Validation Data

The Water Irrigation Scheduler for Efficiency (WISE) system was tested on 20 site years for the following crops: alfalfa, corn, potato and sugar beets. Field testing was conducted on research stations and producers fields using a lysimeter (on station) and soil water balance methods.

The first testing was done on a center pivot-irrigated corn field near Greeley, CO from 2010 to 2012. The dominant soil series in the field is Olney fine sandy loam (fine-loamy, mixed, superactive, mesic Ustollic Haplargids) (Soil Survey Staff, 2013). Soil bulk density in the corn field was determined using a Madera probe (Precision Machine Company Inc., Lincoln, NE) and methods described by Evett (2008). Soil bulk density was not a required input for the WISE system, but was used to convert measured gravimetric soil water content to volumetric values. Two samples were taken from each of two sampling points in the field at the end of the 2011 growing season at depths of 0-15, 15-30, 30-45, 45-60, 60-90, and 90-105 cm. The Permanent Wilting Point (PWP) for each soil layer was determined using a WP4-T Dewpoint PotentiaMeter (Decagon Devices, Inc., Pullman WA). A soil water release curve was created to obtain gravimetric water content at 1.5 MPa of tension. Volumetric water content at PWP was then calculated using methods described by Evett (2008). The field capacity (FC) for each soil layer was determined in situ by gravimetric sampling 24 hours after several deep irrigation/precipitation events in 2011. Values for soil properties used in the WISE system evaluation are summarized in Table 1.

Sampling Point Depth Bulk Density Field Capacity Wilting Point Available Water Capacity
(cm) (g cm-3) (cm3 cm-3) (cm3 cm-3) (cm3 cm-3)
0-15 1.16 0.33 0.17 0.16
15-30 1.13 0.27 0.17 0.10
North 30-45 1.33 0.28 0.14 0.14
Point 45-60 1.28 0.26 0.14 0.13
60-90 1.30 0.26 0.11 0.15
90-105 1.24 0.27 0.15 0.12
0-15 1.06 0.33 0.15 0.18
15-30 1.31 0.32 0.18 0.14
South 30-45 1.25 0.27 0.14 0.13
Point 45-60 1.22 0.23 0.13 0.10
60-90 1.40 0.25 0.12 0.13
90-105 1.28 0.28 0.16 0.12

The crop and irrigation system was maintained and managed by a cooperating grower who allowed weekly monitoring of soil water content and crop development in his field.  Actual gross irrigation amounts and dates were provided by the grower.  Irrigation application efficiency of the system was estimated to be 90%.  The crop coefficient (Kcr) curve used for corn was similar to curves reported by Allen et al. (2007), with adjustments made to fit the canopy development of corn grown in the 2010 growing season and days after planting (x-axis) converted to percentage of growing degree days (GDDs) to maturity.  Daily weather data for use in the WISE system were taken from the GLY04 (40.4487° N, 104.638° W, 1427 m above mean sea level) CoAgMet weather station located 5.2 km west of the corn field.

The performance of the irrigation scheduling tool in estimating current total soil water deficit (Dc) in the managed root zone was tested by comparing the scheduler’s calculated Dc to measured Dc on measurement dates (once a week during the growing season).  This evaluation method is similar to the one used by Jensen et al. (1971).  Measured Dc was derived from measured gravimetric soil water content of the managed root zone (Gleason, 2013). The management depth used in the evaluation of the irrigation scheduler was set at 1.05 m (41 in).  In each year, the tool was initialized with the first measured Dc at each sampling point.

In the 2013 growing season, the WISE system was also tested on four center pivot sprinkler-irrigated sugar beet fields in northeast Colorado.  This initial year of testing for sugar beets was primarily done to gather base line irrigation, crop development, and management information to improve upon a seasonal Kcr curve for sugar beets that was initially obtained from Allen et al. (2007).

Evaluation Results

An example of daily Dc for corn in 2010 is shown in Fig. 1. Performance statistics (calculated versus measured Dc on measurement dates) in 2010 – 2012 are given in Table 2. The calculated Dc from the WISE system was from a simplified daily soil water balance model. The WISE system was initialized with the measured Dc for the profile at the start of each growing season. No corrections were made to the daily calculated Dc throughout each growing season. Nevertheless, overall RE was relatively low at 13.6%. The RMSE, with the exception of 2012 at the north sampling point, was less than what could be applied by the center pivot in a single irrigation event (< 19 mm). The RMSE was deemed acceptable because just one irrigation application could easily compensate for it. The positive RE values (Table 2) mean that the magnitude of calculated Dc was overestimated, which indicated that the WISE system was estimating a drier soil profile than actual conditions. Calculated and measured Dc values were both represented as negative values (lack of water in the profile) in the calculation of MBE. Thus, the negative MBE values in Table 2 indicated that calculated Dc values were larger in magnitude than measured Dc values, consistent with the positive RE values. This bias would indicate the need for slightly more irrigation water than actually required, but would not increase the risk of water stress and yield reductions.

To demonstrate potential water savings through irrigation scheduling using the WISE system, Gleason (2013) compared actual (farmer-managed) gross irrigation amounts in 2011 for the above-mentioned corn field with hypothetical gross irrigation amounts that would result from scheduling irrigations based on Dc estimated by the WISE system. Precipitation in late May and the first half of June allowed the farmer to delay irrigations until 24 June 2011. Afterwards the center pivot ran on a fixed schedule, applying 13 to 17 mm of gross irrigation every 2 days. The center pivot was turned off for only 1 irrigation cycle in mid-July when significant rainfall occurred (15 mm on 11-13 July 2011). Gross irrigations were increased to 19 mm every 2 days beginning 28 July 2011 (reproductive growth phase) until 5 September 2011. An irrigation cycle was skipped only when significant rainfall occurred (28 mm on 2-4 August 2011) or was forecasted.

