
EGU2013-2658, 2013.</mixed-citation></ref>
      <ref id="bib1.bib22"><label>22</label><mixed-citation>
Lebedev, S., Adam, J. M. C., and Meier, T.: Mapping the moho with seismic
surface waves: a review, resolution analysis, and recommended inversion
strategies, Tectonophysics, 609, 377–394, 2013.</mixed-citation></ref>
      <ref id="bib1.bib23"><label>23</label><mixed-citation>
LeCun, Y., Bengio, Y., and Hinton, G. E.: Deep Learning, Nature, 521,
436–444, 2015.</mixed-citation></ref>
      <ref id="bib1.bib24"><label>24</label><mixed-citation>
Legendre, C. P., Deschamps, F., Zhao, L., and  Chen, Q. F.: Rayleigh-wave
dispersion reveals crust-mantle decoupling beneath eastern Tibet, Scientific
reports, 2015.</mixed-citation></ref>
      <ref id="bib1.bib25"><label>25</label><mixed-citation>Liu, Q., Hu, X., Ye, M., Cheng, X., and Li, F.: Gas Recognition under Sensor
Drift by Using Deep Learning, Int. J. Intell. Syst., 30, 907–922, <ext-link xlink:href="https://doi.org/10.1002/int.21731" ext-link-type="DOI">10.1002/int.21731</ext-link>, 2015.</mixed-citation></ref>
      <ref id="bib1.bib26"><label>26</label><mixed-citation>
Masters, G., Barmine, M. P., and Kientz, S.: Pasadena: Calif. Inst. of Technol.,
Mineos User's Manual, Computational Infrastructure for Geodynamics, 2014.</mixed-citation></ref>
      <ref id="bib1.bib27"><label>27</label><mixed-citation>
Montagner, J.-P. and Jobert, N.: Vectorial tomography–ii. Application to the
indian ocean, Geophys. J. Int., 94, 309–344, 1988.</mixed-citation></ref>
      <ref id="bib1.bib28"><label>28</label><mixed-citation>
Meier, U., Curtis, A., and Trampert, J.: Global crustal thickness from neural
network inversion of surface wave data, Geophys. J. Int.,
169, 706–722, 2007.</mixed-citation></ref>
      <ref id="bib1.bib29"><label>29</label><mixed-citation>
Qian, H., Mechie, J., Li, H., Xue, G., Su, H., and Cui, X.:
Structure of the crust and mantle down to 700 km depth beneath the
Longmenshan from P receiver functions, Tectonics, 37, 1688–1708, 2018.</mixed-citation></ref>
      <ref id="bib1.bib30"><label>30</label><mixed-citation>
Shapiro, N. M. and Ritzwoller, M. H.: Monte-Carlo inversion for a global
shear-velocity model of the crust and upper mantle, Geophys. J. Int., 151,
88–105, 2002.</mixed-citation></ref>
      <ref id="bib1.bib31"><label>31</label><mixed-citation>
Stolk, W., Kaban, M., Beekman, F., Tesauro, M., Mooney, W. D., and Cloetingh, S.: High resolution regional crustal models from irregularly
distributed data: application to Asia and adjacent areas, Tectonophysics,
602, 55–68, 2013.</mixed-citation></ref>
      <ref id="bib1.bib32"><label>32</label><mixed-citation>
Sun, X.,  Song, X.,  Zheng, S.,  Yang, Y., and  Ritzwoller, M.: Three dimensional
shear wave velocity structure of the crust and upper mantle beneath China
from ambient noise surface wave tomography, Earthquake Sci., 23,
449–463, 2010.</mixed-citation></ref>
      <ref id="bib1.bib33"><label>33</label><mixed-citation>
van der Baan, M. and Jutten, C.: Neural networks in geophysical applications,
Geophysics, 65, 1032–1047, 2000.</mixed-citation></ref>
      <ref id="bib1.bib34"><label>34</label><mixed-citation>
Wang, C. Y., Lou, H., Yao, Z. X., Luo, X. H., and Li, Z. Y.: Crustal thicknesses and poisson's ratios in
Longmen shan mountains and adjacent regions, Quaternary Sciences, 30, 652–661, 2010 (in Chinese).</mixed-citation></ref>
      <ref id="bib1.bib35"><label>35</label><mixed-citation>Wang, S. J., Wang, F. Y., Zhang, J. S., Liu, B. F., Zhang, C. K.,
Zhao, J. R., Duan, Y. L., Song, X. H., and Deng, X. G.: A study of deep seismogenic environment in Lushan Ms7.0
