EPA/600/R-13/004 | March 2013 | www.epa.gov/gateway/science
United States
Environmental Protection
Agency
Stormwater Management for TMDLs in an
Arid Climate: A Case Study Application of
SUSTAIN in Albuquerque, New Mexico
Office of Research and Development
-------
EPA/600-R-13/004
January 2013
Stormwater Management for TMDLs in an Arid Climate: A Case Study
Application of SUSTAIN in Albuquerque, New Mexico
by
Leslie Shoemaker, Ph.D.
John Riverson, Jr.
Khalid Alvi, P.E.
Jenny X. Zhen, Ph.D., P.E.
Ryan Murphy
Brandon Wood
Tetra Tech, Inc.
10306 Eaton Place, Suite 340
Fairfax, VA 22030
In support of
EPA Contract No. GS-10F-0268K
Project Officer
Ariamalar Selvakumar, Ph.D., P.E.
Urban Watershed Management Branch
Water Supply and Water Resources Division
2890 Woodbridge Avenue (MS-104)
Edison, NJ 08837
National Risk Management Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
January 2013
-------
Disclaimer
The work reported in this document was funded by the U.S. Environmental Protection Agency (EPA)
under Task Order 1108 of Contract No. GS-10F-0268K to Tetra Tech, Inc. Through its Office of
Research and Development, EPA funded and managed, or partially funded and collaborated in, the
research described herein. This document has been subjected to the Agency's peer and administrative
reviews and has been approved for publication. Any opinions expressed in this report are those of the
authors and do not necessarily reflect the views of the Agency; therefore, no official endorsement should
be inferred. Any mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
-------
Abstract
This case study for the Albuquerque, New Mexico area was conducted under contract with the U.S.
Environmental Protection Agency (EPA) Office of Research and Development using the System for
Urban Stormwater Treatment and Integration Analysis (SUSTAIN). The effort focuses on investigating
both site- and regional-scale stormwater management questions ahead of a pending watershed-based
municipal separate storm sewer system (MS4) permit. The SUSTAIN modeling system integrates
watershed modeling capabilities, best management practice (BMP) process simulation, and BMP cost
representation within the context of a cost-benefit optimization framework (USEPA, 2009). The system
can be used to evaluate complex decisions about green infrastructure selection and placement,
performance, and costs for meeting flow or water quality targets or both.
With the large degree of variability in physical, hydrological, and chemical characteristic in a watershed,
watershed-scale management for the nonpoint source component of MS4 permits quickly grows in
complexity. Because different land uses had different pollutant levels, and because there were differences
in BMP cost assumptions, there were differences in cost-effectiveness between management strategies
that yielded comparable water quality outcomes. The objective of watershed management within an
optimization framework like SUSTAIN is to identify strategies that meet water quality goals while
minimizing cost.
Tetra Tech has developed two other SUSTAIN case studies for EPA in Kansas City, Missouri and
Louisville, Kentucky, as part of this development phase (USEPA, 2012). These studies investigated the
use of green or gray infrastructure practices to mitigate combined sewer overflows in temperate climate
areas. In contrast, this Albuquerque case study focuses on water quality performance of different
management practices for various storm sizes in an arid climate. The proposed approach documents the
various phases of the SUSTAIN application process, including establishing baseline hydrology and water
quality loads, identifying the critical condition for management, formulating management objectives on
the basis of local design standards, and testing the sensitivity of optimization results to different
formulations of the management objectives. The study estimates the potential range of benefits and
impacts in light of existing Escherichia coli (E. coll) total maximum daily load (TMDL) targets. The
results of this study provide quantitative technical guidance to support the pending watershed-based MS4
permit.
-------
Executive Summary
This case study for the Albuquerque, New Mexico, area was conducted under contract with the U.S.
Environmental Protection Agency's (EPA) Office of Research and Development using the System for
Urban Stormwater Treatment and Integration Analysis (SUSTAIN). SUSTAIN extends the capabilities
and functionality of traditionally available models by providing integrated analysis of water quantity,
quality, and cost factors. SUSTAIN offers a unique evaluation platform for cost-benefit optimization to
evaluate complex decisions about Best Management Practice (BMP) selection, placement, and
performance for meeting specified flow or water quality targets or both. Two previous SUSTAIN case
studies were published for applications in Kansas City, Missouri and Louisville, Kentucky, as part of this
development phase (USEPA, 2012). Those studies investigated the use of green or gray infrastructure
practices to mitigate combined sewer overflows in temperate climate areas. In contrast, this Albuquerque
case study focuses on the water quality performance of both structural and nonstructural BMPs for
managing runoff in an arid climate.
The focus area is in the Albuquerque metropolitan area in north-central New Mexico and lies along a
portion of the Middle Rio Grande watershed that was listed for Escherichia coli impairment as part of a
recent total maximum daily load (TMDL). The goal of this effort was to identify cost-effective
Stormwater management strategies through optimization that reduced E. coli loading by 66 percent, based
on target requirements established by the Middle Rio Grande E. coli TMDL. Given that rainfall events
are few and far between in this arid environment (annual average of about 9.5 inches), it was not
surprising to find that continuous simulation models were neither widely used nor readily available. An
alternative approach was applied to develop the baseline rainfall/runoff response for this case study. This
case study had three primary objectives:
1. Develop a boundary condition using a simple hydrology method (TR-55 Method) in conjunction
with event-mean concentrations
2. Evaluate the trade-offs in cost and management performance of both structural and nonstructural
BMPs during the optimization process
3. Estimate the water quality performance benefits of proposed management practices, based on
critical conditions identified in the Middle Rio Grande E. coli TMDL, to support a pilot
municipal separate storm sewer system (MS4) watershed-based permit being developed by EPA
Region 6.
Because different land uses had different pollutant levels, and because there were differences in BMP cost
assumptions, there were differences in cost-effectiveness between management strategies that yielded
comparable water quality outcomes. The structural BMPs considered included rainwater collection,
xeriscaping, and extended detention ponds; the nonstructural BMPs included pet waste management and
street sweeping. Due to the wide range between various literature-based and local implementation cost
estimates for street sweeping, two scenarios were run to test model sensitivity at lower and higher cost
estimates. Consequently, when street sweeping was weighed against other distributed and centralized
management practices, it was found that the cost assumption affected the order in which the practices
were selected. Given the range of conditions and constraints reflected in this modeling study, it was
estimated that a cost-effective solution for achieving a 66 percent E. coli load reduction target would cost
IV
-------
between $8,736 and $12,955 per 100 acres per year (assuming a 20 year BMP lifecycle). At both ends of
the modeled spectrum, rainwater conservation practices (collection and xeriscaping) are shown to be cost
effective. When street sweeping costs are lower, its use reduces the overall need for structural practices
by controlling loading at the source. The results also suggested that more infrequent sweeping (i.e.,
monthly) is not cost-effective compared to what can be achieved with rainwater conservation practices.
When street sweeping costs are higher, centralized wet ponds become an attractive supplement to
rainwater conservation practices. For the higher-cost scenario, street sweeping was not part of the
recommended suite of practices needed to meet the 66 percent E. coll TMDL reduction target. In fact, it
was only considered beyond the 75 percent load reduction point, and only after all other opportunities for
structural practices had been exhausted.
This case study documents the various steps taken in developing this SUSTAIN application, including
establishing baseline hydrology and water quality representation, identifying the critical condition for
management, formulating management objectives, and synthesizing model outputs with a focus on the E.
coll TMDL and pending watershed-based permit. Through this study a framework was established that
can be applied to support the development of implementation strategies for meeting water quality targets
associated with both the E. coll TMDL and pending watershed-based permit.
-------
Abbreviations
ABCWUA
ASCII
BMP
CFS
cfu/day
CN
d6
DPM
E. coll
EMCs
EPA
ESRI
FDCs
FIPS
FY
GIS
HARN
HRU
HUC 12
in.
