1

Presentation to Landsat Science Team, June 1, 2017

Dennis Helder, USGS EROS Senior Calibration Advisor

EROS Cal/Val Center of Excellence

2

3

Why is EROS doing this?

USGS EROS is closely connected to remote sensing data users.

USGS EROS already has the expertise and is a recognized leader in the calibration of Landsat remote sensing imagery.

As Landsat is looked upon as a reference sensor and data archive for much of the remote sensing industry, a need exists to formalize an entity responsible for elucidating quality standards of, and reference methodologies for, relating to Landsat sensors and data in a way that is extensible to the broader community.

Given the second point, a need exists for an entity to play a major role in the development of standards and procedures for calibration/validation in support of the broader community as well as being consistent with USGS needs.

As the number of remote sensing instruments [small sats] and data sets continues to grow, it becomes increasingly important for a stable reference point to exist that is aware of the heritage of calibrated remote sensing products and also responsive to the new opportunities continually be developed.

4

ECCOE Vision

The EROS Calibration Center of Excellence (ECCOE) is recognized nationally and internationally as a leading organization for effectively improving the accuracy, precision, and efficiency of radiometric, geometric, and spatial characterization, calibration, and cross-calibration of optical remote sensing systems to achieve the highest degree of interoperability of remote sensing data products."

5

ECCOE Mission

Develop advanced methods to perform radiometric/geometric/spatial characterization and calibration of optical remote sensing systems to improve the accuracy and precision of all derived data products.

Develop new and improved methods for the cross-calibration of optical remote sensing systems for continuous improvement of the interoperability of remote sensing data products.

Provide the expertise and advocacy necessary for the acquisition of new measurements that continue to improve optical remote sensing data quality, coverage, and expanded usefulness for science applications.

Develop mutually beneficial partnerships within USGS/DOI, across other government agencies, and throughout the industry.

6

ECCOE Key Aspects/Attributes/Activities

Consistent with and supportive of USGS EROS mission:

“We do land change science using land remote sensing data…”

More than just the Landsat system

Leverage existing expertise

Focus on Optical Land Remote Sensing Systems

Collaborative and Distributed

Research

Publish!

Data emphasis with strong connections to the Instruments

PR—experts, customers, public, congress

Strong connection to science

7

Generic ECCOE Structure

USGS EROS CalVal

Center Of Excellence

Affiliates

Science

Interface

Panel

Collab

orators

Other

Govt. Cal

Groups

USGS

EROS

CEOS WGCV IVOS

Government calibration groups that

work closely with the ECCOE.

Affiliates

Other

Govt. Cal

Groups Organizations that are funded by

ECCOE to conduct CalVal research.

Collab

orators

Organizations and gov’t. entities that

interact with ECCOE to conduct

CalVal activities.

Science

Interface

Panel

External panel of experts that meets

at least annually to review ECCOE

activities and recommend future

directions for CalVal work.

CEOS

WGCV

IVOS

International CalVal organization that

interacts synergistically with ECCOE.

JACIE

8

Current ECCOE Structure (May 2017)

USGS EROS CalVal Center Of Excellence

SDSU

JPL

RIT

UAZ

Affiliates

DG PL Collab

orators I2R

NASA

GSFC

LPSO

NOAA

NCC

USGS

EROS

CEOS WGCV IVOS

GISTDA (partial listing)

Geosci.

Aust.

ESA

Science

Interface

Panel

JACIE

Govt.

Agencies

SSTL

9

ECCOE Structure & Personnel – USGS EROS

Director – Dennis Helder

Subject Matter Experts

Cody Anderson

Mike Choate

Jon Christopherson

Obaidul Haque

Minsu Kim

Esad Micijevic

Tom Stone, USGS Flagstaff

Web/PR/Admin. Support

Anya Hartpence

Ron Morfitt

Rajagopalan Rengarajan

Aparajithan Sampath

Pat Scaramuzza

Greg Stensaas

Jim Storey

Chris Crawford, Science liaison

10

ECCOE Structure & Personnel – Affiliates

University of Arizona

Nik Anderson

Stu Biggar

Jeff Czapla-Myers

+ Students

South Dakota State University

Larry Leigh

Morakot Kaewmanee

Cibele Teixeira Pinto

+ students

Rochester Institute of Technology

Aaron Gerace

Matt Montanero

John Schott

Nina Requeno

Emmett Ientilucci

+ students

Jet Propulsion Laboratory

Simon Hook

11

ECCOE Research & Publication

If a Center is going to be Excellent, it must do research and publish it

Good news—much of the work already being done fits!