earthquake zone by wide-angle seismic reflection/refraction profile, Chinese
J. Geophys., 58, 474–485, 2015.
 </mixed-citation></ref><?xmltex \hack{\newpage}?>
      <ref id="bib1.bib36"><label>36</label><mixed-citation>Wang, Z., Zhao, D., and Wang, J.: Deep structure and seismogenesis of the north – south
seismic zone in southwest China, J. Geophys. Res.-Sol. Ea.,
115, <ext-link xlink:href="https://doi.org/10.1029/2010JB007797" ext-link-type="DOI">10.1029/2010JB007797</ext-link>, 2010.</mixed-citation></ref>
      <ref id="bib1.bib37"><label>37</label><mixed-citation>Xie, J.,  Ritzwoller, M. H., Shen, W., Yang, Y., Zheng, Y., and Zhou, L.: Crustal
radial anisotropy across Eastern Tibet and the Western Yangtze Craton, J.
Geophys. Res., 118, 4226–4252, <ext-link xlink:href="https://doi.org/10.1002/jgrb.50296" ext-link-type="DOI">10.1002/jgrb.50296</ext-link>, 2013.</mixed-citation></ref>
      <ref id="bib1.bib38"><label>38</label><mixed-citation>
Xu, L., Rondenay, S., and Van Der Hilst, R. D.: Structure of the Crust
beneath the southeastern Tibetan Plateau from teleseismic Receiver Functions,
Phys. Earth Planet. In., 165, 176–193, 2007.</mixed-citation></ref>
      <ref id="bib1.bib39"><label>39</label><mixed-citation>
Xu, X., Keller, G. R., Gao, R., Guo, X., and Zhu, X. Uplift of the Longmen
Shan area in the eastern Tibetan plateau: An integrated geophysical and
geodynamic analysis, Int. Geol. Rev., 58, 14–31,2016.</mixed-citation></ref>
      <ref id="bib1.bib40"><label>40</label><mixed-citation>
Yi, G. X., Yao, H. J., Zhu, J. S., and Van, D. H. R. D.: Rayleigh-wave phase velocity distribution in
China continent and its adjacent regions, Chinese J. Geophys.
51, 265–274, 2008.</mixed-citation></ref>
      <ref id="bib1.bib41"><label>41</label><mixed-citation>
Yao, H.,  Van Der Hilst, R. D., and  De Hoop, M. V.: Surface-wave array
tomography in SE Tibet from ambient seismic noise and two-station
analysis–I. Phase-velocity maps, Geophys. J. Int., 166,
732–744, 2006.</mixed-citation></ref>
      <ref id="bib1.bib42"><label>42</label><mixed-citation>
Zeng, R. S., Sun, W. G., Mao, T. E., and Zhong, Y. L.: The depth of Moho in the mainland of China,
Acta Seismologica Sinica, 17, 322–327, 1995 (in Chinese).</mixed-citation></ref>
      <ref id="bib1.bib43"><label>43</label><mixed-citation>
Zhang, Q.,  Sandvol, E.,  Ni, J., Yang,  Y., and  Chen, Y. J.: Rayleigh wave
tomography of the northeastern margin of the Tibetan Plateau, Earth Planet.
Sc. Lett., 304, 103–112, 2011.</mixed-citation></ref>
      <ref id="bib1.bib44"><label>44</label><mixed-citation>Zheng, S.,  Sun, X.,  Song, X.,  Yang, Y., and  Ritzwoller, M. H.: Surface wave
tomography of China from ambient seismic noise correlation, Geochem.
