Period: 2014 – present
Frequency: 10 days
Data can be accessed here.
Algorithm description: Ten day composite of GDMP (gross dry matter productivity) and DMP (dry matter productivity). Both quantities are expressed in units of dry matter instead of kg of C (unit: kg of DM/ha/day).
The algorithm is same as that used for Copernicus 1km DMP and GDMP product, version 2. The only difference is 300m FAPAR product is used as an input. However, the 300m FAPAR product is not gap-filled (unlike GEOV2 1km) and have more stringent constraints on solar zenith angle.
Accuracy: The product has more data gaps than Copernicus 1km DMP and GDMP product pertaining to the missing data in 300m FAPAR product used as input. Comparison with the Fluxnet data (for limited number of sites) and other reference products (MODIS GPP/NPP) showed that the performance of the product is highly dependent on the type of canopy.
The product is not able to capture swift changes in the cropland as well managed grasslands due to its coarser pixel size. Overall accuracy is moderate however there is a broad scatter of points around 1:1 line when compared against FLuxnet data.
- Missing data especially in the northern latitude.
- Other limitations similar to Copernicus 1km DMP and GDMP product.
Swinnen, E. & van Hoolst, R. Algorithm Theorethical Basis Document: Dry Matter Productivity (DMP) collection 300m Version 1. 1–50 (2018).
Swinnen, E., van Hoolst, R. & Tote, C. Quality Assessment Report: Dry Matter Productivity (DMP) collection 300m Version 1. 1–71 (2018).