Enabling proactive agricultural drainage reuse for improved water quality through collaborative networks and low-complexity data-driven modelling

Abstract

With increasing prevalence of Wireless Sensor Networks (WSNs) in agriculture and hydrology, there exists an opportunity for providing a technologically viable solution for the conservation of already scarce fresh water resources. In this thesis, a novel framework is proposed for enabling a proactive management of agricultural drainage and nutrient losses at farm scale where complex models are replaced by in-situ sensing, communication and low complexity predictive models suited to an autonomous operation. This is achieved through the development of the proposed Water Quality Management using Collaborative Monitoring (WQMCM) framework that combines local farm-scale WSNs through an information sharing mechanism.

Under the proposed WQMCM framework, various functional modules are developed to demonstrate the overall mechanism: (1) neighbour learning and linking, (2) low-complexity predictive models for drainage dynamics, (3) low-complexity predictive model for nitrate losses, and (4) decision support model for drainage and nitrate reusability. The predictive models for drainage dynamics and nitrate losses are developed by abstracting model complexity from the traditional models (National Resource Conservation Method (NRCS) and De-Nitrification-DeComposition (DNDC) model respectively). Machine learning algorithms such as M5 decision tree, multiple linear regression, artificial neural networks, C4.5, and Naïve Bayes are used in this thesis. For the predictive models, validation is performed using 12-month long event dataset from a sub-catchment in Ireland.

Overall, the following contributions are achieved: (1) framework architecture and implementation for WQMCM for a networked catchment, (2) model development for low-complexity drainage discharge dynamics and nitrate losses by reducing number of model parameters to less than 50%, (3) validation of the predictive models for drainage and nitrate losses using M5 tree algorithm and measured catchment data. Additionally modelling results are compared with existing models and further tested with using other learning algorithms, and (4) development of a decision support model, based on Naïve Bayes algorithm, for suggesting reusability of drainage and nitrate losses.

More information

Identifiers

ORCID for Nick Harris: orcid.org/0000-0003-4122-2219

Catalogue record

Export record