Top Navigation Bar

Article Request Page ASABE Journal Article

Assessing the Impacts of Future Climate Conditions on the Effectiveness of Winter Cover Crops in Reducing Nitrate Loads into the Chesapeake Bay Watersheds Using the SWAT Model

S. Lee, A. M. Sadeghi, I.-Y. Yeo, G. W. McCarty, W. D. Hively


Published in Transactions of the ASABE 60(6): 1939-1955 (doi: 10.13031/trans.12390). 2017 American Society of Agricultural and Biological Engineers.


Submitted for review in April 2017 as manuscript number NRES 12390; approved for publication as part of the “International Watershed Technology” collection by the Natural Resources & Environmental Systems Community of ASABE in September 2017.

The USDA is an equal opportunity provider and employer. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

The authors are Sangchul Lee, Postdoctoral Associate, Department of Geographical Sciences, University of Maryland, College Park, Maryland, and USDA-ARS Hydrology and Remote Sensing Laboratory, Beltsville, Maryland; Ali M. Sadeghi, ASABE Member, Soil Scientist, USDA-ARS Hydrology and Remote Sensing Laboratory, Beltsville, Maryland; In-Young Yeo, Senior Lecturer, School of Engineering, University of Newcastle, Callaghan, Australia, and Department of Geographical Sciences, University of Maryland, College Park, Maryland; Gregory M. McCarty, Research Soil Scientist, USDA-ARS Hydrology and Remote Sensing Laboratory, Beltsville, Maryland; W. Dean Hively, Research Physical Scientist, USGS Eastern Geographic Science Center, Reston, Virginia. Corresponding author: Sangchul Lee, 10300 Baltimore Ave., BARC-West Bldg. 007, Beltsville, MD 20705; phone: 443-462-2927; e-mail: sangchul.lee84@gmail.com.

Abstract Winter cover crops (WCCs) have been widely implemented in the Coastal Plain of the Chesapeake Bay Watershed (CBW) due to their high effectiveness in reducing nitrate loads. However, future climate conditions (FCCs) are expected to exacerbate water quality degradation in the CBW by increasing nitrate loads from agriculture. Accordingly, the question remains whether WCCs are sufficient to mitigate increased nutrient loads caused by FCCs. In this study, we assessed the impacts of FCCs on WCC nitrate reduction efficiency in the Coastal Plain of the CBW using the Soil and Water Assessment Tool (SWAT). Three FCC scenarios (2085-2098) were prepared using general circulation models (GCMs), considering three Intergovernmental Panel on Climate Change (IPCC) Special Report on Emissions Scenarios (SRES) greenhouse gas emission scenarios. We also developed six representative WCC implementation scenarios based on the most commonly used planting dates and species of WCCs in this region. Simulation results showed that WCC biomass increased by ~58% under FCC scenarios due to climate conditions conducive to WCC growth. Prior to implementing WCCs, annual nitrate loads increased by ~43% under FCC scenarios compared to the baseline scenario (2001-2014). When WCCs were planted, annual nitrate loads were substantially reduced by ~48%, and WCC nitrate reduction efficiency was ~5% higher under FCC scenarios relative to the baseline scenario. The increase in WCC nitrate reduction efficiency varied with FCC scenario and WCC planting method. As CO2 concentrations were higher and winters were warmer under FCC scenarios, WCCs had greater biomass and thus demonstrated higher nitrate reduction efficiency. In response to FCC scenarios, the performance of less effective WCC practices (i.e., barley, wheat, and late planting) under the baseline scenario indicated a ~14% higher increase in nitrate reduction efficiency compared to WCC practices with greater effectiveness under the baseline scenario (i.e., rye and early planting) due to warmer temperatures. The SWAT simulation results indicated that WCCs were effective in mitigating nitrate loads accelerated by FCCs, suggesting the role of WCCs in mitigating nitrate loads will likely be even more important under FCCs.

Keywords.Future climate conditions (FCCs), SWAT, Water quality, Winter cover crops (WCCs).

The Chesapeake Bay (CB) is one of the most valuable natural resources in the U.S. Its drainage areas of 166,000 km2, known as the Chesapeake Bay watershed (CBW), is the residence area for nearly 18 million people and 3,600 species of plants and animals (CBP, 2008). Water quality degradation caused by excessive nutrient loads from agricultural activities and point sources has been of great concern in this region (NRC, 2011). Agricultural land covers only 25% of the CBW but has been reported to be responsible for 42% and 46% of total nitrogen (N) and phosphorus (P) loads, respectively, into the estuary (CBP, 2008). Nutrient loads have been shown to increase substantially during winter seasons in this region due to low evapotranspiration (ET), higher water table, and fallow cropland (Fisher et al., 2010; Yeo et al., 2014). Future climate conditions (FCCs; elevated CO2 concentration, temperature and precipitation changes) are expected to affect the hydrology and N cycle in the CB, which potentially poses a threat to the health of the bay (Najjar et al., 2009, 2010; Lee et al., 2015, 2017a). For example, Lee et al. (2017a) predicted an increase in nitrate loads of ~39% compared to the baseline condition in the Coastal Plain of the CBW. Hence, there is a question of whether current conservation practices implemented in the CBW are sufficient to mitigate the increased nutrient loads caused by FCCs.

Numerous agricultural best management practices (BMPs) have been emphasized throughout the CBW to improve water quality (McCarty et al., 2008). For example, winter cover crops (WCCs) are recommended as a cost-effective agricultural BMP for reducing nitrate loads (McCarty et al., 2008; Ator and Denver, 2012). Due to the high efficiency of WCCs in reducing nitrate loads, both Federal and state governments provide cost-sharing programs to encourage local farmers to adopt WCCs (McCarty et al., 2008). Normally, WCCs are planted during winter seasons, between harvest of summer crops late in the fall and planting of summer crops in the following spring, to immobilize leftover nitrate in the soil profile (Hively et al., 2009). The effectiveness of WCCs in reducing nitrate in this region has been investigated in several studies. The nitrate reduction efficiency has been shown to vary by implementation method, watershed characteristics, and crop rotation (Hively et al., 2009; Yeo et al., 2014; Lee et al., 2016). For instance, previous studies discovered that early-planted WCCs are more effective in reducing nitrate loads than late-planted WCCs due to warmer conditions and longer growing periods (Hively et al., 2009; Yeo et al., 2014; Lee et al., 2016). In addition, rye cover crops were found to outperform barley and wheat cover crops in reduction of nitrate loads, most likely due to their hardiness and more rapid root system development during the winter (Hively et al., 2009; Staver et al., 1989; Prabhakara et al., 2015).

FCCs are expected to affect WCC growth during winter seasons. For example, an increase in precipitation would reduce crop water stress under drought conditions (Zeppel et al., 2014). An increase in temperature would enhance crop growth by reducing frost damage and providing warmer conditions. Elevated CO2 concentrations could also enhance plant growth by increasing water use efficiency (Kimball and Idso, 1983) and stimulating photosynthesis (Parry et al., 2004). These FCCs would likely facilitate WCC biomass growth and therefore nitrate reduction efficiency. WCC nitrate reduction efficiency is known to correlate strongly with WCC biomass because soil nitrate absorbed by WCCs is converted to crop biomass (Hively et al., 2009). However, WCCs might not be effective in coping with increased nitrate loads caused by FCCs. Woznicki and Nejadhashemi (2012) reported that the performances of agricultural BMPs were highly sensitive to climate variability. In addition, little is known about WCC nitrate reduction efficiency under FCCs. Therefore, it is important to examine the effectiveness of current WCC practices in response to FCC variables.

