Mediterranean Journal of Modeling and Simulation
Volume 3, Numéro 1, Pages 087-099
2015-03-30
Authors : Haddad Mahdi . Hachemi Hania . Taibi Hebib .
This paper is concerned with the estimation of the quality of the global free air surfaces anomalies DTU10, EGM2008 and ITG-Goce02 in Western Mediterranean Sea. We performed comparisons between independent in-situ marine gravity measurements (a set of 162374 gravity points supplied by the BGI by the International Gravimetric Bureau - BGI) and those estimated from the three surfaces of free air anomalies. This is done by critically looking of differences between in-situ measurements and estimated values. However, in-situ marine gravity data are not free from errors. Thus, outlier data detection processes are an essential first step in this statistical analysis. In this regard, five approaches are used to identify outliers in BGI data, namely: the Grubbs Test, Generalized ESD Procedure, Z-scores and its modified version and the Box plot approach. After removing the suspicious data identified by Boxplot testing which seems to be the bests (about 3\% of data are rejected), a general result appears over the three set of comparisons. Relative to the marine data, the DTU10 and EGM 2008 offer similar differences with a standard deviation of 19.5 mgal, while the ITG-Goce02 surface offers a standard deviation of about 30.5 mgal.
Outliers; Geopotential model; Altimetry-derived gravity anomalies; BGI marine gravity measurements.
Rami R
.
Khelifa S
.
pages 225-236.
Zaoui Rabah
.
pages 912-921.
Laboudi Hesna
.
pages 797-805.
Meriem Othmane
.
pages 813-819.