A New Approach to Sap Flow Measurement Using 3D Printed Gauges and Open-source Electronics

Abstract:

A new type of sap flow gauge was developed to measure transpiration from herbaceous plants using a modified heat pulse technique. Gauges were fabricated using 3D-printing technology and low-cost electronics to keep the materials cost under $20 (U.S.) per sensor. Each gauge consisted of small-diameter needle probes fastened to a 3D-printed frame. One needle contained a resistance heater to provide a 6 to 8 second heat pulse while the other probes measured the resultant temperature increase at two distances from the heat source. The data acquisition system for the gauges was built from a low-cost Arduino microcontroller. The system read the gauges every 10 minutes and stored the results on a SD card. Different numerical techniques were evaluated for estimating sap velocity from the heat pulse data – including analytical solutions and parameter estimation approaches . Prototype gauges were tested in the greenhouse on containerized corn and sunflower. Sap velocities measured by the gauges were compared to independent gravimetric measurements of whole plant transpiration. Results showed the system could measure daily transpiration to within 3% of the gravimetric measurements. Excellent agreement was observed when two gauges were attached the same stem. Accuracy was not affected by rapidly changing transpiration rates observed under partly cloudy conditions. The gauge-based estimates of stem thermal properties suggested the system may also detect the onset of water stress. A field study showed the gauges could run for 1 to 2 weeks on a small battery pack. Sap flow measurements on multiple corn stems were scaled up by population to estimate field-scale transpiration. During full canopy cover, excellent agreement was observed between the scaled-up sap flow measurements and reference crop evapotranspiration calculated from weather data. Data also showed promise as a way to estimate real-time canopy resistance required for model verification and development. Given the low-cost, low-power, and open-source characteristics of the system; the technology is well suited for applications requiring large number of gauges (spatial scaling or treatment comparisons). While early work was done with agricultural crops, the approach is well suited for other species such as riverine shrubs.