The objective of this study was to assess the impacts of FCCs on WCC nitrate reduction efficiency in the Coastal Plain of the CBW using the Soil and Water Assessment Tool (SWAT). Specifically, we analyzed how future WCC nitrate reduction efficiency may vary under different FCC scenarios and planting methods. For comparison with the baseline scenario (2001-2014), we prepared three FCC scenarios (2085-2098) using data from eight global circulation models (GCMs) under three Intergovernmental Panel on Climate Change (IPCC) Special Report on Emissions Scenarios (SRES) greenhouse gas emission scenarios (A1B, A2, and B1). Six representative WCC implementation scenarios were also developed based on the most commonly used planting dates and species of WCCs in this region. First, because WCC nitrate reduction efficiency is highly dependent on WCC biomass, WCC biomass growth was compared between the baseline and FCC scenarios. A comparison of the nitrate budget between the presence and absence of WCCs was then made for the baseline and FCC scenarios to quantify the effectiveness of WCCs in response to FCCs. Note that this study focused primarily on comparing WCC nitrate reduction efficiency between two periods (i.e., baseline and FCC scenarios). The spatial variability of WCC nitrate reduction efficiency was not taken into account.

Materials and Methods

The two subwatersheds, Tuckahoe Creek watershed (TCW) and Greensboro watershed (GW), occupy drainage areas of approximately 220.7 and 290.1 km2, respectively, within the upper region of the Choptank River Watershed in the CBW (fig. 1). The Choptank River Watershed has been widely investigated as one of the benchmark watersheds of the USDA Conservation Effects Assessment Project (CEAP) (McCarty et al. 2008; Richardson et al., 2008; Sharifi et al., 2016). The Choptank River is recognized as an “impaired” water body by the U.S. Environmental Protection Agency under Section 303(d) of the Clean Water Act due to excessive nutrient and sediment loads (McCarty et al., 2008). Contrasting characteristics of land use and soil type exist between TCW and GW (fig. 2). TCW land use is dominated by agriculture (54%), followed by forest (32.8%), pasture (8.4%), urban (4.2%), and water body (0.6%). The soils are 56% well drained (hydrologic soil group (HSG) A and B) and 44% poorly drained (HSG C and D). In contrast, GW land use comprises forest (48.3%), agriculture (36.1%), pasture (9.3%), urban (5.6%), and water body (0.7%). The soils are 75% poorly drained and 25% well drained.

Soil and Water Assessment Tool (SWAT)

SWAT is a process-based, semi-distributed model designed to simulate the impacts of land management on hydrology and nutrient cycles using multiple components, including weather, hydrology, sedimentation, crop growth, nutrients, pesticides, and pathogens (Neitsch et al., 2011). This model has been applied widely to assess the effectiveness of conservation practices and evaluate climate change impacts on watershed physical processes, such as water balance and water quality (Arnold et al., 1998; Gassman et al., 2007). The basic modeling unit is the hydrologic response unit (HRU), which represents a unique combination of land use and soil type within each subwatershed. Hydrologic variables are generated at individual HRUs, aggregated into the subwatershed, and then routed to the watershed outlet. Cropland comprised 312 of the 542 HRUs in the TCW and 431 of 760 HRUs in the GW. The average HRU size was 41 ha (0.01 to 721 ha) for TCW and 38 ha (0.01 to 2134 ha) for GW.

Figure 1. Locations of Tuckahoe Creek watershed (TCW) and Greensboro watershed (GW) (adapted from Lee et al., 2016).

Surface runoff and infiltration are estimated using USDA Soil Conservation Service (SCS) curve number (CN) method. A daily CN value is determined based on soil permeability, land use type, and antecedent soil moisture condition. Infiltrated water in the soil profile is transported to streams via lateral flow or percolated into subsurface groundwater through the vadose zone. Subsurface groundwater is transported to streams via groundwater flow or enters into the deeper groundwater aquifer (fig. 3). The SWAT simulates both inorganic and organic forms of the N cycle (Neitsch et al., 2011). The addition of nitrate in the system is simulated through nitrification, mineralization, and fertilization, and nitrate loss occurs by denitrification and plant uptake. The N components are transported from soils to nearby streams by surface runoff, lateral flow, and groundwater flow (fig. 3). Portions of the N components can also be lost to deep groundwater aquifers.

Plant growth is simulated in SWAT based on the heat unit theory (Neitsch et al., 2011). The model assumes that plant growth takes place when the average ambient temperature is above the base temperature, which varies by crop species. A daily heat unit is computed as the difference between the average ambient temperature and the base temperature. When the cumulative daily heat unit is equal to the plant heat unit, the plant is assumed to reach maturity. At maturity, the plant is simulated to cease absorption of water and nutrients. SWAT first predicts plant growth under optimal conditions representing adequate water and nutrient supply and favorable climate conditions (Neitsch et al., 2011). The actual growth is then estimated with consideration of temperature and the availability of water, N, and P. Water stress occurs when the actual transpiration is not adequate for the plant. Temperature stress occurs when the average ambient temperature is lower than the base temperature or higher than the maximum temperature. A plant experiences N and P stress when the nutrient contents are =50% of optimal values.

In SWAT, the atmospheric CO2 concentration is simulated to affect ET, stomatal conductance, and biomass growth. We used the Penman-Monteith method in SWAT, which considers the plant canopy resistance for the potential evapotranspiration (PET) calculation (Neitsch et al., 2011). The plant canopy resistance is affected by CO2 concentration as follows:

(1)

where rc is the plant canopy resistance, rl is the minimum effective stomatal resistance of a single leaf, and LAI is the leaf area index of the plant canopy. As the CO2 concentration increases, the plant canopy resistance is reduced according to equation 1, subsequently decreasing ET regarding the relationship between ET and plant canopy resistance (Neitsch et al., 2011).

Figure 2. Physical characteristics of TCW and GW: (a) land use and (b) hydrologic soil groups (adapted from Lee et al., 2016).

(2)

where gl,CO2 is the leaf conductance modified to reflect CO2 effects, and gl is the leaf conductance without CO2 effects. Equation 2 simulates the linear reduction in leaf conductance with increasing CO2 and estimates a 40% reduction in conductance for all plants when the CO2 concentration is doubled (Neitsch et al., 2011). The CO2 concentration also affects plant biomass growth by modifying the radiation use efficiency (RUE), as shown in equations 3 and 4:

(3)

where RUE is radiation-use efficiency, and r1 and r2 are coefficients.

(4)

where ?bio is the potential increase in plant biomass on a given day, and Hphosyn is the amount of intercepted photosynthetically active radiation on a given day.

Baseline SWAT Input Data

The SWAT input data are listed in table 1. The digital elevation model (DEM), land use, and soil maps are required as basic geospatial data. A LiDAR-based DEM was acquired from the Maryland Department of Natural Resources (MD-DNR) and further manipulated by the USDA Agricultural Research Service (ARS) at Beltsville, Maryland, for use for SWAT simulation. The land use map was generated by intersectioning multiple data to accurately delineate the boundaries of each land use type, such as agriculture, forest, urban, etc. (Lee et al., 2016). The soil map was obtained from the USDA Natural Resources Conservation Service (NRCS) Soil Survey Geographic Database (SSURGO). We downloaded daily precipitation and temperature data from the NOAA National Climate Data Center (NCDC) for Chestertown, Royal Oak, and Greensboro (USC00181750, USC00187806, and US1MDCL0009, respectively). Because data were unavailable, humidity, wind speed, and solar radiation were generated using SWAT’s built-in weather generator (Neitsch et al., 2011). Monthly streamflow data collected at the outlets of the two watersheds were provided by two U.S. Geological Survey (USGS) gauge stations: on the Tuckahoe Creek near Ruthsburg (USGS #01491500) for TCW and on the Choptank River near Greensboro (USGS #01491000) for GW (fig. 1). Nitrate grab samples from the Chesapeake Bay Program (CBP, TUK #0181) for TCW and from the USGS (USGS #01491000) for GW were extrapolated into continuous monthly nitrate loads using LOADEST (Runkel et al., 2004).

