4.1 Overview

Submerged metals are continuously affected by the chemical processes of corrosion, the destructive degradation of metal by chemical or electrochemical reactions within the marine environment (Valenca et al., 2022:2–3; Venugopal, 1994:35). Over time, metal ions at anodic sites defuse into electrolytic solutions from the oxidising reactions occurring at cathodic sites, causing the creation of corrosion byproducts, like rust on iron, and the loss of structural mass. The different reduction reactions in the microstructures of alloys and the imperfections found within refined materials, like carbon slag in steel, are targeted by this process, essentially reverting the chemically unstable materials back to their more stable original forms (Moore III, 2015:192; MacLeod, 2016a:90–92). The deterioration of metallic shipwreck hulls has become a growing concern within the field of marine conservation as many of the fuel tankers deliberately targeted in WWII threaten to release trapped fuel and chemical cargoes after nearly eight decades of exposure to a range of corrosive environments (Barrett, 2011:4–5). With the deterioration rate of ship hulls averaging at around ±0.1–0.4 mm of loss per year and the thickness of ship deck plates from the 1940s to the 1960s generally ranging from 1–4 cm in thickness, the window to act on the majority of potentially polluting shipwrecks (PPW) before a catastrophic breach occurs is closing (MacLeod, 2016b:8; Beldowski, 2018:249; Masetti, 2012:33; Masetti & Calder, 2014:139).

While laboratory and field methods of corrosion analysis have become well-established and can provide researchers with malleable models for understanding steel corrosion, a definitive predictive system for the manner of and time to failure of shipwreck hulls remains challenging (Russell et al., 2004:37). Wreck sites are open systems, with constant exchanges of materials and energy creating a state of dynamic negative disequilibrium unique to every environment and situation, but which will all ultimately result in the complete disintegration of the hull (Quinn, 2006:1420). While the full scope of factors affecting a shipwreck is too large to definitively quantify, a wide understanding of the host of corrosion factors encountered in situ, as well as an understanding of the increasing complexity created by structural degradation, is useful when attempting to decide a course of action for a specific wreck (Etkin et al., 2009:3–4; Russell et al., 2004:37). Inevitably, the metal hull of every PPW will corrode to a point of partial or complete failure, allowing oil or other pollutants to escape into the local environment. It is therefore advantageous to have as thorough an understanding as possible of the factors unique to each site and vessel that will accelerate or reduce corrosion rates. This will allow for the development of a timeline to failure for each ship, which is essential to remediating the wreck before it creates an environmental catastrophe.

4.2 Corrosion Types in Shipwrecks

Corrosion of shipwreck metals can be very loosely grouped into two categories: relatively uniform corrosion found across an exposed surface and more harmful, localised corrosive attacks on specific portions of the hull (Tait, 2012:864–865; Yongjun Tan, 2023:1). Uniform corrosion, otherwise known as general corrosion, is typically characterised by the gradual thinning of surfaces and protective corrosion products by the abrasive electrochemical and mechanical characteristics or contents of a surrounding electrolytic solution (Xia et al., 2021:2). A short period of rapid corrosion due to high rates of oxygen diffusion along the exposed surface will occur upon initial submersion which will then be followed by a slower, longer-term corrosion rate protected by the buildup of corrosion products, biofilms, sand particles, and static organisms, commonly called concretions (Melchers, 2003:272). Concretions form semi-permeable, anaerobic barriers between the bare metal and the seawater, creating a protective outer coating from oxygen diffusion corrosion while also containing heightened acidity and chloride ion concentrations, stabilising the hull at a steady deterioration rate (North, 1976:254–257). Uniform corrosion will not affect the structural integrity of the metal until a large portion of the cross-section has deteriorated, making it a reliable and stable way to measure, predict decay rates, and create corrosive models for steel-hulled vessels (Nürnberger et al., 2007:195; MacLeod & Viduka, 2011:135). Corrosion rates can drastically increase, however, if the concretion layer protecting the hull is stripped off or broken in any way, as the metal becomes exposed again to the marine environment and vulnerable to rapid, targeted attacks until recolonised (MacLeod et al., 2017:270–280).

