Evapotranspiration-based irrigation scheduling tools for use in eastern Colorado

Abstract

Accurate evapotranspiration (ET) information can be used to improve irrigation water management in eastern Colorado. Crop ET information can be used to help an irrigation manager make decisions on when to initiate irrigation and to determine how much water should be applied. ET information can be obtained through the use of specialized equipment, estimated using models, or obtained from sources such as Colorado Agricultural Meteorological Network (CoAgMet) (http://climate.colostate.edu/~coagmet/). This study has one main focus, the testing of tools for use in ET-based irrigation scheduling. The purpose of the first part of this study was to develop and test two irrigation scheduling tools, one for use with annual crops (Colorado Irrigation Scheduler: Annual (CIS-A)) and the other for use with forage crops (Colorado Irrigation Scheduler: Forage (CIS-F)). The tools use ET information calculated using the ASCE Standardized Reference ET equation to track the daily soil water balance in a crop's root zone and make recommendations on irrigation timings and amount of water to be applied. The second part of this study tested the accuracy of a Model E atmometer (ETgage Company, Loveland, CO, USA) in providing estimates of reference ET in southeastern Colorado. In the first part of the study the CIS-A was tested at two sites (north and south) during the 2010 - 2012 growing season in a corn (Zea mays L.) field located near Greeley, Colorado. The results of the study indicated that the performance of the tool was acceptable based on the relatively small magnitude of errors in the estimated deficits compared to total available water (TAW) in the soil profile. RMSE was at most 15.3% of TAW, as was the case in 2012 at the north site, and was as low as 8.6% of TAW in 2011 at the north site. The CIS-A tended to overestimate the observed deficit during all years of the study and across all sites (relative error, RE = 13.58% and mean bias error MBE = -3.41 mm). Overall average error indicated that the CIS-A was within 15.92 mm (root mean square error, RMSE) and 12.61 mm (mean absolute error, MAE) of the observed deficit for the entire study. Satisfactory performance of the CIS-A was observed in all years and across all sites with the exception of 2012 at the north site. In 2012 the performance of the CIS-A was less than acceptable (RMSE = 22.89 mm, MBE = -12.86 mm, MAE = 18.01 mm, and RE = 30.85%). Evaluations of the CIS-F during the 2010 and 2011 growing seasons showed mixed results. The CIS-F was tested on two weighing lysimeters in two different alfalfa (Medicago sativa L.) fields located at the Arkansas Valley Research Center (AVRC) near Rocky Ford, Colorado. During both years of the study the CIS-F tended to overestimate the observed deficit. The CIS-F performed best in 2011(RMSE = 22.02 mm, MBE = -16.95 mm, MAE = 17.65 mm, and RE = 18.73%) with a RMSE within 6.6% of TAW. In 2010 poorer results were obtained (RMSE = 38.21 mm, MBE = -32.84 mm, MAE = 32.94 mm, and RE = 34.11%). However, in 2010 RMSE was still within 11.5% of TAW. Upon further analysis it was found that much of the error encountered during the evaluation of the CIS-F occurred early in each growing season. It was determined that during this period, crop ET (ETc) estimated using the scheduler was higher than lysimeter measured ET. The difference between lysimeter measured ET and ETc estimated using the ASCE (2005) hourly guidelines for a tall crop and crop coefficients developed using data from the lysimeters was determined to be the major source of the error experienced during both growing seasons. ETc was found to be significantly higher than lysimeter measured ET during the initial part of the alfalfa growing season causing the CIS-F to estimate a deficit greater than what was observed. The objective of the second part of this study was to determine if an ETgage Model E atmometer, equipped with a canvas #54 cover, could be used to effectively estimate alfalfa reference ET. The ASCE Standardized Alfalfa Reference ET Equation (ASCE ETrs) was used as the standard for comparison of atmometer ET values to determine atmometer performance. Four years of alfalfa ET, as determined by an atmometer (ETgage), were compared to ASCE ETrs. Daily as well as 2, 3, 5, and 7 day sums of daily ETgage and ASCE ETrs were compared using simple least-squares linear regression. Coefficients of determination (R2) between daily ETgage and ASCE ETrs for all years were greater than or equal to 0.80. Throughout the study, the atmometer tended to underestimate ASCE ETrs. Average seasonal underestimation of ASCE ETrs measured by the atmometer ranged from 9.06% to 18.9%. Root Mean Square Error (RMSE) and Mean Bias Error (MBE) ranged from 1.14 to 1.82 mm d-1 and -0.66 to -1.51 mm d-1, respectively. The atmometer underestimated daily ASCE ETrs 88% of the time, with an average underestimation of 1.30 mm d-1. Under estimation of ASCE ETrs measured by the atmometer occurred most often on days when mean daily horizontal wind speeds were greater than 2 m s-1 or when mean daily air temperatures were below 20 °C. The atmometer performed best when the alfalfa was at reference condition. Localized calibration equations for reference and non-reference conditions with a temperature correction were developed to improve accuracy, with average magnitude of MBE reduced from -0.97 mm d-1 to 0.13 mm d-1.