Figure 3. Schematic of SWAT model components (adapted from Neitsch et al., 2011).
Table 1. SWAT input data.
DataSource[a]DescriptionYears
DEMMD-DNRLiDAR-based 2 m resolution2006
Land useUSDA-NASSCropland data layer (CDL)2008-2012
MRLCNational Land Cover Database (NLCD)2006
USDA-FSA-APFONational Agricultural Imagery Program digital orthophoto quad imagery1998
U.S. Census BureauTIGER road map2010
SoilsUSDA-NRCSSoil Survey Geographical Database (SSURGO)2012
ClimateNCDCDaily precipitation and temperature1999-2014
StreamflowUSGSMonthly streamflow2001-2014
Water qualityUSGS and CBPDaily grab nitrate samples2001-2014

    [a] MD-DNR = Maryland Department of Natural Resources, USDA-NASS = USDA National Agricultural Statistics Service, MRLC = Multi-Resolution Land Characteristics Consortium, USDA-FSA-APFO = USDA Farm Service Agency Aerial Photography Field Office, and USDA-NRCS = USDA Natural Resources Conservation Service.

The NoWCC scenario (i.e., baseline land management) was prepared using multiple geospatial data sources. Major crop rotations for the land management scenario were identified from USDA National Agricultural Statistics Service (NASS) Cropland Data Layer (CDL) data (Lee et al., 2016). Detailed agronomic scheduling and activities were developed from literature review and communication with a local expert (Lee et al., 2016). Six WCC scenarios were prepared based on the guidelines of the Maryland Agriculture Cost Share (MACS) cover crop program, county-level statistics, and WCC expert knowledge (table 2). Refer to Lee et al. (2016) for further details on the land management scenarios. We considered three commonly used cover crop species, i.e., wheat (Triticum aestivum L.), winter barley (Hordeum vulgare L.), and rye (Secale cereale L.), and two planting times, i.e., early (Oct. 1) and late (Nov. 2). The summer cropping (i.e., cash crops including corn and soybean) was assumed to be the same in the NoWCC and WCC scenarios. However, during winter seasons, WCCs were present in the WCC scenarios and absent in the NoWCC scenario. Detailed agricultural management schedules for the NoWCC and WCC scenarios are presented in table 3. The crop rotations and land use types were assumed to not vary over the periods to maintain the same amount of agricultural N input for the baseline and FCC scenarios.

Table 2. Winter cover crop scenarios.
ScenarioSpeciesPlanting Timing
NoWCCNoneN/A
WEWinter wheatEarly planting (Oct. 3)
BEWinter barleyEarly planting (Oct. 3)
RERyeEarly planting (Oct. 3)
WLWinter wheatLate planting (Nov. 2)
BLWinter barleyLate planting (Nov. 2)
RLRyeLate planting (Nov. 2)
Table 3. Management schedules for NoWCC and WCC scenarios (adapted from Lee et al., 2016).
ScenarioCropPlantingFertilizer[a]Harvest
NoWCCCorn (after corn)Apr. 30
(no-till)
157 kg N ha-1 (140 lb N acre-1) of poultry manure on Apr. 20;
45 kg N ha-1 (40 lb N ha-1) of sidedress 30% UAN on June 7
Oct. 3
Corn (after soybean and
double-crop soybean)
Apr. 30
(no-till)
124 kg N ha-1 (110 lb N acre-1) of poultry manure on Apr. 20;
34 kg N ha-1 (30 lb N ha-1) of sidedress 30% UAN on June 7
Oct. 3
SoybeanMay 20
(no-till)
-Oct. 15
Double-crop winter wheatOct. 1034 kg N ha-1 (30 lb N acre-1) of sidedress 30% UAN on Oct. 8;
45 kg N ha-1 (40 lb N acre-1) of sidedress 30% UAN on Mar. 1;
67 kg N ha-1 (60 lb N acre-1) of sidedress 30% UAN on Apr. 5
June 27
Double-crop soybeanJune 29-Nov. 1
WCCCorn (after corn)Apr. 30
(no-till)
157 kg N ha-1 (140 lb N acre-1) of poultry manure on Apr. 20;
45 kg N ha-1 (40 lb N acre-1) of sidedress 30% UAN on June 7
Oct. 1 or 30
Corn (after soybean and
double-crop soybean)
Apr. 30
(no-till)
124 kg N ha-1 (110 lb N acre-1) of poultry manure on Apr. 20;
34 kg N ha-1 (30 lb N ha-1) of sidedress 30% UAN on June 7
Oct. 1 or 30
SoybeanMay 20
(no-till)
-Oct. 1 or 30
Double-crop winter wheatOct. 1034 kg N ha-1 (30 lb N acre-1) of sidedress 30% UAN on Oct. 8;
45 kg N ha-1 (40 lb N acre-1) of sidedress 30% UAN on Mar. 1;
67 kg N ha-1 (60 lb N acre-1) of sidedress 30% UAN on Apr. 5
June 27
Double-crop soybeanJune 29-Nov. 1
Winter cover cropOct. 3 and Nov. 2-Mar. 31
(killing)

    [a] UAN = urea-ammonium nitrate. The typical nitrogen content for poultry manure is 2.8% (Glancey et al., 2012).

Baseline Model Calibration and Validation

We adopted SWAT simulations for streamflow and nitrate loads that were previously calibrated and validated by Lee et al. (2017a) for the two watersheds under the NoWCC scenario. Lee et al. (2017a) ran SWAT at a monthly time step for 16 years, including a two-year warm-up period (1999-2000), an eight-year calibration period (2001-2008), and a six-year validation period (2009-2014). They calibrated the model following the technical guidelines for SWAT suggested by Arnold et al. (2012). They manually modified the parameters identified in previous literature as sensitive for this region and selected a set of parameters that produced the best performance measures, i.e., Nash-Sutcliffe efficiency coefficient (NSE), root mean squared error (RMSE) to standard deviation ratio (RSR), and percent bias (P-bias; Lee et al., 2017a). The uncertainty analysis was also performed using 95% prediction uncertainty (95PPU), which represents the range between the top and bottom 2.5% of all simulated results (Lee et al., 2017a). Model simulation results were acceptable based on the model criteria for monthly streamflow and nitrate loads suggested by Moriasi et al. (2007). Thus, the calibrated model was applicable to replicate the hydrology and N cy-cle under different modeling scenarios. The calibrated parameters, performance measures, and a graphical comparison are provided in tables A1 and A2 and figure A1, respectively, the Appendix. Further details on the calibration and validation processes are provided by Lee et al. (2017a).

Simulating WCC biomass growth accurately is important because WCC nitrate reduction efficiency is highly dependent on WCC biomass (Hively et al., 2009). Yeo et al. (2014) estimated WCC parameters that represent the typical growth patterns of rye, barley, and wheat for this region based on landscape-level observed biomass reported by Hively et al. (2009). The WCC biomass simulated using the parameters from Yeo et al. (2014) was consistent with field observations conducted at different periods in this region (Lee et al., 2016). Therefore, we used the calibrated WCC biomass growth parameters provided by Yeo et al. (2014). In addition, we set the base and optimal temperatures for the three WCC species (i.e., rye, barley, and wheat) as 4°C and 18°C, respectively, based on literature and local knowledge (Feyereisen et al., 2006; Prabhakara et al., 2015).

Future Climate Condition (FCC) Scenarios

GCM-based climate scenarios have been widely used to predict the future performance of BMPs (Mearns, 2001). Simulation results using GCM data have been shown to provide insights for preparing effective management practices in response to FCCs (Gassman et al., 2014). We downloaded future daily precipitation and maximum and minimum temperatures from the bias-corrected and downscaled World Climate Research Program (WCRP) Coupled Model Intercomparison Project 3 (CMIP3) climate projection archive (http://gdo-dcp.ucllnl.org/downscaled_ cmip3_projections/). Data from the eight GCMs were downloaded separately for three greenhouse gas emission scenarios (i.e., A1B, A2, and B1) (table 4). Davidson and Metz (2000) explained that the A1B, A2, and B1 scenarios differ by future economic development and energy demand. More specifically, the A1B scenario illustrates a rapid growth of economic and population and balanced dependence of energy between fossil and non-fossil sources, the A2 scenario illustrates regionally oriented economic development and fragmented technological change, and the B1 scenario illustrates rapid changes in economic structures toward eco-friendly technologies (Davidson and Metz, 2000). In our assessment, a 14-year future period (2085-2098) was simulated for comparison with the 14-year baseline period (2001-2014). The CO2 concentrations for the future period were set as 700 ppm (A1B), 820 ppm (A2), and 550 ppm (B1) regarding the future CO2 concentration assumed in CMIP3 (Meehl et al., 2007). The baseline CO2 concentration was set as the SWAT default value of 330 ppm (Neitsch et al., 2011). Future humidity, wind speed, and solar radiation were prepared using SWAT’s built-in weather generator as a result of data unavailability. Because great variations existed among the GCM data, the ensemble means of simulated outputs for the eight GCMs were calculated separately for the A1B, A2, and B1 scenarios (Shrestha et al., 2012; Van Liew et al., 2012). The ranges of possible variations were also represented using the maximum and minimum values of the simulated outputs.

