A topic in the Open Knowledge Graph — a free, open map of 15,290 topics and the order to learn them in.
The Gutenberg-Richter relation log₁₀(N) = a − b·M describes the frequency-magnitude distribution of earthquakes, where N is the cumulative count of earthquakes with magnitude ≥ M. The b-value (typically ~1.0) indicates that earthquakes follow a power-law distribution: roughly 10 times fewer earthquakes for each unit increase in magnitude. Deviations indicate changes in stress or fault behavior.
Plot earthquake catalogs from different regions on log-linear graphs and fit the Gutenberg-Richter relation to compute b-values. Compare b-values before and after large earthquakes to observe stress-related changes.
The b-value is the same everywhere (it varies regionally and temporally). A higher b-value indicates more large earthquakes (it actually indicates more frequent smaller events relative to large ones).
From your study of moment magnitude, you know that each earthquake has a size that can be precisely quantified. The Gutenberg-Richter relation answers the next natural question: how often do earthquakes of each size occur? The answer turns out to be strikingly regular. If you take an earthquake catalog for any well-monitored region — say, Southern California over 20 years — and count how many events exceed each magnitude threshold, then plot those counts on a logarithmic vertical axis against magnitude on the horizontal axis, you get an almost perfectly straight line.
The equation for that line is log₁₀(N) = a − bM, where N is the cumulative number of earthquakes at or above magnitude M. The a-value is the y-intercept and reflects the overall seismicity rate: a region with many earthquakes of all sizes has a high a-value. The b-value is the slope of the line and is the more physically interesting parameter. Because the logarithm of properties matters here, a b-value of 1.0 means that for every unit increase in magnitude, the number of earthquakes drops by a factor of 10. So if a region produces 1,000 magnitude-3 events per year, it produces roughly 100 magnitude-4 events, 10 magnitude-5 events, and 1 magnitude-6 event. This is a power-law distribution — the same mathematical pattern found in many natural phenomena from river floods to asteroid impacts.
The b-value is not a universal constant. It typically hovers near 1.0 globally, but it varies meaningfully between tectonic settings. Volcanic and geothermal areas often show elevated b-values (1.2–1.5), meaning small earthquakes are disproportionately common relative to large ones — a signature of heterogeneous, thermally weakened rock generating many small fractures. Subduction zones locked and accumulating strain before a great earthquake may show depressed b-values (0.7–0.9), indicating that a larger fraction of the seismic energy is released in bigger events. Monitoring temporal changes in b-value is one tool seismologists use to track evolving stress states, though it has not proven reliable enough for deterministic earthquake prediction.
The practical power of the Gutenberg-Richter relation lies in seismic hazard assessment. If you can estimate the a- and b-values for a fault zone or region from decades of catalog data, you can extrapolate to estimate how often rare, large events occur — even if none have been observed in the instrumental record. For example, if the catalog implies one magnitude-7 event every 200 years, that probability feeds directly into building codes, insurance models, and emergency planning. The key assumption is that the linear relationship continues to hold at high magnitudes, which it generally does until you approach the maximum magnitude a fault can physically produce, at which point the distribution tapers off.