USGS Perspectives on

USGS Perspectives on

A Framework for Analyzing the Needs for Continuity

of NASA-Sustained Remote Sensing Observations of

the Earth from Space

Tim Newman

Sarah Ryker

09.18.2013-Ryker – 2

USGS/NASA Landsat Partnership

USGS Perspectives

•

Our organizational perspective is based on:

–

41-year land imaging record.

–

A global change portfolio that includes a focus on

long-term change.

–

Department of the Interior responsibility to provide

data & science on immediate issues in natural

resource management.

–

Responsibility to gather, study, & represent users'

needs

–

2008 free data policy & the Landsat Global Archive

Consolidation have enabled new studies accessing the

full period of record for more locations on the globe.

Defining Continuity

•

Many sources:

–

The law

–

The Landsat Science Team

–

The 2013 NRC report on Landsat

–

Global, continental and national-scale change

studies

Continuity and the Law

•

In Public Law 102-555, the “Land Remote Sensing Policy Act of 1992,”

Congress found:

– “The continuous collection and utilization of land remote sensing data from space are of major benefit in studying and understanding human impacts on the global environment….”

– “…it is in the best interest of the United States to maintain a permanent,

comprehensive Government archive of global Landsat and other land remote sensing data for long-term monitoring and study of the changing global

environment.”

•

The 1992 law defines “data continuity” as the continued acquisition and

availability of data which are, from the point of view of the user:

– “sufficiently consistent (in terms of acquisition geometry, coverage

characteristics, and spectral characteristics) with previous Landsat data to allow comparisons for global and regional change detection and

characterization; and

– compatible with such data and with methods used to receive and process such data.”

USGS-NASA Landsat Science Team

•

The USGS-NASA Landsat Science Team (LST) defines Landsat data

continuity as the collection, archival, and distribution of image data

of the Earth’s continents and surrounding coastal regions with the

content, quality and coverage needed to map, monitor and assess

the Earth’s characteristics and its response to natural and

human-induced change. To accomplish this, continuity includes:

–

Long-term calibrated measurements that are consistent across the

changing instrument record.

–

A continuous record since the initiation of observations with no

significant temporal or geographic data gaps.

–

Measurements that enable backward and forward assessments of the

conditions and changes in the Earth’s surface.

–

Measurements with comparable spectral, spatial, temporal, and

geographic properties that result in sufficiently consistent and

accurate documentation of surface characteristic and dynamics.

LST Statement on Continuity

•

Landsat continuity is needed to:

–

Ensure the scientific integrity and objectivity of global

land and climate change research.

–

Ensure the continued accrual of benefits from past

Landsat investments by extending the long-term

record.

–

Protect the investments of agencies, organizations,

and businesses that rely on Landsat data to meet their

objectives.

–

Deliver the benefits to society for which the Landsat

program was established.

Landsat Science Team (continued)

•

The Landsat 1-8 mission goals have been unchanged with respect to the

continuity purpose in support of science understanding of local to global

land change and providing information needed for natural resource

management.

•

There have generally been consistent basic data characteristics since

Landsat 1, and especially since Landsat 4 and the beginning of the

Thematic Mapper (TM) era.

•

There have also been evolutionary advances in capabilities resulting from

new technologies and ideas that have improved applications without

compromising continuity. Since 1972, there have been incremental

improvements in image acquisition capacity, measurement capabilities,

and data quality, and as a result, Landsat science and applications have

expanded and improved. Now, both scientific and operational

applications are based on the latest Landsat capabilities (e.g., Landsat 8),

and those capabilities have been shown to meet design and operations

aims and offer a

de facto

definition of minimum target specifications.

Landsat Science Team (continued)

• Data Accessibility: Landsat data must be freely available to anyone in order to guarantee the return on investment.

• Geographic Coverage: Must be near-global and include continental surfaces, ice sheets, coastal regions, islands, and coral reefs imaged using the Long Term Acquisition Plans implemented with the start of the Landsat 7 mission.

• Temporal Frequency: Eight or 9-day coverage of the global land surface by two orbiting Landsat satellites has occurred over 2/3s of the program’s history. This is the standard for continuity.

• Latency: Twenty-four hour turn around from acquisition to downloadable data, with capabilities for near real time access in emergencies is required.

