112-1 Does Potassium Management Buffer Crop Yield from Water Stress? a Meta-Analysis.
Monday, October 23, 2017: 1:35 PM
Tampa Convention Center, Room 9
Super-optimal rates of potassium (K) application have been suggested as a fertility management strategy to mitigate the negative impacts of drought stress on crop yields. Following a scoping study, screening criteria were developed to conduct a systematic review and meta-analysis of published field studies that evaluate crop yield response to water stress and K availability. Studies were compiled into a database, including published treatment yield averages and variance measures for each treatment level. A network-based meta-analysis was conducted using a linear, mixed-effects model in R-software via the ‘metafor’ package. Preliminary results validated the methodological approach for categorical analysis of multiple factors as a network. Yield losses due to moderate and severe water stress were substantial compared to well-watered controls, and yield increased incrementally with increasing K rate application. Under well-watered conditions, super-optimal rates of K did not produce a yield increase compared to optimal rates. A pairwise comparison of optimal and super-optimal fertilization rates was conducted to quantify the extent of drought stress mitigation by K fertilization. Preliminary results indicate that while the summary effect of super-optimal application of K improved yield response slightly under drought stress, the significance of such gains is difficult to quantify due to the systemic limitations of sampling variance measurement and reporting in published studies. Crop type and relevant moderating variables were explored, as were several approaches to improve variance calculations and estimated confidence intervals. Results of meta-analysis will be presented, including a brief overview of systematic review screening criteria. Discussion will summarize results and key statistical assumptions in the context of soil fertility and water management recommendations, specifically with respect to their influence on categorical frameworks required for future analysis. This includes consideration of the opportunities and challenges of synthesizing journal articles into actionable knowledge using statistical meta-analysis.
Previous Abstract | Next Abstract >>