LID
LSPC
MFR
mg/L
mL
MS4
NAD
NCDC
NM0234
NM-60
NMED-SWQB
NOI
NRC
NRCS
O&M
PEVT
SFR
SSURGO
SUSTAIN
SWMM
TMDL
TR-20/55
TR-55
TSS
USDA
USGS
WBP
Albuquerque Bernalillo County Water Utility Authority
American Standard Code for Information Interchange
Best management practice
Cubic feet per second
Colony forming units per day
Curve Number
Distribution of mean precipitation
Development Process Manual
Escherichia coll
Event mean concentrations
Environmental Protection Agency
Environmental Systems Research Institute
Flow duration curves
Federal Information Processing Standards
Fiscal Year
Geographic information systems
High Accuracy Reference Network
Hydrologic response units
12 digit hydrologic unit code
Inches
Low Impact Development
Loading simulation program in C++
Multi-family residential
milligrams per liter
milliliter
Municipal separate storm sewer system
North American Datum
National Climatic Data Center
Climate station at Albuquerque International Airport
New Mexico Type IIA - 60 - 24 hour storm
New Mexico Environment Department - Surface Water Quality Bureau
Notice of Intent
National Research Council
Natural Resources Conservation Service
Operation and maintenance
Potential evapotranspiration
Single family residential
Soil Survey Geographic Database
System for Urban Stormwater Treatment and Integration Analysis
Stormwater management model
Total maximum daily load
Technical Release 20/55
Technical Release 55
Total suspended solids
United States Department of Agriculture
United States Geological Survey
Watershed based permit
VI
-------
WIN TR-5 5 Windows version of TR-5 5
WQ Water quality
WQCV Water quality capture volume
VII
-------
Contents
Disclaimer ii
Abstract iii
Executive Summary iv
Abbreviations vi
Chapter 1. Background 1
1.1. Overview of Objectives 2
1.2. Watershed Based MS4 Permit 3
Chapter 2. Data Review 5
2.1. Spatial Data Sets 6
2.2. Regional Weather Patterns 12
2.3. Water Quality Monitoring Data 14
2.4. Linkage to TMDL Framework 17
Chapter 3. Establishing a Modeled Baseline Condition 21
3.1. Baseline Hydrology 21
3.2. Baseline Water Quality 25
3.3. Baseline Validation: Adobe Acres 28
Chapter 4. Proposed Management Activities 31
4.1. Structural BMP Performance and Cost 32
4.2. Nonstructural BMP Performance and Cost 35
Chapter 5. SUSTAIN Application for Representative Area 40
5.1. BMP Scenarios 42
Chapter 6. Evaluation of BMP Performance 54
Chapter?. BMP Optimization 62
7.1. Optimization Target 62
7.2. Optimization Scenarios and Results 62
Chapter 8. Summary and Conclusions 73
8.1. Baseline Representation 73
8.2. Evaluation of Structural and Nonstructural BMPs 74
8.3. Applicability to Watershed Based MS4 Permit 76
Chapter 9. References 77
Appendix: HRU Runoff Hydrographs 79
VIM
-------
Figure 1. Map of Bernalillo County, Albuquerque, and the Rio Grande watershed 2
Figure 2. Albuquerque local land use map 7
Figure 3. Land use map of the Adobe Acres drainage basin 8
Figure 4. Two-foot topographic contour layer for Albuquerque 11
Figure 5. Annual precipitation totals (2000-2010) for selected NCDC precipitation gages 12
Figure 6. Average monthly precipitation distribution (2000-2010) for selected NCDC
precipitation gages 13
Figure 7. Distribution of event storm size at the Albuquerque International Airport (2000-
2009) 14
Figure 8. Albuquerque TSS and E. coll sample counts at each water quality monitoring gage 15
Figure 9. Histogram of monitored against the long-term historical precipitation record 17
Figure 10. Plots of daily precipitation vs. daily streamflow after baseflow separation for USGS
08330000 Rio Grande at Albuquerque, New Mexico (1974-2009) 19
Figure 11. Summary of precipitation statistics by FDC zone 20
Figure 12. NM-60 rainfall distribution used in the WinTR-55 model 23
Figure 13. Unit-area (1 acre) runoff hydrographs representing the parking lot HRU 24
Figure 14. Statistical summary of Bernalillo County E. coll monitoring data 26
Figure 15. Comparison of model E. coll EMCs against monitoring and literature values 28
Figure 16. Adobe Acres SUSTAIN baseline hydrology and water quality validation model 29
Figure 17. Plots of modeled E. coll and TSS concentrations versus observed data for Adobe
Acres 30
Figure 18. Unit cost per gallon of collection system storage 34
Figure 19. Roads in and outside the Albuquerque jurisdiction 38
Figure 20. Fecal coliform bacteria contribution distribution among sources 39
Figure 21. Map comparing the spatial extent of jurisdictional boundaries used for developing
conceptual HRU distributions 40
Figure 22. Comparison of HRU distributions summarized using the MS4 and Albuquerque's
boundaries 41
Figure 23. Structural BMP routing network configuration 43
Figure 24. Commercial retail land use (Category 2) example plot with impervious surfaces
highlighted 45
Figure 25. Commercial service land use (Category 3) example plot with impervious surfaces
highlighted 46
Figure 26. Industrial/manufacturing land use (Category 5) example plot with impervious
surfaces highlighted 47
Figure 27. Multifamily land use (category 6) example plot with impervious surfaces
highlighted 48
Figure 28. Public/institutional land use (category 9) example plot with impervious surfaces
highlighted 49
Figure 29. Single family land use (category 10) example plot with impervious surfaces
highlighted 50
Figure 30. Distribution of mean precipitation, d6, in inches for the United States 52
Figure 31. Comparison of nonstructural BMP scenarios and their effect on TSS load reductions
for various storm event sizes 54
Figure 32. Comparison of nonstructural BMP scenarios and their effect on E. coll load
reductions for various storm event sizes 55
Figure 33. Comparison of nonstructural BMP scenario costs and their reduction of TSS loads
for various storm event sizes 56
IX
-------
Figure 34. Comparison of nonstructural BMP scenario costs and their reduction of E. coll loads
for various storm event sizes 56
Figure 35. Commercial land use—rainwater collection system and xeriscape performance and
cost-effectiveness curve 57
Figure 36. Industrial land use—rainwater collection system and xeriscape performance and
cost-effectiveness curve 58
Figure 37. Multi-family residential land use—rainwater collection system and xeriscape
performance and cost-effectiveness curve 58
Figure 38. Public/Institutional land use—rainwater collection system and xeriscape
performance and cost-effectiveness curve 59
Figure 39. Parking lot land use—rainwater collection system and xeriscape performance and
cost-effectiveness curve 59
Figure 40. Single-family residential land use—rainwater collection system and xeriscape
performance and cost-effectiveness curve 60
Figure 41. Detention basin pollutant removal efficiency (100-acre composite drainage area) 61
Figure 42. Cost-effectiveness curves of optimization scenarios without street sweeping 64
Figure 43. Literature-based cost optimization with monthly street sweeping (5% load removal) 65
Figure 44. Local-cost based optimization with monthly street sweeping (5% load removal) 65
Figure 45. Literature-based cost optimization with biweekly street sweeping (35% load
removal) 66
Figure 46. Local cost based optimization with biweekly street sweeping (35% load removal) 66
Figure 47. Literature-based cost optimization with weekly street sweeping (70% load removal) 67
Figure 48. Literature-based cost optimization with weekly street sweeping (70% load removal) 67
Figure 49. Composite optimal cost-effectiveness curve and selected solution meeting target 68
Figure 50. Literature-based street sweeping cost data: BMP cost composition of the optimal
solutions 69
Figure 51. Local street sweeping cost data: BMP cost composition of the optimal solutions 69
Figure 52. Unit-area (1 acre) runoff hydrographs representing the Parking Lot 80
Figure 53. Unit-area (1 acre) runoff hydrographs representing the Drainage/Flood Control 80
Figure 54. Unit-area (1 acre) runoff hydrographs representing the Road 81
Figure 55. Unit-area (1 acre) runoff hydrographs representing the Single Family Residential 81
Figure 56. Unit-area (1 acre) runoff hydrographs representing the Multi-Family Residential 82
Figure 57. Unit-area (1 acre) runoff hydrographs representing the Industrial 82
Figure 58. Unit-area (1 acre) runoff hydrographs representing the Commercial 83
Figure 59. Unit-area (1 acre) runoff hydrographs representing the Public-Institutional 83
Figure 60. Unit-area (1 acre) runoff hydrographs representing the Transportation 84
Figure 61. Unit-area (1 acre) runoff hydrographs representing the Vacant 84
Figure 62. Unit-area (1 acre) runoff hydrographs representing the Parks 85
Figure 63. Unit-area (1 acre) runoff hydrographs representing the Agricultural 85
-------
Table 1. Mapping of Albuquerque land use categories to SUSTAIN case study HRUs 10
Table 2. Inventory of available TSS monitoring data 15
Table 3. Inventory of available E. coll monitoring data 16
Table 4. Summary of zones identified along the FDC for the TMDL 18
Table 5. Curve number assumptions by land use 23
Table 6. EMC values for E. coll by HRU and storm depth 27
Table 7. EMC values for TSS by HRU and storm depth 27
Table 8. Summary of modeled flow and pollutant concentrations for Adobe Acres 30
Table 9. BMP simulation parameters used in SUSTAIN for structural BMPs 32
Table 10. Summary of rainwater harvesting rebates in Albuquerque 33
Table 11. Summary of nonstructural BMPs 35
Table 12. Summary of literature values for street sweeping costs 36
Table 13. Albuquerque FY2012 street sweeping cost 36
Table 14. Street sweeping costs in California 36
Table 15. Street sweeping removal rates and cost estimates by sweeping frequency 37
Table 16. HRU distribution for the 100-acre SUSTAIN conceptual model 42
Table 17. Summary of nonstructural BMP scenarios 42
Table 18. Estimated potential area that can be converted to xeriscape 51
Table 19. Drain time coefficients forWQCV calculation 51
Table 20. Summary of percent imperviousness in Albuquerque by land use category 51
Table 21. Maximum distributed BMP performances and composition by land use categories 60
Table 22. E. coll Load targets by storm scenario depth for single sample and geometric mean 62
Table 23. Optimization run scenarios 63
Table 24. Optimal literature-based costs solution achieving 66% E. coll load reduction (0.78-
inch storm) 71
Table 25. Optimal local costs solution achieving 66%E. coll load reduction (0.78-inch storm) 71
Table 26. Distributed BMP composition of the optimal solution achieving 66 percent E. coll
load reduction forthe 0.78-inch storm 71
Table 27. BMP composition of the optimal local cost solution achieving 66 percent E. coll load
reduction forthe 0.78 inch storm 72
Table 28. Summary of rainfall events at the Albuquerque International Airport (2000-2009) 86
XI
-------
Chapter 1. Background
The case study began by first engaging a group of local stakeholders, in conjunction with EPA Region 6,
to help (1) define the context of the study including location, (2) identify the pollutant(s) of interest, (3)
set reasonable management objectives for evaluation, and (4) provide relevant models, data sets, and
existing studies as supporting information for the case study.
The study area focused along a portion of the Middle Rio Grande watershed listed for E. coll impairment
as part of a recent TMDL in the Albuquerque metropolitan area in north-central New Mexico. The Rio
Grande watershed begins with headwaters in Colorado. By the time the river reaches Albuquerque, New
Mexico, the drainage area encompasses approximately 14,000 square miles; however, the scope of this
case study is limited to stormwater contributions originating from the Albuquerque metropolitan area
rather than flow volumes and pollutant loads accumulated from upstream sources. Figure 1 presents a
map of the general study area highlighting its spatial relation to the City of Albuquerque, Bernalillo
County, and the larger Rio Grande watershed.
-------
A
Legend
A USGS Streamflow
^ Rio Grande Watershed
| | BernalilloCounty
^ City of Albuquerque
HUC 12 Boundary
Albuquerque, New Mexico
Hydrologic Unit Codes (HUC12)
TETHATECH
Figure 1. Map of Bernalillo County, Albuquerque, and the Rio Grande watershed.
Because the case study scope is limited to local stormwater contributions, it was essential to establish a
framework that links surface water flow, water quality, and regulatory measures (i.e., TMDL, watershed-
based permitting [WBP]) to watershed drivers such as precipitation, pollutant generation, and stormwater
management strategies. The SUSTAIN platform provide a means for making this linkage and assessing
cost-benefit relationships of various stormwater management strategies. The final exercise involved
extrapolating SUSTAIN site-scale performance to the watershed scale to quantify potential regional
impacts of widespread adoption of both structural and nonstructural management practices.
1.1. Overview of Objectives
A number of analytical objectives distinguish this study from previous SUSTAIN case studies. First, it
represents management in an arid climate environment, which affects watershed behavior and changes the
types of stormwater management approaches that work well. Collaboration with a stakeholder group was
initiated early in the process to help set objectives and modeling direction. During discussion with this
group a number of unique study features were either proposed or discovered through data collection.
