IMPROVING GEOMAGNETIC-FIELD MODELS FOR THE PERIOD 1980-1999 USING ORSTED DATA

The Danish satellite Orsted is due to be launched in 1998, and should provide, for the first time since the Magsat mission (1979-1980), a de nse and global coverage of the Earth's surface with vector measurement s of the magnetic field. In this paper, we compare the expected error in the main field models computed for the 1970-1999 time interval usin g observatory data, with or without the a priori information given by the knowledge of the field at both Magsat and Orsted epochs. This work is based on the reasonable hypothesis that the main field models deri ved from Orsted data will be as accurate as the Magsat models. The a p riori information given by the Magsat and Orsted models is based on a linear behaviour of the rate-of-change of the field throughout this pe riod, plus a noise level which can be estimated as a function of time and degree from past field changes. The expected error in the models c omputed for the 1980-1999 period with a priori information appears to be significantly smaller than the expected error in the models compute d without this information. This result is related to the heterogeneou s distribution of the observatories over the Earth surface. Consequent ly, when the Orsted data is available, improved models can be computed for the 1980-1999 period particularly in regions without observatory data. This method with a priori information may allow the use of the s ame set of observatories throughout the entire period. Indeed, our met hod alleviates the requirement of a very dense data distribution.