The problem of validity of the absolute gravimetry data on the mirrored data base AGrav of BGI (France) and BKG (Germany) is not so simple.
Ideally the data of absolute gravity measurements are valid if the absolute gravimeter is calibrated or participated in the comparisons organized ubder the auspices of International Committee of Weights and Measures and Regional Metrology Organizations (e.g. EURAMET) with a proper Technical Protocol (see e.g. The 8th International Comparison of Absolute Gravimeters 2009: the first Key Comparison (CCM.G-K1) in the field of absolute gravimetry, Z.Jiang et al, METROLOGIA, vol.49 (2012), pp 666–684; The European Comparison of Absolute Gravimeters 2011 (ECAG-2011) in Walferdange, Luxembourg: results and recommendations, O. Francis et al, Metrologia, vol. 50 (2013), pp 257–268 ).
Currently the system of calibrations of absolute gravimeters is at the beginning of its development. There are some National Metrology Institutes and Designated Institutes (see on the website of the BIPM http://kcdb.bipm.org/) which are declared their capabilities for the calibration of absolute gravimeters.
As for all the data of absolute gravity measurements on the data base obtained even by uncalibrated gravimeters we have accept them as they are taking into account the typical uncertainty in the measurements by appropriate type of gravimeters.
I am sorry for so long answer but iy is not easy to respond in a few words to such Q.
Do not hesitate ask me any question on this subject.
I mean that the the experimental data are valid if these data are obtained using the measuring instruments with the confirmed performance properties or with the confirmation that legal requirements of a measuring system are achieved (International vocabulary of metrology — Basic and general concepts and associated terms (VIM); see on http://www.bipm.org/en/publications/guides/)
My first question is for Walid because he is the first to start the discussion on validity of BGI data. I think he wants to use the Bouguer anomaly data for geological purposes not the full gravity acceleration(g). Error on (BGI) Bouguer gravity data in the case of North Africa is probably (1-2 mGal) or even more. But for this regional scale data, this error is acceptable.
The BGI data will help you on large scale studies: Regional Tectonic features , regional lineaments. For study of karst (small area), you need detailed gravity studies.
Well I come to this conversation a bit later than others... At BGI as well as here in the USA (with NGA and NGS), there are three 'categories' of gravimetric data : absolute gravimeters, base stations and point data (latter being widespread observations typically used for Bouguer anomalies or geoid determination). Each have their own concerns for precision (few microGals, tenths of milliGals and a few milliGals), tracability in metrologic sense and validity over time. As Leonid describes the requirements for absolute meters and the national networks they support require inter-comparisons and good metrology; and then can demonstrate temporal change of gravity. The latter two rely on 'freshness/staleness', meaning can you trust tracability and lack of temporal change over possibly decades to give researcher valid current data. At the milliGal level staleness is likely not a worry. And if a Bouguer anomaly researcher is merging his data with 40 year old data; then using a 50 year old base station in same area makes sense.