A spatial model of climate change effects on yields and break-even prices of switchgrass and miscanthus in Ontario, Canada
Aaron V. De Laporte,
Department of Food, Agricultural and Resource Economics, University of Guelph, Guelph, ON, Canada
Correspondence: Aaron V. De Laporte, Department of Food, Agricultural and Resource Economics, J.D. MacLachlan Building, University of Guelph, 50 Stone Road East, Guelph, ON N1G 2W1, Canada, tel. +1 519 824 4120 (Office), fax +1 519 767 1510, e-mail: firstname.lastname@example.org
Concerns over global climate change have led many jurisdictions to implement strategies aimed at reducing greenhouse gas levels. One example is the replacement of coal with dedicated energy crops, such as switchgrass and miscanthus. The yields and costs of these potentially valuable bio-energy crops have been evaluated in only a few cases, and previous studies have not focused on climate change effects. This article assesses the potential yields and costs of growing switchgrass and miscanthus on the agricultural land base in Ontario, Canada, under different climate assumptions, using a GIS-based integrated biophysical and economic simulation model. The model shows that miscanthus has a mean peak yield that is 88.5% (29.6 t ha−1 compared with 15.7 t ha−1) higher and a mean farm gate break-even price that is 25.9% ($58.20 per tonne compared with $73.29 per tonne) lower than switchgrass. The impact of climate change on the yield and break-even price of switchgrass and miscanthus is dependent upon the climate model. CGCM3.1 predicts that mean peak yields of switchgrass and miscanthus could drop by 17.8% and 14.9%, whereas CCSM3.0 predicts that mean yields could increase to 41.4% and 44.9%, from 2071 to 2100, in the A2 climate scenario respectively. Both crops show promise as biomass sources for bio-energy production, but a changing global climate, along with cultivar and planting technology developments, could affect crop choices.
Concerns over global climate change have increased demand for sources of green energy. The substitution of coal for agricultural biomass presents an opportunity to lower greenhouse gas, sulphur dioxide, ash, and mercury emissions (Tillman, 2000), and has the potential to increase soil carbon storage if annual food crops are replaced by perennial biomass crops, including switchgrass and miscanthus (Post & Kwon, 2000). Many green energy policies have been put in place to discourage fossil fuel use in energy production. The number of countries with policy targets for renewable energy increased from 79 in 2008 to 98 in 2010 (REN21, 2011).
One jurisdiction that has adopted green energy policies is the province of Ontario, Canada. In the Green Energy Act of 2009, the Government of Ontario mandates the elimination of all coal-fired electricity production by 2014. According to various Government of Ontario news releases, coal provides approximately 2–3% of the province's energy needs, down from about 25% in the early 2000s. The Ontario Power Generation Corporation (OPG) has been examining agricultural biomass as a potential replacement for a portion of the energy previously provided by coal and estimates their annual demand at approximately 2 million tonnes of dry matter biomass (tDM). Biomass demand could also come from other sources, such as ethanol production or smaller scale biomass burners, especially if the price of oil increases sufficiently. Although it is assumed that dedicated energy crops grown on less desirable agricultural land will be the major source of the necessary biomass, it is not known how much switchgrass and miscanthus could be grown in Ontario.
Even if there is a demand for agricultural biomass, it is not guaranteed that any will be provided. Producer and landowner decision-making on a microeconomic scale will determine the amount of biomass produced. Landowner decisions will hinge on the costs of production, the biomass price offered, the presence of policy incentives, and personal characteristics. Global climate change could also alter many production factors and have a significant impact on the yields and costs of growing dedicated energy crops. As climate change could have an impact on crop yields in Ontario, the relative attractiveness of different agricultural options could change significantly with the climate.
This article considers Ontario's potential ability to grow two perennial warm season grasses for biomass combustion, switchgrass and miscanthus, using a GIS-based biophysical growth model that incorporates site specific climate factors, including temperature, precipitation, and solar radiation. Similar models have been developed in other contexts, including Illinois (Khanna et al., 2008, 2011), the Midwestern US (Jain et al., 2010) and Ireland (Clifton–Brown et al., 2000), but these simulations do not incorporate the potential impacts of different future climate scenarios, or consider the case of Ontario, Canada.
