Abstract
Recent studies indicate that hydrologic regionalization can benefit from explicitly accounting for large-scale ocean–atmosphere teleconnections. However, traditional approaches often rely on fixed indices such as the canonical Niño regions, which may capture only part of the variability associated with Pacific sea surface temperature anomalies. In addition, linear techniques commonly used to identify coupled ocean–land patterns may struggle to represent evolving or non-stationary teleconnection structures. In this study, we apply a structural similarity framework, termed point–slope (PS) similarity, to identify ocean regions that exhibit synchronized regime behavior with precipitation variability in California. The PS metric evaluates paired time series based on the alignment of changepoints and the agreement of trend directions between successive segments. Using monthly SST anomalies from 1981 to 2022 sampled at every 1.5° grid node across the Pacific and precipitation observations from 178 rain gauges in California, we compute a spatial field of PS similarity linking ocean nodes to land precipitation regimes. To assess the robustness of the detected teleconnections, we complement the PS analysis with a surrogate-based significance test and a sensitivity analysis to SST detrending. Statistically significant relationships are primarily found for January, where 287 ocean nodes remain significant after false discovery rate correction. These nodes form a broad corridor of influence extending across the equatorial Pacific, partially overlapping but not fully confined to the canonical Niño regions. Finally, we compare the PS teleconnection field with patterns obtained from maximum covariance analysis (MCA), a conventional linear benchmark. While MCA captures a similar ENSO-related structure, the PS framework reveals a more spatially heterogeneous teleconnection pattern, highlighting additional ocean regions linked to precipitation regime shifts in California. These results suggest that structural similarity metrics can complement traditional covariance-based methods when identifying non-stationary climate–hydrology relationships..
