6.6 Control Networks, Observation Reductions, and Quality

The Control Hierarchy

Control surveys establish the framework of horizontal and vertical positions on which boundary, topographic, route, construction, and mapping surveys depend. The governing rule is work from the whole to the part and from higher order to lower order: primary (high-accuracy) control is established first, and secondary control is tied to it. You never densify or extend control from points weaker than the work you are about to perform, because errors propagate downward.

Control points are monumented and documented so they can be recovered. In the US, the National Geodetic Survey (NGS) maintains the National Spatial Reference System (NSRS) of passive marks and Continuously Operating Reference Stations (CORS) that anchor GNSS work. The Online Positioning User Service (OPUS) returns NSRS coordinates for a static GNSS occupation by processing it against CORS.

Classes of control include:

• Horizontal control - latitude/longitude or grid northing/easting.
• Vertical control - benchmarks with orthometric elevations.
• 3D control - modern GNSS marks carrying horizontal position and ellipsoid height together.

Redundancy, Reductions, and Adjustment

Redundancy is the presence of more observations than the minimum needed to fix the unknowns. It is the engine of quality: only redundant observations allow you to compute closures, detect blunders, and run a least-squares adjustment that yields a best estimate plus error statistics. A traverse with no redundancy can be computed but cannot be checked.

Before adjustment, raw field observations are reduced into comparable quantities:

Reductions must precede adjustment: you cannot least-squares-adjust a network whose observations are still a mix of slope, ground, and grid values. After reduction, the least-squares adjustment distributes random error, and the resulting residuals (observation minus adjusted value) reveal whether observations agree and whether a blunder remains.

Judging Network Quality

Network quality is more than a single closure number. The FS exam expects you to weigh several factors together:

• Geometry - well-distributed points, strong angles, and (for GNSS) good satellite geometry/low PDOP.
• Redundancy - enough extra observations to check and adjust.
• Instrument and field procedures - calibrated equipment, multiple sets, balanced backsight/foresight, and atmospheric corrections.
• Adjustment statistics - small, randomly signed residuals and a reference variance near 1 indicate the error model fits.
• Accuracy classification - traditional NGS work is rated by order and class (e.g., first-order, second-order class I); modern GNSS work is reported as network accuracy (relative to the NSRS datum) and local accuracy (relative to neighboring marks), at the 95% confidence level.
• Documentation - datum, epoch, geoid model, units, combined factors, and adjustment report - without metadata, even an accurate network is unusable to the next surveyor.

A low misclosure does not by itself prove a good network if redundancy is poor; conversely, a slightly larger misclosure with strong geometry and clean residuals may be the more trustworthy result.

Closures, Standards, and Modern GNSS Control

The FS exam connects network quality to specific, computable checks. For a closed traverse, the angular misclosure is compared with the sum of interior angles, (n - 2) x 180 deg, and a typical specification limits it to a small multiple of the instrument's least count times the square root of the number of angles. The linear misclosure is the length of the closure vector found from the sum of latitudes and departures, and dividing the perimeter by that misclosure gives the relative precision (for example 1:10,000), a dimensionless ratio that scales the closure to the size of the survey.

For a level loop, the allowable misclosure is commonly expressed as a constant times the square root of the loop distance, reflecting that random leveling error accumulates with distance, not linearly.

Modern control is increasingly GNSS-based, which changes how quality is reported:

• CORS and OPUS anchor a static occupation to the NSRS without running conventional traverse from a distant mark.
• Network (RTN) and RTK methods deliver real-time positions whose quality depends on baseline length, satellite geometry (PDOP), and the geoid model used for heights.
• Results are reported as network accuracy (relative to the datum) and local accuracy (relative to neighbors) at 95% confidence, replacing the older order-and-class language for new work.

Reading any FS control scenario, first ask what is redundant, what closure it produces, and whether the metadata is complete - the same three questions that define a defensible network.