The Curve Number (CN) method is used in many models to predict surface runoff depth and transport of dissolved agrochemicals. CNs were determined on 70 small plots at 8 sites and different crop stages with artificial rain. The measured CNs deviated greatly from the commonly used CNs in most cases. For growing crops, CN correlated closely with cover, regardless of whether the crop was spring or fall barley or rape. The CNs measured with artificial rain agreed well with CNs measured on larger plots with natural rain. A new table was developed that accounts for the resulting seasonal changes in CNs of different small grain crops. The use of this table will greatly improve runoff predictions under German cropping conditions. Predictions will be poor between harvest and subsequent plowing, because of the fast and unpredictable changes in CNs during this generally short period (average CN: 75; standard deviation: 15). On a very stony site, CNs were much lower than would be expected for the hydrological soil group A. If, however, stone cover (23-35%) was included in total cover, the CNs fell into the range of the regression developed for crop cover. In cases where stones are not embedded into a surface seal, but rather protect the soil as would a crop or mulch cover, they can similarly reduce runoff. (C) 1996 Academic Press Limited.
