Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
This study integrates a biophysical model with a county-specific economic analysis of breakeven prices of bioenergy crop production to assess the biophysical and economic potential of biofuel production in the Midwestern United States. The bioenergy crops considered in this study include a genotype of Miscanthus, Miscanthus×giganteus, and the Cave-in-Rock breed of switchgrass (Panicum virgatum). The estimated average peak biomass yield for miscanthus in the Midwestern states ranges between 7 and 48 metric tons dry matter per hectare per year ( t DM ha−1 yr−1), while that for switchgrass is between 10 and 16 t DM ha−1 yr−1. With the exception of Minnesota and Wisconsin, where miscanthus yields are likely to be low due to cold soil temperatures, the yield of miscanthus is on average more than two times higher than yield of switchgrass. We find that the breakeven price, which includes the cost of producing the crop and the opportunity cost of land, of producing miscanthus ranges from $53 t−1 DM in Missouri to $153 t−1 DM in Minnesota in the low-cost scenario. Corresponding costs for switchgrass are $88 t−1 DM in Missouri to $144 t−1 DM in Minnesota. In the high-cost scenario, the lowest cost for miscanthus is $85 t−1 DM and for switchgrass is $118 t−1 DM, both in Missouri. These two scenarios differ in their assumptions about ease of establishing the perennial crops, nutrient requirements and harvesting costs and losses. The differences in the breakeven prices across states and across crops are mainly driven by bioenergy and row crop yields per hectare. Our results suggest that while high yields per unit of land of bioenergy crops are critical for the competitiveness of bioenergy feedstocks, the yields of the row crops they seek to displace are also an important consideration. Even high yielding crops, such as miscanthus, are likely to be economically attractive only in some locations in the Midwest given the high yields of corn and soybean in the region.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Concerns about energy security, exhaustible oil supplies and global warming in the United States have led to an ambitious Renewable Fuel Standard (RFS) of producing 136 billion liters of biofuels by 2022 under the Energy Independence and Security Act (EISA), 2007. Although corn has been the main feedstock used for ethanol production, there is growing recognition that relying on corn as the only feedstock for ethanol is not sustainable because of its impact on the environment and food prices (Abbott et al., 2008; Khanna et al., 2009). As a result, the RFS mandates that 79 billion liters of biofuels should come from noncorn feedstocks and at least 61 billion liters from cellulosic feedstocks, such as crop residues, dedicated energy crops and wood products. Moreover, these dedicated energy crops and crop residue must be harvested from agricultural land cleared or cultivated before December 2007 to qualify for renewable fuel credit under RFS.
Although conversion of cellulosic biomass to fuel is not yet commercially viable, considerable research is underway on high-yielding feedstock sources that could provide abundant biomass for large scale cellulosic biofuel production in the United States and minimize the amount of land that needs to be diverted from food to fuel production. Two perennial C4 crops, switchgrass (Panicum viragatum) and miscanthus (Miscanthus×giganteus), have been identified as among the best choices for low-input and high dry matter yield per hectare in the United States and Europe (Lewandowski et al., 2003; Gunderson et al., 2008; Heaton et al., 2008). There has been field research on switchgrass in the United States since 1991 (McLaughlin & Kszos, 2005). Research on miscanthus in the United States, on the other hand, was initiated in 2002 following the establishment of field trials of miscanthus and switchgrass at the University of Illinois Agricultural Research and Education Centers (Heaton et al., 2008).
Switchgrass is a warm season perennial grass native to North America that has historically been used as forage. The Cave-in-Rock Switchgrass cultivar (hereinafter referred to as switchgrass) studied here is an upland variety that originated in Southern Illinois with excellent potential in the northern states of the United States (Lewandowski et al., 2003). [Cave-in-Rock is more cold tolerant and suited for the Midwest than the more high yielding lowland varieties like Alamo (T. Voigt, personal communication, University of Illinois, Urbana-Champaign)] Miscanthus is a perennial rhizomatous grass; the miscanthus variety being evaluated here as a feedstock for biofuels is the sterile hybrid genotype Miscanthus×giganteus (hereinafter referred to as miscanthus). It has been studied extensively through field trials in several European countries but is nonnative to the United States. [Miscanthus×giganteus is a cross between two species and has three sets of chromosomes instead of the normal two. This prevents the normal pairing of chromosomes needed to form fertile pollen and ovules and makes it sterile. It has been grown in the European Union on a very large scale for over 20 years with no evidence of becoming invasive (for more details, see Long et al., 2007).] The majority of growth for switchgrass occurs during the warm summer months June to August, whereas miscanthus biomass accumulation normally peaks between August and October. These grasses have high efficiency of converting solar radiation to biomass and in using nutrients and water, and have good pest and disease resistance (Semere & Slater, 2007; Clifton-Brown et al., 2008). Field trials indicate that miscanthus has relatively high yields in the US Midwest, more than twice those of switchgrass and higher than miscanthus yields observed in Europe (Heaton et al., 2008; Miguez et al., 2008).
While these perennial grasses have great biophysical potential to help meet future energy needs, the extent to which this potential can be realized will depend on the economic viability of converting existing cropland to produce these crops, as required by the RFS. Farmers will produce these crops only if they can receive an economic return that is at least equivalent to net returns from the most profitable conventional crops. The economic viability of bioenergy crops is likely to differ across locations due to variations in yields, which depend on location specific climatic and biogeochemical conditions, costs of production, and the opportunity costs of land.
This paper examines the biophysical and economic potential of producing bioenergy crops (switchgrass and miscanthus) and the determinants of spatial variability in this potential in the Midwestern US (Iowa, Illinois, Indiana, Michigan, Minnesota, Missouri, Ohio, Wisconsin). This research is interdisciplinary and requires linking biogeochemical models that can simulate energy crop yields with economic decision models that reflect the incentives of landowners to adopt crops that lead to the highest return to the land. Biophysical model simulations are conducted at 0.1° scale, using the Integrated Science Assessment Model (ISAM), and then aggregated to a county level to obtain average yields at the county level for each of these energy crops. These yields are used to calculate the county-specific costs of production and the breakeven price of producing bioenergy crops. This study seeks to advance our understanding of the spatial and temporal dynamics of bioenergy crop yields together with the spatial variability in the breakeven farm-gate price of bioenergy crops across the Midwestern United States.
