Embedded Network Sensing of Moisture and Nitrate Propagation During Irrigation with Reclaimed Wastewater

Center for Embedded Networked Sensing Embedded Network Sensing of Moisture and Nitrate Propagation During Irrigation with Reclaimed Wastewater J. Eric Haux 1 , Thomas C. Harmon 1 , Jose Saez 2 , Juyoul Kim 3 , Yeonjeong Park 3 , Naim D. Busek 4 , Thomas Schoellhammer 4 , Deborah Estrin 4 University of California, Engineering Division, Merced, CA; 2 Loyola Marymount University, Civil and Environmental Engineering, Los Angeles, CA; University of California, Civil and Environmental Engineering, Los Angeles, CA; 4 University of California, Department of Computer Science, Los Angeles, CA Overview : A real-world application of sensor networks in soil and irrigation with wastewater A wireless sensor network monitors soil moisture and nitrates in an agricultural plot irrigated with reclaimed wastewater. An embedded sensor network that operates in situ over large scales in time and space is advantageous as it increases data acquisition rates, distributes the computation and analysis process, improves the modeling and prediction of nitrates, and reduces human intervention -- all in real-time. Once established, the automated network can optimally control nitrate and moisture inputs. The objective of this project is to systematically develop sensor networks and design a control system to monitor and respond to nitrate propagation in soil being irrigated with reclaimed wastewater. Problem Description: Simulation and Control model development using sensor networks Simulation and Control Models Description of the transport and fate of nitrate in the unsaturated zone requires flow, temperature, and nitrate simulation models. Soil is homogenous fine to medium sand; Flow expected to be mainly vertical (one dimension). Irrigation must be scheduled with simulation models and data feedback from the pylons to control pivot operation, thus optimizing discharge of nitrate-laden wastewater. rain gauge D ata acquisition a nd wireless t ransmission gr ound surface 30-50 cm 150-200 cm Thermistor Soil moisture Nitrate 250-300 cm PVC pipe Sensor Network System A Multi-level sensing station (pylon) for temperature, moisture, and nitrate was deployed at a test site in Palmdale, CA. The pylon measures temperature, moisture, and nitrate in its one-dimensional setting. Pylons will communicate with other nearby pylons to delineate nitrate concentration distribution in time and space. Data is captured at a base-station and relayed to a database. Proposed Solution: System characterization and optimization of irrigation scheduling 1D simulation results Volumetric water content (cm3/cm3) Irrigation Control Management step: 20 Optimal value Volumetric water content (cm3/cm3) 1D model Depth (cm) Concentration (ppm) 3m Depth (cm) Water content at the top node Management step Total application rate Depth (cm) Darcian water flux (cm/hr) Application rate (cm/hr) • Pylon installation (above). •The irrigation pivot (right). Management step Management step Time (hr) Depth (cm) • Flow and nitrate simulation is performed with 1D Richards equation and 1D advection- dispersion equation using a finite difference method. • Application rate is assigned as sine function to imitate the irrigation pattern when the pivot passes. • The simulation will be extended to 2D to examine settings in which horizontal flow is significant. Parameter estimation • Parameter estimation for the simulation model is an important part of sensor network calibration. • Soil samples are collected to measure hydraulic conductivity and moisture retention and their spatial variability. • Each pylon provides real-time sensor data for local network node calibration. • Deterministic and geostatistical algorithms for scaling up toward sensor network error resiliency will then be tested. • The objective of irrigation control is to determine the application rate such that wastewater usage is maximized and the nitrate regulatory level is not violated. • The control scheme (measurement, decision, and action) is executed by using the on-line data feedback from the pylons and providing control to the watering pivot. UCLA – UCR – Caltech – USC – CSU – JPL – UC Merced