An efficient time-dependent equation for predicting ground surface temperature devised by Bhumralkar (1975) and Blackadar (1976) is tested against a 12-layer soil model and compared with five other approximate methods in current use. It is found to be generally superior if diurnal forcing is present and very much superior to the use of the insulated surface assumption. An analogous method of predicting ground surface moisture content is presented which allows the surface to become moist quickly during rainfall or to become drier than the bulk soil while evaporation occurs. These improved methods are not of much relevance unless the main influences of a vegetation layer are included. An efficient one-layer foliage parameterization is therefore developed that extends continuously from the case of no shielding of the ground by vegetation to complete shielding. It includes influences of both ground and foliage albedos and emissivities, net leaf area index, stomatal resistance, retained water on the foliage, and several other considerations. When it is tested against data for wheat measured by Penman and Long (1960), it appears quite adequate despite the many simplifying assumptions. The parameterization predicts that errors of up to a factor of 2 in evapotranspiration can be incurred by ignoring the presence of a vegetation layer.