Crop Data Ludemann et al 2022

Summary statistics data from a range of articles found in the scientific literature related to crop harvest index and nutrient concentrations of crop products (e.g. beans and grain) and crop residues (straw and stover).

Provider's website

https://doi.org/10.5061/dryad.djh9w0w4m Downloaded file summary_statistics_data_from_articles.csv

https://doi.org/10.1016/j.fcr.2022.108578

Documentation

Ludemann, C.I, Hijbeek, R., van Loon, M., Murrell, T.S., Dobermann, A. and van Ittersum, M. (2024), Published field experimental data of crop yields, nutrient concentrations, and harvest indices from around the world.

Ludemann CI, Hijbeek R, van Loon MP, Murrell TS, Dobermann A, van Ittersum MK. 2022. Estimating maize harvest index and nitrogen concentrations in grain and residue using globally available data. Field Crops Research 284: 1-25.

Pre-processing

Original data filtered:

• in case the location is missing (latitude and longitude)
• in case both sowing and harvest date are missing
• in case the crop type (attribute Crop_standardised) is missing

The object name in AGROSTAC is a combination of the record ID (simply the row number of the original dataset named Summary_statistics_data_from_articles.csv) and the string “Ludemann_et_al_2024” e.g. Ludemann_et_al_2024_2 (first number is header in original file).

The comment is based:

• DOI
• Record ID (of the original dataset)
• Location (latitude and longitude in decimal degrees)
• Objective of the experiment

The combination DOI and location determines an experiment in the Ludemann dataset. Each experiment can have one or multiple treatments. Each treatment is one record in file Summary_statistics_data_from_articles.csv and this is defined as an AGROSTAC object (name) identified by the record ID (simply the row number of the original dataset named Summary_statistics_data_from_articles.csv).

The geometry accuracy is based on the accuracy of the co-ordinates. However we have seen examples where these co-ordinates point to a location within a city clearly not the location of the experiment. So the given geometry accuracy must be used with care and checked against topograhic map layers.

CROP_CODE

Attribute: Crop_standardised
Remark: -

CUL_NAME

Attribute: Crop_variety
Remark: -

CROP_DEV_BBCH

Attribute: Date_sowing_YYYYMMDD (Sowing date in YYYYMMDD format) and Date_harvest_YYYYMMDD (Harvest date in YYYYMMDD format)
Remark: Harvest date is also used as the date of the plant density, yield and above ground biomass

PLANT_DENSITY_CNT_M2

Attribute: Harvest_density_Plants_m2 (Density of plants at harvest, plants per m2)
Remark: Harvest date is used as the date of the observed plant density

SO_FWT_KGHA / SO_DWT_KGHA

Attribute: CPY_mean_kg_fresh_ha (Mean crop product, kg per ha in fresh weight), CPY_mean_kg_DM_ha (Mean crop product, kg per ha in dry weight)
Remark: Harvest date is used as the date of the observed crop yield

TOPS_FWT_KGHA / TOPS_DWT_KGHA

Attribute: AGY_mean_kg_fresh_ha (Mean above ground biomass, kg per ha in fresh weight), AGY_mean_kg_DM_ha (Mean above ground biomass, kg per ha in dry weight)
Remark: Harvest date is used as the date of the observed above ground biomass

Management

The management was treated as follows:

• FieldManagementType = FIELD_TRIAL
• NutrientsNType = see below
• NutrientsPType = see below
• NutrientsKType = see below
• NutrientsManagementType = see below
• WaterManagementType = see below
• PestsDiseasesManagementType = UNKNOWN

With regard to nutrient and water management, we tried manually to indicate if the treatment of an experiment was optimal or suboptimal for nitrogen (N), phosphorus (P), potassium (K), nutrient (N/P/K) and water management. Therefore, we used the values of the attributes:

• Experimental_objective_Description
• Fertiliser_N_kg_N_ha
• Fertiliser_P_kg_P_ha
• Fertiliser_K_kg_K_ha
• Actual_irrigation_mm

This was done for experiments that studied varying N, P, K or irrigation applications. The assumption is that experiments that vary nutrient or water have at least one treatment that is optimal for the varying factor. The experiments were identified by looking at:

• the objective of the experiment (e.g. studying effect of varying N-application)
• variation of N, P, K or irrigation application

The treatment with the highest application was labelled optimal while the other treatments, having lower application values, were labelled as suboptimal. In case these experiments had relative high application value of the other main inputs (nutrients, water), these inputs were also labelled as optimal. In case all N, P and K were labelled as optimal the nutrient management as a whole was also labelled optimal, otherwise suboptimal. Note that some experiments were not yet included because of time constraints (multiple experiments that share one objective).

In all other cases (other experiments) the management was labelled as unknown.

Note that this indication of optimal or suboptimal management is a first rough approach and must be used with care. We advise to also consult the original data (Summary_statistics_data_from_articles.csv) which is easily found through the record ID.