|Causarano, H - VISITING SCIENTIST|
|Izaurralde, R - UNIVERSITY OF MARYLAND|
Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: December 16, 2008
Publication Date: February 17, 2009
Citation: Doraiswamy, P.C., Causarano, H.J., Izaurralde, R.C., Daughtry, C.S., Hatfield, J.L. 2009. Mapping soil organic carbon in the U.S. corn belt [abstract]. The North American Carbon Program, 2nd All-Investigators Meeting. 2009 CDROM. Technical Abstract: Soil carbon sequestration in the U.S. Con Belt was studied to evaluate the impact of land-use, soil and crop management that can play a significant role in promoting the mitigation of atmospheric CO2 and improve soil quality such as reduced erosion and also increase soil organic carbon (SOC) content. The objective of this research was to study the rate of soil carbon sequestration across the study region to understand the potential sinks and recommend soil and crop management practices to enhance or maintain soil carbon sequestration rates. The Environmental Policy Integrated Climate (EPIC) model was adapted to study the long term impact of soil and crop management practices on soil carbon sequestration in the U.S. Corn Belt. Landcover classification from Landsat imagery was use to develop the baseline soil carbon and to project the potential future trends in soil carbon sequestration rates. A detailed and computationally intensive approach integrating the EPIC model with soil, climate, land use, and management data was developed to conduct simulations at a grid-cell level of 2.59 km2. Based on the soils data, current SOC stocks are estimated to be 11–157 Mg C ha-1 in the top 20 cms across the study region. Uncertainties were 26–30% when simulations were conducted at the regional scale but reduced to 8–11% with site-specific simulations. The suggested potential enchantments in the rates of soil carbon sequestration are based on adaptation of long-term prescribed tillage practices, residue management, clay content, slope and elevation. A web-based decision support system was developed for end-users at the farm and regional scales to optimize management practices to maximize the potential increase in SOC and maintain crop production. The potential use of visible/near-infra red imagery at multiple resolutions was evaluated for mapping soil tillage and surface residue levels. The spectral angle mapping (SAM) supervised classification method was applied to map tillage practices using QuickBird, SPOT and AWiFS imagery. Ground truth data provided training data and the accuracy levels for till and no-till classification were between 80-90 %. Limitations in the use of these sensors and the differences between fall and spring assessments are discussed.