HR: 1340h
AN: V33C-1470 [Abstracts]
TI: Volcanic Landform Classification of Iwate Volcano from DEM-Derived Thematic Maps
AU: * PRIMA, A O
EM: prima@soft.iwate-pu.ac.jp
AF: Faculty of Software & Information Science, Iwate Prefectural University, 152-52, Sugo, Takizawa, Iwate,
020-0193
Japan
AU: YOSHIDA, T
EM: tyoshida@mail.tains.tohoku.ac.jp
AF: Institute of Mineralogy, Petrology and Economic Geology, Graduate School of Sciences, Tohoku University,
Aramaki-aza-Aoba, Aoba-ku, Sendai, 980-8578
Japan
AB:
Over the last three decades, digital elevation models (DEMs) have been developed as surface data instead of contour lines to
allow numerical analysis or modeling of terrain by computer. DEMs have allowed the development of algorithms to rapidly
derive slope, relief, convexity, concavity and aspect of any points of surface, and also have allowed the definition of a
number of new morphometric measures i.e. openness (Yokoyama et al., 2002). Openness is an angular measure of the relation
between surface relief and horizontal distance. Openness has two viewer perspectives. Positive values, expressing openness
above the surface, are high for convex forms, whereas negative values describe this attribute below the surface and are high
for concave forms. The emphasis of terrain convexity and concavity in openness maps facilitates the interpretation of
landforms on the Earth_fs surface. Prima et al. (2003) proposed automated landform classification using openness and slope
with genetic factors. This method had been proved to produce good classification for constructional (alluvial plains,
alluvial fans and volcanoes) and erosional (hills and mountains) landforms. The capability of this method to classify
landforms from DEMs with genetic factors is important because it allows landform evolution to be numerically analyzed. In
this study, we adopted this method to classify volcanic landforms of Iwate Volcano from Honshu, Japan, where volcanic
landforms were categorized referring to geological map of Iwate Volcano (Doi, 2000). This process took three steps. First,
the characteristic of each category was evaluated against the mean and standard deviation of slope, and both positive and
negative openness, in two dimensional feature spaces. Second, the characteristic of each category were observed and the
combinations of mean and standard deviation of slope and openness showing high separabilities were selected. We found that
the standard deviation of slope, positive and negative openness yielded high separabilities of each category and indicated
consistency between the trend of categories_f distribution and their geological successions. Third, Mahalanobis distance was
used as classification rule to automatically classify the volcanic landforms of Iwate Volcano with those categories. The
result shows fine interpretation of volcanic landforms of Iwate Volcano according to their geological successions. The
northeastern sector of Iwate Volcano that has relatively young strata was clearly identified against other sectors of the
volcano. Although some misclassification occurred in places where the ages of landform formation are relatively close, we
considered that the present result is provisionally acceptable because the classified landforms in major accurately
replicated many components of volcanic landforms of Iwate Volcano.
DE: 8404 Ash deposits
DE: 8450 Planetary volcanism (5480)
DE: 9320 Asia
DE: 9604 Cenozoic
DE: 1824 Geomorphology (1625)
SC: Volcanology, Geochemistry, Petrology [V]
MN: 2004 AGU Fall Meeting