The impact of scale on remote sensing of forests

The role of forests within the global ecosystem is something we’ve discussed quite a bit through our blog. From the ecosystem services that forests provide, to the proposed implementation of REDD+ and the concept of natural capital. Carbomap specialises in remote sensing monitoring techniques, collecting key data for effective and robust management of these environments, and we thought we would take a closer look at the differences between scales of monitoring.


Paracou, French Guiana – 1km2 spatial resolution

Satellites provide an ideal solution for providing synoptic views of forests at very large scales. For example, free global data products from NASA’s MODIS satellite are available (up to) every 8 days. This satellite provides optical images (light we can see, plus infrared) which can be easily combined with other data to extract useful information. However, a downside is the poor resolution of the images. The best products available from MODIS are 250m squares- though typically they are 1km. To demonstrate how this would look for the project in French Guiana we created this canopy height model image.

Although there are satellites which can provide a much higher level of detail, the cost of data from these satellites can be expensive. Instruments such as MODIS are perfect for looking at the international or regional scale, and provide a great approach for giving a broad snapshot over a large area.

Paracou, French Guiana - 1ha spatial resolution

Paracou, French Guiana – 1ha spatial resolution

However, here are Carbomap we feel that if you have a monitoring budget, and the area of forest doesn’t cover vast areas, then other types of data are better fit for purpose. Other sources, such as radar, can give you  measurements of the forest structure, which is more useful for estimating carbon (the focus of REDD+).

One of the most widely used radar datasets for such applications is the SRTM dataset. This is also free, and has an approximate resolution of a hectare (1 hectare = 100x100m2), which is good for looking at the landscape to regional level. To give you an idea of what a CHM from such data might look like in French Guiana, we’ve created this image.

There are other types of radar data more accurate than SRTM, for example TerraSAR-X can provide data at a resolution down to the metres level, but again cost starts to creep in. However, given the amount of information that you can get for large areas, this is an incredibly economic way of monitoring your forest.

But what about if you need much finer information? What if your forest is on a smaller scale?


Paracou, French Guiana – 1m2 spatial resolution

The answer that Carbomap would give you is LiDAR. Aircraft and Unmanned Aerial Vehicle laser scanning instruments are ideal platforms for measuring forests on the plot to landscape scale. The resolution can be down to the centimetre level, and the information is in 3D – showing details about the canopy height and structure that is not possible from other remote sensing technologies. In the image here it is easy to see the improved level of detail is compared to the above images from satellite resolution data.

Full Airborne LiDAR systems can cover hundreds of square kilometres in a single flight, meaning that you can collect a lot of detailed data very quickly. They are ideally suited to the landscape and regional scale. UAVs on the other hand have a more limited range, covering tens of square kilometres in a single flight. These limitations are partly due to the size of the craft and battery life, but also related to UAV flight legislation. We have discussed the differences between full Airborne LiDAR and UAV-LiDAR in an earlier blog post.

In summary, there are a range of different techniques that you can apply, and each has it’s own strengths and weaknesses; depending on the area of forest and the type of data you’re interested in. If you have some forests that need mapped, get in touch with Carbomap and we can talk about the best approach for you.


Co-authored by Gordon Moran

One thought on “The impact of scale on remote sensing of forests

Leave a Reply

Your email address will not be published. Required fields are marked *