Importance of complex LULCC representation for N cycling and productivity in the EC-Earth framework

With human land-use activities expected to increase in the future, it is important to understand how LULCC (Land-Use Land-Cover Change) activities affect the Earth’s surface, climate and biogeochemical cycles. Here we use the CMIP6 version of the EC-Earth3 Earth System Model (ESM) to assess the impacts of LULCC on surface fluxes of carbon (C) and nitrogen (N). EC-Earth is one of the first ESMs to interactively couple a 2nd generation dynamical vegetation model (LPJ-GUESS) with mechanistic C-N dynamics in soil and vegetation to an atmospheric model. The size, age structure, temporal dynamics and spatial heterogeneity of the vegetated landscape are represented and simulated dynamically in LPJ-GUESS. Such functionality has been argued to be essential to correctly capture biogeochemical and biophysical land-atmosphere interactions on longer timescales and has been shown to improve their representation compared to more common area-based vegetation schemes. The patch-based structure of LPJ-GUESS also makes it possible to represent the history (soil, litter status) of a single patch as it might have been involved in several land-use transitions. We examine the effects on surface fluxes of carbon and nitrogen in three LUMIP simulations, for both offline land-only and fully coupled ESM runs. We focus on the effects of gross land-use transitions (“land-hist”), net land-use transitions (“land-noShiftcultivate”) and fixing land-use at 1850 levels (“land-noLu”).

In general, EC-Earth shows a higher historical C loss due to LULCC than other ESMs, but our results are still in line with LULCC emissions constrained by biomass observations. Gross transitions result in a higher historical C loss compared to net transitions, while runs without LULCC (noLu) show a constant gain in C. LULCC also affects the total N content of the system and hence soil nutrient content. Gross (net) LULCC leads to a loss of 1.5 (1.0) PgN over the historical period whereas 1.5 PgN is gained in noLu runs. As increases in global fertilization and harvest fluxes more or less offset one another for both gross and net LULCC, the differences in total N pools derive from biological nitrogen fixation (BNF), soil fluxes, leaching and land-use associated fluxes of N. Changes in soil fertility result in a higher productivity for net compared to gross transitions, mainly in the Tropics. Net transitions also results in less N lost through land-use change and hence a higher net mineralisation rate. This is mainly notable in the Tropics where the initially organic matter content is lower compared to temperate and boreal regions. The productivity and harvested biomass from crops are similar for gross and net transitions as their N source mainly comes from N fertilization, with the exception of some developing countries where N fertilisation is not as high as in industrialised countries. Based on these examples of how the N cycle and productivity are affected by LULCC, we argue that the full complexity of gross transitions is required to accurately predict how LULCC affects the N cycle, productivity and biogeophysical feedbacks to the climate.