Environ. Sci. Technol. 2010, 44, 7138–7144
Modeling Miscanthus in the Soil and Water Assessment Tool (SWAT) to Simulate Its Water Quality Effects As a Bioenergy Crop TZE LING NG,† J. WAYLAND EHEART,† X I M I N G C A I , * ,† A N D FERNANDO MIGUEZ‡ Department of Civil and Environmental Engineering, University of Illinois at Urbana-Champaign, 205 N. Mathews Avenue, Urbana, Illinois 61801, and Department of Agronomy, Iowa State University, 2101 Agronomy Hall, Ames, Iowa 50011
Received February 9, 2010. Revised manuscript received June 28, 2010. Accepted July 13, 2010.
There is increasing interest in perennial grasses as a renewable source of bioenergy and feedstock for second-generation cellulosic biofuels. The primary objective of this study is to estimate the potential effects on riverine nitrate load of cultivating Miscanthus x giganteus in place of conventional crops. In this study, the Soil and Water Assessment Tool (SWAT) is used to model miscanthus growth and streamwater quality in the Salt Creek watershed in Illinois. SWAT has a built-in crop growth component, but, as miscanthus is relatively new as a potentially commercial crop, data on the SWAT crop growth parameters for the crop are lacking. This leads to the second objective of this study, which is to estimate those parameters to facilitate the modeling of miscanthus in SWAT. Results show a decrease in nitrate load that depends on the percent land use change to miscanthus and the amount of nitrogen fertilizer applied to the miscanthus. Specifically, assuming a nitrogen fertilization rate for miscanthus of 90 kg-N/ha, a 10%, 25%, and 50% land use change to miscanthus will lead to decreases in nitrate load of about 6.4%, 16.5%, and 29.6% at the watershed outlet, respectively. Likewise, nitrate load may be reduced by lowering the fertilizer application rate, but not proportionately. When fertilization drops from 90 to 30 kg-N/ha the difference in nitrate load decrease is less than 1% when 10% of the watershed is miscanthus and less than 6% when 50% of the watershed is miscanthus. It is also found that the nitrate load decrease from converting less than half the watershed to miscanthus from corn and soybean in 1:1 rotation surpasses that from converting the whole watershed to just soybean.
source of feedstock for second generation cellulosic biofuels. Current first generation biofuels, namely corn ethanol and biodiesel from soybeans, are restricted in their potential to offset the use of liquid fossil fuels in transportation (1). Miscanthus x giganteus, a C4 photosynthesis plant, is a perennial grass that possesses many characteristics desirable in an energy crop, e.g., high water and nutrient use efficiencies, positive soil restoration and carbon sequestration, and low nutrient contents (2). Further, it has been shown (3) to give yields superior to switchgrass, another perennial grass that has received attention for its potential as a model bioenergy crop (4). In addition to reductions in greenhouse gas emissions, perennial grasses, under the right conditions, also have the advantage of reducing waterborne pollutants, including nitrate (5). Excessive nitrate concentrations in drinking water have health implications (methemoglobinemia), and municipal water facilities are obligated by law to meet a maximum nitrate concentration standard of 10 mg/L, which many utilities, especially in the Corn Belt, struggle to do. Further, nitrate from agricultural operations in the Midwest has been implicated in hypoxia in the Gulf of Mexico; (6) estimated 43% of the total annual average nitrate flux into the gulf to originate from the Upper Mississippi River Basin (of which the watershed modeled in this study is a part) even though it constitutes just 16% of the total area draining into the gulf. Thus, the primary objective of this study is to estimate the potential effects on riverine nitrate load of cultivating Miscanthus x giganteus in place of conventional crops. To achieve this, the Soil and Water Assessment Tool (SWAT) (7) is used to model miscanthus growth and streamwater quality in the Salt Creek watershed in East-Central Illinois (see Figure 1). SWAT is widely accepted and has been applied in a number of water quality assessments (5, 8) that have some common features with the present work. Other studies (9-14) have successfully developed mathematical models of miscanthus growth. (Some background information on these models is provided in the Supporting Information (SI).) These models are however, standalone models that are not linked to a hydrologic model. In contrast, the crop growth model in SWAT (which is based on the