These items were substantial enough to develop into study objectives which are not only important to this
study, but also are highly relevant and transferable to SUSTAIN applications in general.
-------
Second, the other two case studies in Kansas City and Louisville focused on water quantity (flow volume
or peak flow reduction) (USEPA, 2012); however, the focus of this case study is water quality. Unlike
the other two case studies, no existing continuous simulation model was available to link with SUSTAIN.
Given that rainfall events are few and far between in this arid environment, it was not surprising that no
continuous simulation models had been developed. An alternative approach was explored for developing
the baseline rainfall/runoff response. Finally, stakeholders were interested in evaluating structural BMPs
and the tradeoffs between a mix of structural and nonstructural practices. On the basis of these findings,
the following three modeling objectives were defined:
• Develop a boundary condition using a simple hydrology method (such as TR-20/55 or Rational
Method), in conjunction with event-mean concentrations;
• Evaluate the trade-offs in cost and management performance of both structural and nonstructural
(programmatic) BMPs during the optimization process; and
• Estimate water quality performance of proposed management practices based on critical
conditions identified in the Middle Rio Grande E. coll TMDL to support a pilot MS4 watershed-
based permit under development by EPA Region 6.
1.2. Watershed Based MS4 Permit
As documented by the National Research Council (NRC) in a report titled Urban Stormwater
Management in the United States, stormwater discharges can negatively affect water quality through
increases in stormwater volume and, consequently, pollutant loads to the receiving waters (NRC 2008).
The NRC report concludes that adopting the MS4 permitting approach using the physical topography of
watersheds is a more effective means to halt and reverse damage to waterbodies, rather than basing the
permits on political boundaries. With this, the NRC recommends that EPA use a watershed based permit
(WBP) approach to improve the stormwater program.
A pilot program was suggested as a first step to allow EPA to explore the many complexities of WBP.
EPA selected Region 6 and EPA Region 6 subsequently selected the Middle Rio Grande valley as one of
three pilot WBP projects nationwide. This area was chosen largely because of existing water quality
impairment in the Rio Grande and the opportunity to work on the challenges of permitting unique to arid
and semi-arid parts of the country.
The proposed watershed-based MS4 permit in the Middle Rio Grande is designed to accommodate a
general permit approach using a Notice of Intent (NOI) to the general permit in lieu of an individual
permit application. The operator of a regulated MS4 must include in its permit application, or NOI,
chosen BMPs and measurable goals for each minimum control measure. The NOI can include schedules
to fully develop and implement the stormwater program over the initial 5-year permit term. To help
identify the most appropriate BMPs, EPA Region 6 is working with the Middle Rio Grande workgroup to
develop a menu of BMPs to serve as guidance. The permit will encourage the use of green infrastructure,
such as swales, rain gardens, and porous pavement.
According to EPA Region 6, Appendix A of the proposed permit will identify a number of
implementation options for regulated MS4 operators. This includes sharing responsibility for program
development with a nearby regulated small MS4, taking advantage of existing local or state programs, or
participating in implementing an existing Phase I MS4's stormwater program as a committee. These
options are intended to promote a regional approach to stormwater management coordinated on a
watershed basis. Operators of regulated MS4s in the Middle Rio Grande are required to design their
programs to do the following:
-------
• Reduce the discharge of pollutants to the maximum extent practicable;
• Protect water quality; and
• Support the applicable water quality goals of the Clean Water Act.
This case study presents a methodology for applying SUSTAIN to support and inform certain aspects of
permit implementation, which include refining the suite of proposed BMPs. In addition to providing
different degrees of management practices on the basis of local design standards and practices, the model
can highlight the merits of the various BMP options during the planning process. The conclusions from
this case study can serve as documented guidance for those MS4 operators tasked with implementing
management practices in fulfillment of the WBP. Recommendations and conclusions will be made that
synthesize these case study findings related to the (1) cost-effectiveness of specific BMPs, (2) critical
conditions under which BMPs are and are not effective, and (3) tradeoffs between structural and
nonstructural BMPs. Although a representative subwatershed unit is evaluated for BMP implementation
using SUSTAIN, the case study findings are not intended to be globally representative of all localized
hydrologic, water quality, or land use context within the MS4 permit boundaries. This report presents a
methodology for applying SUSTAIN to study management opportunities in a data-challenged
environment.
-------
Chapter 2. Data Review
The quality of a SUSTAIN model application, as is true of models in general, depends on the quality of the
input data sets used during model development. Therefore, acquisition, review, and synthesis of a robust
data set is critical for developing a sound, defensible model baseline representation of local hydrologic
and water quality conditions. The initial data collection phase of this study involved an iterative process
of collaboration and data request in conjunction with the local stakeholder group. The following items
were requested:
• Geospatial data sets describing physical characteristics of the study area (land use, topography,
soils, subwatershed/catchment boundaries);
• Local flow, water quality, and climate monitoring data sets;
• Documents with local BMP guidance; and
• Local BMP cost data.
Each provided data set was reviewed using a multi-step process, beginning with compiling all provided
data in a centralized location. After this compilation, each data set was reviewed and assessed for specific
utility to this case study application with a focus on the locations and pollutant constituents identified
during stakeholder discussions. This section presents key data sets provided by Bernalillo County, the
city of Albuquerque, EPA Region 6, and other stakeholder groups which have been identified through
data review as critical to the baseline model development. The following items were specifically
evaluated:
• Geospatial data sets describing the physical characteristics of the study area (land use and
topography) provided by Bernalillo County and the city of Albuquerque;
• Local precipitation and climate data obtained from NCDC sources; and
• Water quality monitoring data provided by Bernalillo County.
Spatial data describing land use, slopes, and soils were evaluated in conjunction with watershed
boundaries to construct the geospatial portion of the SUSTAIN model representation of the watershed.
Precipitation data were used to evaluate representative storm volume, intensity, and duration for
accurately reflecting baseline rainfall-runoff conditions. Water quality monitoring data were used to
inform characterization of an appropriate pollutant loading response for selected storm intervals.
Two pollutants, E. coll and total suspended solids (TSS), were identified as the focus of this study during
conference calls with the stakeholders. E. coll was selected as the primary pollutant of concern because
of its relevance to an existing TMDL. TSS was also evaluated as a secondary pollutant; however,
because no numeric target exists, it was not used as an objective for optimization. Monitoring data sets
provided by Bernalillo County focus on four drainage basins ranging from 20 to 1,000 acres with mixed
land use distributions. The four basins were (1) Adobe Acres, (2) Sanchez Farms, (3) Alameda
Boulevard, and (4) Paseo Del Norte. These sites were considered as candidate locations for calibration
efforts to establish a baseline watershed characterization, rainfall-runoff response, and water quality
signature.
-------
2.1. Spatial Data Sets
Representative characterization of physical aspects of the study area is important when modeling the local
rainfall-runoff response in a watershed and developing hydrologic response units (HRUs) for organizing
areas of similar response in the context of a model. During SUSTAIN model setup, these spatial data sets
were used as the basis for representing HRUs. HRUs are discrete spatial units that embody unique
physical or environmental characteristics and human influences that result in a unique hydrologic or water
quality signature. HRUs are defined by physical characteristics such as slope or soil infiltration capacity,
both of which affect the timing and magnitude of runoff. Anthropogenic elements such as impervious
cover and land use often affect not just the hydrology but also pollutant loading.
Nationally scoped data sets are readily available as land use raster, digital elevation model, and soil
surveys that provide a starting point for model development; however, local data sets maintained by local
agencies often provide a higher resolution. Higher resolution, local data are more desirable than the
coarse national data sets because they will usually provide more accurate, site specific representation of
watershed characteristics. Bernalillo County provided drainage area boundaries for the four drainage
areas with coincident water quality monitoring data. The following spatial data sets from Albuquerque's
online geographic information systems (GIS) database were collected:
• County wide Land Use;
• Land Use Categories Mapping Table; and
• Two-Foot Elevation Contours.
Two land use data sets were available: (1) National Land Cover Dataset 2006 hosted by the Multi-
Resolution Land Characteristics Consortium, and (2) local Albuquerque land use. The Albuquerque local
land use data set is divided into 13 categories. This locally sourced data set is more appropriate than the
national data set because it is (1) updated continually and made public twice a month, and (2) more
reflective of local nuances in the land use categorization. Figure 2 presents a map of Albuquerque's land
use data set rendered using the predefined land use categories, and Figure 3 presents a detailed, site-
specific view of the same data set zoomed in to the Adobe Acres drainage basin.
-------
City of Albuquerque
Land Use
Agriculture
| Commercial
| Drainage-Flood Control
Industrie
Multi-Family Residentia
Parking Lot
Parks
Public-Institutional
Single Family Residential
Transportation
Vacant
Albuquerque, New Mexico
City of Albuquerque Land Use
Source: Albuquerque CIS 2012
Figure 2. Albuquerque local land use map.
-------
, Drainage Area
Land Use
Agriculture
Commercial
Drainage-Flood Control
Industrial; Wholesale-Warehousing
Multi-Family Residential
Parks & Open Space
Public-Institutional
Single Family Residential
Vacant
Albuquerque, New Mexico
Adobe Acres Land Use (218 ac.)
0 0.05 0,1 0,2
Kilometers
0.1 0.2
NAD_1983_StatePlane_Nei»_Me«ico_C»nlral_FIPS_30(]2_reol
Map producQd 01-06-2012
Source: Albuquerque GIS 2012
Figure 3. Land use map of the Adobe Acres drainage basin.
-------
The land use distribution shown in Figure 3 for the Adobe Acres drainage basin includes 11 of the 13 land
use categories included in the complete city of Albuquerque land use data set (Albuquerque GIS, 2012).
Albuquerque land use data set provided a clear framework for representing the local HRU texture as it
relates to this study and was preferred over the national Multi-Resolution Land Consortium coverage.
The simplifying assumptions imposed while deriving the final HRU layer used in SUSTAIN were as
follows:
• The existing commercial categories that distinguished between service and retail have been
grouped into a single Commercial category.
• The warehousing category was reclassified as Industrial.
• Local roads were not represented explicitly as polygons in this data set; therefore, they rendered
as void space between polygons. When the data set was converted to a raster for use in
SUSTAIN, void spaces were assigned a unique value to represent the roads.
