Canopy N across European forests: comparing spatial patterns of canopy N retrieved from remote sensing, environmental variables and global vegetation models

Humans have dramatically increased atmospheric CO₂ concentration as well as biologically available nitrogen (N). Nitrogen is an essential nutrient for vegetation growth and N availability represents a limiting factor on carbon (C) sequestration by the terrestrial ecosystems.  While there is a large infrastructure for measurements to constrain the C cycle, data to constrain the N cycle are less readily available. Using a combination of remote sensing products (MODIS), canopy N concentration data (ICP forest), plant functional type and environmental variables including soil, climate (WorldClim) and elevation (EU-DEM), we generated a canopy N map across European forests using a random forest statistical method (hereafter RF canopy N map).

Most current Global Vegetation Models (GVMs) have integrated C and N cycles, to account for the link between C and N for plant growth and respiration. Leaf N concentration is also important for other biomass compartments as N allocations are prescribed relative to leaf N.  The objective of this study is to compare canopy N of two GVMs, O-CN and LPJ-GUESS, and the RF canopy N map in European forests.

The obtained canopy N maps show contrasting spatial patterns. The RF canopy N map shows higher canopy N values, i.e. between 1.8 and 2.2 %N, in mid-western and eastern Europe, while showing lower values, i.e. 1.2 and 1.6 %N, around the Mediterranean region and in the south of Sweden. The canopy N map obtained from the O-CN simulation shows relatively lower canopy N values, ranging from 1.0 to 1.8 %N, in central and northern Europe, while in the Mediterranean region the values are higher, between 1.8 and 2.4 %N. Similar to the RF map, the LPJ-GUESS canopy N map shows relatively higher canopy N values in mid-western Europe compared to southern and northern Europe, however, the LPJ-GUESS canopy N values show little spatial variation in the Mediterranean region.  Also, the LPJ-GUESS values are higher, with canopy N values ranging between 2.0 and 2.8 %N in mid-western Europe, and canopy N values ranging between 1.6 and 1.8 %N in the Mediterranean region.

The analysis yields insight into spatial differences in RF canopy N and canopy N predicted by GVMs, with especially a mismatch in arid and warm regions.