|Senaviratne, G.m.m.m. - UNIVERSITY OF MISSOURI|
|Lory, John - UNIVERSITY OF MISSOURI|
|Udawatta, Ranjith - UNIVERSITY OF MISSOURI|
|Nelson, Nathan - KANSAS STATE UNIVERSITY|
|Williams, Jimmy - TEXAS A&M UNIVERSITY|
|Anderson, Stephen - UNIVERSITY OF MISSOURI|
Submitted to: Transactions of the ASABE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 12/18/2017
Publication Date: 4/24/2018
Citation: Senaviratne, G.A., Baffaut, C., Lory, J.A., Udawatta, R.P., Nelson, N.O., Williams, J.R., Anderson, S.H. 2018. Improved APEX model simulation of buffer water quality benefits at field scale. Transactions of the ASABE. 61(2):603-616. doi:10.13031/trans.12655.
DOI: https://doi.org/10.13031/trans.12655 Interpretive Summary: Collecting flow and water quality data to estimate how much buffer strips reduce sediment and nutrient transport from agricultural fields is long and expensive. Computer simulation models like the Agricultural Policy Environmental eXtender (APEX) offer alternatives. However, verification that these models simulate the effects of buffers correctly is needed. This study utilizes 16 years of monitoring data from three field-scale row crop watersheds. Two watersheds had buffers installed in fall of 1997, one with upland agroforestry buffers (grass plus trees), and the second with upland grass buffers; the third watershed was maintained as a control. The objectives of this study were to: i) determine the effect of upland buffers on APEX parameters, ii) evaluate the ability of APEX to estimate upland buffer effectiveness, and iii) evaluate the impact of alternative buffer placement options on water quality outcomes. Runoff and total phosphorus were well simulated before and after buffer implementation. Measured and simulated sediment yields had a similar range of values but lack of large sediment loss events caused uncertainties in the values of the sediment related input parameters. Defining additional areas in order to simulate buffers had an impact on the results that was as large as the buffer vegetation for runoff, and half as large for TP. These effects need to be considered when estimating buffer effectiveness by adding or removing buffers to a calibrated model. These results should help scientists and conservation planners interpret model results on effectiveness of management practices that require definition of smaller subareas.
Technical Abstract: Watershed models offer cost-effective means to quantify environmental benefits of conservation practices. This study evaluated the ability of the Agricultural Policy Environmental eXtender (APEX) to simulate upland buffer effectiveness for reducing event runoff, sediment and total phosphorus (TP) loadings. The study used 16 years of monitoring data (1993-2008) from three field-scale row crop watersheds. Two watersheds had buffers installed in fall of 1997, one with upland agroforestry buffers (grass plus trees), and the second with upland grass buffers; the third watershed was maintained as a control. Objectives were to: i) determine the effect of upland buffers on APEX parameters, ii) evaluate the ability of APEX to estimate upland buffer effectiveness, and iii) evaluate the impact of alternative buffer placement options on water quality outcomes. After modification of the APEX code to improve simulation of infiltration in the buffers, we successfully calibrated APEX for no-buffer and upland buffer conditions for event runoff and TP. However calibrated parameters from the no-buffer model simulated buffer conditions poorly and resulted in an over-estimation of buffer effectiveness. Buffer effectiveness estimated with upland buffer parameters was similar to that obtained from measured data. Scenario analysis indicated that a combination of backslope and footslope agroforestry buffers was more effective than contour placement for reducing average annual edge-of-field runoff and total P losses. These results highlight the complexity of using APEX to assess upland buffers and highlight potential problems using APEX to quantify the benefits of conservation practices not included in the calibration dataset.