Few crop productivity modeling studies have estimated the yields for miscanthus and switchgrass. The MISCANMOD has been used in conjunction with a GIS to calculate yields of miscanthus across Europe (Clifton-Brown et al., 2004). The ALMANAC model (Kiniry et al., 1992) is a general crop growth model that has been used in several site-specific studies to estimate the yields of switchgrass (Kiniry et al., 1996, 2005; McLaughlin et al., 2006), but is primarily a single-point model. More recently, Gunderson et al. (2008) used results from various switchgrass growth studies across the United States to statistically estimate the potential for switchgrass production.
Many studies have developed estimates of the costs of producing switchgrass at specific locations in the United States under given assumptions about the yield potential of the crop, such as Iowa (Duffy, 2007), Arkansas (Popp, 2007), Tennessee (Mooney et al., 2009; Epplin et al., 2007) and Indiana (Brechbill et al., 2008). Others report yield and cost data obtained from field experiments in regions, such as Iowa (Hallam et al., 2001), the northern plain states (Perrin et al., 2008), Oklahoma (Haque et al., 2008) and Wisconsin (Vadas et al., 2008). [The assumed switchgrass yield in these studies varies between 9.0 (Duffy, 2007) and 12.3 metric tons dry matter per hectare (t DM ha−1) (Epplin et al., 2007) whereas the observed yields in field trials range widely from 5 t DM ha−1 in the Upper Plain states (Perrin et al., 2008) to 28 t DM ha−1 in Mississippi (Busby et al., 2007). The reported farm-gate costs of switchgrass production also vary significantly from study to study with $90.7 t−1 DM in Duffy (2007) being the highest and $38.9 t−1 DM in Hallam et al. (2001) being the lowest. A few studies have compared the costs of growing switchgrass with other potential cellulosic feedstocks including other herbaceous crops (Hallam et al., 2001), willow and poplar (Turhollow, 2000; Ugarte et al., 2003) and short rotation woody crop (Downing & Graham, 1996), and found these costs to be lower. Busby et al. (2007) compare the costs of growing miscanthus and switchgrass in Mississippi and Oklahoma using data from field trials and find that while the former has a yield and cost advantage in Oklahoma, the latter has an advantage in Mississippi. Khanna et al. (2008) compare the costs of producing switchgrass and miscanthus and their spatial variability across counties in Illinois and find that miscanthus has a consistent cost advantage over switchgrass across Illinois. James et al. (2010) compare the breakeven prices of several herbaceous and woody biomass crops, including switchgrass and miscanthus in Michigan.
This paper builds upon and extends the approaches of previous studies. While we use a similar crop growth modeling approach, we calibrate and evaluate model parameters using observed field data for miscanthus and switchgrass from experimental trials in Illinois. Following calibration and validation, we use this model in combination with measured climate and solar radiation data to predict the potential yields and breakeven prices for miscanthus and switchgrass for a larger geographical area and to provide insights on the factors influencing the spatial variability in these costs across counties in the Midwestern United States.
Models and methods
Crop growth model description
The annual peak dry matter yields (Yp) of miscanthus and switchgrass for each 0.1°× 0.1° grid cell in the United States. Midwestern region are calculated using equations based on the principal developed by Monteith (1977) and used by Clifton-Brown et al. (2000, 2004) to estimate the yield of miscanthus for Ireland:
where St is the integral of incident solar radiation (MJ m−2); ɛi is the efficiency with which the radiation is intercepted by crop canopy, which is defined as a function of radiation extinction coefficient and leaf area index (LAI) (Clifton-Brown et al., 2000).
and ɛc is the efficiency with which the intercepted radiation is converted into biomass energy. The model is used to calculate the yields across the Midwestern region of the United States under both limiting and nonlimiting water resources. The soil water balance between daily rainfall and potential and actual evaporations are calculated using climatic water budget model of Pastor & Post (1985) as implemented by Jain & Yang (2005). The soil hydraulic characteristics for the soil moisture function and the water balance calculations are derived from soil depth and texture information for each FAO soil type (Zobler, 1986, 1999) and relationships between soil texture and water content at the critical pressure (Rawls et al., 1982).
The Yp equation described above is calibrated using observed yields of miscanthus and switchgrass from data collected in 2005 and 2006 crop years from the experimental field trial site located at the University of Illinois Agriculture Research and Education Centers (UIAREC), Urbana, IL site. The model was evaluated for six UIAREC sites across Illinois: DeKalb, Havana, Orr, Brownstown, Fairfield and Dixon Springs. The field trials for DeKalb, Urbana and Dixon Springs were established in year 2002 (Heaton et al., 2008) and for the other sites in year 2004. The data for the Urbana site is available starting 2003, DeKalb and Dixon Springs starting 2005, and for the rest of the sites starting 2006. The seven field sites used in this study are located in North, Central and Southern Illinois, spanning almost 5° of latitudes, about 5 °C variation in mean temperature and a range of soils. Although the data are from Illinois, they are applicable to most of the Midwestern United States, given the similarity of cropland and cropping systems in this region (Heaton et al., 2008). Measures of canopy light interception, leaf area index, intercepted photosynthetic active radiation (PAR), and aboveground biomass are used to calibrate the model as discussed next.
The daily temperature and precipitation data for the years 2000 through 2007 at 0.1°× 0.1° are generated from a monthly 4 km × 4 km resolution grid climate dataset from the PRISM Group continental US analysis (PRISM Group, 2009). Monthly solar radiation is obtained by interpolating point measurements from the US National Solar Radiation Database over 1961–2005 (NREL, 1995, 2007). The daily values for temperature and precipitation are calculated using a climate generator built on MODAWEC (Liu et al., 2009), where monthly precipitation, maximum and minimum temperature, and wet days data are used as an input. Lastly, daily solar radiation is calculated using the WXGEN weather generator implemented in the EPIC productivity model (Sharpley & Williams, 1990).
