Historic Structure Report, Volume 2, Castillo de San Marcos National Monument: Stabilize For Pkg 116


Material Information

Historic Structure Report, Volume 2, Castillo de San Marcos National Monument: Stabilize For Pkg 116
Series Title:
Castillo de San Marcos Terreplein Project
Physical Description:
Architectural Data by: C. Craig Frazier and Randall Copeland; Historical Data by: Luis R. Arana
Publication Date:
Physical Location:
Folder: Prior excavations, research, & reports

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
97-3003; 93-3000; 94-3000
Castillo de San Marcos
System ID:

Full Text



Stabilize Fort Pkg 116

August 1986

Architectural Data
Historical Architects C. Craig Frazier and Randall Copeland
Denver Service Center

Historical Data
Historian Luis Arana Castillo de San Marcos National Monument

U.S. Department of. the Interior / National Park Service

periodically patched by the park staff and well maintained. Those that lie-below present surfaces are presumed to be appropriately protected. (For further discussion of tabby slabs, their properties, and duplication, see appendix J).

2. Terreplein

Technically, the terreplein is an earthen deck upon which defensive cannon are operated. At the Castillo, the terreplein is an architecturally complex system including a tabby slab surface, stone firing steps, sentry boxes, chimney penetrations, and a rainwater management system consisting of drainage ports, scuppers, and flashing. Any recommendations about this system must take into account the elevational relationships of all these elements and also the functioning of the terreplein as a rainwater management feature and walking surface for thousands of visitors. The preservation and condition of the stone parapets at the terreplein level have been discussed above. This subsection will examine the deck itself and several related features that interrelate with the function--and dysfunction--of the terreplein today.

The existing condition of the terreplein is analyzed below in terms of the various functions for which it was built and by addressing relevant historical and contemporary design questions.

a. Floor Function

Does the terreplein serve as a functional floor surface, i.e., permit the mounting and operation of cannon and walking use by


visitors? Yes. The structure supporting the terreplein can accommodate the load of cannon and visitors. The conservative engineering calculations (see appendix L) give the terreplein a loading capacity of 100 to 250 pounds per square foot. The current tread surface has stood up fairly well to millions of visitors in its 20-year life.

b. Proper Elevation

Does the terreplein, being a succession of layers, replacements, and restorations, have the proper design elevation? Maybe not. The terreplein is currently about 28.5 feet above mean sea level. The "correct" elevation of the terreplein is not an easy matter to determine. When originally installed, the crest of the mid-18th century parapets was said to be 6 feet above the deck. However, the historical Spanish foot is not equal to our present 12-inch foot; its length, in fact, has not been confirmed.* Furthermore, such a statement as "the parapets are six feet high," must be considered a generalized statement, since the level of the terreplein deck surely sloped then, as it does today, for drainage and therefore varied in elevation several

*Note: In 1568 Felipe II established the standard Spanish yard or vara at 33.755 inches, i.e., just over 11 inches per foot. For various reasons this length fluctuated throughout the empire (e.g., in Nicaragua the vara measured 27.507 inches, while in Costa Rica it stretched to 38.375); however, in Mexico it seems to have maintained its standard at just under 33 inches (32.96 or 32.9731 inches, depending on temperature) (Bowman 1961). Deagan (1976) states on p. 105 that "the St. Augustine vara was determined by the author through correlation of archeological house footing measurements with the vara dimensions given by Roque (1788) for the same buildings" to be 1.02 meters or 40.1574 inches, or 13.38 inches per Spanish foot. Thus, the Spanish, foot probably equalled something between 11 and 13 inches.


inches, while the crest of the parapets was quite likely consistent. On the one hand if today's parapets are within 4 to 6 inches of 6 feet in height, the terreplein elevation may be considered "correct." In fact, the parapet crest consistently measures 5'10" to 6'0", quite satisfactory.

On the other hand, one could conclude that the terreplein is 8 or 9 inches higher than it was in 1756 and that most of one historic floor layer (tabby floor C) is currently below the modern layers. Over the centuries the elevation of the terreplein has fluctuated measurably--as much as 10 inches--because of repairs, replacements, remodeling, and restorations (see figs. 10, 11, and 12). The lowest colonial terreplein surface identified by Manucy in his December 1939 report occurred some 10 inches and three distinct floor layers below the (then) existing surface (at about 27.76 feet). The deck installed after his investigation in 1939 created an elevation about 2 inches lower than the one it replaced, which would make it about 8 inches above the lowest historic surface. The present concrete slab (installed in 1960) and the Laykold wear coat waterproofing system (installed in 1964) vertically match the post 1939 elevation (approximately 8 inches above the lowest colonial surface).

