Submitted to: Journal of Hydrometeorology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/1/2005
Publication Date: 12/5/2005
Citation: Meek, D.W., Prueger, J.H., Kustas, W.P., Hatfield, J.L. 2005. Determining meaningful differences for SMACEX eddy covariance measurements. Journal of Hydrometeorology. 6(6):805-811.
Interpretive Summary: Measurements of energy exchanges near the earth's surface are critical to understanding the potential impact of humankind's effect on the global climate. There are a number of questions about the variation in the energy exchanges across landscapes and different production systems. One of the critical components in accurately assessing whether the observed differences observed across different management practices or within landscape positions of the same management practice. These observations are made with similar instruments and one of the questions that occur is what level of difference indicates a real difference among sites that is greater than the instrument error. We conducted an experiment as part of a larger field-scale study to determine the variability that exists among the same instruments when placed over the same surface. Comparisons were made before and after the field study. These comparisons and subsequent analyses provided a base for evaluating differences among instruments. These analyses revealed there was no bias among instruments and the differences among instruments were quite small. These results provide scientists with techniques to conduct comparisons among instruments and users of the larger field scale data base with the confidence of significant differences among instruments.
Technical Abstract: Two eddy covariance instrument comparison studies were conducted before and after the SMACEX field campaign to (1) determine if observations from multiple sensors were equivalent for the measured variables over a uniform surface and to (2) determine a least significant difference (LSD) value for each variable to discriminate daily and hourly differences latent and sensible heat and carbon dioxide fluxes, friction velocity, and standard deviation of the vertical wind velocity from eddy covariance instruments placed in different locations within the study area. In central Iowa, two intercomparison studies were conducted in early June over an alfalfa field and in mid-September over a short grass field. Several statistical exploratory, graphical and multiple comparison procedures were used to evaluate each daily variable. Daily total or average data were used to estimate a pooled standard error and corresponding LSD values at the P=0.05 and P=0.01 levels using univariate procedures. There were no significant sensor differences in any of the daily measurements for either intercomparison period. Hourly average data were used to estimate a pooled standard error and corresponding LSD values at the P=0.05 and P=0.01 levels using mixed model procedures. Computed LSD values were used to determine significant daily differences and threshold values for the variables monitored during the SMACEX campaign. Sensor differences for pre and post intercomparisons were minimal for hourly and daily values of CO2, water vapor, sensible heat, friction velocity and standard deviation for vertical wind velocity.