As shown by Clifton-Brown & Lewandowski (2000), miscanthus rhizomes are killed in laboratory freezing tests in which the specimen is kept at −3 °C for 3 h. The same study also reports the cold effect on the miscanthus yield for four different sites, showing that miscanthus rhizome dies when temperature reaches on average about −3 °C. To account for this, we used data from the NASA Global Land Data Assimilation System (Rodell et al., 2004) to modify miscanthus yields to zero at grid points where the 6 h average soil temperature over the period of 2002–2007 in the 4.5–9 cm soil layer reaches −3 °C. In the case of switchgrass, Nobumasa et al. (2002) find that switchgrass rhizomes die if chilled down to −20 °C. These soil temperatures are not seen in the study region, so we do not account for this effect for switchgrass. Since there are no field trials in the United States high latitudes where the soil temperature could routinely reach this critical temperature, we have estimated yields for with and without cold effect (see ‘Model estimated yield for biofuel crops for the US Midwest region’).
Model calibration and evaluation
Following the method of Clifton-Brown et al. (2000) and using the data discussed above, four model parameters are calibrated. First, we calibrated thermal leaf area coefficient (thc) by regression of LAI on accumulated degree days above base temperature (DDTB) (i.e., LAI=thc× DDTB). The DDTB is calculated with daily minimum and maximum air temperatures. Based on the correlation between LAI and DDTB for different base temperature, we find the highest correlation coefficient of r2=0.90 with a base temperature of 12 °C for miscanthus and r2=0.91 with a base temperature of 10 °C for switchgrass (Table 1). Using these DDTB, the regression values of thc coefficients are found to be 0.0192 for miscanthus and 0.0127 for switchgrass (Fig. 1a and Fig.1c and Table 1).
Table 1. Crop growth model parameters for miscanthus and switchgrass that were calibrated using the measured data
Base temperature (°C)
Thermal leaf area coefficient (unitless)
Light extinction coefficient (unitless)
Radiation use efficiency (g MJ−1 PARi)
Second, the radiation extinction coefficient k is calculated using the relationship between LAI and radiation intercepted coefficient (ɛi) [Eqn (2)].The calculated k for switchgrass and miscanthus is 0.44 and 0.57, respectively (Table 1). As shown in Fig. 1b and e, an LAI of about 5 for switchgrass and 4 for miscanthus is sufficient to intercept about 90% of the incident radiation.
Third, the radiation use efficiency (ɛc) is estimated from the regression of the aboveground dry matter on intercepted radiation (Fig. 1c and f). The estimated ɛc for switchgrass and miscanthus is 1.7 and 3.4 g MJ−1 of intercepted PAR (PARi), respectively (Table 1).
Fourth, thermal time to maturity is calculated by establishing a relationship between latitude and peak biomass. At the Urbana site, with latitude of 40.11°N, the thermal time for miscanthus was determined to be 1260 DDTB, with switchgrass requiring 1320 DDTB to reach maturity.
Finally, the length of growing season is determined by the number of days between the last spring air frost and the first autumn air frost. The air frost threshold temperature is assumed to be 0 °C.
Overall, the model was able to replicate the observed peak yields of switchgrass and miscanthus for the calibration and evaluation sites in Illinois. Considering all six evaluation sites, modeled miscanthus yields were approximately 12% higher, and switchgrass yields were 4% higher than the observed data (see Fig. 2). Average observed peak yields were 37.2 metric tons of dry matter per hectare per year (t DM ha−1 yr−1) for miscanthus and 14.7 t DM ha−1 yr−1 for switchgrass, whereas average modeled yields were 42.5 t DM ha−1 yr−1 for miscanthus and 15.3 t DM ha−1 for switchgrass.
The modeled yields of miscanthus at DeKalb, Urbana and Fairfield are found to be in close agreement with the observed data, with differences being <2%; whereas the modeled yields for switchgrass are in close agreement with observations at the DeKalb, Urbana and Orr sites, with the difference at these sites being <2%. However, we found large differences in calculated (43 t DM ha−1) and observed (17 t DM ha−1) yield for miscanthus at the Brownstown site; the model yields overestimated observed yields by 60%. This can be attributed to poor soil conditions at the site, causing very low observed yields in every year of sample data (F. Dohleman, personal communication). If we exclude the Brownstown site for calculation of average error in miscanthus production, the average error in miscanthus estimation drops to 4%. The largest percentage difference (12%) for switchgrass was found at Brownstown as well. However, the difference between measured and modeled yields is not as large as for miscanthus, suggesting that the poor soil quality effect is greater for miscanthus than for switchgrass, which our model is not able to capture.
We also compare our model estimated yields of switchgrass with those observed at two other sites in the Midwest. Casler & Boe (2003) estimate Cave-in-Rock switchgrass yield at a site in Arlington, WI (lat. 43.33°, long. −89.38°) to be 13.3 t DM ha−1. Lemus et al. (2002) report Cave-in-Rock switchgrass yield of 12.5 t DM ha−1 yr−1 at the McNay Farm, IA (lat. 40.97°, long. −93.43°). Both studies reported the measured yield in the third year after planting. Our model estimated yield at both sites is 13.4 t DM ha−1 yr−1, which is in close agreement with the observed yields.
Both costs and yields of biomass during the establishment years differ from those over the remaining lifetime of these crops. To compare costs of production at different points in time over the life of the crop, and across crops with different lifetimes, we first calculate the present discounted value of the sequence of annual costs, Ct, over the life of each crop using a discount rate of 4%. We similarly calculate the present value of yields, given the sequence of annual yields over the life of the crop, using the same discount rate as above. Note that Yt is yield after losses during harvesting and storage in year t and we refer to the annualized yields after losses during harvesting and storage as yield at farm-gate. The breakeven farm-gate price PB ($ t−1 DM) for each crop is the minimum price per dry metric ton of the bioenergy crop that a cropland owner would need to receive each year to cover all the costs of production over the life of the crop. This price would result in the present value of revenues from the crop being just equal to the PV of costs of producing it over its life as follows:
where T is the life of the crop, Ct is the cost of the bioenergy crop per hectare in period t, d is the discount rate and Yt again is yield per hectare after harvesting and storage losses in year t. Ct includes the cost of producing the crop at time t (Cpt) and opportunity cost of land (CLt), both measured in $ ha−1. We estimate CLt as follows:
where Pct, Qct and Cct are the price ($ t−1), yield (t ha−1) and production cost of corn ($ ha−1), respectively, whereas Pst, Qst and Cst are the corresponding values for soybeans at time t.
