Medieval Japan: Destruction

In a study from the western Pacific, paleotempestological evidence was used to independently verify historical accounts of two “Kamikaze” typhoons that were said to have divinely struck southern Japan in the thirteenth century CE and in so doing destroyed the fleets of the Kublai Khan, whose vast armada had set out to conquer Japan (Woodruff et al. 2014). Legends state that despite being significantly outnumbered, the Japanese defenses were saved by the fortuitous destruction of the Mongol fleets by intense typhoons occurring in November 1274 CE and August 1281 CE, respectively (Sasaki 2008). Noting that some historical accounts can be prone to exaggeration, Jonathan D. Woodruff and colleagues (2014) attempted to assess the presence and possibly the magnitude of these typhoons.

The researchers began by analyzing the most recent sediments in a coastal lake in the region where the ships were supposed to have landed and correlating distinctive sedimentary layers found there against known typhoon strikes that occurred during the last 100 years. These “modern analogues” helped establish a unique sedimentary signature for typhoons in the area. Then, the authors identified two prominent sand layers in cores collected in the lake that dated to the thirteenth century CE (Woodruff et al. 2014). These layers were very similar to the ones left by the twentieth-century typhoons and were also enriched in the element strontium, an indictor of marine sediment over-wash. These data strongly suggested that two major marine flooding events occurred around the time the historical records indicate the fleets were destroyed.

Woodruff and colleagues (2014) argue that it is impossible to unequivocally link the geochemical and sedimentological data to the typhoon events, in part due to the precision of radiocarbon dating, which typically produces ages within a range of a few decades. Still, they argue that the most likely explanation for the sand layers is the Kamikaze typhoons, due to their timing and a lack of evidence for any other natural event that could have produced a similar deposit (such as a tsunami). Their work then places the two typhoon strikes into a broader paleoclimatic context and shows that this period—the thirteenth century—was one in which more typhoons were steered toward southern Japan by changes in El Niño activity in the tropical Pacific Ocean.

The authors thus argue that the typhoons serve as an example of how past events of extreme weather associated with climatic change appear to have had significant geopolitical influence (Woodruff et al. 2014). In this particular case, the impacts were severe, and the response was one of destruction and dispersal of the Mongol fleets—likely due to a combination of high wind, rain, and wave activity. While their study represents an example of the effects of hurricane impacts, it illustrates how geological data can be integrated with the historical record to assess the magnitude of natural events that had significant impacts on humans.

The typhoon strikes in the Japanese historical example also illustrate the concept of vulnerability—the degree to which natural and/or human systems are susceptible to, and unable to cope with, adverse impacts of climate change (IPCC 2014:54). The scale of the destruction indicates that the Kublai Khan’s fleets must have been highly vulnerable and leads to questions concerning the quality of the ships and the degree to which disaster preparedness and mitigation was included in the invasion planning. For example, in both cases, the fleets were destroyed in late summer/autumn of 1274 and 1281 CE, respectively, which is the height of typhoon season in the western Pacific. To what extent was seasonality considered when the invasions were planned? Following the destruction of the first armada in 1274 CE, what was done to limit potential typhoon impacts during the second invasion in 1281 CE? Whatever the answers, both events are examples of human systems being overwhelmed by natural phenomena.

The geological evidence itself provides only the starting point for understanding the human response to typhoon activity. The paleotempestological data produced by Woodruff and colleagues (2014) validate and refine the historical accounts of two high-magnitude typhoons occurring around the time of the destructions of the Mongol fleets. But understanding the nature of the impacts themselves requires data from historical and archaeological sources. For example, underwater excavations led by Kenzo Hayashida of the Kyushu Okinawa Society for Underwater Archaeology have produced at least one of the Khan’s ships and a range of maritime-related material culture (Delgado 2003). While it is now believed that the size of the Khan’s armada was probably exaggerated (Delgado 2010), the archaeological evidence is consistent with historical reports of the incident and geological evidence of the storms themselves.