Microbiologically influenced corrosion (MIC) is a form of corrosion that can develop underneath, or even help to build, a concretion. It can be difficult to predict the effect that a biofilm, a conglomeration of synergistic communities of microorganisms, will have on the surface that it is found on. On the one hand, biofilms can reduce corrosive potential by decreasing oxygen concentrations, halting diffusion by active transport, and producing corrosion inhibitors. On the other hand, however, biofilm communities have also been observed producing oxygen, sulfides, ammonia, and highly concentrated acids, using various structural metals as electron donors in metabolic processes, ennobling metals to produce galvanic couples, altering anions to a more aggressive, corrosive state, deactivating corrosion inhibitors, and deriving energy by oxidising metals, all processes that accelerate the corrosion of metals (Little & Lee, 2009:2–21). Ultimately, whether a microbial community is going to protect or target a hull is going to depend on the type of metal that the MIC is found on, the biofilm community itself, and the surrounding electrolytic solution in which the vessel is submerged, with some seasonal variation seen in certain locations (Little & Lee, 2009:2–3).

Localised corrosion can be initiated in the natural environment in many ways, as described by Galvele (1983:2). Corrosion from the metal’s electrochemical reaction with the environment is thought to be initiated by the chemical breakdown or mechanical disruption of the protective oxide film of a metal or alloy. Aggressive anions can become trapped in this small break between the film and the surface of the metal, with corrosion rates reaching several thousand to millions of times higher than in the surrounding passive metal, causing serious sectional and penetrative damage to the material structure (Kruger & Rhyne, 1982:206–207; Vargel, 2020:164–166; Melchers, 1999:6). Pitting corrosion, one of the most common and harmful types of localised chemical attack, is confined to these small breaks, on the order of square millimeters or less, and is covered with corrosion products that restrict ionic species flow in or out of the pit, drastically increasing the corrosion rate (Kruger & Rhyne, 1982:206–207; Galvele, 1983:1–15). Crevice corrosion is a similar chemical attack that occurs under corrosion product deposits and is found in component crevices, like structural support couplings or joints (Kruger & Rhyne, 1982:206–207; Makhlouf, 2015:541–543). Cavitation corrosion is a physical corrosive attack initiated by the environment and is caused by the collapse of gas bubbles on the surface of the metal. The subsequent increased velocity of the electrolyte fluid creates miniature shock waves on the surface, thus creating breaks in the protective oxide film and initiating pitting corrosion (Makhlouf et al., 2018:111).

Several forms of localised corrosion occur within the metal itself and can be caused by dynamic pressures or material differences in the components or alloys. Galvanic coupling, also referred to as bimetallic or proximity corrosion, is a static reaction that occurs when metals with different corrosive potentials touch each other or are connected by the same concretion in an electrolytic solution (MacLeod, 2019:871; North, 1984:133–134). While the metal with the higher electro-reactivity potential (E corr.) in the coupling will begin to erode more quickly, the lower potential metal will be protected and begin to corrode more slowly (North, 1984:133–134). This reaction between materials has been used as a form of temporary protection for fragile artifacts with the use of sacrificial anodes, as demonstrated by corrosion scientist Ian MacLeod (2016b:9), in collaboration with North (McCarthy, 2000:86–88) and Steyne (MacLeod & Steyne, 2011:347–349; Steyne & MacLeod, 2011:67). Intergranular corrosion is another static, localised attack at and adjacent to grain boundaries between the microstructures of a metal alloy, caused by chemical differences in the grain, impurities at the grain boundaries, and reduction or enrichment of an alloying element in the grain boundary area. This form of corrosion is only prevalent in alloys and can severely impact the material’s structural strength or even cause it to disintegrate (Karlsdottir, 2022:259–260).

Stress corrosion cracking, a dynamic attack, can be particularly damaging to a crumbling shipwreck that is experiencing changing environmental pressures. It is characterised by internal, residual stresses, created when the metal is formed, welded, or processed, and external, environmental stresses, applied to a metal component by the environment. This form of corrosion can result in intergranular or transgranular cracking of the source material (Makhlouf et al., 2018:114). Corrosion fatigue, a similar dynamic attack, is caused by repeated, cyclical exposures to a corrosive environment and external stresses. Eventual catastrophic structural collapse can potentially occur at any time, even when those stresses are not presently in action (Komai, 2003:345).

