| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Department of Agricultural Economics, Kansas State University, Manhattan, KS 66506-4011
b Department of Agronomy, Kansas State University, Manhattan, KS 66506-5501
c Kansas Industrial Extension Service, Kansas State University, Manhattan, KS 66506-2508
* Corresponding author (jwilliam{at}agecon.ksu.edu)
Received for publication April 26, 2005.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND ANALYSISCONCLUSIONSREFERENCES |
|---|
Abbreviations: CT, conventional tillage • CT84M, conventional tillage, with 84 kg N/ha from beef cattle manure • CT84N, conventional tillage, with 84 kg N/ha from commercial N fertilizer • CT168M, conventional tillage, with 168 kg N/ha from beef cattle manure • CT168N, conventional tillage, with 168 kg N/ha from commercial N fertilizer • NT, no-tillage • NT84M, no-tillage, with 84 kg N/ha from beef cattle manure • NT84N, no-tillage, with 84 kg N/ha from commercial N fertilizer • NT168M, no-tillage, with 168 kg N/ha from beef cattle manure • NT168N, no-tillage, with 168 kg N/ha from commercial N fertilizer
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND ANALYSISCONCLUSIONSREFERENCES |
|---|
Most of the previous economic analysis has estimated the costs of sequestering C to derive a carbon supply curve at the regional level. Examples include those by Antle et al. (2001) and Feng et al. (2004). They estimated the costs of sequestering C in production systems of the Northern Great Plains region and the state of Iowa, respectively. Both studies concluded that the costs depend on site-specific characteristics, the total amount of C sequestered or desired, and the various policy instruments used. Antle et al. (2001) reported the costs ranged from $12 to $500/Mg. Feng et al. (2004) found that the costs ranged from $10/Mg to as much as $800/Mg.
Although there have been some recent studies that have examined the economic feasibility of reducing tillage (Bushong and Peeper, 2004; DeVuyst and Halvorson, 2004; Harman et al., 1996; Parsch et al., 2001; Ribera et al., 2004; Williams et al., 2004b), there has been only one study that examined both the economic feasibility of reducing tillage and sequestering carbon at the crop enterprise level (Williams et al., 2004a). Williams et al. (2004a) reported that payments to induce producers to use no-tillage rather than conventional tillage in wheat and grain sorghum production to sequester higher amounts of carbon range from $8.62 to $64.65/Mg of C/yr. Additional research is needed to improve the understanding of the economic feasibility of adopting alternative tillage and fertilizer systems to enhance C sequestration in soil.
Soil carbon sequestration costs vary widely due to location, soil type, estimated C uptake, land rental rate, management techniques, and resulting crop yields. Marginal costs of C sequestration rise as forest or agricultural establishment moves from land with poor productivity and/or low opportunity costs to areas of greater productivity and/or opportunity costs (Richards, 2001). McCarl and Schneider (1999) and Caspers-Simmet (1999) reported that the marginal costs in U.S. agriculture to sequester C were in the range of $8.82 to $23.15/Mg/yr. Antle et al. (2002) found payments to induce producers from crop/fallow to continuous cropping began at $4.41/Mg/yr, and increased to $63.93/Mg/yr as more acres in continuous cropping were desired. Lewandrowski et al. (2004) estimated that managers would adopt conservation tillage at $9.92/Mg and begin to convert land to forest at $25.35/Mg. Because of the number of factors affecting soil carbon sequestration costs, examining the economic feasibility of alternative management techniques by using experimental data for specific locations is useful.
Many of the previously mentioned studies used econometric models that focus on C changes due to implementation of a single sequestration technology across a wide range of environments. Others used region-based mathematical programming models. Eve et al. (2002) suggested that basing economic analysis of carbon sequestration on experimental work is also important. Analysis of experimental work can provide useful information for constructing mathematical programming models and validating econometric model results, as well as evaluating technologies for specific production regions.
The objective of this study was to use field experiment data to determine the dollar value of a carbon credit incentive needed for managers to use no-tillage, as opposed to conventional tillage, and/or manure applications, rather than commercial N fertilizer, to enhance soil carbon sequestration in corn. The values of C credits needed to motivate adoption of practices that sequester additional C in the soil were derived, while accounting for C released to the atmosphere from production inputs.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND ANALYSISCONCLUSIONSREFERENCES |
|---|
Study Region and Production Systems
The yields and soils data were from an experimental field located in the Kansas River Valley in northeastern Kansas. The landscape is fairly level and consistent with a river valley. Average annual precipitation at the site during the study period was 810 mm. The soil is a Kennebec silt loam (fine-silty, mixed, mesic Cumulic Hapludolls).
