Point Classification (Ridges and valleys)

• Selected on Geomorphometry analysis of Analyze menu
• On Ridges choice of Calculate menu.
• Used with DEM map masking

Now also includes:

• Pennnock Landform Classification (Pennock,D.J., Zebarth,B.J., and De Jong,E., 1987. Landform Classification and Soil Distribution in Hummocky Terrain, Saskatchewan, Canada. Geoderma, 40: 297-315.), using WhiteBox.
• Iwashashi and Pike Classification

The top parameters apply only to the Wood (1996) algorithm.  Because it uses 4 parameters, this can take some adjustments to get the tuning right for the DEM and geomorphic characteristics. Mask radius applies the Simple and Simple 2 algorithms.  For each point classified as a ridge, a radius of this size will be masked. Algorithm:  Wood   Simple   Simple 2 Ridge valley tolerance: for the Simple 2 algorithm, how many neighbors must be higher/lower.  Smaller values lead to more classified points. Draw shows the ridges and other point classifications, and optionally shows counts Ridge map: creates grid using the current algorithm, with the elevations of the points classified as ridges and peaks. Valley map: creates grid using the current algorithm, with the elevations of the points classified as pits and valleys. Ridge mask creates mask with the ridge and peak points.  This can be used with map masking. Valley mask creates mask with the pit and valley points.  All points map: creates  grid using the current algorithm, with each category with a different number. High/Low neighbors Close: exit form.  Do not hit this early.

Draw Ridge/valley Map
Wood (1996) algorithm	Simple algorithm	Simple 2 algorithm  (tolerance 5)

Draw Simple 2 Ridge/valley Map
Simple 2 algorithm (tolerance 7)	Simple 2 algorithm  (tolerance 6)	Simple 2 algorithm (tolerance 5, same as above)

Ridge Mask
Wood (1996) algorithm	Simple algorithm	Simple 2 algorithm

Simple:

 if (z < znw) and (z < zn) and (z < zne) and  (z < zsw) and (z < zs) and (z < zse) and (z < zw) and (z < ze) then Result := PitPoint
else if (z = znw) and (z = zn) and (z = zne) and (z = zsw) and (z = zs) and (z = zse) and (z = zw) and (z = ze) then Result := FlatPoint
else if (z > znw) and (z > zn) and (z > zne) and (z > zsw) and (z > zs) and (z > zse) and (z > zw) and (z > ze) then Result := PeakPoint
else if ((zn > znw) and (zn > zne)) and ((z > zw) and (z > ze)) and ((zs > zsw) and (zs > zse)) then Result := RidgePoint {NS trend}
else if ((zw > znw) and (zw > zsw)) and ((z > zn) and (z > zs)) and ((ze > zne) and (ze > zse)) then Result := RidgePoint {WE trend}
else if ((zn < znw) and (zn < zne)) and ((z < zw) and (z < ze)) and ((zs < zsw) and (zs < zse)) then Result := ValleyPoint {NS trend}
else if ((zw < znw) and (zw < zsw)) and ((z < zn) and (z < zs)) and ((ze < zne) and (ze < zse)) then Result := ValleyPoint {WE trend}
else Result := OtherPoint;
 

Simple 2:

Of 8 neighbors, count those higher and lower

 if Higher = 8 then Result := PitPoint
else if Lower = 8 then Result := PeakPoint
else if (Higher = 0) and (Lower = 0) then Result := FlatPoint
else if (Higher >= RidgeValleyTolerance) then Result := ValleyPoint
else if (Lower >= RidgeValleyTolerance) then Result := RidgePoint
else Result := OtherPoint;

Compare to:

• TPI: Topographic position index

Last revision 3/2/2023