Vegetation vigour evolution in the Romanian forests, at 10m resolution, 2022-2023

Data description

These datasets were generated for the use case "Monitoring forest vegetation vigour evolution over phenological cycles" in the context of the Open Earth Monitor Cyberinfrastructure project.

We used Sentinel2 L2A products, with the provided cloud cover/contamination masks to generate the vegetation vigour product for years 2022, 2023 and to obtain the evolution 2022-2023 of this parameter, for the Romanian territory. The ancillary data employed was the ESA Land Cover 2021 v200 product for the delimitations of the tree cover class regions.

The vegetion vigour (VV) product is a custom type of pixel-based index that characterises the vegetation during long periods of time, preferably during a complete phenological cycle, represented by the integral (i.e. cumulative, the area of the subgraph) measure of NDVI intensity. The product does not represent a snapshot of the vegetation, but it characterises its cumulative state over time. For each pixel, the VV is proportionally higher as the NDVI values are higher, over the entire analysed period. Thus, the VV can be used to qualitatively compare regions of classes of vegetation, over multiple phenological cycles. 

The evolution of the vegetation vigour - which is the main output - represents the comparison between 2 consecutive analysed periods (in this case, calculated for 2022 and 2023). For visualisation purposes, the authors propose 10 sequential classes, low value classes indicate less forest vegetation vigour, high value classes indicate intense forest vegetation vigour.

However, to allow forestry specialists to work on visualisations based on their field knowledge, the authors have uploaded the obtained vegetation vigour for year 2022 and 2023, 8bit unsigned raster files (0-255) as well. 

No data regions or areas where the data was insufficient are represented with 0. 

The vegetation vigour products have an auxiliary trust layer, with values from 0 to 100, 100 representing 100 % certainty of the obtained value of VV in the pixel. The trust layer is strongly correlated with the number of revisits and the medium level of scene contamination. 

Naming Convention

To ensure consistency and ease of use across and within the projects, we follow the standard Open-Earth-Monitor-Cyberinfrastructre file-naming convention. The convention defines 10 fields that describe important properties of the data, as follows: 

Generic variable name: vegetation_vigour

Variable procedure combination i.e. method standard: yearly_integral_ndvi

Position in the probability distribution / variable type: c = classes

Spatial support: 10m

Depth reference or depth interval e.g. below ("b"), above ("a") ground or at surface ("s"): s

Time reference begin time (YYYYMMDD): 20220101

Time reference end time: 20230101

Bounding box (2 letters max): eu

EPSG code: epsg.3035

Version code i.e. creation date: v20250123

Thus, the file name is:

vegetation_vigour_yearly_integral_ndvi_c_10m_s_20220101_20230101_eu_epsg.3035_v20250123

Source Data

Copernicus Sentinel 2 L2A data [2022-2023]

To identify the foresty vegetation areas where the algorithm would run:ESA WorldCover 10 m 2021 v200. https://doi.org/10.5281/zenodo.7254221