(610d) Impact of Sensor Placement in Soil Water Estimation of Agro-Hydrological Systems: A Real-Case Study

Freshwater scarcity is becoming a serious issue worldwide primarily due to population growth, climate change, and increasing pollution. Of the total amount of freshwater, about 70% is consumed in the agricultural activities, with the main consumer being irrigation [1]. Currently, the water-use efficiency in irrigation is estimated to be between 50% to 60% due to poor irrigation strategies [2]. In order to mitigate the freshwater supply crisis, the water-use efficiency in agriculture irrigation needs to be improved, through precision irrigation and implementing a closed-loop irrigation system [3].

One important element in implementing the closed-loop irrigation system is obtaining the soil moisture of the entire field which is required for feedback control. Since the agriculture fields usually are of large scale, installing sensors in the whole field is impractical. Soil moisture estimation techniques are required to reconstruct the full soil moisture information based on a small number of sensors. Because of the limited number of sensors, the determination of the best locations to install the sensors such that good state estimation can be obtained is an important problem. In our recent work [4], this issue has been addressed by employing the modal degree of observability based on extensive simulations. It was found that optimally placed sensors can lead to much improved soil moisture estimation performance. However, it is unclear whether the significantly improved estimation performance can still be observed in actual applications.

In this work, we consider an actual agricultural field in Lethbridge, Alberta, Canada, and study the impact of sensor placement in soil water estimation. Soil moisture measurements from 36 soil moisture sensors installed at different depths were collected for one growing season. First, a three-dimensional agro-hydrological model with heterogeneous soil parameters of the research farm is developed. Then, a state estimator designed based on the extended Kalman filter (EKF) is adopted to estimate the soil water content. Subsequently, we apply the modal degree of observability to the three-dimensional system and determine where the best sensor locations. Estimation results based on different measurements are obtained and analyzed. In this presentation, we will share our findings on the impact of sensor placement in soil moisture estimation.

References:

[1] T. Sauer, P. Havlik, U. A. Schneider, E.Schmid, G. Kindermann, M. Obersteiner, “Agriculture and resource availability in a changing world: The role of irrigation” Water Resourc. Res.,46 (2010).

[2] “Water in a changing world,” The United Nations World Water Development Report 3, (2017).

[3] S. L. Shah, B. R. Bakshi, J. Liu, C. Georgakis*, B. Chachuat, R. D. Braatz, and B. R. Young. Meeting the challenge of water sustainability: the role of process systems engineering, AIChE Journal, 67:e17113, 2021.

[4] S. R. Sahoo, X. Yin, and J. Liu, “Optimal sensor placement for agro-hydrological systems” AIChE Journal. vol. 65, no. 12, (2019).