Least-squares network adjustment is the process for solving the system of equations that represent all of the survey observations (relative-gravity differences, and absolute-gravity observations), along with any additional parameters that describe relative-gravity meter drift and calibration.
In brief, network adjustment solves the system of equations, represented in matrix form as:
L+ V=AX
where L is an n x 1 vector of relative-gravity (delta-g's) and absolute-gravity measurements , V is a n x 1 vector of residuals, Ais the n x m design matrix, and Xis a m x 1 vector of unknowns, that is, the gravity value at each station and relative-gravity meter parameters that describe drift and meter calibration. The value nis the total number of observations (both delta-g's ad absolute-gravity), and m is the number of unknowns (number of gravity stations plus the number of gravity-meter parameters).
GSadjust can solve the system of equations using the least squares routine in the Python Numpy package (using the linalg.inv method), or using the program Gravnet. The solution to use is selected with the Adjustment > Adjustment type... menu. Prior to adjustment, adjustment options, including whether to include a calibration coefficient, are set using the Network adjustment options dialog.
After successful adjustment, station values are shown in the Adjustment results table in the upper right part of the Network adjustment tab. Statistics about the adjustment results and meter parameters (drift and calibration, if included in the adjustment) is shown in the lower right part.
Note that the delta-g's shown in the Delta table on the Network adjustment tab are effectively static once the Delta table is populated. If changes are made to the samples that comprise each station occupation, or the drift method, the Delta table must be cleared (Adjustment menu > Clear delta table) and repopulated.
Oftentimes it is helpful to adjust the network in pieces (individual loops or surveys) prior to adjusting the campaign (or a survey) as a whole. This can be accomplished using the "Populate delta table - selected survey" and (or) "Populate delta table - selected loop" options on the Adjust menu.
