National Aeronautics and Space Administration

Hyperspectral Data for Land and Coastal Systems

Amber Jean McCullum, Juan L. Torres-Pérez, and Zach Bengtsson

January 19, 2021

NASA’s Applied Remote Sensing Training Program (ARSET)
https://appliedsciences.nasa.gov/what-we-do/capacity-building/arset/about-arset

• Part of NASA’s Applied Sciences

Program
• Empowering the global community Water Resources
through remote sensing training
• Seeks to increase the use of Earth
science in decision-making through
training for:
Disasters
– Policy makers
Air Quality
– Environmental managers
– Other professionals in the public
and private sector
Eco

NASA’s Applied Remote Sensing Training Program 2

Course Structure and Materials

• Three, 1.5-hour sessions on January 19, January

26, and February 2
• The same content will be presented at two
different times each day:
– Session A: 11:00-12:30 EST (UTC-5)
– Session B: 16:00-17:30 EST (UTC-5)
– Please only sign up for and attend one
session per day.
• Webinar recordings, PowerPoint presentations,
and the homework assignment can be found
after each session at:
– https://appliedsciences.nasa.gov/join-
mission/training/english/hyperspectral-data-
land-and-coastal-systems
• Q&A following each lecture and/or by email at:
• [email protected]
• [email protected] or
• [email protected]

NASA’s Applied Remote Sensing Training Program 3

Prerequisites

• Prerequisites:
– Please complete Fundamentals of
Remote Sensing or have equivalent
experience.
• Course Materials:
– https://appliedsciences.nasa.gov/j
oin-
mission/training/english/fundament
als-remote-sensing

NASA’s Applied Remote Sensing Training Program 4

Homework and Certificates

• Homework:
– One homework assignment
– Answers must be submitted via Google
Forms
– HW Deadline: Tuesday February 16

• Certificate of Completion:
– Attend all live webinars
– Complete the homework assignment by the deadline (access from ARSET
website)
– You will receive certificates approximately three months after the completion
of the course from: [email protected]

NASA’s Applied Remote Sensing Training Program 5

Course Outline

Session 1: Overview of Session 2: Hyperspectral Session 3: Hyperspectral

Hyperspectral Data Data for Land Data for Coastal and
Management Ocean Systems

NASA’s Applied Remote Sensing Training Program 6

Learning Objectives

• By the end of this session, you will be

able to…
– Recognize hyperspectral data and
how it differs from multispectral
data
– Identify current hyperspectral
sensors and future hyperspectral
satellite missions of interest
– Locate available hyperspectral
data and identify data processing
platforms
Image Credit: NASA JPL

NASA’s Applied Remote Sensing Training Program 7

Introduction to Hyperspectral Data
NASA’s Applied Remote Sensing Training Program 9
Spectral Resolution

• The ability of a
sensor to define
wavelength
intervals.
• Each “band”
represents a
different part of the
electromagnetic
spectrum.
• The finer the
spectral resolution,
the narrower the
wavelength range
for a particular
channel or band.

Image Credit: Cristina Milesi

NASA’s Applied Remote Sensing Training Program 10

Spectral Resolution

• Each image band is a different layer in an image.

Image credit: USGS

NASA’s Applied Remote Sensing Training Program 11

What is hyperspectral remote sensing?

• The acquisition of imagery in

hundreds of contiguous
spectral bands such that a
radiant spectrum can be
derived for each pixel.
– Measuring reflectance at
close intervals on the
electromagnetic spectrum
– Bands are usually spaced 10
nm or less from one another Image Credit: NSF NEON

NASA’s Applied Remote Sensing Training Program 12

Multispectral vs. Hyperspectral Data

Multispectral Hyperspectral

Image Credit: CIRES, University of Colorado

NASA’s Applied Remote Sensing Training Program 13

Multispectral vs. Hyperspectral Data

Spectrum representation including: (A) Multispectral example, with 5 wide bands; and (B)
Hyperspectral example consisting of several narrow bands. Some hyperspectral sensors
have several hundred bands. Image Credit: Adao, et al., 2017

NASA’s Applied Remote Sensing Training Program 14

Multispectral vs. Hyperspectral Data

Multispectral: Hyperspectral:
– Popular sensors – Limited in numbers of
such as Landsat satellite sensors
and MODIS – Some mission-specific
sensors aboard the ISS
– Limited number of
– Airborne sensors flown
spectral bands
during flight campaigns
– Relatively high – Narrow bands that
temporal resolution measure more
– Global spatial characteristics of surface
extent reflectance
– Low temporal resolution
– Less spatial coverage

NASA’s Applied Remote Sensing Training Program 15

Why is spectral resolution important?
Percent Reflectance

60

Pinewoods
Grasslands

20 Red Sand Pit

Silty Water
0.4 0.6 0.8 1.0 1.2
Wavelength (µm) Adapted from an image from the Indian Institute of Science.

