We optimized the baseline length of SQUID-based symmetric axial gradiometer
s using a computer simulation. The signal-to-noise ratio (SNR) was used as
the optimization criteria. We found that in most cases the optimal baseline
is not equal to the depth of the primary source, rather it has a more comp
lex dependence on the gradiometer balance and the ambient magnetic noise. W
e studied both first and second order gradiometers in simulated shielded en
vironments and second order gradiometers in a simulated unshielded environm
ent. The noise source was simulated as a distant dipolar source for the shi
elded case. We present optimal gradiometer baseline lengths for the various
simulated situations.