GeoDataFarm Premium Crop Simulation

GeoDataFarm Premium adds a crop simulation workspace to the free GeoDataFarm QGIS plugin. It combines field records, weather, soil, fertilizer, irrigation, crop, and variety information to produce planning estimates for crop water stress, fertilizer timing, and yield.

Important: This is a planning model, not a guaranteed harvest forecast or a substitute for local agronomic advice.

Before You Start

Create a farm, connect to it, and add at least one field. The simulation works best when the field has:

• A planting record with crop, planting date, and variety where available.
• Soil clay and organic-matter measurements.
• Dated fertilizer applications with numeric rates.
• Irrigation records with spatial coverage, when irrigation is used.
• Weather data covering the selected growing period.
• Harvest data if you want to compare the estimate with an actual result or train the model.

Missing soil or fertilizer information does not always prevent a run, but the plugin may use a simpler fallback or omit that factor from the estimate. The result indicates what information was available.

Activate Premium

1. Open the Crop simulation tab in QGIS.
2. Purchase a Premium license through the Get a pro licence button (through Lemon Squeezy).
3. Paste the license key into the license field.
4. Click Activate.

License activation sends the license key and an installation identifier to the license-verification provider. Farm records and simulation data are not uploaded by the plugin. A previously validated license can continue working offline for a limited grace period; the plugin checks the license again when a connection is available.

Run a Simulation

1. Select a field.
2. Choose the crop detected from the planting records, or select a crop manually.
3. Set the From and To dates. Use a range that covers planting through the expected harvest or the date you want to inspect.
4. If the crop was deliberately stopped before natural maturity, enable Growth stopped early and choose the stop date.
5. Click Run simulation.

The simulation fetches or reads the weather needed for the selected period. It then calculates a field-wide season estimate and a cell-by-cell map.

Read the Field Map

The Field map selector has three views:

Water stress

Shows each grid cell’s soil-water condition on the selected date. The map is calculated cell by cell using local soil, crop, variety, and irrigation records where available. Gray cells indicate that no usable value was available.

Total rain + irrigation

Shows cumulative rain and irrigation received by each cell through the selected date. Irrigation is assigned using the logged operation’s spatial coverage, so only cells covered by an operation receive that operation’s water.

Predicted yield

Shows the estimated harvestable yield for each cell on the selected date, in tonnes per hectare. The map can vary between varieties, soil zones, and irrigation zones. The legend shows the lowest and highest estimates for the current date.

The field is divided into a grid. Very large fields are automatically drawn with a coarser grid so the map remains responsive. This changes map resolution, not the underlying field-wide season estimate.

Use the Date Slider

After a successful run, drag the date slider to inspect the season day by day. The stress and water views show the selected day’s state. The predicted-yield view shows yield formation through the season.

Crop development is driven by growing degree days rather than by a linear percentage of calendar time. The model uses the crop’s base temperature and season GDD parameters. Harvestable yield is delayed until the crop’s formation phase:

• Potatoes remain near zero during early vegetative growth and form most harvestable yield during tuber bulking.
• Barley and other cereals remain near zero until late grain filling, then approach the final estimate near maturity.

These formation thresholds are simplified crop-level assumptions. Local climate, cultivar, planting date, and management can shift the real timing.

Planned Fertilizer Applications

Use Add planned application to test a possible fertilizer date and rate without writing it to the database. Run the simulation again after adding or removing a planned application.

The application detail reports a timing-risk category. When numeric nitrogen rate, soil information, and weather data are available, the advanced balance estimates water movement, uptake, and possible leaching. Otherwise, the plugin uses a simpler rainfall-risk index based on the rain after application.

Varieties and Crop Settings

Variety information is read from planting records. When spatial planting data identifies different varieties in different parts of a field, each grid cell can use its own variety settings. A cell without a variety match falls back to the field-wide crop model.

Open Crop model settings to inspect or edit model values. Settings can be saved for a crop or for a specific variety. Variety-specific settings are used by the per-cell map; the single season summary remains a field-wide estimate.

Teach Your Model

The Teach your model section fits selected historical examples to your data.

1. Click Scan farm.
2. Review the field/year rows and compare predicted with actual harvested yield.
3. Uncheck unreliable or non-representative examples.
4. Click Train selected.
5. Review the fitted values and error before saving.
6. Select the crop or variety fit and click Save this fit.

The fitting currently adjusts potential yield, nitrogen response, and the minimum nitrogen yield factor. Fits are grouped by crop and variety. A fit with too few or poor-quality examples can be misleading, so use multiple representative field/year records and validate the result against seasons that were not used for fitting when possible.

The Allow multi-year crops option is off by default because a planting record from an older year may be stale for an annual crop. Enable it only for genuinely multi-year crops such as grassland or other perennial systems.

Scientific Basis and Limitations

The simulation is literature-informed and intentionally planning-grade. It uses:

• FAO-56-style crop coefficients and growth stages for crop water demand.
• FAO-33 and Jensen-style multi-period yield response for water stress.
• Growing degree days to represent crop development.
• Crop-specific nutrient uptake and leaching approximations for nitrogen and potassium.
• Literature-informed attainable-yield baselines, which can be refined using the training workflow.

The model does not fully represent every source of yield variation. In particular, results do not comprehensively model pests, disease, weeds, frost, runoff, detailed mineralization, salinity, winter vernalization, equipment performance, or every cultivar’s physiology. Phosphorus and magnesium are reported as season-total supply status rather than tracked as daily yield-limiting pools.

Weather quality, soil measurements, planting records, spatial coverage, and harvest data all affect the result. A trained model is only as useful as the historical examples used to fit it.