A real-time cost-effective wireless sensor network for monitoring peatlands with high spatial and temporal resolution

Abstract

Peatlands exist across the world and act as a vital carbon store. Comprehensive monitoring of peatlands is essential for tracking degradation and restoration but current techniques are limited by high cost and low spatial or temporal resolution. This research aims to overcome many of the limitations imposed by the current state-of-the-art environmental monitoring technologies by developing a resilient and modular multi-purpose wireless sensor network (WSN). As a demonstration of this new solution, factors affecting greenhouse gas emissions were also investigated.The developed WSN measured carbon dioxide and methane concentrations, air temperature, humidity, ambient light & soil moisture. The sensor nodes were solar-powered and automatically transmitted collected data via cellular networks to a cloud-hosted platform. The final iteration of the sensor nodes also implemented power-adaptive transmission scheduling, location tracking and orientation detection.To accommodate the power and size constraints of the sensor nodes, NGM2611-E13 methane sensor modules, designed for industrial applications (~5000 ppm), were characterised and adapted. Instead of being powered continuously, the sensors were briefly powered for each measurement period. A machine learning-based approach was devised and employed to calibrate the sensors at lower methane concentrations than the original sensor specifications. This work demonstrated that the response of the sensors to 0 -- 10 ppm methane was measurable and repeatable in this configuration. A low-cost co-calibration method for calibrating SCD30 CO2 sensors was also demonstrated.The performance of the WSN was assessed over 3 field deployments at peatland sites in the New Forest (Hampshire, England), CAFRE Hill Farm (Ballymena, Northern Ireland) and Whitelee Moor (Northumberland, England). For two of the sites, the WSN recorded the first known greenhouse gas metrics for each location. The prototype WSN deployed in the New Forest demonstrated that WSNs can reduce monitoring costs by up to 90%. Strong overall corroboration between the WSN and flux tower measurements was demonstrated in the CAFRE Hill Farm study. At the time of writing, the Whitelee Moor WSN has been in continuous operation for almost a year (September 2023--2024), demonstrating the long-term feasibility of the technique. The collected data highlighted diurnal and seasonal emission trends, both of which varied by location within the site. This final WSN deployment will extend beyond the duration of this PhD project and the collected data will be used by the Northumberland Peat Partnership to monitor the success of restoration activities.The presented sensor network shows potential as a new, low-cost means of carrying out detailed studies of peatland emissions and associated climate variables, with minimal human intervention and site disturbance. Future work should investigate ways to synthesise the data collected by the WSN with other monitoring techniques, such as remote sensing and flux towers.

More information

Identifiers

ORCID for Hazel Louise Mitchell: orcid.org/0000-0002-8461-2949

ORCID for Hugh Lewis: orcid.org/0000-0002-3946-8757

Catalogue record

Export record