Assessment of the Predicted Biomass Production in the Billion Study ...

An ASABE Meeting Presentation Paper Number: 152181831

Assessment of the Predicted Biomass Production in the Billion Study Update Daniela S. Gonzalesa, Stephen W. Searcya, Laurence M. Eatonb a

Department of Biological and Agricultural Engineering, Texas A&M University, TAMU 2117, College Station, TX 77843-3131, USA b

Environmental Sciences Division, Oak Ridge National Laboratory, P.O. Box 2008, Oak Ridge, TN 37831-6335, USA

Written for presentation at the 2015 ASABE International Meeting Sponsored by ASABE New Orleans, LA July 26-29, 2015 (The ASABE disclaimer is in a table which will print at the bottom of this page.) Abstract. The expansion of the renewable fuels industry in the US requires advances on a number of focus areas to ensure a reliable production and feedstock delivery system. The US DOE published a strategic analysis that estimates if the US agriculture and forest resources have the capability to produce at least one billion dry tons of biomass annually, in a sustainable manner. The Billion Ton Study (BT2) provides a means to predict the development of the biofuels industry. The estimates of potential biomass were obtained through the Policy Analysis System (POLYSYS) agricultural modeling framework and, were based on numerous assumptions about current and future inventory, production capacity, availability, and technology. The objective of this study is to assess the underlying assumptions for the range of conditions across a nation and determine their influence on total biomass predicted. Particularly, we look at the following assumptions: the production of perennial grasses is limited to rain-fed lands classified as cropland, cropland pasture and permanent pasture, and all conversion from pastureland to perennial grasses is limited to counties east of the 100th Meridian. The biomass predictions for the states bisected by the 100th Meridian were questioned based on the knowledge of local conditions. In collaboration with the Oak Ridge National Laboratory, we have carried out additional simulations in the POLYSYS framework, using the annual average precipitation data as a bound for land conversion to perennial grasses. We concluded that the potential perennial grass production in the US, and particularly in Texas, was overestimated by allowing cropland in regions with less than 635 mm (25 in.) precipitation to be converted into perennial grasses. The perennial grass production was reduced in the US and Texas by 7.7 % and 86.8%, respectively under the farmgate price of $60 per dry ton scenario. Keywords. Biomass, Billion Ton Study, Perennial grasses, Switchgrass

The authors are solely responsible for the content of this meeting presentation. The presentation does not necessarily reflect the official position of the American Society of Agricultural and Biological Engineers (ASABE), and its printing and distribution does not constitute an endorsement of views which may be expressed. Meeting presentations are not subject to the formal peer review process by ASABE editorial committees; therefore, they are not to be presented as refereed publications. Citation of this work should state that it is from an ASABE meeting paper. EXAMPLE: Author's Last Name, Initials. 2015. Title of Presentation. In 2015 ASABE International Meeting. St. Joseph, Mich.:

ASABE. For information about securing permission to reprint or reproduce a meeting presentation, please contact ASABE at [email protected] or 269-932-7004 (2950 Niles Road, St. Joseph, MI 49085-9659 USA).

Introduction Several studies in academia, government and industry sectors focus on the feasibility of increasing renewable energy production in response to the Energy Independence and Security Act (EISA) of 2007. Current biofuel production in the United States is primarily corn grain conversion to ethanol, which is considered the firstgeneration biofuel. The Renewable Fuels Standards (RFS2), proposed by EPA in response to EISA, states that starting in 2016, all of the increase production in renewable fuels must be met with advanced (or secondgeneration) biofuels, such as cellulosic ethanol and other biofuels from feedstock other than starch or sugars. The targeted production is for 57 billion liters (15 billion gallons) per year of conventional renewable transportation fuel (such as corn grain based fuel) and 79 billion liters (21 billion gallons) per year of advanced biofuels by 2022. The production of renewable transportation fuels would displace conventional petroleum use and offer a cleanburning alternative. The RFS2 proposes that the American agriculture not only produce food, feed and fiber, but also biofuel feedstock for renewable energy. The decisions on how to best support all the agricultural roles and allocate agricultural lands are vital to the US and hence, extended research has been performed on this topic. Given that the national Biomass Program strategy has already determined the starch-based ethanol (corn grain based) as a well-established commodity fuel, research has recently switched focus to introducing alternative fuels, such as advanced and cellulosic biofuels into the marketplace. The feedstock used for advanced and cellulosic biofuels includes crop residues, annual energy and perennial energy crops. The collection of crop residues for bioenergy must be limited to protect the soil by controlling erosion from water and wind, retaining moisture, increasing or maintaining organic matter and nutrients. Annual energy crops will most likely be established in cropland as an alternative crop in rotations. Switchgrass (panicum virgatum), the most promising biomass crop in the US, can be established on marginal land such as pastureland because of its inherent resistance to drought and heat (Gunderson et al., 2008). Miscanthus (Miscanthus giganteus) has also been identified as one of the best choices for low-input and high dry matter yield per hectare in the USA and Europe (E. A. Heaton, Dohleman, & Long, 2008). Miscanthus x giganteus best adapts to the region below the US plant hardiness zone (PHZ) 5 (USDA Agricultural Research Service, 2012). The Department of Energy (DOE) has published a strategic analysis that estimates whether the US agriculture and forest resources have the capability to produce at least one billion dry tons of biomass annually, in a sustainable manner. The Billion Ton Study (BT2) assessment provides a means to predict the development of the biofuels industry. The estimates of potential biomass within the contiguous US were obtained through the Policy Analysis System (POLYSYS) agricultural modeling framework. These estimates were based on numerous assumptions about current and future inventory, production capacity, availability, and technology. The objective of this study is to assess the underlying assumptions for the range of conditions across a nation and determine their influence on total biomass predicted. Our goal is to examine the influence of those POLYSIS assumptions on the predicted biomass production (particularly for perennial grasses) and to examine the change in production estimates under alternative assumptions. A reliable resource assessment is paramount to calculate realistic estimates of future feedstock supplies.

