Assessment of a Rapid Approach for Estimating Catchment Areas for Surface Drainage Lines - PowerPoint PPT Presentation

Loading...

PPT – Assessment of a Rapid Approach for Estimating Catchment Areas for Surface Drainage Lines PowerPoint presentation | free to download - id: 72fc1-ZDc1Z



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

Assessment of a Rapid Approach for Estimating Catchment Areas for Surface Drainage Lines

Description:

Lawrence Stanislawski, Science Applications International ... Lower Prairie Dog Town Fork Red. Interior Highlands of Ozark Plateaus. Hilly Humid. 10290107 ... – PowerPoint PPT presentation

Number of Views:49
Avg rating:3.0/5.0
Slides: 33
Provided by: larrystan
Learn more at: http://cegis.usgs.gov
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: Assessment of a Rapid Approach for Estimating Catchment Areas for Surface Drainage Lines


1
Assessment of a Rapid Approach for Estimating
Catchment Areas for Surface Drainage Lines
  • Lawrence Stanislawski, Science Applications
    International Corporation (SAIC)
  • Michael Finn, U.S. Geological Survey
  • E. Lynn Usery, U.S. Geological Survey
  • Mark Barnes, U.S. Geological Survey

ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
2
  • Brief overview of the National Hydrography
    Dataset (NHD)
  • Generalization Process for NHD
  • Pruning of Drainage Network
  • Preprocessing Requirements
  • Methods
  • Thiessen-polygon-derived (TPD) catchments
  • Elevation-derived (ED) catchments
  • Catchment comparisons
  • Results
  • Between subbasin comparisons
  • Within subbasin comparisons
  • Extended results
  • Summary

ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
3
NHD Features
Artificial Paths
ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
4
NHD Features
ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
5
NHD Features
Example surface water flow network (NHDFlowline
feature class)
ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
6
  • National Hydrography Dataset (NHD)
  • Vector data layer of The National Map
    representing surface waters of the United States.
  • Includes a set of surface water reaches
  • Reach significant segment of surface water
    having similar hydrologic characteristics, such
    as a stretch of river between two confluences, a
    lake, or a pond.
  • A unique address, called a reach code, is
    assigned to each reach, which enables linking of
    ancillary data to specific features and locations
    on the NHD.
  • Reach code from Lower Mississippi subbasin
  • 08010100000413
  • region-subregion-accounting unit-subbasin-reach
    number

08010100000413
08010100000696
ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
7
  • National Hydrography Dataset (NHD)
  • Divided and distributed at watershed basin and
    subbasin boundaries.
  • Stored in an ArcGIS geographic database
    (geodatabase) model.
  • Three levels of detail (resolutions)
  • Medium (1100,000-scale source)
  • only complete layer
  • High (124,000-scale source)
  • 90 percent complete
  • Local (112,000 or larger source)
  • (expect to include in high resolution layer)

Subregions along northern shore of Gulf of Mexico
ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
8
  • NHD Generalization
  • Develop a generalization strategy that can be
    implemented on subsets of the NHD.
  • Strategy should produce a dataset in the NHD
    model format that maintains
  • feature definitions,
  • reach delineations,
  • feature relationships, and
  • flow connections between remaining generalized
    features.
  • Extracted dataset should function with NHD
    applications
  • less detail
  • faster processing speed
  • Extracted level of detail
  • user-specified

Development of such a generalization process
could eliminate the need to store and maintain
all but the highest resolution NHD data layer.
ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
9
  • NHD Generalization Strategy
  • Base data highest resolution NHD that covers
    desired area.
  • Feature pruning removal of features that are
    too small for desired output scale.
  • select a subset of network features
  • select a subset of area features
  • remove point features associated with pruned line
    or area features
  • Feature simplification
  • removal of vertices
  • aggregation, amalgamation, merging, linearization
    of area features, etc.

ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
10
  • NHD Generalization

Catchment The area associated with a segment of
a drainage network is referred to as the
segments catchment area, or just catchment.
Surface runoff in the catchment flows into the
associated network segment.
Catchments (cyan) associated with each network
segment (red) of a hydrographic network. The
network pruning strategy of our NHD
generalization process is based on upstream
drainage area, which requires catchment area
estimates for each network segment.
ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
11
  • NHD Generalization

Upstream drainage area (UDA) for any network
segment is the sum of all upstream catchment
areas, including the segment of interest.
For instance, the UDA for the network segment
marked with the green square is the yellow shaded
area ( 11.2 sq km).
ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
12
  • Generalization Network Pruning Example

Gasconade-Osage subregion (1029) falls in the
Interior Plains and Interior Highlands
physiographic divisions.
ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
13
  • Generalization Network Pruning Example
  • Feature pruning removal of feature that are too
    small for desired output scale.
  • select a subset of network features

Pruning test on Gasconade-Osage subregion
(1029) Green, blue, red 1100,000 Blue, red
1500,000 Red 12,000,000
ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
14
  • NHD Generalization
  • Preprocessing requirements for network pruning
  • Catchment area estimates
  • Upstream drainage area estimates
  • Values not available for high resolution NHD
    Layer
  • Therefore, we developed a rapid approach to
    estimate catchment areas using Thiessen polygons.

ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
15
MethodsThiessen-polygon-derived (TPD) catchments
ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
16
MethodsComparison to elevation-derived (ED)
catchments
  • Each ED catchment is precisely geospatially
    associated to one segment of a surface drainage
    network that is derived from the same elevation
    model.
  • Associating ED catchments to each segment of a
    hydrographic network, such as that in the NHD, is
    a complex and imprecise process.
  • Thiessen-polygon derived catchments can be
    precisely associated with individual segments of
    any network regardless of how the network is
    derived.
  • Generate surface drainage network from an
    elevation model.
  • Compute ED catchments for ED network.
  • Compute TPD catchments for ED network.
  • 4. Compare ED and TPD catchments through an
    overlay process.

ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
17
Methods
Study subbasins
Subbasin name State NHD subbasin number Regime Physiographic division1
Upper Suwannee FL, GA 03110201 Flat Humid Atlantic Plain of Coastal Plain
Lower Beaver UT 16030008 Flat Dry Intermontane Plateaus of Basin and Range
Pomme De Terre MO 10290107 Hilly Humid Interior Highlands of Ozark Plateaus
Lower Prairie Dog Town Fork Red TX 11120105 Hilly Dry Interior Plains of Great Plains and Central Lowland
South Branch Potomac WV 02070001 Mountainous Humid Appalachian Highlands of Valley and Ridge
Piceance-Yellow CO 14050006 Mountainous Dry Intermontane Plateaus of Colorado Plateaus
  • Six NHD subbasins that fall in one of six regimes
    based on climate and topography were evaluated.
  • For each subbasin
  • A 30-meter resolution DEM was extracted from the
    National Elevation Dataset.
  • ED Streams and catchments were derived for
    several (7) stream formation thresholds.
  • TPD catchments were generated for all ED network
    segments
  • ED catchments and TPD catchments were compared
    through a spatial union.

ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
18
Methods
  • Catchment comparison computations
  • For each TPD catchment in all networks
  • percent correct,
  • percent omission, and
  • percent commission
  • For all TPD catchments in each stream-formation
    threshold
  • Mean percent correct
  • Mean percent omission, and
  • Mean percent commission
  • Total percent correct
  • For all stream-formation thresholds in each
    subbasin
  • Average mean percent correct
  • Average mean percent omission, and
  • Average mean percent commission
  • Average total percent correct

Thiessen-derived catchment (red outline)
overlaying associated elevation-derived catchment
(gray outline) with correct area in green, and
areas of commission error in purple and omission
error in pink .
ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
19
Methods
Computations Coefficient of areal correspondence
(CAC) is computed for any two associated areas as
the area of intersection, divided by the area of
union. In the figure, CAC is the computed as the
green area divided by the sum of all colored
(pink, purple, and green) areas. CAC was computed
for all catchments of each subbasin and
stream-formation threshold, and summarized in the
same manner as percent correct values.
Thiessen-derived catchment (red outline)
overlaying associated elevation-derived catchment
(gray outline) with correct area in green, and
areas of commission error in purple and omission
error in pink .
ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
20
Results
  • Number of catchments computed ranged from 957 to
    24,603, with catchment density increasing with
    decreasing stream-formation threshold.
  • Processing (Pentium 4 CPU, 3.0 GHz, 1 GB RAM)
  • Speed
  • ED catchments 10 minutes to 2 hours.
  • TPD catchments 2 to 14 minutes (5 to 10 times
    faster)
  • Reliability
  • ED process failed for some of the more
    dense network computations.
  • TPD process never failed.
  • Applicability
  • ED catchments requires an integrated DEM
    and ED network.
  • TPD catchments can be applied to any network.

ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
21
Results
  • Between subbasin comparisons
  • Averages of mean percent correct values range
    from about 50 to 65, with averages better than 60
    on hilly and mountainous subbasins.
  • Average total percent correct values
  • range from about 58 to 75.
  • greater than average mean percent correct for all
    subbasins.
  • Average mean omission errors are about 7 percent
    larger that average mean commission errors.

ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
22
Results
Distribution of percent correct values for all
catchments from the 100-cell stream-formation
threshold for the mountainous humid subbasin
(WV). Mode of distribution is 71.
Distribution of percent correct values compared
to catchment size for the 100-cell
stream-formation threshold in the mountainous
humid subbasin (WV).
ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
23
Results
Distribution of percent correct values compared
to network segment length for the 100-cell
stream-formation threshold in the mountainous
humid subbasin (WV).
Distribution of percent correct values compared
to catchment size for the 100-cell
stream-formation threshold in the mountainous
humid subbasin (WV).
ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
24
Results
  • Between subbasin comparisons
  • Average mean coefficient of areal correspondence
    (CAC) ranges from 0.34 to 0.51, with better
    correspondence in the hilly and mountainous
    subbasins.
  • CAC Co / (CoOmCm) and Co Om 1
  • Flat subbasins 0.5 / (1 0.93(0.5)) 0.34
  • Hilly and Mountainous subbasins 0.67 / (1
    0.93(0.33)) 0.51

ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
25
ResultsWithin subbasin comparisons
Dry Mountainous (CO) Hilly (TX) Flat (UT)
Humid Mountainous (WV) Hilly (MO) Flat (FL,GA)
ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
26
Results
Within subbasin comparisons Catchments were
separated into headwater (light green) and
non-headwater catchments (light blue) based on
whether or not they contained a dangling node
(cyan) of a stream line. Mean percentages were
recomputed for headwater and non-headwater
catchments.
ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
27
Results
Headwater average, average, and non-headwater
average of the mean coefficient of areal
correspondence (CAC) for each formation threshold
is shown for each subbasin.
ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
28
Results Pruning Tests
  • 200-cell stream-formation threshold network
    (green, pink, blue) for hilly dry subbasin (TX).
  • Pruned networks using UDA values based on ED
    catchments
  • Pruned to 1100,000-scale (pink, blue)
  • Pruned to 1500,000-scale (blue)
  • 200-cell stream-formation threshold network
    (green, purple, cyan) for hilly dry subbasin
    (TX).
  • Pruned networks using UDA values based on TPD
    catchments
  • Pruned to 1100,000-scale (purple, cyan)
  • Pruned to 1500,000-scale (cyan)

ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
29
Results Pruning Tests
ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
30
Results Pruning Tests
ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
31
  • Summary
  • Results suggest that the TPD catchment process
    is
  • Less likely to fail because of hardware or
    software limitations than ED process,
  • About 5 to 10 times faster than the ED catchment
    process,
  • Logistically much simpler to implement than ED
    process which requires a network integrated to an
    elevation model.
  • And that the
  • Fractional part that TPD catchments overlay ED
    catchments is about ½ for subbasins in flat
    terrain and about 2/3 for subbasins in hilly or
    mountainous terrain.
  • Headwater TPD catchments exhibit better areal
    correspondence (up to 17 percent) with ED
    catchments than do non-headwater catchments.
  • The lowest areal correspondence of TPD catchments
    to ED catchments occurs on relatively small
    catchments or on very short network segments.
  • Better than 80 percent linear correspondence can
    be expected between networks pruned to
    1100,000-scale or smaller using UDA based on TPD
    catchments and UDA based on ED catchments.

ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
32
Questions?Assessment of a Rapid Approach for
Estimating Catchment Areas for Surface Drainage
Lines
  • Lawrence Stanislawski, Science Applications
    International Corporation (SAIC)
  • Michael Finn, U.S. Geological Survey
  • E. Lynn Usery, U.S. Geological Survey
  • Mark Barnes, U.S. Geological Survey

ACSM-IPLSA-MSPS 2007, March 9-12, St. Louis, MO
About PowerShow.com