HighF v14.12 Data Comparison

From SCECpedia
Revision as of 19:35, 20 May 2016 by Maechlin (talk | contribs)
Jump to navigationJump to search

In this page we show a comparison between seismograms generated by Hercules and AWP for some of the stations in the La Habra station list.


Simulation Results - 17 May 2016

AWP simulation results compared to Center for Engineering Strong Motion Data (CESMD) processed V2 data for La Habra Earthquake for large simulation region. Matching synthetics and observations for 217 sites.

Simulation Results - 11 May 2016

In these plots, Red are data, blue are synthetics. This corresponds to comparisons of 0–0.25 Hz synthetics with data from BH channels downloaded from SCEDC site. The synthetics correspond to the last run Hercules using the large domain. That run was only for 60 s of simulation, so several stations appear cut and in some this introduces artifacts in the signals because of filtering abruptly cut traces. This will be repair in next release.

Simulation Results - 01 May 2016

Data from the small region against observations.

Simulation Results - 21 March 2016

Same as 3 March but using corrected AWP data. Updated results using the small domain region. There are two sets of AWP results, one using RWG's rules for clamping Vp/Vs and another one using Bill's rules, which can be described by conditions where Vp/Vs is limited between 1.45 and 3. VpMin is clamped at 1500 m/s and VsMin is clamped at 500 m/s. RWG's rules, on the other hand can be described with conditions that limits Vp/VsMin to 1.45 and VsMin to 500 m/s. Hercules data was calculated with the following simulation parameters: Max Frequency: 4 Hz, Minimum Vs: 500 m/s, Minimum Vp:1500 m/s, Q = 100 Vs, Domain: 28*28*14 km. RWG data was calculated with the small model region. These use Vsmin = 500 m/s, Vpmin = 1500 m/s.

First set of data includes RWG, Hercules, and AWP-RWG:

Simulation Results - 3 March 2016

Don't use these plots, use data from 21 March above instead.

Updated results using the small domain region. There are two sets of AWP results, one using RWG's rules for clamping Vp/Vs and another one using Bill's rules, which can be described by conditions where Vp/Vs is limited between 1.45 and 3. VpMin is clamped at 1500 m/s and VsMin is clamped at 500 m/s. RWG's rules, on the other hand can be described with conditions that limits Vp/VsMin to 1.45 and VsMin to 500 m/s. Hercules data was calculated with the following simulation parameters: Max Frequency: 4 Hz, Minimum Vs: 500 m/s, Minimum Vp:1500 m/s, Q = 100 Vs, Domain: 28*28*14 km. RWG data was calculated with the small model region. These use Vsmin = 500 m/s, Vpmin = 1500 m/s.

First set of data includes RWG, Hercules, and AWP-RWG:

Second set of plots also includes AWP-WHS:

Simulation Results - February 22 - Using small domain

The following results correspond to runs done with Hercules comparing previous results obtained with the large simulation domain against new results obtained with the smaller simulation domain. The former are shown in red, whereas the latter are shown for two conditions in blue and green. The results shown in blue are equivalent to those in red with the only difference that the red results were obtained with the original large 180 km x 135 km domain, but the blue were obtained with the smaller 28 km x 28 km domain. In both these cases the minimum Vs is equal to 500 m/s but there is no cap on the minimum Vp. Then the results shown in green correspond to a run done with the smaller domain but now applying a constrain to a minimum Vp equal to 1500 m/s.

Proposed Small Simulation Region

Simulation Results - 5 January 2016

Same data as 14 December 2015, but updated RWG data for 3 closest stations computed using a 3D velocity model generated directly from "ucvm_query".

Simulation Results - 15 December 2015

Same data as 14 December 2015, but adding observation data.

Simulation Results - 14 December 2015

Same data from 7 December 2015, updated RWG data for the three closest stations only.

Simulation Results - 7 December 2015

Same data from 20 November 2015. Here are the steps used to process the simulated time series:

  • Read all seismograms for a given station, find out largest dt, it becomes new_dt
  • Filter at 10Hz back and forward, 4 poles, using filtfilt
  • Interpolate using 1D linear to new_dt (in this case, AWP's 0.025 dt)
  • Trim longest seismograms to match the shortest one
  • Taper at the last 2 seconds: hanning with 161 points, then take 81 last values and apply this to last 81 points of each time series
  • Pad last 5 seconds with zeroes
  • Filter at 4Hz 4 poles back and forward with filtfilt
  • Plot time series and FFT using 1.25 filter frequency
  • Repeat last 2 steps with filter frequency set to 0.5, 1, 2, and 3 Hz

Related Entries