Geophy. Geosys., 9, Q05020, <ext-link xlink:href="https://doi.org/10.1029/2008GC001981" ext-link-type="DOI">10.1029/2008GC001981</ext-link>, 2008.</mixed-citation></ref>
      <ref id="bib1.bib45"><label>45</label><mixed-citation>
Zhou, L.,  Xie, J.,  Shen, W.,  Zheng, Y.,  Yang, Y.,  Shi, H., and  Ritzwoller, M. H.:
The structure of the crust and uppermost mantle beneath South China from
ambient noise and earthquake tomography, Geophys. J. Int., 189,
1565–1583, 2012.</mixed-citation></ref>
      <ref id="bib1.bib46"><label>46</label><mixed-citation>Zhou, Y., Nolet, G., Dahlen, F., and Laske, G.: Global upper-mantle structure
from finite-frequency surface-wave tomography, J. Geophys. Res., 111,
B04304, <ext-link xlink:href="https://doi.org/10.1029/2005JB003677" ext-link-type="DOI">10.1029/2005JB003677</ext-link>, 2006.</mixed-citation></ref>
      <ref id="bib1.bib47"><label>47</label><mixed-citation>
Zhu, J. S., Zhao, J. M., Jiang, X. T., Fan, J., and Liang, C. T.: Crustal flow beneath the eastern
margin of the Tibetan plateau, Earthquake Science, 25, 469–483, 2012.</mixed-citation></ref>

  </ref-list></back>
    <!--<article-title-html>Inverting Rayleigh surface wave velocities for crustal thickness in eastern Tibet and the western Yangtze craton based on deep learning neural networks</article-title-html>
<abstract-html><p>Crustal thickness is an important factor affecting lithospheric structure and
deep geodynamics. In this paper, a deep learning neural network based on a
stacked sparse auto-encoder is proposed for the inversion of crustal
thickness in eastern Tibet and the western Yangtze craton. First, with the
phase velocity of the Rayleigh surface wave as input and the theoretical
crustal thickness as output, 12 deep-sSAE neural networks are constructed,
which are trained by 380&thinsp;000 and tested by 120&thinsp;000 theoretical models. We
then invert the observed phase velocities through these 12 neural networks.
According to the test error and misfit of other crustal thickness models, the
optimal crustal thickness model is selected as the crustal thickness of the
study area. Compared with other ways to detect crustal thickness such as
seismic wave reflection and receiver function, we adopt a new way for
inversion of earth model parameters, and realize that a deep learning neural
network based on data driven with the highly non-linear mapping ability can
be widely used by geophysicists, and our result has good agreement with
high-resolution crustal thickness models. Compared with other methods, our
experimental results based on a deep learning neural network and a new
Rayleigh wave phase velocity model reveal some details: there is a
northward-dipping Moho gradient zone in the Qiangtang block and a relatively
shallow north-west–south-east oriented crust at the Songpan–Ganzi block.
Crustal thickness around Xi'an and the Ordos basin is shallow, about 35&thinsp;km.
The change in crustal thickness in the Sichuan–Yunnan block is sharp, where
crustal thickness is 60&thinsp;km north-west and 35&thinsp;km south-east. We conclude
that the deep learning neural network is a promising, efficient, and
believable geophysical inversion tool.</p></abstract-html>
<ref-html id="bib1.bib1"><label>1</label><mixed-citation>
Bassin, C., Laske, G., and  Masters, G.: The current limits of resolution for
surface wave tomography in north America, EOS T. Am. Geophys. Un., 81,
F897, 2000.
</mixed-citation></ref-html>
<ref-html id="bib1.bib2"><label>2</label><mixed-citation>
Bengio, Y.: Learning deep architectures for AI, Foundations and trends in
Machine Learning, 2, 1–127, 2009.