Assessments of possible future climate change are difficult and can be somewhat controversial (Cox & Stephenson, 2007). Many modern climate change predictions are conducted by Atmospheric-Oceanic General Circulation Models (AOGCMs), which are complex computer simulation models incorporating the basic laws of physics, fluid dynamics, and chemistry, in addition to other factors (Heffernan, 2010). Various AOGCMs exist, such as the National Center for Atmospheric Research's (NCAR), Community Climate System Model (CCSM), and the Canadian Center for Climate Modelling and Analysis' (CCCMA) Coupled Global Climate Model (CGCM), but each has multiple stages of development and includes slightly different assessments of the relevant factors in future prediction. Therefore, different results are obtained from different AOGCMs and different iterations of each model. Various authors have explored the differences between these models. The effects of climate on traditional crops in Ontario have also been examined (Cabas et al., 2010). Changes in the growth range of different plant species have been studied (Bertin, 2008; Kelly & Goulden, 2008; Woodall et al., 2009), particularly the climate envelopes of various tree species (e.g. McKenney et al., 2011a). However, the effects of possible climate change on purpose grown biomass crops have been less well studied, especially in the context of Ontario.
The demand for agricultural biomass in Ontario is increasing, and could continue to increase long into the future, when climatic conditions could be different. Therefore, an assessment of the yields and costs of dedicated biomass crops, such as switchgrass and miscanthus, under different climate possibilities, is necessary. The purpose of this study is to assess the yields and costs of two biomass crops, switchgrass and miscanthus, on the Ontario agricultural land base, under different climate conditions, using a GIS-based biophysical and economic model.
Materials and methods
Ontario is located in central-eastern Canada, bordered by the provinces of Manitoba to the west and Quebec to the east. Its borders extend from approximately 56°51′ N in the North, to 41°41′ N in the South, to 74°19′ W in the East, to 95°10′ W in the West. The total area of the province is approximately 917,741 km2, larger than France and the United Kingdom combined. The southernmost parts of Ontario are aligned with, for example, Northern California, Barcelona, Spain, Rome, and Italy, whereas parts of Northern Ontario are in the Subarctic climate zone. The agricultural zones of Ontario are shown in Fig. 1.
Ontario has approximately 2.8 million hectares of land growing annual crops such as corn, soybean, and wheat. Using satellite imagery from 1998 to 2002 at a 30 m × 30 m spatial scale, Fig. 1 shows that agricultural production is focused in the southern region of the province (AAFC, 2009). There are also approximately 2.5 million hectares of land growing perennial crops, including pasture land.
High resolution spatial climate models have been developed to help assess the potential impacts of climate and climate change on Ontario growing conditions (see McKenney et al., 2011b; and Price et al., 2011). This analysis makes use of projections of average minimum and maximum mean monthly temperatures, precipitation, and solar radiation from three climate change scenarios, A2, A1B, and B1, from two different AOGCMs, CGCM3.1, (McFarlane et al., 2005) and CCSM3.0 (Collins et al., 2006). The three climate change scenarios have different projected amounts of aerosols, with A2 (high) resulting in the most extreme changes in climate, followed by A1B (medium), and finally B1 (low) (Solomon et al., 2007). Predicted data are available for three time periods: 2011–2040, 2041–2070, and 2071–2100. The base scenario uses monthly means of daily minimum and maximum temperatures, precipitation, and solar radiation from 1971 to 2000.
From 1971 to 2000, mean daily maximum July temperatures ranged from 16.4 °C to 28.5 °C, with an average of 22.8 °C (Fig. 2; McKenney et al., 2011b). Figure 2 shows the differences in July maximum mean daily temperatures between the scenarios for the CGCM3.1 and CCSM3.0 models (see also McKenney et al., 2011a).
Figure 2 shows the differences in mean daily maximum temperatures between recent normals, different climate models, and different climate scenarios. Normals from 1971 to 2000 show the lowest temperatures as mentioned in the previous paragraph. In CGCM3.1, mean July maximums increase from 22.8 °C in the base case to 25.3 °C in the B1 (Low) and to 27.7 °C in the A2 (High) climate change scenarios, from 2071 to 2100. The CCSM3.0 predicts that mean daily maximum July temperatures increase to 27.6 °C, which is slightly lower than the prediction of CGCM3.1. However, the range of Ontario temperatures is smaller in CCSM3.0 (22.4–32.4 °C) than in CGCM3.1 (21.3–33.2 °C).
