Sampling variability effects in drop‐resolving disdrometer observations

Rain events are often studied with raindrop disdrometers—instruments designed to measure and record individual drop properties. Data from these instruments reveal that the instantaneous raindrop arrival rate is highly variable, with drop counts in consecutive time intervals occasionally changing by an order of magnitude or more. These drop count fluctuations result from some unknown combination of real physical changes in the underlying rainfall in conjunction with statistical fluctuations due to instrumental imperfections, limited sampling area, and the correlated nature of drop arrivals. Here empirical observations are used to drive a simulation that explores the degree to which observed variability in bulk rainfall observations might be produced by statistical fluctuations. It is shown that meaningful conclusions about true underlying rain properties require more drops than previously thought. Revisions to the interpretation of disdrometer data are discussed. Key Points Previous simulation studies of disdrometer sampling variability may have been too optimistic Small disdrometric samples tend to underestimate all moments of the drop size distribution Sampling can have interesting effects on disdrometer‐derived Z‐R relationships