1. Introduction Due to global warming and energy independence concerns, there is increasing interest in perennial grasses as a renewable source of bioenergy. Energy from combusting biomass can be thought to be carbon-neutral (or nearly so), because the carbon released during the combustion process is absorbed by the biomass as it is growing. Perennial grasses are also a * Corresponding author phone: (217) 333-4935; e-mail: xmcai@ illinois.edu. † University of Illinois at Urbana-Champaign. ‡ Iowa State University. 7138
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 18, 2010
FIGURE 1. Locator map of the Salt Creek watershed, Illinois with the four gaging stations (∆) from which data are used to calibrate and validate the SWAT model. 10.1021/es9039677
2010 American Chemical Society
Published on Web 08/03/2010
TABLE 1. SWAT Crop Growth Parameters and Their Estimated Values for Miscanthus
a Optimal growth parameter estimated by fitting the optimal crop growth equations in SWAT to simulated data from a second model; refer to Section 2.1 for more details. b Nutrient parameter derived from field data in 20. c Miscellaneous parameter; assume all above-ground biomass is harvestable. d Miscellaneous parameter based on 23. e Miscellaneous parameter estimated based on personal communication with F. Dohleman and C. Bernacchi, 2009. f Miscellaneous parameter estimated based on the default values for other crops of the same class (perennial grasses) in the SWAT crop database.
Erosion Productivity Impact Calculator (EPIC) model in 15) is embedded within a larger hydrologic model, enabling the modeling of the surrounding water quality and quantity in response to crop choice. Note however, that, as miscanthus is relatively new as a potentially commercial crop, data on the SWAT crop growth parameters for it are lacking. This leads to the second objective of this study, which is to estimate those parameters to facilitate the modeling of miscanthus in SWAT.
2. Method The SWAT model of the Salt Creek watershed applied in this study has been calibrated and validated in a previous study (16) for the calibration period 1996-2003 and validation period 1988-1995. To calibrate the SWAT model, the Generalized Uncertainty Estimation (GLUE) method (17), a stochastic calibration method, was used. The model is deemed sufficiently accurate for use in this study, based on the coefficient of determination and Nash-Sutcliffe efficiency statistics for several time-series of empirical data. Those data are of daily streamflow at four locations within the watershed, annual corn and soybean yields, and monthly nitrate load at the watershed outlet. Additional work is carried out in this study to parameterize the crop growth model in SWAT for miscanthus. Even though
SWAT comes with a database of default parameters for a number of crops, including corn, soybean, wheat, etc., default values for miscanthus are unavailable as it is relatively new as a potentially commercial crop. To estimate them, the crop growth parameters are divided into three subsets: those describing optimal biomass growth under zero stress conditions, nitrogen and phosphorus stress parameters, and miscellaneous parameters not included in the first two subsets. Refer to Table 1 for a list of the parameters and their definitions. 2.1. Optimal Biomass Growth Parameters. There are ten parameters in SWAT describing optimal crop growth, i.e., the theoretical maximum growth achievable when there are no nutrient and water limitations and when ambient temperature is in the optimal range. These parameters for miscanthus are unknown and their values are identified in this study by fitting SWAT’s optimal crop growth equations (as given by 18) to simulated biomass and leaf area index (LAI) data from a second model. That second model is an improved version of the miscanthus model by 14. The model has been previously parameterized and tested using empirical data from European studies (as described by 14). The current implementation is called BioCro and will be available in the near future as an R (19) package. The primary difference between BioCro VOL. 44, NO. 18, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7139
