Ground-water sapping and the generation of natural amphitheaters   Donald Rosenberry   Geography Department University of Colorado Boulder, CO 80309   U.S. Geological Survey MS 413, Bldg. 53, DFC Lakewood, CO  80225 http://www.colorado.edu/geography/ geomorph/geog6241/GWsapping.htm

The spectacular theater-headed canyons common throughout the Colorado Plateau of the southwestern United States are due in part to the process of ground-water sapping.  Click on the links below to find out more about ground-water sapping, how it occurs, and how ground-water sapping acts to create these beautiful features of the American southwest.

What is ground-water sapping?

Where do we find ground-water sapping?

How does sapping work?

What does sapping have to do with natural amphitheater formation?

Some relevant references on ground-water sapping

 

Headward erosion of a side channel along the Snake River in Idaho as a result of ground-water sapping.

Chemical precipitates are often a good indicator that ground-water sapping is occurring.

Thomas Dunne (1990) proposed that processes that generate sediment transport resulting from subsurface flow be lumped into two main groups: (1) development of a critical drag force that entrains particles in water seeping through or out of a porous medium.  This causes liquifaction or Coulomb failure (especially on steep slopes); (2) application of a shear stress to the margins of a macropore.  The former process is associated with unconsolidated sediments and the latter with consolidated sediments, especially where fractures in the consolidated sediments provide the main conduits for flow.  The critical stress for either process may be reduced by chemical or mechanical weathering caused by the exfiltrating water.  Few data are available, but it is thought that long periods of chemical erosion in between sapping-induced failures contribute to reduction in intergranular cohesion (Dunne, 1990).  There is much evidence that mineral decomposition is more thorough in sapping areas than in surrounding rock (Howard, 1986).  However, the weathering of chemically erosive materials, such as karst formation in limestone, is not generally considered as part of the sapping process.  No matter the process, Dunne (1990) also pointed out that the weathered residuum must be removed by some erosional process in order for sapping-induced erosion to continue.

Back to top

Where does ground-water sapping occur? -- In many locations and settings, from streambanks to river canyons, from ocean bottoms to ocean coastlines, from sandy beaches to escarpments on Mars.	Erosion of cohesionless sediment adjacent to downgradient stream causes eventual capture of upgradient stream (Pederson, 2001).	Ground-water discharge in the Thousand Springs area of the Snake River, Idaho, causes accelerated erosion of canyon wall.
Discharge of ground water or gas through ocean sediments suspends ocean sediments so they can be removed by ocean currents, creating a pockhole or "seabed pockmark" (Hovland and Judd, 1988).		Headward erosion caused by ground-water sapping creates fjord-like bays on Cape Cod (Fitzgerald et al., 2002).
Erosion rivulets in beach sand created by ground water discharge following a waning tide.		Short, deep, stream-like channels on Mars may be the result of ground-water sapping (Gulick, 2001).  Features on Mars generated substantial interest in studies of ground-water sapping during  the 1980s and early 1990s.

Back to top

(Some characters do not display correctly with older versions of Netscape.  Please click on this link if you see question marks instead of symbols in the text below.)  Symbols-friendly view

The critical seepage force required to move sediments can be envisioned using variables shown above.  In the upper cartoon, q_z represents the vertical component of seepage, q.  q_z is proportional to the hydraulic potential

  .  

λ is the angle between the seepage-force vector and the vertical plane.  This force is resisted by the immersed weight of the grains which can be expressed as 

, 

where F/V = force per volume, ρ_s = density of the solid grains, ρ_f = density of the fluid, g = acceleration due to gravity, and p = porosity.  If we represent the vertical component of the hydraulic-head gradient as i_z, then sediment sapping will occur when

  .

