Spatially delineated carbon credit potential and implied nutrient reduction co-benefit: An assessment with integrated ecological and economic modeling framework

FINAL REPORT

What were the key research questions this project hoped to answer? 

The main research questions of this project are how farmers respond to natural factors such as soil and weather conditions, as well as socio-economic contexts including market and policy incentives, and what the nutrient reduction potential is of various carbon initiatives.

Findings:

We used data from the Iowa Farm and Rural Life Poll (IFRLP), a longitudinal panel survey of Iowa farmers. We analyze data from the 2014, 2016 and 2018 waves of the IFRLP, three waves that asked about farmers’ adoption of cover crops. The IFRLP is unique due to its long-running nature and its focus on Iowa farmers’ perspectives on various agricultural issues. The dependent variable is the binary adoption decision in 2017, with a total sample size of 501 after filtering out missing values (386 non-adopters and 115 adopters).

Past adoption behavior, an important predictor constructed based on the adoption status in 2013 and 2015 is categorized as follows: non-adoption in both years (0,0), adoption in 2015 but not in 2013 (0,1), adoption in 2013 but not in 2015 (1,0), and adoption in both years (1,1), with non-adoption in both years as the reference group. Two other key dummy variables are crop insurance purchase decision in 2015 and cost-share program participation in 2013. The dummy variable for cost-share received four years ago allows us to assess the long-term effectiveness of the cost-share program and determine whether respondents discontinue cover crop use after the cost-share payment contract ends. Past adoption behavior is a strong predictor of adoption in 2017. Farmers who adopted cover crops in both previous periods (1, 1) show a 25.7% higher likelihood of adoption in 2017 compared to those who never adopted. Moreover, farmers who adopted in the later period (0, 1) demonstrate a 29.3% increase in adoption probability, relative to the same reference group of non-adopters in both periods.

While our model uses adoption in 2017 as the dependent variable (with adoption = 1 and non-adoption = 0), it also helps us understand the potential for past adopters to discontinue the practice, which we refer to as disadoption. Based on the full set of respondents, farmers who were non-adopters in 2013 but adopted in 2015 have a 30.9% possibility of disadopting by 2017. This suggests that recent adopters have a relatively high likelihood of continuing the practice, with about one-third reverting to non-adoption. For farmers who adopted in 2013 but not in 2015, the chance of being a non-adopter (disadopting) in 2017 reaches 79.1%. Consistent adopters (adopted in 2013 and 2015) show the lowest disadoption rate, with only a 28.9% chance of becoming non-adopters (disadopting) in 2017, indicating that they are more likely to continue the practice, though there is still a notable probability of disadoption even among this group. %.

Cost share participation in 2013 decreases the adoption probability by 6.1%, but it is not statistically significant. Using the DLEM-catchment model developed in Cao et al. (2023), we evaluated multiple in-field practices (cover crops, fertilizer rate reductions, optimized timing, and crop rotation) and edge-of-field measures (saturated riparian buffers and wetlands) in four Iowa catchments during 2000–2019.

Cover crops and fertilizer rate reductions emerged as the most effective in-field strategies, substantially lowering nitrate losses with only minor yield penalties. Wetlands contributed an additional 19–36% reduction at the catchment scale, whereas saturated riparian buffers accounted for 1% due to limited suitable siting and the small fraction of tile drainage intercepted. The magnitude of nitrate reductions and associated yield responses varied substantially among catchments and across years, underscoring the importance of locally tailored strategies (Huang, in prep).

Because these practices have minor impacts on SOC stock, we did not use them to examine carbon payment systems. Using cover crops as our case study, we evaluate soil carbon stock (SOC) changes and crop yield impacts using model simulation by Cao et al. (2023) based on historical climate data from 2000 to 2019. We assume SOC stock changes primarily reflect CO₂ dynamics, as crop-assimilated carbon is harvested, consumed and released elsewhere. To estimate SOC change variability, we develop five climate scenarios that alter the frequency and sequence of historically observed weather conditions and evaluate SOC under conditions with and without cover crops, keeping all other model input drivers and parameters unchanged.

