EOSDA Forest Monitoring: An Interview On The Update

What Has Changed In The Updated Version Of EOSDA Forest Monitoring?

Kostiantyn Chernyshevskyi: First and foremost, now this platform is customizable to address specific customer requests more effectively. If a client of ours is interested in forest monitoring, we first learn their needs and expectations, set up an instance of the platform with the required features enabled for their areas of interest, and then provide this client with access to it. This way, we ensure the analytics our platform presents relies on various factors, including the use of accurate neural networks, allowing us to provide insights with approximately 90% accuracy (depending on the region).

Apart from that, we’ve made several significant enhancements to improve user experience and functionality. While maintaining the same user-friendly patterns, we’ve streamlined the interface, added more relevant data, and removed any information that could potentially create noise or confusion for our users.

From a functionality standpoint, we’ve expanded our solution offerings within the platform. Previously, we had four solutions: Deforestation, Forest Cover, Forest Productivity, and Water Stress. Now, we’ve improved our Forest Cover and Deforestation solutions with a new technology that guarantees getting satellite data analytics of any area of interest at least once a month. This technology uses various imagery including satellite radar images, making it independent of cloud-free optical images, which is crucial for regions with high cloud cover.

Our Deforestation technology now provides data on both deforestation and forest cover, while the Forest Cover technology focuses solely on forest cover. This differentiation caters to the needs of different types of users. Some need to understand where logging is occurring, while others need to know the overall forest cover, especially during the initial stages of purchasing forest companies or investing in forests.

We’ve also added new solutions to the platform. Our Forest Health solution is based on NDVI (Normalized Difference Vegetation Index), comparing the current index with the average index over the past five years to determine changes in forest health. We’ve also introduced Burned Area, a solution for assessing fire damage, which is crucial for planning recovery or land-use changes after large-scale forest fires. Lastly, our newly developed technology for Tree species identification, initially tested in Ukraine, can recognize up to ten species within a similar climate zone. In general, this solution has the ability to identify any dominant species within a specified region if sufficient relevant ground data about it is available.

Our Reforestation solution has also been improved. We can now detect changes in forest cover in terms of tree growth, which is important for understanding the progress of large-scale reforestation efforts. Satellite data can provide a cost-effective and faster alternative to ground expeditions for monitoring these areas.

In summary, the updated EOSDA Forest Monitoring platform offers more comprehensive, user-friendly, and effective solutions for sustainable forest management.

Vera Petryk: Let me also add that the importance of new Forest Health and Burned Area solutions within EOSDA Forest Monitoring cannot be overstated, especially in today’s context.

It was found that forest fires result in 3 million more hectares of tree cover loss per year compared to 2001 ^{Tyukavina A, Potapov P, Hansen MC, Pickens AH, Stehman SV, Turubanova S, Parker D, Zalles V, Lima A, Kommareddy I, Song X-P, Wang L and Harris N (2022) Global Trends of Forest Loss Due to Fire From 2001 to 2019. Front. Remote Sens. 3:825190. https://doi.org/10.3389/frsen.2022.825190.}. In 2023 alone, wildfires in Canada have burned a staggering 25 million acres, as reported by the New York Times ^{Nadja Popovich (July 18, 2023). How Canada’s Record Wildfires Got So Bad, So Fast. New York Times.}. In July, Greece has been fighting dozens of devastating forest fires during the longest hot spell the country has ever seen ^{Thousands flee Greek island fires as heatwave continues around the world. (Jul 23, 2022). Le Monde with AP and AFP.}.

The Burned Area solution is thus vital for tracking and assessing the impact of forest fires: by providing accurate, real-time data on the extent of fire damage, this tool can guide effective recovery efforts and help mitigate future fire risks.

Reforestation, on the other hand, is an effective method to combat climate change while also maintaining the many benefits forests provide. For instance, India plans to increase its forest cover from 25% to 33% and restore 26 million hectares of land by 2030, as part of its ambitious climate goals ^{Nicole Schwab, Bhaivrani Jani (Mar 23, 2022). Here's why forest restoration is key to India's ambitious climate goals. World Economic Forum.}.

The reforestation feature in EOSDA Forest Monitoring can play a crucial role in such initiatives by providing accurate and timely data on forest cover and health, enabling better planning and implementation of reforestation efforts. By leveraging satellite analytics, we can ensure that our efforts to restore and protect our forests are based on solid data, ultimately helping us make the Earth a better place for future generations.