The farm-gate production cost (Cpt), of switchgrass and miscanthus include (i) the cost of inputs, such as chemicals, fertilizers and seeds, (ii) the cost of field operations, such as planting and harvesting, and (iii) the costs of storage. Costs of production for each county are obtained using state-specific input prices and machinery costs for 2007. The per hectare costs of land, overhead (such as farm insurance and utilities), building repair and depreciation, and labor are not included in the costs of perennials or row crops since they are assumed to be the same for all crops and do not affect the relative profitability of alternative crops. Instead, these are included as the opportunity costs of using existing farmland, labor and capital to produce bioenergy crops.
The agronomic assumptions about the temporal pattern of crop yields, reseeding rates and input application rates are based on a review of existing agronomic and economic studies presented in Khanna et al. (2008, 2009) and supplemented with information in Brechbill et al. (2008) and Duffy (2007) and experimental field research conducted at UIAREC. Application rates are assumed to be the same across all states in the Midwestern US and are given in Table 2. Some input application rates are dependent on yield and thus their application rate per hectare varies across counties. Other input application rates per hectare do not vary across counties. In practice, fertilizer application rates, replanting probabilities, second-year yields and harvest losses may differ across locations. To allow flexibility in input requirements for these variables we consider two alternative scenarios, a low-cost and a high-cost scenario. The agronomic assumptions for these two scenarios are described in Table 2. The low-cost scenario considers a low fertilizer application rate, low replanting probability, high second-year yield and low harvest loss and economies of scale in harvesting costs whereas the high-cost scenario captures the opposite case of production. In particular, in the high-cost scenario we assume that the costs of baling per ton are the same for switchgrass and miscanthus; thus baling costs per hectare are much higher for miscanthus than for switchgrass given its higher yield per hectare. In the low-cost scenario, we differentiate between the fixed costs of baling (for tractor and other implements) and variable costs (fuel, lube and labor); thus baling costs per ton of dry matter decrease as the tons of biomass per hectare increases [see Figure S1 in the Supporting Information (SI) provided]. These assumptions are used together with input prices and costs of various field operations obtained from Khanna et al. (2008), Duffy (2007), Brechbill et al. (2008) and FBFM (2007, 2008) to construct costs of production for switchgrass and miscanthus in each of the Midwestern states. The costs of production differ in the establishment phase (year 1 for switchgrass and years 1 and 2 for miscanthus) and the maintenance years, defined by the remaining life of the crop (see Table S1 for the breakdown of these costs for various field operations for Illinois). Similar methods were used for calculating the costs for other states, together with state-specific prices of inputs and yields. (Details of the costs of production of miscanthus and switchgrass for the other Midwestern states are available from the authors on request.) These prices are provided in Table S2. Annualized breakdown of costs of production for each crop for each of the Midwestern states, under the assumption of a 10-year life of switchgrass and 15-year life of miscanthus and in the low and high-cost scenarios, are reported in Tables S3 and S4, respectively.
Table 2. Agronomic assumptions for miscanthus and switchgrass production
Miscanthus low-cost – high-cost scenarios
Switchgrass low-cost – high-cost scenarios
Application rate is measured in kg t−1 DM of biomass removed.
**Assumptions about harvesting costs between the low cost and high cost scenarios are also different as explained in the text.
***Farm-gate yield is defined as annualized yield after losses during harvesting and storage.
There is considerable uncertainty about the cost of miscanthus rhizomes as they are not yet commercially available for large scale plantations. Personal communication with developers and producers of miscanthus varieties, rhizomes and plugs indicates that they expect miscanthus plugs to cost between $0.30 and $0.80 plug−1 as they begin commercial sales in 2010 and close to $0.25 plug−1 by 2011. For a large scale plantation with 10 000–12 000 plugs ha−1, the costs of establishment (including labor, equipment and chemicals) are expected to be $3000–$4000 ha−1. (Personal communication with representatives from Pyramid Farms Ltd, Ontario, Canada, Mendel Biotechnology, California and with Tom Voigt, University of Illinois.) Plugs are vegetatively produced from rhizomes and have a dominant meristem. Costs are expected to be lower for rhizomes compared with plugs and to decrease as the scale of production increases. Cost of propagating rhizomes at the University of Illinois is estimated to be $0.10 rhizome−1. We assume a rhizome cost of $0.25 and a planting rate of 10 000 rhizomes ha−1. Our costs of establishment, including planting, fertilizer and chemical costs amount to $2957 ha−1 in Illinois. We examine the sensitivity of the breakeven prices of miscanthus to a doubling of the cost of rhizomes by considering a scenario with $0.50 rhizome−1. We also consider a scenario where these costs are $0.10 rhizome−1.
We define CLt as the foregone profits per hectare from the next best alternative use of that land. In the case of cropland, this alternative use is defined by choice of crop, rotation and tillage practice. A landowner is assumed to be using the land in its most profitable use, which in turn depends on the yields of the alternative crops, their prices and costs of production (which in turn depend on prices of inputs such as fertilizers and chemicals). The most profitable land use and the opportunity cost of the land is likely to change over time (with changes in input and output prices and yields) and across locations (due to spatial variation in yields of alternative crops).
