Modeling nitrate leaching with a biogeochemical model modified based on observations in a row-crop field in Iowa

Abstract

Prediction of nitrate leaching from cropland is crucial for preventing surface or ground water degradation. Accurate modeling of nitrate leaching requires simulations of both soil hydrological and biogeochemical processes. This paper reports an attempt to improve an existing biogeochemical model, Denitrification–Decomposition or DNDC, for estimation of nitrate leaching from crop fields with tile drainage system. DNDC was equipped with detailed biogeochemical processes of nitrogen turnover but a simple module for one-dimensional movement of soil water. Observations from nine drainage tiles with three different fertilizer treatments in 4 years (1996–1999) at an experimental field in Iowa were used for model modifications. Preliminary comparisons with observed tile discharge flow indicated that the original DNDC lacked the water leaching recession character. To correct this deviation, new water retention features were added to DNDC by: (1) adopting a recession curve to regulate the gravity drainage flow in the explicitly simulated soil profile (0–50 cm) and (2) introducing a virtual water pool for the space between the bottom of the modeled soil profile (50 cm) and the tile lines depth of placement (145 cm) to control the tile discharge flow. With these modifications, model prediction of water leaching fluxes from the tile drainage lines was improved. An adsorbed N pool was created in DNDC to simulate the buffering effect of soil on the amount of nitrate available for leaching. The Langmuir equation was adopted to simulate adsorption and desorption of ammonium ions on the soil absorbents. This modification enhanced the model capacity for simulating free ammonium dynamics, nitrification, and nitrate leaching. Sensitivity tests of the modified DNDC showed that the modeled impact of differences of precipitation, soil texture, soil organic carbon content, and fertilizer application rates on nitrate leaching rates were consistent with observations reported by other researchers. This study indicated that a biogeochemical model with limited modifications in hydrology could serve nitrate leaching prediction and be useful for sustainable agricultural management.

Introduction

Modern agricultural practices are strongly linked to fertilizer application for maintaining optimum yields. However, inefficient fertilizer use has led to a significant portion of the nitrogen (N) applied to farm fields reaching surface or ground water systems (Karlen et al., 1998, Tilman et al., 2001). Surface and ground water quality in the Midwestern Corn Belt, the most productive region in the US, has been negatively affected by the increasing application of N fertilizers, particularly from 1950 to 1980s (Burkart and James, 1999). In this region it is common to have subsurface artificial drainage systems installed to improve the soil moisture conditions to allow for row-cropping operations, and hence create channels to lead the excess soil water rapidly into the surface water bodies. The intensive row-crop agricultural practices in the Midwestern Corn Belt are a significant source of N contamination of water resources in this region (Keeney and DeLuca, 1993, Cambardella et al., 1999, Jaynes et al., 2001, Dinnes et al., 2002). It is an increasing challenge to sustain agricultural production and the environmental quality by adopting best management practices including those for N management (Dinnes et al., 2002).

Nitrate (NO₃⁻) leaching from row-crop fields is directly controlled by water discharge flow and residual soil-NO₃⁻ that is at risk for leaching, which are affected by numerous factors such as climate conditions, soil properties, and management (e.g., tillage, fertilization, irrigation, manure application, crop rotation, etc.). A large number of experiments have been conducted to observe the correlation between NO₃⁻ leaching and the environmental or management factors. However, with limited time and funding for field experiments, estimation of NO₃⁻ leaching, especially at regional scale, has to rely on mathematical models. Some of the models, such as MIKE SHE (DHI, 1999) and MODFLOW (Harbaugh et al., 2000) are hydrology-oriented with less details about N biogeochemical processes; and some, such as CENTURY and SOILN (Liu et al., 2000, Johnsson et al., 1987), have N turnover functions but with marginal hydrological features.

The goal of our project was to merge the two kinds of models for improving our modeling reliability. This paper reports how we adapted a biogeochemical model with limited modifications to serve as a NO₃⁻ leaching prediction tool that can be used for farm management planning.