Table 4. GCMs used for FCC scenarios (2085-2098).
ModelFull NameAgency
CCCMA_CGCM3.1.1Canadian Centre for Climate Modelling and Analysis Coupled GCM 3.1.1Canadian Centre for Climate Modelling and Analysis, Canada
CNRM_CM3.1Centre National de Recherches Météorologiques Coupled Global Climate Model, version 3.1National Center of Meteorological Research, France
GFDL_CM2.0.1Geophysical Fluid Dynamics Laboratory Climate Model, version 2.0.1Geophysical Fluid Dynamics Laboratory, U.S.
GFDL_CM2.1.1Geophysical Fluid Dynamics Laboratory Climate Model, version 2.1.1Geophysical Fluid Dynamics Laboratory, U.S.
IPSL_CM4.1L’Institut Pierre-Simon Laplace Coupled Model, version 4.1L’Institut Pierre-Simon Laplace, France
MIROC3.2 (medres)Model for Interdisciplinary Research on ClimateMarine-Earth Science and Technology, Japan
MIUB_ECHO_GMeteorological Institute of the University of Bonn ECHAM4 Hamburg Ocean Primitive Equation GMeteorological Institute of the University of Bonn, Germany
MRI-CGCM2.3.2Meteorological Research Institute Coupled GCM 2.3.2Meteorological Research Institute, Japan Meteorological Agency, Japan

Assessment of FCC Impacts on WCC Nitrate Reduction Efficiency

We assessed the FCC impacts on WCC nitrate reduction efficiency at the cropland and watershed scales. Multiple simulations were performed for the baseline (2001-2014) and FCC scenarios (2085-2098) when WCCs were present and absent (table 3). At the watershed scale, we investigated FCC impacts on water and nitrate budgets prior to implementing WCCs and then analyzed how these nitrate budgets were altered when WCCs were planted. We calculated the relative differences in nitrate budget between the NoWCC and WCC scenarios for the baseline and FCC scenarios to determine whether the WCC nitrate reduction efficiency increased or decreased. At the cropland scale, we further investigated how future WCC nitrate reduction efficiency differed by FCC scenario and WCC planting method. The cropland scale indicated all HRUs classified as cropland. All simulated outputs were normalized by the cropland or watershed size.

Statistical analyses were performed to determine if significant differences existed in simulated outputs depending on climate conditions (baseline versus FCC scenarios) and WCC treatment (NoWCC versus WCC scenarios). Paired sample t-tests were used to evaluate the effects of FCC scenario and WCC treatment on simulated output (Ficklin et al., 2010, 2013; Lee et al., 2016, 2017a). This method was used to (1) examine whether FCCs significantly increased nitrate loads compared to the baseline scenario and to (2) determine whether WCC treatment significantly reduced nitrate loads under the baseline and FCC scenarios.

Results and Discussions

Annual and monthly baseline climate data were compared with the ensemble mean of the eight GCMs (figs. 4 and 5). Relative to the baseline scenario, annual cumulative precipitation decreased by 37 mm (A1B), 53 mm (A2), and 40 mm (B1). Annual mean temperature increased by 3.4°C (A1B), 4.2°C (A2), and 2.1°C (B1) (fig. 4). The largest changes in annual climate data relative to the baseline were shown under the A2 scenario, followed by A1B and B1. Monthly cumulative precipitation was more variable than monthly mean temperature. Monthly cumulative precipitation increased by up to 28 mm (A1B), 21 mm (A2), and 20 mm (B1) in January and decreased by up to 42 mm (A1B), 45 mm (A2), and 43 mm (B1) in October compared to the baseline scenario (fig. 4a). Monthly mean temperature was higher than the baseline throughout the year. During WCC growing seasons, the cumulative precipitation values for A1B (587 mm), A2 (590 mm), and B1 (567 mm) scenarios were similar to or less than the baseline precipitation (590 mm), and the mean temperature increased by 3.3°C (A1B), 4.1°C (A2), and 2.1°C (B1) compared to the baseline scenario (fig. 4b).

Winter Cover Crop Biomass

Comparisons of biomass for the three WCCs under the baseline and FCC scenarios are provided in figure 6. Overall, WCC biomass growth was greater under FCC scenarios

Figure 4. Comparison of annual average (a) cumulative precipitation and (b) mean temperature between baseline and FCC scenarios. Vertical black lines indicate the range between maximum and minimum values of eight GCMs. Annual precipitation was 1221 mm (baseline), 1184 mm (A1B), 1168 mm (A2), and 1181 mm (B1). Annual temperatures were 13.9°C (baseline), 17.3°C (A1B), 18.1°C (A2), and 16.0°C (B1).

than under the baseline for all three WCC species. Biomass increased from 1.04 (baseline) to 1.64 (FCC scenarios) Mg ha-1 for wheat, from 1.19 to 1.7 Mg ha-1 for barley, and from 1.41 to 1.92 Mg ha-1 for rye. Elevated CO2 concentrations and warmer winter temperatures both resulted in increased biomass under FCC scenarios because elevated

CO2 concentrations contributed to increased biomass growth by reducing water demand and enhancing radiation use efficiency. The warmer winter temperatures also accelerated heat unit accumulation and curbed temperature stress. Basche et al. (2016) found that warmer temperatures contributed to increasing biomass of future rye cover crops in Iowa, compared to current conditions. Relative to the baseline, the biomass of wheat, barley, and rye increased by 24% (B1) and 58% (A2), by 12% (B1) and 43% (A2), and by 7% (B1) and 37% (A2), respectively. Conditions more favorable to WCC growth (i.e., higher CO2 concentrations and warmer winter temperatures) led to greater increase in WCC biomass (A2 > A1B > B1) (figs. 6d, 6h, and 6l).

Figure 6. Biomass growth of early-planted (a to d) winter wheat, (e to h) barley, and (i to l) rye under the baseline and FCC scenarios. Gray areas indicate the range between the maximum and minimum values of simulated WCC biomass.

Watershed-Scale Analysis

Figure 7 presents the 14-year averages of annual streamflow, ET, and nitrate loads calculated for watershed assessment. When WCCs were not implemented (i.e., NoWCC), streamflow increased significantly by 40% (A1B), 43% (A2), and 33% (B1) for TCW and by 23% (A1B), 27% (A2), and 19% (B1) for GW, compared to the baseline scenario (p < 0.01; fig. 7a). Increased streamflow was mainly attributed to a reduction of ET resulting from reduced stomatal conductance under elevated CO2 concentrations (fig. 7b). Note that annual precipitation remained approximately consistent for the baseline and FCC scenarios. Higher CO2 concentrations were shown to decrease plant stomatal conductance, subsequently reducing plant transpiration (Field et al., 1995). Several studies have shown similar results where elevated CO2 concentrations greatly increased the water balance, although temperatures were warmer (Ficklin et al., 2009, 2010). The A2 scenario with the highest CO2 concentration showed a greater increase in streamflow than the other two FCC scenarios (fig. 7a). Similar to the streamflow variations, nitrate loads increased significantly by 43% (A1B), 42% (A2), and 35% (B1) for TCW and by 28% (A1B), 27% (A2), and 23% (B1) for GW, compared to the baseline scenario (p < 0.01, fig. 7c). Although the increased level of streamflow was most pronounced for the A2 scenario, nitrate loads were greater under A1B compared to A2 because of the greater seasonal precipitation during fertilizer application seasons (i.e., March and April) under the A1B scenario (fig. 5a). Lee et al. (2017a) showed that increased precipitation during periods around fertilizer application led to substantial increases in nitrate loads. Therefore, the substantial export of nitrate occurred around fertilizer application under the A1B scenario, causing greater nitrate loads compared to the A2 scenario.

