CSEP2 Community Responses - Revision history

Wsavran: /* Global Forecasts (GEAR1) */

2018-08-01T22:19:26Z

Global Forecasts (GEAR1)

Wsavran: /* Global Forecasts (GEAR1) */

2018-08-01T22:17:43Z

Global Forecasts (GEAR1)

Wsavran at 22:15, 1 August 2018

2018-08-01T22:15:44Z

Wsavran: /* USGS Aftershock Forecasts (R&J and ETAS) */

2018-07-31T22:46:19Z

USGS Aftershock Forecasts (R&J and ETAS)

Wsavran: /* UCERF3-ETAS */

2018-07-31T22:19:46Z

UCERF3-ETAS

Wsavran: /* USGS Aftershock Forecasts (R&J and ETAS) */

2018-07-31T21:02:25Z

USGS Aftershock Forecasts (R&J and ETAS)

Wsavran: /* USGS Aftershock Forecasts (R&J and ETAS) */

2018-07-31T20:57:52Z

USGS Aftershock Forecasts (R&J and ETAS)

Wsavran: /* UCERF3-ETAS */

2018-07-31T17:50:50Z

UCERF3-ETAS

Wsavran: /* UCERF3-ETAS */

2018-07-31T17:43:05Z

UCERF3-ETAS

Wsavran: /* UCERF3-ETAS */

2018-07-31T17:40:14Z

UCERF3-ETAS

@@ Line 171: / Line 171: @@
 ; Object of forecast? Epicenters, Hypocenters, Faults? Cast as rates/probabilities, Sets of simulated catalogs? Forecast horizons, updates?
 ; Required data inputs to compute forecast?
 ;  Authoritative data source for testing forecasts?
 ; Available/unavailable tests, metrics etc. If unavailable, what scientific/computational developments need to occur to implement these tests?
 ; Specific ideas for Retrospective/Prospective testing? Timescales?
 ; Community involvement? competing models? New models/extant CSEP models?
 == General Responses ==

@@ Line 160: / Line 160: @@
 ; What are the scientific questions to be answered by experiments involving your model?
-: * Are larger earthquakes (m7.5+) occurring in proportion to the rate of smaller ones (m5.8+), which would be revealed by constructing the model as in the past, but testing only for m7.5+),
+: Are larger earthquakes (m7.5+) occurring in proportion to the rate of smaller ones (m5.8+), which would be revealed by constructing the model as in the past, but testing only for m7.5+)?
-Does surface strain rate represent the occurrence of larger earthquakes before the beginning of the catalog used to construct the forecast. If so, then we expect that the success of the forecast would improve if the weight given to surface strain rate is increased for longer test periods.
-Can the magnitude distributions of earthquakes over magnitude 5.8 be described in terms of uniform values on 5 tectonic region types, as assumed in the model, or do corner magnitude and b-value differ significantly within each tectonic type? For example, do all subduction zones have the same magnitude and b-value, or are there significant differences. If the latter, what controls the size distribution?
+: Does surface strain rate represent the occurrence of larger earthquakes before the beginning of the catalog used to construct the forecast. If so, then we expect that the success of the forecast would improve if the weight given to surface strain rate is increased for longer test periods?
 ; Model software requirements? Will the model be computed internal/external to CSEP? Required computing/memory/storage? Automated/On-demand? Versioned code? Status of code?
 ; Object of forecast? Epicenters, Hypocenters, Faults? Cast as rates/probabilities, Sets of simulated catalogs? Forecast horizons, updates?

← Older revision		Revision as of 22:15, 1 August 2018
Line 155:		Line 155:

	: all other short, medium and long-term models.		: all other short, medium and long-term models.
		+
		+	== Global Forecasts (GEAR1) ==
		+	Submitted by: Dave Jackson
		+
		+	; What are the scientific questions to be answered by experiments involving your model?
		+	: * Are larger earthquakes (m7.5+) occurring in proportion to the rate of smaller ones (m5.8+), which would be revealed by constructing the model as in the past, but testing only for m7.5+),
		+	Does surface strain rate represent the occurrence of larger earthquakes before the beginning of the catalog used to construct the forecast. If so, then we expect that the success of the forecast would improve if the weight given to surface strain rate is increased for longer test periods.
		+	Can the magnitude distributions of earthquakes over magnitude 5.8 be described in terms of uniform values on 5 tectonic region types, as assumed in the model, or do corner magnitude and b-value differ significantly within each tectonic type? For example, do all subduction zones have the same magnitude and b-value, or are there significant differences. If the latter, what controls the size distribution?
		+
		+	; Model software requirements? Will the model be computed internal/external to CSEP? Required computing/memory/storage? Automated/On-demand? Versioned code? Status of code?
		+
		+	; Object of forecast? Epicenters, Hypocenters, Faults? Cast as rates/probabilities, Sets of simulated catalogs? Forecast horizons, updates?
		+
		+	; Required data inputs to compute forecast?
		+
		+	; Authoritative data source for testing forecasts?
		+
		+	; Available/unavailable tests, metrics etc. If unavailable, what scientific/computational developments need to occur to implement these tests?
		+
		+	; Specific ideas for Retrospective/Prospective testing? Timescales?
		+
		+	; Community involvement? competing models? New models/extant CSEP models?

