Averages yield across all years at each location. A sample with 180 bu in 2024 and 220 bu in 2025 becomes 200 bu. Best for: Finding patterns that persist across years. Smooths out weather variation.
All Years (Combined)
Each year treated as separate data point. Same location counted multiple times (once per year). Best for: Maximum sample size when years had similar growing conditions. β οΈ Combines good and bad years together - results may be misleading if yields varied significantly.
All Years (Year-Normalized)
Each year treated separately, but yields converted to % of that year's average. Example: 200 bu yield in a 180 bu avg year = 111%. 250 bu yield in a 250 bu avg year = 100%. Best for: Comparing across years with different weather. Isolates nutrient effect from seasonal variation.
Individual Year (e.g., "2024")
Only uses yield from that specific harvest year. Best for: Analyzing a specific season's results.
π€ How should I choose?
Similar yields across years? β "Combined" gives you more data points
Very different yields (drought vs good year)? β "Year-Normalized" controls for weather
Want simplicity? β "Averaged" is most straightforward
Analyzing specific season? β Select individual year
π Understanding Field Normalization×
Why do correlations change when normalized?
WITHOUT normalization:
Correlations include differences between fields. A high-yielding field with high P will inflate the P-yield correlation, even if P isn't the reason that field yields well (could be better soil type, drainage, etc.)
WITH normalization:
Yields are converted to "% of field average", removing field-to-field differences. This isolates the nutrient effect by asking: "Within each field, do higher nutrient levels produce above-average yields?"
Why correlations often decrease:
Some of the original correlation was driven by field productivity differences, not the nutrient itself. The normalized correlation is stricter but more trustworthy for fertilizer decisions.
Example:
Raw correlation: P = +0.35 (includes field effects)
Normalized: P = +0.18 (nutrient effect only)
The 0.18 is more reliable for predicting if adding P will help within a field.
π‘ Recommendation:
Use normalized correlations when deciding on variable-rate fertilizer applications within fields. Use raw correlations when comparing overall field performance.
π
Individual Correlations
See how each nutrient independently relates to yield. Higher RΒ² = stronger relationship. Positive correlation means more nutrient = more yield.
π
No Yield Data Available
Import yield maps on the Import page to see correlations.
Scatter Plot: Nutrient vs Yield
bu/ac
Threshold Analysis
Data Verification Report
π¬
Multivariate Regression
Analyzes all nutrients together. Shows which nutrients significantly affect yield after controlling for the others - answers "What matters when everything is considered?"
π Select Variables for Model
π Model Summary
β οΈ Collinearity Warning
Some variables are highly correlated with each other, which can make coefficient estimates unreliable.
π Regression Coefficients
Each coefficient shows the nutrient's effect on yield while holding other nutrients constant
Variable
Coefficient
Std Error
t-value
p-value
Sig.
Significance: *** p < 0.001, ** p < 0.01, * p < 0.05, . p < 0.1
π― Yield Predictor
Enter nutrient values to predict yield based on the model.
π
Yield by Nutrient Level
Groups soil samples into Low, Medium, and High categories for each nutrient, then shows the average yield for each group. Helps identify which nutrient levels are associated with the best yields.
Loading yield bucket analysis...
π
Breakpoint Analysis
Find the critical threshold where yield response changes. Below the breakpoint = yield penalty from deficiency. Above = diminishing returns on fertilizer investment. Uses crop-specific thresholds (Corn: 5 bu/ac, Soybeans: 2 bu/ac).
Breakpoint Result: Zn
Breakpoint--
Yield Penalty--
Confidence--Based on penalty size, stability %, sample count, and years of data. High = strong evidence, Medium = likely real, Low = needs more data.
Stability--Bootstrap test: how often does the breakpoint land in the same range when we resample your data? Higher % = more reliable.
Below breakpoint:-- points, avg yield -- bu/ac
Above breakpoint:-- points, avg yield -- bu/ac
π Impact Summary
Category
Points
Est. Acres
Avg Nutrient
Avg Yield
vs Above BP
π΄ Below Breakpoint
--
--
--
--
--
π‘ Near Breakpoint
--
--
--
--
--
π’ Above Breakpoint
--
--
--
--
baseline
π° Estimated Yield Opportunity:--
πΎ Fields to Address
Fields ranked by number of points below breakpoint
π Management Zones
Clusters of nearby points below breakpoint - target these areas
π Individual Point Details
(click to expand)
Point ID
Field
Nutrient Value
Yield
Flags
Action
Hinge-MVR Results
Linear model using bent-curve transformation at breakpoint
RΒ²:--
Below-breakpoint effect:--
Above-breakpoint effect:--
β οΈ No clear breakpoint detected for this nutrient.
This could mean:
Yield penalty below threshold is less than 5 bu/ac
Not enough data points on one side of potential thresholds
Nutrient levels are relatively uniform across the field
Select a field and nutrient to analyze spatial changes over time.