Potential Suppy and Cost of Biomass from Energy Crops in the TVA Region
Potential Suppy and Cost of Biomass from Energy Crops in the TVA RegionMethods |
Because energy crops[1] are not currently grown in the TVA region, organized markets for SRWC and switchgrass feedstocks do not exist. Consequently, quantity and price information for biomass energy crop feedstock must be modeled. The model of energy crop price and supply used in this analysis is based on the assumption that farmers will convert their agricultural land to biomass production when the profit received from producing biomass meets or exceeds current profit margins received from producing conventional agricultural crops or using their land for pasture [Equation (1)].
( YLDc * PRICEc ) - COSTc=( YLDw * BEP ) -
COSTw ,
where
c=conventional crop displaced by the energy crop,
w=type of energy crop (switchgrass or SRWC species),
YLDc=yield expected for conventional crop c (bushels or bales per acre),
PRICEc=expected market price of the conventional crop c ($ per
bushel or bale),
COSTc=cost of producing conventional crop c ($ per acre),
YLDw=annual harvested yield of energy crop w (tons per acre per year),
BEP=break-even price of harvested energy crop biomass ($ per ton),
COSTw=annual cost of producing energy crop w ($ per acre).[2]
The left side of the equation may be considered land rent or returns to land and management from growing conventional crops on cropland or maintaining pastureland in pasture production. Agricultural commodity payments or other government subsidies are not included in this land rent because existing programs are likely to change in the near future.
The right side of the equation, along the same line, is the land rent received from growing biomass crops on the same land. If all other variables can be quantified, the break-even price can be solved.
For cropland, the farm-gate price of biomass--the price a utility would have to pay a farmer for harvested biomass (not transported)--was assumed to be the same as the biomass break-even price associated with the most profitable dominant conventional crop that could be grown on the land. Using the break-even price associated with the most profitable crop rather than the actually grown crop provided some assurance that the farm-gate prices were not underestimated. Dominant conventional crops (such as corn, soybeans, wheat, and cotton) were used to define the farm-gate price under the assumption that these crops, rather than minor crops such as barley, rye, and sorghum, would define the regional economics of energy crop production.
The analysis is
also based on the assumption that differences in soil quality, climate, and
land use across the region will create geographic variation in the cost and
potential supplies of energy crop feedstocks. Yields of both conventional crops
and energy crops are sensitive to land quality and climate. Consequently,
farm-gate prices of biomass should show geographic variability reflective of
that sensitivity. Energy crop supplies will be a function not only of yields
but also of the availability of agricultural land. The geographic distribution
of agricultural land should also be very important in defining regional supply
patterns. Thus, the model structure is designed to capture the effect of these
three spatial factors on energy crop price and supply.
Two elements not incorporated in the model structure were transportation
costs from field to conversion plant and agricultural risk to farmers.
Transportation costs are conversion plant specific, and other analyses for this
region have shown over-the-road transportation costs for wood to range from $10
to $20 per (dry) ton, depending on transport distance (
The impact of agricultural risk on farmer's decision-making processes is difficult to quantify. Its effects would depend on the supply/demand infrastructure which is unknown at this point. If risk is lower than it is for conventional crops, the farm-gate price could be lower than projected using the concept of a break-even price. If risk is perceived as the same or higher, the farm-gate price might be higher.
The approach
used to project potential biomass supplies in the TVA region is diagrammed in
and described in the following five sections of text. The approach is divided
into five interconnected stages. In Stages 1-3, geographic-specific values for
the components of Eq. (1) are derived and the TVA region land base is
characterized. In Stage 4, the farm-gate prices for switchgrass and SRWC wood
are calculated. In Stage 5, the projected farm-gate prices and energy crop
yields are linked with data on land use and land quality in the region. The
potential supply of grass or wood feedstock from energy crop production is
projected at the county level, and regionwide supply curves of energy crop
feedstock are derived.