Topographic data can be used to derive slope, which affects the timing and magnitude of the rainfall-
runoff relationship. Also, where water movement is determined by topography, the data are useful for
establishing flow direction and delineating drainage areas. This information is generally important when
developing detailed baseline hydrology models. Albuquerque provides a spatial data set of 2-foot
topographic contours, which increases in resolution to 1-foot contours for some areas. Figure 4 presents a
map of percent slope derived from 2-foot contours provided by the city (Albuquerque GIS, 2012). The
map also locates the four monitored drainage basins with available water quality data referenced for
guiding water quality representation in this study. Those four sites are all located close to the Middle Rio
Grande in areas with generally lower slopes and lower elevation by over 1,000 feet than seen elsewhere in
the city and county.
Soils spatial data sets provide a basis for assessing infiltration potential (1) from pervious land types when
developing a baseline rainfall-runoff model, and (2) through infiltration-based BMPs. No soils data set
was initially available through the Albuquerque or Bernalillo County websites. National soil survey data
can be used to develop a general assessment of infiltration potential. The national data set, Soil Survey
Geographic Database (SSURGO) is maintained by the U.S. Department of Agriculture (USDA) Natural
Resources Conservation Service (NRCS). This data set was not available in digital format for all areas.
Because this case study focuses on urban areas and New Mexico restricts stormwater harvesting and
infiltration-based practices, a generalized soil survey data set is of lower importance than layers
representing other physical characteristics. In addition, this case study focused primarily on urban water
quality management for single events (instead of continuous simulation, which would involve inter-event
dewatering). For those reasons, only the Albuquerque land use coverage was used as the basis of HRU
representation. Applying the previously listed transformations and assumptions to the Albuquerque land
use data set resulted in 12 unique HRU categories on which the SUSTAIN model will be based (Table 1).
Conservative assumptions will be applied when developing runoff time series to address likely soil
conditions.
-------
Table 1. Mapping of Albuquerque land use categories to SUSTAIN case study HRUs
Albuquerque
land use categories
Agriculture
Commercial Retail
Commercial Services
Drainage/Flood Control
Industrial/Manufacturing
Wholesale/Warehousing
Multi-Family Residential
Parking Lots/Structures
Parks/Recreation
Public/Institutional
Single-Family Residential
Transpiration/Utilities
Vacant/Other
No Representation
SUSTAIN
case study HRU
Agriculture
Commercial
Drainage Flood Control
Industrial
Multi-Family Residential
Parking Lot
Parks
Public/Institutional
Single-Family Residential
Transportation
Vacant
Roads
10
-------
City of Albuquerque
Monitoring Location
Almeda Blvd. ^
88 «r^ LI
Paseo del Norte .-
Adobe Acres t
*^ai < ^ * «
Albuquerque, New Mexico
Percent Slope
NAD_1983_Slat6Plane_New_Me»ico_Centfa!_FIPS_3002_feel
Map produced 08-16-2012
Figure 4. Two-foot topographic contour layer for Albuquerque.
11
-------
2.2. Regional Weather Patterns
SUSTAIN requires runoff time series which represent the unique rainfall-runoff response of each HRU.
These runoff time series serve as the primary input boundary condition. Understanding the regional
weather patterns is essential to accurately reflect expected volume, intensity, and duration of expected
storm events with high spatial variability in meteorology. As previously noted, Albuquerque is in the arid
desert climate of the southwest, which differs from the wetter temperate climates from previous case
study locations. On average, the area has 300 days of sunshine and about 9 inches of rainfall annually,
with a 90th percentile storm depth estimated near 0.41 inch (Penttila, 2011). In such a unique
environment, the typical practices for stormwater management (largely derived for more temperate
climates) are not always applicable.
To categorize the uniqueness of Albuquerque's climate in the context of this case study, three NCDC
daily precipitation gages were selected for further analysis. Selecting these stations was on the basis of
the length of record, record quality, and proximity to the urbanized areas in Albuquerque and Bernalillo
County.
• Albuquerque International Airport (NM290234)
• Albuquerque Valley (NM290231)
• Petroglyph National Monument (NM296754)
Analysis spanned in temporal resolutions from annual precipitation totals to the distribution of individual
storm depths. Figure 5 and Figure 6 present the annual precipitation totals and average monthly
precipitation distribution for these three NCDC stations.
20
18
16
14
-»-Albuquerque International Airport (NM0234)
-*-Albuquerque Valley (NM0231)
-•-Petroglyph National Monument (NM6754)
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Figure 5. Annual precipitation totals (2000-2010) for selected NCDC precipitation gages.
12
-------
3.00
2.50
2.00
1.50
1.00
0.50
0.00
-Albuquerque International Airport (NM0234)
Albuquerque Valley (NM0231)
-Petroglyph National Monument (NM6754)
Figure 6. Average monthly precipitation distribution (2000-2010) for selected NCDC precipitation
gages.
Figure 5 shows that the range of annual precipitation at the three NCDC gages from 2000 to 2010 varied
between 6 and 14 inches, demonstrating the arid climate of this case study. The annualized monthly-
aggregated plot in Figure 6 (for the same period of 2000-2010) shows that about 50 percent of annual
precipitation volume occurs between July and October. Because of its proximity to the study area and
relatively high data quality, rainfall data from the Albuquerque International Airport were selected for
further analysis. Data for calendar years 2000-2009 were coincident with water quality monitoring data
provided by Bernalillo County.
Using observed data from the Albuquerque International Airport from January 1, 2000, through
December 31, 2009, individual precipitation events were categorized. A 72-hour antecedent period and a
minimum of 0.1 inch were used to classify the hourly rainfall time series into discrete storm events. The
resulting precipitations event distribution from that period is presented in Figure 7. Of the precipitation
events summarized, over 80 percent were less than 0.75 inch, and almost 88 percent were less than 1 inch.
Knowing this distribution of event magnitudes allows the focus of this study to be further refined. A
summary of events has been included as Table 28 in the Appendix.
13
-------
<0.25 0.25 0.50 0.75 1.00 1.50
-0.49 -0.74 -0.99 -1.49 -1.99
Storm Depth (in.)
>=2.0
Figure 7. Distribution of event storm size at the Albuquerque International Airport (2000-2009).
Understanding the frequency and magnitude of precipitation events will help to reveal critical conditions
when (1) formulating numeric objectives, (2) analyzing the water quality monitoring data, and (3)
establishing a linking framework to flow duration curves (FDCs) developed for the Middle Rio Grande E.
coli TMDL (NMED-SWQB, 2010). Coincident climate observations are also available from the
Albuquerque International Airport for other parameters, including temperature, which can be used to
estimate potential evapotranspiration (PEVT) rates for representing PEVT from the surface of some
structural BMPs like extended detention basins as needed.
2.3. Water Quality Monitoring Data
Local water quality monitoring data are useful when characterizing the baseline pollutant loading time
series representing each HRU category. These pollutant loading time series, in conjunction with the
runoff time series discussed previously, act as the input boundary conditions for SUSTAIN. Additional
observed water quality data help to validate assumptions used in generating these model inputs, ensuring
that they are representative of the local conditions.
The U.S. Geological Survey (USGS) urban stormwater monitoring program was identified as a robust
source of water quality data. From 2004 through 2009, Bernalillo County, in conjunction with the USGS,
performed water quality monitoring of urban stormwater runoff at four sites discharging to the Middle
Rio Grande mainstem. Bernalillo County provided a digital copy of this data set. Water quality
monitoring data were analyzed to distinguish specific spatial and temporal patterns of TSS and E. coli
water quality metrics. This data set was also evaluated in parallel with precipitation data to identify
temporal patterns and ensure that an adequate range of event depths were sampled.
Figure 8 presents a map of raw sample counts from the Bernalillo County. Flow or runoff depth were not
included as part of this monitoring effort; however, pumps are used at each of the four sites to convey
14
-------
collected stormwater to the Rio Grande. A log of pump hours was made available and pump rating
curves are known, which allow for reasonable estimation of storm volumes (Glass 2012).
Legend
WQ Sample Count
TSS & E. Coli
• <30
• 31-45
>45
Albuquerque Water Quality Sampling
Figure 8. Albuquerque TSS and E. coli sample counts at each water quality monitoring gage.
The four unique sites represent a mix of land uses and drainage area sizes intended to reflect a near-
complete distribution of the expected water quality response to the climate patterns in Albuquerque.
After an initial review of the water quality monitoring data set and discussions with Bernalillo County, it
was found that some samples were quality control flagged or designated as dry weather samples. The
flagged values were removed from consideration. Water quality samples were tagged with only an
analysis time but no sampling time. Table 2 presents summary statistics for the water quality monitoring
data set for TSS, and Table 3 presents a similar summary for E. coli.
Table 2. Inventory of available TSS monitoring data
Description
Alameda Blvd.
Paseo del Norte
Adobe Acres
Sanchez Farms
Count
24
23
24
14
Start date
2/23/04
2/23/04
3/4/04
4/16/05
End date
12/16/10
8/23/09
12/16/10
12/16/10
Min.
(mg/L)
70
100
80
18
Max.
(mg/L)
1,920
3,216
2,096
296
Median
(mg/L)
478
524
604
134
Mean
(mg/L)
526
703
670
149
mg/L = milligrams per liter
15
-------
Table 3. Inventory of available E. coli monitoring data
Description
Alameda Blvd.
Paseo del Norte
Adobe Acres
Sanchez Farms
Count
22
22
18
12
Start date
2/23/04
2/23/04
3/4/04
4/16/05
End date
12/16/10
8/23/09
12/16/10
12/16/10
Min.
(#/100 ml)
45
220
286
2,420
Max.
(#/100 ml)
241,960
241,960
51,720
77,010
Median
(#/100 ml)
2,420
1,497
2,566
5,669
Mean
(#/100 ml)
53,126
12,664
7,085
13,006
ml = milliliter
Those four monitoring sites are at lower elevation and lower sloped sites along the Middle Rio Grande.
Some factors that potentially influence sample values include natural behaviors like higher seasonal
groundwater levels and waterfowl population migrations in addition to anthropogenic influences. When
considering the use of these samples for benchmarking a set of representative event mean concentrations
(EMCs), it will be important to also compare against regional literature values to validate the applicability
of the monitoring data.
Corresponding flow values for these samples were not reported, so nearby rainfall records were reviewed
to associate samples taken with coincident storm size. Both TSS and E. coli water quality samples were
mapped to associated rainfall events using observed rainfall data from the Albuquerque International
Airport (NM0234). As noted previously, a 72-hour antecedent period and a minimum of 0.1 inch were
used to classify the hourly rainfall time series into discrete storm events (as presented in Figure 7). The
water quality samples were then associated with the storm that spanned the dates for which they were
taken. Because the water quality monitoring events were not tagged with a sampling time, a 12-hour
buffer was applied to either side of the sample date when associating with precipitation events. This
analysis provides a context for evaluating the distribution of the monitoring data across the spectrum of
observed storm events. Figure 9 presents a distribution of TSS and E. coli sample counts categorized by
associated precipitation depth.
