Methods for Onsite Infiltration Controls Ruben Kertesz and James P. Heaney Dept. of Environmental Engineering Sciences University of Florida May 2006

1
Methods for Onsite Infiltration Controls
Ruben Kertesz and James P. HeaneyDept. of
Environmental Engineering Sciences University of
FloridaMay 2006
2
Acknowledgements

• Research during the past five years on LID-type
  systems has been funded by
• US EPA Edison to develop spreadsheet methods to
  simulate and optimize BMPs.
• National Cooperative Highway Research Program
  -Evaluation of methods for highway BMPs including
  LID
• Water Environment Research Foundation-Principles
  of BMP evaluations including LID
• City of Gainesville-LID options for Tumblin Creek
• Corps of Engineers, Jacksonville-Water quality
  impacts of EAASR/STA 3/4 as part of Everglades
  Restoration

3
Collaborators

• Wayne Huber, Oregon State U.
• Eric Strecker and Marcus Quigley, Geosyntec, Inc.
• Neil Weinstein, LID Center
• John Sansalone, Betty Rushton, and Joong Lee, U.
  of Florida
• Brett Cunningham, Jones, Edmunds, Assoc.
• Alice Rankeillor and Stewart Pearson, City of
  Gainesville Stormwater Utility

4
Outline

• Evolution of LID
• Previous LID Studies
• Case Studies in Gainesville
• SWMM Modeling
• Results
• Summary and Conclusions

5
Evolution of LID

• Dissatisfaction with long-term maintenance
  problems associated with centralized ponds in
  Prince Georges County, MD
• Switched focus to decentralized, on-site,
  infiltration controls that are integrated with
  landscaping techniques

6
Conventional vs. LID
Centralized vs. Distributed
7
Total Annual Runoff

• About 80 of the total annual runoff comes from
  small storms that occur frequently.
• These smaller storms tend to have the largest
  impacts on urban water quality.
• Control strategy - focus on the smaller storms
  for protecting water quality.

8
Previous Studies

• Lee and Heaney (2003) used detailed GIS data to
  estimate the rainfall-outflow relationship for
• Kings Creek Apartments in SE Florida
• Calibrate against flow measurements for existing
  development
• Wonderland Creek in Boulder, CO
• Calibrate against detailed SWMM run
• Ref. Lee, J., and J.P. Heaney. 2003. Urban
  imperviousness and its impacts on stormwater
  systems. Jour. Of Water Resources Planning and
  Management, Vol. 129, No. 5, p. 419-426.

9
Event Based Rainfall DepthMiami, 1948-01
10
Kings Creek Apartment CatchmentSouth Florida
(Lee Heaney 2003)
11
Results for Kings Creek

• Location Miami
• Area 5.95 ha
• DCIA 44.1

12
Conclusions for Kings Creek

• With accurate GIS coverage for Kings Creek
• SWMM (old) produced very accurate outflow
  hydrographs that compared well with measured
  flows.
• Possible to develop accurate cause-effect
  linkages regarding the sources of outflows

13
Wonderland CreekBoulder, CO

• Apply variable degrees of GIS data to estimate
  the value of better spatial data

Wonderland Creek Study Site
Area 5.81 ha (14.36 ac) Single-family
residential
14
Wonderland CreekBoulder, CO

• Apply variable degrees of GIS data to estimate
  the value of better spatial data

15
(No Transcript)
16
Detailed sub-catchment discretization in one block
17
Results for Wonderland Creek

• A 1-year return period 1-hour design storm is
  modeled with 5 different levels of detail in DCIA.

Runoff modeling by SWMM (Lee 2003)
18
Conclusions for Wonderland Creek

• Use of inaccurate aggregate spatial data
• Led to estimates of peak flows that were almost
  three times too large
• Relatively expensive to develop detailed spatial
  coverages
• For existing developments

Runoff modeling by SWMM (Lee 2003)
19
Current Watershed Studies in Gainesville

• Tumblin Creek
• Lake Alice

20
Tumblin Creek Redevelopment
21
University Corners across from campus(future
development)
Design View from the South
Design View from the East
140,000 sq ft of shopping and dining, over 300
condominiums, penthouses and hotel rooms,
swimming pools on third and fifth floors, 1,200
parking spaces underground. In contrast to the
namely 1 and 2 story single use buildings
currently present.
22
Tumblin Creek Redevelopment
23
Heritage Oaks

