By William S. Ginell

In 1963 a devastating earthquake struck Skopje, Yugoslavia, destroying a large part of the city. Not only was the city's economic, social, and political life virtually paralyzed—its cultural property loss was catastrophic.

Skopje, today the capital of the Former Yugoslav Republic of Macedonia, is in the midst of one of the most seismically active areas of the world, a region that stretches from the western end of the Mediterranean Sea through North Africa, Italy, the Balkans, Greece, and Turkey, and on into central Asia. Temples, monuments, defensive structures, and buildings have all been subjected to destructive earthquakes that have raked the area for countless centuries.

The southern Balkan region historically has been ravaged not only by quakes but by a series of invading armies that preceded the establishment of a variety of cultures whose art and architecture survive to this day. Important remnants of this heritage include numerous Byzantine churches dating from the 9th to the 14th century. The interiors of these churches, still in daily use, were originally covered with frescoes, many of which have survived. The earthquake risk remains for these structures, as well as for other religious and historical monuments. Their destruction or damage would constitute an irreplaceable loss not only to the culture of the region but to the world.

The Skopje Project

Conservation image

Since 1985 the Getty Conservation Institute has promoted emergency preparedness and the development of preventive measures to protect cultural heritage from disasters such as earthquakes. As part of its efforts, the GCI in 1990 established a collaborative research program with the Institute of Earthquake Engineering and Engineering Seismology (IZIIS) of the University "St. Cyril and Methodius" and the Republic Institute for the Protection of Cultural Monuments (RZZSK) in Skopje to study seismic stabilization of Byzantine churches. These institutes offered the expertise of highly experienced professionals and modern experimental facilities to test retrofitting measures that could benefit structures in the Balkans and elsewhere.

Dr. Predrag Gavrilovic, Professor of Seismic Engineering at IZIIS, and Lazar Sumanov, a conservation architect and Deputy Director of RZZSK, were the project's principal investigators. Together they assembled a team of specialists in art history, architectural conservation, chemistry, archaeology, geophysics, and engineering. From the beginning, the objective was to develop methods that minimized physical intervention and that preserved the cultural values of the churches.

Typical Byzantine construction in the Balkan region is characterized by walls consisting of two outer faces of stone and brick set in lime mortar as well as a core between the faces filled with rubble set in lime mortar. About every meter up the wall were horizontal belts of brick or timber which provided continuity and ductility to the structure. Because the rubble cores of the walls are structurally weak, building stability was achieved by making walls and mortar joints very thick. However, as the mortar deteriorated with age and the timbers rotted away, the buildings' vulnerability to earthquakes increased.

The Skopje research program had three phases. In the first phase, data were collected on 50 existing Byzantine churches, including information on the topology of the structures, past conservation efforts, soil and meteorological conditions, and the historic, cultural, and artistic features of the churches. Preliminary structural analyses were made and possible repair and strengthening concepts explored. Ultimately, four churches were selected for detailed investigation, each representing a major type of Byzantine church architecture typical of the region: basilica, single nave, multidome apses, and single or five-dome churches with a cross-shaped interior (an "inscribed cross"). Most of the 50 churches studied fell into the last category.

The program's second phase included: definition of seismic parameters for the four specific church locations, studies of the physical-chemical properties and bearing capacity of building materials, experimental measurements of the dynamic characteristics of selected structures, and development of approaches for repair and strengthening. A major component of this phase was testing, on a seismic simulation shake table, of a scale model of one prototype church, before and after retrofitting.

In the final phase of research, analysis of the data, development of mathematical models, and vulnerability functions for various parts of church buildings were used to create general recommendations for retrofitting Byzantine churches.

The Church Of St. Nikita

Conservation image

The prototype chosen for modeling and testing was the Church of St. Nikita, located on Skopska Crna Gora, west of the village of Banjani, near Skopje. While there is no record of its date of construction, indirect evidence based on the church's frescoes (painted by two prominent medieval fresco painters, Mihajilo and Evtikie) suggests that St. Nikita was built in the early 14th century on the foundations of an earlier, Byzantine Empire church.

The church, part of a monastic complex, is a single-dome structure constructed in the shape of an inscribed cross. The walls are made of tuff (a weak volcanic rock) alternating with layers of brick set in lime mortar. Four interior brick columns support the massive brick tambour and dome.

The test model of St. Nikita, built on the 5-by-5 meter shake table, was 4.5 meters (14.8 feet) tall and weighed 21 tons. To create an accurate model, the project team measured St. Nikita's response to minor seismic activity in a series of simulation experiments, then, using the results, constructed a model that duplicated the real church's behavior during a seismic event.

To determine the actual seismic risk for St. Nikita, the team conducted geophysical surveys and, based on historical records, estimated the characteristics and return frequency of various types of earthquakes for the region. (As a rule, larger quakes occur less frequently than smaller ones.) These studies formed the basis for the computer-controlled earthquake simulations used in testing the model.

The original, undamaged model—extensively instrumented—was subjected to three types of earthquakes of varying intensity and duration. Twelve tests were performed, with intensity levels progressively increased. During the testing, the dome vibrated strongly. The model's first cracks appeared in the dome at a low intensity level, but no cracks occurred in the lower part of the structure. The final test at four times this level caused numerous cracks and the model's structural failure.

Following these tests, the model was repaired by injection grouting of cracks and by structural strengthening. Bolted horizontal steel tie rods were incorporated at three levels within the rubble part of the walls in the areas where wood timber belts originally existed in the prototype.

At the base of the dome a horizontal band was attached. Vertical steel ties were applied in the tambour and anchored to the main walls. In addition, exterior walls were anchored to the foundation with steel ties, and the spaces around the ties were filled with grout to provide positive connections with the walls. Thus, in an actual retrofit of a real church, no evidence of the engineering changes would be visible.

The retrofitted model was subjected to the same earthquake simulations used in the original tests. It was found that the maximum above-ground-level forces decreased and the cracking pattern and crack location changed. Displacement values—the amount of movement of building elements—were reduced by half.

To study the structure's behavior after cracking and to estimate the damage from higher-than-expected earthquake intensities, the earthquake amplitude was increased to a level that statistically would occur in the region only once in a thousand years. This resulted in some additional wall cracking and damage in the upper parts of the church, but no structural damage.

According to Dr. Gavrilovic, the tests in Skopje "showed that in the case of an earthquake of maximum expected intensity in the vicinity of St. Nikita, structural failure of the church, as it now exists, is likely and that seismic retrofitting should be undertaken."

Indeed, the experiments on the model of St. Nikita indicate that the church's structural stability can be increased enough to prevent structural damage during a major earthquake. The project also demonstrated that through careful design, structural stabilization can be achieved while the ethics of conservation are respected. The techniques used would neither alter the external appearance of the church nor damage its interior frescoes or other significant features.

It is expected that the analytical methods and retrofitting designs developed during this project will be applicable to the many historic structures of similar construction that are now at risk in earthquake regions throughout the world.

William S. Ginell is Head of Architecture and Monument Conservation Research in the GCI's Scientific Program.


Aisle One of several walkways of a basilica that extends the length of the church in an east-west direction.

Apse Semicylindrical or polygonal extension on the east end of a church.

Basilica Type of church having aisles, apse, nave, and often a narthex.

Ductility The ability of a material or structure to deform under stress without fracture.

Narthex Western entry area into a church leading to the nave and aisles.

Nave High central aisle of a basilica.

Rubble Uncut stone and brick used as random fill between two masonry walls.

Tambour Cylindrical or polygonal drum base for a dome.