• Spectral Bands: The full set of 11 Landsat 8 spectral bands (VNIR, SWIR, TIR) collected in a near-simultaneous mode is the standard for data continuity. The LST priorities for spectral bands are from greatest priority to least priority: VNIR / SWIR, Cirrus, TIR, Enhanced blue (coastal/aerosol band), Panchromatic.

• Spatial Resolution: The heritage 30 m spatial resolution of the TM, ETM+, and OLI sensors has proven eminently suitable for the applications of Landsat data and the scientific objectives of the Landsat program. Continuation of the 30 m resolution for reflective bands is one of the highest priorities for the future of land imaging. The

heritage 15 m spatial resolution of the ETM+ and OLI panchromatic bands is considered a maximum ground sample distance for future panchromatic images, and 120 m spatial resolution is the coarsest acceptable resolution for the TIRS thermal bands.

Landsat Science Team (continued)

• Radiometric Calibration and Accuracy: Radiometric accuracy and stability are essential for

monitoring change over time, a major Landsat program objective and the rationale behind the goal of continuity for future land imaging. Continuity requires either an uncertainty of less then 5% with respect to absolute spectral radiance or less than 3% with respect to top-of-atmosphere reflectance in the case of images for reflective spectral bands. Continuity requires an uncertainty of less than 2% with respect to at-sensor spectral radiance in the case of thermal bands.

• Radiometric Performance: The LST considers the Landsat 8 radiometric performance specifications to be the continuity standard. The radiometric performance requirements of the Landsat 8 OLI and TIRS instruments were stringently specified in terms of signal-to-noise ratios (SNR’s), cross-track radiometric and spectral response uniformity, coherent noise, stray light, and a number of

additional characteristics.

• Geometric and Geodetic Accuracy and Stability: Geometric/geodetic accuracy and stability are critical to observing change over time and other land image analyses, and the LST considers the Landsat 8 performance with respect to geometric accuracy, geodetic accuracy (registration to a cartographic project with ground control), orthorectification, band-to-band registration, and multi-temporal image-to-image registration, to be the continuity standard.

NRC report (2013): Landsat and Beyond:

Sustaining and Enhancing the Nation's Land

Imaging Program

Continuity:

•

Spatial resolution

– 30 m except in the thermal band, which would have coarser spatial resolution.

– Finer resolution (10-15 m), perhaps in a panchromatic band, was desired by some.

•

Spectral requirements

– Visible and near-infrared region (VNIR, 0.4-1.1 μm).

– Shortwave infrared region (SWIR, 1.2-2.8 μm).

– Thermal infrared region (TIR, 8-12 μm, with some interest in 3.5-4.0 μm).

– Calibration sufficient to allow backwards-compatible comparisons of future image products to previous collections.

– A larger dynamic range in the VNIR region to prevent saturation over snow and clouds; this requirement has been met in the Landsat 8 OLI, with its 12-bit quantization instead of 8.

NRC Report (Continued)

•

Coverage and repeat cycle

– Ability to acquire and make available imagery anywhere on Earth, except

perhaps for areas very near the poles, at approximately weekly frequency. The 705-km Landsat orbit, at 98° inclination, provides 16-day repeat. The temporal frequency is not necessarily to acquire weekly data but for cloud-free images.

– Increased temporal frequency could be achieved with a slightly larger swath and consequently slightly larger off-nadir view angles at the edge (there was no objection to this among the users queried).

•

Data management and distribution

– A free data policy, as is currently in place, provides huge benefits to the nation as well as the international user community by supplying imagery to

operational programs critical to U.S. needs as well as spurring innovation in the private sector.

– The USGS data distribution system is successful and effective but has opportunities to continue to improve with technological advances and to streamline methods for managing Landsat imagery and derived products.

Continuity - Radiometry

•

An essential building block for continuity in land imaging:

–

Maintaining the same instrumentation over a 40+ year

observational record isn't possible.

–

Consistent radiometry is essential to maintain continuity across

long time periods with changing instrumentation.

–

This is clearly demonstrated by a number of quantitative

biophysical analyses that rely on consistent radiometry to

identify slow-moving trends over long periods.

–

These analyses provide fundamental understandings of global

change in a number of fields.

–

Examples in land change and vegetation characteristics at the

global, continental, national, and regional scales are referenced

at the end of the presentation.

Operational Continuity

A rapidly growing area of Landsat use: Environmental

decision-making, from adjudicating water rights disputes to

urban planning.

