The El Niño – Southern Oscillation (ENSO) is a natural variation of ocean temperature in the tropical Pacific and a major driver of global climate variability, including precipitation extremes and tropical cyclone activity. ENSO has a diversity of spatial patterns that can alter its teleconnections. However, no single index can capture ENSO’s diversity, leading to busts in seasonal climate forecasts. The 2015-2016 El Niño is a case in point, producing unexpectedly weak impacts on California precipitation despite sea-surface temperature (SST) warming comparable to historical El Niño events that led to extreme precipitation. We demonstrate that this response can be explained by El Nino’s spatial pattern, and present a new ENSO index that can capture such ENSO diversity. Unlike traditional ENSO indices based on SST anomalies in a fixed region, the new “ENSO Longitude Index” tracks the east/west shifts in tropical Pacific deep convection that drive climate teleconnections and accounts for both the non-linear response of deep convection to SST as well as background SST changes associated with climate change. This physically-based index (1) reveals future changes in ENSO that are not apparent from traditional indices and (2) improves the value of ENSO as a predictor of California’s seasonal hydroclimate extremes compared to traditional ENSO indices.