From 1971 to 2000 mean total annual precipitation in Ontario ranged from 455 to 1136 mm, with an average of 718 mm (Fig. 3). In CGCM3.1, mean total annual precipitation increased to between 781 mm (B1) and 826 mm (A2). Mean total annual precipitation increased to 822 mm in the A2 scenario of CCSM3.0. From Fig. 3, total annual precipitation appears to have increased to a greater degree in Southern Ontario in CCSM3.0 compared with CGCM3.1, in the A2 scenario.
Growth models for switchgrass and miscanthus similar to those presented here have been used in different contexts, including miscanthus growth in Illinois (Khanna et al., 2008, 2011), the Midwestern US (Jain et al., 2010), Ireland, (Clifton–Brown et al., 2000) and at a coarse European scale (Clifton–Brown et al., 2004). The method is based on the growth principles established by Monteith (1977), using an application of the Beer–Lambert Law of light interception (Monsi & Saeki, 1953). In this model, yield is given by the equation:
where S(t) represents the integral of photosynthetically active radiation, ei represents the interception efficiency, and ec is the energy conversion efficiency. Equation (1) states that the daily yield of biomass is equal to the amount of sunlight received times, the amount of light intercepted, and the plant's ability to build mass using solar energy. The interception efficiency is given by the equation:
where LAI is the leaf area index and k is the radiation extinction coefficient. To calculate total yield in a season, a daily stepping model was constructed for the two crops using temperatures present in the climate data discussed previously. Spatial temperature data were interpolated at a 300 m × 300 m scale using kriging to integrate with the land use data available at a 30 m × 30 m scale. Daily biomass production was summed over the growing season to create total biomass yield at the end of the growing season.
Switchgrass (Panicum virgatum) is a perennial warm season grass found in the remnant tall grass prairies of North America (McLaughlin & Kszos, 2005). As such, its growth in an Ontario context is relatively more certain compared to miscanthus species, which are Asian in origin. For example, Madakadze et al. (1998), examined the growth of switchgrass in Southern Quebec, located near Ontario. Switchgrass growth in this model is adapted from Illinois field data (Jain et al., 2010). For the cave-in-rock cultivar of switchgrass:
where DD10 is growing degree days above 10 °C. The ec for switchgrass is 1.7 g MJ−1 and the extinction coefficient, k, is 0.44.
Giant Miscanthus (Miscanthus x giganteus), an Asian warm season grass, has been identified as a potential source of biomass (Heaton et al., 2008). However, its growth in an Ontario context is uncertain as current field trials have only published data from the second and third growth seasons (Deen et al., 2011). Therefore, growth parameters from Illinois were used (Jain et al., 2010). In this case, LAI is given by the equation:
where DD12 is growing degree days above 12 °C. The ec for miscanthus is 3.4 g MJ−1 and the extinction coefficient, k, is 0.57.
Soil moisture considerations are incorporated into the growth model using the Hargreaves & Allen (2003) method to calculate potential evapotranspiration from temperature and solar radiation data and monthly average precipitation in the various climate scenarios. Due to the lack of available soil information for the entire region being considered, soils were assumed to have a moisture holding capacity of 150 mm (Clifton–Brown et al., 2000). Growth could only occur when soil moisture deficit was less than 150 mm. In this manner, the yield results from the growth models can be considered to be the maximum climate potential yields of switchgrass and miscanthus.
Cost data and break-even calculations
Many supply chain complications exist when attempting to create a fully functioning biomass market and many of these, such as wide-scale farm gate biomass prices, have not been addressed in a North American context. Therefore, this study examines the cost side of growing biomass by examining break-even farm gate prices, as the revenue side is relatively more uncertain.
The costs of many agricultural tasks and products in Ontario are adapted from Vyn et al. (2012). Nitrogen, phosphorus, and potassium costs are estimated at $1.33 kg−1 applied nitrogen, $1.57 kg−1 replacement phosphorous, and $0.96 kg−1 replacement potassium respectively. The cost of fertilizer spreading is fixed at $19.77 ha−1. Fixed harvest costs are $42.01 ha−1, baling is $21 per tonne harvested and on-farm transport and storage costs are $4.5 per tonne harvested. These costs are shown in Table 1.