–

New Landsat User Survey finds that, depending on how we

define terms, approx. 1/3 of Landsat data users are 'operational'

users. These users generally need, for example:

• 8-day repeat; In 2012 analysis of representative Landsat applications,

2/3 of all users required 8-day repeat.

• Examples: crop productivity, fire assessment, flood monitoring, crop

insurance verification, irrigated ag; others

–

These operational users invest heavily in applications & decision

processes; they require high confidence that a compatible data

stream will continue into the future

Operational vs. Science Continuity

•

Does operational continuity require the same

temporal history that science does?

–

How many applications are synoptic (never look

back)? E.g. look at a field & move on; look at flood

inundation extent & move on.

•

A long & continuous observational record enables

us to bring science to bear on operational issues.

–

E.g. with 41 years of Landsat data, looking back at

previous events can enable an improved prediction of

the next event, not only a snapshot after the fact.

Continuity Definition

•

Continuity has to refer to data products and supported applications, not to

the satellite technology.

•

Definition of continuity should be based on the ability to produce

information products that can look back across all previous records &

continue to generate related products into the future, supporting existing

operational applications.

– For example, if the US Forest Service has built an infrastructure to monitor burned area that depends on particular imaging characteristics, data

continuity should allow USFS to continue using Land Imaging data products for that application.

•

Is "continuity“ fundamentally backward-looking?

– Not if continuity is defined as based on uses of the data. The uses may be stable over time, but be supported by new measurements.

Use-Based Continuity

•

A stricter definition would emphasize the science and applications need

for long-term, comparable measurements:

– "the ability to identify and measure short-term change and long-term trends in land surface variables due to both human and natural causes.“

•

Critical in this definition is the idea that "system" noise (i.e. variability

across the entire data record) can't swamp the ability to identify change

and trends. The variables of interest might be those identified by GCOS as

Essential Climate Variables. This definition emphasizes multi-decade

temporal consistency in derived products.

•

Similar concept to the definition provided in the 1992 Act:

– "sufficiently consistent (in terms of acquisition geometry, coverage

characteristics, and spectral characteristics) with previous Landsat data to allow comparisons for global and regional change detection and

Additional Resources

We’re happy to put you in touch with the Landsat

Science Team, and provide further examples of science

and operational uses.

Examples of Global Studies

•

Hansen, M.C., Potapov, P., Moore, R., Hancher, M., Turubanova, S.,

Tyukavina, A., Thau, D., Stehman, S, Goetz, S., Loveland, T.R.,

Kmommareddy, A., Egorov, A., Chini, L., Justice, C., and Townshend, J.,

2013.

High-resolution global maps of 21

_{-century forest cover}

change

.

Science

(in press).

– More than 650,000 Landsat images were analyzed to document annual global forest gains and losses from 2000-2012. The results of this study will be

published on November 15, 2013.

•

Sexton, J.O., X.-P. Song, M. Feng, P. Noojipady, A. Anand, C. Huang, D.-H.

Kim, K.M. Collins, S. Channan, C. DiMiceli, J.R. Townshend. 2013. Global,

30-m resolution continuous fields of tree cover: Landsat-based rescaling

of MODIS continuous fields and lidar-based estimates of

error.

International Journal of Digital Earth

6(5): 427-448.

– A 2000 and 2005 global 30-m resolution dataset of percent tree cover was developed from Landsat imagery, resulting in the highest resolution global tree cover dataset currently available for use in global investigations.

Global Studies (continued)

•

United Nations Food and Agriculture Organization, 2012.

Global Forest

Land-Use Change: 1990-2005.

FAO Rome, Italy 53 p.

– Landsat data were used to produce the first independent FAO estimate of global forest land use and forest cover change. The study was based on a

sampling framework and involved the interpretation of nearly 10,000 Landsat images. Results showed that forest loss is greatest in the tropics.

•

Williams, R.S., Jr. and Ferrigno, J.G., eds., 2012.

Satellite Image Atlas of

the World: State of the Earth’s Cryosphere at the Beginning of the

21

_{Century – Glaciers, Global Snow Cover, Floating Ice, and Permafrost}

and Periglacial Environments.

U.S. Geological Survey Professional Paper

1386-A 496 p.

– This is one of ten volumes in a long-term study of the status of the cryosphere based on Landsat and other satellite image documentation. The report

summarizes the continuing reduction of the Earth’s snow and ice and the pervasive footprint of human activity as detected in the satellite record since 1972.