Literature Review In 2005, the Department of Energy (DOE) and U.S. Department of Agriculture (USDA), published the Billion-Ton Study (or 2005 BTS), to determine if the contiguous U.S. agriculture and forest resources could sustainably produce at least one billion dry tons of biomass annually (Perlack et al., 2005). In the 2005 BTS, the agriculture and forestry resource potential was not restricted by price. In other words, all identified biomass was potentially available, even though some potential feedstock would more than likely be too expensive to be economically viable. This and several shortcomings of the 2005 BTS were addressed in the 2011 Billion-Ton Update (BT2) (Downing et al., 2011). Specifically, the update incorporated spatial county-by-county biomass inventory of primary feedstock, analyzed available biomass based on several farmgate prices (between $40 and $80 per dry ton with $5 increments) for the individual feedstock, and implicated a more rigorous treatment and modeling of resource sustainability. In the BT2, farmgate price is defined as a basic feedstock price that includes cultivation (or acquisition), harvest, and delivery of biomass to the field edge or roadside. It excludes on-road transport, storage, and delivery to an end user. For grasses and residues, this price includes baling. The BT2 identified sufficient biomass resources to meet the volumetric requirements of the EISA (Downing et al., 2011). The assessment was based on several assumptions about macroeconomic conditions, policy, weather, and 2015 ASABE Conference Meeting Paper