In the face of escalating threats to our forests, we’ve been designing these innovative solutions as essential instruments in our collective effort to safeguard the health of our planet and ensure a sustainable future for generations to come.

And That’s How The Updated Version Of The Product Aligns With EOSDA’s Commitment To Sustainability?

Vera Petryk: At EOSDA, our dedication to sustainability is deeply ingrained in our product development, guided by our alignment with 10 of the 17 United Nations’ Sustainable Development Goals (SDGs). Our forest monitoring platform embodies this commitment, providing tools to combat illegal logging — a major contributor to global greenhouse gas emissions — and address climate change.

The platform allows users to monitor forest health from space, using special indices to analyze historical and current data. This enables us to identify illegal logging, track reforestation efforts, and provide concrete evidence of forest conditions. In doing so, we’re not just observing the problem; we’re equipping users with the tools to make a positive impact.

As we relaunch EOSDA Forest Monitoring to a wider audience, I believe it will play a significant role in promoting sustainable forest management and combating climate change.

Vera, What Exactly Do You Mean By A Wider Audience?

Vera Petryk: EOSDA Forest Monitoring’s updated version is designed to cater to a wide range of audiences, given the global importance of forest conservation. The primary users are not individuals, but businesses, governments, and decision-makers who have the capacity to implement substantial changes. This includes:

1. Landowners and Forest Managers: They can utilize the system for effective forest management planning, monitoring forest health, identifying tree types, and tracking illegal deforestation.
2. Financial Institutions and Investors: They can use the system to assess the suitability of land for loans or insurance, minimizing risks by understanding the land’s history and potential profitability.
3. Food Producers: They can track the progress of forest landscape restoration efforts and generate zero deforestation reports based on satellite data, which is crucial for carbon trading.

The system is designed to be adaptable, capable of addressing specific problems or needs, making it customer-tailored. While the main geographical focus is North America, South America, Africa, Australia, and Europe, the system’s relevance extends to anyone living on Earth due to the interconnected nature of environmental challenges.

Kostiantyn, Tell Us More About The Technologies Fueling Forest Monitoring To Serve So Many Different Users?

Kostiantyn Chernyshevskyi: The product leverages two distinct approaches for data provision.
The first approach is applied when the client’s territory is not excessively large, and they require regular forest monitoring. For such cases, we employ neural networks to model the forest locations, which are then verified by satellite imagery and our team of GIS experts. This method allows us to review the entire territory and make necessary adjustments, resulting in an accuracy of around 90% depending on the region.

The second approach is semi-automated and is used when dealing with very large territories, such as those spanning three to seven million hectares. Here, we use larger neural networks and other technologies like land cover classification to determine forest locations. This is followed by the application of our neural network to verify whether the area is forested or not. The results are then validated by a team comprising over 60 in-house scientists and GIS experts. This approach yields an accuracy of over 80% as physically verifying such large territories is challenging.

In both approaches, the use of neural networks and satellite imagery is crucial. Neural networks help model and predict forestation patterns, while satellite imagery provides the necessary data for verification and validation. This combination of technologies ensures reliable and accurate forest monitoring, which is critical for sustainable forest management.

Vera, Given That EOSDA Forest Monitoring Is Tailored To Each User, Does This Imply That Potential Users Need To Have A Clear Understanding Of The Specific Issues They Want To Address Before Reaching Out To Us?

Vera Petryk: While it can be beneficial for potential users of EOSDA Forest Monitoring to have a specific problem in mind, it’s not a requirement. The system is designed to provide a comprehensive overview of a forested area, which can be particularly useful for users who have recently acquired land and need to understand its condition.

For instance, the system can identify tree types, detect burnt areas, assess forest health, and provide historical data. This information can help users identify potential issues and make informed decisions, even if they didn’t have a specific problem in mind when they first contacted us.

Kostiantyn, Vera Touched On The Forest Health Feature. Could You Delve Deeper Into What Exactly ‘Forest Health’ Encompasses And How This Particular Feature Functions?

Kostiantyn Chernyshevskyi: The Forest Health feature primarily evaluates the vegetation of a forest, as trees are the main components of a forest and their health can be determined by their vegetation. If a tree’s vegetation is normal, the tree is healthy. If something is wrong, such as pest infestation or fungal growth, the vegetation will deteriorate, indicating the tree is not healthy.