4.3 OCP and General Corrosion Factors

The kinetic driving force behind shipwreck corrosion can be described as the Open Circuit Potential (OCP), or corrosion potential. This value compares the equilibrium potential between a metallic wreck, referred to as the working electrode, and its electrolyte or environment to a reference electrode in the same environment, providing a measurable rate of real corrosion behavior (Siddaiah et al., 2021:7–9; Rasol et al., 2015:294). An increase in OCP results in a depolarisation of the cathode and an increase in corrosion, whereas a decrease in OCP will exhibit the opposite effect (Rasol et al., 2015:294). Well-documented are the physical, chemical, and biological corrosion factors that positively increase corrosion potential and threaten the structural stability of shipwreck hulls (Kuroda et al., 2008:3–6; MacLeod & Viduka, 2011:136; Liddell & Skelhorn, 2019:83; Gilbert et al., 2003:178–179). For example, the type of metal used in the reaction, a mild steel in the case of most PPWs, and its associated microscopic structure, along with the microorganisms forming biofilms on the surface of the material, can influence OCP (Eyres & Bruce, 2012:45–49; Rasol et al., 2015:294–297). Differing areas of the ship may also have unique open circuit potentials due to dissimilar microstructures in the materials or passive film layers, contributing to non-uniform corrosion across the vessel (Mischler & Munoz, 2018:508). Areas that have been exposed to physical stresses, including physical damage or heat, may change enough in material structure to positively shift the OCP of that component (Moreto et al., 2018:2–5). Hull-rivet interactions, welded seams, or damage due to impact or fire may result in accelerated corrosion of an area or interface relative to an intact, undamaged section of the hull, resulting in a higher risk of corrosion (Mischler & Munoz, 2018:508; Rasol et al., 2015:294–297; Chaves et al., 2022:195). Every shipwreck, however, is being acted upon by a unique set of environmental characteristics that should be properly understood before attempting to create a roadmap to failure for the ship.

4.4 Acute Environmental Factors

In the face of destructive natural events, such as catastrophic storms and sudden coastal changes, metal shipwrecks can suffer damages complicated by their age and preexisting structural degradation. Volcanic tremors affecting shipwrecks, while rarely studied, have been shown to affect site structural integrity by shifting vessels deposited on sloped bedforms. This can produce wear through mechanical damage and abrasion or by exposing shallow water wrecks to corrosion factors related to coastal erosion (North & MacLeod, 1987:74–75; Ridwan, 2019:1624). Events such as earthquakes and subsequent tsunamis can have severe impacts on site dynamics as well, drastically altering the surrounding landscape and exposing hulks not just to more dynamic surface currents and tidal action, but also to more anthropogenic activity in the form of plundering, fishing, vandalism, or accumulated rubbish and waste (Ridwan, 2019:1624–1625). Estuarine and riverine wrecks are prone to experience acute damage during floods and droughts as higher rates of corrosion impact abandoned watercraft found at air-water interfaces (North & MacLeod, 1987:75).

Large tropical storms, cyclones, hurricanes, and typhoons can cause the physical degradation of a wreck by dislodging loose pieces of the vessel from the whole, reinforcing the effects of tidal currents to enact scour damage, altering the composition of the water column, subjecting the structure to abnormally powerful wave action, washing away supportive surrounding sediments, damaging or stripping away concretions, or even causing the collapse of the hull (MacLeod & Viduka, 2011:135; North & MacLeod, 1987:75; Steyne & MacLeod, 2011:68; MacLeod et al., 2017:270). Sudden shifts in the surrounding environment, as well as collapses inside of the wreck due to large weather events, have been shown to create structural damage that can expose new sections of the ship to corrosion, altering existing corrosion patterns. This was seen with damage to the wreck of USS Mississinewa caused by a large summer storm in 2001, which subsequently initiated a series of oil leaks that grew in frequency and severity until the wreck had to be relieved of its oil cargo several years later (U.S. Navy, 2004:1-1—1-2). Structural collapses can also cause changes in fluid dynamics around the wreck, which will negatively affect areas of the ship that were previously spared these velocities (MacLeod et al., 2017:273).

4.5 Cyclical Environmental Factors

Where they exist, regular fluctuations in water column characteristics must be considered in models of corrosive behavior. Cycles in the biochemical characteristics of a shipwreck site will regularly occur throughout the year, as salinity, dissolved oxygen, and microbial activity in the water column will experience high and low points due to seasonal changes and the minor associated shifts in ambient water temperature, velocity, and microbial content (Li et al., 2019:6056–6060; Olson et al., 2022:1–2; Mestre et al., 2020:1–2; Liao et al., 2022:4–17). More dramatic site-specific cyclical characteristics, like the variable seasonal temperature changes witnessed at most inshore coastal sites, can profoundly influence corrosion potential (Zintzen et al., 2008:330). Strong water velocities, caused by particularly aggressive tidal action and recurrent storm patterns, can accelerate corrosion or destabilise the vessel, creating leaks. The SS Jacob Luckenbach wreck oiled birds and beaches along the San Francisco Bay area for decades because seasonal current variations and regular winter storms disturbed the wreckage, releasing fuel into the water column (Moffatt, 2004:65; Duerr et al., 2016:1). Current flow which is not uniform across the structure may result in variable corrosion rates, reducing the value of predictions made using a uniform corrosion model. The Cerberus, for example, offshore Melbourne, Australia, has experienced more deterioration on its starboard side from exposure to the open ocean and the associated higher fluxes of oxygen than on its port side, which faces into Port Philip Bay (MacLeod & Steyne, 2011:341).