NASA’s Applied Remote Sensing Training Program 16

Applications of Hyperspectral Imagery

- Geology
- Invasive Species
- Coastal/Ocean monitoring
Left: The island of Fordate in Tanimbar
- Carbon monitoring (Indonesia) by CASI hyperspectral data
- Microbial life in the Arctic (resolution 2,5 m). Right: The bottom-
- Volcanic Activity type classification map made from the
hyperspectral data. Image Credit: SEOS

NASA’s Applied Remote Sensing Training Program 17

Applications of Hyperspectral Imagery

This graph compares the reflectance of hematite (an iron ore) with malachite and chrysocolla (copper-
rich minerals) from 200 to 3,000 nanometers. Image Credit: NASA/USGS by Robert Simmon.

NASA’s Applied Remote Sensing Training Program 18

Satellite & Airborne-Based Hyperspectral Imagers
NASA Hyperspectral Imagers

Satellite/Space-Based Airborne
– EO-1 Hyperion – Airborne Visible/Infrared
Imaging Spectrometer
– Test missions onboard the (AVIRIS)
International Space Station • Campaigns flown
(ISS): around the world
• Hyperspectral Imager for • Proof of concept for
the Coastal Ocean future hyperspectral
(HICO) satellite sensors
• ECOsystem Spaceborne – Portable Remote Imagining
Thermal Radiometer Spectrometer (PRISM)
Experiment on Space • COral Reef Airborne
Station (ECOSTRESS)* Laboratory (CORAL)
– Thermal imagery • Mission flown 2016-2019
to analyze coastal
prototype for future ecosystems and reef
hyperspectral missions conditions

NASA’s Applied Remote Sensing Training Program 20

EO-1 Hyperion

• Date Range: 2000-2017

• 220 spectral bands
• 357 to 2567 nm
• 10 nm bandwidth
• 30 m spatial resolution
• 7.75 km swath
• 12-bit

Hyperion image of Mount. Fuji, 2000 (left), the

Hyperion sensor (top), and a comparison of the
Landsat bands and a spectra from Hyperion.
Image Credit: USGS

NASA’s Applied Remote Sensing Training Program 21

EO-1 Hyperion

Hyperion Images of
outcrops near one of
the largest mining
operations in Jordan
(Khirbat en-Nahas)

With a natural color

image (top) the
minerals appear
uniformly dark, but
different rock types
can be identified with
the many spectral
bands in Hyperion
(bottom).
Image Credit: Robert Simmon and NASA

NASA’s Applied Remote Sensing Training Program 22

Hyperspectral Imager for Coastal Ocean (HICO)
• First spaceborne imaging spectrometer designed to sample the
coastal ocean
– Onboard the International Space Station (ISS)
• Date Range: 2009-2014
• Spatial Resolution: 90 m
• Spectral Resolution: 128 bands (400-900nm every 5.7nm)
• Temporal Resolution: ~3 days
HICO image of a
massive Microcystis
bloom in western
Lake Erie, Sept. 3,
2011. Image Credit:
NASA

NASA’s Applied Remote Sensing Training Program 23

HICO Image Credit: NASA

• Bermuda, August
2013
• This animation
displays all 128
HICO bands, 3 at a
time, to produce
color.
• Island
characteristics,
shallow water
components, and
coral signatures
can be examined.

NASA’s Applied Remote Sensing Training Program 24

ECOSTRESS

• Onboard the ISS

• Date Range: Aug 2018-
Present
• Spatial Resolution: 70 m
• Spectral Resolution: 6
bands (160-1200 nm)
• Range: 53.6° N latitude
to 53.6° S latitude
• Find Data At: Data
Pool, NASA Earthdata
Search, AppEEARS,
and USGS EarthExplorer

Image Credit: NASA

NASA’s Applied Remote Sensing Training Program 25
ECOSTRESS

ECOSTRESS data displaying

evaporative stress from 2019
(top), 2020 (middle), and
percent change in
evaporative stress from 2019-
2020 (bottom) near
Farmington, NM in the NAPI
farms region.