16
-------
<0.25 0.25 0.50 0.75 1.00 >= 1.50
-0.49 -0.74 -0.99 -1.49
Associated Storm Depth (in.)
Figure 9. Histogram of monitored against the long-term historical precipitation record.
Figure 9 suggests that the Bernalillo County monitoring data set has samples taken across the full range of
storm conditions; however, considerably more E. coll samples were taken for storms less than 0.75 inch.
Nevertheless, several samples were collected for some of the largest storms. Using data that are well
distributed across the full range of critical events gives confidence that the baseline runoff and water
quality time series will represent the range of expected conditions in the watershed. Because sampling
coincides with precipitation events that characterize a spectrum from less than 0.25 inch to more than 2
inches and 23 E. coll and 22 TSS samples were taken with wet intervals of more than 2 inches, this data
set should be adequate for developing a robust model baseline.
2.4. Linkage to TMDL Framework
One objective of this case study is to create a linkage between management strategies evaluated with
SUSTAIN and the recently adopted E. coll TMDL. This linkage will provide a context for interpreting
case study findings in the broader, regional TMDL framework. Because the TMDL was developed using
FDCs, the SUSTAIN model baseline must be structured in a way that provides a linkage or basis of
comparison with the TMDL baseline. The TMDL was developed on the basis of 2005 monitoring data
collected by the New Mexico Surface Water Quality Bureau. FDCs were developed for four assessment
units, and five zones were identified along each FDC. These zones, corresponding to flow percentile
ranges, help to classify the FDC on the basis of similar hydrologic conditions. Table 4 summarizes the
five FDC zones used in developing the TMDL (NMED-SWQB, 2010).
17
-------
Table 4. Summary of zones identified along the FDC for the TMDL
FDC zone
Low flows
Dry conditions
Mid-range flows
Moist conditions
High flows
Flow exceedance
percentile3
90 to 100
60 to 89
40 to 59
10 to 39
Oto9
a TMDL lists percentile ranges as High Flows (0-10), Moist conditions (10-40), Mid-range flows (40-60), Dry conditions (60-90),
and Low flows (90-100). Percentile breaks used in Table 1 were selected to avoid data point overlapping.
For consistency with the established TMDL, the approach used for this study was to identify storms of
varying magnitudes that are representative of critical zones along the FDC. Typically, TMDL
impairments in the right-most zones indicate the influence of point sources; impairments in the left-most
zones indicate pollutant contribution from nonpoint sources. This analysis focuses on the three highest
zones representing the critical condition for nonpoint source impairment, which can be managed by
stormwater BMPs. Representative storm depths will be selected for each zone, and hydrographs for these
three storm depths will be generated for each HRU using a simplified modeling approach.
Identifying representative storm depths for the mid-range flow, moist condition, and high-flow zones
began with reconstructing an FDC used in developing the TMDL with daily streamflow data for the
USGS 08330000 Rio Grande at Albuquerque, New Mexico, gage. Streamflow records from January 1,
1974, through December 31, 2009, were used to ensure consistency with the TMDL. Because this
streamflow gage collected data for a drainage area of more than 14,000 square miles, storm-influenced
flows were parsed out of the data set using a sliding interval hydrograph separation technique (USGS,
1996). Daily values were matched with coincident daily precipitation at the Albuquerque International
Airport (240234) from January 1, 1974, through December 31, 2009. The FDC is plotted with
corresponding precipitation events in Figure 10.
18
-------
10,000
1,000
.1
1
1
1
— Streamflow (cfs)
• Precipitation (in)
_O
w—
re
e
'ra
Q
100
1
^-^ 1
• ' ^*^***fc^^^^
*
1.25
o
ra
.Moist
Conditions
1.75
1.5
20%
40% 60%
Percent of Days Flow Exceeded
80%
100%
Figure 10. Plots of daily precipitation vs. daily streamflow after baseflow separation for USGS
08330000 Rio Grande at Albuquerque, New Mexico (1974-2009).
Several studies have attempted to quantify a BMP treatment depth for the Albuquerque area using a
rainfall percentile analysis. One study presented by Albuquerque analyzed precipitation data from 1891
through 2009 and identified the 90th percentile storm depth as 0.44 inch (Penttila, 2011). A 24-hour
rainfall water quality control volume for BMP design based on a 0.41-inch depth was identified in a
recent Bernalillo County BMP study (Radian, 2011).
The objective of this analysis is to identify representative rainfall depths corresponding to each of the
three critical-condition FDC zones for further evaluation with the management practices in SUSTAIN.
Precipitation summary statistics were calculated uniquely by FDC zone from Figure 10 to reveal a
correlation between the FDC and rainfall depth. Figure 11 presents the summary statistics for the mean,
75th, 85th, 90th, and 95th percentile precipitation events associated with the five FDC zones.
19
-------
Mean
75th Percentile
85th Percentile
90th Percentile
95th Percentile
90 to 100
Low Flows
Figure 11. Summary of precipitation statistics by FDC zone.
Because the approximate 90th percentile rainfall depth (or 10th percentile flow exceedance) has loosely
been identified for further evaluation as a potential threshold treatment depth for water quality purposes, it
is recommended that this case study evaluate storm depths that bracket the 90th percentile value. Figure
11 highlights three depths that bracket the 90th percentile storm of 0.44 inch, providing insight on the
range of expected BMP performance around this critical point. The 0.27-, 0.47-, and 0.78-inch storm
depths are associated with the mid-range flows, moist conditions, and high-flow zones along the FDC,
respectively. These depths brackets identify the 0.44-inch treatment depth of interest, and they
correspond to critical precipitation conditions (Figure 7) and coincident water quality monitoring data
(Figure 9). Developing runoff hydrographs for these three storm depths, which constitutes the SUSTAIN
runoff baseline, is discussed in the next section.
20
-------
Chapters. Establishing a Modeled Baseline Condition
In SUSTAIN stormwater runoff and pollutant washoff time series simulated using an external watershed
model are typically the forcing functions that drive BMP simulation. Watershed models use site-specific
spatial and temporal elements to characterize the rainfall runoff response, buildup-washoff of pollutants,
or other water quality loading processes. The watershed model time series represent the existing
condition (or baseline), which serves as the reference point from which water quality improvement will be
evaluated.
A critical first step of a SUSTAIN application is to establish or confirm a representative baseline condition
with a high degree of confidence in its applicability. Model baselines should be validated as vigorously
as possible against a local, observed data set. This becomes especially important in the context of cost-
benefit optimization of future management objectives, because the model baseline is foundational to
results interpretation and resulting conclusions. It is important for the model baseline condition to
appropriately represent variability throughout the study area. It needs to consider the influence of
physical features associated with both surface and subsurface behavior and human activities that could
affect pollutant load generation.
The following sections describe the process of developing baseline runoff and pollutant loading time
series by HRU representative of the Albuquerque study area. Without an existing continuous-simulation
watershed model, discussion is focused around an alternative method to runoff time series development
using a design storm approach. An analysis of local water quality monitoring data, in conjunction with
referenced literature, was used to inform the representation of pollutant loading by HRU. Fully
developed runoff and pollutant time series were then used in a validation model of the Adobe Acres
neighborhood to assess the distribution and magnitude of water quality responses against observed data.
This validation was limited to a statistical assessment comparing model outputs against the central
tendency of water quality monitoring data.
3.1. Baseline Hydrology
Hydrographs were developed for three storm depths as representative events along the FDC. These
hydrographs were used as the hydrology boundary condition inputs representing surface runoff for the
SUSTAIN baseline. SUSTAIN offers the flexibility to integrate with different external model output
formats. In previous case studies, continuous-simulation watershed models (such as LSPC) or rainfall-
runoff models (such as SWMM) have been the most widely integrated external data sets; however, the
runoff and pollutant loading boundary conditions applied in SUSTAIN are not limited to continuous-
simulation models. For example, the recent E. coli TMDL completed for the Middle Rio Grande that was
established using flow and load duration curves did not require development of any continuous-
simulation models.
Because rainfall in the area is relatively infrequent compared with other parts of the country, developing a
continuous-simulation watershed model is not necessary to adequately characterize and represent the
critical conditions for the SUSTAIN model baseline. An alternative approach involves developing
hydrographs using a design storm approach. Albuquerque Development Process Manual (DPM), volume
II, Design Criteria, Chapter 22, describes locally accepted methods that are used to generate design
storms and hydrographs. These methods are mostly based on the Rational Method. The DPM also
includes language describing accepted alternatives, including continuous simulation with SWMM and
design storm approaches using TR-20/55.
21
-------
The WinTR-55 computer model was used to generate runoff hydrographs for the three storm depths
corresponding to the mid-range flow, moist condition, and high-flow zones along the FDC. WinTR-55 is
a graphical version of the DOS-based TR-55, which is described as a small watershed hydrology
application (NRCS 2012). WinTR-55 is used to generate a storm hydrograph response given a user-
specified rainfall depth. Runoff can also be routed through channels or structures to an outlet location.
An application can be developed with up to 10 subbasins. The program uses the Curve Number (CN)
approach, which applies a unique CN by land use. The CN describes runoff generating potential of a site
and can range between 0 (low runoff potential) and 100 (high runoff potential). This is a lumped
parameter that considers physical characteristics such as soil infiltration capacity. Runoff is predicted in
the TR-55 model using the following equation:
(P-0.2S)2
(P + 0.8S)
where Q is the runoff depth in inches, P is the rainfall depth in inches, and S is the potential maximum
retention depth (in inches) after runoff begins (NRCS, 2012).
In the above equation, S is largely a factor of the ground cover, vegetation, and soil conditions that affect
infiltration capacity and initial abstraction. S is a function of CN via the following relationship in which
GVcan range between 0 and 100:
CN
A typical use of WinTR-55 is with user-specified, 24-hour rainfall totals for a specific return interval. A
variety of rainfall distributions are included in the program; the user can also prescribe the rainfall
distribution. This study applied the NM-60 24-hour rainfall distribution, a local variation of the Type-IIa
24-hour distribution, which is specified for use with the TR-20 program by Albuquerque's DPM, Chapter
22, Section E. 1(4), Programs for Alternative Procedure Acceptance. Figure 12 presents the NM-60
rainfall distribution.