This simplified analysis is for a horizontal slope.  In most sapping conditions, the sediments are located on a sloping surface, and a rotational or torque analysis is more appropriate.  Panel b, shown above, from Howard and McLane (1988), shows the forces acting on a grain that needs to rotate out of position in order to be moved by ground-water sapping forces, indicated here as Fs.  Fw represents the fluid drag or shear stress due to fluid flow along the sloping surface, and Fg is the gravitational force imparted on the particle.  θ is the angle of the sloping surface, ø is the angle of internal friction (or the angle necessary for the particle to have rotated to the point of incipient motion), and v is the angle from horizontal of the direction of seepage force.  If we include coefficients C₁ and C₂ to represent the effects of particle shape and particle packing, we can define the point at which the hydraulic gradient, i, is sufficient to mobilize sediment as 

(σ in this equation = ø  as earlier defined).

The magnitude of i required for sediment-stability failure is minimized when v = ø - θ.

A case study  (Collison, 2001) Ground-water sapping in marl located near the Sierra Nevada mountains in Spain is related to formation of tension cracks that accelerate the process.  Collison used SEEP/W to model shear stress, tension head, and pressure head in the absence of tension cracks, and in the presence of tension cracks of various sizes.  He found that theater-head retreat is associated with a loss of suction resulting from surface flow through tension cracks.  Headward erosion was found to be coincident with shear-stress buildup and the timing of significant rainfall events.  He identified a cycle of headward erosion related to buildup of shear stress and strain; development of a tension crack; growth of the crack; preferential flow of surface and ground water, and piping flow; failure of a sediment block; and erosion of the collapsed debris and redevelopment of shear stress.

Shifting of the water table from the break in slope at the left side of the hillslope to the upper land surface in response to the development and enlargement of a tension crack.

Shear stress is extremely large in the vicinity of the sapping area where the discharging point of the ground water is exposed to atmospheric pressure.

The cycle of headward erosion.

Another case study   (Bryan et al., 1998) Water was rained on sand in a 7.1 x 2.4 m sand tank.  Fourteen TDR probes were used to measure the water table and the soil moisture conditions above the water table.  Rainfall rates of 50-60 mm/hr were simulated.  Some rill incision was associated with surface-flow processes, but most rill formation occurred when rainfall was sufficient to saturate the sand.  Sediment yield increased greatly upon soil saturation.  Seepage forces alone (5.3 E-5 m/s) were not sufficient to transport sediment, but seepage force, when combined with positive pore pressure and reduced shear strength at the water table, was sufficient to move sediment.  Sediment eroded rapidly headward at the water table, causing undercutting and slope steepening.  The cliff-like slopes were maintained by the slightly drier and more cohesive unsaturated sediments above the water table.

This is an image from the surface of Mars (Jakosky and Mellon, 2004).  Note the similarity of these features and those from the sand-tank experiment to the right.  This indicates that ground-water sapping likely occurs over a great range of scales.

Note the theater-headed valleys in the drainages near the upper portion of the photo.

Back to top

What does sapping have to do with natural amphitheaters?  -- Sapping typically erodes sediments at the base of sandstone, undercutting the formation to create overhanging sandstone or vertical cliffs. Amphitheater development is related to ground-water flow direction jointing and faulting permeability contrasts formation slope and dip angle formation cohesion	Ground water flows along the top of less-permeable formations and is visibly discharging at the base of the more permeable sandstone above.
	Amphitheater near the Colorado River just upstream of the Colorado-Utah border. No evidence of sapping is visible, other than the shape of the amphitheater-like formation.  In this case, sapping may have occurred long ago when the water table was much higher, either in response to a wetter climate or during a time when the nearby Colorado River was not as deeply incised.

Laity and Malin (1985) indicated that two different morphologies are present in the Navajo Sandstone in Utah, Arizona and Nevada.  Where ground water is discharging to a stream channel, ground-water sapping occurs and the stream channels are wide with steeply sloping sides.  Where surface water flows to ground water and infiltrates into the sediment, these sapping features do not appear and a more typical dendritic stream-channel network develops.  This is clearly shown below, where sapping processes are evident in panel A and are absent in panel B.