Page 1 of 12

international developments, with no domestic or external shocks to global agricultural markets (Westcott, 2010) (discussed below). The predicted values in the BT2 were obtained through the POLYSYS agricultural modeling framework. POLYSYS simulates changes in policy, the economy and/or resource conditions and estimates the resulting impacts for the US agricultural sector (De La Torre Ugarte, Daniel G. & Ray, 2000). The model estimates potential crop residue supplies by accounting for how much residue is produced (a function of crop yield, moisture, and residue to grain ratio), residue production costs (a fixed per ton grower payment plus collection costs per ton of residue removed), and how much residue that must remain to keep erosion within tolerable soil loss levels and maintain soil carbon levels. Reliable and realistic estimates of future feedstock supplies are key for business planning and policy development for the expansion of a sustainable biofuels industry. The BT2 estimates are an improvement of the 2005 BTS, however, the assessment still needs to be verified. The BT2 estimations of biomass feedstock are published in the Bioenergy Knowledge Discovery Framework (KDF) (ORNL, 2010). Given that the BT2 assessment scope is of a national level, the land use change assumptions may not represent the realities of all states when evaluating on a county by county basis. Evidently, when the published BT2 predictions for biomass are mapped on a county by county basis (Figure 1(a)), the predictions of perennial grasses for the states that bisect the 100th Meridian were questioned based on the knowledge of local conditions in Texas. High quantities of perennial grasses were forecasted in counties with low average annual precipitation despite the BT2 assumption that energy crops will not be established on land requiring supplemental irrigation. The BT2 evaluated two scenarios: baseline and high yield. The baseline scenario assumes a continuation of the USDA 10-year forecast for the major food and forage crops and an extension to year 2030. The baseline scenario considers an annual increase of a little over 1% for corn yield and energy crops for the 20-year simulation period. A tendency toward no-till and reduced cultivation is also assumed in this scenario. The assumed increase in energy crop yields reflects learning or experience in planting energy crops. In contrast, the high-yield scenario assumes higher corn yields (1.95% annual increase) and a much larger fraction of crop acres in reduced and no-till cultivation. The energy crop productivity increases are modeled at three levels—2%, 3%, and 4% yield increase annually. These gains are attributed to experience in planting energy crops and to more aggressive implementation of breeding and selection programs. In this study, we focused specifically on the baseline scenario and the simulated year 2022 to evaluate the contribution of these states to the RFS2 targeted goals of 21 BGY of advanced biofuels for year 2022. Crop acres per county in the BT2 are estimated from a four-year average survey data from the National Agricultural Statistics Service (NASS). Pasture acres are based on the 2007 census (USDA NASS, 2009). The scenarios presented in the BT2 also assume that perennial grasses can be potentially grown on cropland, cropland pasture, and permanent pasture, if the market price is sufficient to cause land use change. Switchgrass is used as the model perennial grass in POLYSYS. Approximately 80 years of experience with switchgrass as hay and forage crop suggest that it will be productive and sustainable on rain-fed marginal land east of the 100th Meridian, except for the Pacific Northwest (Mitchell, Vogel, Anderson, & McAndrew, 2005; Mitchell, Vogel, & Schmer, 2013). This experience guided the BT2 in limiting the conversion of pastureland to perennial grasses to counties east of the 100th Meridian, except for the Pacific Northwest. However, some of the counties east of the 100th Meridian have a low average annual precipitation, which challenges the BT2’s assumption about rain-fed sustainable production. Similarly, a rain-fed sustainable production will be challenging with the BT2 allowance to all counties in the US to convert cropland to perennial grasses, regardless of the annual average precipitation in the area. For example, in the baseline scenario at the $60 per ton farmgate price in year 2022 a total of 54,600 tons are predicted to be produced in Borden County, TX, which averages 508 mm (20 in.) annually. Thus, it is important to address the irrigation assumption for perennial grasses as the assumption greatly impacts the resulting inventories. Successful establishments of switchgrass production can be grown west of the 100th Meridian with irrigation (Washington State University, 2009). Likewise, a compilation of published literature on 1,190 observations for switchgrass yields of lowland and upland ecotypes grown at 39 field sites across the US concluded that precipitation limits yield west of the Great Plains (Wullschleger, Davis, Borsuk, Gunderson, & Lynd, 2010). Most of the observations of lowland cultivars analyzed by Wullschleger et al., were planted in the south, whereas, the upland switchgrass ecotypes were planted across the full range of latitudes. Wullschleger et al. and Heaton et al. concluded that there is a positive response of yield to precipitation and nitrogen (E. Heaton, Voigt, & Long, 2004; Wullschleger et al., 2010). But, in contrast to Heaton’s analysis, Wullschleger concluded that biomass yields do increase with higher temperatures up to a point and then decreased. The response patterns suggested a broad optimal temperature, hence the study by Wullschleger et al. was not able to completely explain the yield response to temperature. The state of Texas is one of the states bisected by the 100th Meridian, hence, we aimed our attention to this 2015 ASABE Conference Meeting Paper

Page 2 of 12

state. Upland switchgrass ecotypes occur in upland areas that are not subject to flooding, whereas lowland ecotypes are found on floodplains and other areas that receive run-on water (Vogel, 2004). The Alamo switchgrass, a lowland ecotype, was found to be the best adapted commercially available switchgrass cultivar for biomass feedstock production in Texas (Sanderson et al., 1999). Muir et al. analyzed the yield and stand responses of Alamo switchgrass in Texas during 1992 to 1998 and suggested a southern limit of adaptation for Alamo between Stephenville, TX (32.22◦N) and Beeville, TX (28.4◦N) that is unrelated to rainfall or soil type since average rainfall is similar in both sites (Muir, Sanderson, Ocumpaugh, Jones, & Reed, 2001). In addition, at latitudes northern than 38.2◦N, the yield of lowland cultivars declines at a rate of about 12.5% per degree latitude (Wullschleger et al., 2010). To accelerate the commercialization of cellulosic ethanol in the US, ample feedstock should be accessible by biorefineries at adequate times and at competitive prices (Carolan, Joshi, & Dale, 2007). To secure the feedstock supply system for a biorefinery, it is necessary to understand the complexities of the industry properties and design a reliable distribution network.