22
-------
Incremental Rainfall Percentage —Cumulative Rainfall Percentage
25%
100%
20%
o
'•3
15%
10%
5%
0%
0%
20
22
24
Figure 12. NM-60 rainfall distribution used in the WinTR-55 model.
A suite of WinTR-55 models were developed representing 12 HRUs as derived from Albuquerque's land
use spatial data set (Albuquerque GIS, 2012). The objective was to generate a set of 1-minute
hydrographs by land use, representing each of the three storm depths identified in Figure 11 (i.e., 0.27,
0.47, and 0.78 inch). These hydrographs will be used as model inputs for SUSTAIN to represent runoff
from each model HRU. A number of simplifying assumptions were adopted while developing these
runoff time series:
1. A square drainage area of 10 acres
2. There is an estimated time of concentration of 0.132 hour
3. A conservative infiltration (assuming Hydrologic Soil Group C) was applied
Future modeling studies could adjust the time of concentration to account for the effects of slope or
unique drainage features. Case study HRUs were mapped to land use categories available in the WinTR-
55 model as presented in Table 5. It should be noted that multiple HRUs with similar hydrologic surface
characteristics can use the same curve number; however, they are represented by a separate hydrograph
because of variations in water quality responses.
Table 5. Curve number assumptions by land use
Case study HRU
Roads
Drainage Flood Control
Parking Lot
Commercial
Transportation
Win TR-55 land use
Paved Curb
Paved Curb
Paved Parking Lot
Commercial and Business
Paved Open Ditches
Curve number
98
98
98
94
92
23
-------
Case study HRU
Industrial
Vacant
Multi-Family Residential
Single-Family Residential
Public-Institutional
Agriculture
Parks
Win TR-55 land use
Industrial
Developed Urban
1/8-Acre Townhomes
1/4-Acre Lots
Open Space, Poor
Pasture, Grassland, Range (Fair)
Open Space, Fair
Curve number
91
91
90
87
86
79
79
The outputs from the WinTR-55 applications were post-processed to obtain the runoff results generated
from the land before routing through the reach consistent with the input needed for SUSTAIN.
Intermediate output hydrographs were reported on a 30-second timestep for the 10-acre WinTR-55 model.
Final hydrographs were aggregated by averaging to a 1-minute reporting timestep and downscaled to 1-
acre for use in SUSTAIN. Figure 13 presents hydrographs from the selected 0.27-, 0.47-, and 0.78-inch
storm depths for the parking lot HRU.
High Flows (P=0.78 in.)
—Moist Conditions (P=0.47 in.
—Mid-range Flows (P=0.27 in.)
0 100 200 300 400 500 600 700 800 900 1,000 1,100 1,200 1,300 1,400
Simulation Minute
Figure 13. Unit-area (1 acre) runoff hydrographs representing the parking lot HRU.
Figure 13 shows three distinct hydrographs with varying peaks and timing for the selected storm depths.
Although the parking lot FIRU generated runoff for all three storms, note that not all HRUs generate
runoff for all storms. As parameterized, some HRUs, such as Parks, do not generate any runoff—even for
the 0.78-inch storm. It is probable that a storm with a larger return interval, such as the 2- or 5-year
event, would produce runoff for all HRUs; however, those events fall within the highest 5th percentile of
rainfall events and are likely outside the focus of this case study evaluating stormwater quality
management with BMPs. A complete set of hydrograph plots for all 12 HRUs is presented as Appendix
A. Developing EMCs by HRU and storm depth for representing pollutant loading is discussed in the next
section.
24
-------
3.2. Baseline Water Quality
Pollutant loading time series were developed for use in parallel with the one-minute surface runoff
hydrographs to represent the complete SUSTAIN model baseline. Local TSS and E. coli monitoring data
provided by Bernalillo County were referenced to provide context for selecting a range of event-mean
concentrations associated with different HRUs. These data sets were first assessed to determine whether
the observed data are reflective of the range of critical conditions (0.27-, 0.47-, and 0.78-inch selected
storm depths). Because the EMC values are unique by HRU and storm size, the monitoring data must
capture at a minimum the magnitude of storms for which hydrographs were developed. Because
coincident runoff flow values were not reported with these water quality samples, nearby rainfall records
from the Albuquerque International Airport were used to map the TSS and E. coli samples to rainfall
events.
Under natural conditions, E. coli bacteria can be killed off with exposure to light or other environmental
factors; however, in certain urban settings, they can regrow between storms (a physical process that is not
explicitly represented in SUSTAIN). Because the focus of this analysis is limited to the immediate storm
events and their associated recession periods, it is reasonable to model the behavior of E. coli as that of a
relatively conservative material that is mobilized with runoff energy and decays with time.
Because this case study evaluates both E. coli and TSS, a literature search was performed to see if any
studies showing any linkages or interactive relationships between the two should be considered during
modeling. With respect to the relationship between bacteria and sediment loadings, Reeves et al. (2004)
examined fecal coliform loadings in the Talbert and Lower Santa Ana watersheds near Huntington Beach,
California. Between 1999 and 2003, several field studies were conducted to sample fecal coliform
(among other) bacteria in stormwater runoff throughout the watershed. The results of these studies
suggest that fecal coliform loadings are poorly correlated with turbidity. Sediments collected from the
storm drain infrastructure had high concentrations of fecal coliform bacteria; however, surface eroded
sediments had low concentrations. Overall, fecal coliform loading rates could be related to flow via a
power function:
L~Qn
where L is the loading rate [71"1], Q is the volumetric flow rate [Z3/7] and n ranges from 1 to 1.5.
Subsequent work by Ahn et al. (2005) suggests that bacteria and viruses in stormwater runoff are either
associated with particles smaller than 53 micrometers (\m\) in diameter or are not associated with
particles at all. This study assessed fecal coliform indicator bacteria concentrations in stormwater runoff
from the Santa Ana River and filtered sediment through a 53 \am sieve. Measuring total organic carbon
before and after this filtration exhibited virtually no change in total organic carbon.
Surbeck et al. (2006) support this finding by examining the load response of fecal coliform bacteria and
sediment to changes in flow, otherwise known as flow fingerprinting. Sampling sites on the Santa Ana
River mainstem (in arid Southern California) and several tributaries yielded data for several storm
hydrographs in 2003 and 2004. The findings suggest that storm flow initially increases fecal coliform
bacteria indicator concentrations and then remains more or less constant throughout the storm
hydrograph. This differs from the flow/load relationship for sediments, which exhibits a power-law
relationship with the storm hydrograph. The conclusion of the literature search is that while fecal
coliform loads are related to storm flow, they do not appear to be strongly affected or influenced by
sediment loadings.
25
-------
Because the EMCs typically vary with storm depth, a statistical analysis was performed on local
monitoring data to assess correlation between storm depth and pollutant concentration. TSS and E. coll
concentration data were divided into bins by storm depth consistent with the monitoring summary
presented in Figure 9. Then the average concentration was calculated for each bin. Figure 14 presents
plots of this storm-associated concentration analysis forE1. coll and TSS. Since the Sanchez Farms
drainage basin is managed by a bio swale, water quality samples taken from that location were excluded
from the figures below.
8,000
E 7,000
8
I
<0.25
0.25
-0.49
0.50
-0.74
0.75
-0.99
1.00
- 1.49
>= 1.50
Associated Storm Depth (in.)
,£
c
o
1,000
soo
i= 600
I
400
200
0
<0.25
0.25
-0.49
0.50
-0.74
0.75
-0.99
1.00
-1.49
>= 1.50
Associated Storm Depth (in.)
Figure 14. Statistical summary of Bernalillo County E. coli monitoring data.
26
-------
Figure 14 highlights a relatively constant average E. coll concentration with increasing storm depth, with
the exception of the largest storms where concentrations increase three to four-fold. Conversely, Figure
15 shows little evidence of a positive correlation between increasing TSS concentration and increasing
storm depth on the basis of this monitoring data set; however, a basic understanding of the physical
processes involved with sediment detachment and transport dictates that increasing energy (from rainfall
and flow) will mobilize more sediment. Given the limited number of samples in the 0.75 in to 1.49 in.
range relative to the others, it is difficult to draw a statistically definitive conclusion through simple data
analysis about what is occurring; however, both E. coll and TSS concentrations appear to trend downward
to a point as storm size increases (dilution effect), but then sharply increase for the largest events.
Literature estimates of EMCs and export coefficients by land use were referenced to stratify the average
concentration for each storm depth on the basis of the relative potential for contribution by each HRU
(Lin, 2004). Table 6 and Table 7 present the baseline EMC values by HRU and storm depth for E. coll
and TSS, respectively.
Table 6. EMC values for E. coli by HRU and storm depth
HRU
Parking Lot
Drainage Control
Transportation
Public-Institutional
Parks
Industrial
Vacant
Commercial
Single Family Residential
Multi-Family Residential
Agricultural
E. coli concentration
(#/100 ml)
0.27-in. storm
188
150
188
56
63
56
94
150
75
113
75
0.47-in. storm
750
600
750
225
250
225
375
600
300
450
300
0.78-in. storm
1,500
1,200
1,500
450
500
450
750
1,200
600
900
600
Table 7. EMC values for TSS by HRU and storm depth
Land use
Parking Lot
Drainage Control
Transportation
Public-Institutional
Parks
Industrial
Vacant
Commercial
Single Family Residential
Multi-Family Residential
Agricultural
TSS concentration
(mg/L)
0.27-in. storm
320
120
320
80
80
280
240
280
160
280
256
0.47-in. storm
700
262
700
175
175
613
525
613
350
613
560
0.78-in. storm
1,042
469
1,042
313
313
911
938
911
625
911
834
EMC values for E. coli from Table 6 were compared against the median observation from the Adobe
Acres drainage basin and a sample of literature values reported as part of the National Stormwater Quality
Database vl.l (Pitt, 2004). Figure 15 presents model EMC values by HRU against these literature and
27
-------
monitoring benchmarks as an additional check to ensure they fall within a reasonable range of values
observed in the field. These EMC values will be used in conjunction with the previous presented storm
hydrographs to validate the baseline representation.
: Case Study EMC
10,000
1 1
1 Pitt, R. 2004. The National Stormwater Quality Database v1.1 \
— Single Family Residential Mixed Residential — Highway
a 1,000
B
§
3
a
uJ
100
Figure 15. Comparison of model E. coli EMCs against monitoring and literature values.