Blue colors represent low stress

and high water use, whereas
red colors represent high stress
and low water use.
Image
Credit:
NASA

NASA’s Applied Remote Sensing Training Program 26

Airborne Visible/Infrared Imaging Spectrometer (AVIRIS)

• Flown on four aircraft platforms: NASA's

ER-2 jet, Twin Otter International’s
Turboprop, Scaled Composites' Proteus,
and NASA's WB-57
• Flown in North America, Europe, portions
of South America, and Argentina
• 224 continuous spectral bands
• 400 to 2500 nm
• Bandwidth: < 10 nm
• Objective: Identify, measure, and monitor
constituents of Earth’s surface and
atmosphere based on molecular
absorption and particle scattering
signatures
Image Credit: Lu et al., 2020

NASA’s Applied Remote Sensing Training Program 27

Airborne Visible/Infrared Imaging Spectrometer (AVIRIS)

Hemlock health
monitoring in Catskill
State Park using AVRIS
imagery from July 2001

Hemlocks in this region

are prone to elongate
hemlock scale, Fiorinia
externa Ferris, which
can result in branch
dieback and tree
mortality.

Image Credit: Hanavan et al, 2015

NASA’s Applied Remote Sensing Training Program 28

COral Reef Airborne Laboratory (CORAL)
• Airborne mission flown using the Portable Remote Imaging
Spectrometer (PRISM) to evaluate health and conditions of coral
reef ecosystems
• Date Range: 2016-2019
• Spectral Resolution: 349.9-1053.5 nm (3.5 nm sampling)

Image
Credit:
NASA

NASA’s Applied Remote Sensing Training Program 29

COral Reef Airborne Laboratory (CORAL)

Six sub-campaigns near the Mariana Islands, Palau,

portions of the Great Barrier Reef, and the Hawaiian
Islands (top). CORAL image and classification (right)
from the French Frigate Shoals in northwestern Hawaii.
Image Credit: NASA

NASA’s Applied Remote Sensing Training Program 30

Additional Hyperspectral Missions
Instrument TianGong-1 PRISMA HISUI EnMAP SHALOM HypXIM
Organization Chinese Italian Japanese German Italy-Israel France
Academy Space ministry of GFZ-DLR Space Space
of Science Agency Economy, agencies Agency
and Physics (ASI) Trade, and (ASI-ISA) (CNES)
Industry
Date Range 2011-2013 2020- 2021 2021 2022 2021/2022
Present
Spectral 400-2500 nm 400-2500 400-2500 420-2450 400-2500 400-2500
Range nm nm nm nm nm

Spectral 128 249 185 244 275 210

Bands
Spatial 10-20 m 30 m 30 m 30 m 10 m 10 m
Resolution
Objective Land Natural Energy, Earth Land and Soil, urban,
imaging in resources vegetation observation ocean coastal
China and monitoring observation applications
atmospher
e
NASA’s Applied Remote Sensing Training Program 31
Future NASA Hyperspectral Satellite Initiatives
• Plankton, Aerosol, Cloud, and Ocean Ecosystem (PACE)
─ Observations of the global oceans, atmosphere, and terrestrial ecosystems
─ Ultraviolet through the visible and into the shortwave infrared region of the
electromagnetic spectrum, specifically from 340-890 nm sampled at every 2.5
nm with 5 nm resolution
• Surface Biology and Geology (SBG)
– Applications across a variety of focus areas
– Precursor to SBG: Hyperspectral Infrared Imager (HyspIRI) mission concept
activity (2007-2018)
– Imaging spectrometer measuring from the visible to short wave infrared
(VSWIR: 380 nm - 2500 nm) in 10 nm contiguous bands
• Geosynchronous Littoral Imaging and Monitoring Radiometer (GLIMR)
– Observations of ocean biology, chemistry, and ecology in the Gulf of Mexico,
portions of the Southeastern US coastline, and the Amazon River plume
– Hyperspectral ocean color radiometer

NASA’s Applied Remote Sensing Training Program 32

Plankton, Aerosol, Cloud, and Ocean Ecosystem (PACE)
https://pace.gsfc.nasa.gov/
• PACE is NASA’s next great investment in
hyperspectral earth imagery and multi-angle
polarimetry.
– Launch Date: 2023
– 3-year design life; 10-year propellant
• Hyperspectral Imager: Ocean Color Instrument
(OCI)
– Spectral Resolution: UV to SWIR (340-890 nm
every 2.5 nm, with 940, 1038, 1250, 1378, 1615,
2130, & 2250 nm)
– Temporal Resolution: 2 days
– Spatial Resolution: 1-km2 at nadir
• Two Multi-Angle Polarimeters
– HARP-2: Wide swath, hyper-angular, 4 bands
across the VIS & NIR
– SPEXone: Narrow swath, hyperspectral (UV-
NIR), 5 viewing angles