In this stage,
land use, soils, current crops, and crop yields in the 276-county region are
characterized. To capture intraregional physiographic and climatic differences
that influence energy crop and conventional crop yields, the region is broken
into eight subregions. Boundaries of subregions are based on current land use
and physiographic features and largely followed the U.S. Department of
Agriculture's (USDA's) Major Land Resource Areas (MLRA) (
Soil Conservation Service (SCS) agricultural capability classes are used to characterize soil quality. The classes signify limitations in crop choice and are based on soil and site characteristics. There are eight general classes ranging from 1 (few limitations restricting cropland use) to 8 (precluded from cropland use). Classes 5-8 are generally unsuitable for growing crops. Within classes 2-8, four subclasses define the primary cause of cropland limitations. These subclasses are w (excessive water), e (erosion potential), s (soil restrictions--salinity, shallowness, or texture problems), and c (climate restrictions). For the analysis, the 29 SCS classes are aggregated into nine soil categories by grouping classes 5-8 into one category and grouping classes 3w and 4w together and classes 3s and 4s together. This grouping results in nine categories--1, 2e, 2s, 2w, 3e, 3-4s, 3-4w, 4e, and 5+. No climate-restricted soils are present in the study area, and 2s and 3-4s soils are rare or absent.
The 1982 National Resources Inventory (NRI) data base (
Land-use data from the 1987 Agricultural Census (
For conventional crops--soybeans, cotton, corn, and wheat, The University of Tennessee (UT) Agricultural Extension Service publication Guide to Farm Planning manual (Johnson 1990) is used as the basis for defining crop management activities and crop production costs. These include application and timing of fertilizer, harvest date, labor costs, and equipment usage.
In lieu of determining the value of production from pastureland and the
costs of maintaining pasture [left side of Eq. (1)], the USDA Economic Research
Service's dollar-per-acre value of pastureland rent is used (
Several tree species are suitable for SRWC production within the TVA region: sweetgum (Liquidambar styraciflua), poplar (Populus spp.), sycamore (Platanus occidentalis), and black locust (Robinia pseudoacacia). In this analysis, each species is assigned to the soil categories for which it is projected to have the best growth relative to the other species. Poplar is assigned to the 1 and 2w soil categories, sweetgum to the 2e and 2s soil categories, black locust to the 3e and 3-4s categories, and sycamore to the 3-4w category. No SRWC species is considered suitable for growing on soil categories 4e and 5+.
Previous information collected by field researchers from experimental field production of SRWC was used to outline management scenarios and construct production costs for each of the species. Rotations varied in length from 6 to 10 years, and coppice rotations were assumed for poplar, sweetgum, and sycamore. Although species are managed differently, all species are assumed to have uniform harvest loss of 15% of the standing yield. The management schemes for pasture conversion are the same as for plantation establishment on cropland but include more site preparation. The SRWC management schemes do not vary between the two SRWC production scenarios (current yields and 25% increase in current yields).
Two management schemes were developed for switchgrass--one for establishment on pasture and one for establishment on cropland. Both are based on work by Bransby and Parrish (Bransby et al. 1990; Parrish et al. 1990). In both cases, harvest/storage losses of 14% are assumed and switchgrass is reestablished every 10 years. Unlike SRWC, switchgrass production is considered possible on all nine soil categories, including 4e and 5+. Thus, the potential land base for switchgrass production is slightly larger. As with SRWC, management schemes for switchgrass do not vary for the two switchgrass scenarios (current and 25% increased yields).
Production costs for both SRWC and switchgrass include materials (seedlings or seeds, fertilizer, and pesticides), equipment, and hourly labor costs. SRWC production costs also include contract aerial spraying of pesticides and contract harvesting and chipping. The discount rate is set at 6%. Machine labor rates are assumed to be fixed at $7.80/h. Harvesting and chipping costs account for nearly half of the total cost of SRWC production and are sensitive to the amount of wood harvested per acre. SRWC harvest and chipping costs are assumed to be $17 per (dry) ton if 30 (dry) tons are harvested per acre and to increase if fewer tons are harvested (Anthony Turhollow, Oak Ridge National Laboratory, Oak Ridge, personal communication). Harvesting costs account for about 40% of the production costs for switchgrass and are sensitive to the amount harvested per acre.
