Understanding the hydrologic processes of rangeland, plant communities
is essential to determine if water augmentation through shrub managem
ent is feasible. Vegetation manipulation studies are costly, difficult
to accurately replicate, and often require more than 10 years to dete
rmine treatment effect on the water budget. If properly applied, hydro
logic simulation models are an attractive alternative for assessing ve
getation manipulation practices. The ERHYM-II model was evaluated to d
etermine if it was capable of simulating the water balance for honey m
esquite shrub clusters, grass interspaces, and bare soil in south Texa
s. The simulated water budget was within 2 percent of the measured eva
potranspiration for the shrub clusters and grass interspaces. The mode
l underestimated the number of runoff events and overestimated runoff
volume for the grass interspace and shrub clusters. Simulated runoff w
as overestimated by approximately twofold for the grass interspace and
threefold for the shrub clusters. Although simulated runoff was subst
antially overestimated, observed and simulated runoff only accounted f
or 3 to 6 percent of annual rainfall for the grass and shrub dominated
areas, respectively. Simulated evapotranspiration was underestimated
by 18 percent and soil water content was overestimated by 82 percent f
or the bare soil. The model underestimated evapotranspiration for the
bare soil as a result of restricting evaporative losses to the first s
oil layer. Based on our analysis, the ERHYM-II model has the potential
for simulating the annual water balance for semiarid rangeland plant
communities where runoff and deep drainage are limited components of t
he water balance.