Metal-hulled ships permanently or cyclically exposed to the atmosphere by wave action in the tidal zone, like the Civil War era Sub Marine Explorer found in the Bay of Panama, will not only experience the physical stresses caused by the crashing of the water, but will additionally undergo periodic wetting and drying, increased oxygen availability at the air-water interface, physical destruction of protective concretions and corrosion products, impacts from dissolved or transported debris, larger variations in temperature, expedited cycles of erosion, and concentrations of aggressive salts, which can all contribute to an advanced decay rate (North & MacLeod, 1987:75; Evans et al., 2009:46; Johnson et al., 2010:58–59). On the other hand, cyclical exposure to an anaerobic environment has been observed contributing to accelerated MIC for certain metals. When seasonal erosion patterns expose and re-cover a shipwreck, corrosive bands are formed, like those seen on the copper wires found in the SS Xantho, an abandoned steamship found off the coast of Western Australia (MacLeod, 2002:702; McCarthy, 2000:91–92). Similarly, corrosion rates of wrecks in areas of high seasonal runoff or seasonal hypoxic conditions due to the migration of pollutants or high microbial or marine growth activity can also be expected to vary (North & MacLeod, 1987:75).

4.6 Long-Term Environmental Factors

The natural characteristics of the wrecking location can often be a strong determinant in how long a shipwreck will be preserved. Vessels that are deposited in a high-energy environment will experience more rapid movement of currents and tidal actions across the metal surface, causing the hull to be impacted by a greater flux of dissolved oxygen and preventing the formation of protective corrosive films, causing non-uniform corrosive potential to increase (North & MacLeod, 1987:75; MacLeod et al., 2017:280). In contrast, if seawater is still or stagnant for extended periods, as is often encountered in deep bays, atolls with narrow, shallow entrances, or at very deep-water wrecks, corrosion rates will rapidly decrease (North & MacLeod, 1987:75). Alternatively, deep-water wrecks are prone to developing ‘rusticles’, or icicle-shaped iron oxide accumulations, that can be found throughout the vessel and seemingly mobilise structural mass to different areas of the ship. Famously having been found in the Titanic, the corrosion phenomena, which appears to develop more extensively on Atlantic deep-water wrecks than Pacific, has been theorised to be the result of MIC-related biodeterioration, dissolved iron accumulation deposited on the wreck, or natural corrosion processes interacting with intense hydrostatic pressures (Salazar & Little, 2017:26–30; Cullimore et al., 2002:117–120; Silva-Bedoya et al., 2021:10–23). Seabed composition can also be important for the long-term preservation of a shipwreck. Frequent current and storm action can create bedforms predominantly composed of rocky, gravelly sand and shells, which can be highly abrasive (Wheeler, 2002:1151). Bedforms consisting of larger-sized sand grains and gravel can still be affected by strong bottom currents which can carry sediment to wash the hull, effectively sandblasting it, or covering it and causing it to collapse under the excess weight (Hac, 2018:175; North & MacLeod, 1987:75). Meanwhile, sediments washed away from under a structurally significant portion of the vessel can create additional weights and strains to become apparent on the metal as it is forced to stay rigid, eventually causing that section to buckle and break away (Hac, 2018:175).

Coastlines that are regularly indented, have high bathymetric relief, and numerous estuaries, as seen along Ireland’s southern, western, and northern coasts, will provide stable micro-environments that can protect shipwrecks from rough wave action and mechanical corrosion (Wheeler, 2002:1151). Protective atolls and shelters, including Ulithi Atoll, final resting place of USS Mississinewa, and Wardang Island, South Australia, the resting place of Songvaar, can provide wave-breaks that will preserve the vessel from storm action, current velocity, or strong tidal surge (Moore et al., 2014:21–22; MacLeod, 2002:703). Similar complications can be found on the seabed depending on the topography of the area. A vessel or artifact deposited in a narrow reef gully may experience a higher water velocity and thus a greater corrosion potential, while a wreck on the leeside of a reef may be protected from these same corrosion pressures (MacLeod, 2018:65–68). Alternatively, a crevice or depression may protect a vessel or artifact from high surges and extensive wave action that would normally cause artifacts to physically shift or roll along the seabed, collecting mechanical damage and abrasion (North & MacLeod, 1987:75).