</mixed-citation></ref-html>
<ref-html id="bib1.bib3"><label>3</label><mixed-citation>
Bengio, Y., Lamblin, P., Popovici, D., and Larochelle, H.: Greedy layer-wise
training of deep networks, in: Proceedings of the Advances in Neural
Information Processing Systems, Vancouver, BC, Canada, 153–160, 2006.
</mixed-citation></ref-html>
<ref-html id="bib1.bib4"><label>4</label><mixed-citation>
Bishop, C. M.: Neural Networks for Pattern Recognition, Oxford University
Press, Oxford, UK, 1995.
</mixed-citation></ref-html>
<ref-html id="bib1.bib5"><label>5</label><mixed-citation>
Chen, S. and Wilson,  C. J. L.: Emplacement of the Longmen Shan Thrust – Nappe
Belt along the eastern margin of the Tibetan Plateau, J. Struct. Geol., 18,
413–430, 1996.
</mixed-citation></ref-html>
<ref-html id="bib1.bib6"><label>6</label><mixed-citation>
Chen, Y., Niu, F., and Liu, R.: Crustal structure beneath China from receiver function
analysis, J. Geophys. Res., 115, B03307, <a href="https://doi.org/10.1029/2009JB006386" target="_blank">https://doi.org/10.1029/2009JB006386</a>, 2010.
</mixed-citation></ref-html>
<ref-html id="bib1.bib7"><label>7</label><mixed-citation>
Das, T. and Nolet, G.: Crustal thickness estimation using high frequency
Rayleigh waves, Geophys. Res. Lett., 123, 169–184, 2001.
</mixed-citation></ref-html>
<ref-html id="bib1.bib8"><label>8</label><mixed-citation>
de Wit, R. W., Valentine, A. P., and Trampert, J.: Bayesian inference of
Earth's radial seismic structure from body-wave travel times using neural
networks, Geophys. J. Int., 195, 408–422, 2013.
</mixed-citation></ref-html>
<ref-html id="bib1.bib9"><label>9</label><mixed-citation>
de Wit, R. W. L., Käufl, P. J., Valentine, A. P., and Trampert, J.: Bayesian inversion of free oscillations for Earth's
radial (an)elastic structure, Phys. Earth Planet. In.,
237, 1–17, <a href="https://doi.org/10.1016/j.pepi.2014.09.004" target="_blank">https://doi.org/10.1016/j.pepi.2014.09.004</a>, 2014.
</mixed-citation></ref-html>
<ref-html id="bib1.bib10"><label>10</label><mixed-citation>
Deng, J., Zhang, Z., Marchi, E., and Schuller, B: Sparse Autoencoder-Based
Feature Transfer Learning for Speech Emotion Recognition, in: Proceedings of
the 5th Biannu Humaine Association Conference on Affective Computing and
Intelligent Interaction, Geneva, Switzerland, 511–516, 2013.
</mixed-citation></ref-html>
<ref-html id="bib1.bib11"><label>11</label><mixed-citation>
Devilee, R., Curtis, A., and Roy-Chowdhury, K.: An efficient, probabilistic
neural network approach to solving inverse problems: inverting surface wave
velocities for Eurasian crustal thickness, J. Geophys. Res., 104,
28841–28857, 1999.
</mixed-citation></ref-html>
<ref-html id="bib1.bib12"><label>12</label><mixed-citation>
Dziewonski, A. M. and  Anderson, D. L.: Preliminary reference Earth model, Phys. Earth Planet. In., 25, 297–356, 1981.
</mixed-citation></ref-html>
<ref-html id="bib1.bib13"><label>13</label><mixed-citation>
Fang, J.: Global crustal and lithosphere thickness inverted by using satellite
gravity data, Crustal deformation and Earthquake, 19, 26–31, 1999 (in Chinese).
</mixed-citation></ref-html>
<ref-html id="bib1.bib14"><label>14</label><mixed-citation>
Flesch, L. M.,  Holt, W. E.,  Silver, P. G.,  Stephenson, M.,  Wang, C. Y., and
Chan, W. W.: Constraining the extent of crust – mantle coupling in central Asia using
GPS, geologic, and shear wave splitting data, Earth Planet. Sc. Lett., 238,
248–268, 2005.
