Agricultural enterprises growing crops under irrigation are promoted in the Czech Republic since 1974 by so-called Agrometservis program providing weekly. information on potential water deficits under main crops. The program was prepared at the previous Research Institute of Fundamental Agrotechnology (VUZA) at Hrugovany u Brna and it is operated by the Czech Hydromcteorogical Institute (Ulehla, 1983). This program was later extented by so called SIRHOZ program (Raszka et al., 1987) providing water balance data for individual irrigated fields. Crop evapotranspiration is calculated in both programs using the basic formula Penman (1963) for the-potential evapotranspiration of short grass well-supllied with water'', which is corrected by taking in account the temperature dependent growth of LAI of the respective crop, as well as the respective crop and soil coefficients. The soil moisture model is similar to those applied elsewhere (Ritchic, 1972, for example). Models of both potential and actual crop evapotranspirations were also applied in the draft of a standard for estimating water requierements for crop irrigation at the stage of irrigation systems planning. However, it appeared suitable to update the basic approach and to compare the VUZA model. already being 15 years in use, with some newer ones. We report here mainly on results obtained with the Meteorogical Office Rainfall and Evapotranspiration Calculation System (MORECS: Thompson et al.. 1981). Soil moisture estimates calculated according to both VUZA and MORECS models were compared with actual soil moisture data which were obtained at Hrusovany during 1978-1982 by Odlozilik (1983). Soil moisture was estimated by weighing fresh and dried soil samples taken by 10 cm to the depth of 0.6 m in winter wheat stands (after forecrops winter wheat and maize), spring barley and lucerne. Since both models calculated the soil moisture in this case for the soil layer of 1 m, the actual soil moisture measured at 0.5 to 0.6 m was supposed to represent the 0.5 to 1.0 m layer. Soil water balance was calculated in daily steps from January, starting with the actually measured soil moisture. The comparison of calculated and measured soil moisture data has shown that the VUZA model proposed for estimating the irrigation water requirements of crops underestimates the evapotranspiration in comparison with the MORECS model. The original Penman formula was found previously (Winkler, 1990) to give higher values of potential evapotranspiration than the Penman-Monteith modification (used in the MORECS model) at a crop resistance of 40 s/m. This shows that the discrepancy between the two models is not due to differences in the computation of potential evapotranspiration proper, which would be particularly relevant for the evapotranspiration estimates of closed crop stands. The situation is different at low L-41 values, where the calculated evaporation from barc soil is the main component of the evaporative water loss. The MORECS model uses for the evaporation from wet bare soil the surface resistance of 100 s/m, The evaporation from bare soil calculated in this way corresponds under our conditions approximately to the potential evapotranspiration according to the original Penman formula multiplied by a ''crop'' (soil coefficient of 0.8). This suggests that the coefficient of the value 0.4 used in the VUZA model for estimating the average evaporation from bare soil is too low. and that it underestimates the actual evaporation parcticularly during and immediately after rainfall. This is also why a more realistic two-layer model for bare soil evaporation was compiled and tested at VUZA by U l c h l a (1978). The possibility that the coefficient 0.4 is too low appears to be corroborated by the observation that the underestimation of evapotranspiration for April and May is more pronounced with barley that with winter wheat and lucerne, which compares with differences in the bare soil exposure. However, the main difference between the VUZA and MORECS models has to be looked for in the conception of the soil model. After rainfall. the VUZA model reduces at first the soil moisture deficit and that calculates the actual evapotranspiration as a function of potential evapotranspiration and the new soil moisture. On the contrary. in the MORECS program the evapotranspiration after rainfall goes on at a potential rate. The actual situation is simplified in both cases. since neither model takes into account such characteristics as the retention and infiltration curves of the soil, the simplification being necessary for practical reasons. Nevertheless. the MORECS model gives better results for the soil at Hrusovany, which has a relatively high usable water capacity, different results may be expected on light textured soils with a high rate of infiltration. The soil moisture calculated according to the MORECS model was higher than the actually measured in all five years with spring barley, in four years out of five with winter wheat, and in three years out of five with lucerne. The results also suggest that the crops are able to transpire at a high rate during certain periods in spite of the soil moisture falling close to the wilting point.
