|Donoghue, Ann - Annie|
|Van santen, E|
Submitted to: International Soil Tillage Research Organization Proceedings
Publication Type: Abstract only
Publication Acceptance Date: 11/25/2002
Publication Date: 7/13/2003
Citation: Terra, J.A., Reeves, D.W., Shaw, J.N., Raper, R.L., Van Santen, E., Mask, P.L. 2003. Soil management, terrain attributes and soil variability impacts on cotton yields. International Soil Tillage Research Organization Proceedings. Interpretive Summary:
Technical Abstract: Soil management practices impact soil quality and crop productivity differentially across landscapes but these impacts are rarely assessed at the landscape level. We evaluated cotton (Gossypium hirsutum L.) yield response to soil management practices and their interactions with landscape and soil attributes in a 9 ha strip trial in Alabama, USA (Aquic and Typic Paleudults) during 2001 and 2002. A soil survey, geo-referenced elevation, and electrical conductivity (EC) maps were developed for delineating field variability. Four treatments in a corn (Zea mays L.)-cotton rotation [conventional tillage (CT), conventional tillage + dairy bedding manure (CTM), no-till (NT) and no-till + manure (NTM)] were established in strips intercepting zones of landscape variability in a RCB design with 6 replications. Eight terrain attributes were calculated. Data was analyzed using analysis of variance and regression analysis. Due to 27% less rainfall in 2002, seed cotton yield was 50% lower than in 2001 (2983 kg ha-1). No-till and NTM yields (2433 kg ha-1) were 14.5% higher than CT and CTM and the coefficient of variation was greater in conventional systems (13.4%) compared to no-till systems (10.8%). Electrical conductivity, slope, soil texture and elevation explained between 15 to 60% of yield variability but simple subjective terrain attribute information was similarly effective in delineating variability. Elevation, slope, EC and texture were the main variables explaining yield spatial variability within treatments. However, when terrain attributes were used as covariates, little improvement in model fit was obtained. The no-till systems had greater impact in the drier year and in zones with lower yield potential.