</mixed-citation></ref-html>
<ref-html id="bib1.bib15"><label>15</label><mixed-citation>
Grad, M. and Tiira, T.: The moho depth map of the European Plate, Geophys. J.
Int., 176, 279–292, 2009.
</mixed-citation></ref-html>
<ref-html id="bib1.bib16"><label>16</label><mixed-citation>
Guo, L. H., Meng, X. H., Shi, L., and Chen, Z. X.: Preferential filtering method and its application to
Bouguer gravity anomaly of Chinese continent, Chinese J. Geophys.,  55, 4078–4088, 2012 (in
Chinese).
</mixed-citation></ref-html>
<ref-html id="bib1.bib17"><label>17</label><mixed-citation>
Hansen, T. M., Cordua, K. S., Looms, M. C., and Mosegaard, K.: SIPPI: A
Matlab toolbox for sampling the solution to inverse problems with complex
prior information: Part 1 – Methodology, Comput. Geosci., 52, 470–480,
2013.
</mixed-citation></ref-html>
<ref-html id="bib1.bib18"><label>18</label><mixed-citation>
Hinton, G. E. and Salakhutdinov, R. R.: Reducing the dimensionality of data
with neural networks, Science, 313,  504–507, 2006.
</mixed-citation></ref-html>
<ref-html id="bib1.bib19"><label>19</label><mixed-citation>
Huang, J. P., Fu, R. S., Xu, P. ,Huang, J. H., and Zheng, Y.: Inversion of gravity and topography data
for the crust thickness of China and its adjacency, Acta Seismologica
Sinica, 28, 250–258, 2006.
</mixed-citation></ref-html>
<ref-html id="bib1.bib20"><label>20</label><mixed-citation>
Lampinen, J. and Vehtari, A.: Bayesian approach for neural networks-review
and case studies, Neural Networks, 14, 257–274,2001.
</mixed-citation></ref-html>
<ref-html id="bib1.bib21"><label>21</label><mixed-citation>
Laske, G., Masters, G., Ma, Z., and Pasyanos, M.: Update on CRUST1.0 – A
1-degree Global Model of Earth's Crust, Geophys. Res. Abstracts, 15, Abstract
EGU2013-2658, 2013.
</mixed-citation></ref-html>
<ref-html id="bib1.bib22"><label>22</label><mixed-citation>
Lebedev, S., Adam, J. M. C., and Meier, T.: Mapping the moho with seismic
surface waves: a review, resolution analysis, and recommended inversion
strategies, Tectonophysics, 609, 377–394, 2013.
</mixed-citation></ref-html>
<ref-html id="bib1.bib23"><label>23</label><mixed-citation>
LeCun, Y., Bengio, Y., and Hinton, G. E.: Deep Learning, Nature, 521,
436–444, 2015.
</mixed-citation></ref-html>
<ref-html id="bib1.bib24"><label>24</label><mixed-citation>
Legendre, C. P., Deschamps, F., Zhao, L., and  Chen, Q. F.: Rayleigh-wave
dispersion reveals crust-mantle decoupling beneath eastern Tibet, Scientific
reports, 2015.
</mixed-citation></ref-html>
<ref-html id="bib1.bib25"><label>25</label><mixed-citation>
Liu, Q., Hu, X., Ye, M., Cheng, X., and Li, F.: Gas Recognition under Sensor
Drift by Using Deep Learning, Int. J. Intell. Syst., 30, 907–922, <a href="https://doi.org/10.1002/int.21731" target="_blank">https://doi.org/10.1002/int.21731</a>, 2015.
</mixed-citation></ref-html>
<ref-html id="bib1.bib26"><label>26</label><mixed-citation>
Masters, G., Barmine, M. P., and Kientz, S.: Pasadena: Calif. Inst. of Technol.,
Mineos User's Manual, Computational Infrastructure for Geodynamics, 2014.
