Agricultural Monitoring with EVI and Phenology

Your First Phenology Chart in 60 Seconds

Let's get a win right away. This script charts MODIS EVI for an agricultural point in central Iowa across the 2023 growing season. We'll see the characteristic rise, peak, and decline of crop growth.

// Define an agricultural point in central Iowa
var point = ee.Geometry.Point([-93.5, 42.0]);

// Load MODIS 16-day EVI (1 km resolution)
var modisEVI = ee.ImageCollection('MODIS/061/MOD13A2')
  .filterDate('2023-01-01', '2023-12-31')
  .select('EVI');

// Scale EVI values (MODIS stores EVI * 10000)
var eviScaled = modisEVI.map(function(image) {
  return image.multiply(0.0001)
    .copyProperties(image, ['system:time_start']);
});

// Create the phenology chart
var chart = ui.Chart.image.series(eviScaled, point, ee.Reducer.mean(), 1000)
  .setOptions({
    title: 'MODIS EVI Phenology - Iowa Cropland (2023)',
    vAxis: {title: 'EVI', viewWindow: {min: 0, max: 1}},
    hAxis: {title: 'Date'},
    lineWidth: 2,
    pointSize: 4
  });
print(chart);
Map.centerObject(point, 10);

Why Should We Care About Crop Monitoring?

Traditional crop surveys require field agents visiting farms, counting plants, and estimating yields. That approach is slow, expensive, and limited in coverage.

Satellites solve this by observing millions of hectares every few days. Here's what agricultural remote sensing supports:

• Food security: Organizations like the USDA and FAO monitor global crop conditions to anticipate shortages before they become crises.
• Precision agriculture: Farmers use vegetation maps to identify stressed areas within fields, enabling targeted irrigation and fertilization.
• Crop insurance: Insurance programs use satellite data to verify crop losses and process claims more efficiently than on-site inspection.
• Yield estimation: Peak vegetation greenness correlates with final yield, allowing harvest forecasts months before crops are harvested.

The key to all of these? Measuring how green crops are over time. That's exactly what vegetation indices like EVI give us.

EVI vs NDVI: What's the Difference?

We already know NDVI, which uses only the NIR and red bands. EVI takes it a step further by adding the blue band and applying correction coefficients that reduce two common problems in agricultural monitoring.

The EVI formula:

EVI = 2.5 * ((NIR - RED) / (NIR + 6*RED - 7.5*BLUE + 1))

Why does this matter for agriculture? Let's break it down:

• Soil background: In early growth stages, when crop cover is sparse, bare soil between rows strongly influences NDVI. Think of it like background noise in a recording. EVI's correction coefficients reduce this soil signal, giving us a cleaner measure of actual plant greenness.
• Saturation: NDVI approaches a maximum of about 0.9 even as vegetation continues to thicken. EVI remains more sensitive at high biomass levels, which is critical for distinguishing a good corn crop from an excellent one.
• Atmospheric effects: Haze, smoke, and aerosols scatter light and affect the red band. EVI's blue-band correction compensates for these atmospheric distortions.

Here's how we calculate EVI from Sentinel-2 imagery using .expression():

// Load Sentinel-2 SR and filter
var s2 = ee.ImageCollection('COPERNICUS/S2_SR_HARMONIZED')
  .filterDate('2023-06-01', '2023-08-31')
  .filterBounds(ee.Geometry.Point([-93.5, 42.0]))
  .filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE', 20))
  .median();

// Scale reflectance values
var scaled = s2.divide(10000);

// Calculate EVI using .expression()
var evi = scaled.expression(
  '2.5 * ((NIR - RED) / (NIR + 6 * RED - 7.5 * BLUE + 1))', {
    'NIR': scaled.select('B8'),
    'RED': scaled.select('B4'),
    'BLUE': scaled.select('B2')
  }).rename('EVI');

Map.addLayer(evi, {min: 0, max: 0.8, palette: ['brown', 'yellow', 'green']}, 'EVI');
Map.centerObject(ee.Geometry.Point([-93.5, 42.0]), 12);

Reading the Rhythm: Phenology Curves

A phenology curve plots a vegetation index over time for a specific location. Think of it as a crop's heartbeat across the year. The shape tells us everything about the crop's life cycle.

For a typical summer crop in the US Midwest, the curve follows a predictable pattern:

1. Dormancy (January to April): Low EVI values because fields are bare or covered in crop residue.
2. Green-up (May to June): EVI rises rapidly as crops emerge and begin growing.
3. Peak (July to August): Maximum greenness, when the crop canopy is fully developed.
4. Senescence (September to October): EVI declines as crops mature, dry down, and are harvested.