FIGURE 2. LAI, above-ground biomass, and total new biomass optimal curves as obtained using the SWAT optimal crop growth equations (s) and BioCro (shaded gray line). and SWAT is in their treatments of the crop’s radiation-tobiomass conversion efficiency. SWAT treats it as a userprovided constant, while BioCro estimates its effects endogenously from fundamental processes. This makes BioCro more mechanistic; it thus can be expected to perform better in locations where no data are available for site-specific calibration. Using this model, optimal above-ground and belowground biomass and LAI data for 14 years are generated. The data are generated with input solar radiation and temperature data from the Illinois State Water Survey’s (ISWS) Water and Atmospheric Resources Monitoring (WARM) database (www. isws.illinois.edu/warm/) for the weather station at Kilbourne, IL. This approach of using simulated data is appropriate, as simulated data are not subject to the uncontrollable factors to which field data are vulnerable. Using field data, it would not be possible to isolate the effects of water, nutrient, and temperature stress on biomass growth to identify the optimal biomass and LAI curves. 2.2. Nutrient and Miscellaneous Parameters. SWAT calculates nutrient stress in a crop by comparing actual nitrogen and phosphorus levels in the crop against ideal optimal levels. The crop’s actual nitrogen and phosphorus contents are endogenous variables within SWAT that are computed based on the availability of the nutrients in the soil. On the other hand, the optimal nitrogen and phosphorus contents in the crop are computed based on six parameters provided by the user, three for nitrogen and three for phosphorus. The six parameters refer to the optimal fractions of the nutrients in the plant at emergence, 50% maturity, and maturity, and are used to interpolate for the optimal fractions of the nutrients throughout the growing season. Further, there are two other user-provided parameters related to the nitrogen and phosphorus content of the plant. These parameters refer to the fractions of the nutrients in the aboveground biomass at harvest. SWAT uses these parameters to calculate the amounts of nutrients removed from the system at harvest, which affect the content of nutrients in the soil. To estimate these eight parameters for miscanthus, data reported in the literature (20) are used. Literature data, 7140
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 18, 2010
together with expert opinion and the default values for other crops of the same class (perennial grasses) in the SWAT crop database, are also used to estimate the miscellaneous parameters not included in the first two subsets. Refer to Table 1 for the specific source, assumption, or method used to estimate each individual parameter.
3. Results and Discussion 3.1. SWAT Crop Growth Parameters for Miscanthus. Refer to Table 1 for the optimal crop growth parameter values for miscanthus determined by following the steps outlined above. Refer also to Figure 2 for the LAI, above-ground biomass, and total new biomass (new in the current year, equal to total minus carryover from the previous year) curves obtained when applying those values to the optimal biomass growth equations in SWAT. As can be seen from Figure 2, the parameter values determined are able to reproduce satisfactorily the “ideal” optimal curves produced using BioCro (as described in Section 2.1 above). For the calibration period, viz., 2002-2008, the R2 values for LAI, above-ground biomass, and total new biomass are, respectively, 0.77, 0.90, and 0.94. As for the validation period, viz., 1995-2001, the R2 values are 0.92, 0.96, and 0.98, respectively. However, the SWAT optimal growth equations tend to predict more total new biomass than BioCro. This is because the procedure in SWAT to predict partitioning between above- and below-ground biomass is simpler than the procedure in BioCro; SWAT does not predict both total new and above-ground biomass to the same extent. In this study, above-ground biomass is deemed to be of greater interest than total new biomass and hence is given greater emphasis when adjusting the optimal growth parameters. Note that the bio_e for miscanthus is set to be the same as the SWAT default bio_e for corn. Theoretically, crops of the same photosynthesis type should have about the same optimal radiation use efficiency (RUE) values (21). Like miscanthus, corn is a C4 photosynthesis plant. Miscanthus RUE values reported in the literature range from 1.91 to 4.2 g per MJ/m2 (11, 12, 9, 22, 10). These are however, empirical values projected from field data (that are subject to water,
FIGURE 3. Optimal biomass growth (- · - · ) and actual biomass growth when nitrogen fertilizer application rate in kg-N/ha is 0 (s), 40 (---), and 80 (...) as predicted by SWAT; and actual biomass growth (∆) when nitrogen fertilization is zero as measured in field trials.
