SURFACE SOIL-WATER LOSS AFTER SUMMER RAINFALL IN A SEMIDESERT GRASSLAND

Surface soil water contents were analyzed for bare and litter- or grav el-covered soils for 20 drying periods after summer rainfall in southe astern Arizona. Water contents were measured every minute by fiberglas s cells calibrated for the sandy-loam soil, and 30- to 60-min averages were stored by microloggers. Water loss to equivalent matric potentia ls of -1.5 MPa for bare surface soils (1-3 cm) was fast (1.2 +/- 0.5 d ays), moderate (2.6 +/- 0.9 days), or slow (5.9 +/- 1.4 days) when the upper 15 cm of soil at the start of the drying period had 13.3, 27.1, and 36.1 cm of water, respectively. Those water contents are near sat uration and near field capacity for slow and moderate drying rates, re spectively. For fast drying rates, water content was nearfield capacit y at 1-3 cm but dry below 8 cm. Litter or gravel mulches increased the time of surface soil water availability by 0.7 to 1.9 days for fast a nd moderate drying periods, respectively. Drying from the surface down into the soil profile averaged 3.6 cm day(-1). Short summer rainstorm s may germinate warm-season grasses which will desiccate if subsequent rainfall is nor persistent enough to maintain surface water availabil ity and allow adventitious root development. Estimates of lime that th e drying front reaches the depth of seminal roots of warm-season grass es were used to suggest rain-rainless day patterns that might result i n seedling mortality. Sowing during the summer rainy season when the p robability of these patterns is minimal may increase revegetation succ ess. High correlation of available water periods with meterological va riables suggests that soil water dynamics could be accurately estimate d using physically based models, and may permit more detailed modeling of seedling establishment.