Finally, the typhoon strikes must qualify as natural disasters, due to infrastructure damage and lives lost (Burton et al. 1978). However, this also is an example of the idea of natural disasters being relative to the actors impacted by the event. From the point of view of the Japanese, the Kamikaze typhoons and their impacts were seen as “divine winds” that essentially saved their civilization, whereas these events could only have been viewed as disasters from the perspective of the Mongols. While most natural disasters are typically considered to have negative consequences for human systems, it suggests that the concept of natural disaster can be more nuanced and can occasionally produce positive outcomes as well. Examples include the creation of opportunities for upgrading infrastructure and the stimulation of economic growth (e.g., Hallegatte and Dumas 2009), as well as the political and social opportunities for urban renewal (see Pickett, chapter 4, this volume).

The Maya Realm: Destabilizing Force or Unlikely Benefit?

The Caribbean coast of Central America experiences frequent hurricanes, many of which have had devastating consequences for people in both coastal and inland areas. Hurricane Mitch, for example, struck the coast of Honduras in late October 1998 and resulted in more than 11,000 deaths and billions of dollars in damage, much of it due to heavy rainfall that caused severe mudslides in interior mountainous regions (Hellin et al. 1999). In addition, in August 2007, Hurricane Dean struck the Yucátan coast of Mexico as a category 5 storm. Despite causing over $1.6 billion in damage, fatalities from this hurricane were relatively low, with only forty-five deaths recorded in the Caribbean islands and fewer than two dozen in Mexico. This relatively low death toll was attributed in part to reliable forecasts of the hurricane’s track and effective warning and evacuation procedures (Franklin 2008).

The frequency of high-intensity hurricanes affecting Central America prompts the obvious question: what impacts did these storms have on pre-industrial people, in particular the ancient Maya, with their sophisticated cities and intricate political alliances? The answer, however, is not entirely clear, in part because of limited paleotempestological research done in the region (although see McCloskey and Keller 2009; Denommee et al. 2014) but also because the effects of hurricanes would have certainly varied given the region’s topographic heterogeneity and diverse landscapes and cultures.

Some of the earliest work examining this question argued that hurricanes may have been “trigger mechanisms” for configurations of settlement patterns, subsistence strategies, warfare, trade, and migrations and demographic stability (Konrad 1985). As a result, “The hurricane was an integral feature of the pre-Hispanic Maya cosmology and ecological paradigm” in the first millennium CE (Konrad 1996:99). Herman W. Konrad (116–120) also argued that the presence of hurricanes led to local adaptations, such as household architecture characterized by rounded roofs and walls—features that are common along the Yucátan coast—that would have better withstood the impact of hurricane-force winds (123). More recently, others have argued that hurricanes were a destabilizing force for the ancient Maya and may have precipitated episodes of warfare (Dunning and Houston 2011:64). While some evidence—in the form of epigraphic, iconographic, and geoarchaeological data—exists to support this view (Dunning and Houston 2011), links between hurricane impacts and the ancient Maya are still mostly tentative.

One notable example of the effects of hurricanes on ancient Mayan civilization involves their possible role in the Terminal Maya collapse. This event, which occurred around 900 CE, has been attributed to a number of interrelated factors, including a series of severe droughts (Hodell et al. 1995). Martín Medina-Elizalde and Eelco J. Rohling (2012) analyzed four high-resolution paleoclimatic records (including one stalagmite and three lake sediment records) for the Yucátan Peninsula and concluded that droughts coincident with the Terminal Maya collapse were driven largely by a reduced frequency and intensity of hurricanes (as well as tropical depressions and other storms). Moreover, they argue that the groundwater table in the Yucátan Peninsula is particularly sensitive to tropical storm frequency and that only a modest reduction in these events would be necessary to explain the drought signal present in isotopic records for the region (958).