The pre-1939 level was established when 2 inches of concrete with an asphalt impregnated paper underlayment was installed between 1889 and 1891. This deck received a coating of paraffin and petroleum in 1892, followed by subsequent repairs and partial replacement in some locations. It was then totally removed in 1939. Two of the other three floors documented by Manucy (tabby floors A and B, totaling about


7-3/4 inches) were also removed in 1939 because they did not provide a satisfactory base for the proposed terreplein restoration. When tabby floors A and B were installed is unconfirmed. It is known that the original deck (probably the 1-foot-thick deck labeled C in 1939) was in place by 1756, and records indicate that the Spanish replaced the terreplein (probably installing a new tabby slab over the existing) on three sides and two bastions in 1800-1802 (this is very possibly Manucy's floor A). Many other repairs--mostly patches--are noted in historical accounts, but no other clearly identified floor installation is recorded. It is possible that floor B was installed by the British, since they installed a new terreplein at Fort Matanzas, but the records are lacking to confirm or disprove this hypothesis.

It is clear that the current surface is above the historic surface (about 8 inches); but it is also true that the general elevational relationship between terreplein and parapets is acceptable. A restoration of the terreplein could be considered appropriate if it resulted in a decrease in elevation (approaching the historic level) and an increase in the height of the parapet (approaching 6 Spanish feet, which is perhaps

6 feet 6 inches).


Fig. 71. General Cross Section of the Terreplein, 1891-1939. This drawing is based on the findings of Albert Manucy (1939 NPS file report) and shows typical conditions prior to the replacement project of 1939.

Fig. 72. Terreplein Section Detail, 1939. This illustration shows the terreplein design of 1939. Notice that tabby layer C--possibly the original 1756 deck--was left in place, and any future work should make

I every effort to keep it intact.





I 97

f N ( 71

7 6_p-T E 16811

--- 7
TA, 'Y tc
27 .-.o
j N

TAs ec VA.1

4-8!' 7fICK k LOW 'C:

palabr jCYIAp id OALL

7. 57
Z-7,7 T-pbby C A

Af HALT Lp Al A o41

Fig. 73. Contemporary Terreplein Section Details. The 1939 18"x18"

concrete pavers and tarpaper were removed except under the firing steps and at the parapet and replaced with an 8'x8' grid of reinforced concrete in 1960. In 1964, the 1960 slab joints were cut out and filled with okum and sealant, and the entire terreplein surface, including the firing steps, was coated with the present Laykold wear coat membrane (hot asphalt-impregnated layers of burlap and Laykold waterproof paint). At the perimeter the 1939 copper flashing was restored and incorporated into the new work. The 1940 reconstructed firing steps had been patched with coquina concrete in 1947. These had to be repaired again with a topping of concrete in 1964 prior to installation of the Laykold system.


AFPA op ct?cl-I T A P Ja6eT LAycoLp 6 T lM PAILNA4 L

5sLA (co) -- ,

"i 4 9" CO ,', 14

S-AgP FAC ,0 c 'A


c. Roof Function

Does the terreplein properly function as a roof, i.e., manage rainwater? No. Based on historical records, there is only one documented period of time during which the casemates below the terreplein were dry: a few years after the 1939 system was installed (NPS file report by Manucy, 1963c). Leakage was recorded as a problem by the Spanish as early as 1707, by the British when they came in 1764 and again in 1766, by the Spanish in their second period in the 1790s and 1820, and by the U.S. military upon their arrival in 1821 and again in 1833, 1834, 1839, and later in 1884 and 1886, leading to the 1889-91 repair work and the 1892 coating. Reports of problems associated with leakage have continued into the present century. Shortly after each case of leakage problems was reported, attempts were made to repair the terreplein to alleviate the problem--but always, only temporary results were achieved.

