Susan A. O’Shaughnessy

Integration of Wireless Sensor Networks into Moving Irrigation Systems for Automatic Irrigation Scheduling

A six-span center pivot irrigation system was used as a platform for testing two wireless sensor networks (WSN) of infrared thermometers. The cropped field was a semi-circle, divided into six pie-slice sections of which three were irrigated manually and three were irrigated automatically based on the time temperature threshold method. One network was mounted on masts fixed to the pivot arm (Pivot-WSN) and was programmed with mesh networking firmware. The second wireless network was comprised of sensors programmed with non-mesh firmware and was deployed in the field (Field-WSN). Our objectives were to: (1) compare the performance of a mesh and non-mesh networking systems of wireless sensors on a center pivot platform; (2) investigate the relationships between crop canopy, sensor body, and air temperatures; and (3) investigate automatic irrigation scheduling using data from wireless sensor networks.

Two-Source Energy Balance Model: Refinements and Lysimeter Tests in the Southern High Plains

A thermal two-source energy balance model (TSEB-N95) was evaluated for calculating daily evapotranspiration (ET) of corn, cotton, grain sorghum, and wheat in a semiarid, advective environment. Crop ET was measured with large, monolithic weighing lysimeters. The TSEB-N95 model solved the energy budget of soil and vegetation using a series resistance network, and one-time-of-day latent heat flux calculations were scaled to daily ET using the ASCE Standardized Reference ET equation for a short crop. The TSEB-N95 model included several refinements, including a geometric method to account for the nonrandom spatial distribution of vegetation for row crops with partial canopy cover, where crop rows were modeled as elliptical hedgerows. This geometric approach was compared to the more commonly used, semi-empirical clumping index approach. Both approaches resulted in similar ET calculations, but the elliptical hedgerow approach performed slightly better. Using the clumping index, root mean squared error, mean absolute error, and mean bias error were 1.0 (22%), 0.79 (17%), and 0.093 (2.0%) mm d-1, respectively, between measured and calculated daily ET for all crops, where percentages were of the measured mean ET (4.62 mm d-1). Using the elliptical hedgerow, root mean squared error, mean absolute error, and mean bias error were 0.86 (19%), 0.69 (15%), and 0.17 (3.6%) mm d-1, respectively, between measured and calculated daily ET for all crops. The refinements to TSEB-N95 will improve the accuracy of remote sensing-based ET maps, which is imperative for water resource management.

Single- and Dual-Surface Iterative Energy Balance Solutions for Reference ET

The concept of a reference evapotranspiration (ETr) calculated from daily or hourly weather data, and multiplied by a crop coefficient (Kc) in order to estimate crop water use (ETc), is widely established in agricultural science and engineering. To find region- and variety-specific values of Kc from field-measured ETc values, the equation is inverted to: Kc = ETc/ETr. Forms of the Penman-Monteith (PM) formula for calculation of reference alfalfa or grass evapotranspiration (ETr and ETo, respectively) were promulgated by ASCE in 1990, FAO in 1998, and ASCE in 2005. The PM formulations are sensitive to climatic conditions, producing estimates of ETr and ETo that are more or less close to measured values depending on regional climate, and yielding values of Kc that vary from region to region and so are not transferrable. Theoretical shortcomings may be the basis of some of these problems, including the explicit nature of the calculation, which relies on the implied assumption that canopy and air temperatures are equal. We examined the ETr estimation of two surface energy balance formulations that stipulated different air and canopy temperatures: a two-layer (soil and canopy) approach, and a one-layer (big leaf) approach that included soil heat flux. Since canopy temperature is implicit in these formulations, they must be solved iteratively. Iterative solutions of ETr were compared with the ASCE PM formulation and against lysimeter-measured ETr. All three methods of ETr estimation produced ET values that compared very well with field-measured ET for alfalfa grown under reference ET conditions. Errors may occur with any of the three approaches to ETr estimation when stomatal resistance changes due to weather conditions; thus, assumptions of constant daytime and nighttime surface resistances cause mis-estimation of surface energy fluxes. It appears that a surface resistance value of 200 s m-1 at night for alfalfa grown under reference ET conditions is too large. It also appears that assuming constant daytime surface resistance of 30 s m-1 is probably not ideal, and that presenting daytime surface resistance as a function of vapor pressure deficit might improve ETr calculation.

Using an integrated crop water stress index for irrigation scheduling of two corn hybrids in a semi-arid region

Different thermal-based plant feedback systems have been used for irrigation management of cotton and grain crops in the Texas High Plains region, producing yields that are similar or better than irrigation scheduling using the neutron probe. However, there are limited studies using plant feedback systems to actively scheduling irrigations for corn. In this 2-year study, a drought tolerant and a conventional hybrid were managed under a variable rate center pivot irrigation system. The main treatments were manual and plant feedback irrigation scheduling based on weekly neutron probe readings and an integrated crop water stress index (CWSI), respectively. In each main treatment, three irrigation treatment levels were established. Crop responses were compared between irrigation methods and levels. Results demonstrated that overall grain and biomass yields and grain WUE for the plant feedback-control plots were similar to those from the manual-control plots for both years. These results indicate that a plant feedback system using a CWSI could be used to manage corn in a semi-arid region and over a large-sized field. The plant feedback system could provide convenience and time savings to farmers who manage multiple center pivot fields.

