Milestone climate agreements like the Kyoto Protocol and The Paris Agreement rely heavily on data from Earth observation (EO) technologies to monitor and track progress toward their ambitious goals. From astronauts manually capturing photographs of the Earth from satellites in the '60s, we have come a long way in mapping our planet not just visually, but on different superspectral bands that tell us about our Earth’s health (such as surface temperature, vegetation indices, cirrus cloud detection, etc.) using remote sensing and in-situ data methods from an array of sensors and sources.
In making accurate models and predictions of the Earth’s complex climate ecosystem, robust data of the Earth is therefore necessary. Sensor tools are rapidly evolving to capture higher fidelity data, but data by itself has no value. By 2032, the cumulative EO data taken is estimated to be 2 exabytes (that's enough to stream 57,000 years of HD video non-stop!) and manually parsing through that much data to make actionable insights is no longer tenable.
Why AI?
We utilise AI to bridge the gap to unlock critical information that would guide policy and decision-making, not just in commercial carbon projects, but in saving human lives from climate-related disaster preparedness and/or mitigation.
To work through the petabyte-scale datasets to track offset projects, traditional models take an average of 8 months to create and train. Using AI and proprietary algorithms, our workflow has successfully reduced that time to just 1 month, including data visualisation and recommending actionable insights for our clients. The full potential of AI and ML to speed up the process is up to 1,000 times faster than traditional models while significantly reducing computational demands and remaining precise.
What’s the science behind it?
How exactly does AI reduce processing times by that many orders of magnitude? Geospatial AI foundational models are trained on various data types in a self-supervised way, learning patterns from raw untagged data, enabling the models to understand atmospheric dynamics. They can create highly accurate models of global patterns while remaining computationally efficient by incorporating advanced ML methods to process and normalise data.
Don’t just take our word for it. Here are cases of AI being used by others to greatly improve our collective climate action:
ML models were used to parse through EO data after Hurricane Ian, resulting in rapid evaluation and $80 million in expedited assistance for disaster survivors. In comparison, manually assessing geospatial damage takes 30 seconds per structure, which means evaluating 4.1 million structures after the Hurricane Ian disaster would take one person 16 years.
Developed by Microsoft, Aurora is a cutting-edge AI foundational model that can forecast global levels of six key air pollutants – CO, NO2, NOx, SO2, O3 and particulate matter – in under a minute. Aurora is set to generate high-resolution weather forecasts for up to ten days, surpassing both traditional simulation tools and specialised deep-learning models. According to its developers, Aurora is about 5,000 times faster than state-of-the-art numerical integrated forecasting systems.
The use of EO data in ML-based and foundational models can decrease the build time of a flood map by as much as 80%, helping to convert data into more precise predictions of floods, and EO data was proven to increase the accuracy of predicting flood susceptibility by up to 20%. In the case of flash flooding, a 12-hour lead time can potentially reduce damage by up to 60%. In comparison, just a one-hour advance lead time can reduce damage by 20%.
What’s next?
Finally, AI can make climate insights more accessible, which is critically important when communicating to decision-makers, stakeholders, and impacted community members. It’s fast coming in: intuitive climate dashboards and AI chatbots at the front capable of answering, near instantaneously, user questions after combing through the dataset, à la the ChatGPT everyone’s familiar with.
As satellite and sensor technology becomes better and easier to deploy, we are rapidly growing the arsenal of tools to diagnose our planet earth faster and come up with better, more timely solutions to mitigate climate change.
Nika.eco specialises in harnessing AI and Earth observation data to inform impactful climate actions. Whether you’re involved in offset projects or disaster management, our solutions drastically reduce delivery times and improve accuracy. Get in touch with our team to learn how our AI-powered models can give your project the edge in today’s climate challenges.
Comments