Today nearly every casemate has leakage problems. The actual dripping-type leakage is a symptom of too much moisture being allowed into the coquina vault masonry. This water infiltration, in conjunction with fluctuating temperature and stagnant air with excessive humidity, causes alternating drying/wetting cycles, which are detrimental to the masonry fabric: The coquina is weakened, the mortar is leached, and the plaster softens and spalls. In addition, the excessive moisture permits vegetation growth (especially ferns), which mechanically accelerates masonry deterioration. There are other sources of excessive moisture and associated fabric deterioration problems in the casemates


(see the discussion above); but, clearly, the major source of water is the leaking terreplein. This leakage is also a nuisance to visitors, a hazard to exhibitry, and disruptive of park staff functions (see figs. 74 and 75).

An even greater concern than leakage through the terreplein into the casemates is leakage into the bastion fill. Water entering the bastion fill contributes to the more serious problem of structural destabilization. When water enters the fill, which can be likened to a mass packed tightly in a box, it increases in weight and may expand and overload the walls, then it drys, contracts, and erodes out through the cracks and below the foundations, creating voids in the fill; then the destructive cycle repeats itself. This line of thinking was first reported by Lieutenant Black in his assessment of the bastion cracks in April 1886 (cited in the 1983 HSR).

Waterproofing the bastions' terreplein is, structurally speaking, very important. In fact, the bastions (especially the northwest and southwest bastions, where the deck is most deteriorated) exhibit the most obvious terreplein Laykold membrane deterioration. There are cracks in the deck, some with vegetation growing out of them, and areas where water puddles. In the past few years the park staff has attempted repeated repairs with a decreasing degree of success. Now they commonly attempt'repairs of previous repairs.


Leakage through the vaults is not restricted to one location (such as at the vault crown or vault spring); but, there is a predominance of leakage at the rear of the casemates, near the scarp, where the 1960 slab joins the 1940 firing steps. It is suspected that the 1960-64 terreplein system failed most rapidly and cdnsistently at its intersection with the portion of the 1939 work that was not removed in 1960 (see fig. 76). One can observe, too, that the copper flashing is not secure at many locations along both the scarp and courtyard wall parapets and that cracks are present at most scuppers. In Manucy's 1963 file report "Urgent Need to Waterproof the Terreplein," he recommended that, in order to control the excessive leakage, the 1939 work be removed entirely and a new system installed on the pre-1939 base. As a less expensive alternative, Manucy thought something like the Laykold system might work. The Laykold approach was selected, but it has not performed well.

In most areas the terreplein has a proper slope, and most of the rainwater is channeled away from the courtyard and the center of the bastions toward drainage ports and scuppers below the parapet at the scarps. However, cracks have developed between the terreplein and the drainage ports, and several of the 29 existing concrete scuppers (installed in 1890) are broken. Thus, water is allowed to pass directly into the top of the scarp walls.


d. Appearance and Safety

Does the terreplein have other problems? Yes (see figs. 76 and 77). Besides its function as a floor and roof the terreplein has a historic aesthetic and interpretive role to play. Here, there are currently major flaws. The appearance of the terreplein is that of a modern waterproofing membrane painted battleship gray. This is out of character and incongruous with the adjacent historic fabric, and it is inconceivable today that such a system was ever permitted on so significant a historic structure as the Castillo. The only good thing about this Laykold system is that in less than 20 years it has proven a failure and its replacement is necessary. (This is a rather short life for a Castillo terreplein compared, for example, to the deck installed in 1800-1802, which apparently lasted nearly 90 years.) In all fairness to the Laykold system and to the previous deck, which lasted only 21 years, it is quite likely that they may have served longer if it had not been for several post-construction deck modifications, such as the channeling of the deck for installation of electrical wiring.

The firing steps pose another problem area on the castillo terreplein. Besides the fact that they suffer the same aesthetic incongruity as the terreplein deck--being covered by. the Laykold membrane--several of them (installed in 1940) are mislocated or should not have been installed at all. Because the Eastillo parapets and embrasures were not simultaneously restored to the same period, the restoration of all the firing steps to the 1740s period now creates a situation that never existed historically. At two locations, firing steps were reconstructed in


Front of embrasures, mistakenly inviting visitors to step into the embrasures and creating a safety hazard. As a temporary measure, two embrasures have been boarded up with wooden barriers. Manucy rationalized (November 8, 1939) that a future restoration of the parapets and embrasures to the 1740s period would rectify this problem. However, current management policy inclines us away from such a restoration and toward a compromise firing-step arrangement that could accommodate the several periods that are represented today within the historic fabric and

-1 interpretive plan of the castillo.