Minimal changes in streamflow and ET between the NoWCC and WCC scenarios (i.e., WE to RL; table 2) were shown for the baseline and FCC scenarios (figs. 7a and 7b). However, nitrate loads were significantly reduced by implementing WCCs (p < 0.01; fig. 7c). This trend was consistent with two previous studies conducted in this region, showing that the effects of WCCs on water budget were negligible compared to their impacts on nitrate budget (Yeo

et al., 2014; Lee et al., 2016). The WCCs reduced annual nitrate loads by ~43% (baseline) and ~48% (FCC scenarios) for TCW and by ~25% (baseline) and ~30% (FCC scenarios) for GW. Nitrate loads under FCCs with WCCs were ~28% lower than those under the baseline without WCCs. This comparison revealed substantial reduction in nitrate loads by WCCs, considering that nitrate loads under FCCs increased by 43% compared to the baseline when WCCs were not planted in both periods. Overall, WCC nitrate reduction efficiency increased by ~5% under FCC scenarios relative to the baseline. This improved efficiency indicates that WCCs under FCC scenarios immobilized ~3.7 and ~0.8 kg N ha-1 more nitrate for TCW and GW, respectively, compared to WCC performance under the baseline scenario (table 5). This result matches well with the simulated WCC biomass, considering that WCC nitrate reduction efficiency is proportional to WCC biomass growth. Improved WCC growth conditions increased crop biomass, resulting in enhanced nitrate reduction efficiency under FCC scenarios.

Table 5. Reduction efficiency and nitrate load for winter cover crops under the baseline and three FCC scenarios.[a]
WCC
Scenario
Tuckahoe Creek Watershed, % (kg N ha-1)Greensboro Watershed, % (kg N ha-1)
BaselineA1BA2B1BaselineA1BA2B1
WE21 (2.6)43 (7.7)44 (7.8)41 (7)22 (1.1)28 (1.9)28 (1.9)26 (1.7)
BE38 (4.8)45 (8.1)46 (8.3)44 (7.4)23 (1.2)29 (2)29 (2)27 (1.8)
RE43 (5.3)48 (8.6)48 (8.6)47 (7.9)25 (1.3)30 (2)30 (2)29 (1.9)
WL21 (2.6)35 (6.3)38 (6.9)32 (5.4)13 (0.7)22 (1.5)24 (1.6)20 (1.3)
BL25 (3.1)39 (7)41 (7.4)36 (6.1)15 (0.8)25 (1.7)26 (1.8)22 (1.5)
RL32 (4.1)43 (7.8)44 (7.9)41 (7)18 (1)26 (1.8)27 (1.8)25 (1.6)

    [a] Values outside parentheses are the reduction efficiency (%), and values inside parentheses are the amount (kg N ha-1) of nitrate load by the winter cover crops. The six WCC scenarios (WE to RL) are defined in table 2.

Figure 7. Fourteen-year average of annual (a) streamflow, (b) ET, and (c) nitrate loads for NoWCC and six WCC scenarios (WE to RL) under the baseline and FCC scenarios at the watershed scale. Vertical black lines indicate the range between the maximum and minimum values of simulations with eight GCMs. The six WCC scenarios (WE to RL) are defined in table 2.

Cropland-Scale Analysis

To accurately assess the effectiveness of individual WCC scenarios, we estimated 13-year averages of the winter nitrate budget (Oct. to Mar.) at the cropland scale (fig. 8). We extracted nitrate yield and leaching from cropland only; nitrate yields and leaching from other land use types (e.g., forest and urban) were excluded. Water yield and percolation results are shown in figure A2 in the Appendix. Nitrate yield is the summation of nitrate fluxes transported to streams by surface runoff, lateral flow, and groundwater flow. When WCCs were not planted, nitrate yield and leaching increased by 27% (~2.5 kg N ha-1) and 71% (~5.8 kg N ha-1), respectively, under FCC scenarios compared to the baseline values (fig. 8). The increases in nitrate yield and leaching were greatest under the A2 scenario due to the highest CO2 concentration and greatest winter precipitation (table 6). When WCCs were planted, nitrate yields decreased by 60% (baseline scenario) and 69% (FCC scenarios), and nitrate leaching decreased by 78% (baseline scenario) and 86% (FCC scenarios). Overall, WCCs showed greater nitrate reduction efficiency with increased reduction in nitrate yield (~3.1 kg N ha-1) and leaching (~10.7 kg N ha-1) at the cropland scale under FCC scenarios compared to the baseline scenario values (table 6).

Figure 8. Thirteen-year averages of (a) nitrate yield and (b) leaching during winter seasons (Oct. to Mar.) for NoWCC and six WCC scenarios (WE to RL) under the baseline and FCC scenarios at cropland scale. Nitrate yield is the summation of nitrate fluxes transported to streams by surface runoff, lateral flow, and groundwater flow. Vertical black lines indicate the range between the maximum and minimum values of simulations with eight GCMs. The WCC scenarios (WE to RL) are defined in table 2.
Table 6. Reduction efficiency and amount of nitrate yield and leaching by winter cover crops under the baseline and three FCC scenarios.[a]
WCC
Scenario
Tuckahoe Creek WatershedGreensboro Watershed
BaselineA1BA2B1BaselineA1BA2B1
Nitrate yield, % (kg N ha-1)
WE47 (4.2)57 (6.6)57 (6.5)55 (6.1)52 (2.8)63 (4)63 (4)61 (3.7)
BE51 (4.6)61 (7)61 (7)58 (6.5)56 (3.1)66 (4.2)67 (4.3)64 (3.9)
RE57 (5.1)64 (7.4)64 (7.4)63 (6.9)60 (3.3)69 (4.4)69 (4.4)67 (4.1)
WL27 (2.4)45 (5.2)47 (5.4)41 (4.5)29 (1.6)49 (3.1)52 (3.3)44 (2.7)
BL31 (2.8)51 (5.8)52 (5.9)47 (5.2)34 (1.8)55 (3.5)57 (3.6)51 (3.1)
RL41 (3.7)56 (6.5)56 (6.5)54 (5.9)42 (2.3)59 (3.8)61 (3.9)56 (3.5)
Leaching, % (kg N ha-1)
WE56 (11.2)72 (18.3)74 (19.1)69 (17.1)65 (5.3)78 (10.5)79 (10.9)76 (9.8)
BE61 (12.3)77 (19.7)79 (20.6)74 (18.3)72 (5.9)82 (11.1)84 (11.6)80 (10.3)
RE68 (13.7)82 (20.8)83 (21.5)80 (19.6)78 (6.4)85 (11.5)86 (11.9)84 (10.8)
WL29 (5.9)58 (14.7)62 (16.1)52 (12.8)33 (2.7)60 (8.2)65 (9)54 (6.9)
BL35 (7.1)65 (16.6)68 (17.8)60 (14.8)39 (3.2)68 (9.3)72 (9.9)63 (8.1)
RL47 (9.5)73 (18.5)75 (19.4)69 (17.2)51 (4.2)74 (10)76 (10.6)71 (9.2)
WE56 (11.2)72 (18.3)74 (19.1)69 (17.1)65 (5.3)78 (10.5)79 (10.9)76 (9.8)

    [a] Values outside parentheses are the reduction efficiency (%), and values inside parentheses are the amount (kg N ha-1) of nitrate yield or leaching by the winter cover crops. The six WCC scenarios (WE to RL) are defined in table 2.