We estimate the costs of corn and soybean production for each county in the Midwestern United States using information from the crop budgets compiled for that state by state extension services. The revenue from corn and soybean production is estimated using the county-specific 5-year (2003–2007) average yields and state-level price statistics published by the National Agricultural Statistics Service (NASS) for 2007. The sources for this data and the assumptions underlying the estimation of the opportunity costs of land under representative conditions for each of the Midwestern states are reported in Table S2. For comparison, we also estimated the profits from wheat and hay production (Table S2). We find that, with the exception of Michigan, a corn–soybean rotation with conventional tillage is the most profitable land use on average in each of the Midwestern states. (Corn production with conventional tillage or no-till typically involves about the same costs per hectare. However, soybean production costs with no-till are typically higher than with conventional till due to higher seed and chemical costs which more than offset the lower fuel and machinery costs. We assume that yields per hectare for each crop are the same under conventional till and no-till. For more details regarding input and fuel uses under different types of till, see http://www.ers.usda.gov/Data/ARMS/app/Crop.aspx and http://ecat.sc.egov.usda.gov/.) Average profits range between $366 ha−1 in Michigan and $785 ha−1 in Illinois (Table S2). Not surprisingly, it has a dominant share in total crop acreage in the Midwest. In Iowa, its share in total crop acreage is 84%; on average its share is 53% in total crop acreage in the eight Midwestern states (Padgitt et al., 2000). An exception to this is Michigan in which only 22% of acreage is in a corn–soybean rotation; on average, we find that hay alfalfa is about 13% more profitable (on average) than the corn–soybean rotation.
Given the diversity of observed crops, rotations and tillage practices within and across locations in the Midwest, it is evident that a corn–soybean rotation with conventional tillage is not the most profitable land use in all locations. To the extent that other land uses are more profitable than a corn–soybean rotation, the opportunity cost of converting that land to bioenergy crops would be higher. We use representative profits (for each county) from a corn–soybean rotation with conventional tillage to provide an estimate of the cost of converting land currently under a corn–soybean rotation to bioenergy crops in a county. This opportunity cost would be higher if corn residues have value as a feedstock for biofuels or biopower. We discuss the sensitivity of our breakeven price estimates to assumptions about the discount rate, input and output prices, yields and the use of corn residues below.
Even with a corn–soybean rotation, the opportunity cost of land differs across counties in a state since the yields and production costs of corn and soybeans differ across counties. The map in Fig. 3a shows the spatial variation in the per hectare yield of corn across the Midwestern United States. Yields of corn are much higher in central and northern Illinois and in Iowa and relatively lower in Missouri. The spatial distribution of soybean yields per hectare is highly correlated with corn yields per hectare and is not shown for brevity. Figure 3b shows the estimated opportunity cost of corn–soybean land in terms of dollars per hectare in this region. In general, states with a high corn and soybean yield per hectare also have a high opportunity cost per hectare of using that land for bioenergy crops. Illinois has the highest opportunity costs of land as its corn and soybean yields per hectare and prices are high while the costs of production of corn and soybeans per hectare are relatively low. On the other hand, Michigan has the lowest opportunity cost of land per hectare under a corn–soybean rotation due to its low corn and soybean yields per hectare and relatively high costs of production per hectare of these crops.
Model estimated yield for biofuel crops for the US Midwest region
While the model was run for the entirety of the continental United States, this paper presents estimates for the Midwestern United States only. In regions such as the Midwest, model results are largely driven by input temperature and solar radiation. On a per hectare basis, our model results suggest that the miscanthus peak biomass yield in the Midwest is about three times the yield of switchgrass. The estimated peak biomass yield for miscanthus in the Midwest ranges between 0 and 62 t DM ha−1 yr−1 and for switchgrass ranges between 8 and 40 t DM ha−1 yr−1. The model results show a north to south gradient, which suggest that yields in this region are largely driven by temperature and solar radiation (Fig. 4). As such, Missouri and the southern portions of Illinois and Indiana show strong yields for miscanthus and switchgrass, with northern Minnesota and Michigan's Upper Peninsula being the region where the estimated yields for both biofuel crops are the lowest (Fig. 4). As discussed in Clifton-Brown & Lewandowski (2000), extremely cold soil temperatures will kill rhizomes of Miscanthus×giganteus. Using the criteria implemented in the model, most of Minnesota, Wisconsin and the Michigan Upper Peninsula are shown to be very poor performers in miscanthus production, with averaged county-level production being zero for much of this region of the Midwest. Given the lack of field data from these regions and uncertainty of the degree to which there is cold tolerance, yields could be in between the range of values provided here.
Water limitation, as implemented in the model has only a small effect on yields of miscanthus and switchgrass. Over the study region, the average yield increases by 1.0 t DM ha−1 for switchgrass and 3.1 t DM ha−1 yr−1 for miscanthus when water limitation is disabled. The spatial pattern of yield is the same with and without water limitation across the Midwest.
Without some frost in late autumn, miscanthus does not senesce properly (Clifton-Brown et al., 2001a), reducing translocation of nutrients from the aboveground biomass to the rhizome. This can result in nutrient limitation as the aboveground biomass is harvested along with the necessary nutrients for next year's growth. This effect is simulated within the model, but it does not affect yields within the study region. While year to year variations in climate do not change the regional pattern of yield they do affect peak yield levels, our model results indicate that the variation in peak yields for the two biofuel crops was between ±5% and 10% due to variations in local climate over the period 2003–2007 (Table 3).
Table 3. Crop growth model estimated peak miscanthus and switchgrass biomass yields averaged for the period 2003–2007
Cold effect yield (t DM ha−1 yr−1)
No cold effect yield (t DM ha−1 yr−1)
Yield (t DM ha−1 yr−1)
Cost of bioenergy crops in Midwestern United States
The state-level estimates of costs as well as breakeven prices in the low-cost and the high-cost production scenarios are reported in Table 4 and the results of the low-cost production scenario are also depicted in Fig. 5. The costs of production of switchgrass in the low-cost scenario ranges from $39 to $58 t−1 DM across this region while the costs of miscanthus vary more widely between $34 and $80 t−1 DM among the Midwestern states. In the high-cost scenario, these costs range between $62 and $90 t−1 DM for switchgrass and $58 and $131 t−1 DM for miscanthus. In general, the costs of production for miscanthus are considerably lower than those for switchgrass (Table 4) except for the Northern Midwestern states such as Minnesota and Wisconsin, which have significantly low miscanthus yields. Although the establishment cost for miscanthus is much higher than for switchgrass, its annualized cost of production per dry ton is still lower than that of switchgrass primarily because of its higher yield (more than twice that of switchgrass on average) and longer lifetime (15 years instead of 10 years for switchgrass). As a result of the difference in yield per hectare of miscanthus and switchgrass, the opportunity cost of land per dry ton for these two crops differs greatly at any location. It accounts for a large part of the total cost per dry ton in the Midwest. This opportunity cost of land also differs across states due to differences in the productivity of conventional crops across these states.
