ASSUMPTIONS USED IN PA	REALITY	IMPLICATIONS
The Culebra dolomite is a confined aquifer throughout the region. Recharge and lateral flows are limited to small values by matrix permeabilities of the Rustler and Dewey Lake.	At the shafts and to the west, fractures and dissolution channels interconnect most strata from top of the Salado across all Rustler and Dewey Lake beds. Storm infiltration events and evaporation at Laguna Grande support an order of magnitude greater flux than PA used.	Unconfined flow definitely occurs W of the LWA, and occurs to decreasing degree west to east and from the repository to the surface. A simple 2-D model confined to the Culebra cannot suffice, whereas the entire saturated interval is heterogeneous, requiring transmissibility that varies strongly with water levels.
Karst conditions are unimportant east of Livingston Ridge, and do not influence transport in the sole significant aquifer, the Culebra.	Karst features at the surface exist throughout the LWA, except perhaps east of the repository. Saturated karst is found at WIPP-33, and must exist also beneath the LWA. Open fractures prove that karst occurs above the repository (Figure. 8)	Rapid, episodic recharge through the vadose zone to a water table of uncertain position induce rapid lateral flows in karst channels on and off the LWA, but mainly above the Culebra. A rising water table will enhance the karst aquifer. Groundwater flow from WIPP cannot be modeled with continuous Darcyan properties, if at all.
Lateral conveyance of solutes will be significant only from intruding boreholes via the Culebra dolomite.	Pathways from subsidence or hydrofractures up to the Rustler will intercept solution channels below or above the Culebra, so lateral conveyance will be through karst features to Nash Draw.	Because the location, size, continuity and orientations of solution channels in the Rustler are unknown, transport cannot be modeled, but should be assumed more rapid than Culebra flows. Compliance cannot be met.
It is unnecessary to collect hydrologic data from known karst features nor of boundary conditions such as the recharge or discharge rates or locations.	Karst flows may dominate over matrix and fracture flows, so data evaluating each are essential.	PA modeling of flow and transport via the Culebra is unconservative, misleading, and must be replaced.
In the Culebra itself, flow is dominated by horizontal fractures, with diffusion into matrix retarding transport.	Horizontal fractures are actually shaley partings formed upon sampling. Two sets of vertical fractures partially dissolved from the Culebra conduits.	A conduit system largely confined to dissolution channels following filled fractures may coalesce downstream to a small number of large conduits for rapid, occasional flows. Integration of the Culebra with Rustler karst features is also significant. No known model suffices.