Continental-Regional Studies

•

Masek, J. G., Huang, C., Wolfe, R., Cohen, W., Hall, F., Kutler, J., and Nelson,

P., 2008.

North American forest disturbance mapped from a decadal

Landsat record.

Remote Sensing of Environment

112/(6): 2914–2926.

– A wall-to-wall record of stand-clearing forest disturbance (clearcut harvest, fire) for the United States and Canada was produced for the 1990–2000 period using the Landsat satellite archive. Results indicate disturbance rates of up to 2–3% per year are common across the US and Canada due primarily to harvest and forest fire, with highest rates in the southeastern US, the Pacific

Northwest, Maine, and Quebec.

•

Skole, D.L. and Tucker, C.J., 1993.

Tropical deforestation and habitat

fragmentation in the Amazon: satellite data from 1978 to

1988.

Science

260, 1905-1910.

– One of the first studies of large-area forest loss, this study used Landsat data to provide definitive information on the rages of humid tropical forest loss from the Amazon (this paper), and in Africa and Southeast Asia.

National-Regional Studies

• Drummond, M.A. and T.R. Loveland. 2010. Land-use pressure and a transition to forest-cover loss in the Eastern United States. BioScience 60: 286-298.

– Nearly 30 years of Landsat imagery were interpreted to document Eastern US land cover change. Results show that forest cover is declining due to the increasing rates of timber harvest associated with Southeast US silvaculture practices, and also documented that a significant percentage of the Eastern US is in transitional bare ground status due to

forest clearing and urban development.

• Fraser, R. H., Olthof, I., Carrière, M., Deschamps, A., and Pouliot, D., 2011.

Detecting long-term changes to vegetation in northern Canada using the Landsat satellite image archive. Environmental Research Letters 6(4), 045502.

– The study of vegetation in Canadian national parks using Landsat imagery spanning 17-25 years showed changes in fractional shrub and other vegetation covers and a

consistent pattern of greening (6.1–25.5% of areas increasing) and predicted increases in vascular vegetation in all four parks that corresponded with positive temperature

National-Regional Studies (continued)

• Huang, C., Kim, S., Song, K., Townshend, J.R., Davis, P., Alstatt, A., Rodas, O., Yanosky, A., Clay, R., Tucker, C.J., and Musinsky, J., 2008. Assessment of

Paraguay’s forest cover change using Landsat observations. Remote Sensing of Environment 67:1–12.

– The research produced a comprehensive assessment of Paraquay’s forest cover loss from the 1970’s through mid-2000’s using Landsat imagery. The results showed that the Atlantic Forests ecoregion experienced the greatest loss.

• Olmanson, L.G., Bauer, M.E., and Brezonik, P.L., 2008. A 20-year Landsat water clarity census of Minnesota's 10,000 lakes. Remote Sensing of

Environment 112(11): 4086–4097.

– A 20-year comprehensive water clarity database was created using Landsat imagery for Minnesota lakes larger than 8 ha in surface area contains data on more than 10,500 lakes at five-year intervals over the period 1985–2005. Mean water clarity in the Northern Lakes and Forest and North Central Hardwood Forest ecoregions remained stable from 1985 to 2005 while decreasing water clarity trends were detected in the Western Corn Belt Plains and Northern Glaciated Plains ecoregions.

National-Regional Studies (continued)

•

Paul, F., Kääb, A., Maisch, M., Kellenberger, T., and Haeberli, W.,

2004.

Rapid disintegration of Alpine glaciers observed with satellite

data

.

Geophysical Research Letters

31(21): L21402-21406.

– Analyses of Landsat data from the 1970’s through 2000 indicated accelerated decline of alpine glaciers in Switzerland and other regions beginning in the 1980s. An 18% area reduction was observed for the period 1985 to 1999, a seven times higher loss rate compared to the 1850–1973 decadal mean.

•

Sleeter, B.M., Sohl, T.L., Loveland, T.R., Auch, R.F., Acevedo, W.,

Drummond, M.A., Sayler, K.L, and Stehman, S.V., 2013.

Land-Cover

Change in the Conterminous United States from 1973-2000.

Global

Environmental Change 23: 733-748.

– Landsat images from 1972-2000 were analyzed to document the rates and causes of contemporary US land cover change. Results showed the slow decline in forest area, the steady increase in developed lands, and the significant reductions in cropped area due to the 1985 Farm Bill.