The production systems studied included the use of either conventional tillage (CT) or no-tillage (NT) with applications of either 84 or 168 kg of N equivalent from commercial N fertilizer (N) or from beef cattle manure (M).
The eight systems studied were:
The CT system field operations consisted of disking in the spring, field cultivating before planting, row cultivating after planting, and chiseling in the fall after harvest. The NT systems did not have any tillage operations. Either commercial N fertilizer or beef cattle manure was applied in both tillage treatments shortly after the final disking in the CT systems. Most herbicides were applied at the same time and rate to both CT and NT systems. In some years, the NT system required additional herbicide applications due to greater weed pressure. The beef cattle manure was analyzed for N content one week before application to determine the amount of manure equivalent to the nitrogen from commercial N fertilizer. The average annual application rate of manure was 3.2 Mg/ha for the 84-kg N treatment and 6.4 Mg/ha for the 168-kg N treatment. All strategies received applications of phosphate (P) to prevent P from being the limiting yield factor (R. Lamond, Department of Agronomy, Kansas State University, personal communication, 2002).
Yields, Prices, Costs, and Net Returns
Yield data collected from 1991 to 1999 were used for simulating correlated empirical yield distributions. Table 1 provides a summary of yields. Northeastern Kansas average annual corn prices from the USDA, for the period of 1991–1999, were used to form an empirical price distribution. The average price of this distribution was $0.10/kg, with a standard deviation of $0.02/kg. Simulated yield and price distributions were used to calculate gross returns and net returns to land and management by subtracting costs (2002 $). Costs for each field operation were obtained from Beaton et al. (2005). Custom rates from custom applicators for beef cattle manure loading and application were used (Bar Six Construction, Protection, KS; Jones Construction, Leoti, KS). The average custom charge was $2.39/Mg. Input costs were based on actual experiment application rates. Prices for seed, commercial N fertilizer, and herbicides were from input dealers and Kansas State University (2002).
|
 | [1] |
 | [2] |
The empirical distribution and shape were specified by the data used because too few observations exist to estimate parameters for another distribution (e.g., normal distribution). A multivariate distribution has been shown to appropriately correlate random yields based on their historical correlation (Richardson et al., 2000). The multivariate distribution is a closed form distribution, which eliminates the possibility of values exceeding values observed in history (Ribera et al., 2004). Yield distributions were correlated in the simulation. The correlations ranged from 0.73 to 0.96, with the average being 0.88. Statistically significant correlation (95% level) was found among all the historical yield series. The average correlation between price and the yield series was 0.06. Because prices were not statistically correlated with any of the historical yield series, correlation between price and yield was not included in the simulation.
Soil Carbon Data
Carbon data for the top 30 cm of the soil in the experiment was from soil sampling of organic C content taken post-harvest in 1992 and again in 2002, rather than in 1999, because 1999 soil samples were not available for the 84-kg treatments. Yield data after 1999 was not used because the objective of the experiment changed to studying the impact of residual N. During 2000 and 2001, beef cattle manure was not applied in the experiment. Total soil C was converted to Mg of C/ha using measurements of bulk density taken at the time of sampling for soil C. Soil samples were taken to 90 cm, but there were no differences in bulk density and soil C below 30 cm. The annual average of soil C changes, calculated by taking the difference between the soil C samples taken post-harvest in 2002 and again in 1999 divided by the number of years between the soil samples, is reported in Table 1.
Carbon Release from Energy Use in Crop Production
Carbon in the form of CO2 is also released into the atmosphere from direct energy use. Carbon is emitted from direct energy consumption, primarily from diesel fuel combustion in field operations. In addition, C releases associated with energy expenditures in the production of fertilizers, chemicals, and equipment in the crop production system (embodied or indirect energy) as well as C contained in the hydrocarbon feedstocks for each fertilizer and chemical input (feedstock energy) were calculated. Released and consumed kg of CO2 per kilojoule (kJ) expended were converted to kg of C per kJ from direct, embodied, and feedstock energy for the fertilizers, chemicals, and equipment used in the production systems.
Carbon-equivalent emissions with respect to direct energy consumption for field operations were estimated by using Eq. [3]. The embodied and feedstock emissions were determined by using Eq. [4]:
 | [3] |
The same format was used in calculating emissions for all field operations, although the fuel consumption in L/ha differs, depending on the field operation performed:
 | [4] |
Energy consumption rates in kilojoules (kJ) and their corresponding carbon emissions were determined for both direct and embodied energies relevant to field management operations and chemical inputs for the production method. These include diesel fuel for tillage, planting, and harvesting (direct energies) of each crop as well as energy in the form of natural gas, fuels oil, and electricity (embodied energies) required to obtain, manufacture, and distribute fertilizers, herbicides, and pesticides. Energy consumption rates were from West and Marland (2002).
