Google Earth Engine (GEE)

for Urban Research

One-day workshop
April 25, 2024

Presented by Anup R. Joshi

What is Urban Sprawl?
• A form of urban growth in which low-density development (such as
single-family homes) of large plots of land takes place over a large
area, extending the urban limits.
• A very common type of development in the US and many other
countries around the world.
• Is often seen as a problem in many parts of the world, due to its
negative effects on the environment, social life, and economy of cities.
How to monitor urban sprawl?

Using satellite data & remote sensing

• Classifying satellite image
• Building spectral indices based on different bands
• Remote sensing of nighttime light emissions
Using satellite images
Landsat 9 Satellite Image (True Color)
• Landsat 9 satellite
Image (Bands
4,3,2)
• 11 bands
• 30 m x 30 m
Sentinel -2 Satellite Image (True Color)
• Sentinel -2
satellite Image
(Bands 4,3,2)
• 12 bands
• Bands 2,3,4 &
8 are 10 m x10
m
• Other bands
20 m x 20 m
resolution
Planet Satellite Basemap

• Daily ~3 m
spatial
resolution
• Produced from
2014 - present
• Has 4 bands
(red, green,
blue, and near
infrared)
Google Earth Engine Satellite Image
Classifying satellite images
• Land use change maps in regular intervals
Classification of satellite imagery to generate land use cover
Identifying land use change between two time periods

Classification

Landsat 2020 Land cover map 2020

 Classified images 2015 & 2020

Identify land
cover change

Land cover map 2015 Land cover map 2020

Using Indices for classification
NDVI, NDBI & NDWI Calculation Using
Landsat
• Landsat data having several bands based on their wavelength (blue
band, green band, red band, infrared band, thermal band,
panchromatic).
• Panchromatic band is used for increase the resolution of data.
• Landsat 7 images have total of 8 band while Landsat 8/9 data have 11
bands.
• For analysis of Normal Difference Vegetation Index (NDVI), Normal
Difference Built-up Index (NDBI) and Normal Difference Water Index
(NDWI), only four bands are used (Green, Red, NIR, SWIR).
Landsat 7 and Landsat 8 data’s bands, wavelength
& their resolution
Indices to identify 3 major land cover classes

• Normalized Difference Vegetation Index (NDVI),

• Normalized Difference Water Index (NDWI)
• Normalized Difference Build up Index (NDBI) respectively
 Spectral reflectance curve:
Spectral reflectance curve shows the relationship between electromagnetic spectrum
(distribution of the continuum of radiant energies plotted either as a function of wavelength or
of frequency) and the associated percent reflectance for any given material.
Normalized Difference Vegetation Index (NDVI):

• Healthy vegetation is good absorber of electromagnetic

spectrum in visible reason.
• Chlorophyll contains in a greeneries highly absorbs
Blue (0.4 - 0.5 µm) and Red (0.6 - 0.7 µm) spectrum
and reflects Green (0.5 – 0.6 µm) spectrum.
• Healthy plants having high reflectance in Near Infrared
(NIR) between 0.7 to 1.3 µm.
• This is primarily due to internal structure of plant
leaves.
• High reflectance in NIR and high absorption in Red
spectrum, these two bands are used to calculate
NDVI.
 Normalized Difference Vegetation Index (NDVI)

NDVI = (NIR – Red) / (NIR + Red)

• For Landsat 7 data, NDVI = (Band 4 – Band 3) / (Band 4 + Band 3)
• For Landsat 8 data, NDVI = (Band 5 – Band 4) / (Band 5 + Band 4)
• The NDVI value varies from -1 to 1. Higher the value of NDVI
reflects high Near Infrared (NIR), means dense greenery. Generally,
we obtain following result:
• NDVI = -1 to 0 represent Water bodies
• NDVI = -0.1 to 0.1 represent Barren rocks, sand, or snow
• NDVI = 0.2 to 0.5 represent Shrubs and grasslands or senescing crops
• NDVI = 0.6 to 1.0 represent Dense vegetation or tropical rainforest
Normalized Difference Built-up Index (NDBI):
NDBI = (SWIR – NIR) / (SWIR +
NIR)
• For Landsat 7 data, NDBI = (Band 5 –
Band 4) / (Band 5 + Band 4)
• For Landsat 8 data, NDBI = (Band 6 –
Band 5) / (Band 6 + Band 5)
• Normalize Difference Build-up Index
value lies between -1 to +1.
• Negative value of NDBI represent
water bodies
• Higher value represent build-up
areas.
• NDBI value for vegetation is low.
 Normalized Difference Water Index (NDWI)
• Water bodies having low reflectance. It only reflects
within visible portion of the electromagnetic spectrum.
• Water bodies in their liquid state are generally high
reflectance on Blue (0.4 - 0.5 µm) spectrum than Green
(0.5 -0.6 µm) and Red (0.6 – 0.7 µm) spectrum.
• Clear water having greatest reflectance in the blue
portion of the visible spectrum. So, water appear blue.
• Turbid water has higher reflectance in visible spectrum.
• There is no reflection in Near Infrared (NIR) and
beyond.
NDWI = (NIR – SWIR) / (NIR + SWIR) • NDWI was developed by Gao(1996) to enhance the
For Landsat 7 data, NDWI = (Band 4 – Band 5) / (Band water related features of the landscapes.
4 + Band 5) • This index uses the near infrared (NIR) and the Short-
For Landsat 8 data, NDWI = (Band 5 – Band 6) / (Band Wave infrared (SWIR) bands.
5 + Band 6)
But result appear from above formula is poor in quality.
The pure water neither reflects NIR nor SWIR.
Modified Normalized Difference Water Index (MNDWI):

