Landsat multispectral scanner data for Brazos County, Texas, were evaluated in terms of effectiveness for classifying soils on vegetated landscapes at three times during the year: a time of normally adequate soil water, a time of expected soil water deficit, and a time when soil water is normally being replenished. Six test sites were used to evaluate LARSYS supervised and unsupervised classification of vegetated soil landscapes. Open grassland soils were best separated in the fall during a period when soil moisture was being replenished after the summer period of soil water deficit. Woodland soils were separated by Landsat data in late spring when adequate moisture was available. However, a high degree of accuracy was not achieved using Landsat for separating soil map units. Accurate separation of soil mapping units on vegetated landscapes was not possible during late summer when soil water was deficient. Selected soil properties important to plant growth were separable on the test sites using June and October Landsat data. Particle size and soil moisture regime were separated at both dates. Soils with argillic horizons were separated from soils without argillic horizons.
