U.S. flag

An official website of the United States government

Dot gov

Official websites use .gov
A .gov website belongs to an official government organization in the United States.


Secure .gov websites use HTTPS
A lock ( ) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.


Main content area

Development and testing of watershed-scale models for poorly drained soils

Fernandez, G.P., Chescheir, G.M., Skaggs, R.W., Amatya, D.M.
Transactions of the ASAE 2005 v.48 no.2 pp. 639-652
water pollution, simulation models, water quality, forested watersheds, nitrate nitrogen, streams, coastal plain soils, poorly drained soils, pollution load, subsurface drainage, hydrologic models, North Carolina
Two watershed-scale hydrology and water quality models were used to evaluate the cumulative impacts of land use and management practices on downstream hydrology and nitrogen loading of poorly drained watersheds. Field-scale hydrology and nutrient dynamics are predicted by DRAINMOD in both models. In the first model (DRAINMOD-DUFLOW), field-scale predictions are coupled to the canal/stream routing and in-stream water quality model DUFLOW, which handles flow routing and nutrient transport and transformation in the drainage canal/stream network. In the second model (DRAINMOD-W), DRAINMOD was integrated with a new one-dimensional canal and water quality model. The hydrology and hydraulic routing components of the models were tested using data from a 2950 ha drained managed forest watershed in the coastal plain of eastern North Carolina. Both models simulated the hydrology and nitrate-nitrogen (NO3-N) loading of the watershed acceptably. Simulated outflows and NO3-N loads at the outlet of the watershed were in good agreement with the temporal trend for five years of observed data. Over a five-year period, total outflow was within 1% of the measured value. Similarly, NO3-N load predictions were within 1% of the measured load. Predictions of the two models were not statistically different at the 5% level of significance.