A constrained, least-squares method for analyzing multiple-station mea
surements of lightning field changes (DELTAEs) is introduced. Previous
methods have attempted to fit the spatial pattern of lightning DELTAE
s using nonlinear models, such as a point charge (Q) or a point dipole
(P) model. With the linear method, the DELTAEs are described not by m
odels but by a general volume charge distribution that is deposited on
a large (40 x 40 x 20 km3) Cartesian grid above the measuring network
. A linear system of equations is used to relate the measured DELTAEs
to the charges that are deposited at each grid point. With this approa
ch, the information content of the measurements can be quantified by a
n eigenanalysis of the covariance matrix of the liner system. Constrai
nts can be used to reduce the infinity of possible solutions to the li
near system and also to reduce systematic biases that can be introduce
d by the method of solution. It is shown that a Landweber iterative me
thod, derived from the general method of steepest descent, can be used
to solve the linear system and that the resulting volume charge distr
ibutions are generally consistent with computer-simulated charge sourc
es, when these sources are over the measuring network. The Landweber i
teration has also provided solutions for natural lightning events that
are consistent with Q- and P-model results.