Methodology - Response of Flood Events to Land Use and Climate Change - page 62

Characteristic of elevation error: To find possible errors, the contour lines for
each modeled DEM were compared with the contour lines of original DEM.
Q-Q plot: In statistics, a Q-Q plot (''Q'' stands for quantile) is a probability
plot, which is a graphical method for comparing two probability distributions by
plotting their quantiles against each other [ 39 ].
Descriptive statistic between original, estimated, and residuals values:In
this part, the Moran's I analysis [ 40 ] was applied to analyze the spatial autocor-
relation analysis. Moran's I is a measure of spatial autocorrelation developed by
Patrick A. P. Moran. Spatial autocorrelation is characterized by a correlation in a
signal among nearby locations in space [ 40 ]. This analysis explores the spatial
patterns in the model especially in model residuals. Spatial Analysis in Macroe-
cology (SAM) was applied to explore spatial patterns in the model [ 41 ].
3.2.3 Hydrological Analysis
Different methods have been proposed and suggested by many researchers to
address the problem of drainage analysis. These methods range from simple DEM
smoothing to flood direction assignment (Dutta et al. 2001; [ 42 - 46 ]). Hutchinson
et al. [ 20 , 47 ] developed one of the most satisfactory methods which assigns to
drainage network direction. The innovation behind this method is a drainage
enforcement algorithm to improve DEM hydrologically. The method is adopted in
an ArcGIS environment, which automatically removes spurious sinks or pits from
a fitted grid (Dutta et al. 2001; [ 20 , 47 ]). To delineate watersheds and stream
networks the process developed by Hutchinson [ 20 , 47 ] is defined as follows:
(a) construct the stream network by selecting a drainage area threshold; (b)
divide the stream network into individual stream links; (c) find the outlet cell at the
lower end of each link; (d) delineate the watershed for each of these outlet cells.
This method works well in hill slopes, but it may have some flaws in flat areas
[ 20 , 47 , 46 ].
To test the quality of DEM hydrologically, the best DEMs from the previous
step (applying interpolation methods) were selected to create the stream network
for each surface using the Arc Hydro tools as an extension of ArcGIS. The
modeled drainage network derived from each DEM from the previous sector was
compared with the observed stream network. The comparison was based on
qualitative evaluation of the level of agreement between the modeled and observed
drainage network.
The Arc Hydro tool was applied to derive several data sets which describe the
drainage networks of the catchment. Raster analysis is performed to generate data
on DEM reconditioning, fill sinks, flow direction, flow accumulation, stream
definition, stream segmentation, catchment delineation, and watershed delineation.
A short definition for each data set is defined in Table 3.2 .