Table 4. Costs of bioenergy crop production (US$ t−1)
Production cost (US$ t−1 DM)
Opportunity cost of land (US$ t−1 DM)
Breakeven price (US$ t−1 DM)
Share of opportunity cost in breakeven price (%)
Production cost (US$ t−1 DM)
Opportunity cost of land (US$ t−1 DM)
Breakeven price (US$ t−1 DM)
Share of opportunity cost in breakeven price (%)
The opportunity costs of land per dry ton range between $46 and $92 t−1 DM across the Midwestern states in the low-cost scenario for switchgrass and between $19 and $74 t−1 DM in the case of miscanthus. Some states have much higher opportunity costs of land per dry ton for switchgrass or miscanthus because they have a much lower yield per unit of land than others (Fig. 4). Iowa has the highest opportunity cost of land ranging between $92 and $103 t−1 DM for switchgrass and Minnesota has the highest opportunity cost of land for miscanthus at $74–$103 t−1 DM across the low and high-cost scenarios. Michigan has the lowest opportunity cost of land ranging between $46 and $54 t−1 DM for switchgrass and Missouri has the lowest cost of land for miscanthus at $19 and $27 t−1 DM.
A comparison across states shows that Missouri has the lowest breakeven price per dry ton for both bioenergy crops; Minnesota and Iowa have the highest breakeven price per dry ton for switchgrass due to its high opportunity cost of land and Minnesota and Wisconsin have highest breakeven costs for miscanthus due to their much lower miscanthus yield levels (Table 4). In Missouri, the breakeven price of production of switchgrass ranges between $88 and 118 t−1 DM whereas that of miscanthus ranges between $53 and 85 t−1 DM. The average breakeven price of switchgrass, across all Midwestern states, is $115 t−1 DM whereas that of miscanthus is $88 t−1 DM, in the low-cost scenario (Table 5).
Table 5. Sensitivity of breakeven prices for the low cost scenario*
Percentage changes are calculated relative to the breakeven price for that crop in the baseline case provided in row 1 of this table.
This reflects the lack of change in opportunity cost of land for Missouri where corn stover production is not profitable at the breakeven price of miscanthus of US$52.95 t−1 DM calculated here.
(US$ t−1 DM)
25% Increase in corn-soybean price
(US$ t−1 DM)
25% Decrease in corn-soybean price
(US$ t−1 DM)
25% Increase in preharvest cost
(US$ t−1 DM)
25% Decrease in preharvest cost
(US$ t−1 DM)
25% Increase in harvest cost
(US$ t−1 DM)
25% Decrease in harvest cost
(US$ t−1 DM)
25% Increase in fertilizer price
(US$ t−1 DM)
25% Decrease in fertilizer price
(US$ t−1 DM)
10% Increase in bioenergy crop yield
(US$ t−1 DM)
10% Decrease in bioenergy crop yield
(US$ t−1 DM)
Change discount rate from 4% to 8%
(US$ t−1 DM)
Change rhizome price from US$0.25 to US$0.50 rhizome−1
(US$ t−1 DM)
Change rhizome price from US$0.25 to US$0.10 rhizome−1
(US$ t−1 DM)
Change life time from 15 to 10 years for miscanthus
(US$ t−1 DM)
Inclusion of corn stover profits in opportunity cost of land
Corresponding figures for the high-cost scenario are $151 t−1DM for switchgrass and $136 t−1DM for miscanthus.
If miscanthus is more cold tolerant than assumed here, the average yield of miscanthus would change from <10 to 25.65 t DM ha−1 yr−1 in Minnesota and to 26.79 t DM ha−1 in Wisconsin (Table 3). At these new yield levels, the breakeven prices of miscanthus would decrease dramatically (by about 45%) from $153–$234 t−1 DM to $86–$134 t−1 DM in Minnesota and from $140–$211 t−1 DM to $78–$120 t−1 DM in Wisconsin. If we exclude Minnesota and Wisconsin, then the breakeven price of production of miscanthus across the remaining states ranges between $53–$79 t−1 DM in the low-cost scenario and $85–$128 t−1 DM in the high-cost scenario. When the two states are excluded, the average breakeven price of miscanthus in the remaining Midwestern states is $68 t−1DM in the low-cost scenario and $107 t−1DM in the high-cost scenario.
Mooney et al. (2009) estimate the costs for the Alamo variety in Tennessee with a yield of 17.7 t DM ha−1 yr−1 to range between $46 and 67 t−1 DM. This includes an estimate of land costs in Tennessee that ranges between $9 and 19 t−1 DM. Our estimates for the costs of switchgrass production per ton (excluding land costs) for Missouri, in Table 4, are very similar to Mooney et al. (2009) ($39–62 t−1 DM). However, our estimated breakeven price in Missouri is higher than their estimate for Tennessee, partly due to a somewhat lower yield (15.5 t DM ha−1 yr−1) and partly due to a much higher cost of land ($49–56 t−1 DM of switchgrass). Our breakeven prices for switchgrass production in Indiana, $103–138 t−1, are more than twice as high as those in Brechbill et al. (2008), $54–57 t−1, in large part due to our higher costs of land ($61–67 t−1 DM as compared with their cost of land of $14 t−1 DM). James et al. (2010) estimate of switchgrass breakeven price of $115 t−1DM in Michigan (with a yield of 9 t−1 DM ha−1 yr−1) is within the range we find for Michigan, between $102 and $142 t−1 DM (with an average yield of 10 t DM ha−1 yr−1). Their estimate of the breakeven price of miscanthus in Michigan ranges between $45 and $200 t−1DM as rhizome costs vary between $0.5 and $1.80 rhizome−1 and qwith an assumed life of 10 yrs and a yield of 22.4 t−1 DM ha−1 yr−1for miscanthus. Our breakeven prices for miscanthus production in Michigan range between $79 and $128 t−1DM; while we have lower rhizome costs and a longer lifetime for miscanthus, we also consider a lower yield of 17 t DM ha−1 yr−1 that raises costs of production.
