JAIC , Volume 39, Number 1, Article 12 (pp. to )
JAIC online
Journal of the American Institute for Conservation
JAIC , Volume 39, Number 1, Article 12 (pp. to )

SEISMIC STABILIZATION OF HISTORIC ADOBE STRUCTURES

WILLIAM S. GINELL, & E. LEROY TOLLES



1 1. INTRODUCTION

Many of the early structures that date back to the Spanish Colonial period in the southwestern United States were built of mud brick or adobe. The materials available for construction of many of the early churches, missions, and houses were generally limited to those that were locally available and easily worked by local artisans. Adobe has many favorable characteristics for construction in arid regions: it provides good thermal insulation; the clayey soil from which adobe bricks are made is ubiquitous; the skill and experience required for building adobe structures are minimal; and building construction does not require the use of scarce fuel. As a consequence of their age and design and the functions they performed, surviving adobe buildings are among the most historically and culturally significant structures in their communities.

Historically, the seismic performance of adobe structures, as well as those of stone and other forms of unreinforced masonry, has in many cases been very poor. The seismic behavior of such buildings is characterized by sudden, dramatic, and catastrophic failures. The local or general collapse of adobe structures is often accompanied by a high likelihood of serious injuries and loss of life. Yet it has been observed that some adobes have withstood repeated severe earthquake ground motions without total collapse. Generally, the evaluation of the engineering community is that adobe buildings, as a class, pose the highest earthquake risks among the various building types, and the standard engineering analyses and approaches to the design of retrofits are not applicable to these structures.

Following the emergence of modern construction methods in which steel and reinforced concrete replaced brick and stone as principal building materials, structural designs were developed that could withstand environmental loads (wind and earthquake) and perform in a relatively predictable and acceptable way. Steel and reinforced concrete are ductile materials that have linear elastic properties and good postelastic strength characteristics. After yielding, these materials maintain most of their strength while undergoing substantial plastic deformations, and they can be analyzed with reasonable accuracy using analytic or computational methods. In contrast, after cracks are initiated, the behavior of brittle, unreinforced materials, such as stone, brick, or adobe, is extremely difficult to predict even with today's advanced computational capabilities. Even if computational results could be generated, the results would not be accurate. In a brittle material, once yielding occurs, cracks develop and there is a complete loss of tensile strength. After cracks have developed, the seismic behavior of adobe buildings is dominated by the interactions among large, cracked sections of the walls that rock out-of-plane and collide or rub against each other in-plane.

The fundamentals of the postelastic behavior of adobe are entirely different from those of ductile building materials because adobe is a brittle material. Once a typical unreinforced adobe wall has cracked, the tensile strength of the wall is completely lost, but so long as the wall remains upright and stable, it can still carry vertical loads. Cracks in adobe walls may occur from seismic forces, from settlement of the foundation, or from internal loads such as roof beams. However, even though the tensile strength of the wall material has been lost, the structure may retain its ability to remain standing. The thickness of typical historic adobe walls makes these walls difficult to destabilize even when they are severely cracked. The support provided at the tops of the walls by a roof system may add additional stability to the walls, especially when the roof system is anchored to the walls. In many adobe buildings, the height-to-thickness (slenderness) ratios may be less than 5 and the walls can be 1–2 m thick; both factors make wall overturning unlikely. Some seismic retrofit techniques can be used on other, thinner wall structures to improve stability and to reduce the differential displacements of the cracked sections during an earthquake.

Reinforced concrete bond beams, placed at the tops of walls below the roof, are often recommended for the upgrading of existing adobe buildings (California State Historical Building Code 1999). The function of bond beams is to provide lateral support and continuity at the tops of the walls. However, the installation of a bond beam usually requires the removal of the roof system, a very invasive and destructive procedure. The design of bond beams is often based upon elastic design criteria, which result in a very stiff bond beam. After cracks in the adobe walls develop during an earthquake, the stiffness of the bond beam may exceed the stiffness of the walls by 2 or 3 orders of magnitude. Adobe walls have been observed to pull out from underneath bond beams during an earthquake due to the difference in stiffness between the bond beam and the cracked wall sections and the lack of a positive connection between the bond beam and the adobe walls.

