Researchers combine active airborne lidar sensors, passive optical sensors in space and machine learning – ScienceDaily

About 40 percent of interior Alaska is covered by ice-rich permafrost — permanently frozen soils of earth, gravel, and sand — held together by ice. Certain conditions, such as global warming, have intensified wildfires on the tundra, which is having a profound impact on permafrost thawing.

Surface vegetation plays a dominant role in protecting permafrost from summer heat, so any change in vegetation structure, particularly after severe wildfires, can cause dramatic top-down thawing.

Severe wildfires remove vegetation and surface soil organic matter, and the loss of this insulation increases soil heat flux and promotes permafrost thawing. This thawing triggers ground subsidence and the development of thermokarst (collapsing of the soil surface by thawing permafrost), leading to surface water inundation, vegetation shifts, changes in soil carbon budgets, and carbon emissions, all of which impact global warming.

The permafrost-fire-climate system has been a research hotspot for decades. The large-scale effects of these wildfires on land cover change, postfire resilience, and subsequent thaw colonization remain unknown. Thaw settlements are difficult to measure because there are often no absolute frames of reference to compare with the subtle but widespread topographical changes in permafrost landscapes.

Florida Atlantic University researchers, in collaboration with the United States Army Corps of Engineers’ Cold Regions Research & Engineering Laboratory and Alaska Ecoscience, have systematically analyzed the onshore effects of six major fires that have occurred in the lowlands of the Tanana Flats of interior Alaska since 2000 Cover changes, vegetation dynamics, and terrain subsidence or subsidence.

For the first time, they have developed a machine-learning-based ensemble approach to quantify fire-induced thaw settlement across the Tanana Flats, which cover more than 3 million acres. Researchers combined repeated airborne LIDAR data with time series from Landsat products (satellite imagery) to delineate thawing settlement patterns in the six burn scars. This novel approach helped explain about 65 percent of the variance in the altitude change detected by the lidar.

Study results published in environmental research letters, showed that the six fires from 2000 to 2014 resulted in a total loss of nearly 99,000 acres (about 400 km2) of evergreen forest, while nearly 155,000 acres (about 590 km2) were fire-affected forests of varying fire severity. The fires provided favorable conditions for the development of shrub bogs (low-growing shrubs), resulting in comparable postfire cover of shrubland and evergreen forest, and increased encroachment of shrubland into areas with sparse vegetation.

Importantly, based on Landsat observations, the researchers observed no regrowth of forests after 13 years of the oldest fire in 2001.

“Our study has shown that linking airborne repeater lidar to Landsat products is an encouraging tool for large-scale quantification of fire-induced thaw formation,” said Caiyun Zhang, Ph.D., senior author and professor in the department there of Geosciences Charles E. Schmidt College of Science at FAU. “As airborne lidar measurements are increasingly conducted in northern permafrost regions, our method is a valuable tool to project elevation changes across entire burn scars within uniform permafrost-affected landscapes using data-driven machine learning techniques.”

The Tanana Flats, covering more than 6 million acres (about 2,500 km2), are representative of the lowland landscape south of Fairbanks in interior Alaska. Composed of a complex mosaic of ice-rich permafrost and permafrost-free ecosystems, it is a hotbed of thermokarst. Much of the country is part of a military training area administered by the US Department of Defense.

For the study, researchers evaluated three commonly used machine learning algorithms, including an artificial neural network, a support vector machine, and a random forest for modeling fire-induced thaw settlements.

“Machine learning has been widely used for modeling in earth science,” Zhang said. “The idea is that each algorithm has its advantages and disadvantages and an overall analysis of comparison models can provide a more robust estimate than applying a single model.”

Current and future projected increases in mean annual air temperature, the length of the summer growing season, and the severity and magnitude of wildfires are expected to result in an increasingly dominant role of wildfires in permafrost ecosystems.

“Mapping thaw settlements resulting from wildfires is critical because it is associated with subsequent thermokarst development, snow accumulation, hydrology, vegetation shifts, and corresponding changes in land-atmosphere exchanges of water, energy, and greenhouse gases,” Zhang said. “Combining active airborne lidar sensors with passive optical sensors in space will allow scientists to measure widespread and large areas affected by wildfires in cold regions, especially in times of global warming and increased fire events.”

This research was funded by the US Army Corps of Engineers, Engineer Research and Development Center, Applied Research Program Office for Installations and Operational Environment and Basic Research Program (PE 0601102/AB2), the US Department of Defense Strategic Environmental Research and Development Program (Projects RC2110 and RC18-1170) and the US Department of Energy, Office of Science, Environmental System Science program (0000260300).

Study co-authors are Thomas A. Douglas, Ph.D., research chemist and senior scientist, US Army Cold Regions Research & Engineering; David Brodylo, a Ph.D. Student at the Department of Geosciences at FAU; and M. Torre Jorgenson, Alaska Ecoscience.

Leave a Reply

Your email address will not be published. Required fields are marked *