Comparison of LiDAR-Derived Curvature and Corn and Soybean Yield
Useful curvature maps for a field can be derived from LiDAR, but the data may need to be further smoothed after the initial download to make it more useful for certain application (as previously shown). In the examples shown through the link to the PDF that follows, curvature is compared to corresponding corn and soy yield maps. Yield are to the extent of Landsat pixels that are valid and only contain the more valid yield data as was previously described previously in the about Landsat yield correlation section. Fields with at least thirty contiguous pixels of yield data are included. Pixel boundaries are included because they are useful for reference. (If a very small gap can be seen between the yield map and the outside pixel boundary it is because the raster yield map is made to extent of 4-meter spaced yield points which can leave a small, but sometimes visible gap with a pixel boundary, while curvature grids have 2.5 foot spacing. Yield monitor data are cleaned by virtually the same method described in yield monitor data cleaning section; a main difference being that the initial extent of the yield dataset is to valid pixels. Curvature ranges from the most concave (darkest shade), to the most convex (brightest shade). In most cases, it can be seen that the higher convex areas correspond to lower yielding areas. The following is a link to a PDF that shows how LiDAR curvature spatially compares to corn and soybean yield:
(pdf) LiDAR curvature and yield comparison