Fig. 74. Laykold Membrane Detail, August 1984. The 1964 Laykold membrane is no longer a water-controlling system; it is full of cracks. Fig. 75. Terreplein Flashing Detail, August 1984. Mortar at the terreplein flashing raglet has fallen out. This type of failure is common enough to make the present roof function of the terreplein a failure.

~C 7

K j.

--- '- "t,.n


Fig. 76. Laykold System at Firing Step, August 1984. The cracking Laykold membrane seen here follows the line of intersection between the 1939 paving left in place and the 1960 concrete slab. This failure is virtually universal.

Fig. 77. Firing-Step Safety Hazard, August 1984. The park-installed wooden infill at the embrasure was necessary because the firing step built at this location in 1940 mistakenly invited visitors to step into the embrasure, some 30 feet above the moat. Note also the Laykold membrane failure.



r i~

~ .,tP ;iCF

:-4' "

1 i









e. Terreplein Summary

Collectively the deficiencies present in the terreplein easily justify a recommendation for corrective action. The extent of the treatment intervention is the only question left to be resolved. Table 2 summarizes the nature of the situation.

No costs or immediate impacts would be associated with a decision to not take the corrective actions noted in the summary chart. However, over the long term, damage to the fort would become more obvious and irreversible. To attempt only a partial treatment--to reduce the water management deficiencies by replacing the Laykold system with a similar new system--would have a moderate price tag and only a minimal physical impact on the resource. However, such a solution would be short lived, would require replacement every 15 to 20 years, and would not correct the other deficiencies discussed. Replacement of the entire deck, flashing, scuppers, and firing steps is the logical action plan. This approach would entail installation of a moisture membrane at the lowest possible level without excessive impact on the historic tabby layer, followed by installation of a historically appropriate (tabby duplicate) tread surface correcting the appearance, elevation, and firing-step deficiencies.


Table 2: Summary of Terreplein Deficiencies and Corrective Actions

Area of Concern Deficiencies Corrective Actions

1. Design elevation Present deck surface is Without adversely impacting the higher than historic single remaining historic floor surface. layer (C), install a new deck at a lower elevation, improving
functional and historical

2. Rainwater Rain leaks through the Replace the present membrane,
management present deck, permitting flashing, and scuppers to excessive moisture in the ensure proper drainage bastion fill, accelerating gradients and flow of water casemate fabric deteriora- out of the fort; install the tion, and creating problems new membrane on a sound for park operations. base.

Specifically, the junction
between the remaining
portions of the 1939 deck
and the present (1960-64)
system has failed, flashing
and scuppers have become
disconnected, the Laykold
membrane is cracked, and
repeated repairs are no
longer effective.

3. Aesthetics Present appearance of the Install a deck composed of deck, flashing, and firing materials that duplicate the steps (color, texture, and color and texture of historic materials) is incongruous materials, are compatible with adjacent to the historic the adjacent features, and fabric. provide a sound wear surface.

4. Firing Steps Locations of several steps Replace firing steps as part of create a situation that never terreplein replacement, installexisted historically and that ing them above the new moisture poses a safety hazard for barrier so that they can be visitors. Also the steps exposed masonry construction; are aesthetically inappro- do not reinstall a number of priate (see above). the steps for both historical accuracy and safety reasons.


3. Foundation and Structural Condition

This section of the report addresses the integrity of the foundation and superstructure--walls, scarps, and arches--of the Castillo. The engineering design characteristics are followed by a discussion of each group of problems: scarp cracks, vault cracks, and covered way wall cracks. It should be pointed out that these problems have been manifest at the Castillo for many years and that extensive archival recordation was consulted as part of the current assessment. However, the analysis presented here does not rely entirely upon the historical records and observations, but also upon a recent crack-monitoring program conducted by DSC and the park maintenance staff (see appendix K).

a. Design Characteristics

The engineering design concepts used in the Castillo may be categorized as relying on (1) the principles of mass construction to support both vertical and horizontal loads (see fig. 78), and (2) the principles of the arch and vault as vertical load-carrying structures (see fig. 79). These are sound principles and well executed in the case of the Castillo except for two corollary provisions or potential agents of weakness: (1) that stone masonry laid up in courses with lime mortar as used in the Castillo has poor horizontal load resistance and (2) that the foundation substrate for mass-type vertical load bearing is crucial to the
----'--------------success of this design approach and is in fact of questionable reliability. The following figures illustrate the design principles and inherent weaknesses of the approach and can be used to predict types of


deformation or structural failure. By comparing these graphic models with- observations, it is possible to answer the question: Are the observed--and in some cases measured--deformations in structural fabric a matter of preservation concern?


