Datasets Reference#

Comprehensive guide to the 7 satellite and climate reanalysis platforms supported by ndvi2gif v1.0.0.

Overview#

Platform

Type

Coverage

Resolution

Variables

Best For

Sentinel-2

Optical

2015-present

10-30m

40+ indices

High-res vegetation

Sentinel-3

Optical

2016-present

300m

40+ indices

Large-scale ocean/land

Landsat

Optical

1982-present

30m

40+ indices

Long-term change

MODIS

Optical

2000-present

500m

40+ indices

Daily global coverage

Sentinel-1

SAR

2014-present

10m

7 indices

All-weather monitoring

ERA5-Land

Climate

1950-present

~11km

47 variables

Climate analysis

CHIRPS

Climate

1981-present

~5.5km

1 variable

Precipitation monitoring

Optical Sensors#

Sentinel-2 (S2)#

European Space Agency multispectral imaging mission

processor = NdviSeasonality(
    roi=roi,
    sat='S2',
    index='ndvi',
    start_year=2020,
    end_year=2023
)

Technical Specifications:

  • Satellites: Sentinel-2A (launched 2015), Sentinel-2B (launched 2017)

  • Revisit Time: 5 days (combined constellation)

  • Swath Width: 290 km

  • Spatial Resolution:

    • 10m: Blue, Green, Red, NIR

    • 20m: Red Edge (3 bands), SWIR (2 bands)

    • 60m: Coastal Aerosol, Water Vapor, Cirrus

  • Temporal Coverage: June 2015 - present

  • Earth Engine Collection: COPERNICUS/S2_SR_HARMONIZED

Spectral Bands:

Band

Wavelength (nm)

Resolution

Description

B1

443

60m

Coastal Aerosol

B2

490

10m

Blue

B3

560

10m

Green

B4

665

10m

Red

B5

705

20m

Red Edge 1

B6

740

20m

Red Edge 2

B7

783

20m

Red Edge 3

B8

842

10m

NIR

B8A

865

20m

NIR Narrow

B9

940

60m

Water Vapor

B11

1610

20m

SWIR 1

B12

2190

20m

SWIR 2

Best For:

  • High-resolution vegetation monitoring (10m)

  • Agricultural field-level analysis

  • Red Edge indices for detailed vegetation health

  • Water quality assessment (chlorophyll, turbidity)

  • Urban green space mapping

Advantages:

  • Excellent spatial resolution (10m)

  • Free and open data

  • Frequent revisit (5 days)

  • Red Edge bands for advanced vegetation analysis

  • Active mission with ongoing data collection

Limitations:

  • Shorter temporal archive (2015+)

  • Cloud contamination in tropical/temperate regions

  • Data volume requires cloud processing

Recommended Statistical Methods:

  • key='median': Robust to clouds (recommended)

  • key='percentile', percentile=75: Good cloud-free balance

  • key='max': Maximum greenness/NDVI

Sentinel-3 OLCI (S3)#

Ocean and Land Color Instrument for large-scale monitoring

processor = NdviSeasonality(
    roi=roi,
    sat='S3',
    index='ndvi',
    start_year=2020,
    end_year=2023
)

Technical Specifications:

  • Satellites: Sentinel-3A (2016), Sentinel-3B (2018)

  • Revisit Time: <2 days (combined)

  • Swath Width: 1270 km

  • Spatial Resolution: 300m

  • Temporal Coverage: 2016 - present

  • Earth Engine Collection: COPERNICUS/S3/OLCI

Spectral Bands:

  • 21 bands from 400-1020 nm

  • Optimized for ocean color and land applications

  • Includes atmospheric correction bands

Best For:

  • Regional to global-scale monitoring

  • Coastal and inland water quality

  • Rapid coverage applications

  • Complementing Sentinel-2 for large areas

Advantages:

  • Very high temporal frequency (<2 days)

  • Wide swath (1270 km)

  • Excellent for large-scale monitoring

  • Free and open data

Limitations:

  • Lower spatial resolution (300m)

