Dear José Renato Silva de Oliveira:

This is a very interesting question, then before giving an answer, it is necessary to make some contextual considerations to understand the question.

First, I assumed that if you are using EMAG2 dataset it is because your study objectives are primarily related to Geodynamics. Maybe, your objective of study should be the magnetic characterization of the Lithosphere, since the EMAG2 data has been oriented to know the response of the Earth's Lithosphere.

Remember that EMAG2 is a magnetic intensity anomalies grid at 4 km above the Geoid with 2 arc-min resolution, that combines marine and airborne track line observations, satellite data, and magnetic observatory data to map the location, intensity, and extent of lithospheric magnetic anomalies with 250 km / 300 km wavelengths. It means: the local structures below the minimum wavelength content cannot be resolved.

Remember this because the RPT method is effective to make the best imaging of the anomalies at local scales. (Maybe you can compare the local surveys dataset with the EMAG2 responses by approaches such as upward continuation, spectral analysis and statistical evaluation.)

On the other hand: if you are working in a large area, you must do another attention: the superposition of the influences of lithosphere features and Upper Mantle sources. It means: it makes a merge or combined effects in the anomaly map for a large area where these contributions are likely to overlap in wavelengths.

If my considerations were correct, perhaps you probably have a large study area (ZOI, zone of interest), that is, a project of regional or semi-regional scope and your geological target is the Lithosphere´s magnetic characterization.

In that case, to be able to apply the RTP correction in the Fourier domain, it would be difficult to obtain a constant inclination (I) and declination (D) values ​​of the entire area that best honor the parameters of the pole conversion. If your ZOI has a preferentially elongated shape in a N-S path, since the RTP requires a constant value of I and D, it is probable that there are distortions in the RTP anomalies between the whole ZOI. You must remember that the RTP filter tends to put both, geomagnetic field and the magnetization of the source in vertical position. When reducing to the pole from equatorial latitudes, features normal to the declination direction can blow up due to the strong amplitude correction that is applied. A strong discontinuity occurs in the direction normal to the declination D. By specifying a higher latitude for the amplitude correction alone, this problem can be reduced. This is however at the expense of under-correcting the amplitudes of features along the declination direction.

An amplitude inclination of 90 causes only the phase component to be applied to the data (no amplitude correction), and a value of 0 (zero) causes phase and amplitude components to be applied over the entire range.

If a date is entered to calculate the inclination and declination, and if the grid does not have a coordinate system assigned, you will be prompted to defined it. The IGRF date and parameters, once defined, become the values used in other controls that require them. The field strength is only required for apparent susceptibility calculation but is re-calculated along with the inclination and declination in the other controls, so the values are always synchronized.

When you ask: “Which reference time I use to obtain the Inclination and Declination parameters?”, please, remember that this is an optional field if you use a program such as Geosoft Oasis Montaj.

There are also two theoretical questions that are always very well explain by Professor Alan Reid: the numerical problem and the physical problem. In Reid's words: “The NUMERICAL problem is that the "reduction to pole" filter has a divisor which gets smaller as you get nearer to zero magnetic inclination (i.e. near the geomagnetic equator) and becomes zero at zero inclination. So RTP is unstable at inclinations less than about 20 degrees and infinite at zero inclination. Some commercial software has a stabilization technique which is somewhat effective, but there's no TRUE solution. The PHYSICAL problem is that 2D source bodies with strike exactly aligned N-S have zero induced magnetic anomaly at zero geomagnetic inclination. And there's no filter that can make "something" out of "nothing".

PS: there are another way to walk: Differential Reduction to Magnetic Pole assumes that B and by extension M vary across the survey area, such as you can read in:

Arkani - Hamed L, 1988 Differential reduction to the pole or regional magnetic anomalies, Geophysics, Vol 53, No 12, pp 1529, 1600.

Arkani - Hamed L, 2007 Differential reduction to the pole: Revisited, Geophysics, Vol 72, No 1, pp 13, 20.

Swain C, 2000 Reduction to the pole of regional magnetic data with variable field direction and its stabilisation at the low inclinations. Exploration Geophysics, Vol 31, pp 78-83.

In conclusion: if you are interesting to Lithosphere´s studies, the RTP filter do not have significant differences. See the picture that I attached.

Best regards, Mario E. Sigismondi