ABoVE Science Team Members Study Boreal Forest Patterns on
ADAPT Using LVIS Data, Field Sensors, and Satellites


A sled dog team helps Arctic Boreal Vulnerability Experiment (ABoVE) researchers access a carbon-measuring instrument station at the Eight Mile Lake study area, located just outside the boundaries of Denali National Park. Dr. Ted Schuur from the University of Northern Arizona, Flagstaff is the Principal Investigator on the Eight Mile Lake Project, a National Science Foundation (NSF)-funded study site that is important to NASA scientists. Image credit: NASA/Peter Griffith.

At the Eight Mile Lake study site, located in the northern foothills of the Alaska range, a suite of instruments on a “flux tower” measures the movement of carbon dioxide and methane from the tundra ecosystem to the atmosphere. This data helps researchers estimate how fast the plants of the tundra are growing each summer, how much carbon the plants remove from the atmosphere, and how much carbon returns to the atmosphere because of respiration and decay of organic matter in the soils. Image credit: NASA/Peter Griffith.

Background


Why study the boreal forest? The biome boundary at the boreal (coniferous forest, a.k.a. taiga) to tundra (cold, vast, flat, treeless region) ecological transition zone in North America and Eurasia—also called the taiga-tundra ecotone—is recognized as being a bellwether of Arctic change. This transition zone is experiencing Arctic amplification—a phenomenon observed when Earth’s climate produces a larger change in temperature near the poles than in the planet’s lower latitudes.

The drivers of Arctic amplification emerge from sources originating near the poles down to the tropics. According to NASA climatologist Anthony Del Genio, one of the major factors in Arctic amplification is a warming climate and the resulting loss of bright, reflective sea ice. When ice melts, it gives way to a darker ocean surface that absorbs more heat from sunlight, rather than reflecting it back to space. Another factor is tropical thunderstorms, which transport heat from Earth’s surface to higher levels of the atmosphere, where global wind patterns sweep it toward the poles in a near-constant flow of heat away from the tropics, a process that dampens warming near the equator and contributes to Arctic amplification.


Using observational (satellite) data from the Aqua and Terra missions, NASA's Scientific Visualization Studio created this scientific visualization of the daily Arctic sea ice and seasonal land cover change progress through time, from the yearly maximum sea ice extent on March 21, 2021 through its yearly minimum extent on September 16, 2021. The yellow boundary represents the minimum sea ice extent from 1981 to 2010, averaged over the 30-year period. The average minimum extent has significantly declined since NASA satellite records began in 1978. The last 15 years have seen the lowest sea ice extent minimums in the 43-year record. Visualization by Trent Schindler, NASA Scientific Visualization Studio.

Along with sea ice loss, scientists have observed evidence of Arctic amplification, including noticeable changes to vegetation and other ecosystem components taking place at the taiga-tundra ecotone, with polar regions warming at a rate two or more times faster than other Earth regions. Measurable changes to vegetation include the length of the growing season, "shrubification" (increasing shrubs), changes in composition (increasing proportion of deciduous species), and changes in wildfire frequency and extent. The first step in measuring these changes is to define what and where the ecological transition zone is and use those parameters as a baseline for future research—a challenging problem for NASA researchers.

Two members of the Arctic-Boreal Vulnerability Experiment (ABoVE) Science Team, geographers Paul Montesano of NASA Goddard Space Flight Center in Greenbelt, Maryland, and Matt Macander of Alaska Biological Research (ABR) in Fairbanks, Alaska, and their collaborators have taken on this challenge via several recent projects leveraging NASA Center for Climate Simulation (NCCS) resources. Ultimately, the ABoVE team wants to better understand the future of ecosystem functioning in Earth’s high northern latitudes.

Geographers Paul Montesano of NASA Goddard Space Flight Center and Matt Macander of Alaska Biological Research (ABR, Inc.), members of the Arctic-Boreal Vulnerability Experiment (ABoVE) Science Team.

Montesano and his colleagues wanted to measure the area of the cold edge of the boreal forest at locations where previous global estimates often show high discrepancies and provide a basis for closely examining ongoing changes in the diverse circumpolar biome boundary. Their methodological approach was to find a way to use Landsat tree cover data in a geospatial analysis system to classify forest patterns in and near the northern limit of the boreal forest, doing so in a way that can help identify and map the extent of the biome boundary that the team wants to study.

The approach used existing Landsat-derived tree canopy cover maps to explain the spatial rate of change in forest structure, which varies across this transition zone. This rate, along with the actual tree cover estimates themselves, helped identify areas where the ecotone gradually and abruptly transitions from forest to tundra. This analysis resulted in a classification of forest structure patterns that, when examined, helps explain the how forest structure changes across the landscape, identifies abrupt edges of forested extents, and provides a spatial bounding of an ecological region that is otherwise difficult to quantify.

