Exciting news! We're transitioning to the Statewide California Earthquake Center. Our new website is under construction, but we'll continue using this website for SCEC business in the meantime. We're also archiving the Southern Center site to preserve its rich history. A new and improved platform is coming soon!

Testable Earthquake Forecasts for 1999

David D. Jackson, & Yan Y. Kagan

Published 1999, SCEC Contribution #484

We describe in this paper a “long term” forecast of earthquakes with magnitudes greater than or equal to 5.8, evaluated for calendar year 1999 over two active seismic regions. We also describe a “short term” daily forecast that explicitly includes clustering. These forecasts are based on earthquake data only, with no explicit use of tectonic, geologic, or geodetic information or any physical model.

Our primary purposes are to develop a quantitative standard against which geological constraints and physical theories of earthquake potential can be compared, to offer an example of a general format for rigorously testable forecasts, and to give our best present estimate of seismic potential. We describe a likelihood-function method for testing the consistency of probabilistic forecasts against the ensuing earthquake record and for comparing forecasts. We have archived specific forecasts for the Northwest Pacific and Southwest Pacific regions of the Earth. At the year’s end, we will test these forecasts against the 1999 earthquake catalog. In the near future, we will extend these forecasts to other seismically active regions and eventually the entire Earth.

Our forecasting methods are evolving. They depend on some arbitrary functional forms and a handful of parameters estimated from earthquake data. We expect that further experimentation and optimization will lead to improved forecasts and that yearly open testing of the forecasts will speed that improvement. We hope that open testing will also stimulate competition for more skillful forecasts.

Earthquake potential models that include spatial variations of earthquake rate, reasonable magnitude distributions, and spatial and temporal clustering are not new. What is new here is a formal statement of these effects and a quantified, testable model that has undergone some optimization to improve its forecasting effectiveness.

We present these forecasts for scientific evaluation only. They are not earthquake predictions or warnings and they carry no official endorsement.

Citation
Jackson, D. D., & Kagan, Y. Y. (1999). Testable Earthquake Forecasts for 1999. Seismological Research Letters, 70(4), 393-403.