The Impact of Regime Shifts on Long‐Range Persistence and the Scaling of Sea Surface Temperature Off the Coast of California

We employ power law scaling using Detrended Fluctuation Analysis to examine sea surface temperature (SST) in Monterey Bay and at Scripps Pier. SSTs from 1920 to 2014 provide the database. The data were first detrended to remove the annual cycle and long‐term trends. Next, Detrended Fluctuation Analysis was applied to the data to estimate the temporal scaling exponents (TSEs). El Niño–Southern Oscillation appeared to influence our scaling results at scales of 2.8 and 5.6 years, but this influence did not adversely affect the interpretation of the results that form the basis for this study. Next, 76‐year time series of TSEs were generated for the period from 1930 to 2005 using a 20‐year moving window. They reveal maxima during the 1930s and 1940s, significantly lower values during the 1950s and 1960s, and higher values from the 1970s to the early 1990s. The period of lower TSEs falls within a 30‐year window bounded by major regime shifts in 1945–1946 and 1976–1977, clearly suggesting a cause and effect relationship. Changes in the SSTs and coastal winds off southern California before and after these events further support this interpretation. Although the annual cycle at Scripps Pier far exceeds that at Pacific Grove, the scaling properties are similar. The TSEs at both locations converge toward unity at longer maximum timescales consistent with long‐range persistence and stationarity and with similar results obtained elsewhere in the North Pacific. Finally, we conclude that power law scaling can be an effective tool for detecting changes in system behavior. Key Points El Niños and the Pacific Decadal Oscillation (PDO) did not impact the results significantly in contrast to several other similar studies Regime shifts in 1945–1946 and 1976–1977 led to significant, long‐term changes in system behavior off the California coast We conclude that power law scaling can be an effective tool for detecting changes in system behavior