Interpretation of County-Scale Observed Trends in Extreme Rainfall

Extreme rainfall events are experienced locally, but most attempts at trend detection and attribution rely upon regionally aggregated observations.  This study examines temporal changes in extreme rainfall events within a portion of the United States to investigate the correlation length scale of temporal trends, to identify spatially coherent trends at sub-regional resolution, and to document trend variability at finer scales.  The specific example of historical variations in extreme rainfall in the vicinity of Houston, Texas, is examined in detail.  It is shown using a variety of approaches that extreme rainfall has increased dramatically in and around Houston in recent decades.  However, the observed trend is unusually large compared to trends in nearby counties and is also much larger than would be expected were the trend driven by atmospheric water vapor capacity.  It thus appears that, while there is a long-term regional upward trend in extreme rainfall, the specific trend observed in Houston is largely a consequence of random temporal variability.  Implications for local-scale perceptions and responses to climate change are discussed.