The spatial distribution of county-level breakeven prices of the bioenergy crops under the low-cost scenario is shown in Fig. 6. In general, switchgrass is more costly to produce than miscanthus in most of the Midwestern counties but for some counties in Minnesota, Wisconsin and Michigan, miscanthus is much more expensive than switchgrass due to the lack of yield advantage and its high establishment cost. Southern Missouri has the lowest breakeven price of energy crop production ($40 t−1 DM of miscanthus) and in most Midwestern counties, the breakeven prices of miscanthus are <$80 t−1 DM. On the other hand, the breakeven prices of switchgrass in most of the Midwestern counties are around $100 t−1 DM or above and; as these are largely determined by the opportunity cost of land, the breakeven prices in Iowa and Northern Illinois are the highest while in Missouri and Northern Minnesota and Wisconsin the breakeven prices of switchgrass are the lowest. These results suggest that there is considerable spatial variation in the breakeven prices of cellulosic feedstocks in the Midwest and that some areas are better suited to provide low price and high yield feedstock than others and will specialize in the production of different bioenergy crops. Bioenergy crops are more likely to be attractive to landowners in locations with high yields of bioenergy crops and with low yields and high costs of production of conventional crops they seek to displace.
We analyze the sensitivity of the breakeven price, including the opportunity cost of land, to several assumptions made above (Table 5). We report and discuss the results for the low-cost scenario only since there is not much difference between the effects in the low and high-cost cases. The opportunity cost of the land is affected by changes in the price of corn and soybeans and a 25% change in the prices of corn and soybeans changes the breakeven price (on average across states) by 27% (ranging between 16% and 32%) for switchgrass and 20% (between 11% and 23%) for miscanthus. The breakeven price of miscanthus is somewhat less sensitive to the price of corn and soybean because its higher yield per hectare allows changes in fixed costs to be spread over a larger number of tons per hectare. Breakeven prices are not sensitive to assumptions about costs for preharvest and harvest operations. A 25% increase in preharvest costs would increase the breakeven price of switchgrass and miscanthus by only 0.4% on average (between 0.3% and 0.5%) and the same percentage increase in harvest costs would increase the breakeven price by about 6% on average (between 5% and 8%) for switchgrass and 7% (between 4% and 10%) for miscanthus. A 25% increase in fertilizer costs also has only small effects on the breakeven price of either crop, which would increase on average by 3% (between 2% and 5%) for switchgrass and around 2% (between 1% and 3%) for miscanthus. Changing discount rate from 4% to 8% would have a very modest impact on the breakeven price of switchgrass (0.5% increase on average and varying between 0.4% and 0.7%) but a more remarkable impact on the price of miscanthus (7.9% increase and varying between 6.7% and 9.6%) because miscanthus has much higher establishment costs. A 10% increase in bioenergy crop yield per hectare reduces the breakeven prices of both switchgrass and miscanthus by 7% on average. When the lifetime of miscanthus is shortened from 15 to 10 years, the breakeven price of miscanthus would increase on average by 12% (between 11% and 15%). Moreover, the breakeven price of miscanthus is also very sensitive to rhizome price. If rhizome costs are $0.50 rhizome−1 instead of $0.25 rhizome−1, the breakeven price of miscanthus would increase on average by 19% (between 15% and 25%). If the rhizome price is $0.10 rhizome−1 then the breakeven price of miscanthus would decrease by 11% on average. In general, the breakeven price of bioenergy crops is more sensitive to the assumptions examined above in the states with low-crop yields than in the states with high-crop yields. We also find that in all scenarios examined here, the average (and the minimum) breakeven price of switchgrass is higher than that of miscanthus in the Midwestern states (Table 5). However, in almost all cases, the maximum estimated price of switchgrass is lower than the maximum price for miscanthus. These high prices for miscanthus are typically those observed in two states, Minnesota and Wisconsin, where yields are very low.
In addition to the factors considered above, such as crop prices and fertilizer prices, the opportunity cost of land for bioenergy crops will also be affected by the value placed on corn stover for cellulosic biofuel production. We examine the impact of including this value in the opportunity cost of land per dry ton for energy crops using the approach in James et al. (2010) and calculate the breakeven price of perennial grasses assuming equal biomass price across cellulosic sources. Corn stover costs depend on a number of factors: stover yields, the amount of replacement nutrients to be applied, the proportion of the residue that can be sustainably harvested and the price of the corn stover. We estimate the costs of harvesting stover, using assumptions about yields and nutrient replacement rates in Sheehan et al. (2003), a residue collection rate of 30% with a corn–soybean rotation and conventional tillage in each state. We then find the breakeven price of perennial grasses assuming the same price would apply to stover. We find that the inclusion of corn stover in the opportunity cost of land raises the average breakeven price for switchgrass by 7% and for miscanthus by 3% across the Midwestern states. The maximum increase in the breakeven price is 11% for switchgrass and 9% for miscanthus. (The breakeven price for miscanthus is relatively low in Missouri while the costs of harvesting corn stover are relatively high. As a result, inclusion of corn stover production does not add value to the opportunity cost of land in Missouri) (see last row, Table 5).
This paper integrates a biophysical model of bioenergy crop yields with economic analysis of the breakeven prices of bioenergy crop production to assess how these prices differ across bioenergy crops and across different locations in the Midwestern United States. The crop growth model was calibrated with a range of site-specific observed field data for miscanthus and switchgrass. This model is then used in combination with measured climate and radiation data to obtain yields of switchgrass and miscanthus in the Midwestern United States. This together with information about returns to land from existing crops (corn and soybeans) grown in the Midwestern states is used to calculate the breakeven prices of these bioenergy crops at a county-specific level.
