Methodology - Response of Flood Events to Land Use and Climate Change - page 63

Table 3.2 Definition and description of several data sets to create drainage networks (modified
and summarized based on the Arc Hydro tools manual)
Data set Definition & objective

DEM reconditioning To modify the DEM based on observed drainage networks

Fill sink To fill the sinks in a grid

Flow direction To calculate the direction of flow for each grid

Flow accumulation To analyze the flow accumulation grid that contains the accumulated

number of cells located in the upstream of a selected cell

Stream definition To define a stream grid value as ''1'' for all the cells in the input flow

accumulation grid that have a value greater than the given threshold.
The rest of cells in the stream grid contain no data or zero

Stream segmentation To define a grid of stream segments with unique identification

Catchment grid
delineation

To create a grid code indicating to which catchment the cell belongs

Catchment polygon
processing

To create the raster and the vector data stored in a feature dataset named
''Layers''

Drainage line
processing

To convert the input stream link grid into a drainage line feature class

Table 3.3 Error Matrix resulting from the comparison between modeled and observed drainage
networks (modified table according to Peralvo [ 7 ])
Modeled

Observed River No River

a b

River Correct river Omission error

d Correct no river
a Representing agreement in river in both modeled and observed is ''correct river''
b Representing river in observed not defined as such in the modeled river showing ''omission
error''
c

c Commission error

No River

Representing river in modeled not present in observed indicate ''commission error''

d

Agreement in non river in modeled and observed river indicate ''correct no river''

The results of modeled and observed drainage networks derived from all
interpolation methods were estimated visually and tallied into an error matrix
(Table 3.3 ).
According to Table 3.3 , representing agreement in river in both modeled and
observed is ''correct river''; representing river in observed not defined as such in
the modeled river showing ''omission error''; representing river in modeled not
present in observed indicate ''commission error'', and agreement in nonriver in
modeled and observed river indicate ''correct no river''. Regeneration of DEM,
visual interpretation, and hill shade analysis was done to check any unnecessary
pits or sinks in the surface.
Additionally, Arc Hydro tools were applied to generate drainage networks
correctly and to perform pour point tests to check any unnatural variations within
the train.