Evaluating the interactions of crop management, carbon cycling, and climate using Earth system modeling and remote sensing

Abstract

Crop management practices, such as soil tillage and crop residue management, are land management activities with potentially large impacts on carbon (C) cycling and climate at the global scale. Improvements in crop management practices, such as conservation tillage or 'no-till' (NT), have been proposed as climate change mitigation measures because such practices may alter C cycles through increased sequestration of soil C in agricultural soils. Despite their potential importance, regional to global scale data are lacking for many crop management practices, and few studies have evaluated the potential impact of the full range of crop management practices on C cycling and climate at the global scale. However, monitoring of crop management practices is crucial for assessing spatial variations in management intensity and informing policy decisions. Inclusion of crop management practices in Earth system models used for assessing global climate is a key requirement for evaluating the overall effects of different crop management practices on C cycling and their potential to mitigate climate change. Studies in this dissertation seek to address these issues by: (1) evaluating the efficacy of remote sensing methods for monitoring differences in soil tillage and crop residue management practices in Iowa; (2) incorporating soil tillage practices into an Earth system model and assessing the potential for soil C sequestration and climate change mitigation through adoption of NT practices; (3) assessing the historical impact of including the full range of crop management practices (residue harvest, grain harvest, soil tillage, irrigation, and fertilization) on changes in C cycling associated with land use and land cover change (LULCC) to crops in an Earth system model. The remote sensing study found that performance of the minimum Normalized Difference Tillage Index (minNDTI) method for assessing differences in tillage and residue management was below average compared to previous studies, even when using imagery from both Landsat 8 and Sentinel-2A sensors. Accurate assessment of these practices using minNDTI was hindered by issues with image quality and inability to obtain sufficient cloud-free, time series imagery during the critical planting window. Remote sensing research aimed at obtaining regional to global scale data on tillage and residue management practices is likely to continue to face these issues in the future, but further research should incorporate additional sensors and assess the efficacy of the minNDTI method for multiple locations and years. Adoption of NT practices in the Community Land Model, which is the land component of the Community Earth System Model, resulted in a cumulative soil C sequestration of 6.6 – 14.4 Pg C from 2015 – 2100 under a future climate change scenario (Representative Concentration Pathway 8.5), and cumulative soil C sequestration was equal to approximately one year of present-day fossil fuel emissions. Adjusting for areas where NT is already practiced had minor impacts on cumulative soil C storage, reducing gains in soil C from NT adoption by 0.4 – 0.9 Pg C globally. These results indicate that soil C sequestration and potential for climate change mitigation through NT may be more limited than has been anticipated elsewhere. Soil C sequestration via NT adoption was highest in temperate regions of developed countries with high initial soil C contents, indicating these areas should be targeted for NT adoption. Simulating the full range of crop management practices in the Community Land Model resulted in an increase in C emissions due to LULCC of 29 – 38 Pg C compared to scenarios with generic crops and model defaults. Individual crop management practices with the largest impact on LULCC emissions were crop residue harvest (18 Pg C), followed by grain harvest (9 Pg C) and soil tillage (5 Pg C). Although implementation of crop residue harvest and soil tillage was extreme in this study, these results imply that Earth system models may underestimate emissions from LULCC by excluding the full range of crop management practices. Studies in this dissertation corroborate the importance of crop management practices for C cycling and climate, but further research on these management practices is needed in terms of data collection, improving process-level understanding, and inclusion of these practices in Earth system models.