Modeling the effect of soil structure on water flow and isoproturon dynamics in an agricultural field receiving...

Transport processes in soils are strongly affected by heterogeneity of soil hydraulic properties. Tillage practices and compost amendments can modify soil structure and create heterogeneity at the local scale within agricultural fields. The long-term field experiment QualiAgro (INRA-Veolia partnership 1998-2013) explores the impact of heterogeneity in soil structure created by tillage practices and compost application on transport processes. A modeling study was performed to evaluate how the presence of heterogeneity due to soil tillage and compost application affects water flow and pesticide dynamics in soil during a long-term period. The study was done on a plot receiving a co-compost of green wastes and sewage sludge (SGW) applied once every 2 years since 1998. The plot was cultivated with a biannual rotation of winter wheat-maize (except 1 year of barley) and a four-furrow moldboard plow was used for tillage. In each plot, wick lysimeter outflow and TDR probe data were collected at different depths from 2004, while tensiometer measurements were also conducted during 2007/2008. Isoproturon concentration was measured in lysimeter outflow since 2004. Detailed profile description was used to locate different soil structures in the profile, which was then implemented in the HYDRUS-2D model. Four zones were identified in the plowed layer: compacted clods with no visible macropores (Δ), non-compacted soil with visible macroporosity (Γ), interfurrows created by moldboard plowing containing crop residues and applied compost (IF), and the plow pan (PP) created by plowing repeatedly to the same depth. Isoproturon retention and degradation parameters were estimated from laboratory batch sorption and incubation experiments, respectively, for each structure independently. Water retention parameters were estimated from pressure plate laboratory measurements and hydraulic conductivity parameters were obtained from field tension infiltrometer experiments. Soil hydraulic properties were optimized on one calibration year (2007/08) using pressure head, water content and lysimeter outflow data, and then tested on the whole 2004/2010 period. Lysimeter outflow and water content dynamics in the soil profile were correctly described for the whole period (model efficiency coefficient: 0.99) after some correction of LAI estimates for wheat (2005/06) and barley (2006/07). Using laboratory-measured degradation rates and assuming degradation only in the liquid phase caused large overestimation of simulated isoproturon losses in lysimeter outflow. A proper order of magnitude of isoproturon losses was obtained after considering that degradation occurred in solid (sorbed) phase at a rate 75% of that in liquid phase. Isoproturon concentrations were found to be highly sensitive to degradation rates. Neither the laboratory-measured isoproturon fate parameters nor the independently-derived soil hydraulic parameters could describe the actual multiannual field dynamics of water and isoproturon without calibration. However, once calibrated on a limited period of time (9 months), HYDRUS-2D was able to simulate the whole 6-year time series with good accuracy.