The land-surface flux model (PROGSURF) designed jointly at the Universities of Vienna and Budapest is reviewed; it belongs to the broad spectrum of PILPS¹ models. PROGSURF comprises one vegetation layer and three soil layers. Temperature prediction is made by the heat conduction equation in conjunction with the force-restore method. Turbulent heat fluxes are parameterized by gradient laws using the resistance concept. The formula for the canopy surface resistance involves both a parameter describing atmospheric demand and one describing moisture availability. Soil moisture prediction is made with Richards' equation. PROGSURF is tested in off-line mode for the Cabauw data set. The observed annual mean values of the state and flux quantities at the earth's surface are well reproduced. For example, the model yields latent and sensible heat fluxes of −35.3 and −2.4 W/m², respectively; evapotranspiration and runoff is −449 and 326 mm/yr; and root zone soil moisture content is 0.344 m³/m³. Further, the seasonal changes of water and energy balance components are well simulated. The sensitivity of PROGSURF to the canopy resistance formulation is analysed. We find that the atmospheric demand is largely represented by the saturation value of the evapotranspiration/soil moisture curve with maximum summer impact upon the annual value and further that the moisture availability is represented by the slope of the evapotranspiration curve. Both saturation value and slope control the amplitude of the seasonal fluctuation of the water balance components; at Cabauw site the saturation value is the governing parameter. These results fit satisfactorily into the other PILPS models. In particular, we are able to reproduce with PROGSURF the total variability of most other PILPS models by simply changing the atmospheric demand and soil moisture availability parameters. PROGSURF presently serves to simulate observed surface fluxes for an atmospheric diagnostic model.