• Scenario 1 uses historical climate data from as recorded.
• Scenarios 2 and 3 replace normal weather years with dry and wet years, respectively.
• Scenario 4 creates intensified wet-dry cycles by removing normal years.
• Scenario 5 eliminates extreme conditions, using only normal years.

Building findings of a prior INRC project, we further refined our analysis of payment schemes and impacts on farmers’ response to the schemes. We show that practice-based payments achieve substantially higher acreage enrollment. The performance-based schemes show progressively declining enrollment as behavioral factors are incorporated. At a $300,000 budget, practice-based payments enroll nearly twice as many acres (9,630 acres vs. 4,830 acres) as performance-based payments under risk and ambiguity aversion. At a $600,000 budget level, the difference grows to 17,000 vs. 5,430 acres, which is more than a threefold difference. In our simulated watershed, this translates to an additional 5,000-11,000 acres enrolled using practice-based payments. Performance-based payments show progressively steeper curves, particularly when behavioral factors are incorporated. The supply curve under ambiguity is steepest, followed by risk aversion and profit maximization, demonstrating how uncertainty aversion substantially increases the marginal cost of achieving additional carbon sequestration. These supply curves further support our central findings about the accountability-cost-effectiveness tradeoff in carbon payment design. While performance-based payments may theoretically offer better targeting of high-potential carbon sequestration, the compensation requirements for uncertainty-averse farmers make these schemes substantially less cost-effective in practice.

Besides analyses for different carbon payment mechanisms and climate scenarios that we develop and simulate, we also assessed the policy impacts of our findings across a broader geographic scale. We analyzed a scenario where Iowa’s 259,079 acres of 2019 EQIP cover crop contracts expire simultaneously, and farmers choose between carbon payment alternatives. Our watershed simulation results indicate that achieving complete Iowa enrollment requires $9.164 million for full participation under practice-based payments while performance-based payments under ambiguity plateau at 82,825 acres (32% participation) regardless of available funding. This participation gap would expand substantially when scaling to the Upper Mississippi River Basin, which contains approximately twice the agricultural area of Iowa. It further suggests that payment design may ultimately be more consequential than price level in determining the success of carbon programs. As agricultural carbon markets transition from pilot efforts to large-scale deployment, the tradeoff between accountability and cost-effectiveness may become a binding constraint on market development.

Related accomplishments and activities 

1 presentation: “Risk and Ambiguity Aversion in Conservation Practice Adoption and the Effectiveness of Carbon Payment Systems.” (with Du, Zhushan, Peiyu Cao and Crystal Lu) (presented at AERE annual conference, May 2025) 

Publications:

• Du, Z., H. Feng and J. Arbuckle. 2025. “Not ready for a long-term commitment? Analyzing temporal variability in cover crop adoption.” Journal of Soil and Water Conservation. Forthcoming. DOI: 10.1080/00224561.2025.2533102.
• Du, Z., H. Feng and J. Arbuckle. 2025. “Exploring the Complementarity Between Traditional Econometric Methods and Machine Learning – An Application to Adoption and Disadoption of Conservation Practices.” Applied Economics 1-16.
• Du, Z., H. Feng and J. Arbuckle. 2025. “The interactions between crop insurance and conservation practices - Insights from Program Initiatives and Farm Surveys.” Choices Magazine. Forthcoming
• Du, Z.H., H. Feng, and L. Schulte Moore. “Conservation Investment and Carbon Payments in US Agriculture: Implications of the Inflation Reduction Act of 2022.” Agricultural Policy Review, Fall 2022. Center for Agricultural and Rural Development, Iowa State University.
• Du, Z., H. Feng, and W. Zhang. “Carbon and Nutrient Co-benefits of Large Conservation Programs: An Illustration with EQIP in Iowa.” Agricultural Policy Review, Spring 2022. Center for Agricultural and Rural Development, Iowa State University.
• Zhushan Du. “Five essays on conservation practice adoption.” 2025. PhD dissertation. Department of Economics, Iowa State University.

Three graduate students and one undergraduate student were employed to assist with this project.