Earthquakes endanger New York more than previously believed

concentrate, and where larger ones could come. The Lamont network, now led by coauthor Won-Young Kim, has located hundreds of small events, including a magnitude 3 every few years, which can be felt by people at the surface, but is unlikely to cause damage. These small quakes tend to cluster along a series of small, old faults in harder rocks across the region. Many of the faults were discovered decades ago when subways, water tunnels and other excavations intersected them, but conventional wisdom said they were inactive remnants of continental collisions and rifting hundreds of millions of years ago. The results clearly show that they are active, and quite capable of generating damaging quakes, said Sykes.

Especially worrisome is one major previously known feature, the Ramapo Seismic Zone, which runs from eastern Pennsylvania to the mid-Hudson Valley, passing within a mile or two northwest of Indian Point, the location of an ageing nuclear power plant. The researchers found that this system is not so much a single fracture as a braid of smaller ones, where quakes emanate from a set of still ill-defined faults. East and south of the Ramapo zone-and possibly more significant in terms of hazard — is a set of nearly parallel northwest-southeast faults. These include Manhattan’s 125th Street fault, which seems to have generated two small 1981 quakes, and could have been the source of the big 1737 quake; the Dyckman Street fault, which carried a magnitude 2 in 1989; the Mosholu Parkway fault; and the Dobbs Ferry fault in suburban Westchester, which generated the largest recent shock, a surprising magnitude 4.1, in 1985. Fortunately, it did no damage. Given the pattern, Sykes says the big 1884 quake may have hit on a yet-undetected member of this parallel family further south.

The researchers say that frequent small quakes occur in predictable ratios to larger ones, and so can be used to project a rough time scale for damaging events. Based on the lengths of the faults, the detected tremors, and calculations of how stresses build in the crust, the researchers say that magnitude 6 quakes, or even 7 — respectively 10 and 100 times bigger than magnitude 5 — are quite possible on the active faults they describe. They calculate that magnitude 6 quakes take place in the area about every 670 years, and sevens, every 3,400 years. The corresponding probabilities of occurrence in any