Read more about the Haiti earthquake in Slate.
The 7.0-magnitude earthquake that rocked Haiti on Tuesday was the most powerful in the region for more than 100 years. A 7.2-magnitude quake struck the island in 1887, and another, from 1751, registered a 7.3. Wait a second—modern seismography dates back only to the late 19th century, and the Richter scale is just 75 years old. How do seismologists know the strength of a 1751 earthquake?
They guess based on contemporary damage reports. To measure earthquakes that happened before 1900, scientists review historical accounts of damage to buildings, the distance at which people felt tremors, and reports on changes in the soil. Each locality for which information is available is assigned a score on the Mercalli Scale, which uses whole number values to describe the effect of an earthquake on a particular locality. (For example, a score of 6 means people have difficulty walking and objects fall from shelves, but no major structural damage is reported. A 12, the highest score, means the ground moves in visible ripples and everything is destroyed.) If investigators can determine the epicenter and assign Mercalli values to a variety of locations around it, they can estimate magnitude by comparing the historical earthquake to modern quakes of known Richter magnitude and Mercalli values. The method isn't very reliable, and there often aren't enough data to come up with any estimate at all. Of the 10 deadliest earthquakes that occurred before systematic seismography, scientists have been able to assign a magnitude to only three.
The most important step in the process is finding the quake's epicenter. In many cases, seismologists find reports of relatively uniform damage across a broad geographic range. Other times, the historical record is unclear. Chinese documents from the 11th century describe an earthquake centered on Juyu Mountain (PDF) that caused boulders from the summit to fall directly into the sea. Two problems arose: First, there is no mountain that still goes by that name. When researchers eventually identified the peak on the Shandong Peninsula using other historical texts, they found it more than 10 miles inland, making it impossible for boulders to fall directly into the sea. Further research indicated that centuries of landfill have since separated the range from the water.
Once the epicenter is located, researchers have to sleuth for key facts about damage. Reports from areas very distant from the epicenter are particularly valuable. Descriptions of damage to buildings—which can take the form of written accounts, paintings, drawings, or woodcuts—are golden if the methods of construction are known, because scientists can model the effects of quakes of varying magnitudes on such a structure. In rare cases, scientists have found descriptions of changes to the area's topography, such as new water springs or cracks in the earth. A contemporary report on the 1690 earthquake in Villach, Austria, for example, gave the precise dimensions of a resulting landslide.
Even with a fairly detailed set of accounts, mistakes are often made. For example, if two major earthquakes happen within a few years, it's difficult to separate the damage from the two quakes.
Explainer thanks Charles A. Langston of the University of Memphis.