Search

Experimental approaches for directly determining preferential water flow in field soils usually require artificial interruption of the soil matrix's water flow. This interruption may change the quantitative as well as the qualitative nature of the preferential water flow, particularly in soils such as Andisols where matrix water flow is the dominant process of drainage. To overcome this problem, we applied both the one-dimensional form of Darcy's equation and the soil water balance method, during and shortly after each rain event, for characterizing as well as determining the preferential water flow across a depth of 1 m in an Andisol. The volumetric water contents and pressure potentials at different soil depths were monitored by time domain reflectometry and tensiometry during a 7-yr period. Preferential water flow was detected only two to seven times per year; nevertheless, it accounted for 16 to 27% of the annual total drainage. These flows occurred under macroscopically homogeneous soil water conditions, with pressure potentials mostly below the air-entry pressure in the subsoil. The in situ unsaturated hydraulic conductivity at the depth of 1 m, determined to apparently explain the total drainage during a preferential water flow event, varied over more than one order of magnitude, with little change in water content, and the maximum value reached the saturated hydraulic conductivity. This indicates that preferential water flow occurred mainly through the largest macropores at this depth due to pore-scale physical nonequilibrium. Macropore-mediated preferential water flow and solute transport should be considered in unsaturated subsoil of Andisols. Abbreviations: TDR, time domain reflectometry.

Andosols, distributed widely around the Pacific basin, have unique soil–water and solute transport properties because of their stepwise water retention curves and high anion-adsorption capacity. The model modification and verification for these properties are crucial for evaluating the potential for improved agricultural management (e.g., using organic matter instead of inorganic fertilizer) to reduce N loss from the soils. Here, we improved an existing biogeochemical model, LEACHM, to predict long-term N leaching from Andosols amended with composted manure, without optimization to fit measured field data. The modified model was verified by observations from a 5.6-year lysimeter experiment with different rates of inorganic N fertilizer plus composted manure (100 + 0, 75 + 25, or 25 + 75%) of two different types (cattle, swine) on lettuce, sorghum (as a catch crop), and Chinese cabbage in rotation. Incorporation of Durner’s bimodal model dramatically improved predictions of drainage water volume and evapotranspiration. The non-linear Langmuir adsorption isotherms for soil NH4+ and NO3− improved model performance in simulating crop N uptake and N leaching loss. The RMSE, R2, and index of Agreement were evaluated as satisfactory in all lysimeters. Our model explained reasonably well that improved agricultural management decreased in current available N addition rates by 8.82–35.6% and reduced in the yearly averaged NO3 leaching by 8.70–41.8%. A modified model relating soil hydraulic properties and N adsorption properties could thus accurately predict N leaching under different long-term N application rates/types, and could be useful for supporting agricultural management decisions in cropped Andosols. [ABSTRACT FROM AUTHOR]

The use of N fertilizers for agricultural production acts as a sources of atmospheric $$\hbox {N}_2\hbox {O}$$ . Fertilizer induced $$\hbox {N}_2\hbox {O}$$ emission considerably varies in accordance with environmental factors. We conducted a $$\hbox {N}_2\hbox {O}$$ flux measurement campaign across 10 different experimental sites with various soil types throughout Japan and investigated the fertilizer induced $$\hbox {N}_2\hbox {O}$$ emission factors (FIEFs) of synthetic fertilizers (mainly urea) in multiple growing periods at each experimental site. FIEFs considerably varied among the 10 sites, and measurement periods ranged from 0.00 to 7.13 % in 40 total observations. Soil profile information divided the experimental sites into two major groups through cluster analysis, which are volcanic or non-volcanic soils. According to this classification, FIEFs were clearly differentiated into high and low FIEF groups (mean values: 2.67 and 0.59 %, respectively). Regression trees selected total soil carbon, the depth to the Fe mottling horizon, and mean air temperature as the key parameters to determine the strength of FIEFs out of 10 explanatory variables (e.g., N fertilizer application rates, fertilizer application times, clay content, total precipitation during measurement periods, pH, total soil carbon, and total soil nitrogen). The existence of shallow Fe mottling layer (up to 42.5 cm) induced high FIEFs in this model, suggesting that upward $$\hbox {N}_2\hbox {O}$$ emissions derived from the intermediate Fe mottling layers presumably contributed high $$\hbox {N}_2\hbox {O}$$ emissions in such soils. Our results suggest that the soil profile information associated with the water regime is an important index for synthetic fertilizer induced $$\hbox {N}_2\hbox {O}$$ emissions. [ABSTRACT FROM AUTHOR]