temperature, and/or nutrient limitations) and not theoretical optimal values which are the interest in this study. Similarly, the blai values reported in the literature are observed field values ranging from 5 to 8 (10, 23) and are generally smaller (as they should be) than the optimal blai determined in this study, 11.5 m/m. As for the other parameters, their values here follow values previously reported. For example, for t_base, the 10 °C estimated here is well within the 6-15 °C range found in the literature (10-12). The value for ext_coef is set at the SWAT default for all crops regardless of type at 0.65 which is very close to the values (0.67-0.68) applied by 10 and 11. Note however, for heat units, there is some difference between the 2100 °C-day obtained here and the 1800 °C-day used by 12. Refer again to Table 1 for the values of the miscanthus nitrogen, phosphorus, and miscellaneous crop growth parameters determined in this study. For the nitrogen parameters pltnfr(1), pltnfr(2), pltnfr(3) and the phosphorus parameters pltpfr(1), pltpfr(2), pltpfr(3), note that their values for miscanthus are less than their SWAT default values for corn (which constitutes a large part of existing crops in the Salt Creek watershed and other parts of the Midwest). For corn, pltnfr(1) is 0.0470, pltnfr(2) 0.0177, and pltnfr(3) 0.0138, and pltpfr(1) is 0.0048, pltpfr(2) 0.0018 and pltpfr(3) 0.0014. This means that when compared to corn, miscanthus has a smaller nutrient demand and is, thus, less susceptible to nutrient stress. Similarly, cnyld and cpyld for miscanthus are less than for corn, which has a cnyld of 0.014 and cpyld of 0.0016. Therefore, in terms of per unit biomass harvested, fewer nutrients are removed from the system when harvesting miscanthus than when harvesting corn. All this translates to lower fertilizer requirements for miscanthus than for corn. usle_c is the Universal Soil Loss Equation (USLE) C factor affecting erosion and sediment runoff into surface waters. The higher the value of usle_c, the greater the erosion and sediment runoff. Here, a miscanthus usle_c factor of 0.003 is assumed. This value corresponds to the SWAT default usle_c for most of the perennial grasses in its crop database and is significantly lower than the default for corn and soybean (which is 0.2 for both crops). The higher usle_c value for corn and soybean reflects the fact that they are annuals requiring tillage and replanting every year. Perennials (e.g., miscanthus), on the other hand, do not require yearly replanting but can continue yielding for many years once established. This allows them to provide year round ground cover, thus protecting the ground from wind and water erosion. To validate the parameterization of the crop growth model in SWAT for miscanthus, the parameter values in Table 1 are entered into the Salt Creek watershed SWAT model. The model is then used to predict “actual” biomass growth (which,
unlike optimal growth, occurs under nutrient, water, and/or temperature stress) for 2007 and 2008 when there is zero nitrogen fertilization. Comparison between the simulated results and field data (refer to Figure 3) show that the model is able to produce reasonable results. The field data are from 24, who conducted miscanthus field trials at a site in Champaign County, IL about 130 km east of the Salt Creek watershed. Figure 3 also gives SWAT’s predictions of actual biomass growth when nitrogen fertilizer application rate is 40 kg-N/ha and 80 kg-N/ha. The optimal growth curves for the two years are provided in the figure as well. Note that the data in Figure 3, with the exception of the field data, are not deterministic but are stochastic; they represent the most likely biomass growth that will occur, given the uncertainties in the Salt Creek watershed SWAT model as characterized in the GLUE calibration of the model (16). 3.2. Water Quality Effects. It is useful to know the relationship between yield and fertilizer application rate, to estimate the rate that can be expected, should farmers decide to plant miscanthus. As fertilizer input is a key contributor to nutrient export from agricultural watersheds, this knowledge is helpful in predicting the water quality effects of cultivating miscanthus. Refer to Figure S1 in the SI for an average of the yield-fertilizer relationship based on yield estimates for different fertilizer rates for different years (from 1995 to 2008) as predicted using SWAT and the miscanthus growth parameters presented above. The results are for a mid-December harvest date. From the average relationship, it can be inferred that farmers cultivating miscanthus in the Midwest are likely to apply somewhere between 80 and 100 kg-N/ha of nitrogen fertilizer. This is significantly less than the typical rate for corn, which is about 190 kg-N/ha (25). The results here somewhat contradict those of 26-28 who found fertilization to have no effect on yield. On the other hand, others have found miscanthus to have a positive response to fertilization (29), observing an average nitrogen response rate of about 37-50 kg biomass per kg-N applied. Further (30), by applying a statistical analysis on a database of published data from 31 studies on miscanthus cultivation and growth, nitrogen fertilization was found to affect biomass growth after the first three years of establishment, though not before. Thus, as it can be seen, there is a wide range in the results reported in the literature, which may be attributed to the possibility that the nitrogen response of miscanthus is a strong function of local soil and weather conditions. Moreover, it has been recently suggested (31) that miscanthus has the ability to obtain substantial nitrogen from atmospheric fixation, greatly reducing its need for synthetic fertilizer. The predicted range of 80-100 kg-N/ha of nitrogen fertilization made in this study is within the range of literature VOL. 44, NO. 18, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7141