The hypothesis of hurricane reduction as a trigger for the Terminal Maya collapse has been challenged by Amy B. Frappier and colleagues (2014), who developed a stalagmite-based paleotempestological reconstruction for the northern Yucátan focused on mud layers incorporated into the stalagmite matrix as an indicator of hurricane-induced flood events in a cave. After calibrating recent mud layers to historical hurricanes, it was shown that the Terminal Classic droughts may have coincided with normal or even enhanced hurricane activity and that the severity of the reduction in precipitation during the Maya droughts may have been overestimated (Frappier et al. 2014:5155). Moreover, the study also concludes that hurricane flood events are likely underrepresented in stalagmites at times when droughts do occur, underscoring the complexity of interpreting these records and highlighting the need for more research in this area.

Coupled with the uncertainty concerning the frequency and intensity of hurricanes during the Terminal Maya collapse is the question of what specific effects hurricanes would have had on Maya civilization given the heterogeneous topography of the region. Hurricane activity has been cited as inherently destructive (Dunning and Houston 2011), especially in low-elevation areas such as the Belize coastal plain (Dunning et al. 2012). Medina-Elizalde and colleagues (2016), however, argue that hurricanes could potentially have had “beneficial” attributes, particularly by delivering precipitation to areas that might be drought-sensitive. While hurricane rainfall typically resulted in short, high-intensity bursts that may have damaged field-based agriculture, such storms would have replenished much-needed domestic freshwater supplies that also supported household gardens (see Dine et al., chapter 7, this volume). Moreover, the effects of hurricanes would likely have been different depending on whether settlements were inland or coastal, as sites in the latter locations would be sensitive to the effects of storm surges in addition to rainfall and wind.

The role of hurricanes in ancient Maya civilization requires additional research focused on new paleotempestological data from a range of sources and proxy indicators, including established sedimentological techniques based on the analysis of over-wash layers from coastal lakes and wetlands (Liu and Fearn 2000), geochemical and sedimentological analyses of sediments from cenotes and other karst features that have recently been shown to record past hurricane variability (Brown et al. 2014), speleothem data recording flood events (Frappier et al. 2014) as well as isotopic signals of hurricane rainfall (Medina-Elizalde and Rohling 2012), and tree rings that incorporate isotopically lighter hurricane water into the plant cellulose (Miller et al. 2006). The use of a range of data sources is especially important to document hurricane impacts at non-coastal sites, so that the effects of hurricanes at some of the larger urban centers, located in interior regions, can be better resolved.

The Taíno at Los Buchillones: Proactive Resistance

Ethnohistorical sources reveal that the Taíno, who occupied much of the Greater Antilles at the time of Spanish arrival, had a sophisticated knowledge of hurricanes. Specifically, they sequenced the event into successive stages that included the coming of the winds, the destructive force of the hurricane, and post-hurricane impacts (Cooper 2013:45). Each of these stages was represented by the deities Gatauba, Guabancex, and Coatrisque, respectively (Pané 1990), indicating the extent to which the knowledge and understanding of hurricanes was ingrained in their worldview (Schwartz 2015:8). Indeed, while there is some confusion over the origin of the Spanish word huracán—which later became the English word hurricane (6)—many sources believe it ultimately derived from “Juracán,” the Taíno zemi (or deity) for disorder and chaos who also controlled the weather. Given the close cosmological relationship between hurricanes and Taíno society (8), it is not surprising that the Taíno (especially those living in coastal locations) would be well prepared to deal with these events, both physically and psychologically.