  • Not suitable for field-level agriculture

  • Fewer band options than Sentinel-2

Landsat (4-9)#

USGS/NASA long-term Earth observation program

processor = NdviSeasonality(
    roi=roi,
    sat='Landsat',
    index='ndvi',
    start_year=1990,  # Long-term analysis
    end_year=2023
)

Technical Specifications:

  • Satellites:

    • Landsat 4-5 ™: 1982-2013

    • Landsat 7 (ETM+): 1999-present (SLC-off after 2003)

    • Landsat 8-9 (OLI): 2013-present

  • Revisit Time: 16 days (8 days with L8+L9)

  • Swath Width: 185 km

  • Spatial Resolution: 30m (15m panchromatic)

  • Temporal Coverage: 1982 - present (40+ years!)

  • Earth Engine Collection:

    • LANDSAT/LC08/C02/T1_L2 (Landsat 8)

    • LANDSAT/LC09/C02/T1_L2 (Landsat 9)

    • LANDSAT/LE07/C02/T1_L2 (Landsat 7)

    • LANDSAT/LT05/C02/T1_L2 (Landsat 5)

    • LANDSAT/LT04/C02/T1_L2 (Landsat 4)

Spectral Bands (Landsat 8/9 OLI):

Band

Wavelength (nm)

Resolution

Description

B1

435-451

30m

Coastal Aerosol

B2

452-512

30m

Blue

B3

533-590

30m

Green

B4

636-673

30m

Red

B5

851-879

30m

NIR

B6

1566-1651

30m

SWIR 1

B7

2107-2294

30m

SWIR 2

Best For:

  • Long-term change detection (1982-present)

  • Historical vegetation trends

  • Decadal climate impact studies

  • Consistent 30m resolution time series

  • Comparison with modern sensors

Advantages:

  • Longest continuous record (40+ years)

  • Free and open data

  • Well-documented and validated

  • Consistent 30m resolution across sensors

  • Global coverage

Limitations:

  • 16-day revisit (less frequent than Sentinel-2)

  • Cloud contamination

  • Landsat 7 SLC-off gaps (2003+)

  • Lower spatial resolution than Sentinel-2

Recommended Use Cases:

  • Deforestation and land use change (1980s-present)

  • Agricultural expansion tracking

  • Urban growth analysis

  • Glacier retreat monitoring

  • Climate change impact assessment

MODIS#

NASA Terra + Aqua daily global coverage

processor = NdviSeasonality(
    roi=roi,
    sat='MODIS',
    index='ndvi',
    start_year=2010,
    end_year=2023,
    key='mean'  # Daily data - mean works well
)

Technical Specifications:

  • Satellites: Terra (2000), Aqua (2002)

  • Revisit Time: 1-2 days (daily global coverage)

  • Swath Width: 2330 km

  • Spatial Resolution: 500m (surface reflectance)

  • Temporal Coverage: 2000 - present

  • Earth Engine Collection: MODIS/061/MCD43A4 (NBAR/BRDF-adjusted)

Spectral Bands:

  • 7 bands optimized for land applications (500m)

  • 36 total bands (various resolutions)

Best For:

  • Global-scale monitoring

  • Daily vegetation dynamics

  • Rapid response applications

  • Coarse-resolution regional analysis

  • Phenology studies (daily resolution)

Advantages:

  • Daily global coverage

  • Long archive (2000+)

  • Excellent temporal resolution

  • Free and open data

  • BRDF-corrected products

Limitations:

  • Coarse spatial resolution (500m)

  • Not suitable for field-level analysis

  • Mixed pixels in heterogeneous landscapes

Recommended Statistical Methods:

  • key='mean': Good for daily data

  • key='max': MVC (Maximum Value Composite) for vegetation

Synthetic Aperture Radar (SAR)#

Sentinel-1 (S1)#

ESA C-band SAR for all-weather monitoring

from ndvi2gif import NdviSeasonality, S1ARDProcessor

# Configure SAR preprocessing
s1_proc = S1ARDProcessor(
    speckle_filter='REFINED_LEE',
    terrain_correction=True,
    dem='COPERNICUS_30'
)