FIGURE 4. Expected (s), expected plus/minus one standard deviation (shaded gray line), and time-averaged (---) monthly nitrate load at the watershed outlet under various scenarios of percent land use change to miscanthus (the remainder in corn and soybean in 1:1 rotation) and an all-soybean scenario for 2000-2008, assuming a nitrogen fertilizer application rate of 90 kg-N/ha for miscanthus. recommended rates. It was recommended in 28 a nitrogen fertilizer application rate of 50-70 kg-N/ha to replenish the nitrogen removed from the field during harvest, even though the study found yield to be unaffected by fertilization. It has also been recommended in the literature rates of 50-70 kgN/ha (32), 80 kg-N/ha (33) and 75 kg-N/ha (34). Refer now to Figure 4 which shows SWAT simulations of monthly nitrate load at the Salt Creek watershed outlet under various scenarios of percent land use change from corn and soybean in 1:1 rotation to miscanthus, assuming nitrogen fertilizer application rates of 190 kg-N/ha for corn, zero for soybean, and 90 for miscanthus. Note that for all the scenarios, it is assumed that the land converted to miscanthus is distributed randomly but evenly across the watershed. This simplifies the work but should not cause it to lose any significance for the purpose of exploring the potential impact of planting miscanthus on water quality at the watershed scale. In the figure, nitrate load is a stochastic variable expressed in terms of its expectation and standard deviation given the uncertainties in the Salt Creek watershed SWAT model as characterized in the GLUE calibration of the model (16). For each scenario, the figure also gives the time-average of the loads over all months from 2000 to 2008 based on the expectations of the individual months. Refer also to Figure S2 in the SI, which, for easier comparison between the different scenarios, gives the expected monthly nitrate load for 2001 and 2002 for the different scenarios in one graph. The years 2001 and 2002 have been selected for closer 7142
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 18, 2010
examination as they have the highest peak nitrate loads in the study period (2000-2008). As a reference, the results for an all-soybean scenario (where all cropland in the watershed is converted to soybean) are also provided in Figures 4 and S2. The all-soybean scenario represents a best-case scenario as nitrogen fertilization is zero when planting soybean. Table 2 summarizes the results in Figure 4 to give the mean monthly nitrate loads over the simulation period 2000-2008 and their percent decreases from the baseline load for the different scenarios of land use change. For comparison, the loads and percent decreases for nitrogen fertilization rates of 30 and 60 kg-N/ha are also presented. As can be seen from the table, miscanthus cultivation in place of that of corn and soybean in rotation reduces nitrate export from the watershed. For example, a 10% land use change to miscanthus gives on the average, a 6.4% decrease in nitrate export, assuming a nitrogen fertilizer application rate of 90 kg-N/ha. Further, with a 50% land use change to miscanthus, it is possible to achieve up to a 30% decrease in nitrate export. The amount of potential nitrate load abatement also depends on the amount of nitrogen fertilizer applied to miscanthus, though the effect of nitrogen fertilization is not as great as might be expected. For instance, the difference between percent nitrate load decrease for nitrogen fertilization rates of 30 and 90 kg-N/ha is less than 1% when 10% of the watershed is converted to miscanthus. Similarly, the
TABLE 2. Mean Monthly Nitrate Loads over the Simulation Period 2000-2008 and Their Percent Decreases from the Baseline Load for Various Scenarios of Percent Land Use Change to Miscanthus and Miscanthus Nitrogen Fertilizer Application Rate nitrogen fertilizer application rate (kg-N/ha) 30
60
90
% land use change to miscanthus
nitrate load (kg-N/mo/ha)
percent decrease
nitrate load (kg-N/mo/ha)
percent decrease
nitrate load (kg-N/mo/ha)
percent decrease
0 10 25 50
1.602 1.484 1.298 1.054
0 7.32 18.99 34.19
1.602 1.490 1.313 1.083
0 6.97 18.00 32.40
1.602 1.499 1.339 1.128
0 6.42 16.43 29.58
difference between 30 and 90 kg-N/ha is less than 6% when half of the watershed is converted to miscanthus. Another point to note is that the potential for nitrate load abatement of cultivating miscanthus surpasses that of converting the entire watershed to soybean. The all-soybean scenario gives, on the average, about a 23% decrease in nitrate load. Compared to this, a greater decrease in nitrate load is achievable with less than half of the watershed converted to miscanthus, regardless of whether the miscanthus nitrogen fertilizer application rate is 30, 60, or 90 kg-N/ha. When half the watershed is converted to miscanthus, the percent decrease in nitrate load is about 30% when the nitrogen fertilization rate for miscanthus is 90 kg-N/ha; the decrease in nitrate load is even greater when the fertilization rate is smaller.