4.7 Anthropogenic Factors

There are a range of anthropogenic activities that can impact the corrosive potential of a metallic shipwreck, with a wide scope of agents responsible for the damages. Individual divers can cause damage by penetrating a wreck and trapping vented oxygen in ship crevices or by grabbing the structure for stability (Ridwan et al., 2014:8–9). Dive companies in the past have been responsible for damaging sites by dropping anchors onto them or by tying off their boats to vessel superstructures while mooring (Viduka, 2011:14; MacLeod & Steyne, 2011:340; Henderson, 2019:8). Dredging and trawling activities can have significant impacts on the seabed and the cultural heritage that is found there (Evans et al., 2009:46). Brennan and colleagues have noted the damages that Coimbra suffered from multiple dredge impacts, eventually leading to the formation of a crack in the hull, while Delgado and colleagues discovered multiple sets of trawl gear trapped under the hull of the Coast Trader (Brennan et al., 2023:328; Delgado et al., 2018:27). Fishers and anglers who forcefully retrieve gear stuck on a wreck may damage portions of the vessel in their struggle, while abandoned net hangs, called ‘ghost nets’, can increase dynamic stress on the structure by inducing drag, destabilising more fragile portions of the ship (Firth, 2018:14; Ridwan, 2019:1627). Modern explosion damage is surprisingly common on wrecks in the Pacific, as fishermen will use ordinance to stun and kill fish found around the bountiful artificial reefs while metal scavengers will use explosives to loosen sheet metal for sale (Naughton, 1985:16–17; MacLeod et al., 2017:270; Browne, 2019:2). Larger scale metal scavenging operations have become far more sophisticated, lifting entire wrecks from the seafloor using cranes installed on large ships (Ridwan, 2019:1625). Souvenir hunters can damage a site while pilfering for treasures, robbing the asset of valuable archaeological evidence, while significant amounts of marine litter can ruin the site’s aesthetics and potentially cause physical or galvanic damage to artifacts (Viduka, 2011:14; Ridwan, 2019:1625).

Largescale coastal developments and the associated transformations in the local environment can increase risks to submerged heritage assets as well (Ridwan et al., 2014:8–9; Evans et al., 2009:45–46). Some wrecks that pose threats to modern watercraft or shipping, like the Cleveco, have been damaged by passing ship traffic, causing oil spills, and have suffered shifting structural pressures while being lifted and moved, causing even more spills (Davin & Witte, 1997:783–786). Coastal expansion projects and offshore energy exploration and infrastructure development have led to the rediscovery of many vessels but could also result in damage to the wrecks themselves, to the surrounding environment, or negative attention from curious divers (Ridwan et al., 2014:8–9; Church et al., 2009:51; Evans et al., 2009:45–46; Moore, 2021:157). Oil spills that come into contact with a metallic shipwreck can have significant disruptive effects on the diversity of microorganisms and corrosion product communities that are protecting the metal surface as well as on the porosity of the surrounding sediments, increasing corrosive potential (Salerno et al., 2018:4–12; Hamdan et al., 2018:3–12; Zhang et al., 2022:2–9). Trade routes and shipping lanes, while instrumental in the deposition of many cultural heritage sites, are also significant sources of heavy metals, oil, and human waste pollution in the modern era, leading to the further degradation of these historic vessels on the seabed (Lawrence, 2008:6–14; Chan et al., 2001:581–582). Poor solid and liquid waste management from urban and agricultural landscapes and rainy season runoff near these locations can create significantly elevated nutrient concentrations around shipwreck sites, potentially expediting corrosion rates by up to 50% (Jiminez et al., 2017:3; Melchers, 2014:110–115). Finally, as ocean temperatures and acidity levels rise due to the effects of climate change, the associated physical and chemical changes seen in the environment will cause further damage to shipwrecks, as is already being seen with the expedited chemical breakdown and growth of rusticles found on the Titanic (Wright, 2016:260–263; Mann, 2012:44–49).