Table 1. Switchgrass and miscanthus production costs (OMAFRA, 2011; Vyn et al., 2012)
Establishment cost ($ per ha)
Nitrogen cost ($ per kg N)
Phosphorous cost ($ per kg P)
Potassium cost ($ per kg K)
Fertilizer application cost ($ per ha)
Fixed harvest costs ($ per ha)
Bailing ($ per tonne)
On-farm transport and storage ($ per tonne)
Opportunity cost of land ($ per ha)
Switchgrass specific cost estimates are adapted from the Ontario Ministry of Agriculture, Food and Rural Affairs (OMAFRA) (2011), and from Vyn et al. (2012). Establishment costs, which included the costs of primary tillage, secondary tillage, seeding, weed control, and establishment failure are estimated at $868.08 ha−1. Nitrogen is not applied in the first year, but applied at a rate of 100 kg N ha−1 every year thereafter (Khanna et al., 2008). Phosphorous and potassium are applied at a replacement rate based on harvested yield. 0.17 kg of P and 0.72 kg of K are applied for each tonne of switchgrass harvested from a hectare of land (Khanna et al., 2008). The yield loss from a harvest after first frost is assumed to be 20% and the stand life for switchgrass is assumed to be 10 years (Jain et al., 2010). Switchgrass is assumed to have no yield in the first year, 67% of the maximum in the second year, and full yield from year 3 onwards (Khanna et al., 2008).
Miscanthus specific cost estimates are adapted from Vyn et al. (2012) and Virani (2011). Establishment costs, which included the costs of primary tillage, secondary tillage, rhizome planting, and weed control are estimated to be $2628.53 ha−1. The majority of this establishment cost consists of rhizome purchase. Rhizome costs are controversial, given the limited market, with estimates ranging from $0.03 to $0.60 per rhizome (Khanna et al., 2008; Smeets et al., 2009). This study assumes a cost of $0.129 per rhizome at a planting rate of 1.75 rhizomes per m2 (Vyn et al., 2012). Nitrogen is applied at an annual rate of 80 kg N ha−1 (Deen et al., 2011). Phosphorous and potassium are applied at a replacement rate based on harvested yield. 0.3 kg of P and 0.8 kg of K are applied for each tonne of miscanthus harvested from a hectare of land (Khanna et al., 2008). The yield loss from an early spring harvest is assumed to be 20% and the stand life for miscanthus is assumed to be 15 years (Jain et al., 2010). Miscanthus is assumed to have no yield in the first year, 50% of the maximum in the second year, and full yield from year 3 onwards (Khanna et al., 2008). Switchgrass and miscanthus production costs are summarized in Table 1.
Farm gate break-even prices were calculated as the yearly biomass price (PB) needed to break even on a net present value (NPV) farm-level biomass production project, using a 5% real discount rate. In this model, the NPV of a biomass project is given as:
where PVB is the present value of benefits, PVC is the present value of yearly costs, and C0 is the initial or establishment cost. In this case, the present value of benefits is given as:
where Y is the expected crop yield (varying across the province), i is the discount rate (0.05), and n is the number of years of the project. Therefore, the break-even price is given as:
where BEPB is the n year break-even price of biomass.
Minimum, maximum, and mean peak yields (Table 2), and farm gate break-even prices (Table 3), of the Ontario agricultural land grid cells, for the various climate scenarios and models are reported for both switchgrass and miscanthus. The results in Table 2 and Table 3 focus on the current agricultural lands of Ontario only (Fig. 1). In discussions of the potential future yields and break-even prices of these bio-energy crops, the A2 scenario of the CGCM3.1 model will be used for exposition as there is evidence that current climate trends could be closer to this scenario than the A1B and B1 scenarios (The Guardian, 2011).
Table 2. Minimum, maximum, and mean switchgrass and miscanthus peak yields for agricultural grid cells across Ontario for three climate change scenarios in two climate models, including the current base climate scenario
Period end year
Yields ( t ha−1 yr−1)
Table 3. Minimum, maximum, and mean switchgrass and miscanthus farm gate break-even prices for agricultural grid cells across Ontario for three climate change scenarios in two climate models, including the current base climate scenario
Period end year
Break-even price ($ per tonne per year)
According to Table 2, switchgrass mean peak yields are generally lower than miscanthus mean peak yields on the Ontario agricultural landscape. In the base scenario, switchgrass yields between 9.3 and 20.2, with a mean of 15.7 t ha−1 yr−1. Miscanthus mean peak yields range from 15.7 to 38.9, with a mean of 29.6 t ha−1 yr−1, nearly double the yield of switchgrass. According to Table 3, miscanthus break-even prices are generally lower than those of switchgrass. Switchgrass break-even prices range from $61.90 to $108.12, with an average of $73.29 per tonne per year. These values are higher than those of miscanthus, which range between $49.97 and $98.54, with an average of $58.20 per tonne per year.
