A Timely Reminder

A magnitude 5.6 earthquake in the San Francisco Bay area in the U.S. state of California this afternoon caused problems for the United States Geological Survey (USGS) website and is a timely reminder to avoid too much integration of reporting systems.

The magnitude 5.6 San Francisco Bay quake was located 15 km north-east of San Jose City Hall at a depth of 9 km. The earthquake struck at 8:05 p.m. local time on October 30th, (3:05 a.m. Universal Time) 4:05 p.m. New Zealand Daylight Time on Wednesday October 31st 2007.

The earthquake threw the USGS website into some disarray, either through overloading or system failures. The information for the San Francisco event became linked to that associated with a magnitude 6.8 event located off New Zealand’s Fiordland coast on October 16th. The USGS website ceased serving data more recent than 4:18 p.m. (NZDT) and at 5:44 p.m. the website was showing an update time of 4:42 p.m. but missing vital information.

In the meantime, a magnitude 7.2 earthquake had struck near the Mariana Islands at 4:30 NZDT. Information on this event was promptly notified by the United States National Oceanic & Atmospheric Administration’s (NOAA) Pacific Tsunami Warning Centre in a bulletin at 4:43 which advised that a tsunami was not generated.

The IRIS (Incorporated Research Institutions for Seismology) website managed to report the Mariana event by 5:34 p.m. but the site, which sources its data from the USGS, was unable to completely load. Geoscience Australia did not report the Mariana quake until near 5:54 p.m. By 6:15 p.m. the USGS website had recovered sufficiently to report the latest earthquake events, but the database corruption linking the San Francisco and Fiordland events still existed.

This article is not a criticism of the U.S. Geological Survey which provides a highly professional and timely earthquake reporting service. Rather it is to sound a warning that too much integration, which is often justified in cost savings, would have compromised reporting if the NOAA and USGS systems had been integrated. The value of duplicated infrastructure was amply illustrated by this sequence of events.

Incidentally, the USGS infrastructure has been either resilient enough, or recovered sufficiently to calculate a nearly one-in-ten chance of an aftershock causing strong shaking in the area north of San Jose during the next 24 hours. In the United States, strong shaking is defined as Modified Mercalli Intensity VI, or the level of shaking that throws objects off shelves.

2 Responses to “A Timely Reminder”

  1. Ken says:

    Some 4 hours after the San Francisco event a check of the USGS listings shows that the database corruption is more extensive than I first thought. Data relating to at least two quake events in Fiordland seem to have become linked to more recent earthquake reports on the USGS website. No doubt the database experts have been called out to effect repairs or recover from a backup.

  2. bruce says:

    The USGS site hasn’t posted any event in NZ on its world map since the Fiordland event and there have been several.. obviously something wierd going on in the system there somewhere.

Leave a Reply