Is there anything more exciting than watching grass grow? For a select few scientists among us, there is – watching moss grow. And in a world first, a team led by University of Wollongong (UOW) researchers, has developed a webcam (MossCam) and smart sensor system in Antarctica to remotely monitor moss beds, providing scientists with invaluable and continuous images and data about the Antarctic environment.
In December, a team from Securing Antarctica’s Environmental Future (SAEF) research program left Australia and headed south to study the Antarctic moss beds using drones, highly specialised sensor systems, artificial intelligence (AI) and machine learning. Two of those researchers included UOW Senior Research Fellow Dr Johan Barthélemy and UOW moss biologist Krystal Randall.
Krystal said mosses are important because they are the dominant vegetation in Antarctica.
“Mosses are commonly referred to as the forests of Antarctica because they provide habitat for much of Antarctica’s terrestrial biodiversity, such as tardigrades, fungi, and bacteria,” Krystal said.
“Mosses also play important roles such as taking carbon dioxide out of the atmosphere and insulating permafrost soils in ice-free areas of Antarctica. By monitoring the moss beds, scientists can understand and track the impact of climate change on Antarctica.”
Moss beds near Casey Station have been monitored over two decades and have shown changes in composition relating to ozone depletion and a general decline in health. Change in the Antarctic region creates ripple effects all over the planet.
Scientists at SAEF want to better understand the drivers of this declining health and to be able to monitor the moss beds remotely and predict future changes within these moss beds and others in Antarctica to inform policy and conservation strategies.
That’s where a special research team made up of Krystal, Johan and Douglas Henness came in. The UOW team co-designed an Artificial Intelligence of Things (AIoT) platform, which they set up in Antarctica to provide continuous data to moss biologists during their fieldtrip.
Each researcher brought valuable and different expertise to the project. In addition to his role as a Senior Research fellow at UOW, Johan is a Developer Relations Manager and AI specialist at NVIDIA. He specialises in smart technologies that use AI, machine learning and big data analysis. Krystal is a moss biologist who studies moss microclimates and Douglas is an electrical engineer.
All three experts co-designed the AIoT platform, while Douglas initially assembled and programmed it. Johan led the AIoT team and also played a role in programming the platform; building the dashboard; and designing and configuring the network infrastructure and data pipeline. Once the platform made it to Casey Station it was installed by Johan and Krystal, with Johan doing tests and debugging.
Several other researchers contributed to the project, including Distinguished Professor Sharon Robinson, Andrew Netherwood and Dr Mehrdad Amirghasemi.
Krystal’s role in the design of the platform was to ensure the platform was equipped to collect the most biologically relevant information that scientists need to monitor the moss beds as well as helping to deploy it in Antarctica.
“The platform we’ve developed will help moss biologists fill in the gaps between field seasons, where there is currently no data, to validate microclimate models, to recognise patterns in the data using AI, and to provide imagery that can help determine direct causality for changes in moss health,” Krystal said.
“This will be especially important for investigating the impact of extreme events such as heatwaves, which are increasing in frequency in Antarctica.”
The MossCam and smart sensors are attached to the AIoT platform, which can process data in-situ before transmitting it. The device then only transmits what is relevant for the end user to save bandwidth. The processing can rely on very simple algorithms such as transmitting data only if the temperature is above a given threshold, to state-of-the-art intelligence models such as image processing, intelligent video analytics, anomaly detection and noise processing.
The continuous monitoring of the moss beds will allow Krystal to develop a microclimate model for the area which will help scientists understand how healthy or unhealthy the mosses are, and identify areas that are at risk of becoming stressed in the future.
Johan said this season was about testing the components and sensors going inside the platform, as well as the LoRaWAN (long range, low bandwidth, low power) wireless data transmission protocol.
“LoRaWAN is free to use as it operates on a license-free radio spectrum. This makes it a very suitable for remote AIoT devices as they only transmit small amounts of data,” he said.
In addition to MossCam, the sensors to monitor the moss include a weather station capturing ambient air temperature and humidity; a thermocouple measuring the temperature of the moss canopy; a light intensity sensor measuring photosynthetically active radiation which is most important for plants; and a heat flux sensor to monitor the energy exchange between the soil and the near-surface air.
The platform was piloted during the 2022/23 summer season, with plans to take it back next season and leave it there for 12 months.
The researchers are now on their way back to Australia but have developed a public dashboard displaying historical data of their recent moss monitoring.
“We’d love people to have a look at what we’ve created online and get a better understanding of the importance of monitoring Antarctic mosses,” Johan said.
“It’s not every day you can watch moss grow!”
You can view the MossCam and data the smart sensors are collecting at: https://uow.to/mosscam
Kate is the University of Wollongong’s media and public relations specialists.