A Site Response Map of the Continental U.S

We create a site response map of the continental U.S. using the topographic slope methodology, which uses correlations between slope and Vs30 (Wald and Allen, 2007). We determine new regional correlations for the central and eastern U.S. (CEUS) and western U.S. (WUS) calibrated to the Vs30 observation database of Pacific Engineering and Analysis, and introduce a novel correlation for areas of the WUS that hosted Pleistocene and younger lakes. In the WUS we develop a new correlation by first removing Vs30 observations from Utah basins and the Imperial Valley, California, from the database, and second using a stochastic simulation to choose slope and Vs30 bins and calculate standard deviation and bias. We select the best model based on three criteria: (i) bias and standard deviation is among the smallest; (ii) the correlations will not change dramatically for neighboring bins; and (iii) the Vs30 bin boundaries can be matched to NEHRP categories. Relative to WUS, the Utah and Imperial Valley Vs30 values are low for a given slope. We fit a separate correlation in the manner described above. We attribute the low values to the occupation of these areas by Pleistocene and younger lakes. We posit that when sediment-laden streams entered these lakes, the flow velocity was reduced so that all coarse sediments were deposited at the lake margin, and only very fine grained (and seismically slow) material was distributed over the lake bed. This model is confirmed by: (i) relatively high Vs30 values in the geologically similar Las Vegas Valley that was not occupied by a lake; (ii) ordinary Vs30 values in Utah and Imperial Valley locations above the high lake stands; and (iii) a consistent pattern of Vs30 values in the Reno, Nevada, basin across a paleo-lake boundary. In the CEUS, we use the recent correlation of Silva et al. (2011) that includes better constraints on the correlation at rock sites. In all regions the slopes are determined from SRTM digital elevation data (Jarvis et al., 2008). The CEUS correlation is appropriate for tectonically stable regions, and the WUS correlation for tectonically active regions. We merge the two by an east-west linear weighting over the Rocky Mountain physiographic province. The Rocky Mountains were formed 40 to 80 million years ago and have subsequently experienced only epeirogenic activity since, so represent a tectonic condition intermediate between the CEUS and WUS.