Methodology Detailed data for perennial grasses that is not publicly available was obtained from members at the Oak Ridge National Laboratory (ORNL) to understand the different types of land where perennial grasses were predicted to grow in the BT2. Perennial establishments on cropland, cropland pasture and permanent pasture are illustrated in Figure 1(b-d). The detailed data illustrates the conversion boundary at the 100th Meridian on pastureland (both cropland pasture and permanent pasture), but not for cropland conversion to perennial grasses. Figure 2 delineates the annual average precipitation for each county in the contiguous US based on data from 1981 to 2010 (PRISM Climate Group, Oregon State University, 2004); note that the map indicates that the 635 mm (25 in.) precipitation mark and the 100th Meridian line up in Texas, Oklahoma, Nebraska and Kansas. Hence, perennial grasses estimated to be grown on cropland west of the 100th Meridian in the BT2 might require irrigation given the low annual average precipitation (lower than 635 mm). The Integrated Science Assessment Model (ISAM), which estimates spatial and temporal variations of biomass yields, was used to study the eastern US between years 2001-2012 (Song, Jain, Landuyt, Kheshgi, & Khanna, 2014). The model simulated that high/low yields for Miscanthus, Alamo and Cave-in-Rock (also a switchgrass cultivar) averaged an accumulated precipitation of 749/704, 754/610 and 669/603 mm (30/28, 30/24 and 26/24 in.) respectively. Further discussion with ORNL members led us to the conclusion that county based annual average precipitation data is a better indicator for perennial grass potential than the 100th Meridian used in the BT2.

Figure 1. BT2 Perennial grass production (DMT/year) for 2017 at $50 per dry ton farmgate price (Downing et al., 2011)

2015 ASABE Conference Meeting Paper

Page 3 of 12

Figure 2. County annual average precipitation (mm) (PRISM Climate Group, Oregon State University, 2004).

Members at ORNL provided us with access to a version of the POLYSYS model in order to generate new predictions for biomass with the 635 mm annual average precipitation mark in place of the 100th Meridian boundary. The approach aimed to insure that production of perennial grasses will occur on lands with at least 635 mm of precipitation, based on the 30-year normal precipitation study (PRISM Climate Group, Oregon State University, 2004). Figure 3 exhibits the total available permanent and cropland pasture acreage –obtained from the census data (USDA NASS, 2014)– presented in the BT2 and when replacing the 100th Meridian with the precipitation boundaries of 508, 635, 686 and 762 mm (20, 25, 27 and 30in.). The use of the 635 mm boundary in place of the 100th Meridian results in a reduction of 2.1 million hectares of total available pastureland. After the publication of the BT2, the POLYSYS model has undergone additional improvements, such as the incorporation of the value of time in farmgate prices and updates with the best information available for perennial grass yields, census data and the USDA Baseline. Table 1 illustrates the differences between the POLYSYS version used for the publication of the BT2 report and the one used for this study.

Figure 3. Total available pastureland (ha) (USDA NASS, 2014).

Table 1. Comparison of the POLYSYS framework versions. POLYSYS Framework BT2 Version Version used in this study USDA Baseline 2009 2014 USDA Census 2007 2012 Switchgrass yields 2010 2014 Firs Year Simulated 2009 2014

Since nearly all dedicated feedstocks have insufficient information from which to extrapolate yield nationwide (Miguez, Maughan, Bollero, & Long, 2012), the Sun Grant Western Region GIS Center (PRISM Climate Group) at Oregon State University developed the PRISM environmental model (PRISM-EM) (Halbleib, Daly, & Hannaway, 2012). As opposed to an empirical approach extrapolating plot/field yield data to larger regions, PRISM-EM operates based on a limiting factor approach. The model tracks precipitation, evapotranspiration and soil moisture depletion and estimates relative yields based on water balance simulations, plant injury curves for summer heat and winter cold, and growth constraints due to soil pH, drainage and salinity. The county-based perennial grass yields input to the POLYSYS model are the output of the PRISM-EM. Since the publication of the BT2, improvements have been made to the PRISM-EM. Consequently, we used an adjusted switchgrass yield input dataset to the POLYSYS model. Figure 4 outlines the different yield datasets. Figure 4 (a) exhibits the perennial grass yields used in the BT2; note that higher yield zones are displayed in eastern Tennessee and Kentucky, and some areas in Georgia, North Carolina and South Carolina. In contrast, the PRISM-EM output illustrated in Figure 4(b) and used in this study, discloses the highest perennial grass yields in Mississippi, 2015 ASABE Conference Meeting Paper

Page 4 of 12

Louisiana, Arkansas, Florida and western Tennessee and Kentucky. The yields mapped below are mainly for switchgrass cultivars, except for Florida, where energy cane is assumed to be adopted. The map in Figure 4(b) illustrates switchgrass yields between 3 and 5 tons per acre along southern states bisected by the 100th Meridian (Kansas, Nebraska, Oklahoma and Texas) and yields between 2 and 3 tons per acre for North and South Dakota. Depending on the latitude, US counties bisected by the 100th Meridian are classified as either low and stable or low and unstable yield zones for mischanthus, Cave-in-Rock upland switchgrass and Alamo lowland switchgrass (Song et al., 2014).

