[1] It was the subject of a treatise by Howard and Mitchell in 1985, who were considering the growth and varietal temporal and spatial variability found in forests, but recognized that their work also had application to farming, and the relatively new science (at that time) of precision agriculture.
The premise of Howard and Mitchell is that landforms, or features of the land's 3D topography significantly affect how and where plants (or trees in their case) grow.
There is already a volume of work, although they don't use the term phytogeomorphology specifically, that considers farm field terrain attributes as affecting crop yield and growth, Moore et al. (1991)[2] provide an early overview of the application of terrain features to precision agriculture, but one of the earliest references to this phenomenon in farming is that of Whittaker in 1967.
Work in this area has been happening for some years (see Reuter et al., (2005),[7] Marquas de Silva et al., (2008), and especially Moore et al., (1991)), but it is slow and sometimes tedious work that necessarily involves multiple years of data, very specialized software tools, and long compute times to produce the resulting maps.
The idea is to create a 'recipe map' for variable rate farm machinery to deliver the exact quantity of amendments required at that location (within that zone of the field).