Seasonal Crop Monitoring with NDVI Time Series: From Planting to Harvest
Quick Answer: An NDVI time series over an agricultural field shows a characteristic curve that rises during crop establishment and vegetative growth, peaks at canopy closure, and falls sharply at harvest. Deviations from this expected curve — suppressed peak, early senescence, or mid-season stress dip — indicate crop problems. SAVI should be used during early growth stages when bare soil is visible; NDMI detects water stress before visible NDVI decline.
Why Satellite Crop Monitoring Matters
Traditional crop monitoring relies on field scouts walking through fields, regional surveys, and farmer self-reporting. These methods are slow, expensive, spatially limited, and often delayed by weeks or months. By the time a stress event is documented through traditional channels, much of the damage may already be done.
Sentinel-2 provides 10-meter imagery every 5 days, enabling near real-time NDVI monitoring of individual fields or entire agricultural districts. A sudden NDVI dip at critical growth stages triggers investigation before crop loss is irreversible.
The NDVI Growth Curve for Annual Crops
Every annual crop follows a characteristic NDVI trajectory through its growing season. Understanding this curve is the foundation of crop monitoring.
Key Phenological Stages
1. Bare soil / Pre-emergence After harvest or during fallow period, the field is bare or has crop residue. NDVI is low (0.1–0.3), possibly with some variation from soil moisture and residue conditions.
2. Seedling / Early establishment Germination and early growth. NDVI begins rising slowly. Because the canopy is thin and soil is visible between rows, SAVI is more appropriate than NDVI during this stage (SAVI corrects for soil background contamination).
3. Vegetative growth Rapid NDVI increase as leaf area expands. The rate of increase indicates crop growth rate — a slow rise can signal germination problems, pest damage, or drought stress at early stages.
4. Canopy closure / Peak At peak canopy, NDVI reaches its maximum for the season — typically 0.6–0.9 depending on crop species, density, and health. The timing and magnitude of this peak are key indicators of crop condition.
5. Reproductive / Grain fill After flowering, NDVI may stabilize or decline slightly as the plant redirects resources from leaves to seeds. This phase is critical — drought or heat stress during grain fill has high yield impact.
6. Senescence / Ripening As leaves age and dry down in preparation for harvest, NDVI declines. The timing of this decline indicates approaching harvest readiness.
7. Harvest Abrupt NDVI drop to bare-soil levels as crop is removed.
8. Post-harvest Depending on management practices: bare soil (low NDVI) or cover crop / residue (moderate NDVI).
Detecting Problems in the Crop Cycle
Suppressed Growth Curve
If the NDVI trajectory rises more slowly than expected or peaks at lower values than previous years, this indicates reduced canopy development from:
- Germination failure — Poor seed quality, unfavorable soil temperature, pest damage
- Nutrient deficiency — Low nitrogen or micronutrient availability
- Early drought — Water stress during establishment phase
- Disease — Foliar diseases reducing photosynthetic leaf area
Compare the current year's curve with the same field from 2–3 previous years (same crop type). A peak that is 20–30% below the historical average warrants field investigation.
Mid-Season Stress Dip
A temporary NDVI decline during the growing season followed by partial recovery indicates a transient stress event:
- Drought period — NDMI typically drops before NDVI; NDVI dips then recovers when irrigation or rainfall arrives
- Pest outbreak — Localized damage patterns visible at 10m resolution
- Hail or storm damage — Sudden widespread dip following extreme weather events
- Waterlogging — Low-lying areas flood and NDVI drops; NDWI may show concurrent increase
Early Senescence
If NDVI begins declining earlier than the historical timing, this indicates:
- Terminal drought stress — Crop ran out of soil moisture before physiological maturity
- Disease — Foliar diseases accelerate senescence
- Frost damage — In temperate regions, unexpected late-season frost
Early senescence is particularly damaging because it cuts short the grain-fill period, directly reducing yields.
Missing Season
If the expected growth curve never appears — NDVI stays at bare-soil levels during what should be the growing season — this indicates:
- Crop failure — No germination or complete early crop death
- Fallowing — Deliberate decision not to plant
- Conflict or access issues — In some regions, agricultural abandonment is conflict-related
The Multi-Index Approach for Agriculture
NDVI + SAVI for Reliable Growth Tracking
Use SAVI during early growth stages (sparse canopy with visible soil) and NDVI once the canopy fully covers the soil. Track both indices to get a seamless picture from planting to harvest.
NDMI for Water Stress Detection
NDMI is more sensitive to plant water status than NDVI because it uses the SWIR band, which responds directly to leaf water content. During dry spells:
- NDMI drops 1–2 weeks before NDVI — This is the early warning window
- When NDMI drops significantly but NDVI is still normal, irrigation is needed immediately
- By the time NDVI drops, visible wilting is likely already occurring
NBR for Burn Damage
If fire passes through agricultural areas — from controlled burns, wildfires, or crop residue burning — NBR detects the damage extent. Compare pre- and post-fire dates to assess what fraction of the monitored area was affected.
Understanding Seasonal Rhythms
Single vs. Double Cropping
Some agricultural regions produce two crops per year:
- Single-crop: One NDVI peak per year
- Double-crop (wheat then maize, for example): Two peaks per year with different heights (different crops have different NDVI values)
- Irrigated vs. rainfed: Irrigated crops in arid regions show NDVI peaks regardless of rainfall; rainfed crops show irregular inter-annual variation following precipitation
A multi-year time series reveals the cropping system being practiced, which is itself valuable information for agricultural land use mapping.
Inter-Annual Variation
Comparing the same field across multiple years reveals:
- Whether peak NDVI is trending upward (improving management, better varieties)
- Whether peak NDVI is trending downward (soil degradation, water table depletion)
- Good years vs. bad years following rainfall patterns
Setting Up Agricultural Monitoring
Polygon selection: For small fields (<1 km²), draw tightly around the field boundary to avoid mixing signals from neighboring fields with different crop types. For large agricultural districts, a single larger polygon captures the regional average.
Index selection: For most crop monitoring, start with:
- NDVI (primary crop health indicator)
- SAVI (early growth stage)
- NDMI (water stress detection)
Start date: Capture at least one full growing season — ideally 12–24 months back to see the current season in context.
Beyond Field Monitoring: District-Scale Surveillance
For regional agricultural authorities, monitoring multiple polygons covering different administrative units simultaneously reveals:
- Which districts have the lowest crop development relative to historical norms
- Where drought or pest events are spatially concentrated
- Whether national food security projections based on average conditions capture local vulnerability
This type of systematic surveillance is the basis for early warning systems used by food security agencies.
Summary
Crop NDVI time series monitoring provides a continuous, automated picture of crop development from planting to harvest. The characteristic growth curve — slow rise, rapid growth, peak at canopy closure, decline at senescence, harvest drop — provides the baseline against which stress events, growth failures, and anomalies are measured. SAVI complements NDVI during early growth stages; NDMI provides earlier warning of water stress. A two-to-three year baseline makes current season anomalies clearly visible by comparison with historical normal curves.
