Philip Maechling | University of Southern California (original) (raw)

Papers by Philip Maechling

Journal of Open Source Software, 2022

For government officials and the public to act on real-time forecasts of earthquakes, the seismol... more For government officials and the public to act on real-time forecasts of earthquakes, the seismological community needs to develop confidence in the underlying scientific hypotheses of the forecast generating models by assessing their predictive skill. For this purpose, the Collaboratory for the Study of Earthquake Predictability (CSEP) provides cyberinfrastructure and computational tools to evaluate earthquake forecasts. Here, we introduce pyCSEP, a Python package to help earthquake forecast developers embed model evaluation into the model development process. The package contains the following modules: (1) earthquake catalog access and processing, (2) data models for earthquake forecasts, (3) statistical tests for evaluating earthquake forecasts, and (4) visualization routines. pyCSEP can evaluate earthquake forecasts expressed as expected rates in space-magnitude bins, and simulation-based forecasts that produce thousands of synthetic seismicity catalogs. Most importantly, pyCSEP contains community-endorsed implementations of statistical tests to evaluate earthquake forecasts, and provides well defined file formats and standards to facilitate model comparisons. The toolkit will facilitate integrating new forecasting models into testing centers, which evaluate forecast models and prediction algorithms in an automated, prospective and independent manner, forming a critical step towards reliable operational earthquake forecasting.

Contains data needed to reproduce the figures from the publication of pyCSEP: A Software Toolkit ... more Contains data needed to reproduce the figures from the publication of pyCSEP: A Software Toolkit for Earthquake Forecast Developers. The data can be uncompressed using the following command: <pre><code class="language-bash">tar -xzvf pycsep_esrl_data.tar.gz</code></pre> After uncompressing, the data set contains the following files: data/ evaluation_catalog.json evaluation_catalog_zechar2013_merge.txt forecasts/ bird_liu.neokinema-fromXML.dat ebel.aftershock.corrected-fromXML.dat helmstetter_et_al.hkj.aftershock-fromXML.dat lombardi.DBM.italy.5yr.2010-01-01.dat meletti.MPS04.italy.5yr.2010-01-01.dat werner.HiResSmoSeis-m1.italy.5yr.2010-01-01.dat ucerf3_forecast/ config.json m71_event.json results_complete.bin

Landscape Journal, 1998

... Gary Ferguson, Denny Galvin, Larry Gamble, Catherine Gilliam, Jan Gingold, Dennis Glick, Fran... more ... Gary Ferguson, Denny Galvin, Larry Gamble, Catherine Gilliam, Jan Gingold, Dennis Glick, Frank Gregg, Bill Hedden, Elizabeth Humstone, Michael Kinsley, Lenora Kirby, Jake Kittle, Gil Lusk ... Larry Howe deserves special men-tion for his expertise in editing and wordsmithing. ...

Change-log Removed normalization of rates on CL-Test (#117) Added function to compute bin-wise lo... more Change-log Removed normalization of rates on CL-Test (#117) Added function to compute bin-wise log-likelihood scores (#118) Properly use region to compute spatial counts and spatial magnitude counts in region class (#122) Fix for simplified 'fast' lat-lon ratio calculation (#125) Add feature to read forecasts with swapped lat/lon values in file (#130) Add 'percentile' argument for plotting Poisson evaluations (#131) Modify comparison plot to simultaneously plot T and W tests (#132) Add feature to trace region outline when plotting spatial data sets (#133) Better handling for magnitude ticks in plotting catalogs (#134) Refactor polygon to models module (#135) Added arguments to modify the fontsize for grid labels and basemap plots (#136) Added function to return the midpoints of the valid testing region (#137) Include regions when serializing catalogs to JSON (#138) Add support for spatial forecasts (#142) Upated CI workflows to reduce time required and fix intermitte...

Computational science researchers running large-scale scientific workflow applications often want... more Computational science researchers running large-scale scientific workflow applications often want to run their workflows on the largest available compute systems to improve time to solution. Workflow tools used in distributed, heterogeneous, high performance computing environments typically rely on either a push-based or a pull-based approach for resource provisioning from these compute systems. However, many large clusters have moved to two-factor authentication for job submission, making traditional automated push-based job submission impossible. On the other hand, pull-based approaches such as pilot jobs may lead to increased complexity and a reduction in node-hour efficiency. In this paper, we describe a new, efficient approach based on HTCondor-G called reverse GAHP (rvGAHP) that allows us to push jobs using reverse SSH submissions with better efficiency than pull-based methods. We successfully used this approach to perform a large probabilistic seismic hazard analysis study us...