The results of growth predictions using CGCM3.1, in the A2 scenario, indicate that mean switchgrass peak yields could decrease to 13.2 t ha−1 yr−1 from 2011 to 2040, 12.6 t ha−1 yr−1 from 2041 to 2070, and 12.9 t ha−1 yr−1 from 2071 to 2100 (Table 1). These represent yield decreases of 15.9% (2011–2040), 19.7% (2041–2070), and 17.8% (2071–2100) from the base scenario. Switchgrass mean break-even prices follow the opposite trend, increasing from $73.29 (base), to $81.62 (2011–2040), to $84.19 (2041–2070), and to $82.41 (2071–2100) per tonne per year (Table 3). In other words, switchgrass break-even prices increase from the base scenario by 11.4% from 2011 to 2040, by 14.9% from 2041 to 2070, and by 12.4% from 2071 to 2100. Just as with switchgrass, miscanthus yields could decrease to 24.8 t ha−1 yr−1 from 2011 to 2040, 24.2 t ha−1 yr−1 from 2041 to 2070, and 25.2 t ha−1 yr−1 from 2071 to 2100. These values represent yield decreases of 16.2% (2011–2040), 18.2% (2041–2070), and 14.9% (2071–2100) from the base scenario. Miscanthus break-even prices increase from the base scenario by 9.9% (to $63.97 per tonne) from 2011 to 2040, 11.2% (to $64.74 per tonne) from 2041 to 2070, and 8.0% (to $62.90 per tonne) from 2071 to 2100 (Table 3).
Growth predictions using CGCM3.1 also show that switchgrass yields could decrease to between 13.6 (B1) and 12.9 (A2; A1B) t/ha/y on average between 2071 and 2100, depending on the climate scenario (Table 2). These would represent yield decreases of 13.4% to 17.8%. Similarly, miscanthus yields could decrease to between 26.0 (B1) and 24.8 (A1B) t/ha/y, representing yield decreases of 12.2% to 16.2%. This indicates that miscanthus yields decrease to a slightly lesser degree on average than those of switchgrass, using the CGCM3.1 Climate Model.
Table 3 shows that break-even prices follow the opposite trend of switchgrass yields. Average switchgrass farm gate break-even prices increase to between $79.62 (B1) and $82.90 (A1B) per tonne per year, from 2071 to 2100. These break-even prices are 8.6% (B1) to 13.1% (A1B) higher than those of the base case. Regarding miscanthus, break-even prices increase to between $62.01 (B1) and $63.92 (A1B), representing increases from 6.5% (B1) to 9.8% (A1B).
In each climate scenario, CGCM3.1 generally predicts lower yields and greater break-even prices than does CCSM3.0 (Table 2). More specifically, CGCM3.1 predicts decreases in yield, whereas CCSM3.0 predicts increases in yield on the current Ontario agricultural land base. In CGCM3.1, mean switchgrass yields are 41.9% lower than CCSM3.0, in the A2 scenario, from 2071 to 2100. Similarly, mean miscanthus yields are 41.3% lower. Average break-even prices are 38.7% higher for switchgrass and 30.3% higher for miscanthus in CGCM3.1 compared to CCSM3.0, from 2071 to 2100, in the A2 scenario.
Figure 4 displays the growing potential of switchgrass on the entire Ontario landscape, rather than just on the current agricultural base to provide provincial context and examine shifts in the climate envelope of the species.
Figure 4 shows, in general, that for CGCM3.1, the best growth area for switchgrass moves further North, away from the current agricultural land base. In contrast, for CCSM3.0, the best growth area for switchgrass remains in the current prime agricultural area of Ontario and switchgrass yields increase across the province.
Similar to Fig. 4, Fig. 5 displays the growing potential of miscanthus on the entire Ontario landscape, rather than just on the current agricultural base.
Figure 5 shows the same trend as Fig. 4, where the yields of miscanthus generally increase across the province according to CCSM3.0, whereas the best growing range of the crop appears to move to the North according to CGCM3.1, in the A2 scenario, from 2071 to 2100.