In prior years, some gains had been made by the engineering community in developing retrofit measures that were seemingly more sensitive to the conservation of historic fabric. Nevertheless, there was very little information about how these less-intrusive retrofit systems would, in fact, behave. In the late 1970s, a proposed retrofit system for the Cooper-Molera adobe building in Monterey, California, used interior and exterior wire mesh to enclose the adobe walls (Forell/Elesser Engineers 1977). This system was not implemented because there was little information on the performance of such a retrofit and concerns that the system would destroy too much of the historic fabric of the building. The system actually installed at Cooper-Molera consisted of reinforced concrete posts and beams in and on the walls (Forell/Elesser Engineers 1977–78), but this system, too, had a large negative impact on the historic fabric (Scawthorn and Becker 1986). A modified wire mesh system was tested later and was shown to be effective, but it remains a too highly intrusive technique for use on historic structures. All too often, the design of seismic upgrading modifications to existing hazardous buildings has focused on the goal of providing maximum life safety to occupants with little consideration being given to the limitation of damage to the buildings (Thomasen and Searls 1991).

Structural stability is the fundamental requirement for the adequate performance of adobe buildings during large earthquakes and should be the principal guide for designing appropriate retrofit measures. The walls of adobe buildings will crack during moderate to large earthquakes because adobe walls are massive and both the adobe brick and adobe mortar are low-strength materials. The massive walls respond to the large inertial forces that result from seismic ground accelerations, and they have relatively little strength to resist these forces. After cracks have developed, it is essential that the cracked elements of the structure remain in place and be able to carry vertical loads.

To achieve structural stability, it is necessary to mobilize adobe's very favorable postcracking energy dissipation characteristics while limiting relative displacements between adjacent cracked blocks. From our investigations, we have concluded that a stability-based approach to retrofitting historic adobe buildings can be the most effective method for providing life safety and for limiting the amount of damage during moderate to large earthquakes. The purpose of a stability-based approach is to prevent severe structural damage and collapse. Application of this approach should recognize the limitations of adobe while taking advantage of the beneficial, inherent structural characteristics of many historic adobe buildings—thick walls that are inherently stable and have a great potential for absorbing energy. These stability and energy absorption characteristics can be greatly enhanced by the application of a number of simple seismic improvement techniques.

Implementation of any stability-based approach must take into consideration that the seismic upgrading of historic buildings embraces two distinct and apparently conflicting goals:

  1. seismic retrofitting must provide adequate life-safety protection, and
  2. preservation of the historic architectural fabric of the building is of primary concern.

These goals are often perceived as being fundamentally opposed. Using current seismic retrofitting practices, substantial alterations of structures are usually required and often involve new structural systems and substantial removal or replacement of building materials. Although historic structures so modified and fundamentally altered can be stabilized, they can lose much of their authenticity. As a result, the conflict is often seen as an either/or proposition:

  1. Either an adobe building can be retrofitted to improve life safety during seismic events but destroying much of the historic fabric in the retrofitting process;
  2. or the historic fabric of the building can remain intact while accepting the risk of potential structural failure and collapse during future seismic events.

Faced with the apparent conflict between the unacceptable seismic hazard posed by many adobe buildings and the unacceptable conservation consequences of conventional retrofitting approaches, the Getty Seismic Adobe Project (GSAP) was initiated to develop and evaluate procedures that would recognize the simultaneous needs for achieving life safety while preserving the historic fabric and authenticity of the building.

In this project, we have formulated a retrofitting philosophy that is consistent with these objectives and is based on improving the stability of the adobe structure rather than improving its strength. The physical methods for carrying this philosophy out are described below, and they have been tested using 1:5 and 1:2 scale model adobe buildings that were shaken on earthquake simulation facilities in California and in the Republic of Macedonia. Some of the results of these tests are given below.