In order to plot the data correctly over a map, you must know the exact map projection and all its parameters. You will then plot the data over this map projection without needing to do any kind of mathematical transformation. To plot native grids, you must set the resource tfDoNDCOverlay = True, so that the data doesn’t go through a mathematical transformation to be overlaid on the map. Additionally, you commonly want to set the resource mpLimitMode=’Corners’ to specify the area of the map to zoom in on. This resource must be accompanied by resources for setting the four corners of the map area to view: mpLeftCornerLatF, mpLeftCornerLonF, mpRightCornerLatF, and mpRightCornerLonF. If your data comes with 2D lat/lon arrays, you then have the option of plotting it on a different map projection. In this case, you do not want to set tfDoNDCOverlay (it is False by default). Instead, you will need to set the special lat2d/lon2d attributes. See example 3 on the Ice Model page, example 1 on the EASE page, or example 3 on the Raster contouring page, The four pieces of information that are required for this projection (with example values) are: mpRelativeCenterLon = True mpCenterLonF = lon(147,172) mpRelativeCenterLat = True mpCenterLatF = 90. For this grid, you must set the center lat (-90 or 90) and center lon, sometimes referred to as the ‘orientation’. A common way that these grids are described is to give the orientation, the delta X and delta Y at some arbitrary latitude and the coordinates of the pole in integer grid coordinates. Coming up with corner lat/lons is difficult when this is the only information provided. mpProjection = ‘Mercator’, sets the projection. This projection requires that the ‘corners’ method of specifying the map boundaries be used. (see example 1 above). Source.