NASA’s Applied Remote Sensing Training Program 33

PACE Applications Program

Air Quality Water Resources Disasters

Ecological Forecasting Climate

NASA’s Applied Remote Sensing Training Program 34

Surface Biology and Geology (SBG) Mission
https://sbg.jpl.nasa.gov/

• In development via guidance from the 2018 Decadal Survey

• Potential Parameters:
– Visible to Shortwave Infrared Bands:
• Spectral Range: 350 or 400 to 2500
nanometers
• Spectral Resolution: 10 nm or better
• Global with 2- to16-day revisit times
– Thermal Bands:
• Spectral Range: 800 to1200 or 300 to
500 nanometers
• Spectral Resolution: Greater than 5
bands
• Global with 1- to 70-day revisit times Mount Kilimanjaro Image Credit: JPL SBG

NASA’s Applied Remote Sensing Training Program 35

SBG Mission

SBG Potential Applications; Image Credit: JPL SBG

NASA’s Applied Remote Sensing Training Program 36

SBG Mission

• Get involved with SBG! https://sbg.jpl.nasa.gov/news-events

Thomas Fire; Image Credit: JPL SBG

Larsen-C Ice Shelf; Image Credit: JPL SBG

NASA’s Applied Remote Sensing Training Program 37

Assessing & Processing Hyperspectral Data
Hyperspectral Data Access

• Some hyperspectral data is

available on websites you may
already be familiar with.
– USGS EarthExplorer
– USGS GloVis
– NASA EarthData
– Google Earth Engine

EarthExplorer User Interface; Image Credit: USGS

NASA’s Applied Remote Sensing Training Program 39

Hyperspectral Data Access

• Data are also available

through the NASA Distributed
Active Archive Centers
(DAACs).
– Data is separated by
application area, such as
Land Processes and
Ocean Biology.
– You can find AVIRIS,
HICO, CORAL and other
hyperspectral datasets
through the online
resources of DAACs. Image Credit: USGS

NASA’s Applied Remote Sensing Training Program 40

Data Availability Considerations

Satellite Test Mission Data Airborne Data

– Data access might be – Limited to flight paths set by
limited to specific mission objectives.
geographic region. – Temporal resolution is often
– Raw data may need to be seasonal or yearly.
processed by a NASA – Campaigns can be limited to
scientist. just a few years or a single
– Contact with mission or test mission.
program scientists may be – Many research flights were
necessary to access grounded in 2020 due to
appropriately pre-processed COVID.
data for your region.

NASA’s Applied Remote Sensing Training Program 41

Hyperspectral Data Processing
• Data is available at different
processing levels depending on
the sensor.
– Level 1: Radiance
– Level 2: Surface Reflectance
• Atmospheric correction must be
applied.
• Dimensionality Reduction
Techniques:
– Principal Component Analysis Image
(PCA)
– Minimum Noise Fraction (MNF) Placeholder
• Processing and analysis can be
conducted in:
– ArcGIS
– QGIS
– ENVI
– Erdas Imagine
– Google Earth Engine False color composites (FCCs) of Landsat and Hyperion data.
– R Hyperspectral data provides many possibilities and combinations of
– Python FCCs as illustrated here for a few combinations using some of the
Hyperion bands. Image Credit: USGS
NASA’s Applied Remote Sensing Training Program 42
Data Processing Considerations

• Large data files

– 100-250 bands
– Increased storage and
processing power
needed
• Bands may display similar
reflectance properties/high
correlation among bands
– May be difficult to
determine which bands
are most appropriate for
the specific application
• May be high signal to noise
ratios Noise reduction in hyperspectral imagery. Image Credit: Rasti et al., 2018

NASA’s Applied Remote Sensing Training Program 43

Summary

• Hyperspectral Data: Generally, hundreds of contiguous spectral bands

such that a radiant spectrum can be derived for each pixel.
– Measuring reflectance at close intervals on the electromagnetic
spectrum
– Bands are usually spaced 10 nm or less from one another
• Multiple satellite and airborne sensors (Hyperion, AVIRIS, etc.)
• Data Considerations:
– Benefits: Ability to differentiate different vegetation types, minerals,
drought indicators, etc.
– Limitations: Large datasets, potential for large signal to noise ratio,
regional not global
• Next Two Sessions: Applications of Hyperspectral Data for Land and
Coastal/Ocean Systems

NASA’s Applied Remote Sensing Training Program 44

 Thank You!

NASA’s Applied Remote Sensing Training Program 45