@article{dukes_evans_gilliam_kunickis_2003, title={Interactive effects of controlled drainage and riparian buffers on shallow groundwater qaulity}, volume={129}, DOI={10.1061/(ASCE)0733-9437(2003)129:2(82)}, abstractNote={As a result of recent surface water quality problems in North Carolina, riparian buffers and controlled drainage are being used to reduce the loss of nonpoint source nitrogen from agricultural fields. The effect of controlled drainage and riparian buffers as best management practices to reduce the loss of agricultural nonpoint source nitrogen from the middle coastal plain has not been well documented. The middle coastal plain is characterized by intensive agriculture on sandy soils with deeply incised or channelized streams. A 2-year study was conducted to determine the effectiveness of controlled drainage, riparian buffers, and a combination of both in the middle coastal plain of North Carolina. It was hypothesized that raising the water table near the ditch would enhance nitrate-nitrogen reduction through denitrification. On the sandy soils studied, controlled drainage did not effectively raise the water table near the ditch to a greater degree than observed on the free drainage treatment. Due to random treatment location, the free drainage treatment was installed along a ditch with a shallower impermeable layer compared to the impermeable layer on the controlled drainage treatments (2 m versus 3- to 4-m deep). This resulted in a perched or higher water table on the free drainage treatment. Over 17 storm events, the riparian buffer (free drainage) treatment had an average groundwater table depth of 0.92 m compared to 0.96 and 1.45 m for the combination (riparian buffer and controlled drainage) and controlled drainage treatments, respectively. Nitrate concentration decrease between the field wells and ditch edge wells averaged 29% (buffer only), 63% (buffer and controlled drainage), and 73% (controlled drainage only). Although apparently more nitrate was removed from the groundwater on the controlled drainage treatments, the controlled drainage treatment water table near the ditch was not raised closer to the ground surface compared to the free drainage treatment. Nitrate removal effectiveness was attributed to local soil and landscape properties, such as denitrification in deeper reduced zones of the soil profile.}, number={2}, journal={Journal of Irrigation and Drainage Engineering}, author={Dukes, M. D. and Evans, R. O. and Gilliam, J. W. and Kunickis, S. H.}, year={2003}, pages={82–92} }
 @article{dukes_evans_gilliam_kunickis_2002, title={Effect of riparian buffer width and vegetation type on shallow groundwater quality in the Middle Coastal Plain of North Carolina}, volume={45}, DOI={10.13031/2013.8528}, abstractNote={The effect of riparian buffer width and vegetation type on shallow groundwater quality has not been evaluated 
in the Middle Coastal Plain of North Carolina. Four riparian buffer vegetation types and no–buffer (no–till corn and rye 
rotation or pasture) were established at 8 and 15 m widths as follows: cool season grass (fescue), deep–rooted grass (switch 
grass), forest (pine and mixed hardwood), and native vegetation. Nested groundwater monitoring wells were installed at the 
field/buffer edge and the stream edge in the middle of each riparian buffer plot at three depths. Most deep, mid–depth, and 
shallow wells were 3.0 m, 1.8 m, and 0.6 m deep from the ground surface to the top of the 0.6 m perforated section, respectively. 
Wells were sampled for 23 months beginning July 1998. Although the ditch well nitrate–nitrogen concentrations at the middle 
well depth were significantly lower in the 15 m wide plots compared to the 8 m plots over half the monitoring period, extreme 
flooding as a result of a hurricane in the middle of the study confounded the results. The effect of vegetation was not significant 
at any time, including the no–buffer cropped and fertilized plots. The effect of vegetation was minimized because at the early 
stage in the buffer vegetation establishment, vegetative cover and root mass were not fully developed, the hurricane–induced 
flooding forced the re–establishment of several vegetation types (forest and fescue), and there was likely some mixing of 
groundwater flowing toward the vegetation plots. Establishment of buffers along streams where groundwater flowed away 
from the stream did not result in lower groundwater nitrate levels.}, number={2}, journal={Transactions of the ASAE}, author={Dukes, M. D. and Evans, R. O. and Gilliam, J. W. and Kunickis, S. H.}, year={2002}, pages={327–336} }