MODIS is ideal for phenology because it captures images every 16 days at a consistent scale. The MOD13A2 product provides pre-calculated EVI at 1 km resolution, so we don't need to compute it ourselves. Let's see it in action.

// Build a multi-year phenology comparison
var point = ee.Geometry.Point([-93.5, 42.0]);

// Load three years of MODIS EVI
var evi2021 = ee.ImageCollection('MODIS/061/MOD13A2')
  .filterDate('2021-01-01', '2021-12-31').select('EVI');
var evi2022 = ee.ImageCollection('MODIS/061/MOD13A2')
  .filterDate('2022-01-01', '2022-12-31').select('EVI');
var evi2023 = ee.ImageCollection('MODIS/061/MOD13A2')
  .filterDate('2023-01-01', '2023-12-31').select('EVI');

// Chart one year as a starting point
var chart = ui.Chart.image.series(
  evi2023.map(function(img) {
    return img.multiply(0.0001).copyProperties(img, ['system:time_start']);
  }), point, ee.Reducer.mean(), 1000)
  .setOptions({
    title: 'Crop Phenology - Central Iowa (2023)',
    vAxis: {title: 'EVI'},
    hAxis: {title: 'Date'},
    lineWidth: 2, pointSize: 4
  });
print(chart);

What's Growing in This Field? The Cropland Data Layer

Before we can compare crop phenology, we need to know what crop is actually growing in each field. The USDA Cropland Data Layer (CDL) answers that question for us.

Published annually at 30-meter resolution, the CDL classifies every agricultural pixel in the continental United States into specific crop types: corn, soybeans, wheat, cotton, rice, and dozens more. Let's load it up.

// Load the 2023 Cropland Data Layer
var cdl = ee.ImageCollection('USDA/NASS/CDL')
  .filter(ee.Filter.eq('system:index', '2023'))
  .first()
  .select('cropland');

// CDL comes with its own color palette built in
Map.addLayer(cdl, {}, 'Cropland Data Layer 2023');
Map.centerObject(ee.Geometry.Point([-93.5, 42.0]), 11);

// Print the crop type at a specific point
var cropType = cdl.reduceRegion({
  reducer: ee.Reducer.first(),
  geometry: ee.Geometry.Point([-93.5, 42.0]),
  scale: 30,
  maxPixels: 1e9
});
print('Crop type code:', cropType);

Here are some common CDL class codes to know:

• 1: Corn
• 5: Soybeans
• 24: Winter wheat
• 2: Cotton
• 176: Grassland/Pasture

Every Crop Has a Signature

Here's where it gets interesting. Different crops have different phenological signatures. Corn, soybeans, and winter wheat each trace a unique curve through the growing season. It's like a fingerprint: each crop tells its own story over time.

• Corn: Peaks early (mid-July), reaches the highest EVI values (0.8 to 0.9), and senesces rapidly in September.
• Soybeans: Peaks about two weeks later than corn (late July to early August), reaches slightly lower EVI values, and senesces more gradually.
• Winter wheat: Shows a completely different pattern. It greens up in early spring (March to April), peaks in May, and is harvested by late June, well before summer crops reach their maximum.

Let's put all three on the same chart and compare.

// Compare phenology for three crop types
// Use CDL to locate sample points for each crop
var cornPoint = ee.Geometry.Point([-93.50, 42.00]);
var soyPoint = ee.Geometry.Point([-93.45, 42.05]);
var wheatPoint = ee.Geometry.Point([-97.50, 38.80]); // Kansas for winter wheat

// Load MODIS EVI for 2023
var modisEVI = ee.ImageCollection('MODIS/061/MOD13A2')
  .filterDate('2023-01-01', '2023-12-31')
  .select('EVI')
  .map(function(img) {
    return img.multiply(0.0001).copyProperties(img, ['system:time_start']);
  });

// Create a FeatureCollection of crop sample points
var cropPoints = ee.FeatureCollection([
  ee.Feature(cornPoint, {label: 'Corn'}),
  ee.Feature(soyPoint, {label: 'Soybean'}),
  ee.Feature(wheatPoint, {label: 'Winter Wheat'})
]);

// Chart EVI for all three crop types
var chart = ui.Chart.image.seriesByRegion(
  modisEVI, cropPoints, ee.Reducer.mean(), 'EVI', 1000, 'system:time_start', 'label')
  .setOptions({
    title: 'Crop Phenology Comparison (2023)',
    vAxis: {title: 'EVI', viewWindow: {min: 0, max: 1}},
    hAxis: {title: 'Date'},
    lineWidth: 2
  });
print(chart);

Next Steps

• Lab 25: Agricultural Monitoring - Build phenology curves and compare crop types hands-on
• Urban Analysis - Apply remote sensing techniques to built environments