The soil sequestration data and estimates of C emissions were used to calculate the net change in C resulting from each cropping system. The net change of C for each system equals the sequestered soil C less atmospheric loading of C in Mg/ha. Table 1 provides a summary of the net C sequestration rates.
Carbon Credits
Equation [5] is used to determine the dollars per Mg of C required to make a system with a greater sequestration rate (C ratej), but smaller net returns (NRj) economically equivalent to a system with a lesser sequestration rate (C ratei), but with larger net returns (NRi). The dollar value of carbon credit is the incentive a manager would need to be indifferent between production systems:
 | [5] |
|
RESULTS AND ANALYSIS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND ANALYSISCONCLUSIONSREFERENCES |
|---|
Net Returns and Costs
The average net return to land and management was positive for all systems (Table 1). The NT systems had larger net returns than did CT systems for all fertilization treatments. This result occurred largely because NT had lower costs than CT. Herbicide costs were only slightly higher for NT systems, but field operation costs were substantially less than those in the CT systems (Table 2). Yield differences were relatively small between CT and NT systems, and were not statistically different. The differences ranged from –113 to 226 kg/ha (Table 1).
|
Although the cost of spreading the equivalent amount of N in beef cattle manure was cheaper than paying for commercial N fertilizer and the cost for spreading it, the cost difference was much smaller than the difference in gross return. The sum of the commercial N fertilizer cost and application cost were $55.28 and $98.62 per hectare for the 84- and 168-kg treatments, respectively, whereas the costs for spreading the equivalent amount of N with beef cattle manure were $46.97 and $93.97 per hectare, respectively (Table 2).
Soil Carbon and Net Carbon Sequestration
The NT systems had greater annual soil C gains than CT (Table 1). The beef cattle manure–fertilized systems had greater annual sequestration rates than commercial N fertilizer systems. This is likely due to the extra C added by the manure. Soil C gains were greatest in the NT168M system at 2.66 Mg/ha/yr. The next-greatest gain in soil C was for NT168N, at 2.53 Mg of C/ha/yr. The CT84N had the smallest rate of gain at 1.16 Mg of C/ha/yr. The CT systems had gains in soil C since the inception of the experiment. The gains in the CT systems are likely due to a change to corn from the previous crop (oat) and a reduction in tillage intensity from moldboard plow to the practices used in the experiment. The difference in the rate of soil sequestration between CT and NT systems in the literature ranges from 0.3 to 0.8 Mg/ha (West and Post, 2002). The range in this study was 0.18 to 1.06 Mg/ha. Therefore, the differences in rates of gain between systems used for the economic analysis are reasonable.
Emissions from direct energy use were greatest for the CT systems because there were more trips over the field. Embodied and feedstock emissions were almost equivalent for CT and NT systems because fertilizer use was the same and chemical applications were relatively similar (Table 1). Total C emissions were greatest for the CT systems (Table 1). Once again, this was primarily because there was substantially more tillage in the CT systems; hence, there were significant C emissions from diesel fuel. Systems using beef cattle manure had greater rates of direct energy use and emissions than commercial N fertilizer systems did because there were more trips over the field to apply manure. The beef cattle manure systems had less embodied and feedstock emissions because commercial N fertilizer was not used, resulting in lower total emissions from production inputs. Further work on the impact of C-equivalent emissions from methane and nitrous oxide released from manure is needed, but was beyond the scope of this study. If manure is already available as an output from an existing cattle enterprise, however, there is some evidence that land application of manure may have less methane emissions than long-term storage of manure in piles (USEPA, 1995).
The rate of net C sequestration for NT, relative to CT systems, and beef cattle manure, relative to commercial N fertilizer systems, increased when C emissions were considered, because they had fewer C emissions. The NT168M had the largest net sequestration rate and CT84N had the smallest net rate (Table 1).