4. Implications The results obtained in this study may be of interest to policy makers who are considering incentives for the cultivation of second-generation bioenergy crops (e.g., miscanthus) as representing a favorable trade-off between energy production and nutrient runoff. Currently, there is very little or no demand in the U.S. for such crops. Nevertheless, their demand can be expected to grow with the development of the U.S. bioenergy market (35). As the results of this study show, this can be expected to lead to decreases in agricultural nitrogen runoff into surface waters, which for the case of the Salt Creek watershed is mostly via subsurface tile drainage (36). Policy makers should, however, be aware of the trade-off between the water quality benefits of these crops and the land they take away from conventional crops. Consider for instance, the recommendation of 6 to reduce nitrogen discharges into the Gulf of Mexico by 45% to mitigate the hypoxia problem in the Gulf. Based on Table 2, more than half of the Salt Creek watershed would have to be converted to miscanthus for there to be a 45% decrease in its nitrate export (if no other nutrient management measures are in place). This implies that the amount of cropland needed to be converted to miscanthus to ameliorate the Gulf of Mexico hypoxia problem could lead to shortages in food and feed as conventional crops (such as corn and soybean) are displaced. One conceivable way of reaping the water quality benefits of second-generation bioenergy crops without displacing conventional crops is to cultivate the former on marginal lands where the latter yield poorly (37, 38). Miscanthus is generally held to be better able to cope with poor soils where nutrients are limited than corn or soybean, due to its ability to recycle nutrients from the shoot back to the root system at the end of the growing season for use the next growing season (39). However, there simply may not be a sufficient quantity of such land to meet bioenergy demand and nutrient runoff reduction targets. As noted above, more than half of the Salt Creek watershed would have to be planted with
miscanthus to decrease nitrogen export by 45%. However, no more than a small percentage of the land in the watershed may be considered marginal. This work is helpful in furthering the understanding of the environmental impact of miscanthus cultivation; nonetheless, more work is required before a full understanding is possible. In particular, a comprehensive assessment of the effects of crop pattern change on water quality will require additional information such as the effects of climate variability and physical proximity of crop lands to streams.
Acknowledgments This work is funded by the Energy Biosciences Institute (EBI). We thank Frank Dohleman and Carl Bernacchi for sharing their expertise in miscanthus cultivation and Mark B. David for water quality data.
Supporting Information Available Additional information on models of miscanthus growth; additional figures giving more details on the miscanthus yield-nitrogen fertilizer relationship and water quality effects of miscanthus cultivation for the Salt Creek watershed. This material is available free of charge via the Internet at http:// pubs.acs.org.
