
Adapted GLM gamma parameter estimates for drop size distributions
Author(s) -
Brawn Dan
Publication year - 2015
Publication title -
atmospheric science letters
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.951
H-Index - 45
ISSN - 1530-261X
DOI - 10.1002/asl2.572
Subject(s) - estimator , maximum likelihood , generalized linear model , disdrometer , bin , mathematics , statistics , drop (telecommunication) , estimation theory , computer science , algorithm , physics , meteorology , precipitation , telecommunications , rain gauge
This article outlines an alternative maximum likelihood gamma parameter estimator for grouped and truncated data. Such data might be produced by traditional disdrometer instruments for recording rain drop sizes. The method is an adaptation of a log‐linear generalized linear model ( GLM ), forming an iterative process which quickly converges to very near maximum likelihood estimates. Simulations with a Joss–Waldvogel bin regime show that the estimator has little bias. Application to observed data from two different types of disdrometers yields results practically equal to published maximum likelihood estimates.