Click to expand Figure 1
Click to expand Figure 1
Site – Year Number of Measurements Root Mean Square Error (mm) Mean Bias Error (mm) Mean Absolute Error (mm) Relative Error (%)
North 2010 16 13.0 -0.3 10.6 1.8
South 2010 16 16.4 -1.5 13.1 8.6
North 2011 16 12.1 -1.8 10.8 11.3
South 2011 16 15.7 -1.6 12.6 11.3
North 2012 15 22.9 -12.9 18.0 30.9
South 2012 15 13.4 -2.9 10.8 6.5
All 94 15.9 -3.4 12.6 13.6

In contrast, hypothetical irrigations were applied only when the calculated Dc approached MAD, taking into consideration the system capacity of the center pivot. Calculated water balance components were compared during the period 13 June 2011 – 10 October 2011 (Table 3). The actual gross irrigation for the period was 511 mm. In contrast, the recommended (hypothetical) gross irrigation amount was 372 mm using the WISE system for irrigation scheduling. This represented a 27% (139 mm) reduction in gross irrigation – a significant savings in water and electrical energy (energy savings not calculated). The number of irrigation applications was also reduced from 30 down to 20, which would have reduced electricity costs for running the center pivot. Runoff and deep percolation calculated by the WISE system was 146 mm for the season with the actual irrigation amounts while it was reduced to 37 mm (75% reduction) using the WISE system for irrigation scheduling. The reduced water losses could have potentially reduced nutrient and chemical transport from the field – a benefit for water quality (actual water quality not measured). Following the recommended irrigations from the WISE system would have stored more water in the profile (∆S = 42 mm; 250% increase) compared to actual irrigations (∆S = 12 mm).

Water balance component With actual irrigations (mm) With recommended irrigations (mm)
ETc (mm) 501 501
Gross Irr (mm) 511 372
P (mm) 125 125
DP + SRO (mm) 146 37
Change in soil water storage in managed root zone (mm) 12 42

For the above comparison, corn grain yield would not have been reduced if recommended irrigations were applied because there was no reduction in ETc (Table 3). Corn grain yield has a linear response to cumulative ETc (Tolk et al., 1998), so the same ETc would result in the same yield. This study showed that for the 13 June 2011 – 10 October 2011 period, the corn crop was over-irrigated by approximately 139 mm.

Figure 2 shows an example graphical output of the WISE system for a sugar beet field in 2013. In contrast to Fig. 1, which shows soil water status in terms of water deficits, Fig. 2 shows soil water status in terms of plant-available water (PAW) in the root zone. The WISE system has the capability of presenting soil water status in either mode, depending on the preference of the user. Daily estimated PAW is shown relative to available water capacity (AWC) and management allowed depletion (MAD). Net irrigation, effective rainfall, and deep percolation (DP) (if any) are also shown on the graph. As a criterion for irrigation scheduling, the PAW is ideally kept in the region between AWC and MAD when deciding irrigation amount and timing. A second growing season is still needed to fully evaluate the effectiveness of the WISE system for scheduling irrigations for sugar beets.

More details about the WISE system, including factors that affect its accuracy, are given by Andales et al. (2014).

Created with GIMP
Click to expand Figure 2

References

Allen, R.G., J.L. Wright, W.O. Pruitt, L.S. Pereira, and M.E. Jensen. 2007. Water requirements. Chap. 8. In: G.J. Hoffman et al., editors, Design and Operation of Farm Irrigation Systems, 2nd ed. Am. Soc. of Agr. and Biol. Engrs, St. Joseph, MI, p. 208-288.

Andales, A.A., Bauder, T.A., and Arabi, M. 2014. A Mobile Irrigation Water Management System Using a Collaborative GIS and Weather Station Networks. In: Practical Applications of Agricultural System Models to Optimize the Use of Limited Water (Ahuja, L.R., Ma, L., Lascano, R.; Eds.), Advances in Agricultural Systems Modeling, Volume 5. ASA-CSSA-SSSA, Madison, Wisconsin, pp. 53-84.

Evett, S., 2008. Gravimetric and Volumetric Direct Measurements of Soil Water Content. In: S.R. Evett, et al., editors, Field Estimation of Soil Water Content: A Practical Guide to Methods, Instrumentation, and Sensor Technology. IAEA-TCS-30, 1018-5518. International Atomic Energy Agency, Vienna, Austria (2008), pp. 23–37 (Chapter 2). (Available online at http://www-pub.iaea.org/mtcd/publications/pdf/tcs-30_web.pdf; verified 29 May 2014).

Gleason, D.J. 2013. Evapotranspiration-based irrigation scheduling tools for use in eastern Colorado. M.S. Thesis. Colorado State University, Fort Collins, CO. 225 pp.

Jensen, M.E., J.L. Wright, and B.J. Pratt. 1971. Estimating soil moisture depletion from climate, crop, and soil data. Trans. ASAE 14:954-959.

Tolk, J.A., T.A. Howell, and S.R. Evett. 1998. Evapotranspiration and yield of corn grown on three high plains soils. Agron. J. 90:447-454.