Overview of Theoretical Cave
Relatively little field research has been conducted on subaqueous cave system evolution. Most karst research has concerned itself with the geomorphology of karst terrains and origin of dry caves. Dry caves have been studied in depth, due primarily to their accessibility. With various theories originating in the early 1900's concerning vadose and deep phreatic speculation to the presently accepted Ford-Ewers model stressing the drainage basin and hydrogeologic setting, karst hydrogeologic research has undergone considerable growth. The largest progress in the field has resulted from increased direct observation of this unique environment. Almost all scientific cave literature available today is written with emphasis on the examination of dry cave systems with proposed parallels to subaqueous or phreatic cave evolution.
Historically, there have been three primary theories concerning the origin of cave passages. These include the vadose theory, the deep phreatic theory and the water table cave theory (White, 1988). The origin of cave passages by vadose theory involves the infiltration of surface waters percolating through the unsaturated zone invoking dissolution of carbonate rock above the water table. This movement downward by chemically energetic waters produces easily recognizable dissolution features such as surface stream capture by vertical dissolution pits, dissolving sinkhole drains, and prominent solution canyons which terminate at the top of the saturated zone (White, 1988). While there are many variations on the theme of vadose cave passage formation, many of the elements remain the same. There must be (1) a hydraulic gradient sufficient to drive flow, (2) a developing surface or underground drainage pattern, (3) some relief of carbonate strata above the water table, and (4) a change in base level of the watershed to initiate the process of dissolution through vadose entrenchment. It is typically thought that initial movement of surface waters into the carbonate strata begins through highly permeable joints and/or fractures. These joints and/or fractures are then widened over time, developing vertical dissolution features. After a sufficient time has elapsed, the conduit system produced is one in which evolution of passages has been through dissolution by streams with a free surface (Dreybrodt, 1988).
The theory of deep phreatic conduit evolution has its early origins in the alpine karst of Europe. Later championed by Davis (1930), the deep phreatic theory involves a hydrostatic head which drives flow deep into the phreatic zone along flow lines (White, 1988). The driving force of flow was again thought to be the hydrostatic head. The theory involves uniform dissolution throughout the flow lines leading to base level. With this in mind, many proponents of the deep phreatic theory believed this dissolution would take place tens to hundreds of meters below a stable water table. This stable water table was usually thought to coincide with a peneplain geomorphology, while subsequently being preserved for a sufficiently long period of time (Dreybrodt, 1988). White (1988) mentions that upon careful examination of both the Davis (1930) and the Bretz (1955) hypotheses the reader is left with "the distinct impression that Bretz's deep phreatic was nowhere near as deep as Davis' deep phreatic."
The water table theory of karst conduit development was advanced primarily by Swinnerton (1932). Upon observation of the overwhelming majority of horizontal cave passage arrangement, Swinnerton (1932) hypothesized that passage development must take place near the water table surface and specifically in the vertical region of fluctuations of the water table from periodic seasonal variations (White, 1988). As water moves down through the vadose zone to the water table, it is theorized this water moves along the water table to base level. The largest movement occurs in the shallow phreatic zone near the water table surface (Dreybrodt, 1988). With this large mass transport occurring in such a limited location, the majority of dissolution will thus occur in this horizontal position, eventually developing a conduit system. This is further strengthened by a stable water table surface, in which flow is allowed to migrate down the hydraulic gradient toward base level over a sufficient period of time.
The water table theory is also supported by the observations of tiered caves. A tiered cave is one in which there are several levels of nearly horizontal cave passage. White (1988) suggests a causal relationship between the coincidence of passage levels and river terraces. He concludes there to be a correlation between the position of local base level and the position of the water table in the evolution of the drainage system. Davies (1960) hypothesized this in his theory of shallow phreatic conduit development, in which the cave levels formed at the water table surface of the maturing river valley base level. Davies (1960) and White and White (1974) related river terraces to caves found in the Potomac River Valley. Others such as Sweeting (1950) and Droppa (1957) have also found the same type of correlation in Yorkshire, England and in Demanova Valley, Czechoslovakia, respectively.
Perhaps the most compelling observations for water table cave development came from Miotke and Palmer (1972) and Palmer (1981) concerning the Flint Ridge-Mammoth Cave System in central Kentucky. Palmer systematically distinguishes six distinct cave levels, referring to them as A through F, and correlates their development to periods of stable levels of the Green River. He comments that each of these major levels formed at progressively lower altitudes, in which they can be considered time horizons or markers (Palmer, 1981). The interesting point to note is the younger, more recent levels of passage development are the ones with the lower elevation. Many of these lower levels still function as part of the drainage system, where the older, higher elevation conduits are primarily relict cave stream passages. The tiered caves form when new drainage routes are formed below the older routes, which in turn leave the older, higher elevation passages dry as flow is captured by the lower passages (White, 1988). This type of development is one in which discontinuous downcutting within the drainage basin is dependent on base level lowering (White, 1988).
One of the underlying principals of the preceding theories is the concept of a base level. Base level is defined for a given basin with hydraulic gradient, as the point or place where the gradient is defined as zero. The ultimate base level for all drainage basins is sea level, whereas for a chosen basin, its base level may be defined to be a higher ordered fluvial, lacustrine, or other lower elevation body of water, into which the basin drains, due primarily to the force of gravity.
