BLOG POST
Unlocking the Power of Satellites for Nature-Based Solutions
Last edited: November 8, 2023
Published: November 8, 2023

Orbify Team
Earth Intelligence Specialists
The need for effective and reliable tools for assessing and monitoring the environment has never been greater. Remote Sensing (RS) through satellite technology is emerging as a key solution for addressing challenges related to origination, verification, and due diligence in environmental projects. In this article, we will explore the benefits, limitations, and potential applications of satellite data, as well as the role of machine learning and deep learning in this field.
Why we need Remote Sensing (RS)
Origination, verification, and due diligence in environmental projects can be a daunting task. How can we be certain that a project is viable and environmentally beneficial? The project registration process is often expensive and time-consuming, limiting the accessibility of nature-based solutions. To address this, we turn to remote sensing as a tool to increase trust in these projects.
Trust: Remote sensing offers verifiable transparency, allowing us to track and measure a project's real, measurable, additional, and permanent impact. It also helps in monitoring whether projects are achieving their targets.
Changes in Methodology: In a changing world, where science and environmental conditions evolve, remote sensing provides a means to assess risks more accurately. It enables the production of high-quality carbon credits and aids in biodiversity preservation.
Benefits of using RS data
The use of remote sensing data offers a wide range of benefits:
- Data Availability: Remote sensing provides wide-scale coverage, making it an invaluable resource for environmental assessments. It's non-intrusive, mostly free, and offers access to various tools. Additionally, it provides data with a high frequency and historical context, aiding in long-term assessments.
- Increasing Data Quality: Advances in remote sensing methodologies, such as Machine Learning (ML) and Deep Learning (DL), have enhanced the quality of data available. ML and DL techniques help in the analysis of data, making it more reliable.
While remote sensing is a powerful tool, it's important to recognize its limitations. Fully automated project assessments using remote sensing data and models may never be a reality, as land use issues are complex and interconnected across various realms. It is crucial not to oversimplify the problems but to create tools that help us find the answers.
Passive Satellites for Forest Monitoring
Several passive satellites are used for forest monitoring:

- MODIS (replaced by VIIRS): These satellites offer spatial resolutions ranging from 250 to 1000 meters and provide data every 1 to 2 days across 36 bands, including visible, near-infrared, shortwave infrared, thermal infrared, and microwave.

- Landsat 7 & 8: These satellites provide spatial resolutions ranging from 15 to 60 meters with an 8-day revisit frequency.

- Sentinel 2: With spatial resolutions ranging from 10 to 60 meters and a 5-day revisit frequency, Sentinel 2 offers data in 8 to 11 bands, including visible, near-infrared, shortwave infrared, and thermal infrared regions.

Vegetation Indices and Spectral Indices
Vegetation indices play a critical role in utilizing RS data for assessing forest health, land use classification, disturbance, risk monitoring, and more. These indices provide valuable insights into the state of vegetation and the environment.
Vegetation indices are incredibly versatile, with over 150 indices documented in the literature. They change over time, which increases their utility, though some may need adjustments for local environments. These indices are often visually interpretable and can significantly enhance the performance of machine learning algorithms. However, they can be affected by cloud cover and shadows, as well as saturation issues.

Use cases: Forest health, land use classification, disturbance, risk, monitoring


EVI anomaly data from Orbify

Vegetation Indices

Normalized Burn Ratio (NBR)
Machine Learning and Deep Learning
Machine learning and deep learning are ideal for land cover classification, especially in a dynamic world where the environment is constantly changing. These techniques are particularly useful for solving complex problems, such as tree species identification. Deep learning, which leverages convolutional neural networks, can extract textual information from RS images but requires high-resolution data, typically less than 3 meters per pixel.


Two-stage convolutional neural network-based method (TS-CNN) for oil palm detection in large-scale study area (around 55 km2) located in Malaysia, using high-resolution satellite images (i.e., Quickbird). https://www.mdpi.com/2072-4292/11/1/11
GEDI Global Ecosystem Dynamics Investigation
GEDI, a spaceborne LiDAR instrument, plays a crucial role in monitoring and assessing forest structure and biomass. This technology provides highly accurate and detailed data that is instrumental in environmental research.

Sentinel 1A
Sentinel 1A, a C-band Synthetic Aperture Radar (SAR-C) satellite, offers all-weather monitoring and day-and-night imaging. It can penetrate vegetation, making it invaluable for monitoring forest structure and biomass, as well as detecting deforestation and forest degradation.

C-band (SAR-C): The C-band offers a good compromise between spatial resolution and penetration through vegetation and moderate cloud cover.
ESA Biomass
The upcoming ESA Biomass satellite, set to launch in 2024, will be the first P-band wavelength SAR satellite. This technology will enable the measurement of biomass at a remarkable 200-meter resolution, offering a new level of precision.

LiDAR and SAR Alongside Ground Truth Datasets
For the most accurate assessments, LiDAR and SAR data should be used in conjunction with ground truth datasets. These technologies provide valuable information about forest structure, which can be combined with ground truth data to generate precise Above Ground Biomass (AGB) maps. Projects like GEO-TREES offer global forest plot data for reference, but local ground truthing data is often necessary for specific projects.
Conclusion
In summary, remote sensing data and satellite technology offer a powerful set of tools for addressing environmental challenges. Passive satellites, vegetation indices, LiDAR, and SAR data are essential for monitoring forest health, land use, and environmental changes. Machine learning and deep learning techniques enhance the analysis of this data. By combining these resources, we can unlock the power of satellites for nature-based solutions, tailoring our approach to specific regions and challenges.
If you are interested in learning more about this subject, we recommend watching our recorded webinar session on "Why Are Satellites the Game-Changer for Effective Carbon Credit Assessment and Tree Growth Monitoring?", where we showcase the transformative power of satellites in NbS. Or, alternatively, schedule a personalized demo call with us.
Related blog posts

Using Geospatial Data to Monitor Tree Growth
September 6, 2023 路 4 mins read
Read more
Satellite Data for Mitigating Deforestation Leakage: An Integrated Approach
September 5, 2023 路 7 mins read
Read more
How Can We Increase the Effectiveness of Carbon Credit Markets?
April 20, 2023 路 9 mins read
Read more