The EMCs used to drive this SUSTAIN model application generally fall within the range of values
reported in the National Stormwater Quality Database; however, they are quite lower than the median
observed value at the Adobe Acres site. Discussions with the stakeholder group indicate that proximity to
the river and seasonal groundwater conditions may make the Adobe Acres drainage basin prone to a
denser population of waterfowl, which could influence the background bacteria concentrations at the site.
3.3. Baseline Validation: Adobe Acres
Baseline hydrographs and EMCs were used in a site-scale SUSTAIN validation model of the Adobe Acres
drainage basin to assess the magnitude of the runoff and pollutant generation for each of the three storm
depths. A map of the Adobe Acres baseline drainage area and HRU distribution is presented below in
Figure 16. This model was configured using a single subwatershed. During simulation all flow and
pollutant loads were routed to an assessment point where runoff and water quality were reported. Table 8
presents a summary of modeled peak flow, average TSS concentration, and average E. coli concentration
for each of the three selected rainfall depths.
28
-------
Legend
Subwatershed j] Agricultura
Outlet Type
—»•— Total
• BasinRouting
BMPs
Commercial
Drainage
g Parking Lot
HI Parks
| | Public
SF Residential
Vacant
Transportation p
Adobe Acres
Baseline Model
NAD_1983_StatePlane_New_Mexico_FIPS_3003
Map produced 04-25-2012
Figure 16. Adobe Acres SUSTAIN baseline hydrology and water quality validation model.
29
-------
Table 8. Summary of modeled flow and pollutant concentrations for Adobe Acres
FDCzone
Mid-range Flows
Moist Conditions
High Flows
Rainfall
depth
(in.)
0.27
0.47
0.78
Peak
flow
(cfs)
3.4
15.7
54.2
Average E. coll
concentration
(#/100 ml)
150
641
1,203
Average TSS
concentration
(mg/L)
133
455
778
cfs = cubic feet per second; in. = inches; mg/L = milligrams per liter; ml = milliliters
Modeled E. coll and TSS concentrations presented in Table 8 were compared with observed data sampled
from the Adobe Acres site. Observed data associated with all storms less than 1 inch were summarized
and plotted against the modeled concentrations to assess the ability of the model baseline to reflect the
range of observed water quality data. Figure 17 presents modeled concentration points against boxplots
summarizing observed E. coll and TSS data. Modeled concentrations scatter about the observed median
and are generally reflective of the inter-quartile range (25th to 75th percentile value) of the monitoring
data.
10,000
o 1,000
«H
2?
o
Observed (25th to 75th Percentile)
X Modeled {0.27-in, Storm)
X Modeled (0.78-in. Storm)
— Observed (Median)
x Modeled (0.47-in. Storm)
c
0)
100
10
X
10,000
1,000
100
10
-------
Chapter 4. Proposed Management Activities
It is recognized that stormwater BMPs for the arid and semi-arid region are important tools for preserving
and improving the water quality of the region's precious water resources. Those BMPs should be
designed using the unique climate characteristics and to meet the special needs (water conservation, reuse,
and sustainable land use management) of the region.
Gautam et al. (2010), among others, assert that nonstructural BMPs should play an important role in arid
regions, where structural BMP options are often limited or expensive. Street sweeping was identified in
the Middle Rio Grande-Albuquerque Reach Watershed Restoration Action Strategy as a viable practice
for managing sediment and bacteria loading from urban land (MRGARWG, 2008). Pet waste
management is another nonstructural BMP that is practical and effective in reducing bacteria pollution.
Both Albuquerque and Bernalillo County have laws requiring pet waste to be picked up from any
property other than that of the owners'.
Several references suggest that the best strategy for effective BMPs for the arid Southwest should focus
on stormwater conservation and water reuse, for example rainwater harvesting, local groundwater
infiltration when feasible, and minimizing evapotranspiration losses. LaBadie (2010) makes specific
recommendations regarding the types of LID practices most favorable for the region's arid climate and
identifies those practices deemed least favorable. A summary of favorable and unfavorable LID practices
identified is presented below:
Favorable practices Unfavorable practices
• Harvesting parking lot runoff • Swales
• Extensive use of rain barrels • Flow-through structures
• Harvesting street runoff • Rain gardens
• Detention facilities to capture first flush flows • Green roofs
• Increased urban tree cover
In addition to the BMPs listed above, a water conservation measure, xeriscaping or xerogardening, has
gained popularity in arid and semi-arid regions in recent years. Xeriscaping is a practice that focuses on
reducing water demand by using drought-tolerant plants, efficient irrigation, soil improvement, and
mulching. When compared with traditional landscaping practices, xeriscaping requires lower soil
moisture and can have a higher water retention capability. Therefore, xeriscaping serves as an effective
measure for water conservation and reduces stormwater pollution during storm events by retaining more
water on-site, resulting in less runoff. The Albuquerque Bernalillo County Water Utility Authority
(ABCWUA) is implementing a rebate program to encourage homeowners to convert traditional landscape
into xeriscape (ABCWUA, 2012). In addition to direct rainfall, harvested runoff from rooftops and
impervious surfaces can be used to irrigate a xeriscaped area, which further reduces runoff pollution.
Of this and the other BMPs listed here, street sweeping and pet waste management were deemed suitable
for the Albuquerque area with respect to nonstructural BMP selection. Additionally, rainwater
harvesting, xeriscaping, and detention basins were deemed suitable and are considered in this SUSTAIN
application study.
SUSTAIN requires BMP cost data as a key component of the optimization process. Data sets can be input
either as individual components costs (i.e., length of pipe, volume of soil) or as an aggregate volumetric
cost (i.e., $/ft3). While national cost databases exist (such as those included in the SUSTAIN BMP cost
31
-------
database), up-to-date, local costs are always preferable to accurately reflect realistic materials and labor
costs (for both construction and maintenance).
For comparison purposes, all costs are represented as an annual cost where applicable. Present value
costs are assumed to be paid over the design lifetime in equal installments. These annualized costs are
calculated using the following equation:
A = P
i
where A is the annualized cost, P is the present value of that cost, / is the assumed interest rate (5 percent),
and n is the design life span in years.
4.1. Structural BMP Performance and Cost
Structural BMPs are physically represented in the SUSTAIN model with parameters that can be adjusted
during optimization. The parameters used in SUSTAIN to represent structural BMPs are summarized in
Table 9.
Table 9. BMP simulation parameters used in SUSTAIN for structural BMPs
Parameter
Xeriscaping
Rainwater collection
Detention basin
Initial condition
Initial soil moisture
Initial water depth
0.3
0
-
0
0.3
0
Substrate
Ponding depth (ft)
Substrate layer depth (ft)
Substrate layer porosity
Vegetative parameter, A
Background soil saturated infiltration rate*
(inch/hr), fc
ET rate (in/day)
0.042
2
0.4
1
0.3
0.104
--
--
--
--
--
--
4
1
0.3
1
0.3
0.104
Water quality
TSS 1st order decay rate (I/day), k
E . coli 1st order decay rate (I/day), k
0.8
0.5
0.8
0.5
0.8
0.5
* Soil map shows the majority background soil has hydrologic soil group of C; therefore, a 0.3 inch/hr background infiltration
rate is assumed.
The BMP module in SUSTAIN was configured to represent the water retention and infiltration capacity
provided by xeriscaping. It was recommended that the soil at the site be loosened and amended to a depth
of 24 inches to prompt plant root development and moisture retention. In SUSTAIN a 24-inch substrate
with 0.4 porosity was assumed with no underdrain. Additionally, a 0.5-inch ponding depth was used to
represent depression storage. Also a conservative saturated background infiltration rate of 0.3 inch/hour
(assuming Hydrologic Soil Group C) was applied.
The xeriscape rebate program conducted by ABCWUA pays property owners a credit of $1.00 for every
square foot of qualifying landscape, with a minimum of 500 square feet to participate. While other
32
-------
maintenance costs would be expected with xeriscaping, because the case study was formulated from a
municipal planning perspective, only the rebate cost was considered. Assuming a life span of 20 years
and an interest rate of 5 percent yields a cost of $0.0802 for every square foot of xeriscape per year.
Differing slightly from xeriscape representation, rainwater harvesting can be physically represented in
SUSTAIN in the form of an arbitrary number of rain barrels. Rain barrels in SUSTAIN can be defined in
volume, drainage area, drainage area soil and infiltration properties, and the like. Using a specified
volume and a 2 dry-day release schedule (collected rainwater released to receiving landscape after 2 dry
days), rainwater harvesting was modeled in SUSTAIN.
This involves two components: a rainwater collection and distribution system, and a landscape that
consumes the collected rainwater. Both components have an initial capital cost and operation and
maintenance (O&M) costs. It is important to note, however, that the water saved by rainwater harvesting
results in cost savings. Additionally, most rainwater harvesting practices are implemented by private
industry or individuals and are not a direct burden on MS4 permittees. Considering these factors, this
study uses the cost information derived from the rainwater harvesting rebate program that ABCWUA
implements. Rainwater collection system rebates are based on the amount of rain that can be stored on-
site, with rebate level summarized in Table 10 below.
Table 10. Summary of rainwater harvesting rebates in Albuquerque
Rebate amount
$25
$50
$75
$100
$125
$150
Minimum volume
(gallons/year)
50
150
300
500
1,000
> 1,499
Maximum volume
(gallons/year)
149
299
499
999
1,499
The cost numbers above were further converted into a unit volume cost in dollars per gallon, which are
plotted in Figure 18 below. Using these data, a power function was generated to define a relationship
between unit volume cost and storage volume, shown below:
= 2.9019F,
-0.472
where PRwn,umtis the present unit volume cost in dollars/gallon and Vb is the storage volume of the unit in
gallons. The equivalent annual cost of this assuming a 20-year lifespan and 5 percent interest rate yields:
ARWHJJmt = 0.2327F,
0.472
The rebate amount for rainwater harvesting, similar to xeriscape conventions, was $1.50 per square foot.
The minimum area was 500 square feet, while the maximum possible area was 2,000 square feet.
33
-------
$0.60
$0.00
y = 2.9019x-°-472
R2=0.,
500 1000 1500
Collection System Storage Volume (Gallon)
2GG3
2500
Figure 18. Unit cost per gallon of collection system storage.
Similar to rain barrels, detention basins can be physically represented in SUSTAIN. More specifically, a
detention basin can be represented using the SUSTAIN dry pond template with specified surface area,
ponding depth, and outlet orifice diameter sized to achieve 40-hour drain time.