4.8 Ship Construction, Condition, Impact, and Orientation

Hull joinery technique becomes an important factor in the long-term stability of a wreck and its ability to successfully hold a potentially polluting cargo. Hull plate rivets often see differing corrosion potentials between head, shaft, base metal, and hull, causing them to corrode preferentially to the surrounding metals and release seeps of oil or allow hull plates to separate from the body of the ship (Johnson et al., 2011:7; MacLeod & Steyne, 2011:340; Brennan et al., 2023:327). Welded hulls will typically withhold petroleum pollutants more effectively over a longer period but have been observed bursting along the welded seam during particularly violent impacts with the seabed, releasing a cloud of cargo product (Kery & Stauffer, 2015:7). Vessels sunk in shallow waters will have little time in the water column to gather momentum or shift orientation before eventually depositing in the sediment and thus will have a less energetic impact when landing, as compared to deep-water wrecks that will have a more variable deposition orientation and higher velocity upon impact (Kery & Stauffer, 2015:1–2; Liddell & Skelhorn, 2019:83). High-velocity impacts upon the seabed have been shown to cause deformations, including folds and staving, in various portions of the hull, and can burst sealed doors and hatches, exposing more internal spaces to corrosive processes, accelerating dissolved oxygen transport and increasing opportunities for MIC (Morcillo et al., 2004:122–123).

The vessel's working life and sinking event will also significantly affect its structural integrity. Military and tanker vessels will have accrued residual damage throughout their working lifetimes that will influence OCP and are expected to have suffered extensive structural damage during their sinking events (Gilbert et al., 2003:178–179; McKay, 2005:129). Portions of the vessel that are missing paints or other protective coatings due to blows and scratches, or were intentionally left unpainted, will be targeted first by corrosive processes (Morcillo et al., 2004:125). Vessels that have been broken into multiple pieces will experience a higher percentage chance of inverting or landing on their sides on the seabed, exposing the vessel to a more diverse range of structural pressures (Brennan et al., 2023:325–334; Morcillo et al., 2004:122–123; Russell et al., 2004:36–42). Corrosion cells can also develop when a vessel has been split into multiple pieces, in which portions of the hull will act as a protective anode for other pieces of the ship with lower electro-reactivity potentials (Viduka, 2011:16). Unnatural weight distributions or new stresses created by the deterioration of the ship in directions or areas not considered in the original design will cause the corrosion potential of different sections of the ship to increase (Gilbert et al., 2003:178–179; MacLeod, 2016b:8). A vessel that settles upright on the seabed will experience a generally stable rate of decay due to pressures generally in-keeping with design limitations, and, if sunk with a fuel cargo, may slowly release it from fragile vents and risers, eventually allowing seawater to enter the emptied cavities (Brennan et al., Chap. 9, this volume). A vessel that is inverted or on its side will experience strains that it was not designed to handle, depending on both the topography of the seabed and the strength of currents impacting the broadsides and bottoms of the hull, thus experiencing an expedited corrosion rate (MacLeod et al., 2017:280; Kery & Stauffer, 2015:5). Finally, cargo type, viscosity, and amount can play a part in the protection of the fuel tanks withholding a petroleum or chemical cargo. Thicker, more viscous oils, like a bunker fuel, can create a protective coating along the insides of the tank, delaying corrosion, as observed on the Coimbra wreck (Brennan et al., 2023:327). Lighter oils and ballast water can create humid, corrosive environments that will be detrimental to the integrity of the tank (Zayed et al., 2018:300).

4.9 A History of Shipwreck Corrosion Analysis

The compilation and breakdown of the previously described site-specific characteristics into data points for use in a single universal shipwreck corrosion model has been a lengthy and challenging endeavor. The first forays into quantifying the complex processes occurring at shipwreck sites began in the 1970s and 1980s with in-situ corrosion analysis led by marine archaeologists North and Pearson while they were studying the wreck of Batavia. They attempted to collect and analyze artifact concretions through the lens of corrosion science, allowing them to develop theories on iron artifact preservation and graphitisation (North et al., 1976:192–193; North & Pearson, 1978:180–182). McCarthy would build on this momentum in the 1980s with his survey of the Xantho, a steamship that sank off the coast of Western Australia, by combining data extracted from the marine environment with data gathered from eyewitness accounts and historical documents (McCarthy, 2000:7–61; Moore III, 2015:194–196). This study was essential to understanding the rate and manner of decay for an iron shipwreck, including non-uniform rates of decay, and the difficulties encountered when attempting to excavate, protect, and recover artifacts (McCarthy, 2000:88–177). The Xantho became the first vessel to be protected by sacrificial anodes as well, a method later perfected and used extensively by Australian corrosion scientist Ian MacLeod (McCarthy, 2000:84–88; MacLeod, 1987:50–55, 1989:7–13, 2019:877; Steyne & MacLeod, 2011:67).