Figure 4. Perennial yields used in the BT2 and in this study.

In addition to the differences between the POLYSYS versions presented in Table 1, the farmgate prices in POLYSYS were updated to reflect real values for every predicted year using the 2016 USDA Long Term Projection GDP price index adjusted values. For simplicity, we present our results in nominal prices as outlined in the BT2 scenarios. To compare the several changes between versions of the POLYSYS framework, we developed a scenario named past100 with the same boundaries for conversion of land to perennial grasses as the BT2 (expounded in Table 2). Given the 5 year difference in simulation periods, we compared predictions for 2017 reported in the BT2 with the data from 2022 simulated in scenario past100 from this study. To evaluate the responsiveness of using the 635 mm annual average precipitation mark in place of the 100th Meridian boundary, we developed two additional scenarios: past25 and all25. The past25 scenario only allowed conversion of cropland pasture and permanent pasture on counties with a 635 mm or higher average annual precipitation. Similar to the BT2 assumption, this scenario has no geographical restriction for establishing perennial grasses on cropland. Conversely, the all25 scenario will only allow for conversion of lands (cropland, cropland pasture and permanent pasture) to perennial grasses in counties with a 635 mm or higher average annual precipitation, regardless of land type. It was anticipated that the addition of a geographical restriction for conversion of cropland would result in a significant decrease of predicted perennial grass nationwide. Table 2 illustrates the differences between the three scenarios created for this study. Figure 5 gives a graphical representation of the differences between scenarios. The information about which specific counties in the Pacific Northwest were allowed for conversion to perennial grasses in the BT2 was not publicly available. Hence, we assumed an inclusion of counties with an average annual precipitation of 635 mm or higher. Scenario past100 past25 all25

Table 2. Set of counties allowed for conversion to perennial grasses. Cropland Cropland pasture and permanent pasture Counties east of 100th Meridian and counties with a 635+ No boundary mm average annual precipitation west of 100th Meridian Counties with a 635+ mm average annual precipitation, No boundary regardless of their location with respect to 100th Meridian Counties with a 635+ mm average annual precipitation, Counties with a 635+ mm average annual precipitation, regardless of their location with respect to 100th Meridian regardless of their location with respect to 100th Meridian


Page 5 of 12

Figure 5. Graphic representation of the set of counties allowed for conversion to perennial grasses.

We also analyzed the reported inventories of perennial grasses in the counties south of the 28◦N latitude given that a study in Texas suggested that in counties south of the 28◦N, there is a limit of adaptation for switchgrass that is unrelated to rainfall or soil type (Muir et al., 2001). In other words, we expected little or no production of perennial grasses in this area. The only states bisected by the 28◦N latitude are Texas and Florida, but, in Florida, energy cane is the perennial assumed to be adopted and not switchgrass. While the baseline scenario in the BT2 does not reflect any production of perennial grasses on counties south of the 28◦N, this is not the case for the high yield scenarios in the BT2. Though the scope of this study did not include high yield scenarios, it is important to note that a 28◦N latitude as a limit for conversion of cropland and pastureland to perennial grasses should be considered for high yield scenarios in Texas.

Results and Discussion Regression analysis techniques were used to compare the output presented in the BT2 and the output from the updated version of POLYSYS in the past100 scenario. Figure 6 describes a strong correlation between the two outputs as the lowest R-square value was 0.9002.

Figure 6. Correlation between major crops' production in the BT2 and the past100 scenario.

Figure 7 illustrates the total predicted biomass production for major crop residues (corn stover and wheat straw), annual energy crops and perennial grasses as farmgate price increases from $40 to $80. The differences in predicted values are mainly attributed to the changes presented in Table 1. As year-1 expected, with the rise in farmgate price, higher production of biomass is estimated in both scenarios. As per Figure 7, the past100 2015 ASABE Conference Meeting Paper