Literature Cited (1) Hill, J.; Nelson, E.; Tilman, D.; Polasky, S.; Tiffany, D. Environmental, economic, and energetic costs and benefits of biodiesel and ethanol biofuels. Proc. Natl. Acad. Sci. U.S.A. 2006, 103 (30), 11206–11210. (2) Heaton, E. A.; Clifton-Brown, J. C.; Voigt, T. B.; Jones, B. J.; Long, S. P. Miscanthus for renewable energy generation: European Union experience and projections for Illinois. Mitigation Adapt. Strategies Global Change 2004, 9 (4), 433–451. (3) Heaton, E. A.; Dohleman, F. G.; Long, S. P. Meeting US biofuel goals with less land: the potential of Miscanthus. Global Change Biol. 2008, 14 (9), 2000–2014. (4) McLaughlin, S. B.; Ugarte, D. G. D. L.; Garten, C. T.; Lynd, L. R.; Sanderson, M. A.; Tolbert, V. R.; Wolf, D. D. High-value renewable energy from prairie grasses. Environ. Sci. Technol. 2002, 36 (10), 2122–2129. (5) Nelson, R. G.; Ascough, J. C.; Langemeier, M. R. Environmental and economic analysis of switchgrass production for water quality improvement in northeast Kansas. J. Environ. Manage. 2006, 79 (4), 336–347. (6) EPA Science Advisory Board (SAB). Hypoxia in the northern Gulf of Mexico: an update by the EPA Science Advisory Board; U.S. Environmental Protection Agency: Washington, DC, 2007. (7) Arnold, J. G.; Fohrer, N. SWAT2000: current capabilities and research opportunities in applied watershed modelling. Hydrol. Processes 2005, 19 (3), 563–572. (8) Hu, X.; McIsaac, G. E.; David, M. B.; Louwers, C. A. L. Modeling riverine nitrate export from an East-Central Illinois watershed using SWAT. J. Environ. Qual. 2007, 36 (4), 996–1005. (9) Beale, C. V.; Long, S. P. Can perennial C4 grasses attain high efficiencies of radiant energy conversion in cool climates? Plant, Cell Environ. 1995, 18 (6), 641–650. VOL. 44, NO. 18, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7143
(10) Price, L.; Bullard, M.; Lyons, H.; Anthony, S.; Nixon, P. Identifying the yield potential of Miscanthus x giganteus: an assessment of the spatial and temporal variability of M. x giganteus biomass productivity across England and Wales. Biomass Bioenergy 2004, 26 (1), 3–13. (11) Clifton-Brown, J. C.; Neilson, B.; Lewandowski, I.; Jones, M. B. The modelled productivity of Miscanthus x giganteus (GREEF et DEU) in Ireland. Ind. Crops Prod. 2000, 12 (2), 97–109. (12) Clifton-Brown, J. C.; Stampfl, P. F.; Jones, M. B. Miscanthus biomass production for energy in Europe and its potential contribution to decreasing fossil fuel carbon emissions. Global Change Biol. 2004, 10 (4), 509–518. (13) Hastings, A.; Clifton-Brown, J. C.; Wattenbach, M.; Mitchell, C. P.; Smith, P. The development of MISCANFOR, a new Miscanthus crop growth model: towards more robust yield predictions under different climatic and soil conditions. Global Change Biol. Bioenergy 2009, 1 (2), 157–170. (14) Miguez, F. E.; Zhu, X.; Humphries, S.; Bollero, G. A.; Long, S. P. A semimechanistic model predicting the growth and production of the bioenergy crop Miscanthus x giganteus: description, parameterization and validation. Global Change Biol. Bioenergy 2009, 1 (4), 282–296. (15) Sharpley, A. N.; Williams, J. R. EPIC-Erosion Productivity Impact Calculator, 1. Model Documentation; U.S. Department of Agriculture, Agricultural Research Service: Washington, DC, 1990. (16) Ng, T. L.; Eheart, J. W.; Cai, X. M. Comparative calibration of a complex hydrologic model by stochastic methods GLUE and PEST. Transactions of the ASABE 2010, submitted for review. (17) Beven, K.; Binley, A. The future of distributed models: model calibration and uncertainty prediction. Hydrol. Processes 1992, 6 (3), 279–298. (18) Neitsch, S. L.; Arnold, J. G. Kiniry, J. R.; Williams, J. R. Soil and Water Assessment Tool Theoretical Documentation, version 2005; USDA Agricultural Research Service: Temple, TX, 2005. (19) The R Development Core Team. R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing: Vienna, Austria, 2009. (20) Beale, C. V.; Long, S. P. Seasonal dynamics of nutrient accumulation and partitioning in the perennial C4-grasses Miscanthus x giganteus and Spartina cynosuroides. Biomass Bioenergy 1997, 12 (6), 419–428. (21) Monteith, J. L. Reassessment of maximum growth rates for C3 and C4 crops. Exper. Agric. 1978, 14 (1), 1–5. (22) Tayot, X.; et al. Potential above-Ground Dry Matter Production of Miscanthus in North-Central France Compared to Sweet Sorghum. In Biomass for Energy, Environment, Agriculture and Industry; Chartier, P., Beenackers, A. A. C. M., Grassi, G., Eds.; Elsevier: Oxford, UK, 1995. (23) Scurlock, J. M. O. Miscanthus: A Review of European Experience with a Novel Energy Crop; Oak Ridge National Laboratory: Oak Ridge, TN, 1999. (24) Dohleman, F. G.; Long, S. P. More productive than maize in the Midwest: how does miscanthus do it? Plant Physiol. 2009, 150 (4), 2104–2115.
