Project Name: Developing Automated Technology to Assess Natural Capital on Pastoral Leases
Aim: To develop automated technology for the purpose of assessing natural capital on pastoral leases in the Southern Rangelands of WA.
Project Duration: October 2022 - July 2023
Funding body: This project is funded by the Australian Government’s Agricultural Innovation Hubs Program, with support from the Grower Group Alliance and the South-West WA Drought Resilience Adoption and Innovation Hub.
Aim: To develop automated technology for the purpose of assessing natural capital on pastoral leases in the Southern Rangelands of WA.
Project Duration: October 2022 - July 2023
Funding body: This project is funded by the Australian Government’s Agricultural Innovation Hubs Program, with support from the Grower Group Alliance and the South-West WA Drought Resilience Adoption and Innovation Hub.
Executive Summary:
The project aimed to develop automated technologies for the objective assessment and measurements of the natural capital elements of biodiversity, coupled with weather and soil attributes, to minimise and/or remove the current human observation requirement which has significant impacts on resources, time and costs.
It included the digitisation of historical paper records and leveraged publicly available data repositories, and piloted the use of readily available devices to capture measurements and data related to natural capital elements in the Southern Rangelands.
Four pilot sites were established for granular on ground measurements. All sites had a weather station and soil probes to generate weather, and soil related data into the AxisStream platform. Three of the pilot sites also had network (IP) monitoring cameras to capture project site images.
Exploration of landscape measurements such as satellite imagery, NDVI and drone data was also completed to assess the possibility of broader landscape measurement data.
The establishment of monitoring cameras at project pilot sites also enabled the testing of Machine Learning (ML) tools for object identification.
It included the digitisation of historical paper records and leveraged publicly available data repositories, and piloted the use of readily available devices to capture measurements and data related to natural capital elements in the Southern Rangelands.
Four pilot sites were established for granular on ground measurements. All sites had a weather station and soil probes to generate weather, and soil related data into the AxisStream platform. Three of the pilot sites also had network (IP) monitoring cameras to capture project site images.
Exploration of landscape measurements such as satellite imagery, NDVI and drone data was also completed to assess the possibility of broader landscape measurement data.
The establishment of monitoring cameras at project pilot sites also enabled the testing of Machine Learning (ML) tools for object identification.
Expected Outcomes:
This project will demonstrate the accuracy and ease-of-use that automated devices can provide for the new marketable product – natural capital – which has been historically degraded in the Southern Rangelands. Degradation needs to be urgently addressed and improved through better monitoring and management.
The automation and digitising of objective measurements, management and processing of the natural capital elements of biodiversity and soil organic carbon information by establishing in-field methodologies with readily available devices and data. Automated devices remove the “human error” or “subjectivity” element which can skew records.
This project has the potential to set the standard for the emerging Natural Capital market in the Southern Rangelands. It will serve as a pilot that can be replicated and extended across other properties in the future.
This project will enable high-level conversations with key financial institutions and stakeholders around the inclusion of natural capital on financial statements.
The automation and digitising of objective measurements, management and processing of the natural capital elements of biodiversity and soil organic carbon information by establishing in-field methodologies with readily available devices and data. Automated devices remove the “human error” or “subjectivity” element which can skew records.
This project has the potential to set the standard for the emerging Natural Capital market in the Southern Rangelands. It will serve as a pilot that can be replicated and extended across other properties in the future.
This project will enable high-level conversations with key financial institutions and stakeholders around the inclusion of natural capital on financial statements.
Methodology:
1. Digitisation of historical records and images from Challa Station.
2. Camera connectivity assessment for Challa Station sites.
3. Project site establishment and IoT Installation/Deployment, and research available imagery for Challa Station.
4. Compile list of flora species.
5. Development of Machine Learning for detection of flora and fauna.
2. Camera connectivity assessment for Challa Station sites.
3. Project site establishment and IoT Installation/Deployment, and research available imagery for Challa Station.
4. Compile list of flora species.
5. Development of Machine Learning for detection of flora and fauna.
Results and Conclusions:
- This innovative project has demonstrated the possibilities of using automated technology to carry out what is usually a time-consuming and highly subjective task.
- It has shown that technology can be utilised to provide permanent (or semi-permanent) monitoring in the Southern Rangelands.
- However, it has also shown the current limitations of some technologies depending on their connectivity requirements.
- With continual improvements in connectivity services to remote areas this will be overcome in the future.
- This project has trialled the various equipment needed for monitoring of various elements of natural capital.
- Following the presentation at the SRPA Pastoral Industry Forum held in 2022 and the ongoing display at the same forum in 2023, along with updates provided by Debbie Dowden of Challa Station, there has been considerable interest from pastoralists who are keen to explore the future natural capital market.
- This project has also formed the basis of other projects and funding applications in which the aim is now to expand upon the learnings achieved over the past 18 months.
- The flora and fauna data collection during this project will be shared into citizen science platforms (currently in progress). This contribution to public and open databases will improve the overall data available for the rangelands region, by providing recently generated images and data. This additional data can be utilised by other pastoralists to improve their knowledgebase of the region, while encouraging their own participation to such platforms. We expect this outcome to also benefit the broader community, while also assisting environmentalists, biologists, wildlife rangers and conservation agencies in future research and studies.
