On the estimation of radar rainfall error variance

One of the major problems in radar rainfall (RR) estimation is the lack of accurate reference data on area-averaged rainfall. Radar–raingauge (R–G) comparisons are commonly used to assess and to validate the radar algorithms, but large differences of the spatial resolution between raingauge and rada...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Advances in water resources 1999-02, Vol.22 (6), p.585-595
Hauptverfasser: Ciach, Grzegorz J., Krajewski, Witold F.
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:One of the major problems in radar rainfall (RR) estimation is the lack of accurate reference data on area-averaged rainfall. Radar–raingauge (R–G) comparisons are commonly used to assess and to validate the radar algorithms, but large differences of the spatial resolution between raingauge and radar measurements prevent any straightforward interpretation of the results. We assume that the R–G difference variance can be partitioned into the error of the radar area-averaged rainfall estimate, and the area-point background originating from the resolution difference. A robust procedure to decompose these components, named the error separation method (ESM), is proposed, discussed, and demonstrated. If applied to a sufficiently large sample, it allows the estimation of the radar error part and description of the uncertainties of hydrological radar products in rigorous statistical terms. An extensive data set is used to illustrate the ESM application. Proportion of the error components in the R–G difference variance is studied as a function of rainfall accumulation time. The intervals from 5 min through 4 days are considered, and the radar grid resolution of 4 × 4 km is assumed. The results show that the area-point component is a dominant part of the R–G difference at short time scales, and remains significant even for the 4-day accumulations.
ISSN:0309-1708
1872-9657
DOI:10.1016/S0309-1708(98)00043-8