# Use with NdviSeasonality
sar = NdviSeasonality(
    roi=roi,
    sat='S1',
    index='rvi',  # SAR vegetation index
    periods=12,
    start_year=2020,
    end_year=2023
)

Technical Specifications:

  • Satellites: Sentinel-1A (2014), Sentinel-1B (2016-2021, decommissioned)

  • Revisit Time: 6-12 days (S1A only after 2021)

  • Swath Width: 250 km (IW mode)

  • Spatial Resolution: 10m (IW mode)

  • Frequency: C-band (5.405 GHz)

  • Polarizations: VV, VH (dual-pol)

  • Temporal Coverage: 2014 - present

  • Earth Engine Collection: COPERNICUS/S1_GRD

SAR Indices Available:

Index

Formula

Description

RVI

4×VH / (VV+VH)

Radar Vegetation Index

DPSVI

VV+VH

Dual-Polarization SAR Vegetation Index

RFDI

(VV-VH) / (VV+VH)

Radar Forest Degradation Index

VSDI

VV-VH

Vegetation Scattering Difference Index

Best For:

  • All-weather monitoring (clouds, darkness, smoke)

  • Flood mapping and water extent

  • Wetland monitoring

  • Rice paddy detection

  • Soil moisture proxy

  • Forest structure analysis

Advantages:

  • Cloud-independent (microwave penetrates clouds)

  • Day/night acquisition

  • Sensitive to vegetation structure and moisture

  • Free and open data

  • 10m spatial resolution

Limitations:

  • More complex preprocessing (speckle, terrain effects)

  • Less intuitive than optical imagery

  • Geometric distortions in mountains

  • Requires understanding of SAR backscatter physics

Preprocessing Features:

  • Speckle Filtering: REFINED_LEE (recommended), LEE, GAMMA_MAP

  • Terrain Correction: Radiometric and geometric correction

  • Orbit Filtering: ASCENDING/DESCENDING for temporal consistency

Recommended Statistical Methods:

  • key='median': Reduces speckle (recommended)

  • key='mean': Alternative for smoother results

Climate Reanalysis#

ERA5-Land#

ECMWF high-resolution land surface reanalysis

processor = NdviSeasonality(
    roi=roi,
    sat='ERA5',
    index='temperature_2m_celsius',  # Or any of 47 variables
    periods=12,
    start_year=1980,  # Can go back to 1950!
    end_year=2023,
    key='mean'  # For temperature
)

Technical Specifications:

  • Provider: European Centre for Medium-Range Weather Forecasts (ECMWF)

  • Temporal Resolution: Daily aggregates (from hourly data)

  • Spatial Resolution: ~11 km (0.1° grid)

  • Temporal Coverage: 1950 - present (75+ years!)

  • Global Coverage: Complete global land surface

  • Earth Engine Collection: ECMWF/ERA5_LAND/DAILY_AGGR

  • Update Frequency: ~5 days behind real-time

Variable Categories (47 total):

Temperature (8 core + 16 variants = 24 total):

  • temperature_2m - Air temperature at 2m (Kelvin)

  • temperature_2m_celsius - Auto-converted to °C

  • temperature_2m_min / temperature_2m_max - Daily extremes

  • temperature_2m_min_celsius / temperature_2m_max_celsius - Extremes in °C

  • dewpoint_temperature_2m (+ celsius/min/max variants)

  • skin_temperature (+ celsius/min/max variants)

  • soil_temperature_level_1 (+ celsius/min/max variants)

Precipitation & Water Balance (5 core + 6 unit variants = 11 total):

  • total_precipitation_sum - Total precipitation (meters)

  • total_precipitation_sum_lm2 - In liters/m² (mm)

  • total_evaporation_sum (+ _lm2 variant)

  • potential_evaporation_sum (+ _lm2 variant)

  • runoff_sum (+ _lm2 variant)

  • surface_runoff_sum (+ _lm2 variant)

  • snowfall_sum (+ _lm2 variant)

Soil Moisture (4 layers):