The increase in WCC nitrate reduction efficiency varied by FCC scenario and planting method. For instance, WCCs under the A2 scenario indicated ~4% and ~10% higher efficiency for reducing nitrate leaching than the A1B and B1 scenarios, respectively (table 6). The A2 scenario represented the most conducive climate conditions for WCC growth (i.e., the highest CO2 concentration and warmest winter temperature). Accordingly, WCCs had the greatest biomass and nitrate reduction efficiency under the A2 scenario (figs. 6 and 8). As the CO2 concentration and winter temperature became lower, WCCs were less effective at reducing nitrate yield and leaching (e.g., A1B > B1). Moreover, the WCC practices with less effectiveness (i.e., wheat, barley, and late planting) under the baseline scenario showed considerable increases in nitrate reduction efficiency under FCC scenarios, compared to the WCC practices with greater effectiveness (i.e., rye and early planting). Compared to rye cover crops, wheat and barley cover crops respectively had ~7% and ~8% higher increases in nitrate reduction efficiency under FCC scenarios. Generally, rye is known to grow better at low temperatures than wheat and barley, which makes rye the most effective WCC for reducing nutrient loads (Dabney, 1998). However, warmer winter temperatures curbed the stress that would have suppressed wheat and barley growth, resulting in more improved nitrate reduction efficiency for those two crops in comparison to rye. The biomass growth rates also indicated that wheat and barley both had greater biomass growth than rye (fig. 6). In addition, late-planted WCCs showed substantial increases in nitrate reduction efficiency under FCC scenarios relative to the baseline scenario. The growth of late-planted WCCs was hindered by low temperatures under current climate conditions (Dabney, 1998). Warmer temperatures likely stimulated the growth of late-planted WCCs and their nitrate reduction efficiency under FCC scenarios. Compared to early-planted WCCs, late-planted WCCs indicated 14% higher increases in nitrate reduction efficiency under FCC scenarios.

Implications

Simulation results for the three FCC scenarios showed increases in nitrate loads in this region, indicating the need for implementation of BMPs in this region in the future. Bosch et al. (2014) reported that the performance of BMPs was less effective under FCC scenarios because the capacities of BMPs to control water and nitrate loads were constant under current and FCC scenarios, whereas FCC impacts increased water and nitrate loads. Chiang et al. (2012) reported that the optimal combination of multiple BMPs (e.g., manure application, vegetation filter strips, and grazing and pasture management) was anticipated to provide increased water quality benefits in Arkansas. However, their study did not clearly show the relationship between FCCs and improved BMP performance (Chiang et al., 2012). In contrast, our study demonstrated that WCCs were effective in mitigating increased nitrate export in the future, based on the enhanced nitrate removal mechanisms of WCCs (i.e., biomass growth) in response to FCC scenarios. In particular, the performance of WCCs was more effective when nitrate yield and leaching increased most substantially (e.g., the A2 scenario) because FCCs increased nitrate loads as well as WCC biomass growth. Prior to implementing WCCs, winter nitrate leaching was ~5.8 kg N ha-1 greater under the A2 scenario than the baseline. However, when WCCs were planted, winter nitrate leaching was ~2.0 kg N ha-1 lower under the A2 scenario compared to the baseline scenario value. As a result, these findings strongly indicate that implementing WCCs in this region in the future could help mitigate agricultural nutrient loads.

In addition to the water quality benefits, WCCs could provide economic benefits. Our findings indicate that the WCC practices using less effectiveness (i.e., wheat, barley, and late planting) under the baseline scenario could be suggested as alternatives to the WCC practices with greater effectiveness (i.e., rye and early planting) in the future to some degree, due to substantially increased nitrate reduction efficiency. Current Maryland cost-share programs provide $10 ha-1 more subsidy for both early planting and rye than for late planting and the other two cover crop species such as wheat and barley (MDA, 2016). If WCC practices with lower expenses (i.e., late planting, barley and wheat) could be installed instead of WCC practices with higher expenses (i.e., early planting and rye), then less subsidy would be required. Van Liew et al. (2012) showed that the conversion of corn or soybeans to switchgrass or pasture was an effective BMP for reducing pollutant loadings in Nebraska under FCCs although it was not economically feasible (Van Liew et al., 2012). Thus, WCCs are a promising BMP for the future because they are capable of offering both economic as well as water quality benefits.

Conclusion

With heightened concern about water quality degradation in the CBW, there is increased attention on the reliability of current BMPs for mitigating FCC impacts on hydrology and nutrient cycles. We evaluated the performance of WCCs in reducing nitrate loads under FCC scenarios using eight GCMs and three greenhouse gas emission scenarios (A1B, A2, and B1). The SWAT model was used to simulate the baseline (2001-2014) and FCC scenarios (2085-2098) to compare the effectiveness of WCCs in reducing nitrate loads between the two periods. First, WCC biomass growth under FCC scenarios was compared with the baseline scenario because WCC nitrate reduction efficiency is strongly associated with WCC biomass increase. We then examined the FCC impacts on water, nitrate budgets, and WCC nitrate reduction efficiency at watershed and cropland scales. Our simulation results showed that FCCs provided conditions conducive to WCC growth (i.e., elevated CO2 concentrations and warmer winter temperatures), resulting in 7% to 58% greater WCC biomass than the baseline scenario value. Prior to implementing WCCs, FCCs significantly increased annual nitrate loads by ~43% at the watershed scale. When planted, WCCs reduced annual nitrate loads by ~43% and ~48% under the baseline and FCC scenarios, respectively, at the watershed scale. The WCC nitrate reduction efficiency was ~5% higher under FCC scenarios due to the increased WCC biomass. The increase in WCC nitrate reduction efficiency varied by FCC scenario and WCC planting method. The CO2 concentrations were higher and winter temperatures were warmer under FCC scenarios; therefore, the WCC nitrate reduction efficiency increased because these climate conditions were more favorable for WCC biomass growth. In response to FCC scenarios, the WCC practices that were less effective under the baseline (i.e., barley and wheat crop species and late planting) indicated a ~14% higher increase in nitrate reduction efficiency compared to the WCC practices that were more effective under the baseline (i.e., rye and early planting), due to the warmer temperatures. Thus, our study suggests that WCCs can effectively mitigate increased nitrate loads caused by FCCs, and the role of WCCs in mitigating nitrate loads will be even more important under FCCs.

Acknowledgements

This research was supported by the USDA Conservation Effects Assessment Project (CEAP), the NASA Land Cover and Land Use Change (LCLUC) Program (Award No. NNX12AG21G), and the U.S. Geological Survey (USGS) Land Change Science Program.

data Availability

The data used to support the findings presented in this article are available in Lee et al. (2017b).

References

Arnold, J. G., Moriasi, D. N., Gassman, P. W., Abbaspour, K. C., White, M. J., Srinivasan, R., ... Jha, M. K. (2012). SWAT: Model use, calibration, and validation. Trans. ASABE, 55(4), 1491-1508. https://doi.org/10.13031/2013.42256

Arnold, J. G., Srinivasan, R., Muttiah, R. S., & Williams, J. R. (1998). Large-area hydrologic modeling and assessment: Part I. Model development. JAWRA, 34(1), 73-89. https://doi.org/10.1111/j.1752-1688.1998.tb05961.x

Ator, S. W., & Denver, J. M. (2012). Estimating contributions of nitrate and herbicides from groundwater to headwater streams, northern Atlantic Coastal Plain, United States. JAWRA, 48(6), 1075-1090. https://doi.org/10.1111/j.1752-1688.2012.00672.x

Basche, A. D., Archontoulis, S. V., Kaspar, T. C., Jaynes, D. B., Parkin, T. B., & Miguez, F. E. (2016). Simulating long-term impacts of cover crops and climate change on crop production and environmental outcomes in the Midwestern United States. Agric. Ecosys. Environ., 218, 95-106. https://doi.org/10.1016/j.agee.2015.11.011

Bosch, N. S., Evans, M. A., Scavia, D., & Allan, J. D. (2014). Interacting effects of climate change and agricultural BMPs on nutrient runoff entering Lake Erie. J. Great Lakes Res., 40(3), 581-589. https://doi.org/10.1016/j.jglr.2014.04.011

CBP. (2008). Bay barometer: A health and restoration assessment of the Chesapeake Bay and watershed in 2008. Annapolis, MD: Chesapeake Bay Program. Available at https://www.chesapeakebay.net/images/press_release_pdf/Bay_Barometer_2008_FINAL1.pdf

Chiang, L. C., Chaubey, I., Hong, N. M., Lin, Y. P., & Huang, T. (2012). Implementation of BMP strategies for adaptation to climate change and land use change in a pasture-dominated watershed. Intl. J. Environ. Res. Public Health, 9(10), 3654-3684. https://doi.org/10.3390/ijerph9103654

Dabney, S. M. (1998). Cover crop impacts on watershed hydrology. J. Soil Water Cons., 53(3), 207-213.

Davidson, O., & Metz, B. (2000). IPCC special report: Emissions scenarios. Summary for policy makers. Cambridge, UK: Cambridge University Press. Available at https://ipcc.ch/pdf/special-reports/spm/sres-en.pdf

Feyereisen, G. W., Wilson, B. N., Sands, G. R., Strock, J. S., & Porter, P. M. (2006). Potential for a rye cover crop to reduce nitrate loss in southwestern Minnesota. Agron. J., 98(6), 1416-1426. https://doi.org/10.2134/agronj2005.0134

Ficklin, D. L., Luo, Y., Luedeling, E., & Zhang, M. (2009). Climate change sensitivity assessment of a highly agricultural watershed using SWAT. J. Hydrol., 374(1), 16-29. https://doi.org/10.1016/j.jhydrol.2009.05.016

Ficklin, D. L., Luo, Y., Luedeling, E., Gatzke, S. E., & Zhang, M. (2010). Sensitivity of agricultural runoff loads to rising levels of CO2 and climate change in the San Joaquin Valley watershed of California. Environ. Pollut., 158(1), 223-234. https://doi.org/10.1016/j.envpol.2009.07.016

Ficklin, D. L., Stewart, I. T., & Maurer, E. P. (2013). Climate change impacts on streamflow and subbasin-scale hydrology in the upper Colorado River basin. PLoS One, 8(8), e71297. https://doi.org/10.1371/journal.pone.0071297

Field, C. B., Jackson, R. B., & Mooney, H. A. (1995). Stomatal responses to increased CO2: Implications from the plant to the global scale. Plant Cell Environ., 18(10), 1214-1225. https://doi.org/10.1111/j.1365-3040.1995.tb00630.x

Fisher, T., Jordan, T., Staver, K. W., Gustafson, A., Koskelo, A., Fox, R., ... Stone, J. P. (2010). The Choptank basin in transition: Intensifying agriculture, slow urbanization, and estuarine eutrophication. In M. J. Kennish and H. W. Paerl (Eds.), Coastal lagoons: Systems of natural and anthropogenic change (pp. 135-165). Boca Raton, FL: CRC Press.

Gassman, P. W., Reyes, M. R., Green, C. H., & Arnold, J. G. (2007). The Soil and Water Assessment Tool: Historical development, applications, and future research directions. Trans. ASABE, 50(4), 1211-1250. https://doi.org/10.13031/2013.23637

Gassman, P. W., Sadeghi, A. M., & Srinivasan, R. (2014). Applications of the SWAT model special section: Overview and insights. J. Environ. Qual., 43(1), 1-8. https://doi.org/10.2134/jeq2013.11.0466

Gitau, M.W., & Chaubey, I. (2010). Regionalization of SWAT Model Parameters for Use in Ungauged Watersheds. Water. 2(4), 849-71, 2010. https://doi:10.3390/w2040849

Glancey, J., Brown, B., Davis, M., Towle, L., Timmons, J., & Nelson, J. (2012). Comparison of methods for estimating poultry manure nutrient generation in the Chesapeake Bay watershed. New York, NY: Council of State Governments Eastern Regional Conference.

Hively, W. D., Lang, M., McCarty, G. W., Keppler, J., Sadeghi, A., & McConnell, L. L. (2009). Using satellite remote sensing to estimate winter cover crop nutrient uptake efficiency. J. Soil Water Cons., 64(5), 303-313. https://doi.org/10.2489/jswc.64.5.303

Kimball, B. A., & Idso, S. B. (1983). Increasing atmospheric CO2: Effects on crop yield, water use, and climate. Agric. Water Mgmt., 7(1), 55-72. https://doi.org/10.1016/0378-3774(83)90075-6

Lee, S., Yeo, I. Y., Sadeghi, A. M., McCarty, G. W., Hively, W. D., Lang, M. W., & Sharifi, A. (2017a). Comparative analyses of hydrological responses of two adjacent watersheds to climate variability and change scenarios using SWAT model. Hydrol. Earth Syst. Sci. (preprint). https://doi.org/10.5194/hess-2017-178

Lee, S., Sadeghi, A.M., Yeo, I. Y., McCarty, G. W., Hively, W. D. (2017b). Climate, crop rotation, and stream flow data used to run the SWAT model in the Tuckahoe and Greensboro subwatersheds of the Choptank River watersheds, Maryland: U.S. Geological Survey data release. https://doi.org/10.5066/F7DB80RP

Lee, S., Yeo, I.-Y., Sadeghi, A. M., McCarty, G. W., & Hively, W. D. (2015). Prediction of climate change impacts on agricultural watersheds and the performance of winter cover crops: Case study of the upper region of the Choptank River watershed. ASABE Paper No. 152123528. St. Joseph, MI: ASABE.

Lee, S., Yeo, I.-Y., Sadeghi, A. M., McCarty, G. W., Hively, W. D., & Lang, M. W. (2016). Impacts of watershed characteristics and crop rotations on winter cover crop nitrate uptake capacity within agricultural watersheds in the Chesapeake Bay region. PLoS One, 11(6), e0157637. https://doi.org/10.1371/journal.pone.0157637

McCarty, G. W., McConnell, L. L., Hapeman, C. J., Sadeghi, A., Graff, C., Hively, W. D., ... Fogel, M. L. (2008). Water quality and conservation practice effects in the Choptank River watershed. J. Soil Water Cons., 63(6), 461-474. https://doi.org/10.2489/jswc.63.6.461

MDA. (2016). Maryland’s 2015-2016 cover crop sign-up. Annapolis, MD: Maryland Department of Agriculture. Available at http://mda.maryland.gov

Mearns, L. O., Hulme, M., Carter, T. R., Leemans, R., Lal, M., Whetton, P., ... Wilby, R. (2001). Chapter 13: Climate scenario development. In Climate change 2001: The scientific basis. Cambridge, UK: Cambridge University Press. Available at https://www.ipcc.ch/ipccreports/tar/wg1/pdf/TAR-13.PDF

Meehl, G. A., Covey, C., Taylor, K. E., Delworth, T., Stouffer, R. J., Latif, M., ... Mitchell, J. F. (2007). THE WCRP CMIP3 multimodel dataset: A new era in climate change research. Bull. American Meteorol. Soc., 88(9), 1383-1394. https://doi.org/10.1175/bams-88-9-1383

Moriasi, D. N., Arnold, J. G., Van Liew, M. W., Bingner, R. L., Harmel, R. D., & Veith, T. L. (2007). Model evaluation guidelines for systematic quantification of accuracy in watershed simulations. Trans. ASABE, 50(3), 885-900. https://doi.org/10.13031/2013.23153

Najjar, R. G., Pyke, C. R., Adams, M. B., Breitburg, D., Hershner, C., Kemp, M., ... Wood, R. (2010). Potential climate-change impacts on the Chesapeake Bay. Estuarine Coastal Shelf Sci., 86(1), 1-20. https://doi.org/10.1016/j.ecss.2009.09.026

Najjar, R., Patterson, L., & Graham, S. (2009). Climate simulations of major estuarine watersheds in the Mid-Atlantic region of the U.S. Clim. Change, 95(1), 139-168. https://doi.org/10.1007/s10584-008-9521-y

Neitsch, S. L., Arnold, J. G., Kiniry, J. R., & Williams, J. R. (2011). Soil and Water Assessment Tool theoretical documentation. Ver. 2009. College Station, TX: Texas Water Resources Institute.

NRC. (2011). Achieving nutrient and sediment reduction goals in the Chesapeake Bay: An evaluation of program strategies and implementation. Washington, DC: National Research Council. Available at https://www.chesapeakebay.net/channel _files/21727/2011_nas_report.pdf

Parry, M. L., Rosenzweig, C., Iglesias, A., Livermore, M., & Fischer, G. (2004). Effects of climate change on global food production under SRES emissions and socio-economic scenarios. Global Environ. Change, 14(1), 53-67. https://doi.org/10.1016/j.gloenvcha.2003.10.008

Prabhakara, K., Hively, W. D., & McCarty, G. W. (2015). Evaluating the relationship between biomass, percent groundcover, and remote sensing indices across six winter cover crop fields in Maryland, United States. Intl. J. Appl. Earth Obs. Geoinfo., 39, 88-102. https://doi.org/10.1016/j.jag.2015.03.002

Richardson, C. W., Bucks, D. A., & Sadler, E. J. (2008). The Conservation Effects Assessment Project benchmark watersheds: Synthesis of preliminary findings. J. Soil Water Cons., 63(6), 590-604. https://doi.org/10.2489/jswc.63.6.590

Runkel, R. L., Crawford, C. G., & Cohn, T. A. (2004). Load estimator (LOADEST): A FORTRAN program for estimating constituent loads in streams and rivers. Reston, VA: U.S. Geological Survey.

Seo, M., Yen, H., Kim, M.K. & Jeong, J. (2014). Transferability of SWAT Models between SWAT2009 and SWAT2012. J. Environ Qual., 43(3), 869-880. https://doi:10.2134/jeq2013.11.0450

Sharifi, A., Lang, M. W., McCarty, G. W., Sadeghi, A. M., Lee, S., Yen, H., ... Yeo, I.-Y. (2016). Improving model prediction reliability through enhanced representation of wetland soil processes and constrained model auto calibration: A paired watershed study. J. Hydrol., 541(part B), 1088-1103. https://doi.org/10.1016/j.jhydrol.2016.08.022

Shrestha, R. R., Dibike, Y. B., & Prowse, T. D. (2012). Modelling of climate-induced hydrologic changes in the Lake Winnipeg watershed. J. Great Lakes Res., 38, 83-94. https://doi.org/10.1016/j.jglr.2011.02.004

Staver, K. W., Brinsfield, R., & Stevenson, J. C. (1989). Effect on best management practices on nitrogen transport into Chesapeake Bay. Proc. 2nd Pan-American Regional Conf. Intl. Commission on Irrigation and Drainage (pp. 163-180). Denver, CO: U.S. Committee on Irrigation and Drainage.

Van Liew, M. W., Feng, S., & Pathak, T. B. (2012). Climate change impacts on streamflow, water quality, and best management practices for the Shell and Logan Creek watersheds in Nebraska. Intl. J. Agric. Biol. Eng., 5(1), 13-34.

Woznicki, S. A., & Nejadhashemi, A. P. (2012). Sensitivity analysis of best management practices under climate change scenarios. JAWRA, 48(1), 90-112. https://doi.org/10.1111/j.1752-1688.2011.00598.x

Yeo, I. Y., Lee, S., Sadeghi, A. M., Beeson, P. C., Hively, W. D., McCarty, G. W., & Lang, M. W. (2014). Assessing winter cover crop nutrient uptake efficiency using a water quality simulation model. Hydrol. Earth Syst. Sci., 18(12), 5239-5253. https://doi.org/10.5194/hess-18-5239-2014

Zeppel, M. J., Wilks, J. V., & Lewis, J. D. (2014). Impacts of extreme precipitation and seasonal changes in precipitation on plants. Biogeosci., 11(11), 3083-3093. https://doi.org/10.5194/bg-11-3083-2014

Appendix

Table A1. List of calibrated parameters (adapted from Lee et al., 2017).
VariableParameterDescription (unit)RangeCalibrated Value
TCWGW
StreamflowCN2[a]Curve number-50% to 50%-30%0%
ESCO[a]Soil evaporation compensation factor0 to 110.95
SURLAG[a]Surface runoff lag coefficient0.5 to 240.50.5
SOL_AWC[a]Available water capacity of the soil layer (mm H2O mm soil-1)-50 to 50%-10%-1%
SOL_K[a]Saturated hydraulic conductivity (mm h-1)-50 to 50%50%49%
SOL_Z[a]Depth from soil surface to bottom of layer (mm)-50 to 50%-20%-31%
ALPHA_BF[a]Base flow recession constant (d-1)0 to 10.070.051
GW_DELAY[a]Groundwater delay time (d)0 to 50012045
GW_REVAP[a]Groundwater “revap” coefficient0.02 to 0.20.100.02
RCHRG_DP[a]Deep aquifer percolation fraction0 to 10.010.05
CH_K2[a]Effective hydraulic conductivity (mm h-1)0 to 150020
CH_N2[a]Manning coefficient0.01 to 0.30.290.021
GWQMN[a]Threshold depth of water in shallow aquifer required for return flow to occur (mm)0 to 50001.91.0
NitrateNPERCO[b]Nitrogen percolation coefficient0.01 to 10.50.2
N_UPDIS[b]Nitrogen uptake distribution parameter5 to 505050
ANION_EXCL[b]Fraction of porosity from which anions are excluded0.1 to 0.70.590.6
ERORGN[b]Organic N enrichment ratio for loading with sediment0 to 54.924.1
BIOMIX[b]Biological mixing efficiency0.01 to 10.010.01
SOL_NO3[c]Initial NO3 concentration in soil layer (mg N kg-1)0 to 10011.230
CDN[d]Denitrification exponential rate coefficient0 to 3.00.31.8
SDNCO[d]Denitrification threshold water content0.1 to 1.11.01.0

    [a] Range adapted from Gitau and Chaubey (2010).

    [b] Range adapted from Yeo et al. (2014).

    [c] Range adapted from Seo et al. (2014).

    [d] Range adapted from Neitsch et al. (2011).

Table A2. Model performance measures for monthly streamflow and nitrate loads (adapted from Lee et al., 2017).[a]
Period and
Variable
StreamflowNitrate loads
TCWGWTCWGW
Calibration
NSE0.723**0.686**0.623*0.702**
RSR0.523**0.556**0.610*0.542**
P-bias (%)-5.8***-3.2***-9.8***-4.1***
Validation
NSE0.674**0.790***0.604*0.567*
RSR0.566**0.454***0.624*0.652*
P-bias (%)17.8**13***-5.6***-12.1***

    [a] Based on the criteria of Moriasi et al. (2007): * = Satisfactory (0.5 < NSE = 0.65, 0.6 < RSR = 0.7, and ±15 = P-bias < ±25); ** = Good (0.65 < NSE = 0.75, 0.5 < RSR = 0.6, and ±10 = P-bias < ±15); and *** = Very Good (0.75 < NSE = 1.0, 0.0 < RSR = 0.5, P-bias < ±10).

Figure A1. Simulated and observed monthly streamflow and nitrate loads for (a and b) TCW and (c and d) GW during calibration and validation periods (adapted from Lee et al., 2017); 95PPU = 95% prediction uncertainty.
Figure A2. Thirteen-year averages of (a) water yield and (b) percolation during winter seasons (Oct. to Mar.) for NoWCC and six WCC scenarios (WE to RL) under the baseline and FCC scenarios at the cropland scale. Water yield is the summation of water fluxes transported to streams by surface runoff, lateral flow, and groundwater flow. Vertical black lines indicate the range between the maximum and minimum values of simulations with eight GCMs. The six WCC scenarios are defined in table 2.