Current dominant crop sequences across EU: a typology based on LUCAS dataset

1 Introduction

Assessing benefits expected from crop diversification requires an accurate description of current cropping systems (i.e. crop sequence and crop management) as a baseline. However, such a description is still lacking for Europe, for at least two reasons: (i) existing datasets at the European scale (e.g. LUCAS dataset, see below) provide information about land-use categories but no information about crop sequence and crop management, (ii) national and sub-national datasets that describe crop sequence and/or management practices are available in the EU-state members, but a lack of harmonization (e.g. spatial and temporal resolution, included descriptors) between these datasets makes them difficult to use in a cross-analysis. To overcome this problem, we developed an original method that combines European-level and national-level datasets to create a map of current dominant crop sequences at the European level.

2 Materials and Methods

The method is based on two public datasets: (i) the French national “Pratiques culturales” survey (PK), and (ii) the European Land Use Cover Area frame statistical Survey (LUCAS). The PK dataset contains information about 5-year crop sequences for a sample of fields across France (c.a. 20 000 fields surveyed for arable crops in 2014, Agreste, 2016). The LUCAS dataset contains information about land use for a grid of points in all EU-state members (c.a. 45 000 points with observations in 2008, 2012 and 2015 for agricultural uses, EUROSTAT, 2019). Both datasets are independent. We proceeded in three steps. First, we classified fields from the PK dataset, according to eight variables characterizing crop sequences (i.e. frequencies of cereals, oilseed rape, sunflower, dry pulses, corn, sugar beet plus potato, alfalfa plus clover and temporary grassland). Based on this set of eight variables, a principle components analysis combined with hierarchical clustering distinguished seven groups. Second, based on this first typology, we derived decision rules to assign each point of the LUCAS dataset to one of these groups. To validate these rules, we compared distributions of crop sequence types for France, resulting respectively from PK fields classification, and from LUCAS points assignment. As both distributions were very close, we finally applied assignment rules to LUCAS data for all EU countries.

3 Results

For France, the first type is characterized by crop sequences based on oilseed rape and cereals (Table 1). The second one gathered corn-based crop sequences, which could be monocropping or corn-cereals sequences. Third came sunflower-cereals sequences. The fourth type gathered crop sequences with annual crops and temporary grassland. Crop sequences based on industrial crops, such as sugar beet or potatoes represented the fifth type. Then came crop sequences with dry pulses (e.g. pea, faba bean, soybean) rotating with cereals. The seventh type gathered crop sequences including multiannual forage legumes (e.g. alfalfa or clover). 

The consistency of both approaches validates the assignment of LUCAS observations as a good proxy of crop sequences distribution. Assignment rules applied to other EU countries resulted in a ranking of crop sequences, close to the French one (Figure 1). At the EU scale, oilseed rape-cereals sequences represented 41%, corn-based sequences 21%, sequences with temporary grassland 9%, beet or potato-based sequences 9%, sunflower-based sequences 6%, sequences with dry pulses 6%, and sequences with alfalfa or clover 5%. Most of LUCAS observations across EU fitted into crop sequence types defined for France; only 3% went in none of these groups.

4 Discussion and Conclusions

In the next months, this description of dominant crop sequences, already validated by partners from each country, will be completed with information about crop management, to describe current cropping systems. This typology could be used to assess all types of diversification, for example increasing protein crops, as proposed by the European protein plan.