Computational scientists often need to execute large, looselycoupled parallel applications such a... more Computational scientists often need to execute large, looselycoupled parallel applications such as workflows and bags of tasks in order to do their research. These applications are typically composed of many, short-running, serial tasks, which frequently demand large amounts of computation and storage. In order to produce results in a reasonable amount of time, scientists would like to execute these applications using petascale resources. In the past this has been a challenge because petascale systems are not designed to execute such workloads efficiently. In this paper we describe a new approach to executing large, fine-grained workflows on distributed petascale systems. Our solution involves partitioning the workflow into independent subgraphs, and then submitting each subgraph as a self-contained MPI job to the available resources (often remote). We describe how the

As a part of the California Integrated Seismic Network (CISN) earthquake early warning (EEW) algo... more As a part of the California Integrated Seismic Network (CISN) earthquake early warning (EEW) algorithm development, funded through USGS NEHRP, we have developed the CISN EEW web site to collect the results of multiple EEW algorithms and to display these results in a comparative manner (www.scec.org/eew). During the last year, the CISN EEW algorithm development group defined a set of

... Authors: Minster, J.; Olsen, KB; Moore, R.; Day, S.; Maechling, P.; Jordan, T.; Faerman, M.; ... more ... Authors: Minster, J.; Olsen, KB; Moore, R.; Day, S.; Maechling, P.; Jordan, T.; Faerman, M.; Cui, Y.; Ely, G.; Hu, Y.; Shkoller, B.; Marcinkovich, C.; Bielak ... of the fault to rupture in a single M7.7 event is a major component of seismic hazard in southern California and northern Mexico. ...

Earth science applications such as those being developed within the Southern California Earthquak... more Earth science applications such as those being developed within the Southern California Earthquake Center (SCEC) require the coordination of hundreds of thousands of computation and data management tasks running on the national cyberinfrastructure resources such as the TeraGrid. Such coordination cannot be done by a single person; rather, automated techniques such as those provided by the Pegasus workflow management system

Three earthquake early warning (EEW) algorithms are currently being evaluated within the Californ... more Three earthquake early warning (EEW) algorithms are currently being evaluated within the California Integrated Seismic Network (CISN) with support from the US Geological Survey. The evaluation encompasses two aspects: their operation using data from throughout the state under real time conditions, and assessment of their alerts at a web-accessible testing center. An EEW system rapidly detects the initiation of earthquakes and predicts their ground shaking. Its purpose is to provide warning of potentially damaging ground motion in a target region before the strong shaking arrives. One of the three algorithms implemented at the CISN data centers uses a single station, or 'On-site' approach (Caltech). The other two, 'ElarmS' (UC Berkeley) and 'Virtual Seismologist' (VS, Caltech/ETH), are network-based. Although single station alerts can be delivered more quickly than those from a network-based system, more of them tend to be false warnings. Network-based algorithms for EEW require that data be gathered at a central site for joint processing. The two network-based systems, ElarmS and VS, run 15 s behind real time in order to gather ~90% of station data before processing. The EEW alert testing center was developed by the Southern California Earthquake Center (SCEC). Results from the various algorithms are collected automatically. Automatically generated performance summaries allow the comparison of the EEW alerts with each other and with earthquakes within the region. Performance criteria include promptness of the alert, earthquake location and magnitude. Provisions have also been made to assess false alerts, ground motion predictions and uncertainties. In addition to evaluating the three algorithms in terms for separate and joint reliability, we review the needs of EEW with respect to instrumentation and data latency. Possible warning times will typically range from seconds to tens of seconds, and each second delay means a decrease in the available warning time. Minimal latency is therefore important to warning systems. As testing progresses, we are formulating specifications for geophysical networks that can provide real time data in a robust fashion.

Abstract For the past three years we have presented results of several large-scale earthquake sim... more Abstract For the past three years we have presented results of several large-scale earthquake simulations run by different research groups, but these have been limited to qualitative comparisons. This work presents a preliminary verification of three simulation sets for the ShakeOut earthquake scenario, version 1.1, an Mw 7.8 earthquake on a portion of the San Andreas fault in southern California. Two of the simulation sets use a finite difference approach while the third uses finite elements. The verification is done ...