Develop a paradigm shift

Goal: 1 journal paper + 1 conference paper per person per year

12

ECCOE Events

ECCOE supports calibration conferences

Sessions at CALCON, IGARSS, JACIE, Pecora, SPIE

ECCOE Workshop

Annual event

Topical focus

Develops recommended practices

Publishes proceedings

Conducts CalVal Future Directions session each year

Venue: JACIE or Pecora

13

ECCOE Workshop 2017

Topic: Cross-calibration of Landsat 8 OLI and Sentinel 2 MSI and its impact on data interoperability

Venue: Pecora Conference, 2017, November, Sioux Falls, SD

Monday (all day) and Tuesday (morning) before start of Pecora

Format:

Panel of Experts

Landsat Calibration expertise

Sentinel 2 Calibration expertise

Applications expertise

General audience

Presentations, discussions, Q&A

Outcome: Specific recommendations for Landsat 8 and Sentinel 2, as well as for general sensor cross-calibration

14

ECCOE Science Interface Panel

Purpose: Provide advice on the direction of Optical Remote Sensing CalVal for optimal service to the science community

Meets Quarterly

2 meetings with LST

2 more focused meetings… (?)

6-12 members

Selected by ECCOE Director and CalVal Team Leads

Some members internal to EROS

Significant membership from the next LST

Several other external members

15

ECCOE—what about validation?

This really begs the questions of how calibration and validation are defined, does ECCOE calibrate sensors or data, and where does the mission of ECCOE stop?

CEOS/WGCV provides the following definitions:

Calibration is the process of quantitatively defining a system’s responses to known, controlled signal inputs.

Validation, on the other hand, is the process of assessing, by independent means, the quality of the data products derived from those system outputs.

ECCOE limits its activities to calibration and validation of sensors and data through the Level 2 products of surface reflectance and surface temperature.

16

Cal/Val becomes a data user!

L1

L2 L3

The Imaging Chain Responsibility L1: Calibrated At-Sensor Radiance/Reflectance Cal/Val Team L2: Validated Surface Reflectance (Temperature) Cal/Val and Science L3: Products: ECVs, NDVI, LAI, faPAR, and many more Science

L3

L3

17

A few comments about Landsat 10

Landsat 10 must be the best calibrated moderate resolution optical imager in space or it will completely lose its relevancy to the remote sensing community

Any commercial sensor can provide data with some good qualities

No commercial sensor can provide data with science grade qualities

Only if we have a well-calibrated science-grade Landsat-like sensor in space will it be possible to have a reference by which to integrate commercial, as well as many or most civilian government, satellite imagery into a coherent, well-integrated data set that will be useful to both the science and commercial remote sensing community

The fundamental dataset for users will no longer be a TOA reflectance or temperature product, it will be a surface reflectance or temperature product

This is, for the most part, already the case!

18

Continuing the tradition of Landsat CalVal Excellence…

Today

Calibration based on solar diffuser, accuracy = 3%

Surface reflectance product based on LEDAPS (Landsat 4, 5, 7) and LaSRC (Landsat 8)

Dependent on external information—MODIS, DDV, etc

Accuracy variable, on the order of 5%*

Total uncertainty ≈ 5.8%*

(*estimate)

Feasible goals for Landsat 10 Calibration improved to 2% accuracy

One technology: piggy-back radiometer Another possibility: learning from

CLARREO approach

Surface reflectance product based on additional spectral bands for atmospheric characterization; one approach: 0.6µm for Ozone, aerosol over land,

biogeochemical study 0.74µm for aerosols over land/water, red-

edge 0.945µm for water vapor absorption 0.76µm for O2 absorption/air press. 1.05µm Water vapor, aerosol over water,

biogeochemical study Accuracy improved to 3%*

Total uncertainty ≈ 3.6%*

19

Thank you!