Page 6 of 12

scenario outputs closer values to the BT2 predictions at the $80 farmgate price than at lower farmgate prices (decrease of 13%). The predicted production for every major crop presented in Figure 7 for the $40 price mark value are lower in past100 as compared to the BT2 data. Hence, the past100 scenario simulates less reliance on energy crops than estimated in the BT2. Conversely, with exception of the $40 price mark, there is a general increase of reliance on crop residues in past100 when compared to BT2. At prices higher than $50, higher amounts for wheat straw are predicted in past100 than in the BT2, yet, the change is minimal compared to a corn stover change, on average, of 33%. Energy crop predictions (annual and perennial) are lower in past100 than in the BT2, hence, as the PRISM-EM is refined, there is less reliance on energy crops at any of the farmgate prices simulated. A drastic change in perennial production is observed at the lowest price level, but the change from the BT2 to the past100 minimizes as price increases. Table 3. Differences between major biomass crops’ production presented in the BT2 for year 2017 and production simulated for year 2022 in the past100 scenario. Major Biomass Crops BT2 for year 2017 (DMT) Percent change from BT2 to past100 Corn Wheat Annual Perennial Corn Wheat Annual Perennial Stover Straw Energy Crop Grass Stover Straw Energy Crop Grass 32,098,300 7,829,900 733,400 3,012,500 -92% -75% -46% -100% 92,552,800 22,216,600 3,762,900 40,814,800 31% -13% -66% -78% 105,734,200 26,078,100 5,047,300 90,135,200 37% 3% -57% -39% 112,893,800 30,322,100 6,091,700 114,377,000 34% 4% -43% -22% 115,550,900 32,646,600 8,347,700 130,724,500 32% 2% -26% -13%

For convenience and ease in reading, the BT2 presents all feedstock quantities and their composite total at the $60 per dry ton level. This price was selected because it brought in most of the available tons (based on the BT2 results) from all of the feedstocks and because the price represented a realistic, reasonable price for discussion purposes at the time of publication. The price is comparable to the DOE Multi-Year cost targets for cellulosic feedstocks when adjusted to exclude transportation and handling costs (U.S. Department of Energy, 2011). A more recent version of the Multi-Year sets a cost target of $80/DMT ($2011) that includes grower payment/stumpage fee and logistics cost to the throat of the conversion reactor in the form of bales, loose chop, etc. (U.S. Department of Energy, 2015). In this study, we will discuss predicted biomass tonnages for both farmgate prices, $60 and $80. Figure 7 is an illustration of the perennial grass predicted in both scenarios at the $60 and $80 price levels. Despite the national decrease in perennial cultivar production between scenarios BT2 and past100 (from 90,135,200 DMT to 54,829,525 DMT at $60 and 130,724,500 to 113,085,174 DMT at $80 for each scenario respectively), the model still simulates high production of perennial cultivars east of the 100th Meridian. The color scale used in Figure 7 is used across all the maps presented in this paper.

Figure 7. Perennial grass production at the farmgate price of $60 and $80 for the BT2 scenario and the past100 scenario.

Similarly, we evaluated the changes in total perennial grasses estimated to be grown in each scenario: past100, past25 and all25. When compared to past100, a higher change was observed in total perennial grasses 2015 ASABE Conference Meeting Paper

Page 7 of 12

estimated to be grown on cropland for past25 and all25 than perennial grasses estimated to be grown on pastureland (Table 4). Table 4. Changes observed in total US perennial grass predictions on pastureland and cropland in past100, past25 and all25. Perennial on cropland Perennial on pastureland Farmgate past100 past100 Price ($) past25 all25 past25 all25 (DMT) (DMT) 40 0 0% 0% 6,690 0% 0% 50 2, 886,325 0% 0% 5,918.057 0% 0% 60 14,677,306 0% -24% 40,152,219 -2% -2% 70 33,185,244 0% -22% 56,255,441 -2% -2% 80 47,116,682 0% -21% 65,968,492 -5% -5%

In scenario past100 at the $60 farmgate price for year 2022, the total national perennial grass predicted to grow on cropland and pastureland is 14,677,306 DMT/year and 40,152,219 DMT/year respectively. When compared to past100, there is a 24% decrease in total perennial predicted to grow on cropland in scenario all25 and no change in past25. In both scenarios, past25 and all25, there is an observed decrease of 2% in total perennial predicted to grow on pastureland. Higher amounts of perennial grass are estimated to grow on pastureland at the higher farmgate price of $80, but lower amounts of perennial on cropland. A 21% decrease in total perennial predicted to grow on cropland in scenario all25 is observed and no change in past25. Again, both scenarios, past25 and all25, simulated a decrease of 5% in total perennial predicted to grow on pastureland. Figure 8 illustrates the geographical changes between the three scenarios of this study at the $60 farmgate price for year 2022.

Figure 8. US perennial grass prediction changes between scenarios at the $60 farmgate price for year 2022.

