🌾 Applications in Crop and Soil Monitoring

This lesson builds core elective concepts in BSc Agriculture with practical applications and exam-oriented clarity.

---

Applications in Crop and Soil Monitoring

Crop Area Estimation

Accurate crop area estimation is one of the most important applications of remote sensing in India. The traditional approach — ground surveys and crop-cutting experiments — is expensive, time-consuming, and covers only a sample of fields.

Satellite-based approach:

1. Acquire multi-temporal images covering the crop season (e.g., June–November for kharif)
2. Classify pixels into crop types using NDVI time series profiles
3. Sum the pixels of each crop class × pixel area = crop area estimate
4. Validate with ground truth (farmer surveys, revenue records)

FASAL Programme (India): FASAL stands for Forecasting Agricultural output using Space, Agro-meteorology and Land-based observations. Launched by the Ministry of Agriculture with ISRO, ICAR-IASRI, and India Meteorological Department (IMD):

• Uses AWiFS (56 m, 5-day revisit) and LISS-III (23.5 m) from RESOURCESAT, plus MODIS (250 m, daily)
• Provides advance crop production forecasts — 1st pre-harvest estimate months before harvest
• Crops covered: rice, wheat, rapeseed, cotton, jute, sugarcane, potato
• Accuracy: ±5% for major crops (national level); higher error at district level
• Reduces farmer distress by enabling policy interventions (import/export decisions, procurement price notifications) ahead of harvest

---

Crop Condition Monitoring

NDVI time series from MODIS (daily, 250 m) or Sentinel-2 (5 days, 10 m) tracks crop condition through the season:

• Below-normal NDVI → stress (drought, flood, disease, nutrient deficiency)
• VCI (Vegetation Condition Index) = (NDVI_current − NDVI_min) / (NDVI_max − NDVI_min) × 100
  • VCI < 35 indicates severe drought stress; used by National Drought Monitoring Cell

Crop Phenology Mapping: Key dates extracted from NDVI time series — start of season (SOS), peak of season (POS), end of season (EOS) — help identify anomalies in crop development.

---

Yield Forecasting

Remote sensing-based yield forecasting uses statistical relationships between satellite-derived vegetation metrics and historical yield data:

1. Simple regression: Yield = a × NDVI_peak + b (calibrated for each crop/region)
2. Multi-variable models: NDVI + LAI + rainfall + temperature → better accuracy
3. Crop simulation models (DSSAT, ORYZA): Assimilate satellite-derived LAI to reduce uncertainty

ICAR-IASRI and ICAR-IARI routinely develop district-level yield models for kharif rice and rabi wheat using NDVI-based approach.

---

Precision Soil Sampling

Traditional soil sampling takes one sample per farm or per block — missing within-field variability. GPS-based precision sampling captures spatial variation:

Grid Sampling

• Divide field into uniform grids (typically 1 ha, 2.5 ha, or 5 ha)
• Take one composite sample per grid cell at GPS-recorded locations
• Consistent and comprehensive, but expensive for small grids

Zone-Based Sampling

• Delineate management zones using EC map, yield history, or NDVI history
• Take samples from each zone (fewer samples than grid, yet captures major variation)
• More cost-effective; recommended for smallholder farms

Directed Sampling

• Use satellite NDVI or EC map to identify variable areas
• Sample intensively in those areas; use fewer samples in uniform areas
• Lowest cost; best for initial variability assessment

---

Soil Nutrient Mapping

After GPS-tagged soil samples are analyzed in the laboratory, results are imported into GIS:

1. Create point layer with GPS coordinates and soil test values (pH, N, P, K, OC, etc.)
2. Apply spatial interpolation (IDW or kriging) to generate continuous surface maps
3. Overlay with farm boundary → prescription fertilizer map

India's Soil Health Card scheme (2015–) targets providing nutrient recommendations to every farmer; remote sensing supports large-area monitoring while the card scheme handles individual farm prescriptions.

---

Soil EC (Electrical Conductivity) Mapping

Soil EC reflects the combined effect of texture (clay content, CEC), organic matter, salinity, and water content — making it an excellent proxy variable for mapping soil spatial variability.

Methods:

• Veris 3100 or EC Surveyor: Tractor-mounted; measures shallow (0–30 cm) and deep (0–90 cm) EC simultaneously while driving across the field at 8–12 km/h; GPS logs one EC reading per second
• EM38 (Geonics): Handheld or sled-mounted electromagnetic induction sensor; non-contact; 0–150 cm depth
• Output: Dense grid of EC measurements → IDW or kriging interpolation → EC map

EC maps reveal soil type zones, tile drain patterns, organic matter gradients, and past flood deposits — all invisible from the surface but clearly visible in EC contour maps.

---

Soil Salinity Mapping

Soil salinity is a major problem in Indo-Gangetic Plains, coastal areas, and canal command areas — affecting ~6.73 Mha in India.

Remote Sensing Approach:

• Saline soils have distinctive spectral signatures in VNIR and SWIR wavelengths (whitish surface crust)
• Salinity Index: Several formulas; a common one: SI = sqrt(Red_band × SWIR_band); or (Red × Green) / Blue
• High reflectance in red and SWIR in bare soil images often indicates salt accumulation
• Multi-temporal composite of barren season images reduces crop cover interference
• NRSC has mapped salt-affected soils of India at 1:250,000 scale using IRS data

Ground Truth:

EC of saturation extract (ECe) > 4 dS/m = saline; EC < 4 but ESP > 15 = sodic (alkali). Remote sensing identifies probability zones; ground ECe measurement confirms diagnosis.

