The 1972 Clean Water Act serves as a crucial legislation to protect the “waters of the United States.” However, a significant obstacle in fully implementing this act lies in the lack of a precise definition for which streams and wetlands are covered by its regulations. This ambiguity has resulted in an ongoing struggle for presidential administrations, regulators, and courts to determine the exact scope of the Clean Water Act. In response to this challenge, a team of researchers from the University of California, Berkeley, leveraged machine learning to gain insight into the protected waterways.
The research conducted by the UC Berkeley team shed light on the impact of the 2020 Trump administration rule on the Clean Water Act. By training a machine learning model to predict jurisdictional decisions made by the Army Corps, the researchers were able to uncover crucial information about the deregulation of certain waterways. Their analysis revealed that the 2020 rule led to the removal of Clean Water Act protection for one-fourth of U.S. wetlands and one-fifth of U.S. streams. Moreover, this rule also deregulated 30% of the watersheds responsible for supplying drinking water to households across the nation.
The findings of the study published in Science highlight the significant negative impact of the regulatory changes. Under the 2020 rule, an estimated 690,000 stream miles, equivalent to the combined length of every stream in seven states, including California and Texas, were deregulated. Furthermore, the wetlands affected by this rule were estimated to provide over $250 billion in flood prevention benefits to nearby buildings.
The constant fluctuation and ambiguity in Clean Water Act regulations have profound implications for environmental protection. The authors of the study emphasize the detrimental effects of this “game of regulatory ping-pong.” The lack of consistent and predictable regulations not only undermines the core objective of preserving water quality but also creates challenges for regulators and developers.
The prediction capabilities of the machine learning model developed by the UC Berkeley team offer a potential avenue for resolving regulatory uncertainties. By providing immediate estimates of the likelihood of a site being regulated, this model could potentially save over $1 billion annually in permitting costs for both regulators and developers. This timely information would replace the current months-long permitting process, characterized by uncertainty and delays.
The research findings also have implications for the future development of Clean Water Act regulations. It is noteworthy that the 2023 Biden White House rule expanded the jurisdiction of the act, but it was soon followed by the Supreme Court’s 2023 Sackett decision, which contracted its scope. The machine learning methodology demonstrated in this study can provide valuable insights into the impact and implementation of future regulatory changes.
The UC Berkeley-led research, utilizing machine learning, has provided unprecedented insights into the coverage of the Clean Water Act. The study revealed alarming statistics regarding the deregulation of wetlands, streams, and watersheds under the 2020 Trump administration rule. By clarifying the scope of the act and streamlining the permitting process, machine learning presents a promising approach to address the challenges posed by uncertain regulations.
Leave a Reply