Skip to main content

Harvests

Harvesting is a critical event in the cultivation cycle. In the Farm Data Model (FDM), the harvest is not just a single data point but a detailed data structure that captures the nuances of different crops and their specific measurements. This detailed approach allows for accurate agronomic calculations, such as nutrient balances and yield comparisons.

The Harvest Data Model​

In FDM, a harvest is intrinsically linked to a cultivation. A single cultivation can have multiple harvest events, especially for crops like grass that are cut multiple times a season. The core data structure for a harvest involves several related entities:

  • cultivation_harvesting: This table links a harvest event to a specific cultivation (b_lu) and records the b_lu_harvest_date.
  • harvestables: Each harvest event can produce one or more "harvestables", and "harvestable" can be linked to multiple harvest events. For example, a potato harvestable might be linked to multiple fields, but collected in the same batch. Note that while the database schema supports it, due to the complexity fdm-core functions currently only support 1:1 relations between harvests and harvestables.
  • harvestable_analyses: For each harvestable, one or more analyses can be recorded. This is where the specific measurements for the harvest are stored.

This structure provides a flexible and detailed representation of harvest events, accommodating a wide range of crops and harvesting practices.

Variable Harvest Parameters​

A key challenge in modeling harvests is the diversity of parameters measured for different crops. What is relevant for a silage maize grower is different from what a potato farmer measures. For example:

  • For silage maize, a farmer might record the fresh weight yield (b_lu_yield_fresh) and the dry matter percentage (b_lu_dm).
  • For potatoes, the gross yield (b_lu_yield_bruto), tare percentage (b_lu_tarra), and underwater weight (b_lu_uww) are common parameters.
  • For cereals, fresh weight yield (b_lu_yield_fresh), moisture content (b_lu_moist), and crude protein (b_lu_cp) are often measured.

FDM is designed to store these crop-specific parameters as they are measured in the field, ensuring that the raw data is preserved.

Standardization using Harvest Categories​

While storing crop-specific parameters is essential for data fidelity, many agronomic calculations require standardized values. The two most important standardized parameters in FDM are:

  • b_lu_yield: The dry matter yield in kilograms per hectare (kg DM/ha).
  • b_lu_n_harvestable: The nitrogen content of the harvested product in grams of nitrogen per kilogram of dry matter (g N/kg DM).

To bridge the gap between the various measured parameters and these standardized values, FDM uses a system of Harvest Categories (b_lu_harvestcat). Each cultivation in the FDM catalogue is assigned to a harvest category. This category determines which parameters are expected at harvest and, crucially, defines the formulas used to calculate the standardized b_lu_yield and b_lu_n_harvestable.

Example of Harvest Categories and Formulas​

Here are a few examples of how harvest categories are used to standardize harvest data:

  • HC020 (Grassland):

    • Measured parameters: b_lu_yield (already in kg DM/ha) and b_lu_cp (crude protein in g/kg DM).
    • Formulas:
      • b_lu_yield is used directly.
      • b_lu_n_harvestable is calculated from crude protein: b_lu_n_harvestable = b_lu_cp / 6.25.

  • HC042 (Potatoes):

    • Measured parameters: b_lu_yield_bruto (kg/ha), b_lu_tarra (%), and b_lu_uww (g/5kg). b_lu_n_harvestable is often a default value for the crop.
    • Formulas:
      • First, dry matter content (b_lu_dm) is calculated from underwater weight: b_lu_dm = (b_lu_uww * 0.049 + 2.0) * 10.
      • Then, fresh yield is calculated: b_lu_yield_fresh = b_lu_yield_bruto * (1 - b_lu_tarra / 100).
      • Finally, dry matter yield is calculated: b_lu_yield = b_lu_yield_fresh * b_lu_dm / 1000.

  • HC050 (Cereals):

    • Measured parameters: b_lu_yield_fresh (kg/ha), b_lu_moist (%), and b_lu_cp (g/kg DM).
    • Formulas:
      • Dry matter content is calculated from moisture: b_lu_dm = (100 - b_lu_moist) * 10.
      • Dry matter yield is calculated: b_lu_yield = b_lu_yield_fresh * b_lu_dm / 1000.
      • Nitrogen content is calculated from crude protein (using a specific factor for cereals): b_lu_n_harvestable = b_lu_cp / 5.7.

Practical Implications​

This two-step approach of storing raw data and standardizing it through harvest categories offers several advantages:

  • Data Integrity: The original, measured data is never lost.
  • Flexibility: The system can easily be extended to support new crops or measurement techniques by adding new harvest categories.
  • Comparability: By converting all harvest data to a standardized format, it becomes possible to compare yields and nutrient removal across different crops and farms.
  • Agronomic Accuracy: Standardized data is essential for accurate nutrient management, as it allows for precise calculation of nutrient offtake in the harvested biomass.