Time of observation adjustments to daily station precipitation may introduce undesired statistical issues

Abstract

ABSTRACTInterstation observations of daily precipitation are often temporally misaligned due to differences in station time of observation. Several time of observation adjustment methods have historically been applied to improve interstation temporal alignment, but the efficacy of such adjustments has not been fully tested. Here, we examine the ability of several time of observation adjustments to improve observation compatibility under three different adjustment scenarios: adjusting morning observations to midnight, adjusting afternoon observations to midnight, and adjusting afternoon observations to morning. We find that all adjustment methods provide necessary improvements to the temporal alignment of daily precipitation observations, especially with respect to morning versus midnight time of observation totals. However, for a majority of the adjustments, improved temporal alignment comes at the cost of significantly altering observed precipitation intensity, frequency, and extremes. We also find adjustments have the potential to overcorrect and increase both general and extreme event spatio‐temporal coherence.