One‐dimensional infiltration experiments were conducted to study the mechanics of unstable flow in homogeneous soils under non‐ponding infiltration. A mechanistic model is presented which explains soil water pressure gradients that are characteristic of stable and unstable flow in homogeneous soils and is based on a dynamic soil water entry pressure, the Darcy‐Buckingham flux equation, and hysteretic moisture retention functions. Infiltration experiments were conducted with five sand samples under applied fluxes of 2, 5, 20, and 50% of their saturated hydraulic conductivity. Soil water pressures were measured at fixed depths following passage of the wetting front. A trend of decreasing soil water pressure over time following passage of the wetting front is not predicted by Richards' equation and produces unstable flow. Under air‐dry initial soil water conditions, soil water pressures were unstable for all fluxes in the three coarser sands. In the two finer sands, unstable flow occurred at infiltration rates of 20 and 50% of their respective saturated hydraulic conductivities but stable flow occurred at lower fluxes. Soil water pressure measured just behind the wetting front was found to be an increasing function of applied flux and average grain size of the media for infiltration with air‐dry initial soil water content. Additional tests showed that systems that produced unstable flow under air‐dry initial soil water contents exhibited stable flow during infiltration with initial soil water contents that were greater than air dry.

Across the western Great Plains of North America, groundwater pumping for irrigated agriculture is depleting regional aquifers that sustain streamflow for native fishes. We investigated linkages between groundwater pumping from the High Plains Aquifer and stream fish habitat loss at multiple spatial scales during spring and summer [2005][2006][2007] in the Arikaree River, eastern Colorado, USA. Monthly low-altitude flights showed that flowing reaches were reduced from about 65 to Ä15 km by late summer, and long permanently dry segments in the lower basin prevent recolonization. Drying occurred rapidly during summer within three 6Ð4-km river segments, and patterns in habitat connectivity varied among segments owing to hydraulic conductivity. Most refuge pool habitats dried completely or lost more than half their volume, disconnecting from other pools by late summer. On the basis of these empirical habitat data, and historical groundwater and streamflow data, we constructed a MODFLOW model to predict how groundwater pumping will affect water table levels and fish habitat under three future scenarios. Under the most conservative scenario, we predicted that only 57% of refuge pools will remain in 35 years (2045), nearly all isolated in a 1Ð7-km fragment of river. A water balance model indicated that maintaining current water table levels and refuge pools for fishes would require a 75% reduction in groundwater pumping, which is not economically or politically feasible. Given widespread streamflow declines, ecological futures are bleak for stream fishes in the western Great Plains, and managers will be challenged to conserve native fishes under current groundwater pumping regimes.

No abstract

Swelling clay soils undergo volume changes as the moisture content varies. This volume change is typically described by a shrinkage characteristic curve. There is a need for a generally acceptable model to describe these curves without requiring extensive soil shrinkage measurements. Models that are presently available in the literature do not represent the shrinkage characteristic curve data adequately and, thus, have limitations to their application. We have developed an analytical volume change model that was derived using the theoretical boundary conditions at the endpoints of the four shrinkage ranges. Thus, the model describes all four shrinkage ranges better without additional input requirements. The proposed model describes a smooth shrinkage characteristic curve with continuous slopes at the endpoints of each shrinkage range. The analytical model was verified against measured shrinkage characteristic curves and was found to represent the data very well.

No abstract