	== General Responses ==		== General Responses ==

@@ Line 92: / Line 92: @@
 ; Object of forecast? Epicenters, Hypocenters, Faults? Cast as rates/probabilities, Sets of simulated catalogs? Forecast horizons, updates?
-: Forecast is cast as a PDF of the number of event hypocenters within a spatial
+: Forecast is cast as a PDF of the number of event hypocenters within a spatial region, time period, and magnitude range.  There is no expectation that this PDF is any particular kind of distribution.
 : The ETAS forecasts will include something similar to a set of simulated catalogs, but not exactly.  The forecasts are based on temporal simulated event sets, while a static spatial kernel is used to spatially distribute the event rate.  So there are temporal simulated catalogs.  Spatial-temporal simulated catalogs could be created using the spatial kernel, but wouldn't have the full level of spatial correlations of a true ETAS simulation.

@@ Line 35: / Line 35: @@
 ; Model software requirements? Will the model be computed internal/external to CSEP? Required computing/memory/storage? Automated/On-demand? Versioned code? Status of code?
-: U3ETAS presently requires high performance computing.  Each simulation (synthetic catalog) takes from 1 to 20 minutes to generate, and we need some number of these to make robust statistical inferences (with the actual number depending on the metric of interest) . Results published to date have utilized about 10,000 simulations.  Kevin is configuring things so that anyone can run a set of simulations (so this could be done either internal or external to CSEP)(.  Nothing will be automated anytime soon.
+: U3ETAS presently requires high performance computing.  Each simulation (synthetic catalog) takes from 1 to 20 minutes to generate, and we need some number of these to make robust statistical inferences (with the actual number depending on the metric of interest) . Results published to date have utilized about 10,000 simulations.  Kevin is configuring things so that anyone can run a set of simulations (so this could be done either internal or external to CSEP).  Nothing will be automated anytime soon.
 ; Object of forecast? Epicenters, Hypocenters, Faults? Cast as rates/probabilities, Sets of simulated catalogs? Forecast horizons, updates?

@@ Line 90: / Line 90: @@
 : Forecasts will be computed externally to CSEP.  Currently on-demand, but in the process of automating.
-; Object of forecast: Epicenters, Hypocenters, Faults? Cast as rates/probabilities, Sets of simulated catalogs? Forecast horizons, updates?
+; Object of forecast? Epicenters, Hypocenters, Faults? Cast as rates/probabilities, Sets of simulated catalogs? Forecast horizons, updates?
 : Forecast is cast as a PDF of the number of event hypocenters within a spatial

@@ Line 86: / Line 86: @@
 : These forecasts are meant to inform the public and decision-makers, not to address any scientific questions.  As we evolve from Reasenberg & Jones to ETAS, we will be able to tests these two models against each other.
-; Model software requirements: Will the model be computed internal/external to CSEP? Required computing/memory/storage? Automated/On-demand? Versioned code? Status of code?
+; Model software requirements? Will the model be computed internal/external to CSEP? Required computing/memory/storage? Automated/On-demand? Versioned code? Status of code?
 : Forecasts will be computed externally to CSEP.  Currently on-demand, but in the process of automating.

← Older revision		Revision as of 17:50, 31 July 2018
Line 25:		Line 25:

	== UCERF3-ETAS ==		== UCERF3-ETAS ==
		+	Submitted by: Ned Field

	; What are the scientific questions to be answered by experiments involving your model?		; What are the scientific questions to be answered by experiments involving your model?

@@ Line 30: / Line 30: @@
 : Ideally, whether elastic rebound is really needed when adding spatiotemporal clustering to fault-based models, whether ETAS is an adequate statistical proxy for large-event clustering, and whether large triggered events can nucleate well within the rupture area of the triggering event (or only around the edges of the latter).
-The practical question is whether including faults is value added.
+: The practical question is whether including faults is value added.
 ; Model software requirements? Will the model be computed internal/external to CSEP? Required computing/memory/storage? Automated/On-demand? Versioned code? Status of code?

@@ Line 62: / Line 62: @@
 Here is our present to-do list with respect to U3ETAS:
 ) train others to run models on HPC and generate plots
 ) fetch recent M≥2.5 seismicity data from comcat and stitch together with the U3 catalog (which ends around 2014).
 ) associate any large CA events to U3 fault sections (probabilistically because there will never be a perfect fit).
 ) deal with catalog incompleteness?
@@ Line 75: / Line 75: @@
 We know that the rate of little events varies significantly with time, and that the probability of big events correlates with the rate of little events.  But there are uncertainties in what will happen looking forward, so the question is whether our forecasts provide useful information given these uncertainties.  Take any metric we are interested in (e.g., rate of M≥2.5 events or statewide financial losses); how do actual fluctuations compare with, say, the 95% confidence bounds from forecasts..
 Starting from whenever the M≥2.5 catalog is complete, I’d like to make forecasts at increments in time moving forward (monthly or yearly, or at some trigger points such as large event or before particularly quiet periods), and the plot the 95% confidence bounds (or whatever) of the forecast against what actually happens.  We can also project this analysis into the future by randomly choosing one the forecasts as what “actually” happens.  Again, the point is to tests whether our forecasts have value given the aleatory uncertainties looking forward.  Of course the answer will depend on the hazard or risk metric one care's about (so we will need to choose some to test), and the answer will also presumably vary depending on what’s happened recently (e.g., our forecast will presumably have some value after large events such as HayWired, but what about following particularly quiet times? Could CEA lower reinsurance levels during the latter?)  This analysis will require significant HPC resources.
 == USGS Aftershock Forecasts (R&J and ETAS) ==