The pure water neither reflects NIR nor SWIR. The formula of NDWI then modified
by Xu (2005). It uses Green and SWIR band.

MNDWI = (Green – SWIR) / (Green + SWIR)

• For Landsat 7 data, NDWI = (Band 2 – Band 5) / (Band 2 + Band 5)
• For Landsat 8 data, NDWI = (Band 3 – Band 6) / (Band 3 + Band 6)
• MNDDWI value lies between -1 to 1.
• Generally, water bodies NDWI value is greater than 0.5.
• Vegetation has much smaller values which distinguishing vegetation from water
bodies easily.
• Build-up features having positive values lies between 0 to 0.2.
 Urban Expansion Chapel Hill 1990-2020
based on spectral indices

1990 1995

2000 2005

2010 2015

2020

Built-up map
Landsat images
Nighttime light pollution
Global view of Earth's city lights

Global view of Earth's city lights from a

composite assembled from Day/Night
data acquired by the Suomi National
Polar-orbiting Partnership (Suomi
NPP) satellite. The data were acquired
over nine days in April 2012 and 13
days in October 2012. NASA Earth
Observatory image.
City Lights of the United States 2012

https://www.lightpollutionmap.info/
Remote sensing of nighttime light emissions
• The Visible Infrared Imaging Radiometer Suite (VIIRS)
instruments aboard the joint NASA/NOAA Suomi National Polar-
orbiting Partnership (Suomi NPP) and NOAA-20 satellites provide
global daily measurements of nocturnal visible and near-infrared
(NIR) light that are suitable for Earth system science and applications
studies.
• VIIRS Day/Night Band (DNB) data are used for estimating
population, assessing electrification of remote areas, monitoring
disasters and conflict, and understanding biological impacts of
increased light pollution.
VIIRS nighttime lights – data processing steps
• Annual global VIIRS nighttime lights dataset is a time series produced from
monthly cloud-free average radiance grids.
• An initial filtering step removed sunlit, moonlit and cloudy pixels, leading to
rough composites that contains lights, fires, aurora and background.
• The rough annual composites are made on monthly increments and then
combined to form rough annual composites.
• The subsequent steps uses the twelve-month median radiance to discard high
and low radiance outliers, filtering out most fires and isolating the background.
• Background areas are zeroed out using the data range (DR) calculated
from 3x3 grid cells.
• The DR threshold for background is indexed to cloud-cover levels, with higher
DR thresholds in areas having low numbers of cloud-free coverages.
Nighttime light

https://
www.lightpollutionmap.info
/
 Hands-on exercises
• Exercise1 – Introduction to Google Earth Engine (GEE)
• Exercise 2 – Image visualization in GEE
• Exercise 3 – Supervised classification using GEE to generate
classified maps
• Exercise 4 – Identifying land use change
• Exercise 5 – Identifying built-up areas using spectral indices
• Exercise 6 – GEE script to monitor urban sprawl using Landsat
satellite
• Exercise 7A & 7B - GEE script to extract nighttime lights statistics.
All scripts are available at
• https://code.earthengine.google.com/?accept_repo=users/arjoshi003/
Odum
Google Earth Engine (GEE)
Google Earth Engine – Code Editor
Nighttime light emissions

https://
www.lightpollutionmap.info
/
 Urban Expansion Chapel Hill 1990-2020
based on spectral indices

1990 1995

2000 2005

2010 2015

2020

Built-up map
Landsat images
Urban Expansion Chapel Hill 1990-2020

1990 2000 2010 2020

 NDVI MNDWI NDBI
Dense Vegetation 0.6 to 1 Small Low

Vegetation 0.2 to 0.5 Small Low

Water -1 to 0 >0.5 <0
Soil/Built up -0.1 to 0.1 0 to 0.2 Higher value