Opinion: Earth intelligence’s role in managing climate change risk

In parts of Australia, we’ve had more than our fair share of catastrophic floods and fires in the last few years, and the impact of El Niño on weather patterns - particularly in the eastern states – is likely to continue until April 2024.

That means a very hot Australian summer in a year that’s already on track to be the hottest on record.

With the World Economic Forum ranking natural disasters and extreme weather events as the second most severe risk facing the world over the next two years (the cost-of-living crisis was ranked first), the rapid technological innovation in “Earth intelligence” offers great potential in managing and mitigating this risk.

Earth intelligence is a term that describes the gathering of all kinds of data about the planet and turning that data into actionable information. Data from satellites has been used for more than half a century, not least involving military satellites and spy planes collecting photographs for examination by teams of experts to understand the capabilities of enemy nations.

However, Earth intelligence has become increasingly important for commercial and environmental applications – not just for mitigating climate change risks, but also for Australia’s agricultural and mining sectors. Gartner predicts 80 percent of major managed land assets globally will be under active satellite monitoring by 2028, compared with 10 percent in 2022.

Top business value use cases include environmental monitoring, weather forecasting and bushfire mitigation.

An ecosystem of emerging technologies for Earth intelligence is redefining and enabling new insights. These include smarter drones and new satellite constellations combined with advanced analytics that can deliver business intelligence at a far lower cost compared with traditional methods, such as commercial aviation and manual analysis.

New spectral techniques and sensors
From space, using wavelengths outside the visible light spectrum can reveal Earth observation insights not previously available. New satellite constellations and aerial systems are being fitted with sensors capable of collecting data employing different wavelengths and data collection techniques, using computer vision to automate the analysis of the data.

Hyperspectral imaging exploits hundreds or potentially thousands of wavelengths enabling analysts to detect different types of objects, events and materials in the Earth or under the sea.

This is achieved by identifying the spectral signatures of various materials using advanced classification techniques and validation of training data. Applications include plant health monitoring, pollution detection, emissions monitoring and soil moisture content.

Multispectral imaging captures images using two or more wavelengths that go beyond the visible light spectrum, such as infrared and ultraviolet. It can be used for activities such as weather forecasting, forest fire detection, and bushfire monitoring and risk management.

Synthetic aperture radar (SAR) is also being used to spot very small changes in surface features, often before they are visible on the ground or in the ocean and for detecting changes in surface features independent of weather or environmental conditions.

Recently, the Australian Government selected Finnish SAR satellite company ICEYE to supply flood and bushfire hazard data nationally.

New data analytics platforms
Data analytics platforms are starting to unlock greater business value of Earth observation for land management by expanding access to more data, improving pre-processing efficiencies and extending the range of analytical opportunities.

This, in turn, enables users to build observational applications for crop yield forecasting; asset protection via flood and fire prediction; spectral geology to identify mineral deposits; and insurance claims processing and validation if the damage is already done.

These platforms support the business value of Earth intelligence by aggregating data from different providers and delivering it cheaply to users. Additionally, many vendors are offering off-the-shelf models for data analysis, as well as sandboxes for developers to apply their own analytics.

Monitoring is one potential application of these platforms for land management in Australia to ensure landholders are complying with land clearing legislation, or that logging companies are meeting the requirements of their timber harvesting licences.

Overcoming adoption challenges
There are still some key challenges that need to be overcome, however, to ensure widespread adoption and acceptance of Earth intelligence solutions for land management. In particular, the inadequate training of deep learning models and ground truth verification.

Ground truth verification is important in qualifying the accuracy of Earth intelligence techniques and increasing user confidence. Advancements in AI are leading to greater accuracy in data analysis and improvements in ground truth verification.

The other primary adoption barriers are the lack of user education and communicating the growing range and richness of satellite data, along with sensor and data fusion techniques that enable new insights and unlock business value.

Earth intelligence is rapidly becoming an essential tool for governments and enterprise operations. Over the next decade, the quality, cadence and analysis of Earth observation data will deliver unprecedented insights.

At the same time, the economies of scale and increased competitive pressure will reduce the cost of incorporating Earth intelligence into business operations.

Nick Ingelbrecht is a senior director analyst with Gartner's Technology and Service Provider Research organisation, focused on computer vision, emerging trends and technologies, video and image analytics and physical security.