In fact down near the New Forest, magnetic declination is practically zero.
Latitude: 50° 50' 13.8" N
Longitude: 0° 47' 50.5" W
FISHBOURNE
Magnetic Declination: +0° 0'
Declination is POSITIVE (EAST)
Inclination: 65° 54'
Magnetic field strength: 48693.6 nT
https://www.magnetic-declination.com/
There seems to be some confusion about declination, variation, etc. Sites like the one above, when they quote declination, are quoting what sailors call variation, which is the difference between magnetic north (at a given place and time) and true north. Magnetic north is variable with respect to place and time but true north, by definition, points along the local meridian towards the geographic north pole and is essentially invariant. Accordingly, true north is the north which all navigators, surveyors, cartographers and earth scientists can agree about, world wide. It is, in that sense, the gold standard, provided the underlying ellipsoid (datum) is also agreed. The modern gold standard for that is WGS84. The OSGB36 datum on which our OS maps are based is different but the difference is of little relevance to a walker.
Notwithstanding the above, when we use OS maps we use grid north, not true north. These are the same only on the meridian at 2 degrees west, or 400km easting, in OS terms. Because of the way the grid is defined, these two norths are different everywhere else. Where I am, true north is 2 degrees east of grid north. I am lucky because magnetic north here is currently 1.28 degrees west of true north, which means my compass points to grid north within about three quarters of a degree, which I can live with!
Someone walking near Cranbrook, in Kent, is not so lucky. There, true north will be 2.5 degrees west of grid north and the local variation is 0.43 degrees east. The local grid magnetic angle (GMA) there is therefore about 2 degrees west. Some may still regard this as negligible but I don't. It's comparable with the reading error of a mirror compass and greater than that of a prismatic sighting compass. More importantly, it's a systematic error which will never be "averaged out" no matter how many compass readings are taken. It's enough of a GMA to be taken properly into account, in my view.