Accurate and rapid CMT inversion is important for seismic hazard analysis. We have developed an a... more Accurate and rapid CMT inversion is important for seismic hazard analysis. We have developed an algorithm for very rapid full-wave CMT inversions in a 3D Earth structure model and applied it on earthquakes recorded by the Southern California Seismic Network (SCSN). The procedure relies on the use of receiver-side Green tensors (RGTs), which comprise the spatial-temporal displacements produced by the three orthogonal unit impulsive point forces acting at the receiver. We have constructed a RGT database for 219 broadband stations in Southern California using an updated version of the 3D SCEC Community Velocity Model (CVM) version 4.0 and a staggered-grid nite-di erence code. Finite-di erence synthetic seismograms for any earthquake in our modeling volume can be simply calculated by extracting a small, source-centered volume from the RGT database and applying the reciprocity principle. We have developed an automated algorithm that combines a grid-search for suitable epicenter and focal mechanisms with a gradient-descent method that further re nes the grid-search results. In this algorithm, the CMT solutions are inverted near real-time by using waveform in a 3D Earth structure. Comparing with the CMT solutions provided by the Southern California Seismic Network (SCSN) shows that our solutions generally provide better t to the observed waveforms. Our algorithm may provide more robust CMT solutions for earthquakes in Southern California. In addition, the rapid and accurate full-wave CMT inversion has potential to extent to accurate near real-time ground motion prediction based on 3D structure model for earthquake early warning purpose. When combined with real-time telemetered waveform recordings, our algorithm can provide (near) real-time ground-motion forecast.

We have successfully applied full-3-D tomography (F3DT) based on a combination of the scattering-... more We have successfully applied full-3-D tomography (F3DT) based on a combination of the scattering-integral method (SI-F3DT) and the adjoint-wavefield method (AW-F3DT) to iteratively improve a 3-D starting model, the Southern California Earthquake Center (SCEC) Community Velocity Model version 4.0 (CVM-S4). In F3DT, the sensitivity (Fréchet) kernels are computed using numerical solutions of the 3-D elastodynamic equation and the nonlinearity of the structural inversion problem is accounted for through an iterative tomographic navigation process. More than half-a-million misfit measurements made on about 38,000 earthquake seismograms and 12,000 ambient-noise correlagrams have been assimilated into our inversion. After 26 F3DT iterations, synthetic seismograms computed using our latest model, CVM-S4.26, show substantially better fit to observed seismograms at frequencies below 0.2 Hz than those computed using our 3-D starting model CVM-S4 and the other SCEC CVM, CVM-H11.9, which was improved through 16 iterations of AW-F3DT. CVM-S4.26 has revealed strong crustal heterogeneities throughout Southern California, some of which are completely missing in CVM-S4 and CVM-H11.9 but exist in models obtained from previous crustal-scale 2-D active-source refraction tomography models. At shallow depths, our model shows strong correlation with sedimentary basins and reveals velocity contrasts across major mapped strike-slip and dip-slip faults. At middle to lower crustal depths, structural features in our model may provide new insights into regional tectonics. When combined with physics-based seismic hazard analysis tools, we expect our model to provide more accurate estimates of seismic hazards in Southern California. Advances in high-performance computing have recently enabled the application of F3DT to large regional and global data sets [Chen et al., 2007b; Tape et al., 2010; Zhu et al., 2012]. This paper describes how F3DT was used to assimilate more than a half-million measurements from earthquake waveforms and ambient-noise correlagrams to image the 3-D crustal structure of Southern California. We started with the Community Velocity Model of the Southern California Earthquake Center, version 4.0 (SCEC CVM-S4), and conducted 26 LEE ET AL.

The California Integrated Seismic Network (CISN) earthquake early warning (EEW) project is develo... more The California Integrated Seismic Network (CISN) earthquake early warning (EEW) project is developing components of a single, unified, prototype EEW system for the state of California. In summer 2009 the CISN concluded a three-year project testing three EEW algorithms: (1) Onsite, run by the California Institute of Technology, (2) Virtual Seismologist, run by the Swiss Seismological Service, and (3) ElarmS,

The increasing accuracy of 3-D earthquake ground motion simulations and demand for long-period (T... more The increasing accuracy of 3-D earthquake ground motion simulations and demand for long-period (T>1 s) probabilistic ground motions motivate the investigation of a potential method for incorporating simulated ground motions into the U.S. National Seismic Hazard Model. We present preliminary sensitivity results from an amplification-based approach to incorporating simulated ground motions into PSHA. Computations employ simulated ground motions from the Southern California Earthquake Center (SCEC) CyberShake Study 15.4. We outline a method for computing amplified ground motions and for their incorporation in probabilistic seismic hazard analysis. Empirical amplification factors from the analysis of smallto moderate-magnitude earthquakes are employed to identify CyberShake locations where the earthquake ground motions are well characterized by the ground motion prediction equations. Examining ruptures from a single seismic source from the Uniform California Earthquake Rupture Foreca...

In this poster, we present an approach to running workflow applications on distributed resources,... more In this poster, we present an approach to running workflow applications on distributed resources, including systems without support for remote job submission. We show how this approach extends the benefits of scientific workflows, such as job and data management, to large-scale applications on open-science HPC resources such as Blue Waters, Stampede, and USC HPCC. We demonstrate this approach with SCEC CyberShake, a physicsbased seismic hazard application, to run over 470 million tasks via 32,000 jobs submitted to Blue Waters and Stampede.

Journal of Open Source Software, 2022

Landscape Journal, 1998