These greatly different morphologies in nearly adjacent drainage basins were caused by regional scale geology.  Ground water flows generally from east to west across a monocline near the lower Escalante River.  Where the ground-water gradient is less than the land-surface gradient, ground water discharges to the surface and creates sapping features.  This occurrs along the eastern tributaries of the Escalante River as shown below.  Where westward-flowing ground water does not discharge to the surface, along the western flank of the Escalante River watershed, narrow-walled tributaries form.

Laity and Malin also found that jointing was related to the direction of headward stream-channel erosion associated with ground-water sapping.  They showed that stream channels erode in a direction parallel to the jointing orientation, seen below.

Canyon growth is away from the main channel and is parallel to the strike of the jointing that is evident in the upper portion of the photograph (Laity and Malin, 1985)     Wet alcoves are situated in areas where greater headward erosion has occurred.  Many dry alcoves are situated in areas where little or no headward erosion is evident.  The extent of headward erosion is somewhat proportional to the volume of ground-water discharge (Laity and Malin, 1985).

Back to top

References cited

Bryan , R.B., Hawke, R.M., and Rockwell, D.L., 1998, The influence of subsurface moisture on rill system evolution: Earth Surface Processes and Landforms, v. 23, p. 773–789.

Collison, A.J.C., 2001, The cycle of instability: stress release and fissure flow as controls on gully head retreat: Hydrological Processes, v. 15, p. 3–12.

Dunne, T., 1980, Formation and controls of channel networks: Progress in Physical Geography, v. 4, p. 211-239.

Dunne, T., 1990, Hydrology, mechanics, and geomorphic implications of erosion by subsurface flow, in Higgins, C.G., and Coates, D.R., eds., Groundwater Geomorphology: The Role of Subsurface Water in Earth-Surface Processes and Landforms: Boulder, Geological Society of America, p. 1-28.

FitzGerald, D.M., Buynevich, I.V., Davis Jr., R.A., and Fenster, M.S., 2002, New England tidal inlets with special reference to riverine-associated inlet systems: Geomorphology, v. 48, p. 179–208.

Gulick, V.C., 2001, Origin of the valley networks on Mars: a hydrological perspective: Geomorphology, v. 37, p. 241–268.

Hovland, M. and Judd. A.G., 1988, Seabed Pockmarks and Seepages, Alden Press, Oxford , 293 p.

Howard, A.D., 1986, Groundwater sapping on Mars and Earth, in Howard, A.D., Kochel, R.C., and Holt, H.E., eds., Proceedings and Field Guide, NASA Groundwater Sapping Conference, Flagstaff, Arizona, November 1985, p. vi-xiv.

Howard, A.D., Kochel, R.C., and Holt, H.E., 1988, Sapping features of the Colorado Plateau: NASA Special Publication 491, 108 p.

Howard, A.D. and McLane, C.F. III, 1988, Erosion of cohesionless sediment by groundwatetr seepage: Water Resources Research, vol. 24, no. 10, p. 1659-1674.

Jakosky, B.M. and Mellon, M.T., 2004, Water on Mars: Physics Today, April 2004, p. 71-76.

Kochel, R.C., Howard, A.D., and McLane, C.F., 1985, Channel networks developed by groundwater sapping in fine-grained sediments; analogs to some Martian valleys, in Woldenberg, M.J., ed., Models in Geomorphology: Boston , Massachusetts , Allen and Unwin, p. 313-341.

Laity, J.E., and Malin, M.C., 1985, Sapping processes and the development of theater-headed valley networks on the Colorado Plateau: Geological Society of America Bulletin, v. 96, p. 203-217.

Pederson, D.T., 2001, Stream piracy revisited: a groundwater-sapping solution: GSA Today, v. 11, no. 9, p. 4-10.

Parker, G.G. and Higgins, C.G., 1990, piping and pseudokarst in drylands, , in Higgins, C.G., and Coates, D.R., eds., Groundwater Geomorphology: The Role of Subsurface Water in Earth-Surface Processes and Landforms: Boulder , Geological Society of America, p. 77-110.

Sap-happy students and sage skipper studying science and sopping suds