---

Crop Water Stress Mapping

Thermal infrared (TIR) sensors measure canopy temperature — stressed crops transpire less → canopy temperature rises:

• CWSI (Crop Water Stress Index): Compares measured canopy temperature to "well-watered" and "non-transpiring" baselines; CWSI = 0 (no stress) to 1 (maximum stress)
• Satellites: Landsat 8/9 Band 10 (100 m TIR), ASTER (90 m), ECOSTRESS (70 m) on ISS
• Drone-mounted thermal cameras: 5–10 cm resolution; detect individual stressed plants in orchards
• Application: Irrigation scheduling maps — identify high-stress zones for priority irrigation

---

Disease and Pest Detection

Remote Sensing Principles for Disease Detection:

Infected plants show changes in:

• Chlorophyll content → reduced absorption in red; reduced NIR reflectance
• Carotenoid/anthocyanin ratios → changes in blue and green reflectance
• Leaf water content (SWIR) → wilting
• Structural changes → altered texture in high-resolution imagery

Examples:

• Yellow rust (stripe rust) in wheat: Subtle spectral changes at 500–600 nm before visual symptoms are apparent; detectable with hyperspectral sensors or Sentinel-2 Red Edge bands
• Rice blast: Drone multispectral survey identifies early infection foci; NDRE (Normalized Difference Red Edge) sensitive to early chlorophyll loss
• Locust swarm detection: MODIS and Sentinel time series detect sudden NDVI drops; FAO's Desert Locust Monitoring uses this approach

Weed Mapping:

• Drone RGB or multispectral imagery at < 5 cm resolution
• Machine learning (CNN, Random Forest) trained to distinguish crop rows from weed patches
• Output: Prescription map for variable-rate herbicide application (spot spraying)
• Herbicide use reduced by 20–70% compared to blanket application

---

Waterlogging Mapping

Waterlogged areas have characteristic spectral properties (dark tone, low NDVI) and can be mapped using:

• Optical: Dark tone in NIR; negative NDWI; however, cloud cover during monsoon limits utility
• SAR (Sentinel-1, RISAT): Smooth water surface gives very low backscatter (specular reflection away from sensor) → dark areas = open water; C-band SAR penetrates some flood vegetation

Inundation frequency mapping: Time series of SAR images classifies each pixel as permanently flooded, seasonally flooded, or rarely flooded — essential for waterlogging risk zoning and drainage planning.

---

Flood Mapping and Damage Assessment

During major flood events (e.g., 2017 Bihar floods, 2018 Kerala floods):

1. SAR imagery (Sentinel-1, RISAT-1A) acquired immediately; maps flooded extent through clouds
2. Pre-flood crop map overlaid with flood inundation map → identifies affected crop area
3. NDVI change detection (pre vs. post flood) → severity of crop damage
4. NRSC activates National Disaster Management Support within 24–48 hours of a flood event

---

Land Degradation Mapping

The GLADA (Global Assessment of Land Degradation and Improvement) project used MODIS NDVI trends (1982–2006) to map areas where vegetation productivity is declining (degradation) or improving (recovery) globally.

India-specific: NRSC's National Wasteland Map (55+ categories at 1:50,000 scale) using IRS LISS-III; updated periodically; input to land reclamation programmes under NAIP and MGNREGS.

---

Carbon Stock Estimation

Remote sensing supports carbon accounting for UNFCCC reporting:

• Biomass = allometric relationship × canopy height (from LiDAR or ICESat-2 GEDI) × wood density
• Forest above-ground biomass maps for India: ISRO + FSI (Forest Survey of India)
• Soil organic carbon estimated from SWIR-based indices in bare soil images

---

Applications Summary Table

Application	Primary Satellite	Data Type	Key Output
Crop area estimation	RESOURCESAT AWiFS, Sentinel-2	NDVI time series	Crop area (ha) per district
Drought monitoring	MODIS	VCI, SPI, NDVI anomaly	Drought severity map
Flood mapping	Sentinel-1 (SAR)	Backscatter	Flood inundation map
Soil salinity mapping	IRS LISS-III	VNIR + SWIR reflectance	Salinity risk zone map
Crop water stress	Landsat 8 TIRS, ECOSTRESS	TIR (canopy temp.)	CWSI map for irrigation
Disease detection	Sentinel-2, drone multispectral	Red Edge, NDRE	Disease hotspot map
Yield forecasting	MODIS + Sentinel-2	NDVI peak, LAI	District yield forecast
Weed mapping	Drone RGB/multispectral	CNN classification	Prescription spray map
Waterlogging	Sentinel-1 SAR	Backscatter time series	Inundation frequency map
Land degradation	MODIS, IRS LISS-III	NDVI trend	Degraded land map

---

Overview

Remote sensing applications in crop and soil monitoring span the entire agricultural decision cycle — from pre-season planning (soil maps, salinity zones) through crop growth monitoring (NDVI condition, disease detection, water stress) to post-harvest assessment (yield forecasting, damage assessment). India's FASAL programme for advance crop production forecasts and PMFBY for insurance are flagship applications. Drone-based monitoring is extending satellite capabilities to field-level precision — enabling weed maps, disease early detection, and variable-rate application prescription maps at centimeter resolution.

---

Summary Cheat Sheet

Topic	Key takeaway
Main focus	Crop area estimation, soil mapping, drought and flood monitoring, disease detection, precision soil sampling, FASAL programme, and PMFBY remote sensing applications.
Section context	Revise this lesson with the rest of Remote Sensing Applications for stronger conceptual continuity.