Figure 6 shows the amount of area potentially converted to switchgrass or miscanthus as biomass prices increase, based on the break-even price. More specifically, a parcel of land is assumed to change from its current agricultural use into biomass when the price of biomass exceeds the break-even price for that parcel. Figure 6 is similar in concept to a supply curve, but with simplified conversion assumptions.
Figure 6 shows an upward sloping, generally cubic trend for both switchgrass and miscanthus. Area converted to biomass increases slowly initially as price increases due to the fact that there is a comparatively small amount of very well situated agricultural land in Ontario. Area conversion then increases rapidly over a fairly small segment of price values, as much of the agricultural land in the province has somewhat similar weather characteristics. This is followed by a region of very small increases in area as price changes, where very large increases in price are required to convert more area to biomass. This is due to the comparatively small number of agricultural areas in the northern regions of Ontario.
Using climate-driven production models and historical climate averages from 1971 to 2000, this article provides yields and break-even price estimates for switchgrass and miscanthus grown in the province of Ontario. From Tables 2 and 3, the mean miscanthus yield is 88.5% higher than the mean switchgrass yield and the mean switchgrass break-even price is 25.9% higher. Therefore, it would seem that the production of miscanthus would be relatively more desirable, economically, in the Ontario context. Higher miscanthus yields also imply that a greater overall quantity of biomass could be grown, or a smaller area could be used to grow a specific amount (Table 2). Growing miscanthus in Ontario could be particularly advantageous if transportation costs to major aggregation/demand centres were considered in the model, as miscanthus would need to be transported from fewer locations. However, this miscanthus enthusiasm could be tempered with a discussion of its potential as an invasive foreign species, compared with the relatively more localized (less invasive) growing habitat of switchgrass. Cultivar selection and technological advances could also change the relative attractiveness of the two species. Furthermore, as information regarding miscanthus growth in the Ontario context is uncertain, especially in the long term, miscanthus break-even prices could be significantly higher. According to Vyn et al. (2012), sensitivity analysis on the results of their break-even analysis of miscanthus showed that the break-even price could increase 64.3% if miscanthus has delayed peak yield, early termination, and comparatively expensive rhizome costs, all possibilities in the Ontario context.
The farm gate break-even price ranges examined in this study for switchgrass ($61.90–$108.12 per tonne) and miscanthus ($49.97–$98.54 per tonne) are somewhat higher than those observed by Khanna et al. (2008) for miscanthus in Illinois ($41–$58 per tonne). However, the ranges are similar given the spatial differences in the locations. The results are more similar to Jain et al. (2010), a Midwestern US study ($88.27–$144.19 per tonne switchgrass and $53.25–$153.47 per tonne miscanthus). A study of the Midwestern US covers more of the variability in the Ontario landscape than a comparison to Illinois. There are not really any well functioning markets for biomass feedstock in Ontario to date. Anecdotal evidence shows small home heating and animal bedding markets for these grasses, but they are only grown on a very small scale by a small number of producers. In order for these grasses to be effective in heating, it is advantageous to convert them into pellets or briquettes beyond the farm gate, at significant additional cost. Given the small premium markets, the prices paid for these grasses exceed the farm gate costs and this could encourage production. However, the current demand appears to be low and increasing supply would quickly undercut the attractiveness of current prices, without increases in demand. In order for demand to increase significantly, biomass would have to be used for energy and become relatively more attractive to, for example, coal, nuclear, or natural gas, likely in the presence of policy incentives. The break-even prices observed for switchgrass and miscanthus in this study are not very close to natural gas, which, unlike coal, is not being mandated away by policy in the Ontario context.
In the models examined, the impact of climate change on the yield of switchgrass and miscanthus is not clear (Table 2). Increasing heat units in Ontario extends the growing season; however, increasing heat can also make water more scarce by increasing evapotranspiration. In CGCM3.1, soil moisture appears to be a problem in future scenarios of climate change, limiting the growth of both switchgrass and miscanthus in the current agricultural areas of Ontario, compared to the base model. Mean yields of switchgrass and miscanthus could drop by 17.8% (15.7–12.9 t ha−1) and 14.9% (29.6–25.2 t ha−1), from 2071 to 2100, in the A2 scenario respectively. Figures 4 and 5 reveal the possibility of a changing weather pattern, where a relatively more Northern area of the province becomes more climatically attractive to switchgrass and miscanthus growth. However, this area is in a region of rocky soils that may not be suited for the growth of bio-energy crops, making switchgrass and miscanthus less attractive in the Ontario context.
