STATEMENT OF PROBLEM
Land within the City of Columbia, MO is being developed, changing
from agricultural to urban land uses. Urbanization of agricultural
land increases storm water discharges and may cause additional flooding
along the creeks fed by the urbanizing areas. New development must
be kept out of areas subject to flooding as regulated by ordinance.
The ability to accurately calculate peak discharges for particular
rainfall events directly impacts the ability to control development
and limit flooding. Because the City of Columbia has been growing
since the development of the last comprehensive storm water management
plan, existing assessments of land use and land cover maps may not
represent the current conditions.
PRESENT SITUATION
The City of Columbia has a preliminary storm water management
plan prepared by a consulting engineering firm. The report was developed
to analyze storm water management options and techniques and includes
calculations of peak discharges for the 100-year storms, for watersheds
within the city limits. The assessment of land use used in the management
plan, developed from aerial photography and zoning, is represented
in large aggregate blocks. The flood elevations (and the areas flooded)
associated with peak discharges for the 100-year flood are used
by city personnel in decisions on whether to allow particular development
to take place.
NEW APPROACH AND POTENTIAL
BENEFITS
Land cover classifications developed from satellite imagery allow
for an up-to-date assessment of land cover. The level of detail
available from the satellite imagery also allows for greater differentiation
of land cover within an area as to whether specific portions are
covered by woods, grass, open water, bare soil, or impervious surfaces.
Greater accuracy in the representation of land cover should allow
for the calculation of a truer representation of runoff curve number,
and thus a truer representation of peak discharges. Assessments
of peak discharges can be used as a part of the decision making
process to evaluate what property will be inundated by the 100-year
flood and to preclude development in those areas.
TECHNICAL APPROACH TAKEN
A portion of the Bear Creek watershed in Columbia, MO was used
as the test case for incorporating land cover classifications from
satellite imagery into hydrologic modeling. The preliminary calculations
of peak discharges from the consultant were used for comparison
of peak discharges calculated using land cover developed from satellite
imagery. Demonstration of this capability was undertaken for a Bear
Creek tributary where all the required information was available
(missing pieces of information limited the calculations possible):
the entire area contributing to the tributary is in the city limits
and peak discharges were calculated by the consultant, complete
land use information was available for the consultant’s preliminary
report, soil groups delineated for the entire area are available
on MSDIS, and IKONOS imagery available for entire area contributing
to the tributary.
In order to assess the impact of the different methods of determining
land use/land cover, the consultant calculations were rerun changing
only the runoff curve numbers. The Corps of Engineers’ HEC-1
model runs prepared by the consultants were rerun in HEC-HMS (successor
program to HEC-1) in order to import all of the modeling characteristics
from the consultant’s evaluation. This comparison was necessary
in order to ensure that any variations in peak discharges calculated
were only a result of changing the runoff curve number and were
not a result of changing other model input parameters. This means
that all basin delineations and all other basin characteristics
used in the modeling are those established by the consultant rather
than being developed separately from currently available satellite
imagery.
The watershed delineation used is that from digitizing the watershed
basins used in the consultant report. Land cover classes were developed
using IKONOS multispectral images with 4 meter spatial resolution
available for the project area. Land use for the Bear Creek area
was determined by supervised classification using a maximum likelihood
classifier. The hydrologic soil groups used to develop runoff curve
numbers were taken from the MSDIS on-line database. Runoff curve
numbers were calculated from land use classifications and soil groups
using modified script files from CRWR-PrePro3a.
Specifically, the year 1998 runoff curve numbers and peak discharges
developed by the consultant are compared with the year 2000 runoff
curve numbers and peak discharges developed by ICREST from satellite
imagery.
PRODUCTS
(1) Maps showing the land use classifications of the portions
of the Bear Creek watershed contributing to the tributary for
existing 1998 conditions (used by the consultant) and showing
the land cover classifications using the 2000 satellite imagery.
(2) Table displaying the runoff curve numbers from the consultant’s
preliminary report and calculated using satellite imagery (by
subbasin).
(3) Table displaying the peak discharges from the consultant’s
preliminary report and calculated using satellite imagery (by
subbasin).
(4) Table displaying the peak discharges at the tributary analysis
points from the consultant’s preliminary report and calculated
using satellite imagery (by subbasin).
INTEGRATION OF INFORMATION
INTO THE HYDROLOGIC MODELING ACTIVITIES
The hydrologic modeling effort utilizes information developed
by other ICREST team members, specifically the land cover classification
for area contributing to the Bear Creek tributary.
LIMITATIONS ON THE USE
OF THE INFORMATION
A comparison of the runoff curve numbers and peak discharges between
the two methodologies shows considerable variation. Before the information
would be used in decision making by local government, it is anticipated
that end users would need additional information about why the changes
are so great, whether the differences are due merely to differences
in surface coverages or from the methodology of calculating and
aggregating curve numbers. The methodological implications can be
represented in part by the fact that the consultant’s calculations
were based on land use, i.e., agricultural activities, and various
residential, commercial, and industrial developments, the curve
numbers for which incorporate pervious and impervious surfaces,
while the satellite imagery provides an assessment of land cover,
i.e., various crops, open water, bare soil, or impervious surfaces.
ANTICIPATED IMPLEMENTATION
FOR LOCAL GOVERNMENT
The maps and tables comparing runoff curve numbers and peak discharges
can be used in presentations and discussions with the end users.
End users need to indicate what supporting information is necessary
before information based on satellite imagery can replace information
developed from other means. This is particularly important where
the change in information sources could possibly result in changes
in different decisions being made depending on what information
source is used. When sufficient supporting information is available,
satellite imagery based calculations can be used in development
decision making.
ADDITIONAL WORK
(1) Find new locations for analysis where FEMA rating curves
(discharge versus elevation curves) are available in order to
show the differences in property expected to be inundated between
conventional calculations and satellite imagery supported calculations.
(2) Investigate the basis for the differences in the runoff curve
numbers and in the peak discharges.
PROJECT PARTICIPANTS
Dr. R. Lee Peyton, Dr. Kathleen M. Trauth, Mr. Harold E. Johnson,
Dr. Aderbal C. Corrêa, and Mr. Janggam Adhityawarma of the
Center for Environmental Technology (CENTECH), Department of Civil
and Environmental Engineering, carried out the project for ICREST.
|