7144
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 18, 2010
(25) Hu, X. Modeling riverine nitrate export from east-central Illinois watersheds. M.S. Thesis, University of Illinois at UrbanaChampaign, Urbana, IL, 2003. (26) Danalatos, N. G.; Archontoulis, S. V.; Mitsios, I. Potential growth and biomass productivity of Miscanthus x giganteus as affected by plant density and N-fertilization in central Greece. Biomass Bioenergy 2007, 31 (2-3), 145–152. (27) Christian, D. G.; Riche, A. B.; Yates, N. E. Growth, yield and mineral content of Miscanthus x giganteus grown as a biofuel for 14 successive harvests. Ind. Crops Prod. 2008, 28 (3), 320– 327. (28) Himken, M.; Lammel, J.; Neukirchen, D.; Czypionka-Krause, U.; Olfs, H.-W. Cultivation of Miscanthus under West European conditions: seasonal changes in dry matter production, nutrient uptake and remobilization. Plant Soil 1997, 189 (1), 117–126. (29) Ercoli, L.; Mariotti, M.; Masoni, A.; Bonari, E. Effect of irrigation and nitrogen fertilization on biomass yield and efficiency of energy use in crop production of Miscanthus. Field Crops Res. 1999, 63 (1), 3–11. (30) Miguez, F. E.; Villamil, M. B.; Long, S. P.; Bollero, G. A. Metaanalysis of the effects of management factors on Miscanthus x giganteus growth and biomass production. Agric. Forest Meteorol. 2008, 148 (8-9), 1280–1292. (31) Davis, S. C.; Parton, W. J.; Dohleman, F. G.; Smith, C. M.; Del Grosso, S.; Kent, A. D.; DeLucia, E. H. Comparative biogeochemical cycles of bioenergy crops reveal nitrogen-fixation and low greenhouse gas emissions in a Miscanthus x giganteus agro-ecosystem. Ecosystems 2010, 13 (1), 144–156. (32) Lewandowski, I.; Clifton-Brown, J. C.; Scurlock, J. M. O.; Huisman, W. Miscanthus: European experience with a novel energy crop. Biomass Bioenergy 2000, 19 (4), 209–227. (33) Bullard, M. Economics of miscanthus production. In Miscanthus for Energy and Fibre; Jones, M. B., Walsh, M., Eds.; James and James (Science Publishers) Ltd.: London, 2001. (34) Huisman, W.; Venturi, P.; Molenaar, J. Costs of supply chains of Miscanthus giganteus. Ind. Crops Prod. 1997, 6 (3-4), 353– 366. (35) Solomon, B. D.; Barnes, J. R.; Halvorsen, K. E. Grain and cellulosic ethanol: history, economics, and energy policy. Biomass Bioenergy 2007, 31 (6), 416–425. (36) McIsaac, G. F.; Hu, X. Net N input and riverine N export from Illinois agricultural watersheds with and without extensive tile drainage. Biogeochemistry 2004, 70 (2), 251–271. (37) Tilman, D.; Hill, J.; Lehman, C. Carbon-negative biofuels from low-input high-diversity grassland biomass. Science 2006, 314 (5805), 1598–1600. (38) Campbell, J. E.; Lobell, D. B.; Genova, R. C.; Field, C. B. The global potential of bioenergy on abandoned agriculture lands. Environ. Sci. Technol. 2008, 42 (15), 5791–5794. (39) Dohleman, F. G.; et al. Perennial grasses as second-generation sustainable feedstocks without conflict with food production. In Handbook of Bioenergy Economics and Policy; Khanna, M., Scheffran, J., Zilberman, D., Eds.; Springer: New York, 2010.
ES9039677
