Earth Observation (EO) can be defined as the gathering, analysis and interpretation of information about the planet’s resources and behaviours, by using remote sensing technologies supplemented with ground-based observations. Earth Observation is used to detect, monitor and assess changes in natural and built environments. Examples include the monitoring of the state and evolution of the environment (e.g., monitoring of vegetation health) and the ability to rapidly assess situations during crises, such as extreme weather events.
Earth Observation is supported by different data acquisition technologies, like aircraft, drones, satellites and ground-based observations (for instance, smartphone apps can be used to collect field data to dynamically verify remote sensing data).
In the last years, the European Space Agency (ESA), through the Copernicus Programme, has launched several satellites, among which Sentinel-1 A/B and two twin satellites Sentinel-2 A/B which have on-board active and passive sensors, respectively. The large amount of free and open data acquired by these satellites has broadened the range of applications, by allowing analyses on long and dense time-series of imagery systematically with global coverage. For instance, Sentinel-1, a C-band Synthetic Aperture Radar (SAR) satellite, can be used to estimate soil moisture and map deforestation, while Sentinel-2 multispectral data can be used for monitoring and detecting changes in vegetation phenology, crop monitoring and land cover classification.
In this regard, Earth Observation-based analysis can help monitor changes in forest areas. This information can be used to track the efforts to reduce deforestation and forest degradation in developing countries and estimate activity data as part of REDD+ (Reducing Emissions from Deforestation and forest Degradation) initiatives.
Cocoa agroforestry systems are considered forest systems and, as such, part of the REDD+ MRV (Measurement, Reporting and Verification) programme. Earth Observation-derived products can support sustainable management and decision-making in cocoa agroforestry systems.
The MRV4C project has generated a series of geospatial products/services that underpin the sustainable management of cocoa in the Dominican Republic, by providing geospatial insights. The following products are available:
- Forest/Non-Forest product: this product shows the extent of all forest types, meeting the national definition of forest in the Dominican Republic. Forest areas were classified by applying a machine learning algorithm on EO optical imagery. The product could be used as a basis to produce forest/non-forest change maps showing the increase or decrease of forest between years.
- Land Suitability for Cocoa Cultivation map: the product is a thematic map showing the suitability of lands in the Dominican Republic for growing cocoa agroforestry systems. The product was derived by using a machine-learning algorithm on a set of EO-derived layers, which describe environmental factors appropriate for growing cocoa trees. This map could be used to plan cocoa production and support land-use planning.
- Above-Ground Biomass and Carbon Stocks product: this map provides estimates of above-ground biomass in cocoa agroforestry systems, by relating SAR backscatter to above-ground biomass. The carbon stocks layer was derived from the Above-Ground Biomass product and offers a one-time snapshot of the carbon stored within a landscape. The product can contribute to fulfilling REDD+ activities, especially when used in combination with the suite of MRV4C products, for instance, the Forest Loss product (to track scenarios where forest cover loss would contribute to emissions) or the Forest/Non-Forest product (to detect carbon-rich, high-value forested areas for conservation purposes).
- Classification of cocoa agroforestry: this product maps the extent of cocoa agroforestry systems, by making use of machine learning algorithms and EO optical imagery. The Cocoa Classification product can be used in combination with the suite of MRV4C product, for instance with the Forest Loss product (to highlight cocoa agroforestry areas which may have affected forest cover) or Above-Ground Biomass product (to estimate how much biomass is stored within the cocoa agroforestry systems).
- Forest Loss product: this product maps forest cover loss events that occurred in the year of analysis, by applying a probabilistic approach on a SAR data time-series. The product could help certify “zero-deforestation” supply chain commitments and provide evidence for REDD+ projects.
- Agricultural Drought Indicator: this product provides levels of alert for drought, based on the cause-effect relationship of agricultural drought, according to which a shortage of precipitation leads to a deficit of soil moisture, which in turn results in reduced crop productivity. The Agricultural Drought Indicator was derived by using a combination of EO optical imagery and satellite-measured variables and could support prioritise interventions in agricultural areas prone to drought.
Sentinel-1 and Sentinel-2 (© ESA, 2020)