Note that these changes in predictions are exhibited in the states that are bisected by the100th Meridian. We focused in the state of Texas to evaluate these changes. In scenario past100 at the $60 farmgate price for year 2022, the total national perennial grass predicted to grow on cropland and pastureland is 2,631,875 DMT/year and 387,194 DMT/year respectively. The only change observed at this price level when comparing scenarios was a decrease of 99.62% (a total of 9,977 DMT/year) of perennial estimated to be grown on cropland in the all25. No other change was observed at this price in Texas. At the farmgate price of $80 a total perennial grass of 8,206,298 DMT/year was predicted to be grown on cropland in Texas and a total of 9,528,800 DMT/year grown on pastureland in scenario past100. Only a 1% decrease was observed in past25 and all25 when compared with scenario past100. All25 did reflect a decrease of 77% from the predicted total amount of perennial grasses in past100.


Page 8 of 12

Figure 9. TX perennial grass prediction changes between scenarios at the $60 farmgate price for year 2022.

An observation can be made that 635 mm of average annual precipitation might not be enough for a sustainable production of perennial grasses. Figure 10 illustrates the set of counties that will be excluded if placing a 686 mm or 762 mm of average annual precipitation boundary in place of the 635 mm boundary. The scope of this study was limited to simulations of the POLYSYS model with the 635 mm precipitation boundary, but Figure 10 illustrates the impact of using high precipitation thresholds for economically viable yield levels.

Figure 10. Representation of impact to changing the 635 mm precipitation boundary to one of 686 mm or 762 mm.

Finally, the total herbaceous biomass predicted in the US (corn stover, barley straw, sorghum stubble, oat straw, wheat straw, annual energy crop and perennial grasses) for each of the scenarios we discussed in this paper are presented in Table 5. We see a slight increase in total herbaceous biomass from the predictions in the BT2 and our past100 scenario, 0.39% and 6.18% at the farmgate price of $60 and $80 respectively. While the total predicted herbaceous biomass predicted in this study is close to the biomass predicted in the BT2, the geographical location of these biomass changes across the nation.

Table 5. Total available herbaceous biomass estimated in each scenario (DMT/year). Farmgate BT2 past100 past25 all25 price ($) 2015 ASABE Conference Meeting Paper

Page 9 of 12

40 50 60 70 80

44,921,000 161,825,600 229,698,200 266,240,000 289,593,900

5,985,619 152,512,658 230,595,716 277,354,628 307,481,164

5,985,619 152,512,658 229,833,740 276,198,332 304,438,673

5,985,619 152,512,658 226,408,163 269,037,366 295,157,426

Summary and Conclusions We concluded that the potential perennial grass production in the US, and particularly in Texas, was overestimated by allowing cropland in regions with less than 635 mm (25 in.) average annual precipitation to be converted into perennial grasses. Placing a precipitation boundary for land conversion to perennial grasses reduced the total perennial grass production (predicted to be established on both cropland and pastureland) in the US and Texas by 7.7 % and 86.8%, respectively under the farmgate price of $60 per dry ton scenario. At the farmgate price of $80 per dry ton, this reduction was higher nationwide, 11.3%, and lower at the Texas level, 36.6%. Given that the BT2 assessment scope is of a national level, the land use change assumptions may not represent the realities of all states. Changing the 100th boundary to the precipitation boundary was discussed to be a better approach to estimate the inputs to the POLYSYS model. The boundaries presented in scenario past25 in this study will be applied in the upcoming Billion Ton Study of 2016. The revised inventories for crop residues, annual energy crops and perennial crops will be used at Texas A&M to determine the structure of the likely biomass feedstock supply chain that may develop in Texas and estimate the portion of the total biomass produced that may be economically stranded.

Acknowledgements We will like to acknowledge the USDA Agriculture and Food Research Initiative, the Oak Ridge National Laboratory for their support and the Texas A&M AgriLife Research letting us use their resources.