MacLeod would become crucial in the interpretation of chemical and physical modes of shipwreck and metallic artifact degradation using on-site measurements. These measurements would ultimately aid him in his development of corrosion rate linear correlation graphs, used to more easily track and display rates of decay, including for specific environmental factors (Moore III, 2015:196–197; MacLeod, 1989: 7–13, 1995:54–58). MacLeod applied many of these techniques to his study of multiple WWII steel vessels found in Chuuk Lagoon, highlighting the importance of depth, vessel position, breakage pattern, and topography when determining corrosion potential (MacLeod et al., 2011:1–10, 2017:270–281). The 1987 assessment of the USS Monitor, only possible via ROV due to the submarine’s resting depth, produced more data on the electrochemical behavior of a complete shipwreck, examining the galvanic interaction between vessel components as well as the electrical continuity between adjoining sections (Moore III, 2015:197–198; Arnold III et al., 1992:52–55). The multidisciplinary nature of the Monitor study, which combined the efforts of the National Oceanic and Atmospheric Administration (NOAA) and the US Navy, would foretell the continuing combination of research interests that would be necessary to gain a more complete understanding of the complex corrosion processes at play on shipwreck hulls. The development of new probabilistic models and equations that investigate specific environmental influences has given scientists and archaeologists many broad, quantitative tools to use, albeit with the false assumption that the shipwreck is corroding at a uniform rate (Moore III, 2015:199–201; Woloszyk & Garbatov, 2022:2–12; Guedes Soares et al., 2011:529–537; Melchers, 2005:2391–2404). These complex, multifaceted webs of study, however, have since been combined to advance understanding of the complex corrosion environment of a single shipwreck: USS Arizona.

4.10 Case Study: USS Arizona

The USS Arizona is an American battleship that was sunk by the Imperial Japanese Navy on December 7, 1941, in a surprise attack on Pearl Harbor, Oahu, Hawaii (Wilson et al., 2007:14). Due to its historical significance, status as a war grave for over 1170 military personnel, and the approximately 2500 tons of fuel oil trapped within, the vessel has received significant attention in the form of legal protections, surveys, and structural conservation efforts that have allowed it to remain mostly intact since its initial deposition (Johnson et al., 2018:747). With permissions received from the National Park Service and USS Arizona Memorial, the ship has become the first long-term research resource for the study of steel hulled ship corrosion in the marine environment (Johnson et al., 2018:747–750; Wilson et al., 2007:14–18; Russell et al., 2004:37–43; Foecke et al., 2010:1091–1100; Russell et al., 2006:311–317; Murphy, 1987:10–15; Johnson et al., 2011:1–8). Since 1999, the USS Arizona Preservation Project has brought together experts from a variety of disciplines to understand the complex internal and external corrosion and deterioration processes affecting Arizona and how they have impacted the structural integrity of the vessel (Wilson et al., 2007:15; Russell et al., 2004:35). The goal of this interdisciplinary, cumulative, and years-long project has been to model and predict the nature and rate of structural changes up to imminent collapse, to be used by future site managers to minimise environmental hazard from fuel release and inform decisions for long-term preservation (Wilson et al., 2007:15; Russell et al., 2004:35, 2006:310).

Understanding the corrosion rate of the Arizona wreck has involved breaking down the major factors affecting shipwreck decomposition into data that can be quantitatively measured, plotted, and modeled (Russell et al., 2004:37). Finite Element Analysis (FEA) modelling has become the principal receptacle into which all experimental and observational corrosion data is fed to create a highly accurate and realistic calculation of the stresses and changes seen in the structure under load (Russell et al., 2004:37). The ship model, divided into sections or ‘elements’, is mapped with the different corrosion behaviors and stresses that the real vessel experiences, allowing precise calculations and analyses of loads to visualise potential future outcomes for the hulk, up to and including collapse (Russell et al., 2004:37). The calculation of this model around a shipwreck that is constantly experiencing erosion and shifting weight redistributions is obviously very complex, and requires precise data based on direct measurements and observation (Russell et al., 2004:37). Corrosion analysis research on Arizona, therefore, has been painstakingly and regularly collected since 1983 to accurately plot the current corrosion trends of the vessel (Wilson et al., 2007:15; Murphy, 1987:10; Russell et al., 2004:36–38).

