During 1983 the Colorado River experienced high flow conditions due to abnormally high snowmelt and runoff from the upper basin states. Flood stages provided sufficient hydraulic head to force water under the protective levee system around Yuma, Arizona. This caused inundation of adjacent urban and agricultural lands from the subsequent rise in the local water table. A methodology was planned whereby a series of "curtain wells" would be installed and the water table pumped down to below grade conditions. The feasibility of accomplishing this was studied by retrofitting an existing agricultural well located on "Yuma Island", and installing a series of monitor wells in a two-dimensional array perpendicular and tangential to the levee. The production well was pumped continuously for eight weeks, and the subsequent drawdown within the piezometers monitored. Based on the data gathered, it was possible to predict the effectiveness of the proposed curtain well system.

Estimates of the variable costs for groundwater pumping and overdraft in Central Arizona are compared to the price of water delivered via the Central Arizona Project (CAP). The respective marginal costs of supplying irrigation water through each of the two alternative sources are compared to the marginal demand for water by farmers. Finally, the relative cost competitiveness of groundwater versus CAP water is evaluated against a number of alternative rates of energy cost escalation and groundwater decline.

Stormflow and sediment production increased greatly after a wildfire on three small cha3parrlal watersheds in .entral Arizona. Peaks frequently exceeded 5 m³ s⁻¹ km⁻² (450 cfg mi⁻²) when 15-minute rainfall intensity exceeded 50 mm hr⁻¹ (2 in hr⁻¹) on catchments that, before burning, responded little to intense rainfall. Source water for the flashy spates and heavy erosion was surface runoff on the severely burned, unprotected, water -repellent soils. For a few years after the fire, intense summer rains produced a disproportionate amount of the runoff and sediment. Early postfire recovery was rapid; severe flooding and erosion were over in 3 years, and within 5 to 10 years stormflows and peaks declined to near prefire levels. Postfire conversion to grass on one watershed did not appreciably change the rate of recovery.