Yield results from CCSM3.0 are more promising regarding future switchgrass and miscanthus yields. From Table 2, the A2 scenario has a greater impact on switchgrass and miscanthus yield than the A1B and B1 scenarios. Compared to CGCM3.1, the precipitation and temperature changes predicted by CCSM3.0 are more advantageous to mean switchgrass and miscanthus yields on the current Ontario land base. Mean switchgrass yields could increase 41.4% (from 15.7 to 22.2 ha−1) and mean miscanthus yields could increase 44.9% (from 29.6 to 42.9 t ha−1), from 2071 to 2100, in the A2 scenario.
The upward-sloping cubic curve depicted in Fig. 6 is a fairly expected result, given the spatial distribution of heat and solar radiation in Ontario, along with the location of the croplands examined. One of the issues associated with this curve relates to the possibility of mass conversions of agricultural land to switchgrass or miscanthus over a fairly small range of biomass prices. However, this possibility is mitigated by some of the transportation factors considered earlier, non-financial producer decision-making factors, and the fact that soil quality has not been incorporated into this model (Kludze et al., 2013). Soil quality effects would increase variability and likely increase the slope of the curves in the observed horizontal range. It is also possible that some, for example, northern forested areas could be converted to agriculture in future climate scenarios, resulting in a lessening of the slope in the latter portion of the observed curves.
It is difficult to determine whether or not estimated switchgrass and miscanthus growth is within the range of possibilities currently observed in Ontario field trials. According to Deen et al. (2011) approximately 9 tDM ha−1 of switchgrass (Cave-in-Rock) and 19.5 tDM ha−1 of miscanthus (Nagara) was harvested from the Elora research station. This model predicts peak yields of 16.5 t ha−1 switchgrass and 32 t ha−1 miscanthus at the station. Similarly, the model predicts 17 t ha−1 switchgrass and 34 t ha−1 miscanthus, at Ridgetown, whereas the field trial estimates 9.5 tDM ha−1 switchgrass and 23 tDM ha−1 miscanthus. At Simcoe, the model predicts 15 t ha−1 switchgrass and 30 t ha−1 miscanthus, whereas the field trial estimates 10 tDM ha−1 switchgrass and 12.5 tDM miscanthus (2nd year). The field trial results are given in terms of dry matter, whereas the results of the model are of peak theoretical yield. The results of the field trials are also in their second and third years, meaning that the crops may not yet be fully established. Hence, it is difficult to validate the results of the model, but they appear to move with the field trial results. Year-to-year fluctuations in yield are also not considered in the model, given it uses 30 year averages in its calculations, making comparison to second and third year field results more problematic.
The results of this model rely on a number of assumptions. The transfer of model parameters from Jain et al. (2010), an Illinois context, could be problematic, but field data are not yet advanced enough for calibrated Ontario specific data. The selection of cultivars of switchgrass and miscanthus will be particularly important to maximize yields and minimize break-even prices. For example, Vogel & Mitchell (2008) show that Kanlow × Sumner F1 hybrid swtichgrass yields 20.9 tonnes per hectare in Nebraska, between 40% and 70% more than its parent cultivars. The selection of a lowland hybrid of switchgrass (Alamo or Sumner), rather than an upland one (Cave-in-Rock), as used in this study, could also be particularly advantageous to switchgrass yields and break-even prices in the Ontario context (Aravindhakshan et al., 2010). The model does not incorporate the effect of soil quality on yield and does not incorporate a specific soil type. These factors could significantly change yield results, especially where poor soils are present. The use of the agricultural land base for analysis does mitigate this somewhat, as the soils should be at least marginally productive. Regardless, the model shows climatic yield potential for switchgrass and miscanthus. The changes in yield over time only reflect the changing climate. Changes in soil quality, agricultural land area, and crop management techniques have not been considered.
This study provides a detailed, spatially explicit account of potential switchgrass and miscanthus growth in Ontario and an estimation of the break-even costs of acquiring that yield at the farm gate. It also provides estimates of the effects climate change could have on the yields of these two bio-energy crops in Ontario. The incorporation of climate change shows that increased temperatures have an ambiguous impact on yields, due to changing precipitation patterns in different climate models. Overall, both crops show promise as sources of biomass in Ontario, but further research is still necessary to determine when and where the crops would be most appropriately grown.