LAMASUS D5.3 Costing Database

This dataset provides a comprehensive database of economic costs associated with land use management (LUM) across Europe. The database spans cropland, grassland, forest, and wetland systems, and is designed to enable robust cost comparisons and facilitate integration into large-scale ex-ante models such as CAPRI and GLOBIOM. The overarching objective is to support policy evaluation and scenario analysis by quantifying the economic implications of diverse land use practices.

The methodology for deriving cost data differs by land use category, reflecting the diversity in available data, model structures, and expertise. Cropland costs are estimated using a micro-econometric approach based on a translog cost function; grassland costs are simulated via the bio-economic farm model FarmDyn; forest management costs are derived using an engineering-based model adapted from the ForestNavigator project; and wetland rewetting costs are based on a literature review. The database is structured using a harmonised cost scheme – covering fertiliser, machinery, diesel, seed, and labour costs – for cropland, grassland, and forest.

Cropland systems are represented by LUM-specific cost estimates derived from a multi-step micro-econometric approach using a translog cost function framework. Based on harmonised FADN (Farm Accountancy Data Network) data for specialised crop farms, the method estimates input demand (fertiliser nitrogen - N, phosphorus - P, and potassium - K, fuel, labour, machinery, pesticides) and allocates it across ten major crops at the NUTS2 level. Integration with yields from the EPIC (Environmental Policy Integrated Climate) model enables cost estimations for LUM classes, ensuring consistency with agronomic performance. Results, expressed in EUR (2015) per hectare, reveal substantial variation across crops, LUMs, and regions, primarily driven by differences in fertiliser and labour costs, as well as machinery use. This approach bridges the gap of limited availability of crop-specific physical input data, by using cost data and estimated behavioural relationships to derive consistent input demand estimates. This enables a more robust, spatially explicit cost database for cropland costs to be incorporated into large-scale ex-ante models, such as GLOBIOM.

Grassland systems are characterised by eight LUM types, ranging from very high to low intensity, and reflecting the differences between pasture and managed grassland. FarmDyn simulations, using CAPRI-derived 2015 price data, provide detailed cost components (fertiliser, machinery, diesel, seed, and labour) tailored to country-level input prices. Results in EUR (2015) per hectare show significant cost variation across LUMs and countries, mainly driven by differences in labour and machinery depreciation. Notably, synthetic fertiliser is not applied in lower-intensity systems where manure application fulfils the N requirements, and machinery investments significantly affect per-hectare costs even in low-input systems.

Forest cost data are developed using a spatially explicit engineering-based approach covering harvesting, regeneration, thinning, and road infrastructure, with cost differentiation by management type, terrain, and tree species. Five forest LUM types are considered, ranging from primary (unmanaged) to very intensive management, with primary forest incurring no costs. Costs are reported both per cubic metre of roundwood and hectare. Intensive and very intensive systems exhibit the lowest per-unit product costs due to high yields, while closer-to-nature and combined objective LUMs incur higher costs due to lower operational efficiency and reduced yield. Country-level variations are significant, influenced by terrain, the level of mechanisation, and input prices.

Wetland management focuses exclusively on peatland rewetting. Due to limited site-specific data, a uniform annualised cost, based on a detailed literature review, of 205 EUR ha-1 and yr1 (based on a literature-derived EUR 3,262 ha-1 investment) is applied EU-wide. This simplification is acknowledged as a significant limitation, with implications for the accuracy of regional policy assessments.

For more information refer to the Deliverable (see references) or the README.txt.