The result of approximately 4 years of work is a 30-meter-per-pixel map built from Landsat tree canopy cover data, classified as seen in the legend below. This map was a helpful step in directing further studies comparing the kind of changes the study team might expect to changes observed in this domain.


Impact: This 30-meter-per-pixel map built from Landsat tree canopy cover data can be helpful in comparing taiga-tundra ecotones and forest transition in different places. Researchers can then evaluate whether the changes that were predicted happened, or if they're different, and whether they're driven by a different set of factors.


Above: The forest structure pattern map developed in this research, depicts the North American taiga-tundra ecological transition zone across portions of Alaska and Northern Canada. Credit: Paul Montesano, NASA Goddard.

The team’s approach was to develop a static snapshot of the current taiga-tundra ecotone patterns. These current patterns can then be used as a lens to view what's happened in the past, where time series of data help explain site histories.

This map can be used to test the hypothesis: will these current patterns be useful for learning about what has happened and what will happen in the ecotone? Can researchers use this type of map to target future studies or modeling studies in a way that helps us understand the variation of upcoming changes? The question then becomes, what kind of changes in forest structure will our predictive models indicate for portions of the ecotone that feature varying patterns in current structure?

This biome boundary will change in different ways, and the changes that transpire near Gwazhał (a.k.a. the Brooks Range) may be different than what happens on the Canadian Taiga Plains Ecozone. A map such as this provides consistent structure patterns and a snapshot in time across the whole domain of the taiga-tundra ecotone. This map can be helpful in comparing ecotones and forest transitions in different places. Researchers can then use this map to target sites for predictive modeling based on these current observations. Then, they can evaluate how these current patterns may be linked to future landscape change processes.

Next Steps


This account of the taiga-tundra ecotone quantifies the area of the cold edge of the boreal forest where previous global estimates often show high discrepancies and also serves as a basis for closely examining ongoing changes in the varied circumpolar biome boundary. Scientists will discuss those changes and run simulations to compare and contrast what the simulations show will happen in 100 years and what might be driving those changes.

Montesano—along with Dr. Amanda Armstrong and Dr. Batuhan Osmanoglu of NASA Goddard, Dr. Howard Epstein of the University of Virgina, and others—are exploring questions on how the fate of ecotone forest structure may differ spatially across a broad biome boundary. To do this, they are using a detailed forest growth model that simulates such changes as part of a current NASA ABoVE project. The high spatial detail that this type of modeling can offer is enabled through the compute and storage on NCCS ADAPT virtual machines. Here, simulations of individual trees are being initialized for dozens of ecotone sites, each many square kilometers in size. These simulations will track how tree growth changes across the ecotone over hundreds of years.

The LVIS Project


Vegetation structure and other land cover patterns derived from satellites can benefit from reference estimates of surface features that are captured with airborne lidar. The goal of this work was to produce an analysis-ready dataset of vegetation structure that is accessible to a range of scientists in a format that is convenient for those who work with geospatial datasets. It arose from a synthesis activity idea conceived of by the ABoVE Vegetation Structure & Function Working Group. Macander wrote the original R code to process single LVIS flightlines. Montesano amended the code to run in parallel on ADAPT virtual machines in order to efficiently process and reprocess the entire archive collected during the ABoVE airborne campaigns in 2017 and 2019.

NASA Goddard’s LVIS Team used NASA’s Land, Vegetation, and Ice Sensor (LVIS), an airborne scanning laser altimeter, to collect an extensive archive of waveform lidar across northwestern Canada and Alaska in the growing season (June to August) of 2017 and 2019 in support of ABoVE.

Map of LVIS 2017 and 2019 flightlines from airborne campaigns across Alaska and western Canada. Map credit: Paul Montesano.

Working together, Montesano and Macander created analysis-ready (geo-referenced, gridded) versions of these data and enriched them with an additional set of canopy cover estimates at a variety of height thresholds. These 30-meter grids describe the vertical column of the vegetation canopy in detail with relative canopy height metrics.

These data provide important vertical and horizontal vegetation structure data across a variety of boreal forest sites in Alaska and Western Canada, supporting mapping and modeling of boreal forest structure patterns and dynamics. They are available at the Oak Ridge National Laboratory Distributed Active Archive Center.

Compute Used


For their work on the LVIS project, Montesano and Macander used the ADAPT (Advanced Data Analytics PlaTform) Science Cloud, a cloud-based environment at NCCS specifically designed for large-scale data analytics. The team ran code in parallel on ADAPT virtual machines in order to efficiently process and reprocess the ABoVE 2017 and 2019 airborne campaigns archive.

For their work simulating ecotone change, Montesano, Armstrong, and Osmanoglu have reconfigured an individual-based forest gap model called ‘SIBBORK-TTE’ to ingest a variety of multi-scale geospatial data stored on ADAPT; built a custom Python environment to run these simulations; and developed code to summarize results. They are working toward publications describing their model updates and their research findings.

Related Links


Sean Keefe, NASA Goddard Space Flight Center