Duration–amplitude relationships of volcanic tremor and earthquake swarms preceding and during the 2009 eruption of Redoubt Volcano, Alaska

• We generated duration–amplitude plots for the 2009 eruption of Redoubt Volcano.
• Periods of volcanic tremor and earthquake swarms were better fit by an exponential law.
• Periods without significant tremor or swarms occurring were better fit by a power law.
• We propose using the ratio of regression coefficients (R2exp/R2pow) as a monitoring tool.
• We found that stronger volcano-seismic episodes had higher characteristic amplitudes.

Duration–amplitude relationships were studied for tremor episodes and earthquake swarms occurring during the 2009 eruption of Redoubt Volcano, Alaska. Duration–amplitude distribution plots were generated daily from January 1 to May 31 and fit with both an exponential law and power law. Comparing R2 values of the fit for both laws showed that the exponential law fit better for days in which volcanic tremor and earthquake swarms occurred, while the power law fit better for other days. Fitting segments of seismic data with both an exponential and a power law leads to a metric that has potential for volcano monitoring: R2exp/R2pow, the ratio of the R2 fits using the exponential law and the power law. The ratio R2exp/R2pow tended to be greater than 1 when volcanic activity or precursory seismic activity was occurring, and less than 1 when no volcano-seismic activity was occurring. Duration–amplitude plots were generated for episodes of volcanic tremor that were identified by the R2exp/R2pow ≥ 1 method and compared in an attempt to identify changes that may have occurred during the eruption. Stronger episodes of volcanic tremor showed higher characteristic amplitudes. Maximum heights of the plumes generated by the explosions showed a positive correlation with the characteristic amplitude of the concurrent tremor.