  • volumetric_soil_water_layer_1 - 0-7 cm depth

  • volumetric_soil_water_layer_2 - 7-28 cm depth

  • volumetric_soil_water_layer_3 - 28-100 cm depth

  • volumetric_soil_water_layer_4 - 100-289 cm depth

Radiation (3 variables):

  • surface_solar_radiation_downwards_sum

  • surface_net_solar_radiation_sum

  • surface_latent_heat_flux_sum

Wind & Pressure (3 variables):

  • u_component_of_wind_10m - East-west wind (m/s)

  • v_component_of_wind_10m - North-south wind (m/s)

  • surface_pressure - Atmospheric pressure (Pa)

Snow (2 variables):

  • snow_depth_water_equivalent - Snow water content (m)

  • snowfall_sum - Daily snowfall (m)

Best For:

  • Long-term climate analysis (1950-2023)

  • Temperature trend detection

  • Precipitation patterns and drought

  • Soil moisture dynamics

  • Climate-vegetation interactions

  • Heat/cold wave identification

  • Water balance studies

Advantages:

  • Exceptional temporal depth (1950+)

  • Gap-free: No missing data (modeled reanalysis)

  • 47 climate variables in one dataset

  • Daily resolution

  • Celsius conversion built-in

  • Global coverage

  • Free access

Limitations:

  • Coarse resolution (~11 km)

  • Not observed data (reanalysis/model)

  • Some variables have known biases

  • Evaporation bands have data issues (swapped values)

Recommended Statistical Methods:

  • Temperature: key='mean' (daily average)

  • Temperature Extremes: key='max' or key='min'

  • Precipitation: key='sum' (monthly/seasonal totals)

  • Soil Moisture: key='mean'

  • Radiation: key='sum' (energy accumulation)

Usage Tips:

# Temperature in Celsius (automatic conversion)
temp = NdviSeasonality(
    roi=roi,
    sat='ERA5',
    index='temperature_2m_celsius',  # ← Kelvin → Celsius
    key='mean'
)

# Precipitation in mm (L/m²)
precip = NdviSeasonality(
    roi=roi,
    sat='ERA5',
    index='total_precipitation_sum_lm2',  # ← meters → L/m²
    key='sum'  # Sum for totals
)

# Daily temperature extremes
temp_max = NdviSeasonality(
    roi=roi,
    sat='ERA5',
    index='temperature_2m_max_celsius',  # Daily max
    key='max'  # Maximum of daily maxima
)

CHIRPS#

Climate Hazards Group InfraRed Precipitation with Station data

processor = NdviSeasonality(
    roi=roi,
    sat='CHIRPS',
    index='precipitation',
    periods=12,
    start_year=2010,
    end_year=2023,
    key='sum'  # Monthly totals
)

Technical Specifications:

  • Provider: UC Santa Barbara Climate Hazards Center

  • Temporal Resolution: Daily

  • Spatial Resolution: ~5.5 km (0.05°)

  • Temporal Coverage: 1981 - present (40+ years)

  • Coverage: 50°S to 50°N (quasi-global tropics/subtropics)

  • Earth Engine Collection: UCSB-CHG/CHIRPS/DAILY

  • Update Frequency: ~3 weeks behind real-time

Variable:

  • precipitation - Daily precipitation (mm/day)

Data Sources:

  • Satellite infrared cold cloud duration (IR-CCD)

  • In-situ station observations (blended)

  • TRMM precipitation estimates (calibration)

Best For:

  • High-resolution precipitation monitoring

  • Drought early warning systems

  • Agricultural rainfall assessment

  • Hydrological modeling inputs

  • Tropical/subtropical precipitation analysis

  • Comparison with ERA5 precipitation

Advantages:

  • Higher resolution than ERA5 (~5.5 km vs ~11 km)

  • Station-calibrated (better accuracy than satellite-only)

  • Specifically designed for drought monitoring

  • Long archive (1981+)

  • Quasi-global tropical coverage

  • Free access

Limitations:

  • Single variable (precipitation only)

  • Limited to 50°S - 50°N (no polar regions)

  • ~3-week latency

  • No other climate variables

Recommended Statistical Methods:

