Using mainly satellite images of both the Arabian Gulf and portions of the Red Sea, Purkis et al. (2010) applied sophisticated mathematical modeling to explain large polygonal seafloor features, “templates,” that are amplified in some areas by modern coral growth. They conclude from their model that the polygonal features result from karst weathering in a humid climate during lowered Holocene sea level. Purkis et al. also separate the coral-capped features into type 1 and smaller type 2 structures. Our comments are directed mainly toward the origin of the type 2 features in the Arabian Gulf.
Although the mathematical models used by Purkis et al. include sophisticated calculations, we believe they are misapplied because no attempt was made to verify karst processes using traditional petrographic methods. Karst, by definition, implies diagenetic changes caused by meteoric water that invariably leaves diagenetic footprints, such as leached grains, fossil molds, and low-magnesium calcite cement. Without verification of actual karst processes, we conclude that the application of mathematical models is unwarranted. Such polygonal features can form in many non-karst environments, including tundra, permafrost, saltpans, and gypsum flats.
Type 2 polygonal features were examined and explained both above and below sea level, including intertidal flats in the Arabian Gulf, more than four decades ago (Kendall and Skipwith, 1969; Shinn, 1969; Purser and Evans, 1973).
Shinn (1969) attributed the type 2 polygonal features to expansion of rapidly cementing seafloor grainstone. Expansion of lithifying subsea sediment produces bending and folding, and overthrusting of layers best explained by force of crystallization. This lateral expansion develops polygonal patterns the reverse of, but similar to, mud cracks in dried mud layers. Such patterns are also formed on Arabian Gulf sebkas, where they result mainly from expanding halite and gypsum crystal growth. Similar patterns also develop in calcrete or caliche layers, so-called caprock, during expansion and contraction caused by repeated thermal changes. Cross sections of the areas of merged and deformed rock layers seen on outcrops in the United States have traditionally been called teepee structures.
Similar features were found covering hundreds of square kilometers west of Qatar in the shallows near Bahrain during fieldwork conducted in the mid-1960s (Shinn, 1969). Expansion at plate margins was demonstrated by underwater dissection of overthrust layers that revealed striated surfaces produced by sliding of rock on rock. Underwater photographs of these features are shown in figures 13–15 in Shinn (1969). Petrographic examination of the young limestone revealed no evidence of leaching characteristic of karst conditions. The upturned and mounded margins of the plates allowed accumulation of a few centimeters of sediment in the low areas, while waves and currents swept the upturned margins clear of sediment. Corals and especially gorgonians, marine algae, and marine borers, preferentially populate the elevated sediment-free margins. The features were clearly seen from the boat, while diving, in aerial photographs, and during low-altitude aerial surveys. Typically, the polygonal-shaped plates were 30–60 m across and only a few centimeters thick. As many as four such plates were seen one atop the other and separated by uncemented sediment in some areas. The dimensions of polygonal plates appear to be controlled by thickness of the cemented layers, with size increasing with increasing thickness. Thicker layers are more common in deeper water, especially where they form the famous Great Pearl Bank.
We conclude from our earlier observations that that these widespread features, especially the type 2 features, are the product of ongoing submarine processes and are not related to karst processes that would imply climate change. The conclusion of Purkis et al. that the features are the result of early to late Holocene climate change, i.e., humid conditions, is not supported by these earlier observations.