Image Processing Reference
In-Depth Information
Table 6.7
Thermal statistics by
land cover (Wm
−2
) (Gluch et al.
2006
)
Min
Max
Mean
SD
imp_dk
27.82
38.78
31.77
3.43
soil_dk
28.98
32.94
30.37
0.58
soil_lt
27.58
30.70
28.96
0.64
imp_lt
26.07
32.54
28.53
1.69
veg_grass
23.06
27.41
26.08
0.64
veg_trees
22.31
27.54
23.78
0.89
Shadow
20.64
24.88
23.04
1.08
Water
20.65
24.60
21.96
0.72
to land cover types (as in V-I-S), which can be directly measured through remote
sensing. Furthermore, energy flux responds very differently among the three V, I,
and S urban cover types, as well as water (see Table
6.1
). Consequently, values
derived from thermal sensors of vegetation, impervious surface, soil, and water are
directly related to such models as the urban heat island (UHI), evapotranspiration
rates, and other physical phenomena.
Table
6.7
shows the mid-afternoon thermal emittance values by land cover class
according to ATLAS channel 13 expressed in watts per square meter (Wm
−2
). Note
that dark impervious surface has the highest mean, followed by dark soil. Vegetation
is coolest (emitting less energy) of the three terrestrial cover types with trees and
shrubs cooler than grass. Shadows are still cooler and water is the least emissive in
mid-afternoon. It follows that ambient air temperature, as sensed by human popula-
tions, run in the same order.
6.5.2
Thematic Mapper/SPOT-P Data: Storm Runoff Prediction
Chen (
1996
) examined the effect of V-I-S cover types as input to storm runoff
models for runoff prediction of storms of different intensities - using merged
Landsat TM and SPOT-P data re-sampled to 10 m. Widespread conversion of land
cover attendant to urbanization typically increases storm water runoff in amount,
intensity, and in routing. However, there is significant variability across the urban
landscape due to the spatial heterogeneity of cover composition. To be able to predict
runoff rates and patterns would significantly assist in storm water drainage design
and engineering, and other environmental planning.
To perform effective research on the process it is necessary to have empirical
data for both precipitation records and concurrent runoff records. In Salt Lake City
and County runoff records were available at outfall points from specifically engi-
neered drainage basins. Nearby weather stations provided precipitation rates for the
study period. The Soil Conservation Service (SCS) storm runoff model provided
the appropriate input for various sub-classes of V, I, and S. Table
6.8
displays the
results of the SCS model applied to 14 rainfall events across five urban basins over
a 4-year time span. Rainfall amount per event is shown by date. Measured runoff
