^{1}

^{2}

^{*}

^{1}

Gravity observations adjustment is studied having in view to take full advantage of the modern technology of gravity measurement. We present here results of a test performed with the mathematical model proposed by our group, on the adjustment of gravity observations carried out on network design. Additionally, considering the recent improvement on instrumental technology in gravimetry, that model was modified to take into account possible nonlinear local datum scale factors, in a 1900 mGal range network, and to check its significance for microgal precision measurements. The data set of the Brazilian Fundamental Gravity Network was used as case study. With about 1900 mGal gravity range and 11 control stations the Brazilian Fundamental Gravity Network (BFGN) was used as case study. It was established mainly with the use of LaCoste & Romberg, model G, gravimeters and new additional observations with Scintrex CG-5 gravimeters. The observables involved in the model are instrumental reading, calibration functions of the gravimeters used and the absolute gravity values at the control stations. Gravity values at the gravity stations and local datum scale factors for each gravimeter were determined by least square method. The results indicate good adaptation of the tested model to network adjustments. The gravity value in the IFE-172 control station, located in Santa Maria, had the largest estimated correction of ?10.4 μGal (1 μGal = 10 nm/s^{2}), and the largest residual for an observed reading was estimated in 0.043 reading unit. The largest correction to the calibration functions was estimated in 6.9 × 10^{-}^{6}mGal/reading unit.

Gravimetry has experienced an important development, in both absolute and relative measurement technique. The modern gravimeters allows observations with some microgals of precision (1 microgal = 10 nm/s^{2}), which means an improvement of at least one order of magnitude in comparison with previous instruments. This fact suggests a review in the technical processing of gravimetric observations, in order to fully explore the modern instrumental technology. This caution is especially appropriate when the points of observation are designed in network structure for reference to future surveys. In this case, the quality of results should be the best possible, ensuring high precision and homogeneity in the gravity values of the network. Therefore, the observations involve, necessarily, the superabundant data acquisition, using two or more instruments. This implies the need for adjustment of observations with appropriate mathematical model and constraints to compensate ambiguities.

Dias and Escobar [