La Habra Simulation Region - Revision history

Maechlin: /* Proposed Small Simulation Region */

2016-01-27T06:42:49Z

Proposed Small Simulation Region

Maechlin: /* Proposed Small Simulation Region */

2016-01-27T06:41:55Z

Proposed Small Simulation Region

Fsilva at 20:08, 26 January 2016

2016-01-26T20:08:47Z

Maechlin: /* Proposed Small Simulation Region */

2016-01-26T19:54:43Z

Proposed Small Simulation Region

Maechlin: /* Proposed Small Simulation Region */

2016-01-26T19:54:22Z

Proposed Small Simulation Region

Fsilva at 17:37, 26 January 2016

2016-01-26T17:37:26Z

Maechlin at 20:33, 25 January 2016

2016-01-25T20:33:37Z

Maechlin at 20:33, 25 January 2016

2016-01-25T20:33:13Z

Maechlin at 20:32, 25 January 2016

2016-01-25T20:32:20Z

Maechlin at 20:31, 25 January 2016

2016-01-25T20:31:59Z

@@ Line 9: / Line 9: @@
 *set (39.9 azimuth) XLEN = 28.0
-*set () YLEN = 28.0
+*set (129.9) YLEN = 28.0
 *set (points down) ZLEN = 12.0
 *set HH = 0.020

@@ Line 1: / Line 1: @@
 == Proposed Small Simulation Region ==
-High-F modelers have proposed to design a "small" 3D model for La Habra EQ that includes the closest 3 sites.  We can then run tests with this model until the behavior of the codes and agreement of the results.  This would only be a short term endeavor and I do not envision this being part of a paper (unless we discover something HUGE!), but it would help us move forward in a much more efficient and practical manner (as opposed to trying to run test cases with the full 3D volume).
+High-F modelers have proposed to design a "small" 3D model for La Habra EQ that includes the closest 3 sites.  We then run tests with this model until the behavior of the codes and agreement of the results, to move forward in a much more efficient and practical manner (as opposed to trying to run test cases with the full 3D volume).
 An additional benefit is that using a small model would make testing/iterating on UCVM development much more efficient.  For example, modifications for RWG mesh output and testing/implementing velocity thresholds. As a potential starting point, here are the model parameters RWG used for his latest runs:

@@ Line 41: / Line 41: @@
   proj=+proj=utm +datum=WGS84 +zone=11
   rot=-39.9
-  x0=-119.288842
+  x0=-118.20819
-  y0=34.120549
+  y0=33.85173
   z0=0.0
   # Number of cells along each dim
-  nx=9000
+  nx=1400
-  ny=6750
+  ny=1400
-  nz=3100
+  nz=600
   # Partitioning of grid among processors
-  px=30
+  px=25
   py=25
-  pz=10
+  pz=20
   # Vs/Vp minimum
-  vp_min=700
+  vp_min=200
-  vs_min=500
+  vs_min=0
 == Simulation Region Map ==

@@ Line 10: / Line 10: @@
 *set (39.9 azimuth) XLEN = 28.0
 *set () YLEN = 28.0
-*set ZLEN = 12.0
+*set (points down) ZLEN = 12.0
 *set HH = 0.020

@@ Line 8: / Line 8: @@
 *set MODEL_ROT = -50.1		# counter-clockwise rotation of X_rwg from default azimuth of 090
-*set XLEN = 28.0
+*set (39.9 azimuth) XLEN = 28.0
-*set YLEN = 28.0
+*set () YLEN = 28.0
 *set ZLEN = 12.0
 *set HH = 0.020

← Older revision		Revision as of 17:37, 26 January 2016
Line 23:		Line 23:

	The above parameters give NX=1400, NY=1400, NZ=600, and and with set NT=60000 (60 sec). This took about 4 hours to run on the Caltech cluster using 864 processors, not counting queue time.		The above parameters give NX=1400, NY=1400, NZ=600, and and with set NT=60000 (60 sec). This took about 4 hours to run on the Caltech cluster using 864 processors, not counting queue time.
		+
		+	== Current UCVM Configuration File ==
		+
		+	# List of CVMs to query
		+	ucvmlist=cvmsi
		+
		+	# UCVM conf file
		+	ucvmconf=../conf/ucvm.conf
		+
		+	# Gridding cell centered or vertex
		+	gridtype=CENTER
		+
		+	# Spacing of cells
		+	spacing=20.0
		+
		+	# Projection
		+	proj=+proj=utm +datum=WGS84 +zone=11
		+	rot=-39.9
		+	x0=-119.288842
		+	y0=34.120549
		+	z0=0.0
		+
		+	# Number of cells along each dim
		+	nx=9000
		+	ny=6750
		+	nz=3100
		+
		+	# Partitioning of grid among processors
		+	px=30
		+	py=25
		+	pz=10
		+
		+	# Vs/Vp minimum
		+	vp_min=700
		+	vs_min=500

	== Simulation Region Map ==		== Simulation Region Map ==

@@ Line 23: / Line 23: @@
 The above parameters give NX=1400, NY=1400, NZ=600, and and with set NT=60000 (60 sec).  This took about 4 hours to run on the Caltech cluster using 864 processors, not counting queue time.
 == Simulation Region Map ==
-*[[Image:base lahabra-win1.png|left|250px]]
+== Related Entries ==
 *[[High-F]]
 *[[Main_Page]]

← Older revision		Revision as of 20:31, 25 January 2016
Line 1:		Line 1:

−	*[[]]	+	== Proposed Small Simulation Region ==
−	File:Base lahabra-win1.png	+	High-F modelers have proposed to design a "small" 3D model for La Habra EQ that includes the closest 3 sites. We can then run tests with this model until the behavior of the codes and agreement of the results. This would only be a short term endeavor and I do not envision this being part of a paper (unless we discover something HUGE!), but it would help us move forward in a much more efficient and practical manner (as opposed to trying to run test cases with the full 3D volume).
		+
		+	An additional benefit is that using a small model would make testing/iterating on UCVM development much more efficient. For example, modifications for RWG mesh output and testing/implementing velocity thresholds. As a potential starting point, here are the model parameters RWG used for his latest runs:
		+
		+	*set MODEL_LON = -117.99503 # LON at center of model region
		+	*set MODEL_LAT = 33.86771 # LAT at center of model region
		+	*set MODEL_ROT = -50.1 # counter-clockwise rotation of X_rwg from default azimuth of 090
		+
		+	*set XLEN = 28.0
		+	*set YLEN = 28.0
		+	*set ZLEN = 12.0
		+	*set HH = 0.020
		+
		+	The map below shows this region (black rectangle) along with the epicenter and recording stations. Green triangles indicate
		+	sites where we have been comparing waveforms (FUL, DLA, BRE are inside this small region). The 4 corners of the proposed
		+	region (magenta circles) are:
		+
		+	* c1= -118.20819 33.85173
		+	* c2= -118.01410 34.04473
		+	* c3= -117.78178 33.88333
		+	* c4= -117.97604 33.69068
		+
		+	The above parameters give NX=1400, NY=1400, NZ=600, and and with set NT=60000 (60 sec). This took about 4 hours to run on the Caltech cluster using 864 processors (zero wait time in queue, too!).
		+
		+	*[[Image:File:Base lahabra-win1.png\|left\|150px]]
		+
		+	== Related Entries ==
		+	*[[High-F]]
		+	*[[Main_Page]]