Fig. .78. Principle of Mass Construction, Illustrating Methods of Failure.

(1) Overturning: Lateral loads overcome dead weight of wall, causing wall to rotate. (2) Sliding: Lateral load is greater than the vertical load times the coefficient of friction, causing horizontal shear cracks. (3) Settlement: A combination of vertical and lateral loads creates soil-bearing pressures greater than allowable, causing differential settlement and vertical shear cracking. (4) Bulging: This problem is caused either by a large vertical load combined with a large lateral load, or by a lateral load exceeding the mortar joint coefficient of friction (structural engineer Jana Chalk).

(I) Oueru-rn 1

Fig. 79. Principle of Arch/Vault Construction, Illustrating Methods of Failure: (1) Sliding: Horizontal thrust developed by arch at the spring line exceeds the allowable shearing stress of the abutment. (2) Settlement of the abutment: Axial loads exceed the allowable soil-bearing pressures. (3) Rotation of the abutment: Horizontal thrust developed by the arch overturns or rotates the abutment. (4) Crushing: Compressive stresses in the arch exceed the allowable compressive stress of the stone; fm (max) generally occurs at the skewback. Previous failures usually

precede this one.

Sliding, settlement, or rotation of abutments would induce critical stress, such as tensil stress, for which the arch/vault is not designed. This would cause cracks to develop in the intrados of the arch. Crushing would generally occur first at the skewback, and joints would possibly

open in the extrados of the arch (structural engineer Jana Chalk).

I e e Ix tr

ii -.i

ii 111i~j

The following list summarizes the predictable structural deformations and compares them with corresponding observations:

(1) Scarp walls of curtains: Bulges? Vertical shear cracks?

Slippage? Overturning? Not present. Differential settlement?

Yes, slight; but it does not appear to be an active problem.

(2) Scarp walls of bastions: Bulges? Yes. Vertical shear cracks?

Yes. Slippage? Yes, slight. Overturning? Possibly yes.

Differential settlement? Perhaps long ago, but not at present.

The bastions pose a mix of structural concerns.

(3) Counterscarps: Bulges? Vertical shear cracks? Slippage?

Overturning? Not present. Structural overloading does not

appear to be a problem.

(4) Covered way wall: Bulges? Yes. Vertical shear cracks? Yes.

Slippage? Yes. Overturning? Yes. Structural failures are


(5) Casemate vaults: Differential settlement? Perhaps long ago,

but not at present. Cracks? Yes, some cracks are present.

Sliding? Rotation of abutment? Crushing? No.

(6) Sea wall of battery: Bulges? Vertical shear cracks? Slippage?

Overturning? Not present. Structural overloading does not

appear to be a problem.


From this summary it is reasonable to eliminate the fort curtain walls, counterscarp walls, and seawall of the battery from the category of structural concern; they appear to be stable. However, it is important to look more closely at the bastion scarps of the fort, the covered way walls, and the vaults.

b. Scarp Cracks, General

More than 300 years ago the castillo scarp walls were laid directly on the sand in a 5-foot-deep foundation trench with no subfoundation cribbing, grillage, or pilings employed. The enormous spread and square foot distribution of the scarp load--the base being more than 16 feet wide (compared to a height of only about 25 feet)--was largely successful in transferring the vertical mass of the wall load to the substrate without exceeding the bearing capacity of the soil. As a rule, if differential settlement and subsequent cracking are going to occur, they appear shortly after construction.

This apparently did not happen at the castillo with the one possible exception of the bastion scarps at the San Pablo (northwest) bastion, where the engineering inspector observed shoddy construction in his report of 1680. Not only had massive voids been left in the masonry work, but according to the report, the foundation had not been constructed properly. It was reported that these deficiencies had been corrected by 1682; however, one wonders if the initial construction of this bastion (between 1676 and 1682) involved soil-bearing or masonry problems that were never completely or properly corrected. By 1802 (if