Sentinel-2 Data and Vegetation Indices

Module 1

Introduction to spectral indices

Spectral Indices Background

Spectral indices are used to enhance particular land surface features or properties, e.g. vegetation, soil, water.

There are developed based on the spectral properties of the object of interest.

Spectral Indices Background

Spectral signatures of clean/turbid water (in reflectance)

Source: Ma, M., et al. (2007) Change in area of Ebinur Lake during the 1998–2005 period. International Journal of Remote Sensing, 28(24), 5523-5533.

Spectral Indices Background

Spectral signatures of snow and clouds (in reflectance)

Source: Dong, C. (2018) Remote sensing, hydrological modelling and in situ observations in snow cover research: A review. Journal of Hydrology, 561, 573-583.

Spectral Vegetation Signatures

In green healthy plant chlorophyll absorbs large proportion of red and blue spectrum for photosynthesis and strongly reflects in green strong reflectance in near infrared (NIR) due to leaf structure and condition lower reflectance in shortwave infrared (SWIR) influenced by water content, which absorbs infrared energy.

Spectral Vegetation Signatures

Dominant factor controlling leaf reflectance.

Spectral Vegetation Signatures

Monitoring vegetation disease

Rational for Vegetation Indices

To explore and highlight the unique spectral signatures of vegetation, which allows to delineate it from the other earth objects.

To delineate the subtle changes in the spectral signatures of vegetation caused by changes in plant vigour and health that cannot be distinguished by human eye.

Spectral vegetation indices have been found to be related to various biophysical parameters, i.e. Leaf Area Index (LAI), percent vegetation cover, fraction of absorbed photosynthetically active radiation (fAPAR), photosynthetic capacity, and carbon dioxide fluxes.

How to calculate spectral indices?

There are calculated using a mathematical equation applied to two or more wavelengths across the optical spectrum.

Spectral indices varied from a simple spectral rationing to more complicated combination of multispectral bands.

They are used to combine the multiple spectral bands into a single image, which highlights particular land surface features or properties.

Sentinel-2 Spectral Bands

Combination of multispectral bands

NDVI - Normalized Difference Vegetation Index

Effective for quantifying green vegetation. Positively correlated with vegetation greenness. NDVI = (NIR – Red) / (NIR + R) For Sentinel-2 the formula is: (B8 - B4) / (B8 + B4) where: B8 = 842 nm, B4 = 665 nm NDVI range value is -1 to 1

Source: Wu Ch-D., et al. (2014)

Examples of the NDVI related indices

TNDVI - Transformed Normalized Difference Vegetation Index, indicates a relation between the amount of green biomass that is found in a pixel. It has always positive values and the variances of the ratio are proportional to mean values (Senseman et al. 1996).

TNDVI = (sqrt(NDVI + 0.5))

NDI45M - Normalized Difference Index (Delegido et al. 2011), is more linear, with less saturation at higher values than the NDVI

NDI45 = (NIR – R) / (NIR + R)

For Sentinel-2 the formula is: (B5 - B4) / (B5 + B4), where: B5 = 705 nm, B4 = 665 nm

Examples of the NDVI related indices

Vegetation Indices Issues

Vegetation Moisture Sensitive Indices

The moisture sensitive indices are calculated using SWIR and NIR.

The SWIR bands are sensitive to vegetation water content and spongy mesophyll structure in the vegetation canopy. The NIR reflectance is affected by leaf internal structure and leaf dry matter content but not by water content.

There are used to assess the vegetation moisture decline that is particularly useful for drought monitoring.

There are widely applied in agricultural and ecological applications including surface water body characteristics, vegetation water stress, soil water content assessment and wetlands monitoring.

Sentinel-2 Indices

Large collection of Sentinel-2 spectral indices and Java scripts available on https://custom-scripts.sentinel-hub.com/custom-scripts/sentinel-2/indexdb/

Module 2

Application of vegetation spectral indicess

Agriculture

Analysing the crop health and vigour

Variation of maize condition depending on drought conditions. The condition of maize, expressed by the vegetation index - NDVI, calculated using Sentinel-2 data compared with the maps of agricultural droughts for the same decades of the year, calculated on the basis of NOAA AVHRR data.

© IGIK

Agriculture

Using multi-temporal vegetation indices can help management of individual fields during the growing season. They can also be used for within-field monitoring of growth and application of fertilizers.

Forestry

Analysing the crop health and vigour

Forest health and condition Combination of different spectral Spectral-2 indices (NDVI, NDMI and PSSRa) allows to assess and monitor forest health condition over time.

© IGIK

Forestry

Multi-temporal Sentinel-2 vegetation indices are applied to monitor bark beetle outbreak in Białowieża National Park - Poland

DSWI - Disease Water Stress Index sensitive to stress due to water shortage and plant damage DSWI = (NIR-GREEN) / (SWIR1 + RED)

Source: Hoscilo et al. (2016)

Vineyards

Monitoring condition of vineyards using NDVI (April-September 2019) Vineyards Srebrna Góra, Poland Sentinel-2

© IGIK & VINUM 4.0 Sp z o.o

Vineyards

Monitoring condition of vineyards using Sentinel-2 moisture index (NDMI) (April-September 2019) Vineyards Srebrna Góra, Poland NDMI

© IGIK & VINUM 4.0 Sp z o.o.

Natural hazards - Fire

Mapping burnt areas with Normalized Burnt Ratio index (NBR). Example from forest fire 2014 in Sala, Sweden. Sentinel-2 image

© Geografiska informationsbyrån

Natural hazards - Windstorm

Assessment of forest damage caused by the windstorm 11-12 August 2017 in Poland, using Sentinel-2 NDMI - Normalized Difference Moisture Index NDMI = (NIR−SWIR) / (NIR+SWIR) NIR (band 8) = 842 nm SWIR (band 11) = 1610 nm

Source: Hoscilo and Lewandowska (2018)