  • key='sum': Monthly/seasonal precipitation totals (recommended)

  • key='mean': Average daily rainfall

Comparison: CHIRPS vs ERA5 Precipitation

Aspect

CHIRPS

ERA5

Resolution

~5.5 km

~11 km

Method

Satellite + stations

Reanalysis model

Coverage

50°S-50°N

Global

Accuracy

High (station blend)

Moderate (model)

Variables

Precipitation only

47 variables

Latency

~3 weeks

~5 days

Usage Example:

# CHIRPS for detailed precipitation
precip_chirps = NdviSeasonality(
    roi=roi,
    sat='CHIRPS',
    index='precipitation',
    periods=12,
    start_year=2015,
    end_year=2023,
    key='sum'  # Monthly totals
)

# Compare with ERA5
precip_era5 = NdviSeasonality(
    roi=roi,
    sat='ERA5',
    index='total_precipitation_sum_lm2',
    periods=12,
    start_year=2015,
    end_year=2023,
    key='sum'
)

Dataset Selection Guide#

By Application#

Agricultural Monitoring:

  • Field-level (< 50ha): Sentinel-2 (10m)

  • Regional: Landsat (30m) or MODIS (500m)

  • All-weather: Sentinel-1 SAR

  • Climate context: ERA5 (temperature, precipitation, soil moisture)

Forest/Vegetation Change Detection:

  • High-detail: Sentinel-2 (10m, 2015+)

  • Long-term trends: Landsat (30m, 1982+)

  • Rapid monitoring: MODIS (daily)

  • Forest structure: Sentinel-1 SAR

Water Quality & Wetlands:

  • High-resolution: Sentinel-2 (chlorophyll, turbidity)

  • Large water bodies: Sentinel-3 OLCI

  • Water extent: Sentinel-1 SAR (all-weather)

  • Hydrological drivers: ERA5 (precipitation), CHIRPS

Climate & Drought Analysis:

  • Long-term trends: ERA5-Land (1950+)

  • High-res precipitation: CHIRPS (1981+)

  • Vegetation response: Sentinel-2 or Landsat NDVI

  • Soil moisture: ERA5 volumetric soil water

Phenology & Seasonality:

  • Daily dynamics: MODIS (500m)

  • Detailed phenology: Sentinel-2 (10m, 5-day)

  • Multi-decadal: Landsat (30m, 40+ years)

  • Climate drivers: ERA5 temperature + precipitation

By Temporal Coverage#

Need

Dataset(s)

1950-1980

ERA5-Land only

1981-2000

ERA5-Land + CHIRPS + Landsat 4-5

2000-2015

All except S2/S3

2015-present

All 7 platforms

By Resolution#

Resolution

Dataset(s)

Best Use

10m

Sentinel-1, Sentinel-2

Field-level precision

30m

Landsat

Regional monitoring

300m

Sentinel-3

Ocean/large-scale land

500m

MODIS

Global/continental

5.5 km

CHIRPS

Regional precipitation

11 km

ERA5-Land

Climate analysis

Earth Engine Collection IDs#

For direct Earth Engine access:

import ee

# Optical
s2 = ee.ImageCollection('COPERNICUS/S2_SR_HARMONIZED')
s3 = ee.ImageCollection('COPERNICUS/S3/OLCI')
l8 = ee.ImageCollection('LANDSAT/LC08/C02/T1_L2')
l9 = ee.ImageCollection('LANDSAT/LC09/C02/T1_L2')
l7 = ee.ImageCollection('LANDSAT/LE07/C02/T1_L2')
l5 = ee.ImageCollection('LANDSAT/LT05/C02/T1_L2')
modis = ee.ImageCollection('MODIS/061/MCD43A4')

# SAR
s1 = ee.ImageCollection('COPERNICUS/S1_GRD')

# Climate
era5 = ee.ImageCollection('ECMWF/ERA5_LAND/DAILY_AGGR')
chirps = ee.ImageCollection('UCSB-CHG/CHIRPS/DAILY')

References#

See Also#

Edit on GitHub
Show Source