Noise Reduction Methods for Weighing Lysimeters

Authors: Vaughan, Peter; Ayars, James

Submitted to: Journal of Irrigation and Drainage Engineering
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/14/2008
Publication Date: 4/1/2009
Citation: Vaughan,P.J. and J.E. Ayars. 2009. Noise reduction Methods for Weighing Lysimeters. Journal of Irrigation and Drainage Engineering. Vol. 135, No. 2:235-240.

Interpretive Summary: A weighing lysimeter is a device that regularly measures the weight of a large box of soil that is open at the top and has plants growing in it. Normally weighing lysimeters are installed underground in the middle of agricultural fields so that the top of the soil box is flush with the surrounding field. Under these conditions a crop can be grown in the surrounding field and in the soil box. Hourly measurements of the weight of the box can be used to calculate how much water the crop has lost to the atmosphere. This provides important information on how much irrigation water the crop needs. The weighing lysimeters operated at the University of California West Side Field Research and Extension Center in Five Points, California vibrate at a frequency around 10 cycles per second. The cause of this vibration is not well understood but the vibration could cause a reduction in the accuracy of the weight measurements. We have developed two numerical methods to reduce noise associated with this vibration and improve accuracy. After tuning each method we were able to reduce noise by about the same factor. One of the methods which contained a numerical technique known as Fourier transform digital filtering was the better choice because it required only a few minutes to start generating good quality data. The other method was a least-squares filtering technique that required about an hour and a half before generating good quality data after starting the data collection computer program. These results will assist researchers who use weighing lysimeters by providing them with computer programming methods to improve the accuracy of their measurements.

Technical Abstract: Mechanical vibration of the grass and crop weighing lysimeters, located at the University of California West Side Field Research and Extension Station at Five Points, CA generated noise in lysimeter mass measurements and reduced the quality of evapotranspiration (ET) data. Two filtering methods for reducing noise are the Savitsky-Golay (S-G) moving average filtering and frequency domain filtering. Either method can be applied to real-time data. Analysis of data from the grass lysimeter using the 7-point and 15-point S-G methods resulted in noise reduction factors of 0.83 and 0.54, respectively. Frequency domain filtering is a convolution of acquired data with a weighting function. Testing of a parameterized trapezoidal weighting function determined that the optimal function of this type was an asymmetric triangular function. The noise reduction factor for this function was calculated as the ratio of standard deviation of the filtered data to that of unfiltered data. For the crop lysimeter this ratio was 0.53, also a significant improvement. Frequency filtering is preferable because it achieves significant improvement but does not require a startup time during which other arrangements for estimating ET would be needed. The two methods could be combined for further improvement.