The full-scale operations of Sentinel-2 had hardly begun when researchers began to strive for free satellite data with higher resolution. Michael Schmidt from CONABIO, Mexico and Matt Hansen from University of Maryland, USA announced a plan for a constellation of eight satellites in a forestry workshop in Mexico in early June 2016. The workshop ‘Regional solutions to forest type stratification and characterizing the forest state for national forest monitoring and carbon emissions reporting’ was organized by the Global Forest Observation Initiative, CONABIO, and the US aid program SilvaCarbon.
The proposed satellites, six of which would fly in a polar sun-synchronous orbit and two in an equatorial orbit could provide imagery of the tropical region at every 45 minutes. This would enable gathering of frequent cloud free images even in the cloudiest regions. The special resolution of the proposed satellites is 2.5 m in the panchromatic band and five meters in the multispectral bands reaching to the short wave infrared area of the spectrum. Additional thermal infrared bands will have resolution of up to 20 m. Each satellite will have an the image size of 125 km but with the satellites flying in a formation a 250 km swath of near-simultaneous imaging is reached.
The investment cost of the constellation has been estimated as €700 million including image reception, processing and delivery systems. The annual operating costs of the constellation would be approximately €10 million and the data would be delivered free of charge to any user. The funding is proposed to come from national space agencies and development aid programs.
The properties of the images from the proposed constellation fall in between Sentinel-2 that provides high quality free data at 10m resolution and the commercial very high resolution satellites that offer images at up to 30 cm resolution. According to the proposers the rationale behind this constellation is that Sentinel-2 resolution and the frequency of the imaging is considered too low in particular to support the developing countries in monitoring their forest and land cover as the primary data. Investment costs, technical limitations and the commercial market restrict attempting a higher spatial resolution.
In practice the constellation would compete for the users with Sentinel-2 program on one hand and the commercial very high resolution satellite operators on the other hand, being ‘a poor man’s’ data source of very high resolution imagery. Particularly the tropical developing countries will need new procedures for the monitoring of their forest and land use to get financial incentives when they participate in the REDD+ (Reduction of Emissions from Deforestation and forest Degradation) program.
The workshop agreed that fine resolution data has to be used to collect reliable reference to derive statistical data on forest and land cover. However, such data should be collected by applying random sampling which means that no wall-to-wall coverage is needed. Wall-to-wall-mapping is an effective component in the statistical framework but the maps alone do not provide reliable information and uncertainty data.
It is an open question whether the free data from the Sentinel satellites of the Copernicus program with strong supporting infrastructure combined with a sample of sub-meter resolution imagery would be more optimal solution for forest monitoring than wall-to-wall imagery with a resolution of a few meters as an exclusive data source.
Finding an optimal concept for reliable forest cover monitoring is worth investing in now.