In:
Journal of Climate, American Meteorological Society, Vol. 23, No. 19 ( 2010-10-01), p. 5325-5331
Abstract:
Instrumental daily series of temperature are often affected by inhomogeneities. Several methods are available for their correction at monthly and annual scales, whereas few exist for daily data. Here, an improved version of the higher-order moments (HOM) method, the higher-order moments for autocorrelated data (HOMAD), is proposed. HOMAD addresses the main weaknesses of HOM, namely, data autocorrelation and the subjective choice of regression parameters. Simulated series are used for the comparison of both methodologies. The results highlight and reveal that HOMAD outperforms HOM for small samples. Additionally, three daily temperature time series from stations in the eastern Mediterranean are used to show the impact of homogenization procedures on trend estimation and the assessment of extremes. HOMAD provides an improved correction of daily temperature time series and further supports the use of corrected daily temperature time series prior to climate change assessment.
Type of Medium:
Online Resource
ISSN:
1520-0442
,
0894-8755
DOI:
10.1175/2010JCLI3499.1
Language:
English
Publisher:
American Meteorological Society
Publication Date:
2010
detail.hit.zdb_id:
246750-1
detail.hit.zdb_id:
2021723-7
Permalink