Overall, the application of ML in object detection and identification at Challa Station represents a promising avenue for enhancing ecological monitoring and conservation strategies.
With ongoing data collection, model refinement, and collaborative efforts, this technology can contribute significantly to the preservation of the unique ecosystem of Challa Station.
Vegetation/Landscape Classification using Technology
Custom satellite camera solution installation and deployment
Equipment Installation
Updates
We are delighted to provide you with an update on a significant SRPA initiative that is reshaping the assessment of natural capital on pastoral leases. The first stage of the project is centred around biodiversity assessment and has been made possible through an Innovation grant secured by SRPA from the SW WA Drought Resilience Adoption and Innovation Hub.
At the core of this project are four pilot sites located at Challa. These sites are equipped with sophisticated tools including weather stations, soil moisture probes, and soil organic carbon probes. Additionally, most sites are outfitted with remote cameras that play a crucial role in observation. These sites not only capture visual data but also collect pertinent information about weather and soil conditions. The amassed data is transmitted to the cloud and integrated into the AxisStream platform. This platform is the engine behind the data analysis, where complex algorithms extract meaningful insights.
Over the course of more than a year, this endeavour has garnered national attention. However, it's important to acknowledge the challenges we've encountered along the way. Connectivity emerged as a critical factor, particularly in projects entailing technology and cameras. We've learned that thorough connectivity mapping and assessment are prerequisites for successful device implementation.
Our efforts in connectivity exploration led us to experiment with LTE-M (Long Term Evolution for Machines), a technology designed for IoT connectivity. Unfortunately, its performance fell short of expectations, even within Telstra's service coverage area. Point-to-point wifi solutions were also tested, yielding moderate success within limited ranges. Notably, in our most remote site, we opted for a Starlink connection powered by solar energy due to its economic viability and efficiency.
In parallel with technological advancements, it's worth highlighting the role of human involvement. Developing automated technology relies heavily on human input to curate and structure data for Machine Learning tools. We extend our appreciation to the team at AxisTech for their dedication to this substantial undertaking.
This project holds significant implications, particularly in positioning the Southern Rangelands as a leader in sustainable agriculture. Every step taken contributes to a repository of knowledge that will inform future progress. As we move forward, we remain committed to refining our understanding and practices, ensuring a more resilient and ecologically aware agricultural landscape.
Stay tuned for more updates as we continue to navigate the frontiers of innovation and environmental stewardship.
Image: The Starlink connection is installed at a remote location on Challa by the AxisTech team
At the core of this project are four pilot sites located at Challa. These sites are equipped with sophisticated tools including weather stations, soil moisture probes, and soil organic carbon probes. Additionally, most sites are outfitted with remote cameras that play a crucial role in observation. These sites not only capture visual data but also collect pertinent information about weather and soil conditions. The amassed data is transmitted to the cloud and integrated into the AxisStream platform. This platform is the engine behind the data analysis, where complex algorithms extract meaningful insights.
Over the course of more than a year, this endeavour has garnered national attention. However, it's important to acknowledge the challenges we've encountered along the way. Connectivity emerged as a critical factor, particularly in projects entailing technology and cameras. We've learned that thorough connectivity mapping and assessment are prerequisites for successful device implementation.
Our efforts in connectivity exploration led us to experiment with LTE-M (Long Term Evolution for Machines), a technology designed for IoT connectivity. Unfortunately, its performance fell short of expectations, even within Telstra's service coverage area. Point-to-point wifi solutions were also tested, yielding moderate success within limited ranges. Notably, in our most remote site, we opted for a Starlink connection powered by solar energy due to its economic viability and efficiency.
In parallel with technological advancements, it's worth highlighting the role of human involvement. Developing automated technology relies heavily on human input to curate and structure data for Machine Learning tools. We extend our appreciation to the team at AxisTech for their dedication to this substantial undertaking.
This project holds significant implications, particularly in positioning the Southern Rangelands as a leader in sustainable agriculture. Every step taken contributes to a repository of knowledge that will inform future progress. As we move forward, we remain committed to refining our understanding and practices, ensuring a more resilient and ecologically aware agricultural landscape.
Stay tuned for more updates as we continue to navigate the frontiers of innovation and environmental stewardship.
Image: The Starlink connection is installed at a remote location on Challa by the AxisTech team
When this project to revolutionise biodiversity measurement in the Southern Rangelands was first conceived, we knew the nature repair bill was poised to introduced into federal parliament. The bill created a framework for a voluntary biodiversity market that will operate alongside carbon markets. It aims to halt the decline of our natural environment by attracting funds from private investment.
Exactly how the biodiversity certification scheme could operate has not yet been determined, but we wanted to get on the front foot and not only create interest in the biodiversity markets in the Southern Rangelands, but also to try to solve the problem of measuring biodiversity in such a huge, remote landscape.