We have developed a technology that collects data on NDVI over the past few years. We can collect data for comparison of medium-term changes, medium-long-term changes, and long-term changes. We saw that medium-long-term changes cover most of the client’s needs. By comparing the average NDVI value over several years with its current value, we can conclude what is happening with the forest: either its health is improving, i.e., it is growing and developing, and everything is okay with it, or the health is deteriorating, i.e., some trees are drying up, etc.

However, we calculate forest health exclusively within the vegetation season, i.e., from the moment the leaves bloom until the leaves fall. Therefore, there will be no issues of false results, especially in comparison with winter. If we talk about, for example, a fire or something else, yes, we also take this into account as a deterioration of the forest’s health. Because if you’re burned, you have no health.

Currently, this solution provides a basic understanding of changes in forest health. The reasons why these changes occurred will be our next step in the development of this technology.

Vera, Maintaining The Health Of Our Forests Is Usually A Responsibility Of The Government. Could You Elaborate On How EOSDA Forest Monitoring Can Assist In Ensuring Compliance With Local Environmental Laws?

Vera Petryk: It can play an important role in ensuring adherence to a country’s laws related to logging and deforestation. The tool can monitor land and provide precise quantities of logging that can be done in a certain area. This allows businesses and individuals to monitor their activities and ensure they are in compliance with the law.

However, it’s important to note that while Forest Monitoring can provide valuable data, it is not responsible for interpreting or enforcing the law. The responsibility to understand and adhere to local laws lies with the user. If a user from a new country approaches us, it’s their responsibility to inform us about their local laws and how they want to use our platform to comply with them.

While we are primarily a monitoring service and not directly connected to law enforcement, we do have a responsibility towards the environment. If we were informed of an intention to cause significant environmental harm, such as turning a forest into a desert, we would likely notify the authorities.

Kostiantyn, Vera, As A Final Treat, Can You Share Some Exciting New Features We Can Anticipate In The Upcoming Updates Of EOSDA Forest Monitoring?

Kostiantyn Chernyshevskyi: The features currently in development are Biomass Estimation and Carbon Stock. They are designed to provide valuable insights into the health and potential of a forest. These features primarily focus on the carbon stored in trees, which is directly related to their biomass. Using satellite data, machine learning, and ground data, we can estimate the amount of biomass in a 100×100 meter pixel with relatively low error. This biomass can then be easily converted into carbon stock, giving us an estimate of the carbon stored at a given point in time.

While this technology won’t allow for daily monitoring of carbon changes, snapshots every few months or annually can provide significant insights. This can be particularly useful for those who want to evaluate their forest’s carbon stock as a first step towards transitioning from a business model that profits from logging to one that profits from preserving trees and earning carbon credits.

Additionally, this technology can be a valuable tool for investors or buyers interested in forests. Combined with our reforestation technology, it can provide a more comprehensive picture of a forest’s potential. For instance, young trees, which absorb more carbon during their growth phase, may have a small carbon stock initially. However, knowing the species planted, we can estimate how much more carbon they will absorb until they reach maturity, and consequently, how much profit can be made.

In essence, using the Carbon Stock, Tree Species, and Reforestation technologies together can provide a more detailed and accurate picture of a forest than using the Carbon Stock technology alone. All our technologies are designed to work together to provide a more comprehensive view of the forest.

Vera Petryk: The future updates of EOSDA Forest Monitoring will also leverage the capabilities of seven optical satellites that EOS Data Analytics plans to launch into Low Earth Orbit. While the first spacecraft is already operational in the sky, together they will form the EOS SAT constellation that will significantly enhance the amount and resolution of the images we use for analysis, allowing us to move from the current 10-meter resolution of Sentinel-2 to a much finer 1.4-meter panchromatic and 2.8-meter multispectral resolution of our own sensors. This will enable us to provide more precise information on deforestation, forest cover, and burnt areas.

In addition to improved resolution, we are also contemplating the development of a neural network that can operate globally. This ambitious project would require training the network on data from all tree species, climatic zones, and land features – a substantial task indeed. Although this global network is merely a prospective vision at this stage, we remain optimistic. As we acquire clients from different parts of the world, we anticipate the accuracy of our neural network for those regions will gradually enhance. Eventually, this could pave the way to a neural network effective across the entire planet.

Such a technology will be not just about better data or insights—it’ll be about fostering a deeper connection between humanity and the natural world. By harnessing the power of satellite analytics, we at EOSDA are committed to helping make Earth a better place for future generations, and we believe that by taking better care of our forests today, we are investing in a greener, more sustainable future for all.