Why do we need to look at forests, in multiple spectral bands?

The forestry industry plays a significant role in the world economy and our ecosystem. In recent years, satellite imagery has become a critical tool in monitoring the health and management of our vast forest lands, helping us understand that inadequate forest management has been having a very negative impact on our environment and economies.

Protecting our forests is one of the major challenges our planet is currently facing. The destruction of forests, whether it is by human hands or natural disasters, is responsible for around 15% of total greenhouse gas pollution. Considering this and the fact that our world’s population will likely reach 10 billion in the next thirty years, and together with it our need for food, fibre and fuel, it is only realistic to assume that the commercial exploitation of our forests will also exponentially increase. Using forests for other economic activities will always be more attractive to businesses if they are able to profit from deforesting.

Therefore, it is crucial to better understand forest conditions with a special focus on type, structure and history, as well as the forest’s current condition and short-term state. Unfortunately, the vastness and inaccessibility of these forest areas make effective management extremely challenging.

Earth observation via satellite imagery is certainly the answer to most of the forest management complications. The next challenge lies in selecting the right tool. Remote sensing solutions work well to track changes in land use and land cover, the localization of deforestation in near real-time, the understanding and managing of the supply chain and optimized forest operations.

These systems can be optimized to provide valuable data to detect early warning signs of stress in the forest, identify specific changes within the ecosystem in near real-time, classify the levels of stress to prioritize actions, quantify the impact in economic terms and geo-locate the specific area of interest.

Depending on the instrument selected, remote sensing imagery creates a huge amount of data that needs to be transformed into actionable information. This data will cover large areas that may not be easily accessible for continuous monitoring and may require artificial intelligence algorithms to autonomously spot trends and recommend actions.

The main advantage of a good geospatial information system is that it will provide quality information to support decision making and focus actions to where it is most needed. It can even go one step further by analyzing the immediate environment that drives the demand for deforestation, improving transparency, accountability and citizen oversight.

This will, in turn, feed into the decision-making processes behind policies so that forests are governed more efficiently, allowing us to take a great stride forward in the fight against illegal deforestation.

Remote sensing from satellites is quite common in monitoring deforestation, afforestation, natural disasters and the general health of forests. The ability to acquire spatial and spectral information on a temporal basis with remote sensing satellites can be highly beneficial for pre- and post-incident assessment relative to conventional surveys. The satellite data currently being used (such as MODIS, Landsat, Sentinel-2, Worldview-2, Planet Scope) either lack spatial, spectral, radiometric or temporal resolution, or a combination thereof.

Remote-sensing data with adequate spectral and spatial resolution information is frequently used to detect areas in forests with high fire fuel content. Enhanced spectral resolution allows the monitoring of changes in canopy chlorophyll mass per area, a great parameter to track forest health. In general, green plants have a very specific spectral signature, with low reflectance in the visible light, and high reflectance in the near infrared range. This transition from the visible to near-infrared part of the spectrum is known as the red-edge effect and is useful due to the sensitivity to both chlorophyll content (chla+b) and canopy structure. The bands of interest lie in the area from about 680 nm to 780 nm. Figure 1 and Figure 2 provides a visual comparison between the vegetation reflectance curves sampled with a multi-spectral and hyper-spectral imager.

Figure 1: The spectral sampling of a typical 7 band multi-spectral imager supper imposed on a typical vegetation reflectance spectrum (based on Senstinel-2 bands).

 

Figure 2: The spectral sampling of a typical 200 band hyper-spectral imager supper imposed on a typical vegetation reflectance spectrum.

 

Using this data, fire fuel vegetation and burn-hazard areas are identified and classified according to the risk of burn-hazard areas (from low to high) to focus on possible interventions. After a fire disaster, remote sensing data is used to quantify the ecological and economic impact and long-term effects as well as restoration planning. This will also highlight (map) potential areas where flash floods and soil erosion may occur.

Figure 3 and Figure 4 show the forest area around Knysna as captured by Sentinel-2 before a huge forest fire that destroyed large parts of the area. Figure 5 and Figure 6 shows the impact of the forest fire, captured approximately one month after the fire.

Figure 3: Sentinel-2 False Color image of the area around Knysna (South Africa) before the devastated forest fires of June 2017 (Bands 8, 4, 3).

 

Figure 4: Sentinel-2 NDVI image of the area around Knysna (South Africa) before the devastated forest fires of June 2017 (based on combination of bands (B8-B4)/(B8+B4)).

 

Figure 5: Sentinel-2 False Color image of the area around Knysna (South Africa) one month after the devastated forest fires of June 2017 (Bands 8, 4, 3).

 

Figure 6: Sentinel-2 NDVI Color image of the area around Knysna (South Africa) one month after the devastated forest fires of June 2017 (based on combination of bands (B8-B4)/(B8+B4)).

 

Each of the three xScape100 solutions features its own strengths and limitations when applying it to forestry applications:

TriScape100:

Images within the visible spectrum provide a visual representation of the areas of interest. The combination of visual content, relative high-spatial resolution and temporal resolution, when used as part of a constellation, is ideal for mapping areas of deforestation, vulnerability and locating areas of impact after a disaster. Due to only operating in the visible spectrum (RGB), this sensor does have limitations in identifying subtle changes in the ecosystem, classifying and differentiating between specific problems and qualifying the impact.

MultiScape100:

This multispectral sensor combines the best of spatial, spectral and radiometric resolution. This sensor is ideal for detecting areas of interest, identifying changes in the ecosystem and, by carefully selecting the spectral bands, it can classify stress levels, differentiate between vegetation types and quantify severity levels. The MultiScape100’s option of 7 spectral bands in the VNIR spectral range and 4.65 m GSD at 500 km will cover a large range of forest management solutions.

HyperScape100:

Hyperspectral sensors considerably increase the spectral resolution of remote-sensing solutions, at the expense of radiometric and spatial resolution. This sensor allows for forest phenotyping and brings a holistic approach to plant-performance measurement and forest management. Cover changes and plant stress levels can be with extreme accuracy. Although the spatial resolution of the hyperspectral instrument can be 4 to 5 times lower than that of the corresponding hyperspectral solution, it is still more than adequate for forestry applications.

With a unique combination of spectral, spatial and radiometric solutions, the xScape100 series of optical payloads bring cost-effect remote-sensing solutions to CubeSats, and with it a wide range of forest management opportunities.

Add to this a high revisit time when used as part of a constellation of CubeSats and advanced image-processing algorithms, this solution may provide daily monitoring of areas at a fraction of a hectare across the globe, automatically.

Please see the xScape100 series of payloads’ datasheets for more information on the technology features.

0 replies

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

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