Metrics Variants

Understand how to interpret vegetation metrics across different analytical views

Carbonleap provides multiple metric variants for each vegetation index to help you understand field conditions from different perspectives. These variants allow you to evaluate what is happening today, how it compares to past scans, and how it compares to historical performance.

Metric variants are available for all vegetation indices:

  • NDVI

  • NDRE

  • NDMI

  • MSAVI

Each metric can be viewed through four different lenses.


1. Actual Values

“What is happening right now?”

Actual Values display the raw vegetation metric captured by the most recent satellite scan. Use this to understand:

  • Canopy vigor

  • Growth stage consistency

  • Areas of weak or strong development

  • Water stress signals (for NDMI)

  • General field health

This is the primary map most growers start with.

Best for: Early growth, mid-season monitoring, hotspot detection.


2. Change From Previous Scan

“How is the field changing over time?”

This variant compares the current scan to the previous scan and highlights whether vegetation values increased or decreased.

You can quickly identify:

  • Rapid growth zones

  • Areas slowing down or declining

  • New stress hotspots

  • Blocks recovering after irrigation or rain

Reductions in value may indicate:

  • Water stress

  • Pest/disease onset

  • Nutrient deficiency

  • Canopy thinning

  • Leaf senescence (late season)

Best for: Detecting new issues quickly and tracking week-to-week crop behavior.


3. Percentile

“How does this field compare to itself this year?”

Percentile evaluates today’s value against all other scans from the same season.

For example:

  • 90th percentile = higher than 90% of this season’s scans

  • 20th percentile = lower than most of this season

This helps understand whether the field is performing above or below its own typical pattern for the current year.

Best for: Mid-season benchmarking, detecting unusual dips or spikes, identifying areas falling behind.


4. 10-Year Percentile (Historical Percentile)

“How does today compare to the past decade?”

This variant compares the current condition of the field to vegetation patterns from the past 10 years for the same date range.

This helps answer:

  • Is this crop behind or ahead of a typical season?

  • Is canopy development unusually low for this time of year?

  • Are we seeing a historically strong or weak season?

It is especially useful for:

  • Perennial crops

  • Climate-affected seasons

  • Long-term planning and forecasting

Best for: Historical context, early season assessment, anomaly detection.


When to Use Each Variant

Growth Stage
Recommended Variant
Why

Early Season

Actual Values, 10-Year Percentile

Evaluate emergence vs. historical patterns

Rapid Growth

Change From Previous Scan, Actual Values

Identify inconsistencies in canopy development

Mid-Season

Percentile, Change From Previous Scan

Track performance relative to this year

Stress Monitoring

Change From Previous Scan, NDMI Actual

Detect new stress zones quickly

Late Season

Percentile, 10-Year Percentile

Monitor canopy decline and compare to typical ripening curves


How Metric Variants Improve Decision-Making

Metric variants help growers:

  • Spot issues early

  • Compare performance across time

  • Understand whether changes are normal or abnormal

  • Plan scouting more efficiently

  • Focus attention on the right areas of the field

  • Monitor recovery after irrigation, sprays, or cultural operations

By switching variants, you can look at the same field through multiple analytical lenses.


Summary

Metric variants allow you to interpret vegetation data in deeper and more meaningful ways:

  • Actual Values: Current reality

  • Change: Short-term momentum

  • Percentile: This season’s context

  • 10-Year Percentile: Historical context

Together, these views give you a complete picture of how your field is performing now, how it is changing, and how it compares to expected seasonal and historical behavior.

Last updated

Was this helpful?