The North American gravity database as well as databases from Canada, Mexico, and the United States are being revised to improve their coverage, versatility, and accuracy. An important part of this effort is revising procedures for calculating gravity anomalies, taking into account our enhanced computational power, improved terrain databases and datums, and increased interest in more accurately defining long-wavelength anomaly components. Users of the databases may note minor differences between previous and revised database values as a result of these procedures. Generally, the differences do not impact the interpretation of local anomalies but do improve regional anomaly studies. The most striking revision is the use of the internationally accepted terrestrial ellipsoid for the height datum of gravity stations rather than the conventionally used geoid or sea level. Principal facts of gravity observations and anomalies based on both revised and previous procedures together with germane metadata will be available on an interactive Web-based data system as well as from national agencies and data centers. The use of the revised procedures is encouraged for gravity data reduction because of the widespread use of the global positioning system in gravity fieldwork and the need for increased accuracy and precision of anomalies and consistency with North American and national databases. Anomalies based on the revised standards should be preceded by the adjective "ellipsoidal" to differentiate anomalies calculated using heights with respect to the ellipsoid from those based on conventional elevations referenced to the geoid.

S U M M A R YThe International Association of Geomagnetism and Aeronomy (IAGA) on 2004 December 12 has released the 10th-Generation International Geomagnetic Reference Field-the latest version of a standard mathematical description of the Earth's main magnetic field used widely in studies of the Earth's deep interior, its crust, ionosphere and magnetosphere. The coefficients were finalized by a task force of IAGA. The IGRF is the product of a large collaborative effort between magnetic field modellers and the institutes involved in collecting and disseminating magnetic field data from satellites and observatories around the world.

Following the call for candidates for the 10th generation IGRF, we produced and submitted three main field and three secular variation candidate models. The candidates are derived from parent models which use a standard quadratic parameterisation in time of the internal Gauss coefficients. External magnetospheric fields are represented by combined parameterisations in Solar Magnetic (SM) and in Geocentric Solar Magnetospheric (GSM) coordinates. Apart from the daily and annual variations caused by these external fields, the model also accounts for induction by Earth rotation in a non-axial external field. The uncertainties of our candidates are estimated by comparing independent models from CHAMP and Ørsted data. The root mean square errors of our main field candidates, for the internal field to spherical harmonic degree 13, are estimated to be less than 8 nT at the Earth's surface. Our secular variation candidates are estimated to have root mean square uncertainties of 12 nT per year. A hind-cast analysis of the geomagnetic field for earlier epochs shows that our secular acceleration estimates from post-2000 satellite data are inconsistent with pre-2000 acceleration in the field. This could confirm earlier reports of a jerk around 2000.0, with a genuine change in the secular acceleration.