Derived Carbon Credits
Derived C credit values for all technically feasible system comparisons are reported in Table 3, and indicate a substantial range in C credit values. The values in Table 3 where an NT system row intersects with a CT system column are frequently indicated by the letter A. The letter A indicates that a carbon credit is not needed for this system to be preferred to the system in the column because the NT system not only sequesters more C, but also has a larger net return than the CT system it is being compared with. The letter B appears in Table 3 when the system in the row has a smaller sequestration rate than the system in the column; therefore, a credit is not feasible.
|
Carbon credits were also derived without accounting for C in CO2 emissions from production of inputs and energy used in the systems, and are also reported in Table 3. With emissions subtracted from soil sequestration the C credit values required to motivate a switch from commercial N fertilizer systems to beef cattle manure systems were smaller than the C credit values when emissions were not subtracted. This was because the net sequestration rates for beef cattle manure systems were greater, compared with those of commercial N fertilizer systems. The value of the C credit needed declined by $11.08/Mg/yr ($39.84 to $28.76) for CT168M versus CT168N, and declined by as much as $383.68/Mg/yr ($520.29 to $136.61) for NT168M versus NT168N. This is because commercial N–fertilized systems released more C emissions than did beef cattle manure–fertilized systems. The relative difference in sequestration increased for beef cattle manure versus commercial N fertilizer systems; therefore, the credit value was less.
These results indicate that the range in C credits is large. The value of C credits needed will be highly specific to the production system costs and sequestration rates. Relatively small changes in costs between systems that sequester carbon at different rates can have a large impact on the C credit. A $15.00/ha change in cost between two systems, when one system sequesters 0.5 Mg of C/ha/yr more, is a change in C value of $30.00/Mg. The difference in sequestration rates also has an impact. For a $10.00/ha cost difference, the C values for a 0.5 Mg/ha and 1.0 Mg/ha difference in sequestration rates are $30.00/Mg and $15.00/Mg, respectively.
Sensitivity Analysis
Changes in manure application costs and herbicide applications in the conventional tillage systems were examined. Manure application costs, $2.39/Mg, were based on custom application rates. Because availability of custom application may be limited and inconvenient, application costs for owning and operating equipment were also estimated by using a procedure developed by Massey (2002). This analysis found the cost ranges from $2.45 to $4.14/Mg, depending on the field characteristics, equipment operating characteristics, and costs. As a result, the cost of beef cattle manure application may be higher than reflected in the study. As a consequence, the beef cattle manure systems would have relatively smaller net returns than commercial N fertilizer systems. For each $0.25/Mg increase in manure application cost, the net returns from the 84M and 168M systems would decline $5.62/ha and $11.23/ha, respectively.
The NT systems had larger net returns than the CT systems because of lower costs. The NT systems only had two more applications of herbicides than the CT systems did. Additional analysis indicated that, even in the unlikely case in which all herbicides were not needed in the CT systems, the NT systems would still have larger net returns.
The commercial N fertilizer systems had larger net returns than the beef cattle manure systems because of higher yields. At an average price of $0.0989/kg ($2.51/bushel), a 251 kg/ha (4 bushels/acre) change in yield between systems results in a change in net return of $24.83/ha.
Net Return Variability Analysis
Although examining average net returns is useful, it is also important to examine variation in net returns to determine if risk affects the decision to use one system or another. Many farm managers are risk averse and will accept less dollars of return for less dollars of variability or loss. Each decision maker trades off risk and return at their own rate, so it is difficult to prescribe a specific strategy for any one manager, but some general conclusions can be made with the use of decision criteria.
Decision criteria are used by researchers to evaluate and compare the net return variability (risk) of alternative production systems or management strategies. One commonly used decision criterion is the mean-standard deviation. Risk-averse managers generally prefer systems that have both the largest mean net return and smallest the standard deviation. Another criterion that can be employed is the coefficient of variation (CV), which measures risk (as measured by standard deviation) relative to the mean net return. The CV is simply the standard deviation of a distribution divided by its mean. The "maximin" criterion, which compares the minimum net return across systems to determine the largest value, can also be used. This comparison is useful because extremely risk-averse managers select the system with the largest minimum net return or smallest negative loss. In addition, the probability of having a loss can be compared across systems with data from a cumulative probability function for each system. Risk-averse managers select the system that minimizes the probability of loss.
Beginning with the mean-standard deviation criterion, we found that there is no strategy that had the largest mean net return and smallest standard deviation (Table 4). The system with the least amount of net return risk, as measured by standard deviation, was CT84M. This system had the smallest average net return and the largest CV, however, indicating a large degree of risk relative to its expected return. There would be a significant amount of net income given up to reduce risk with this system. It also had the next-to-the-smallest carbon sequestration rate. There are also five other systems that had larger minimum net returns. Therefore, there would be little motivation for a manager to use this system for net return risk reduction or carbon sequestration.
