Terrestrial plant ecosystems face tremendous pressure and stress due to climate change and anthropogenic activities, which alters plant function and productivity at different spatial and temporal scales [1]. A better understanding of such dynamics is a prerequisite for accurate monitoring and prediction of the global carbon cycle and securing food production. The integration of the earth observation data and ground-based measurements are highly applicable for rapid analysis of plant productivity and health. However, traditional reflectance based remotely sensed vegetation indices such as Normalized Differential Vegetation Indices (NDVI) are not sensitive to capture the short-term or diurnal changes in plant functioning [2], [3].

Solar-induced chlorophyll fluorescence (SIF), as a promising and novel product, has become an attractive method for estimating hourly plant productivity and for detecting pre-visual plant stress [4]–[6]. The inherent linkage of vegetation photosynthesis and SIF makes it possible to monitor crop yield and detect plant stress [7]–[10]. However, interpretation of the SIF signal at the canopy scale is still challenging due to the impact of plant canopy structure, especially between diverse forest and crop types, as well as different growth stages.

Synthetic Aperture Radar (SAR) is a mature remote sensing technology which can form high-resolution images of the earth’s surface. This project seeks to examine advanced SAR-based surface and underwater object detection and classification. Underwater objects cannot be directly detected via SAR since electromagnetic radiation is reflected and absorbed by the ocean, however SAR can detect surface currents and ocean swell to indirectly perform bathymetry. These techniques therefore potentially provide an avenue for detecting an underwater object. The detection and classification of surface vessels can also be potentially improved by exploiting ocean phenomenon, such as wakes, in addition to using advanced beamforming approaches. This PhD project seeks to explore these main lines of effort before narrowing the scope in the initial stages of the project.

The proliferation of surface and underwater vessels in the Asia Pacific region is a concern to the Australian Government [1]. Underwater objects are challenging to detect, particularly on a global scale. The detection of underwater objects via remote sensing would undoubtedly be beneficial, particularly using Synthetic Aperture Radar (SAR), which is a day-and-night all-weather remote sensing technology. Increases in global shipping has also increased the number of different types of ships navigating the world’s oceans. Detection and classification of these vessels is an important problem for governments around the world, particularly for addressing piracy and illegal fishing. This project seeks to explore advanced detection and classification of surface and sub-surfaces vessels using imaging radar phenomenologies and exploitation approaches.

Australia’s inland waterways, reservoirs and coastal waters provide significant benefits to the ecosystem, local economy, public health, and food security (Boyd, 2020). Currently, securing water resources and their quality has been regarded as a critical consideration due to its increasing demand with socioeconomic and population growth, shrinking freshwater resources, and aquatic ecosystem degradation (SmartSat, 2021). In this sense, substantial improvements have been achieved to safeguard water resources through various management strategies, plans and regulatory arrangements. For instance, the United Nations Sustainable Development Goal 6 (SDG 6) highlights the importance of water quality and access to clean, safe, and secure water supplies are fundamental for attaining sustainable development by ensuring availability and sustainable management of water at the regional and global level (GEO – Group on Earth Observations, 2018). However, the quality of water resources has been impaired to some degree due to various natural and anthropogenic disturbances such as excessive use of pesticides, harmful chemical substances, soil erosion, organic wastes, and heavy metals from the industry (Briffa, Sinagra, & Blundell, 2020; Issaka & Ashraf, 2017). Therefore, there is an urgent need to continuously monitor the quality of Australian inland and coastal waters by harnessing innovative approaches.

In the last few decades, “Digital Earth” has been widely used as a strategic platform to support national and international cooperation towards reaching sustainable development goals (SDGs). Along with being a global strategic contributor to sustainable development, the Digital Earth is being regarded as a vital approach for addressing the environmental, social, economic and cultural challenges that affect human lives, their nations and the planet Earth, allowing humankind to visualise the Earth, to access information about it and to understand above issues (Dhu et al., 2017; Mazlan, Samsudin, & Yin Chai, 2014). Guo, Goodchild, and Annoni (2020) described the Digital Earth as the combination of massive, multi-resolution, multi-temporal, and multi-typed Earth Observation (EO) and appropriate smart analytical algorithms. The Earth Observation (EO), one of the integral parts of Digital Earth, includes next-generation remote sensing satellites and unmanned aerial vehicles (UAVs) and has become a significant part of environmental sustainability to tackle current and emerging challenges, including climate change, natural resource depletion, water insecurity, and environmental degradation (Alvarez-Vanhard, Corpetti, & Houet, 2021). In this respect, Digital Earth Australia (DEA) was established to provide insights into Australia’s evolving land, coast, and water issues by utilizing EO data and other geospatial data (Dhu et al., 2017).

With the technological advances in computational power and internet capabilities, unprecedented opportunities have emerged. For example, cutting-edge sensor technology evolved the way of monitoring environmental, social, and economic challenges. One of the highly developed sensor technologies is the Internet of Things (IoT), which enables interconnecting Things based on existing and evolving interoperable information and communication technologies at various scales (Guo et al., 2020). In other words, it is a network of infrastructure in which objects equipped with computing capabilities can communicate directly with each other and collect and transmit data to central servers (ITU, 2012; Tzounis et al., 2017). The IoT contains real-time in situ water quality data with high temporal and spatial resolution, capturing dynamic flow characteristics, pollution events, and water quality extremes (Chowdury et al., 2019). Furthermore, the low cost and ease of deploying IoT sensors throughout the area of interest significantly reduce data collection barriers while increasing data transparency.