• 20 units/acre, 0.89 acres

On-site Control Method
Reason

• Exfiltration system
• Pervious paving
• Disconnected roofs
• Small retention area
• Low / no-mortar brick
• Porous parking
• Curbless
• Walkway layout

• New buildings, N. parking
• Trees
• Couldnt send across site
• Infiltrate roof runoff
• Infiltrate water on S. side
• Onsite infiltration
• Aesthetics and infiltration
• Part aesthetics

24
On-site Controls at Heritage Oaks
25
Conclusions from University Heights

• Input from multiple disciplines
• Engineers, developers, landscape architects
• Multi-purpose design not exclusively driven by
  stormwater concerns
• Community Redevelopment Association
• Master Plan, Funding Base, Enforcement
• Little or no credit given for on-site controls
• How effective are the numerous on-site BMPs?

26
Lake Alice Watershed-UF

• Lake Alice highly impacted by development
• Stormwater concerns
• Range from flooding to water quality
• Long-term experimental watershed
• Land use diversity
• From ultra-urban to agricultural experimental
  areas
• Good topographic variability
• Local teaching and research applications
• Scale is an important issue

27
Differences of Scale
Lake Alice Watershed
East Crk Watershed
28
Differences of Scale

• 7.5 acres
• 392 subcatch
• .00004 ? 0.5 ac.
• Functional L.U.
• 10 aggregate
• 392 disaggreg.
• Headwater catchment

• 200 acres
• 7 subcatchments
• 7.5 ? 72 acres
• Functional L.U.
• 7 aggregate
• 1000 disaggreg.
• 1 aggreg. model
• 54 land uses
• 398 variables

• 1,000 acres
• 37 subcatch
• 8 ? 72 acres
• Functional L.U.
• 7 aggregate
• 1000 disaggreg.
• Non-uniform

29
Modeling Strategy for Estimating Effectiveness of
LID BMPs

• Disaggregate the catchment into basic functional
  units using GIS
• Use SWMM 5.0 to estimate the runoff for selected
  precipitation conditions
• Use this runoff estimate as correct answer
• Aggregate this result as needed for various
  analyses

30
EPA SWMM

• OLD
• Block to Block
• Downstream Storage / Treatment
• Implies centralized control

• NEW
• Treatment can occur before or in hydraulic system
• Both centralized and onsite control

31
SWMM 5.0 Capabilities

• Quality quantity
• Flood control
• BMP evaluation
• Control strategies
• Change parcel and transport parameters to
  simulate LID controls that can occur anywhere

• Time varying rain
• Storage, saturation
• Groundwater
• Buildup, washoff
• BMP influence

32
Study Area Attributes

• LID options are evaluated by changing attributes
• Infiltration parameters, Mannings n, depression
  storage, etc.

33
UF GIS to SWMM EPA SWMM 5.0
34
Functional Units on UF Study Site
35
Study Site Partitioned into 8 Owners Estimated
On-site Control
36
UF Study Site
37
Results

• Small difference in onsite control
• Area weighted parameter values from detailed GIS
  system provide solid foundation

38
Study Site Partitioned into 8 Owners Estimated
On-site Control
39
Study Site Partitioned into 8 Owners Estimated
On-site Control

• Sharing the responsibility

40
Summary and Conclusions

• Need efficient ways to estimate impact of
  proposed stormwater management systems
• Simple, highly aggregated calculations can lead
  to wasteful designs
• LID is more complex to analyze due to larger
  number of decentralized controls which rely on
  infiltration
• GIS and CAD datasets are widely available and
  GIS-CAD software is easy to use
• New Developments

41
Summary and Conclusions

• Performance of existing systems can be measured
  by direct monitoring and modeling
• Performance of proposed systems can only be
  evaluated by modeling

42
Ongoing Research

• Develop a hierarchy of methods to evaluate
  individual BMPs based on best available secondary
  data sources and experimental data
• Incorporate BMP models into SWMM for doing
  improved performance evaluations ranging from
• Black box input output models, to
• Advanced process models using computational fluid
  dynamics

43
Questions?