</mixed-citation></ref-html>
<ref-html id="bib1.bib27"><label>27</label><mixed-citation>
Montagner, J.-P. and Jobert, N.: Vectorial tomography–ii. Application to the
indian ocean, Geophys. J. Int., 94, 309–344, 1988.
</mixed-citation></ref-html>
<ref-html id="bib1.bib28"><label>28</label><mixed-citation>
Meier, U., Curtis, A., and Trampert, J.: Global crustal thickness from neural
network inversion of surface wave data, Geophys. J. Int.,
169, 706–722, 2007.
</mixed-citation></ref-html>
<ref-html id="bib1.bib29"><label>29</label><mixed-citation>
Qian, H., Mechie, J., Li, H., Xue, G., Su, H., and Cui, X.:
Structure of the crust and mantle down to 700&thinsp;km depth beneath the
Longmenshan from P receiver functions, Tectonics, 37, 1688–1708, 2018.
</mixed-citation></ref-html>
<ref-html id="bib1.bib30"><label>30</label><mixed-citation>
Shapiro, N. M. and Ritzwoller, M. H.: Monte-Carlo inversion for a global
shear-velocity model of the crust and upper mantle, Geophys. J. Int., 151,
88–105, 2002.
</mixed-citation></ref-html>
<ref-html id="bib1.bib31"><label>31</label><mixed-citation>
Stolk, W., Kaban, M., Beekman, F., Tesauro, M., Mooney, W. D., and Cloetingh, S.: High resolution regional crustal models from irregularly
distributed data: application to Asia and adjacent areas, Tectonophysics,
602, 55–68, 2013.
</mixed-citation></ref-html>
<ref-html id="bib1.bib32"><label>32</label><mixed-citation>
Sun, X.,  Song, X.,  Zheng, S.,  Yang, Y., and  Ritzwoller, M.: Three dimensional
shear wave velocity structure of the crust and upper mantle beneath China
from ambient noise surface wave tomography, Earthquake Sci., 23,
449–463, 2010.
</mixed-citation></ref-html>
<ref-html id="bib1.bib33"><label>33</label><mixed-citation>
van der Baan, M. and Jutten, C.: Neural networks in geophysical applications,
Geophysics, 65, 1032–1047, 2000.
</mixed-citation></ref-html>
<ref-html id="bib1.bib34"><label>34</label><mixed-citation>
Wang, C. Y., Lou, H., Yao, Z. X., Luo, X. H., and Li, Z. Y.: Crustal thicknesses and poisson's ratios in
Longmen shan mountains and adjacent regions, Quaternary Sciences, 30, 652–661, 2010 (in Chinese).
</mixed-citation></ref-html>
<ref-html id="bib1.bib35"><label>35</label><mixed-citation>
Wang, S. J., Wang, F. Y., Zhang, J. S., Liu, B. F., Zhang, C. K.,
Zhao, J. R., Duan, Y. L., Song, X. H., and Deng, X. G.: A study of deep seismogenic environment in Lushan Ms7.0
earthquake zone by wide-angle seismic reflection/refraction profile, Chinese
J. Geophys., 58, 474–485, 2015.

</mixed-citation></ref-html>
<ref-html id="bib1.bib36"><label>36</label><mixed-citation>
Wang, Z., Zhao, D., and Wang, J.: Deep structure and seismogenesis of the north – south
seismic zone in southwest China, J. Geophys. Res.-Sol. Ea.,
115, <a href="https://doi.org/10.1029/2010JB007797" target="_blank">https://doi.org/10.1029/2010JB007797</a>, 2010.
</mixed-citation></ref-html>
<ref-html id="bib1.bib37"><label>37</label><mixed-citation>
Xie, J.,  Ritzwoller, M. H., Shen, W., Yang, Y., Zheng, Y., and Zhou, L.: Crustal
radial anisotropy across Eastern Tibet and the Western Yangtze Craton, J.
Geophys. Res., 118, 4226–4252, <a href="https://doi.org/10.1002/jgrb.50296" target="_blank">https://doi.org/10.1002/jgrb.50296</a>, 2013.