The Erosion Productivity Impact Calculator (EPIC) model developed by the
USDA Agricultural Research Service (ARS) is used to predict the yields
associated with the four conventional crops and switchgrass (
The EPIC model was initially developed by ARS to assist in the 1985 status report on the nation's soil and water resources as required under the Soil and Water Resource Conservation Act. Because the purpose of the model was to address the wide variation in crop productivity and soil resources that exist in the United States, the model was designed to be as mechanistic as possible. Inputs to the model include soil characteristics, daily weather data (precipitation, maximum temperature, minimum temperature, solar radiation, and wind speed), crop parameters (e.g., maximum leaf area index, maximum root depth, and optimal nitrogen concentration in plant tissue), and management practices (date of planting, date of harvest, date and amount of fertilizer application, tillage practice, etc.). The model considers only one crop at a time, although crop rotations over time are allowed. It also assumes uniform field conditions. The model can accept up to ten different soil horizons.
The EPIC model has been widely used and adapted for many crops (
Crop parameters developed for EPIC by ARS were used to model the four dominant conventional crops. The parameters used for switchgrass were adapted from EPIC hay parameters by Burton English (UT Department of Agricultural Economics, personal communication), working with ARS researchers.
The EPIC model, used to simulate 30 years of continuous crop production, uses weather data uniquely characteristic of each subregion, soils specific to each subregion and each of the nine soil categories, and management practices specific to each crop. A total of 288 EPIC runs were simulated. Average values (over 30 years) were calculated for annual crop yield (dry tons, bushels, or bales per acre per year). For each of the conventional crops in each subregion, the yields predicted by EPIC were adjusted by the actual average yield in the subregion (as characterized in Stage 1) by applying the following formula:
(2)
where
Yi=adjusted crop yield for soil category i,
Ei=EPIC predicted yield for soil category i,
Pi=proportion of crop in the subregion planted on lands of soil
category i,
Yavg=average yield of crop within the subregion.
In this manner, the sensitivity of conventional-crop yields to soil quality was predicted.
The soil- and subregion-specific switchgrass yields projected by EPIC were used without modification as estimates of switchgrass yield on former cropland. Estimates of switchgrass yield on former pastureland were developed by decreasing the projected cropland switchgrass yields by 10%.
Because EPIC does not yet include a simulation module for any of the SRWC species and little SRWC field data are available specific to the 276-county region, expert opinion was used to derive expected yields for SRWC plantations.
2.5.1 Pastureland
For each soil category in each subregion, pasture rent values and energy crop production costs determined in Stage 2 and the energy crop yields determined in Stage 3 were used to calculate the break-even price for energy-crop biomass produced on pastureland. This break-even price was then used as the farm-gate price for biomass grown on pastureland of that soil category.
2.5.2 Cropland
Crop yields and production costs generated in Stages 2 and 3 were used to develop break-even prices for biomass for each soil category in each subregion. Equation (1) and the market price of conventional crops were used to calculate the prices. The market price was calculated as the average of the years 1989, 1990, and 1991 and does not vary across the region or with soil category (corn=$2.55/bu, wheat=$3.38/bu, soybeans=$6.53/bu, and cotton=$274.50 per bale). The break-even price associated with the most profitable of the three crops (i.e., the highest break-even price) was used as the farm-gate price for harvested energy crop biomass. If conventional crop yields were such that Eq. (1) predicted a negative land rent, a land rent value of zero was used to calculate the farm-gate price. This ensures that the farm-gate price is sufficient to cover the production costs of the energy biomass.
The farm-gate prices for harvested energy-crop biomass calculated in Stage 4 were merged with information from Stage 1 on the acreage of land in each county (by land use and soil category) and information from Stage 3 on energy crop yields to determine the county-specific potential supplies of biomass at the farm-gate price specific to each soil/land use category. This price and supply information was then linked with locations, species, and acreage information to produce county-level maps of potential energy-crop biomass supplies at various farm-gate prices and for various regional biomass supply curves.
File created: October 21, 1996; Last updated: |