After initial non-invasive photographic and mapping surveys were completed in the early 1980s, archaeologists and conservationists not only had an inventory of what had survived the sinking event and how it was deposited in the sediment, but also the first models and drawings of the ship that researchers and the public could use as interpretive devices (Murphy, 1987:10). Sediment and water samples from within and around the vessel were taken to understand internal and external corrosive factors, visual inspections of the hull and galvanic activity studies were completed, and non-destructive hull measurements were taken, when possible, to find concretion and hull thicknesses. Minimally invasive techniques would eventually have to be used, however, for a more complete picture of the wreck’s decay processes to come together. This initially included studies of biofouling composition and thickness which were supported by the in-lab study of scraped concretion material, penetrative measurement of concretion thickness, the removal of concretions from metallic hull edges, and the installation of a series of cameras along the hull to observe biochemical processes (Murphy, 1987:11–15; Henderson, 1989:117–156).

Sample pieces of the hull found in earlier salvage efforts were tested in the late 90s to understand the hull’s materials (Johnson et al., 2019:1–6). Extractive techniques became necessary and sample coupons were removed from the vessel using a hydraulic-powered saw in 2002. Initially, the coupons were taken to subtract current steel thicknesses from the thicknesses illustrated on the original ship plans to develop an initial corrosion rate (Russell et al., 2004:38, 2006:312). The study of these coupons and the concretions removed in the process, however, allowed conservation scientists to develop several innovative techniques to be used in the study of concretion accumulation and corrosion potential on other shipwrecks. These have included the Concretion Equivalent Corrosion Rate (CECR) methodology, in which concretions are analyzed through x-ray diffractometry to correlate iron content in concretions with an average corrosion rate, and the Weins Number, a predictive formula that can be calculated when temperature, oxygen concentration, and concretion thickness are known (Johnson et al., 2006:55–57; Wilson et al., 2007:15–18; Johnson et al., 2011:1–7). Concretion analysis through x-ray diffractometry also assisted researchers in identifying corrosion product species and, with the environmental scanning electron microscopy of concretions, how they interact with hull metals, ultimately determining the Secant Rate of Corrosion (Wilson et al., 2007:15; Russell et al., 2006:312–317). This method of study calculated concretion growth and loss of structural mass through material transfer away from the ship hull using a referential marker left on the vessel by explosions that occurred during the sinking event. This concretion growth, along with early-stage mass loss, provides a base of comparison between Arizona and a linear corrosion rate previously obtained from the metal coupon samples (Johnson et al., 2018:747–751).

Along with these innovative analytical techniques, more traditional methods of marine archaeological study have been implemented in the analysis and preservation of Arizona. This has included interior corrosion analysis using ROVs, environmental monitoring of corrosion factors affecting the interior and exterior of the ship, structural monitoring through visual analysis and tracking, oil analysis to calculate type, breakdown, and amount of oil in the ship, study of microbiology populating the ship’s hull, geological analyses of the surrounding sediments and their effects on the vessel, and the development of Geographic Information System (GIS) maps of the site, to be used for both study and public engagement (Russell et al., 2004:39–44). The results of this project are twofold: first, the FEA model, which so far has been quite accurate to the real ship in its predictions of changing stresses and visible deterioration, has provided site managers with a rough idea of the gradual rate of decay, weakest points in the hull, and possible pattern of collapse of Arizona moving into the near future (Foecke et al., 2010:1096–1099). Second, the project has become a model for the management and preservation of other leaking iron or steel-hulled shipwrecks, with the research potential of this effort displayed in the new techniques and prediction methods that have been developed and can be used on future shipwreck studies (Russell et al., 2004:35; Wilson et al., 2007:15–17; Johnson et al., 2011:3–6).

4.11 Conclusion

Potentially polluting wrecks link advances in our understanding of corrosion processes to events of historic and modern importance. As a result, a collateral value of shipwreck corrosion study is that it generates awareness of this most often unseen, though omnipresent, natural process and the consequences that it can have in the current environment. As site-specific biochemical and physical factors are given more time to enact their influences on shipwreck hulls, pollution events from PPWs will become a more common occurrence. Communicating this problem with decision makers, more of the scientific community, and the public will allow us to attract more concerned parties, engage it with a more thorough understanding of the processes at play and, hopefully, develop a plan to deal with the results. The longer this engagement takes, however, the more likely that catastrophic collapses will occur, transforming this generally manageable problem into an expensive and potentially deadly environmental disaster.