We have seen some prototype websites that endeavour to measure biodiversity using satellites and we believe that there is a place for satellite technology, but it must work in conjunction with on- ground monitoring. We know that any data collected needs to be robust enough to satisfy the market, and collection of that data needs to be cost and time effective. We also wanted to build a monitoring system that wouldn’t be too expensive, and so we partnered with Axistech to build it. Axistech supplied readily available off the shelf products and put them together to build our fixed position biodiversity monitoring sites. Each site was equipped with a camera, a weather station, a soil moisture probe, and a soil carbon probe. The data we collected was interpreted by either humans or machines, depending on what we were looking at and what machine learning technology was available.
Initially we thought that we could create a machine learning app that would identify plants. I now believe that there will be better and more accurate ways to do that, and we will keep working on solutions. We also thought that the machine learning could measure the growth rate of plants after rain. However, this was our driest year in the past thirty, and the lack of rain has meant that the plants have become dormant rather than grown. Hopefully good rains are coming soon that will allow us to measure some plant growth.
What surprised me most about the project was the creatures that we identified that live out on these randomly chosen spots on the station. Our cameras picked up the big animals like cattle and kangaroos, but also birds, lizards and even insects. We have discovered that there are two different dunnart species that live within the field of vision of the camera at our southernmost site. They share the space with an abundance of geckos, spiders and birds.
At one point, we thought we may have discovered a species of small mammal thought to be extinct on the mainland, but the tantalising images were not clear enough to be sure. Nevertheless, they hinted at the exciting potential of this technology in the rangelands.
In the past when I have put trail cams out to capture photos of wildlife, I have placed them at watering points. But our Axistech monitoring sites were installed well away from watering points and we were able to see the little desert animals that can survive with little water. We know there are hundreds of species of animals and birds sharing our landscape in the rangelands. Being able to spy on them from our office is fascinating, and once the monitoring site is in place, we don’t disturb the animals at all. Capturing the images is relatively simple and storing them for future use in our cloud-based platform means that if we ever need this type of data to demonstrate our biodiversity to the market, we have it on hand.
This current project has come to an end and the final report has been submitted to the Southwest Drought Hub. We are already building the next iteration with improved design and ideas, and we hope that our work today forms the basis for a future biodiversity certification that may be monetised through private investment and shared across the region.
Below (Clockwise): Weather Station Device Report in AxisStream, Soil Probe Report AxisStream, ML Vegetation Landscape - 2, ML Animal Identification - 2
Exactly how the biodiversity certification scheme could operate has not yet been determined, but we wanted to get on the front foot and not only create interest in the biodiversity markets in the Southern Rangelands, but also to try to solve the problem of measuring biodiversity in such a huge, remote landscape.
We have seen some prototype websites that endeavour to measure biodiversity using satellites and we believe that there is a place for satellite technology, but it must work in conjunction with on- ground monitoring. We know that any data collected needs to be robust enough to satisfy the market, and collection of that data needs to be cost and time effective. We also wanted to build a monitoring system that wouldn’t be too expensive, and so we partnered with Axistech to build it. Axistech supplied readily available off the shelf products and put them together to build our fixed position biodiversity monitoring sites. Each site was equipped with a camera, a weather station, a soil moisture probe, and a soil carbon probe. The data we collected was interpreted by either humans or machines, depending on what we were looking at and what machine learning technology was available.
Initially we thought that we could create a machine learning app that would identify plants. I now believe that there will be better and more accurate ways to do that, and we will keep working on solutions. We also thought that the machine learning could measure the growth rate of plants after rain. However, this was our driest year in the past thirty, and the lack of rain has meant that the plants have become dormant rather than grown. Hopefully good rains are coming soon that will allow us to measure some plant growth.
What surprised me most about the project was the creatures that we identified that live out on these randomly chosen spots on the station. Our cameras picked up the big animals like cattle and kangaroos, but also birds, lizards and even insects. We have discovered that there are two different dunnart species that live within the field of vision of the camera at our southernmost site. They share the space with an abundance of geckos, spiders and birds.
At one point, we thought we may have discovered a species of small mammal thought to be extinct on the mainland, but the tantalising images were not clear enough to be sure. Nevertheless, they hinted at the exciting potential of this technology in the rangelands.
In the past when I have put trail cams out to capture photos of wildlife, I have placed them at watering points. But our Axistech monitoring sites were installed well away from watering points and we were able to see the little desert animals that can survive with little water. We know there are hundreds of species of animals and birds sharing our landscape in the rangelands. Being able to spy on them from our office is fascinating, and once the monitoring site is in place, we don’t disturb the animals at all. Capturing the images is relatively simple and storing them for future use in our cloud-based platform means that if we ever need this type of data to demonstrate our biodiversity to the market, we have it on hand.
This current project has come to an end and the final report has been submitted to the Southwest Drought Hub. We are already building the next iteration with improved design and ideas, and we hope that our work today forms the basis for a future biodiversity certification that may be monetised through private investment and shared across the region.
Below (Clockwise): Weather Station Device Report in AxisStream, Soil Probe Report AxisStream, ML Vegetation Landscape - 2, ML Animal Identification - 2