|
When the minimum net returns were compared, and the maximin decision criterion was employed, the NT84N system was preferred (Table 4). The NT168N system had the next-highest minimum net return. The probability of having a negative net return was also derived by forming a cumulative probability function from the simulated net returns. The cumulative probability distributions of net returns from SIMETAR for the four most relevant systems are illustrated in Fig. 1. The other four systems were excluded from the figure for purposes of clarity and relevance. This graph was used to determine the probability of a net return above or below a specific level of net return. Stochastic dominance is a risk analysis technique that is used to choose among a set of alternatives by comparing the distribution of possible returns for each strategy and selecting preferred strategies based on risk preferences and not just the mean and standard deviations.
|
There was also no single strategy selected as preferred with second-degree stochastic dominance criteria. Second-degree holds for those decision makers who are risk-neutral or risk averse. A strategy that is second-degree efficient will have a smaller area under its cumulative probability distribution (for each and every observation of net return) than those that are not, because the area is summed across observation from lowest to highest.
Figure 2 reports the probability of having a net return of $0.00 or less for each system, on the basis of the cumulative probability functions in Fig. 1. The NT84N had the smallest probability of loss, at 17%, with NT168N a close second at 19%. In other words, these two systems had an 83 and 81% chance of having positive net returns in any given year, respectively. Figure 2 also illustrates the probability of having net returns larger than the average net return across all systems of $109.49/ha. The NT168N had the greatest probability (56%) and NT84N had the next-greatest (47%). This figure also shows that the probability of having a net return greater than $0.00 is 83% for NT84N and 81% for NT168N. In addition, the cumulative probability of net returns illustrated in Fig. 1 indicates that NT168N had larger net returns than NT84N 77% (1 – 0.23) of the time.
|
The general results were unaffected by considering risk; NT was still preferred to CT and commercial N fertilizer was still preferred to beef cattle manure. However, the selection of the preferred level of commercial N fertilizer or beef cattle manure for a given tillage system may be influenced by both risk preference and the dollar value of carbon credits. Additional research is being performed that uses more advanced risk analysis procedures to determine the carbon credit value needed by managers with various risk preferences to encourage adoption of systems that increase carbon sequestration. These methods are beyond the scope of this paper.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND ANALYSISCONCLUSIONSREFERENCES |
|---|
Carbon credits were derived with, and without, accounting for CO2 emissions from production of inputs and energy used in the cropping systems. Although excluding CO2 emissions did not change the general results, the value of C credits needed to adopt beef cattle manure–fertilized systems increased. Therefore, consideration of emissions from energy used for each production system is important.
This analysis showed that using manure as a substitute N source was not as profitable as using commercial N, even with only application costs being the only expense for manure. However, managers who currently have to dispose of manure may find that land application of it on cropland is less costly than other methods and choose to apply manure to cropland rather than use commercial N. These managers would be in a position to benefit from selling carbon credits and could collect larger per acre payments than those using commercial N–fertilized systems because manure-fertilized systems have the highest soil C-sequestration rates. This benefit may be even larger if land application of manure has fewer C-equivalent methane emissions than other manure management methods and net C-sequestration rates are used as a basis for carbon credit payments.
There are currently two primary carbon credit markets. The Economist in 2005 reports that the European Union has a market where trades have ranged from $10 to $35/Mg of CO2 ($33 to $117/Mg of C). However, U.S. credit suppliers including farm managers cannot participate in this market as the United States has not ratified the Kyoto Protocol. There is also a pilot carbon credit trading market on the Chicago Climate Exchange (CCX). Farmers in parts of the Midwest that use no-tillage crop production practices are receiving credits to sell on the exchange that are equal to 0.5 Mg of C/acre/year. Carbon credit prices on the CCX in late spring of 2005 ranged from $3.70 to $5.04/Mg of C.
This research can be used by policymakers to help formulate greenhouse gas policy in the United States and by extension educators and farm managers to help determine the feasibility of producing carbon credits for sale in emerging carbon markets. The status of a permanent carbon credit market in the United States is uncertain at this time.
This study makes a new contribution to the literature because it considers the economic impact manure applications and atmospheric release of C have on derived carbon credit values. Several limitations to this study should be noted. First, only two rates of N were applied throughout the duration of this study. Second, the net returns were simulated because only nine years of empirical yield data were available. Finally, the organic C data was sampled three years after the last year of the experiment. Future research should also consider the impact of C equivalent emission from methane due to the use of manure.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND ANALYSISCONCLUSIONSREFERENCES |
|---|
lawecon/group10.html (verified 8 Nov. 2005). School of Public and Environ. Affairs, Indiana Univ., Bloomington, IN.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Soil Science Society of America Journal | Journal of Plant Registrations | The Plant Genome | |||
