EVAPORATION AND DISCHARGE FROM ARABLE LAND WITH CROPPED OR BARE SOILSDURING WINTER - MEASUREMENTS AND SIMULATIONS

Evaporation from cropping systems with and without crop cover during w inter (spring cereals with and without undersown Italian rye grass) wa s studied. Measured daily discharge from field lysimeters and tile-dra ined plots was used to investigate to what extent discharge measuremen ts could reveal differences in evaporation between the cropping system s. The identified differences were interpreted using a physically base d model for simulating water and heat flows in the soil. Measurements were made between 1 April 1988 and 1 April 1991 on a sandy loam in sou thwestern Sweden. Annual precipitation (1 April-31 March) was 910 mm, 697 mm and 677 mm respectively. Total annual discharge showed a large degree of inconsistent variation between field plots within the same y ear and treatment, whereas the discharge dynamics showed more consiste nt differences between treatments. Standard meteorological variables a nd data on soil properties and crop development were used as input to the model. Calculations of soil evaporation, transpiration and evapora tion of intercepted water were based on the Penman-Monteith equation. The model was calibrated against the actual date on which discharge st arted, which consistently occurred later in the autumn in the cropped soil than in the bare soil. Simulated annual total evaporation amounte d to 505 mm, 470 mm and 396 mm from the system with bare soil during w inter, and 552 mm, 510 mm and 423 mm from the system with cropped soil during winter. However, during periods with low temperatures and freq uent precipitation, total evaporation rates for the bare soil were as high as or even exceeded those for cropped soil. Simulated annual soil evaporation from the system with cropped soil during winter constitut ed about 38% of the total annual evaporation, whereas it varied betwee n 55 and 65% for the system with bare soil during winter. Discrepancie s between simulated and measured discharge could have been due to spat ial variability in soil and plant properties in the field, as well as to errors in the estimated unsaturated hydraulic conductivities or roo t water uptake functions used in the simulations. In simulations based on either modified soil properties or root depths, similar changes in water dynamics were obtained, demonstrating the need for independent determinations of the soil hydraulic properties.