The details of Taíno resilience can be evaluated at Los Buchillones, a site occupied between about 1220 and 1640 CE on the north coast of central Cuba (figure 2.3a, b; Calvera Roses et al. 2006; Valcárcel Rojas et al. 2006). The site is presently located under approximately 1 m of water in a lagoon and offshore of a narrow sand barrier. Due to the submerged nature of the site, excavations at Los Buchillones are undertaken by building a sandbag dyke around the area of interest and then pumping out the entrapped water (figure 2.3c). Having lowered the water level and removed surficial sediments, archaeologists have recovered a range of archaeological materials—including the remains of approximately forty collapsed Taíno houses, complete with rafters, thatch, and structural posts driven into the earth (figure 2.3d). The excavations also have yielded well-preserved carved wood (figure 2.3e), in addition to chipped stone and groundstone, shell, and ceramics (Calvera Roses et al. 2006), making Los Buchillones one of the largest and best-preserved Precolumbian settlements in the Caribbean.

Geoarchaeological investigations at the site consisted of the collection of a series of sediment cores in a shore-parallel transect through the archaeological remains to determine why the site was submerged (Peros et al. 2006). Based on these investigations, we developed a model for the evolution of the coastline indicating that the village was likely built over water rather than on dry land—a settlement strategy uncommon among Taíno sites (Curet 1992). Worldwide, the reasons for building settlements on pile dwellings vary and can include better access to marine, lagoonal, and terrestrial resources (Peros et al. 2006). However, at Los Buchillones, the nature of the coastal environment and the active hurricane region the site is located in suggest that measures to adapt to this dynamic environment may have played a role in influencing settlement location and domestic architecture at the site.

Los Buchillones is located on the mainland of Cuba, inside a long and wide archipelago of cayes and mangrove islands (figure 2.3b). The tidal range at the site is comparatively small, due to the damping effect of this offshore reef. On a daily basis, the effect of the reef would mitigate tides, but it would also protect against storm surges during a hurricane (Cooper 2013). Indeed, Hurricane Irma struck the area as a category 5 event in September 2017, and anecdotal evidence indicates that most of the destruction from flooding and wind occurred on the outer reef and that the mainland was much less affected. In addition, a network of limestone caves is located approximately 1 km south of Los Buchillones, which would provide hurricane shelters; Precolumbian artifacts found in these caves indicate their use, or at least knowledge of their location, by Precolumbian populations (Cooper 2012). Thus, at the settlement scale, the physical location of Los Buchillones appears to be well suited to mitigate hurricane risk.

The design of the domestic architecture at Los Buchillones may also have been made with hurricanes in mind. The strategy of pile dwellings (figure 2.3f) would have guarded against the effects of flooding, whether from storm surges or excessive rainfall (Cooper and Peros 2010). In addition, the structures include large-diameter wooden posts driven vertically into the ground, upon which lighter structural material (including palm thatch) was placed to complete the walls and roof. While high winds and rain could easily destroy much of the exterior of these structures, those elements could quickly be replaced with locally procured materials following the hurricane event. Indeed, this kind of house architecture contrasts with today’s poorly constructed cement and brick buildings in the area, which, when damaged, require considerable effort with non-local materials to repair, making post-hurricane reconstruction time-consuming and costly (Cooper 2012). Thus, Los Buchillones serves as an example of pre-industrial adaptation to high-risk meteorological events—a proactive response–which would have enhanced cultural resilience in the face of extreme storms.

Although there is evidence for strategies consistent with hurricane adaptation at Los Buchillones, no paleotempestological research has been undertaken in this region of Cuba. Such work would strengthen the case for a proactive response to hurricane activity by the Taíno, and the numerous lagoons and mangroves that are common in the area likely preserve evidence of past hurricane events. For example, paleotempestology could provide data to help assess the extent to which the Archipiélago de Sabana-Camagüey protected the site from flooding during the period of Taíno occupation. Paleotempestological data from the lagoons located on the offshore islands (such as Cayo Coco) could be compared to similar data from the mainland—including the lagoon Los Buchillones is located in—to identify differences between hurricane impacts on the exposed, ocean-facing side of the reef compared to the protected mainland. In addition, these studies could help elucidate whether hurricane activity was more active during site occupation (~ 1220–1640 CE), which would provide additional evidence for the need for the adoption of adaptive strategies by the Taíno to reduce hurricane risk.