Tree Survey Example

The importance of specifying instrument corrections, even when they are seemingly very small, is often not fully appreciated -- if only because there exist large project data sets for which the compass corrections, for example, were never recorded.  It is not even known, in some cases, whether or not different instruments were used in different parts of the survey.  Otherwise, we could at least try to estimate the corrections based on the consistency of the combined data.  Also unfortunate is the practice of reversing shots after the fact, so that the FROM-TO name sequence that the computer sees doesn't necessarily imply which station the clinometer reading was taken at.

An illustrative example (included with Walls) is TREE.WPJ, a Suunto and fiberglass tape tree survey of a 12-acre lot.  It is a 136-loop mesh of compass and tape vectors connecting 5 stakes at known true north relative positions on the outer boundary.  Since the relative elevations of the stakes are not known, the vertical loop count is smaller: 131.  There are 261 traverses, all but 35 consisting of just one vector.  The average shot length is 62 ft.  Obviously, surface surveying in broad daylight is not exactly comparable to cave surveying, but the results can be interesting nonetheless.  In this case, a flagged stick was held vertically at each station; other conditions hardly varied at all.

With the magnetic declination set at +6.5 and the independently determined clinometer correction (IncV) set at +0.3, we see the following effects of the compass correction (IncA) and vertical adjustment (IncH) on consistency as measured by the horizontal and vertical UVEs (correction units are degrees and feet):

IncA        UveH        IncH        UveV        

+2.00        0.69        -0.50        0.23        

+1.50        0.19        -0.30        0.11        

+1.30        0.16        -0.20        0.08        

+1.00        0.27        -0.10        0.11        

+0.50        0.93        -0.00        0.17        

In this way, a compass correction of +1.30 was obtained (assuming the declination is correct).  Detaching the survey from the stakes drops the horizontal UVE from 0.16 to 0.15, suggesting there is nothing wrong with the fixed stake locations.  The vertical UVE is very small -- not unheard of in a mostly horizontal survey.  While a few bad readings could easily affect the size of UveV, the vertical adjustment that minimizes it (IncH = -0.2) would be less affected.  A systematic vertical positioning error of just over two inches is certainly plausible and accounting for it is enough to change the relative ranking of a few traverses.