References Carolan, J. E., Joshi, S. V., & Dale, B. E. (2007). Technical and financial feasibility analysis of distributed bioprocessing using regional biomass pre-processing centers. Journal of Agricultural & Food Industrial Organization, 5(2), 10. De La Torre Ugarte, Daniel G., & Ray, D. E. (2000). Biomass and bioenergy applications of the POLYSYS modeling framework. Biomass and Bioenergy, 18(4), 291-308. Downing, M. E., Eaton, L. M., Graham, R. L., Langholtz, M. H., Perlack, R. D., Turhollow Jr, A. F., Stokes, B., Brandt, C. C. (2011). US billion-ton update: Biomass supply for a bioenergy and bioproducts industry. ORNL/TM-2011/224. Oak Ridge National Laboratory: Oak Ridge, TN. 227p. Gunderson, C. A., Davis, E. B., Jager, H. I., West, T. O., Perlack, R. D., Brandt, C. C., Wullschleger S. D., Baskaran, L. M., Wilkerson, E. G., Downing, M. E. (2008). Exploring potential US switchgrass production for lignocellulosic ethanol. ORNL/TM2007/183, Oak Ridge National Laboratory: Oak Ridge, TN. 51p. Halbleib, M. D., Daly, C., & Hannaway, D. B. (2012). Nationwide crop suitability modeling of biomass feedstocks. Heaton, E. A., Dohleman, F. G., & Long, S. P. (2008). Meeting US biofuel goals with less land: The potential of miscanthus. Global Change Biology, 14(9), 2000-2014. Heaton, E., Voigt, T., & Long, S. P. (2004). A quantitative review comparing the yields of two candidate C4 perennial biomass crops in relation to nitrogen, temperature and water. Biomass and Bioenergy, 27(1), 21-30. doi: http://dx.doi.org/10.1016/j.biombioe.2003.10.005 Miguez, F. E., Maughan, M., Bollero, G. A., & Long, S. P. (2012). Modeling spatial and dynamic variation in growth, yield, and yield stability of the bioenergy crops miscanthus× giganteus and panicum virgatum across the conterminous United States. Gcb Bioenergy, 4(5), 509-520. Mitchell, R., Vogel, K., Anderson, B., & McAndrew, T. (2005). Renovating pastures with glyphosate tolerant soybeans. Forage and Grazinglands, 3(1) doi:10.1094/FG-2005- 0428-01-BR. Mitchell, R., Vogel, K. & Schmer, M. (2013). Switchgrass (pacinum virgatum) for biofuel production. Retrieved from http://www.extension.org/pages/26635/switchgrass-panicum-virgatum-for-biofuel-production#.VEUPEpQ7uG4 Muir, J. P., Sanderson, M. A., Ocumpaugh, W. R., Jones, R. M., & Reed, R. L. (2001). Biomass production of ‘Alamo’switchgrass in response to nitrogen, phosphorus, and row spacing. Agronomy Journal, 93(4), 896-901. ORNL. (2010). Bioenergy knowledge discovery framework. Retrieved from https://bioenergykdf.net/ Perlack, R. D., Wright, L. L., Turhollow, A. F., Graham, R. L., Stokes, B. J., & Erbach, D. C. (2005). Biomass as feedstock for a bioenergy and bioproducts industry: The technical feasibility of a billion-ton annual supply. PRISM Climate Group, Oregon State University. (2004). PRISM gridded climate data. Retrieved from http://prism.oregonstate.edu Sanderson, M. A., Reed, R. L., Ocumpaugh, W. R., Hussey, M. A., Van Esbroeck, G., Read, J. C., Tischler, C.R., Hons, F. M. 2015 ASABE Conference Meeting Paper

Page 10 of 12

(1999). Switchgrass cultivars and germplasm for biomass feedstock production in texas. Bioresource Technology, 67(3), 209219. doi:http://dx.doi.org/10.1016/S0960-8524(98)00132-1 Song, Y., Jain, A. K., Landuyt, W., Kheshgi, H. S., & Khanna, M. (2014). Estimates of biomass yield for perennial bioenergy grasses in the USA. BioEnergy Research, 8(2), 688-715. US DOE. (2011). Biomass multi-year program plan. Washington, DC: U.S. Department of Energy. US DOE. (2014). Biomass multi-year program plan. Washington, DC: U.S. Department of Energy. USDA Agricultural Research Service. (2012). Plant hardiness zone map. Retrieved from http://planthardiness.ars.usda.gov/PHZMWeb/# USDA NASS. (2009). 2007 census of agriculture. Washington, DC: National Agricultural Statistics Service, U.S. Department of Agriculture. USDA NASS. (2014). 2012 census of agricultureWashington, DC: National Agricultural Statistics Service, U.S. Department of Agriculture. Vogel, K. P. (Ed.). (2004). Switchgrass. Madison, WI: American Society of Agronomy-Crop Science Society of America-Soil. Washington State University. Biofuels Cropping System Project 2009 Final Report: Establishing switchgrass for biofuels in the North Columbia Basin. W.L. Pan and K.E. Sowers, ed. Available from http://css.wsu.edu/biofuels/files/2012/09/2009_Final_Report.pdf. Westcott, P. (2010). USDA agricultural projections to 2019 DIANE Publishing. Wullschleger, S. D., Davis, E. B., Borsuk, M. E., Gunderson, C. A., & Lynd, L. R. (2010). Biomass production in switchgrass across the United States: Database description and determinants of yield. Agronomy Journal, 102(4), 1158-1168. doi:10.2134/agronj2010.0087


Page 11 of 12