</mixed-citation></ref-html>
<ref-html id="bib1.bib38"><label>38</label><mixed-citation>
Xu, L., Rondenay, S., and Van Der Hilst, R. D.: Structure of the Crust
beneath the southeastern Tibetan Plateau from teleseismic Receiver Functions,
Phys. Earth Planet. In., 165, 176–193, 2007.
</mixed-citation></ref-html>
<ref-html id="bib1.bib39"><label>39</label><mixed-citation>
Xu, X., Keller, G. R., Gao, R., Guo, X., and Zhu, X. Uplift of the Longmen
Shan area in the eastern Tibetan plateau: An integrated geophysical and
geodynamic analysis, Int. Geol. Rev., 58, 14–31,2016.
</mixed-citation></ref-html>
<ref-html id="bib1.bib40"><label>40</label><mixed-citation>
Yi, G. X., Yao, H. J., Zhu, J. S., and Van, D. H. R. D.: Rayleigh-wave phase velocity distribution in
China continent and its adjacent regions, Chinese J. Geophys.
51, 265–274, 2008.
</mixed-citation></ref-html>
<ref-html id="bib1.bib41"><label>41</label><mixed-citation>
Yao, H.,  Van Der Hilst, R. D., and  De Hoop, M. V.: Surface-wave array
tomography in SE Tibet from ambient seismic noise and two-station
analysis–I. Phase-velocity maps, Geophys. J. Int., 166,
732–744, 2006.
</mixed-citation></ref-html>
<ref-html id="bib1.bib42"><label>42</label><mixed-citation>
Zeng, R. S., Sun, W. G., Mao, T. E., and Zhong, Y. L.: The depth of Moho in the mainland of China,
Acta Seismologica Sinica, 17, 322–327, 1995 (in Chinese).
</mixed-citation></ref-html>
<ref-html id="bib1.bib43"><label>43</label><mixed-citation>
Zhang, Q.,  Sandvol, E.,  Ni, J., Yang,  Y., and  Chen, Y. J.: Rayleigh wave
tomography of the northeastern margin of the Tibetan Plateau, Earth Planet.
Sc. Lett., 304, 103–112, 2011.
</mixed-citation></ref-html>
<ref-html id="bib1.bib44"><label>44</label><mixed-citation>
Zheng, S.,  Sun, X.,  Song, X.,  Yang, Y., and  Ritzwoller, M. H.: Surface wave
tomography of China from ambient seismic noise correlation, Geochem.
Geophy. Geosys., 9, Q05020, <a href="https://doi.org/10.1029/2008GC001981" target="_blank">https://doi.org/10.1029/2008GC001981</a>, 2008.
</mixed-citation></ref-html>
<ref-html id="bib1.bib45"><label>45</label><mixed-citation>
Zhou, L.,  Xie, J.,  Shen, W.,  Zheng, Y.,  Yang, Y.,  Shi, H., and  Ritzwoller, M. H.:
The structure of the crust and uppermost mantle beneath South China from
ambient noise and earthquake tomography, Geophys. J. Int., 189,
1565–1583, 2012.
</mixed-citation></ref-html>
<ref-html id="bib1.bib46"><label>46</label><mixed-citation>
Zhou, Y., Nolet, G., Dahlen, F., and Laske, G.: Global upper-mantle structure
from finite-frequency surface-wave tomography, J. Geophys. Res., 111,
B04304, <a href="https://doi.org/10.1029/2005JB003677" target="_blank">https://doi.org/10.1029/2005JB003677</a>, 2006.
</mixed-citation></ref-html>
<ref-html id="bib1.bib47"><label>47</label><mixed-citation>
Zhu, J. S., Zhao, J. M., Jiang, X. T., Fan, J., and Liang, C. T.: Crustal flow beneath the eastern
margin of the Tibetan plateau, Earthquake Science, 25, 469–483, 2012.
</mixed-citation></ref-html>--></article>