Integrating Digital Earth and derived products with ground-based, high accuracy & frequency IoT sensor networks in the domain of water quality monitoring is a promising option to determine quality parameters, understand its variability and inform evaluations of water quality prediction. Furthermore, this collaboration can provide an integrative and quantitative water source management necessary for evidence-based decision-making (Loucks & van Beek, 2017). In this sense, SmartSat CRC is implementing a pilot project named “Satcom IoT-enabled automatic groundwater collection and aggregation pilot (SIG WATER)”, which is dedicated to checking the technical feasibility, reliability, and cost-effectiveness of deploying IoT sensors as an end-to-end solution for groundwater bores. Furthermore, another project titled “Next-generation testbed design for Earth observation” aims to enhance the right level of “trust” for EO data and derived products by calibration and validation. In the context of our research, we would like to integrate a state-of-the-art ground-based IoT network with quality-assured, uncertainty quantified EO data for water quality monitoring along the Australian west coastline through the abundance of partnerships with organisations and industry cooperation.

Tracking landscape function with the capacity to identify degradation versus regeneration is fundamental to sustainable land management. Spatial changes in soil biophysical state could be monitored by identifying the parameters that contribute to resilience and functional integrity. Soil health is the foundation of landscape function and is paramount to the sustainable management of savannas and grasslands across Northern Australia that occupy approximately two million km2 (Williams et al., 2005). Climatic extremes are increasingly changing these landscapes which require adaptation in land management strategies to deal with more frequent droughts, floods, and fires. Hence, there is an incentive for accurate and rapid monitoring to quantify spatial changes from soil to pastures and higher vegetation.

This research will apply a range of sensing platforms to characterise ground cover dynamics, analysing signal variations before and after fire, drought, floods and the impacts of grazing. To enhance capabilities and provide detailed analysis, Unmanned Aerial Vehicle (UAV) and a hand-held field spectrophotometer will be used to study the landscape function and soil health. Potentially, the phototrophic fabric called biocrusts that inhabit the soil surfaces in the northern savannas could act as a soil health indicator, by means of its biodiversity and its microbial communities’ capacity to act as ecosystem engineers (Eldridge & Leys, 2003). Both UAV and handheld devices are adept to quantifying chlorophyll in biocrusts nondestructively by detecting specific wavelength bands across the entire electromagnetic hyperspectral range (Gitelson, Gritz, & Merzlyak, 2003). Such spectrophotometry would enable a clearer understanding of the nuances of ‘bare ground’ by mapping the distribution of chlorophyll at a field level and across landscapes both temporally and spatially to recognize the role of biocrusts as a surrogate measure for soil health.

Various studies have shown the utilisation of hyperspectral measurements accurately determine the chlorophyll concentration of plants and biocrusts at several scales (Román et al., 2019; Stephens, Louchard, Reid, & Maffione, 2003).
Furthermore, digital colour (RGB, red, green blue) image analysis is another non-destructive method that can accurately measure chlorophyll content (greenness), a plant health Third Party Project Agreement | SmartSat CRC | P3-32s Page 4
indicator, and nutrient levels. The advent of fixed camera and high-performance computing, over the past decade, has facilitated the use of high-resolution visible photos (Red, Green, Blue) and digital technologies (including machine learning) more readily.

In this research, I will work towards linking proximal and remote sensing to provide soil health management tools. The project will be based on 30-years of fire research at Victoria River Research Station (VRRS), Northern Territory, and the impacts of grazing management regimes (established 24 years) at Wambiana Station (WGT) in North Queensland. This research will enable the connection of soil health to management practices. In a progressively variable climate monitoring ecosystem interactions is critical to guiding land management. Furthermore, sensing technologies have the capacity to enhance insight and decision making to maintain resilience and ongoing landscape function.

This project responds to a request from Defence Science and Technology Group (DSTG) to develop a visualisation of SmartSat CRC research activities to provide context for Defence capability managers. This initial activity will draw on conceptual work to show how Indo-Pacific Connector will deliver maritime domain awareness through space-based sensors and advanced communication technologies.

This project will research a novel deep learning pipeline for achieving robust and reliable forecasting of natural disaster events with hyperspectral satellite imagery. The majority of the existing machine learning algorithms do not possess the ability to forecast the occurrence of natural disasters in advance and they are only able to detect their occurrence when the disaster event happens. To address this limitation, we propose novel research which leverages the spatiotemporal modelling capability of deep learning to forecast the occurrence of natural disasters in advance using a satellite onboard execution environment. The research will be conducted using the example use cases from the application area of bushfire event forecasting considering the practical significance to Australia. The developed framework will be deployable on a CubeSat Kanyini-type mission with Hyperscout-2 payload and will use onboard hardware to execute the algorithm.