Selecting the right satellite sensor for forest carbon monitoring is a critical decision. It’s not about finding the single “best” sensor, but the most appropriate one for a specific project. The choice hinges on balancing the technical capabilities of optical, SAR, and LiDAR sensors against the accuracy requirements for Measurement, Reporting, and Verification (MRV) and, of course, the budget. Each sensor type offers a unique view of the forest, capturing different properties of the vegetation and terrain.

Optical sensors are the workhorses of Earth observation. They capture sunlight reflected from the Earth's surface, much like a digital camera. Two of the most widely used optical satellite missions for environmental monitoring are Europe's Sentinel-2 and the US's Landsat 8/9 programme.

While both provide invaluable data, they have key differences. Sentinel-2 generally offers a higher spatial resolution, with some bands capturing details as small as 10 metres. Landsat's primary bands are at 30 metres. This finer detail from Sentinel-2 can be crucial for mapping smaller-scale forest disturbances. However, the true power of these sensors lies in their spectral capabilities, particularly in bands invisible to the human eye.

The electromagnetic properties of vegetation are key to monitoring its health. Healthy plants are rich in chlorophyll, a pigment that strongly absorbs visible red light for photosynthesis. Simultaneously, the cellular structure of healthy leaves strongly reflects near-infrared (NIR) light. By measuring the difference between the NIR and red light signals, we can create spectral indices that act as a proxy for vegetation density and health.

This is where Sentinel-2 has a distinct advantage: its inclusion of Red-Edge bands. These are narrow bands located at the transition point between red light absorption and NIR reflection. This part of the spectrum is highly sensitive to subtle changes in chlorophyll content, making it excellent for detecting early signs of forest stress or tracking growth stages with high fidelity.

While optical sensors are excellent for assessing canopy health, they can struggle to quantify biomass directly. They mostly see the top layer of the forest. To understand the full 3D structure, we need different tools: LiDAR and SAR.

LiDAR (Light Detection and Ranging) works by sending out laser pulses and measuring the time it takes for them to return. By analysing the return signal, we can map the forest in three dimensions. NASA's [{] mission, mounted on the International Space Station, is specifically designed for this. It provides precise measurements of canopy height, vertical structure, and ground elevation, which are critical inputs for accurate biomass estimation.

Synthetic Aperture Radar (SAR) offers another powerful, all-weather alternative. SAR sensors actively send out microwave pulses and record the backscatter that returns. Because microwaves can penetrate clouds, haze, and rain, SAR provides reliable data regardless of atmospheric conditions, a major advantage in persistently cloudy regions like the tropics.

Different SAR frequencies (like L-band and C-band) interact with the forest in different ways. Longer wavelength L-band signals can penetrate deeper into the canopy and interact with larger branches and trunks, making them more sensitive to biomass. However, SAR data is complex, and its signal can become saturated in very dense forests, meaning it loses sensitivity beyond a certain biomass threshold. This saturation effect is a key challenge that needs to be accounted for in carbon monitoring models.

Beyond sensor physics, temporal resolution is a crucial factor. This refers to how often a satellite revisits the same location. Sentinel-2, with its two-satellite constellation, can image a given spot every 5 days at the equator. This high frequency is essential for change detection, allowing for near real-time monitoring of deforestation, logging, or the impact of events like wildfires.

Finally, there's the issue of data access. Fortunately, data from major government-led missions like Landsat, Sentinel, and GEDI are provided free of charge under open data policies. This has democratised access to Earth observation data, enabling researchers and organisations worldwide to develop and implement robust forest monitoring systems. However, processing these vast datasets requires significant computational resources, which remains a practical constraint.

Choosing the right sensor involves a careful analysis of these trade-offs. Optical sensors provide excellent spectral information for forest health, LiDAR gives unparalleled structural detail for biomass, and SAR offers all-weather reliability. Often, the most powerful approach involves combining data from multiple sensors to overcome the limitations of any single one.