Error analysis and quality improvement of the Coshocton weighing lysimeters

Weighing lysimeter-computed evapotranspiration (ET) is sometimes used to assess ET computer-model predictions and to compare water use under different management or climatic conditions. For any use of lysimeter-computed ET the uncertainty associated with this value should be known and reported. The uncertainty associated with individual lysimeter-mass measurements has been previously presented, but ET-measurement uncertainty is a function of water-budget measurements (percolate volume, water-storage change, etc.), measurement-component uncertainty (uncertainty in mass and percolation measurements), and systematic error (e.g., lend-cell and potentiometer slope bias). The Coshocton lysimeters were not sufficiently accurate to compute hourly ET to the desired depth resolution, thus a quality-improvement plan was initiated that utilized Pareto analysis. The objectives of this article are to: (I) derive expressions for ET uncertainty; (2) determine the effect of the quality-improvement procedures applied to the Coshocton lysimeters on ET-measurement uncertainty; (3) perform an error analysis on the measurement components of the improved Coshocton weighing lysimeters. The quality improvements reduced the ET uncertainty from 0.36 mm/d to 0.032 mm/d under the condition of no rainfall (ET and percolation only). The ET-measurement uncertainty of the improved Coshocton lysimeters was sensitive to various measurement components especially the number of rainfall events and the uncertainty associated with mass and percolation measurements. This indicates that the improved measurement systems of the Coshocton weighing lysimeters are sufficiently accurate to compute hourly ET and that lysimeter-mass and percolation measurements should be monitored regularly (quality-control procedures should be followed).