The walls of the Macellum show repair areas whose shape reflects a common mode of failure observed in masonry walls in earthquakes. The failure results from loads applied perpendicular to the face of the wall, so that it is loaded like a floor slab, and is called out-of-plane failure. There is also limited evidence to show that volcanic pyroclastic flow can induce out-of-plane failure in masonry walls. The following discussion reviews the nature of this failure mode due to seismic and volcanic loadings.
Figures 3.1 and 3.2 show walls of the Macellum that exhibit clear signs of damage and repair, where there is a patch that starts at the high corners of the wall, and scoops down toward the center.
This scooping pattern of damage is commonly observed in unreinforced masonry buildings subjected to earthquakes. Figures 3.3 and 3.4 show a building damaged in the 1954 Fallon Stillwater earthquake in Nevada, with a scooping pattern similar to those noted in the Macellum.
|Figure 3.3: Damage in the 1954 Fallon Stillwater Nevada Earthquake. Note that the height of the wall varies, being taller in the front, resulting in deeper scooping there. (photo by Karl V. Steinbrugge, from the EqIIS collection [EERC 1995])|
|Figure 3.4: A detail of figure 3.3. [EERC 1995]|
Loads acting perpendicular to the face of the wall cause the wall to burst outward. The line of failure tends to scoop upward toward the corners because the intersecting walls act as buttresses restraining the outward movement. Figures 3.5 and 3.6 show two more examples, one from the 1989 Loma Prieta earthquake in Santa Cruz, south of San Francisco, and another from the 1987 Whittier earthquake in southern California.
|Figure 3.5: Damage in Santa Cruz from the 1989 Loma Prieta Earthquake. The damage is partially visible in the upper left corner of the photo. (from the EqIIS collection [EERC 1995])|
|Figure 3.6: Damage from the 1987 Whittier Earthquake. Note that the failure line extends down to the second floor level, and that the scooping pattern is interrupted by a window at the left end. (from the EqIIS collection [EERC 1995])|
Note that the out-of-plane failure in figures 3.3 and 3.6 is limited to the upper story (this may also be true for figure 3.5, but the complete view is obscured); this is probably because the lower stories carry additional gravity loads from the floor framing, and this additional load pushing down on the wall counteracts the effects of the horizontal loads pushing outward, giving the lower walls greater resistance to out-of-plane loading. A one-story wall will tend to respond like the upper story of the examples above, as shown in figure 3.7, from the 1979 El Centro earthquake in California.
St. Pierre lies approximately 8 kilometers from the peak of Mt. Pelée, similar to the distance Pompeii lies from the peak of Mt. Vesuvius. In 1902, Mt. Pelée had two violent eruptions, one on May 8 and another on May 20, which completely devastated the city. Figures 3.8 and 3.9 show the city before the first eruption and after the second.
|Figure 3.8: The city of St. Pierre, before the 1902 eruptions. [LaCroix 1904, p. 232].|
|Figure 3.9: The city of St. Pierre, after the 1902 eruptions, June 1904. [LaCroix 1904, p. 233].|
In figure 3.9, the peak of Mt. Pelée is in the upper right of the photograph, and the direction of the pyroclastic flow was nearly parallel to the shore, generally flowing from the north to the south. Note in the photo that there is a predominance of walls standing parallel to the shore; these walls were parallel to the flow and so received little load from it. Walls not parallel to the flow experience much more damage. Figure 3.10 shows a detail from photograph looking north. The scooping shape associated with out-of-plane failure is visible in many of the east-west walls, oriented parallel to the plane of the photograph.
|Figure 3.10: St. Pierre after the 1902 volcanic eruptions. The photo is looking north; note the characteristic scooping pattern of failure in the east-west walls. [La Croix 1904, pl XIX]|
One of the major differences between the eruption of Vesuvius in 79 and the Eruptions of Mt. Pelée in 1902 is that the pyroclastic flow from Vesuvius was preceded by approximately 2.5 meters of Pumice fall, so that the flow hit only the portions of the walls protruding above this level. Sigurdsson concluded that the pyroclastic flow at Pompeii "toppled the walls of most buildings that protruded above the level of the volcanic deposit, and transported building fragments some distance" [Sigurdsson 1985, p. 352], however inspection of Pompeiian walls reveals that this statement should not be interpreted as meaning that walls were cut at the 2.5 meter level, since there are many examples throughout the city of ancient fabric standing to a heights exceeding four and five meters. Figure 3.11 shows an example to the east of the Macellum. This masonry wall is clearly ancient fabric from base to top. The perspective construction lines show that the height of the wall is three times the height of the door, which is approximately 1.8 meters, giving a wall height of approximately 5.4 meters. Many ancient walls of similar height can be found in Pompeii.
|Figure 3.11: Ancient masonry fabric, the perspective construction shows that the wall height is three times the door height, approximately 5.4 meters|
Figure 3.12 below shows the orientation of this wall with respect to the current peak of Vesuvius, based on measurements with respect to the Macellum and the plan. The Macellum is located in the lower left of the plan, and the wall is highlighted in the lower right. The angle between the perpendicular to the wall face and the direction to Vesuvius is slightly less than 30 degrees, so the face of the wall would be expected to receive most of the impact of pyroclastic flow from Vesuvius.
While the eruption of Mt. Pelée demonstrated the potential of pyrocalstic flow to be highly destructive, there is little evidence to suggest that it was so broadly destructive at Pompeii. Considering the Macellum, the tallest portions of the ancient fabric extend to approximately 5 meters. It is likely that any damage due to pyroclastic flow was above this level, so that the currently visible areas of scooping damage are probably not the result of pyroclastic flow.
Earthquakes and pyroclastic flow are both capable of inducing out-of-plane failure in masonry walls, leaving the scooping lines of damage found in several locations of the Macellum, although it appears unlikely that pyrocalastic flow was the cause of currently visible areas of damage and repair.
One of the conclusions from structural theory born out by observation after modern earthquakes is that, other factors being equal, out-of-plane failure is less likely to occur in walls that carry higher gravity loads: a fact that explains some damage patterns in the Macellum. Chapter 4 examines the Macellum in detail, identifying areas of out-of-plane damage and repair, and assessing possible causes.