Automatic Irrigation Scheduling of Grain Sorghum Using a CWSI and Time Threshold

The crop water stress index (CWSI) has been investigated extensively as a quantification of plant water stress and a threshold to time irrigations. Most studies have calculated this thermal-based index by taking instantaneous measurements or by averaging values over a short period of time, usually near midday. Although useful for quantifying stress, measurements over a short period of time have not been particularly stable. This study was conducted to compare the yield response and water use efficiencies of a short-season hybrid grain sorghum obtained from automatic irrigations triggered by a CWSI and time threshold versus those obtained from manual best-practice irrigation scheduling. Irrigation amounts of 80%, 55%, 30% and 0% of full irrigation were applied. Grain yields across the 80% and 0% treatments were not significantly different between manual and automatic irrigation control methods. The only significant difference in water use efficiency between irrigation control methods occurred in the 30% treatment plots where the largest variability in the initial soil water profile existed. Irrigation water use efficiency was similar between irrigation methods in the 55% and 30% treatments, but significantly greater in the 80% automatically controlled treatments at p = 0.05. These results indicate that the CWSI and time threshold index has the potential to be utilized as a tool for deficit irrigation scheduling of grain sorghum. Further research is required to investigate its stability over different growing seasons.

Performance of a Wireless Sensor Network for Crop Water Monitoring and Irrigation Control

Robust automatic irrigation scheduling has been demonstrated using wired sensors and sensor network systems with subsurface drip and moving irrigation systems. However, there are limited studies that report on crop yield and water use efficiency resulting from the use of wireless networks to automatically schedule and control irrigations. In this study, a multinode wireless sensor network (WSN) system was mounted onto a six-span center pivot outfitted with a commercial variable rate irrigation (VRI) system. Data from the WSN was used for automatic irrigation scheduling and irrigation control to produce an early hybrid variety of grain sorghum in 2011. An integrated crop water stress index (CWSI) was used as a threshold to schedule irrigations. Half of the center pivot field was divided into six sectors, three were irrigated using automatic control, and three were irrigated based on weekly direct soil water measurements. Wireless sensor nodes, i.e. infrared thermometers, GPS unit, and multiband radiometers were integrated onto the center pivot system and field below. The WSN system was scaled to 40 different nodes and was operational throughout 98% of the growing season. An assessment of the reliability of the nodes, demonstrated that delivery rates for data packets from the different nodes ranged between 90% to 98%. Automatic irrigation scheduling succeeded in producing mean dry grain yields and controlling crop water use efficiency (WUE) at levels that were similar to those from soil water based irrigation scheduling. Average seasonal integrated crop water stress indices were negatively correlated to irrigation treatment amounts in both the manual and automatic plots and correlated well to crop water use. These results demonstrate that it is feasible to use WSN systems for irrigation management on a field scale level.

Single- and Dual-Surface Implicit Energy Balance Solutions for Reference ET

The concept of a reference evapotranspiration (ETr) calculated from daily or hourly weather data, multiplied by a crop coefficient, Kc, in order to estimate crop water use, ETc, is widely established in agricultural science and engineering. To find region and variety-specific values of Kc from field-measured ETc values, the equation is inverted to: Kc = ETc/ETr. Forms of the Penman-Monteith (PM) formula for calculation of reference alfalfa or grass evapotranpsiration (ETr and ETo, respectively), have been promulgated by ASCE in 1990, FAO in 1998 and ASCE in 2005. The PM formulations are sensitive to climatic conditions, producing estimates of ETr and ETo that are more or less close to measured values depending on regional climate, and yielding values of Kc that vary from region to region and so are not transferrable. Theoretical shortcomings may be the basis of some of these problems, including the explicit nature of the calculation, which relies on the implied assumption that canopy and air temperatures are equal. We tested two surface energy balance formulations that stipulated different air and canopy temperatures, one a two-layer (soil and canopy) and one a one-layer (big leaf) approach but with soil heat flux included. Since canopy temperature is implicit in these formulations, they must be solved iteratively. Iterative solutions of ETr were compared with the FAO and ASCE PM formulations and against lysimeter-measured ETr. All three methods of ETr estimation produced ET values that compared very well with field-measured ET for alfalfa grown under reference ET conditions. Errors may occur with any of the three approaches to ETr estimation when stomatal resistance changes due to weather conditions; and assumptions of constant daytime and nighttime surface resistances thus cause mis-estimation of surface energy fluxes. It appears that a surface resistance value of 200 s m-1 at night for alfalfa grown under reference ET conditions is too large. It also appears that assuming constant daytime surface resistance of 30 s m-1 is probably not ideal, and that presenting daytime surface resistance as a function of vapor pressure deficit might improve ETr prediction.

Two Source Energy Balance Model – Refinements and Lysimeter Tests in the Southern High Plains

A thermal two-source energy balance model (TSM) was evaluated for predicting daily evapotranspiration (ET) of corn, cotton, grain sorghum, and wheat in a semiarid, advective environment. Crop ET was measured with large, monolithic weighing lysimeters. The TSM solved the energy budget of soil and vegetation using a series resistance network, and one-time-of-day latent heat flux estimates were scaled to daily ET using the ASCE Standardized Reference ET equation for a short crop. The TSM included several refinements, including a geometric method to account for the nonrandom spatial distribution of vegetation for row crops with partial canopy cover, where crop rows were modeled as elliptical hedgerows. This geometric approach was compared to the more commonly used, semi-empirical clumping index approach. Both approaches resulted in similar ET estimates, but the elliptical hedgerow approach performed slightly better. Using the clumping index, root mean squared error, mean absolute error, and mean bias error were 1.0 (22%), 0.79 (17%), and 0.093 (2.0%) mm d-1, respectively, between measured and modeled daily ET for all crops, where percentages were of the observed mean ET (4.62 mm d-1). Using the elliptical hedgerow, root mean squared error, mean absolute error, and mean bias error were 0.86 (19%), 0.69 (15%), and 0.17 (3.6%) mm d-1, respectively, between measured and modeled daily ET for all crops. The TSM refinements will improve the accuracy of remote sensing-based ET maps, which is imperative for water resource management.