Mapping Long-term Land Use Change with Remote Sensing Data

| |

Land use change has been one of the most sustained and obvious results of the Anthropocene, where deforestation, agriculture, and expanding cities have left a clear imprint on our planet.

Monitoring this change is key to understanding its impacts but also to potentially alert scientists and policy makers where change needs to happen to stop destruction of our natural environment.

Monitoring land use change

Now, the USGS’ Land Change Monitoring, Assessment, and Projection (LCMAP) initiative is helping to better monitor our fragile Earth and long-term land use change.

Initially, LCMAP was setup at USGS’s Earth Resources Observation and Science (EROS) Center. The idea was to present to scientists and the public a new generation of land cover mapping and change monitoring using higher quality remote sensing data and more frequent data releases.

The focus is on conterminous (or contiguous) United States, although data on Hawaii has also been released. Overall, data are released as part of collections that have so far covered 1985-2021.

More than satellite data, which have mostly consisted of the Landsat program, information has included processed change detection algorithmic results (using PYCCD) and supervised learning model and results information.[1]

Screenshot from a dynamic map viewer that shows side by side three images: an aerial of Waldorf, Maryland from 2013/2014,  an aerial of Waldorf, Maryland from 2017/2018, and an aerial image of the same area with yellow and red for changes from natural to housing.
Side-by-side panels to show land use/land cover (LULC) change in Charles County, MD from 2014-2018. Screenshot from Dynamic LULC Change, Chesapeake Conservancy taken 25-October-2022.

Accessing Data from the USGS Land Change Monitoring, Assessment, and Projection (LCMAP)

Data from LCMAP are available from EarthExplorer, the LCMAP Web Viewer, LCMAP Mosaic websites, and the LCMAP WMS. (Related: USGS Releases Two Decades of Land Cover Change for the Contiguous United States)

Data in collections, such as collection 1.0 which spans from 1985-2017, provides: land cover information, confidence of land cover results, secondary land cover, secondary land cover confidence, annual land cover change, time of spectral change, change magnitude, spectral stability period, time since last change, and spectral model quality.

Benefits of using remote sensing data for long-term land use monitoring

Among different benefits, these data allow one to not only monitor land use change but also observe the quality of the results, how fast land use has changed over a given period, and what other land cover is present besides the primary observation.[2]

More recently, LCMAP’s 2021 (1.3) release extends coverage from 1985-2021.

A four-paneled diagram showing a vertical slice of a satellite imagery with forest damage, a second panel with land use change, a third panel symbolized to show the date of changes, and the fourth panel shows the magnitude of the change.
The USGS recently released LCMAP 1.3 data. From the USGS: “(A) U.S. Department of Agriculture (USDA) National Agriculture Imagery Program (NAIP) aerial imagery from November 15, 2021 over recently disturbed locations near Newnan, Georgia, located southwest of Atlanta. Examples include what appear to be (1) forest clearing for urban development, (2) part of a tornado scar from an EF4 tornado that struck Newnan on March 26, 2021, (3) a combination of disturbances including reforestation, and (4) forest harvest. (B) LCMAP Collection 1.3 Annual Land Cover Change from 2021 identified all four disturbances, shown in purple as land cover classification changes. (C) The LCMAP Collection 1.3 Time of Spectral Change product from 2020 and 2021 shows when spectral changes occurred in 2020 and 2021, grouped by month, including the tornado scar showing changes in April 2021 after the late March tornado. (D) The LCMAP Collection 1.3 Spectral Change Magnitude product, which shows the intensity of a spectral model break, with more drastic changes such as forest clearing in shades of blue and more subtle changes like forest regrowth or tornado damage in shades of green.” Image: USGS, public domain.

The data can be used to monitor many different types of land use change, including the dynamics of wetlands, look at the impact of hurricanes, see how mining has impacted different parts of the US, look at coastal erosion, and understand the intensity of land use change caused by wildfires.

The benefit of different data releases is that algorithms used for classification and validation of results could be continuously improved, allowing better change detection, for instance, over the period of data coverage. These results and updates to algorithms and their improvement are often published, such as in an article that presents change detection and classification methods.[3]

While much of the data focus on the lower 48 states in the US, data on Hawaii have recently been published and an article discussing some of the results has also been published.[4] 

Monitoring forest regrowth with remote sensing

Other recent results have focused on forest regrowth in the US as a way to mitigate long-term climate impacts.

A photo taken showing clear cut area in front and tall, thin trees in a forest in the background against a blue sky.
A forest clear-cut in the Ouachita Mountains of Arkansas. Photo: Mark A Drummond, USGS, public domain.

Results using LCMAP have shown that the US Southeast had the fastest regrowth in trees, averaging around 5 years to recover. This suggests efforts focused on faster growth areas could have a more rapid payoff by focusing on that region.

On the other hand, other areas might be much harder to recover their overall tree cover or at least take longer, such as the northwest taking decades to recover in cases.[5]

Long-term land use monitoring outlooks

Using LCMAP for various long-term land use monitoring will likely become more important for scientists and policy makers, particularly land use change plays such a major factor in a variety of environmental threats but also benefits.

Efforts to bring critical data together and enable an easier way to have classification and detection of land use change, without requiring this work to be done by researchers, makes the utility of Landsat data more accessible. This also now allows better planning to commence, such as where efforts to conserve forest and enable their regrowth should focus.

References

[1]    For more on PYCCD, see:  Killough, Brian D. 2022. ‘Satellite Analysis Ready Data for the Sustainable Development Goals’. In Geophysical Monograph Series, eds. Argyro Kavvada, Douglas Cripe, and Lawrence Friedl. Wiley, 133–43. https://onlinelibrary.wiley.com/doi/10.1002/9781119536789.ch8 (October 23, 2022).

[2]    For more information on the LCMAP program and products, see:  https://www.usgs.gov/special-topics/lcmap.

[3]    For more on how Landsat data are understood and used for change detection and classification in LCMAP using some updates, see:  Brown, Jesslyn F. et al. 2020. ‘Lessons Learned Implementing an Operational Continuous United States National Land Change Monitoring Capability: The Land Change Monitoring, Assessment, and Projection (LCMAP) Approach’. Remote Sensing of Environment 238: 111356.

[4]    For more on Hawaii and classification approaches used for that state, see:  Li, Congcong et al. 2022. ‘Development of the LCMAP Annual Land Cover Product across Hawaiʻi’. International Journal of Applied Earth Observation and Geoinformation 113: 103015.

[5]    For more on forest recovery across the US using LCMAP, see:  Zhou, Qiang et al. 2022. ‘Tree Regrowth Duration Map from LCMAP Collection 1.0 Land Cover Products in the Conterminous United States, 1985–2017’. GIScience & Remote Sensing 59(1): 959–74.

Related

Share this article


Enter your email to receive the weekly GIS Lounge newsletter: