Submitted to: Agricultural Water Management
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 10/10/2006
Publication Date: 11/15/2006
Citation: Vaughan, P.J., Trout, T.J., Ayars, J.E. 2006. A processing method for weighing lysimeter data and comparison to micrometeorological eto predictions. Agricultural Water Management, Volume 88, Issues 1-3, March 16, 2007, Pages 141-146.
Interpretive Summary: Lysimeters are devices for making continuous measurements of the weight of a column of soil that is enclosed in a box but open at the top. The lysimeter is partially buried so that the soil surface of the lysimeter is flush with surrounding soil. A crop can be grown on the lysimeter’s soil and the record of weight changes during each day provides an accurate way to determine how much water the crop loses by evapotranspiration (the combination of soil surface evaporation and transpiration by plants). Our work involves the development of a new computer program that provides a more accurate representation of the hourly evapotranspiration rate than previous computations. We have applied this program to lysimeter weight measurements made continuously during 2004 and 2005 at the University of California West Side Research and Extension Center. A tall fescue grass was grown in this lysimeter so the hourly evapotranspiration rates could be compared directly to predictions of hourly evapotranspiration rate by the California Irrigation Management Information System which maintains a weather station a few meters from the lysimeter. Our results demonstrated that the predictions based on the weather station data agree closely with evapotranspiration rates computed from the lysimeter weight. This work is important in developing accurate representations of water use by agricultural crops and will enhance the efficiency of irrigation in California.
Technical Abstract: Weighing lysimeters make direct measurements of water loss from a growing crop and the soil surface around a crop (evapotranspiration, ET) and thus provide basic data to validate other ET prediction methods. The objective of this study was the improvement of the lysimeter data processing to eliminate bad data and minimize the effect of random variations in the measurements. Also, grass evapotranspiration rates (ETo) were compared with evapotranspiration rates predicted from data taken at a nearby weather station that is part of the California Irrigation Management Information System (CIMIS). The ETo data were calculated from lysimeter mass data using a 7 point Savitsky-Golay approximation of the derivative at the central point. The grass lysimeter located at Five Points, California generated hourly ETo data for a two-year period that were screened to remove corrupt data and data taken on rainy days (20% of all data). All the remaining data were plotted against the CIMIS hourly ETo predictions and the best-fit line almost matched the 1:1 line indicating that CIMIS ETo predictions were accurate, on average. Slopes of zero-intercept, best-fit lines to CIMIS ETo vs. lysimeter ETo for data representing each hour from 8:00 AM through 2:00 PM the CIMIS ETo were consistently greater than unity reaching a maximum of 1.27 at 10:00 AM. Slopes were consistently less than unity at other times. Analogous slopes of CIMIS Penman-Monteith (PM) ETo plotted against lysimeter ETo showed the same pattern with a maximum of 1.21 at 10:00 AM and were closer to unity during the day. The soil heat flux calculated by the PM model probably accounts for this improvement. Mean wind speed increased during the hours 10:00 AM to 4:00 PM when the CIMIS ETo slopes were decreasing indicating that error in the wind function was not the cause of the discrepancy between CIMIS and lysimeter ETo. Between midnight and 5:00 AM correlation of hourly CIMIS and lysimeter ETo was insignificant (r2<0.14). These results suggest that CIMIS hourly ETo predictions at this site could be improved by characterization of factors that would improve prediction of ETo at night. Daily CIMIS ETo predictions exceed lysimeter ETo in the range 2-7 mm d 1 but are smaller when lysimeter ETo>8 mm d-1. For daily ETo<2 mm d-1 the CIMIS predictions generally agree with lysimeter ETo. This non-linear relationship deserves further study.