Linking the human appropriation of net primary productivity-based indicators, input cost and high nature value to the dimensions of land-use intensity across French agricultural landscapes

The intensification of European land use accelerated substantially in a few decades, particularly in agro-ecosystems that are facing an increasing demand for agricultural products and whose area is also constrained by other uses (e.g. urbanization). Increases in land use intensity (LUI) are characterized by increases of the agricultural outputs per land unit through management practices and/or by increasing amounts of inputs. LUI is a complex and multi-dimensional issue in which each dimension needs to be considered to have a better understanding of the impact of LUI on ecosystem functioning and biodiversity. Here, we focused on five existing LUI indicators: the Input Cost per hectare (IC/ha), assessing the expenses in inputs (fertilizers, pesticides, etc.), the High Nature Value (HNV), a scoring system of agricultural areas accounting for the presence of landscape elements and practices favorable to biodiversity, and three indices of the Human Appropriation of Net Primary Productivity (HANPP) framework, i.e. the harvested biomass (HANPPharv), the living biomass flow remaining available after harvests (NPPeco) and HANPP which combines harvested biomass and effects of land use conversion. First, we discussed how these indicators can relate to the dimensions of LUI. Then, we tested whether HANPP, HANPPharv and NPPeco were redundant with IC/ha and HNV throughout 25,758 French metropolitan municipalities, using Pearson's correlation coefficient and Linear Mixed-effects Models, while accounting for climatic and landscape parameters. As expected, HANPP, NPPeco and HANPPharv were highly correlated with each other, but weakly to HNV and IC/ha. HNV showed a positive relationship with NPPeco but negative with HANPP and HANPPharv. The opposite findings were observed with IC/ha. These three indicators seem complementary to HNV and IC/ha indicators, linking farmland structural properties and inputs intensity. Finally, we showed how these indicators can be linked, i.e. particular combinations of the indicator values could reveal contrasting agro-ecosystems types (e.g. intensive vs extensive crop farming).