Research note
Hydrology algorithms as constructive granular partitions
Conceptual link: hydrological terrain algorithms as constructive granular partition theory.
Claim: drainage/lake extraction algorithms (Priority-Flood #NNC37P FastFlow depression routing #KLSYT3 implement the “fiat parsing” of elevation fields that Smith & Mark explicitly flag as needing formal treatment #PMUYXM
Correspondences:
Watershed labeling = categorical coverage, the maximally transparent partition class: projection/location total, functional, mutually inverse #UXCVKB #NNC37P boundary recovery #L6E5XU
Partition theory’s required cell tree #RFQ94S is supplied canonically by the merge tree of nested basins joined at saddles #KLSYT3 Granularity/selectivity (“tracing over” #HLB7Y6 = persistence/spill-depth threshold for which depressions count as lakes; systems of partitions at different thresholds are related by refinement per the logic in #4QQD4A
Boundary typology per feature: lake waterline = fiat act with field-determined (minimax spill level) placement; watershed divide = fiat grounded in ridge structure; river/lake junction = pure fiat like North Sea/Atlantic #RVURFD mountain extents = vague/graded fiat parsings #HXHZA9 #7NTEUM formalized via egg-yolk fo/po/no #V4PL3D
Field vs object division of labor: simulation stays field-based #UF9K28 cartography is the object-projection layer #PMUYXM SNAP/SPAN #9YMD2E lake as continuant vs flooding as occurrent; “waterfall” names both a continuant landform and an occurrent process.
Thesis-shaped claim: for water features, unlike mountains, fiat boundaries have canonical, computable locations induced by the field’s critical-point structure — hydrology algorithms are projection devices in the Bittner–Smith sense.