Bayesian Belief Network analysis of soil salinity in a peri-urban agricultural field irrigated with recycled water

• Bayesian Belief Network (BBN) model was used to analyze the soil salinization.
• Recycled water (RW) irrigation increased the soil salinization.
• Soil salinization was found sensitive to moisture content and RW salinity.
• The BBN model allows back propagation to keep RW salinity in sustainable limit.

Continuous use of recycled water (treated wastewater) over a long period of time may lead to the accumulation of salt in the rootzone soil. This is due to the relatively higher levels of salt content in the recycled water compared to surface water. In this study, an assessment framework is proposed to evaluate the level of salinity in recycled water that can have significant impact on the rootzone salinity when recycled water is used for irrigating peri-urban agricultural paddocks. The framework is constructed with a probabilistic expert system, more specifically, Bayesian Belief Network (BBN). The BBN model analyses the salt accumulation process with quantifying uncertainty associated with various variables related to this process. The proposed BBN was first developed and tested with the results from laboratory batch study and continuous column study conducted over 264 days. Later, a salt transport model HYDRUS 1D was used to quantify salt accumulation in the paddock over 20 years of study period, and the result was used to update the BBN to accommodate field condition. The salt transport modelling identified that in some year rootzone soil water electrical conductivity (ECSW in dS/m) was more than twice of the maximum threshold of salinity tolerance. The BBN was used to quantify the uncertainty associated with the reduction of salinity in terms of electrical conductivity in recycled water (ECrw in dS/m), which subsequently reduced the probability of exceeding the salt accumulation beyond the maximum threshold limit. It was found that, if ECrw can be reduced by 13% (from 0.92 to 0.8 dS/m), there is around 49% probability that the ECSW would be reduced by around 39% (from 6.5 dS/m to 4 dS/m), and will keep it within the threshold salinity limit of 3.0–5.0 dS/m 100% of the time. The study highlighted that any strategies that help in the reduction of salt in the recycled water will be beneficial in managing the soil salinity as a result of recycled water use for irrigating peri-urban agricultural field. The methodology presented in the study provides much needed knowledge for the development of robust and enlightened management strategies in relation to wastewater reuse.