Â鶹ÒùÔº

If it feels like headlines reporting 100 or 1,000-year floods and megafires seem more frequent these days, it's not your imagination.

A project led by researchers from UConn's Global Environmental Remote Sensing (GERS) Lab has yielded surprising insights into land disturbances and disasters in the United States since the late 1980s, including a shift in what drives those disturbances, and how they are increasing with frightening intensity and frequency. Their findings are in Nature Geoscience.

The research is the result of a decade-long project to perform a CONterminous United States (CONUS)-wide disturbance agent classification and mapping project, explains GERS Director and Associate Professor in the Department of Natural Resources and the Environment in the College of Agriculture, Health and Natural Resources (CAHNR) Zhe Zhu. The ambitious project involved the careful analysis of Landsat satellite data spanning more than 40 years.

Disturbances like hurricanes and fires reshape the landscape and play vital roles in Earth's systems; therefore, understanding what drives these kinds of disturbances is important for projecting what changes may be ahead.

When talking about different types of disturbances, word choice is crucial, because the definition of "natural disaster" can be misleading. The authors are careful to define the trends we are seeing now.

"A lot of disturbances are no longer purely natural, and there is no clear line between human and natural disturbances anymore," says Zhu. "For example, there are so many wildfires, and many are not started by lightning nowadays."

In the case of flooding events, human-directed activities like logging and deforestation, construction, impervious surfaces, or dam failures can amplify these disturbances, and are therefore indirectly influenced by humans as well as anthropogenic climate change. The researchers call this category "wild" disturbances.

"We feel we're no longer able to call these disturbances 'natural disturbances,' so we made this new framework that has human-directed compared to 'wild' disturbances like vegetation stress, geohazard, wind, and fire that we put into another category because they are also greatly influenced indirectly by humans," says Zhu.

Lead author and Department of Natural Resources and the Environment Research Assistant Professor Shi Qiu explains that the study focuses on land disturbance occurring in different land surface types, because much of the research in this area has only focused on forest disturbances.

Using an advanced algorithm called COLD developed by Zhu, the researchers analyzed Landsat data from 1982 through 2023 to better understand the context in which different disturbances happened; for example, when and where the disturbance happened, as well as the causal agents, such as logging, construction, fire, or vegetation stress. You can explore the dataset .

"For example, we can capture wild disturbances, like wildfires or hurricanes, and we can also capture human-directed disturbances, like logging, construction, and agriculture. We used long-term Landsat satellite data to capture those disturbances in the past decades to see how those disturbances have shifted in the U.S.," says Qiu.

For the first time, Zhu says, they can distinguish the cause of the disturbance and analyze and quantify the area impacted by different causal agents, and also track trends, which the researchers found quite surprising.

"We found that human-directed disturbance is huge in the U.S., but we observed that human-directed disturbance has decreased in the past decades. Meanwhile, we found that wild disturbance is increasing. That is a major finding," says Zhu.

It is helpful to look at current and ongoing disasters to understand the magnitude of these findings, says Zhu. For example, the which killed over 100 people, or rapidly growing fires in and , to more local examples of years of that have impacted the region's trees.

"They're all linked together: changing our land is causing major disasters and landscape change at a scale we haven't seen before," says Zhu. "A lot of them are extreme weather events, but one question we have is what are the drivers causing them? Are they getting larger or smaller in the impacted area? Are they getting more frequent than before? We were able to see the trends, including the acceleration or deceleration of the trends."

The researchers found that wild disturbances are not only increasing in frequency, but also in severity.

"They are going wild and that is why we feel like 'wild' is quite useful for describing those disturbance agents."

To continue this important work, the researchers say they are looking for opportunities to collaborate to implement this method to other regions.

"This is not simple research that one person or a few people can do. We have lots of collaborations with remote sensing experts at other universities and outstanding ecologists. All of us worked together to make this happen," says Qiu. "To analyze this dataset, the UConn High-Performance Computing facility also gave us a lot of support."

The increasing trends the researchers observed are not linear, says Zhu, which makes it difficult to forecast the severity of future disturbances. Now is the time for more research like this to help guide resilience management.