Review of REDARS 1.0 Seismic Risk Analysis Software
Ronald T. Eguchi, ImageCat, Inc.
Methodology for Validating REDARS
The following flowcharts illustrate our approach for validating the REDARS program. We plan to perform our validation at two levels: component and system. The purpose of the component evaluations is to identify the extent to which each module deviates from actual earthquake experience. The figure below shows the procedure for comparing REDARS output (at the component level) to actual earthquake data.
We will use data from the 1994 Northridge and 1987 Whittier Narrows earthquakes to perform these validations. If it is determined that improvements are necessary to better match actual earthquake experience, the data from these component evaluations will help decide where to focus these efforts. Currently, we are planning on validating the following modules: ground motion modeling; damage state calculations; traffic state analysis; and direct economic loss. The economic loss evaluation will focus on two areas: repair costs of bridges and increase travel costs due to delays.
As part of the validation effort, we also hope to make a number of additional computer runs in order to test the sensitivity of the final results to changes in the individual modules. By conducting these sensitivity studies, we will be able to recommend to the methodology developer (Werner) which modules should be reviewed for possible improvements. The flowchart below outlines some basic steps that we plan to follow in developing this sensitivity matrix.
Data Collection for Validation Study
We have concentrated our efforts in three areas: ground motion data, bridge inventory information, and damage and loss information. We are focusing on Los Angeles and on two events: 1994 Northridge earthquake and the 1987 Whittier Narrows earthquake. Validating the REDARS model using the Northridge Earthquake as a test case has several major benefits. First, this event is probably the best-documented earthquake in the U.S, perhaps next to the 1971 San Fernando earthquake. There is considerable data and information on damage to all types of structures including bridges. Second, any loss estimation model in the U.S. should be tested against this event, if it is to have any credibility. Third, since the project team is located in the Los Angeles area, collecting additional data for this study should be fairly efficient, i.e., low travel costs, more in-person meetings.
The reason for selecting the Whittier Narrows earthquake as a second test case is two-fold: 1) it minimizes our need to collect hazard and inventory information for another area, and 2) it represents an event where little damage occurred. This last point is important since most current loss estimation models (HAZUS, EPEDAT) tend to overestimate losses at low magnitudes.
Ground Motion Data. We will use the TriNet ground motion map for one of our validation studies. This map, seen below, is available for the following ground motion indices: PGA, PGV, spectral ordinates for selected periods (0.3, 1.0 and 3 secs) and a generalized intensity index (MMI like). The PGA map is provided below.
Northridge PGA Map as developed by TriNet Methodology
Bridge Inventory and Damage State Information. We were able to obtain a bridge inventory database for all Caltrans bridges. Important attribute data includes: bridge name, bridge type, number of spans, deck length and width, year built, year retrofit, and location. The bridge locations are shown on the figure below.
Caltrans Bridge Sites (with Northridge Damage State Designations)
Overlaid onto TriNet PGA Map for the Northridge Earthquake
In addition, using data from Caltrans, we were able to re-classify the Caltrans bridge damage states from the Northridge earthquake into states that are used by the REDARS methodology. The REDARS methodology currently includes the following states: none, slight, moderate, extensive and complete. The following chart shows the distribution of damage states by Modified Mercalli Intensity (MMI). In order to show how these bridges performed on a regional basis, we have normalized the damage data by the total number of Caltrans bridges in the Los Angeles region. Note that the chart only provides a general representation of performance. That is, there is no distinction between different bridge types, ages, or retrofit conditions. We expect that there will be significant differences in the chart when these conditions are considered. This is one of our current research efforts. We are also trying to reconcile cost or repair data that we have received from Caltrans on the Northridge earthquake.
Traffic State Data. Several meetings have been held with Caltrans engineers (both in Sacramento and in Los Angeles) to discuss post-Northridge traffic data. So far, we have collected the following reports:
- Northridge Earthquake Recovery Report -Final Comprehensive Transportation Analysis
- Northridge Interim Transportation Report # 1 - January 17-March 31, 1994
- Northridge Interim Transportation Report # 2 - April 1-June 30, 1994
- Interstate 10 Recovery Report
- Interstate 5/State Route 14 Recovery Report
- Weekly and Bi-weekly Transportation Reports (March 7-July 8)
- Intercept Survey of Trucks on Regional Travel Routes into Los Angeles
- Home Interview Survey of Travelers Impacted by the Earthquake
- Survey of Transit Riders on Metrolink Commuter Rail and Bus Routes
- Follow-up Home Interview Survey of Travelers Impacted by the Earthquake
We are now in the process of developing our validation database for the Northridge earthquake. Current measures that are available from the above reports include: weekly traffic volume trends relative to pre-earthquake levels; detour traffic volumes; mix of vehicles on detour routes; detour travel times and delays; and transit ridership and services. We will most likely select several key routes to focus on. Likely routes include the segment of Interstate 10 (Santa Monica Freeway) that experienced several bridge collapses, and the Interstate 5/ State Route 14 bridge failures. There appears to be sufficient data to model these failures.
In a recent progress report, we outlined some critical issues regarding the objectives and scope of this workshop. Currently, the workshop has the following objectives:
- § Introduce the Seismic Risk Analysis (SRA) methodology and the REDARS 1.0 software
- Examine additional performance measures
- Identify ways in which loss estimation methods can be used for the design and construction of new facilities, the retrofit of existing facilities, and the development of more effective post-earthquake response and recovery strategies
- Identify a mechanism to involve end-users throughout this research project.
In a subsequent project memorandum, we discussed several options for conducting the workshop depending upon what our long-term strategy is with regard to the use and implementation of the REDARS software. We still need to resolve this issue. At any rate, the following user groups were identified:
Direct Users of REDARS:
- Federal Highway Administration
- State Highway Agencies
- Regional Transportation Agencies (Metropolitan Transportation Agency)
- Region Planning Organizations (Southern California of Area Governments)
- State and County Emergency Services Organizations
- Other federal agencies, e.g., FEMA
Indirect Users of REDARS:
- Local businesses
- The public, commuters
- Inter-state and intra-state transporters of goods
- Local police and fire
- Utility companies
- Public works organizations
We hope to begin installing the REDARS software at our site shortly. We will discuss with the developer of the methodology/software (Werner), the specific data requirements for the Los Angeles validation study. Our plan is to begin developing the requisite databases for the Los Angeles study.