Our crop yield model shows that miscanthus yields on average were more than two times higher than those of switchgrass. The estimated yields were substantially higher in southern counties and lower in northern counties. Our modeling results suggest that the continental climate of the United States. Midwest with warmer and wetter summers drive higher yields while its consistently cold winters allow greater translocation of nutrients to the rhizome, which stimulates productivity of biofuel crops. However, extremely cold temperatures in the north could freeze the soil deep enough to kill miscanthus rhizomes over the winter, limiting the crop's potential in northern latitudes. Our model results suggest that rainfall in the US Midwestern states is normally a nonlimiting factor for biofuel crop growth, although peak yield amount could vary from year to year depending upon other climate conditions.
A comparison of model-calculated yields with observed data shows close agreement between the two on average and at a site-specific level for most sites in the study. The model's limitations in explaining yields at some sites, such as the Brownstone site in Illinois, may be because the model does not capture the yield-limiting effects of soil nutrient deficits. Better parameterization of the effects of soil nutrient limitations on the growth of perennial grasses may be required to improve model performance.
Our economic analysis shows that breakeven prices of bioenergy crops are critically dependent on the yields of bioenergy crops and the yields of the annual crops they are likely to displace. High yields of bioenergy crops reduce the per ton opportunity cost of land. The latter is particularly high in locations where the yields of corn and soybeans are high and a land owner would need to forgo considerable returns by switching to an energy crop. Moreover, low fertilizer application rate, low replanting probability, high second-year yield, and low harvest loss also play a significant role in reducing the breakeven prices of bioenergy crop production. We find that the breakeven price, which includes the cost of producing the crop and the opportunity cost of land, of miscanthus ranges from $53 t−1 DM in Missouri to $153 t−1 DM in Minnesota in the low-cost scenario. Corresponding costs for switchgrass are $88 t−1 DM in Missouri to $144 t−1 DM in Minnesota. Breakeven prices at the upper end of this range are primarily due to high opportunity cost of land or low yields of bioenergy crops. On average, the breakeven price of switchgrass in the Midwest is $115 t−1 DM whereas that of miscanthus is $88 t−1 DM in the low-cost scenario. If we exclude Minnesota and Wisconsin as being unviable for producing miscanthus due to their low yields per hectare, then the average breakeven price of miscanthus in the remaining Midwestern states is $68 t−1 DM in the low-cost scenario and $107 t−1 DM in the high-cost scenario.
We find that the breakeven prices of miscanthus and switchgrass are sensitive to factors that change the net returns to corn and soybean and thus the costs of converting that land to bioenergy crops. The additional value added to the value of land by including the net returns to corn stover depends critically on assumptions about the cost of corn stover, but is likely to be relatively small compared with the value of the land derived from corn and soybeans. The returns to miscanthus are also sensitive to the cost of rhizomes and the life span of miscanthus. In all the scenarios examined, we find that the breakeven prices per ton are on average lower for miscanthus than for switchgrass.
Our economic analysis underestimates the breakeven prices in several ways. It does not consider other land uses (besides corn–soybean rotation) that might be more profitable in some parts of these states. It also does not consider the effects of diverting land towards bioenergy crops and reducing production of corn and soybeans, thereby raising their prices and the opportunity cost of converting that land. In other cases, our analysis could be overestimating these breakeven prices. We do not consider other sources of returns to bioenergy crops, such as the value of carbon credits generated through the soil carbon sequestration achieved by them or the savings in greenhouse gas emissions realized by using these crops as feedstocks to produce biofuels and displacing gasoline. We also do not include other returns from bioenergy crops, such as the value of switchgrass for forage or seed production and the potential to produce biochemicals. The opportunity cost of land considered here could be lower if marginal land or pasture land is used for bioenergy crop production; this would result in lower breakeven prices of bioenergy crops.
Our analysis focused on the farm-gate breakeven prices and did not take into account the cost of transporting biomass from farm to the biorefinery or biomass-fired power plant. Searcy et al. (2007) estimate this cost to be $12.75 t−1 DM for a 100 km roundtrip. As this cost is the same for switchgrass and miscanthus, it will not affect their relative breakeven price per ton if they are transported over the same distance. However, to the extent that the transportation distance for miscanthus and switchgrass differ due to the size of the catchment area from which biomass is collected to meet the needs of a biorefinery/power plant, the relative costs of the two grasses could be affected.
Our analysis does, however, show some of the factors that influence the breakeven prices of bioenergy crops and explains the causes of differences in breakeven prices across two bioenergy crops and across locations in the Midwestern United States. We can use the breakeven price estimates provided here to project the breakeven prices of biofuel derived from switchgrass and miscanthus. With an average breakeven price of miscanthus of $68 t−1 DM in the low-cost scenario and $107 t−1 DM in the high-cost scenario (for the Midwest excluding Minnesota and Wisconsin), the cost of cellulosic ethanol is estimated to range between $$0.63 L−1 for the low-cost scenario and $0.74 L−1 for the high-cost scenario. [This calculation is based on the assumption of 330 L of biofuels per ton of dry matter of biomass and a cost of fuel conversion minus coproduct credit for a biorefinery with 189 million liter annual capacity of $0.38 L−1 and is based on pilot demonstrations by National Renewable Energy Laboratory (NREL) (Wallace et al., 2005). It includes the round-trip cost of $12.75 t−1 DM of transporting biomass from the farm to a biorefinery located at a distance of 50 km.] In comparison, the breakeven price for biofuel from switchgrass ranges between $0.77 and $0.88 L−1 in the low and high-cost scenario, respectively, with an average breakeven price of switchgrass of $115 t−1DM in the low-cost scenario and $151 t−1DM in the high-cost scenario across all Midwestern states. Our analysis shows that the costs of cellulosic biofuel production will depend on the yield of the feedstock used, the region where that feedstock is produced and the existing land uses it displaces as well as on the agronomic and economic production conditions that emerge with large scale production of these bioenergy crops.
This research has been funded in part by the Energy Biosciences Institute, University of Illinois, the US Department of Energy under Grant ID ER64503-1030822-0013857, and the US National Science Foundation Grant ID NSF-EFRI-083598.
The soil temperature data used in this study were acquired as part of the mission of NASAs Earth Science Division and archived and distributed by the Goddard Earth Sciences (GES) Data and Information Services Center (DISC). We would also like to thank Tom Voigt for valuable comments and information from field trials of miscanthus and switchgrass at the University of Illinois.