Base level lowering has been cited as the major influence controlling conduit development in the Flint Ridge-Mammoth Cave System. Palmer suggests the only explanation for the grouping of major passage levels, in a vertical section, is directly related to periods of eustatic fluvial base levels of the Green River (Palmer, from White and White 1989). He further postulates that the piezometric surface of an actively forming passage progresses toward an elevation that closely coincides with that of the base level. As passage dissolution evolves through time, the hydraulic efficiency increases to the point where there is an indiscernible drop in hydrostatic head within the conduit, and thus this passage would relate a reasonable elevation of base level (Palmer, from White and White, 1989).
Ford and Williams (1989) point out, however, that base level may not be sea level, as evidenced by the documented karst springs which resurge well below sea level. If sea level is not the ultimate base for karst dissolution, and phreatic conduits are presently observed, then what is the ultimate base for active conduit formation? (Ford and Williams, 1989). Ford concludes that the active base for dissolution must be (1) the state of the system, (2) the age, and (3) the extent of clastic infilling (Ford and Williams, 1989). These arguments may be supported by the mass transport along flow lines which migrate to depth below the free piezometric surface, which are normal to the equipotential lines.
Another consideration of karst conduit development in need of clarification is that of submergence. Quaternary sea level history suggests that sea level height has fluctuated considerably (Shackleton, 1987). With the base level lowering concept in mind, Quaternary sea level height would play a major role in development of karst terraines in coastal carbonate platforms. White (1988) points of that with a lowering of sea level many of the platforms of the Gulf of Mexico were exposed to karst processes, where large conduit systems formed and then with the melting of glacial ice, sea level rose and drowned these conduits. Ford and Williams (1989) use the Florida and Yucatan peninsulas as an example of submerged conduit systems in karst. They summarize that multiple fluctuations of sea level due to glaciation would (1) contract the vadose zone and expand the phreatic zone, (2) create a zone of corrosion which would spread with the fresh-salt water mixing zone, and (3) would flood coastal springs and possibly fill existing conduits with clastic material.
Contemporary theories for the development of karst conduits in carbonate rocks have brought the system as a whole into mind. While considering the entire karst basin as a complex system, whose evolution is influenced by an intricate network of positive and negative feedback loops (Dreybrodt, 1988), the modern encompassing theories have tried to find a general underlying framework which is common to all karst basins. The variables of karst processes can be grouped into three essential domains, the chemical processes, the physical processes and the hydrogeologic setting (White, 1988).
The chemical processes have several independent variables which influence the karstification process. These are the temperature, both seasonal and annual mean, the amount of precipitation, and the amount of carbon dioxide available. These three variables are usually grouped together as climate (White, 1988). The carbon dioxide, usually derived from organic flora, along with precipitation drives the chemical rate-limited reactions causing carbonate dissolution.
The physical processes are controlled by the precipitation and the local relief of the carbonate strata. The precipitation is included in the physical processes because a hydrostatic head must be maintained to continually or periodically replenish the chemically aggressive waters which dissolve the rock (White, 1988). The energy gradient required to drive the movement of groundwater is the earth's gravitational field. The relief of carbonate strata is needed to move this flow of groundwater over and through the strata toward base level.
The hydrogeologic setting is of primary importance and can be thought of as having three major variables: (1) tectonic or structural environment, which encompasses the attitude, folding and faulting characteristics (2) thickness of the soluble rock strata, and (3) stratigraphic and lithologic properties, which include porosity, permeability, texture, and matrix or cement material.
These variables of karst were systematically brought together in a series of articles by Ford (1968, 1971), Ford and Ewers (1978), and Ewers (1982), into what is commonly referred to today as the Ford-Ewers model (White, 1988). The approach to the model is a comprehensive zenith of solution chemistry and fluid mechanics, concordant with observations of passage profiles and cave system patterns and their relation to the hydrogeologic setting. The model incorporates many of the variables of karst verified by White (1988). It incorporates the previous theories of cave passage origin into a single genetic theory (Dreybrodt, 1988), in which the major controls on conduit formation are the fissure (fracture and joint) frequency within the soluble rock material, and the ability of chemically aggressive waters to penetrate these fissures (Ford and Williams, 1989).
The Ford-Ewers model is broken into a four-state approach, and is also a guideline to the variations of actual karst terraines which are a continuum (Ford and Williams, 1989). The model integrates many of the features of observed karst morphologies and varying physiographic cave patterns, into an emphasis of geologic setting and climatic regime. It is important to realize that the model also resolves the concept of base level lowering and the complex ideas of the karst water table (White, 1988). In other words, the model does not require differentiation of phreatic and water table caves, and simplifies assignment of cave passages to the state in which its characteristics dominate.
The Ford-Ewers model while comprehensive, does not address all cave development senarios. It should be kept in mind that these models are promarily constructed from observations and data acquired in dry caves. While many of the mechanisms are similar, subaqueous cave development may follow the same type of controls, but there are bound to be some exceptions. This article is to serve as an overview of these models. More information about these and other theories can be found in the following list of references.
Chris Werner has his BS from University of Pittsburgh in Earth and Planetary Science. He is completing a MS in both Geology and Mathematics, and has begun PhD study in Geophysical Fluid Dynamics at Florida State. Werner has been cave diving for 6 years and works as a Geologist at the Florida Geological Survey.
Chris Werner , Florida State University