Sensor-Controlled Irrigation Scheduling for Pot-in-Pot Nursery Tree Production

Irrigation scheduling in nursery tree production is a complex process, due to the large number of tree species grown, differences in plant size, growth rate, and rooting volume (container size).  To avoid the risk of plant water stress, irrigation requirements are typically overestimated by nursery growers.  Sensor-controlled irrigation scheduling for two species of trees was independently compared to grower-scheduled irrigation in a pot-in-pot nursery from March–November 2012.  Sensor-based control was achieved using a prototype nR5 radio node (Decagon Devices, Inc.), in concert with advanced control software (Sensorweb; Carnegie-Mellon Robotics Institute, Pittsburgh, PA), to independently control irrigation events using solenoids.  One row within a block of dogwood (Cornus florida) and red maple (Acer rubrum) trees were retrofitted for independent control of irrigation with nR5-DC radio nodes, which allowed for direct control of a 12V-DC latching 25mm solenoid (Baccara; ML Irrigation Inc., Fountain Inn, SC).  Soil moisture sensors (10HS; Decagon Devices, Inc.) were inserted laterally at 15 cm depth in the root zone of 58 L containers (dogwood) and 116 L containers (red maple).  Ten trees of each species were monitored in this fashion; five trees in the grower-scheduled row and five in the nR5-controlled row, using a custom sensor calibration for the pine-bark substrate used to grow both species.  There were a total of 133 trees in each row for dogwood and 67 trees per row in the maple block.  Irrigation water was applied by the nR5-DC node whenever the average substrate moisture content of the five sensors fell below a pre-determined set-point (46% VWC for dogwood and 49% for red maple).  The grower-scheduled irrigation was typically 3 to 4 six-minute irrigations per day.   Irrigation water was applied to each tree using a single micro-sprinkler spray-stake (Netafim Corp., Fresno CA).  Irrigation application and leachate volumes were continuously recorded with flow meters (Badger Corp., WI) for each treatment.  For dogwood, the grower-scheduled row received 2.69 times more water than the control row (averaging 3.49 L vs. 1.30 L per tree per day, respectively); for maple, this was 1.45 times greater (6.20 L vs. 4.29 L per tree per day, respectively).  Tree height and diameter were measured for the five sensored trees in both rows every two months. Differences in plant height and tree diameter between treatments were not significant for both tree species after nine months, illustrating that sensor–controlled irrigation can achieve significant water savings in commercial production without affecting growth rates.