Summary Band 23

Pätzold, Stefan (1998): Herbicide application in horticulture - Investigation and simulation of degradation, sorption, and leaching of simazine and diuron in loess soils. Bonner Bodenkundl. Abh. 23 (1998), 231 pp.


Over three years of field trials degradation, sorption and leaching of the herbicides simazine and diuron as well as of the diuron degradation product DCPMU were investigated in orchard soils after applying the recommended amounts with a common spraying device. Most of the research was conducted in a 12 year old apple orchard on a Haplic Luvisol near Bonn, FRG. Diuron degradation and sorption were also studied in corresponding laboratory experiments. Both adsorbed and dissolved amounts of simazine, diuron and DCPMU were analyzed in order to characterize the herbicide mobility in soil. For the same reason sorption isotherms were derived from batch experiments. Degradation and leaching of simazine and diuron were simulated by using the computer model PELMO which is based on the model PRZM; PELMO is used for registration purposes by german authorities. For six months after applying the herbicides, degradation of the chemicals in the 0-5 cm layer conformed to first-order-kinetics. The DT50-values ranged from 20 to 64 days for simazine and from 17 to 30 days for diuron. As the degradation rate gradually decreased, a carry-over of herbicide residues from the preceeding growing season occured at the time of reapplication in the following year. The long-term development of the herbicide residue level was best characterized by two combined exponential functions representing fast and slow degradation. The formation of DCPMU was observed in the laboratory and field investigations. Both soil saturating extracts and CaCl2-extracts were equally suitable for the determination of simazine and diuron mobility. Under natural conditions, mobility of herbicide residues decreased significantly with time. In sterile soil samples the decrease of diuron mobility diminished after 28 days indicating the importance of simultaneous microbial degradation and diffusion processes for the immobilisation of herbicide residues. Soil moisture revealed significant effects on herbicide mobility and formation of bound residues. Up to 40 % of the previously unextractable residues were remobilised after drying the moist soil. In general, sorption and desorption followed Freundlich-isotherms. Sorption and desorption phenomena at low concentrations were partially best represented by a separate Freundlich-isotherm. This was caused by sorption of small herbicide amounts on inner surfaces of aggregates and organic matter; slow desorption kinetics from these sites was found to be the reason for the observed carry-over phenomena. Formulation additives enhanced simazine and diuron mobility when high concentrations of the herbicides were added in the batch experiments. In autumn only low concentrations of the simazine and diuron applied in the previous spring remained in the upper soil layers. Consequently, only small amounts of these strongly adsorbed residues were leached into the subsoil due to convective-dispersive flow during the winter months. In contrast, a few days after the herbicide application in May 1994, while high herbicide concentrations occured in the 0-5 cm soil layer, residues of both chemicals were translocated into the subsoil. This rapid movement was due to preferential flow in soil macropores like earthworm channels and shrinkage cracks. The importance of preferential flow was proved by conducting a dye tracer experiment. Small amounts of simazine and diuron were detected in the groundwater. The simulation model PELMO was capable of predicting degradation and leaching of simazine and diuron, as far as no preferential flow occured. PELMO does not account for this process. The degradation and sorption parameters commonly used in simulation for registration purposes led to less realistic simulation results than optimized parameters for the characterization of the chemicals. In particular, the kinetics of herbicide degradation has to be taken into account. The description of herbicide mobility by time-dependent partition coefficients improved simulation results as compared to Freundlich-parameters determined in batch experiments.