The Bernalillo County BMP study, completed in November 201 1, provides estimates of construction
costs for four extended detention basin BMPs and detailed cost estimates for these facilities' annual
O&M. The construction cost of a typical extended detention basin includes a unit volume cost of $0.67
per cubic foot of storage and a fixed cost of $29,000 for inlet, outlet structure, pipe, and riprap. With an
additional 30 percent of the total base cost added as contingency cost, the overall construction cost can be
expressed as
pmf = 1.3 x ($0.67F + $29,000)
where PDB,c is the present construction cost of a detention basin, and Fis the volume of the detention
basin in cubic feet.
The annual O&M cost consists of two components: per facility associated costs and BMP size associated
costs. Per facility costs, not associated with the facility size, cover compliance inspection, inlet/outlet
cleaning, nuisance control, and outlet maintenance. BMP size associated costs include sediment removal
and vegetation/lawn care. Considering these two components, the annual O&M cost of each facility can
be expressed as
= $0.0503F + $1,458.81
where ADB!O&M is the annual O&M costs of a detention basin, and Fis the volume of the detention basin in
cubic feet.
34
-------
Annualizing construction costs at a 5 percent interest rate and assuming equal annual payments for the 35-
year lifespan of a detention basin yields the following annual construction associated cost equation:
YI 4- 'V
*
ADBfC =1.3 x($0.67F +$29,000) >
=> ADBfC ~ $0.0532F + $2,303
where Amc is the annualized construction cost of the detention basin, and Vis the volume of the
detention basin in cubic feet. Combining this annualized construction cost with the annual O&M costs
yields the following equation to calculate total annualized cost for a detention basin as a function of
detention basin volume:
A Total = ADB c + ADB 0&M x $0.103 5V + $3,762
where ADB:Total is the annualized total cost of a detention basin, and Vis the volume of the detention basin
in cubic feet.
4.2. Nonstructural BMP Performance and Cost
Nonstructural BMPs, i.e., street sweeping and pet waste management, are represented by modifying the
corresponding HRU time series to reflect pollutant removal efficiencies (Table 11).
Table 11. Summary of nonstructural BMPs
Nonstructural BMPs
Street sweeping
Pet waste management
HRUs affected
Roads
Multifamily, Single-family, Park and
Recreation, Roads
Pollutants
Sediment and Bacteria
Bacteria
Literature suggests that street sweeping can remove 5 to 70 percent of sediment depending on sweeping
method and frequency (Radian, 2011). No literature values were found with regard to bacteria removal
effectiveness of street sweeping. However, since bacteria naturally die off with sunlight exposure, their
quantities tend to increase asymptotically to some limit and diminish overtime with no additional inputs.
The studies reviewed previously (Reeves et al., 2004; Ahn et al., 2005; Surbeck et al., 2006) all suggest
that once bacteria are in the water, they tend not to be associated with particulate matter. However,
before it reaches the water body, it is commonly recognized that trash and other debris on the land surface
are places where bacteria are harbored. Street sweeping will remove the trash, sediment, and other debris
from the road surface, and consequently, prevent those bacteria from being mobilized into the stormwater
system. For this reason, the same percent removal rates for land-based sediment were also applied for
bacteria.
The cost of street sweeping is a function of sweeping method and frequency. Cost data were compiled
and compared from a number of sources. Some variation existed among the sources, as shown in Table
12. To better understand the impact of these cost assumptions on the recommended management
strategies derived through BMP cost-benefit optimization, two sets of costs were applied. In addition to
local Albuquerque cost data, published cost data from other regional municipalities with similar arid
climate (i.e., Southern California) were also applied. Table 13 provides a breakdown of cost components
for Albuquerque street sweeping costs, whereas Table 14 summarizes street sweeping costs from several
cities in Southern California, according to a weekly sweeping schedule. For this study, two scenarios
35
-------
were considered using the most regionally relevant values: (1) the lower cost literature values from
Southern California and (2) the higher cost local values from Albuquerque.
Table 12. Summary of literature values for street sweeping costs
Street sweeping cost
($ per curb mile swept)
Original
$25
Vacuum: $25
Brush: $45
$68
$164
Present valuet
$30
Vacuum: $36
Brush: $64
$156
$164
Note
2009 dollars, total agency/
contractor cost for sweeping
and disposal
2005 dollars, include
equipment, operation and
maintenance costs
1995 cost, Include labor,
equipment, and material cost
2012 dollars, includes owner,
employee, equipment and
operations and disposal costs
Locale
Southern
California
Minnesota
Michigan
City of
Albuquerque
Source
City of Costa Mesa
City Council Report
2009
RMWMD 2005
USEPA1999
City of Albuquerque
2012
t Present value assumes a 5% interest rate from original date of published costs, and rounded up to nearest dollar
Table 13. Albuquerque FY2012 street sweeping cost
Street sweeping cost components
Equipment replacement cost
Fleet cost (including owner cost and
operational cost)
Employee wages cost
Disposal cost
Total
Component cost
($ per curb mile swept)
$1.77
$127.82
$29.80
$4.14
$163.54
Table 14. Street sweeping costs in California
Agency/contractor
City of Costa Mesa by city staff
Clean Sweep (contractor)
Nationwide (contractor)
Dickson (contractor)
City of Newport Beach by city staff
City of Irvine by city staff
City of Irvine by contract
City of Huntington Beach by contract
Average
Cost components
($ per curb mile per week)
Sweeping
$11.34
$24.00
$23.00
$26.00
$14.00
$20.00
$24.00
$26.28
$21.08
Disposal
$2.89
$4.66
$4.88
$6.00
$3.91
$3.76
$3.76
$3.47
$4.17
Total
$14.23
$28.66
$27.88
$32.00
$17.91
$23.76
$27.76
$29.75
$25.11
Source: City of Costa Mesa City Council Report 2009
To determine a reasonable annual cost estimate for street sweeping in Albuquerque, the length of road
requiring sweeping needed to be estimated. Considering only roads in the Albuquerque jurisdiction,
36
-------
shown in Figure 19, approximately 2,400 miles of road were found. This results in an approximate road
density of 0.02 mile/acre. Assuming that all roads have two curbs, the curb miles per acre are 0.04
mile/acre. Applying this density to a 100-acre area, costs were estimated on the basis of 4.1 curb miles
per 100 acres.
Using street sweeping costs from Southern California, the present value cost of $25.11 in 2009 dollars is
$29.07 in 2012 dollars. This value was previously rounded up to $30 in Table 12 for comparison
purposes. Assuming 4.1 curb miles per 100 acres, this estimate of curb miles results in a street sweeping
costs ranged from $119.19 per 100 acres per week swept (for the literature-based cost data scenario) to
$670.51 (for local cost data scenario). These rates can then be annualized using the frequency of
sweeping in a year (i.e., $119.19x« weeks or $670.51 x n weeks). Table 15 shows the final sets of street
sweeping costs and estimated removal rates by sweeping frequency for both scenarios that were applied
for optimization.
Table 15. Street sweeping removal rates and cost estimates by sweeping frequency
Sweeping frequency
Monthly
Biweekly
Weekly
Sediment/bacteria
removal efficiency
5%
35%
70%
Sweeping cost
($/percurb mile per year)
California3
$348
$755
$1,510
City of Albuquerque'3
$1,962.48
$4,252.04
$8,504.08
Derived from average cost values from the City of Costa Mesa City Council Report 2009, converted to 2012 dollars
' Derived from street sweeping cost provided by City of Albuquerque through EPA Region 6, in 2012 dollars
37
-------
Legend
J City Jurisdiction
Roads within City Jurisdiction
All Roads
Albuquerque Roads
NA.D !S93 HARM Sta!«P!arw New Uatr.a Centra I FIPS 3002 Fool
M.ip proiiuc«d M12-OS 31
Figure 19. Roads in and outside the Albuquerque jurisdiction.
Less information was found regarding pet waste management programs. Although no literature values
were discovered regarding the bacteria removal effectiveness of pet waste management, a bacteria source
tracking study conducted in 2004 in Middle Rio Grande-Albuquerque watershed revealed that 21.9
percent of fecal coliform bacteria originate from dogs/canine (Figure 20). Therefore, for pet waste
management, it was assumed that a combination of actions-based measures representing behavioral
changes within the population could achieve a maximum bacterial removal efficiency of 20 percent. The
scope of this study was limited to evaluating impact that such a program might have toward pollutant load
reduction relative to other practices within the context of a cost-benefit optimization framework. A
maximum efficiency of 20 percent is assumed because it was recognized that the benefit of pet waste
management would only be applicable on a certain subset of land uses types within any given watershed.
No specific implementations strategies or programs are being recommended.
While pet waste management programs are effective when implemented and followed, quantifying the
associated costs are somewhat subjective and indirect. The cost of a pet waste management campaign
varies depending on several factors, including the materials produced (signs, ads, cleanup stations) and if
any funding mechanism can be established to offset the cost. For these and related factors, no local data
were found. Considering the significant uncertainty associated with pet waste management costs, this
study did not explicitly represent this cost component in the optimization analysis. Although to assess the
impact of pet waste management programs, optimization scenarios with and without pet waste
management were compared.
38
-------
Sheep, 0.5
Raccoon, 0.5
Porcine, 1.4
Source: NMED-SWQB 2010
Figure 20. Fecal coliform bacteria contribution distribution among sources.
39
-------
Chapter 5. SUSTAIN Application for Representative Area
While this case study focuses on the proposed MS4 watershed-based permit area, the extent of this area
spans 842.5 square miles (approximately 540,000 acres), which is beyond the practical application scale
of a single SUSTAIN model. The MS4 watershed-based permit area also encompasses areas outside of
Bernalillo County that are not fully represented in the datasets discussed in Chapter 2. As with any
numerical construct, modeled SUSTAIN networks are simplifications of actual hydrologic and hydraulic
networks. To minimize error associated with modeled time of concentration on an hourly time step,
earlier SUSTAIN applications have shown that 100 acres is an appropriate spatial scale for subwatersheds
(USEPA, 2012). In lieu of modeling the entire MS4 permit area, a representative subwatershed model
was constructed for this analysis that (1) represents the urban land use distribution typically found in the
MS4 permit area, and (2) satisfies the requirements of model spatial scale for use in SUSTAIN.
To develop this representative model of the MS4 permit area, the HRU distribution in available
jurisdictional boundaries was tabulated. The distribution was then normalized down to a 100-acre size for
modeling purposes. Two jurisdictional boundaries were used as the basis for summarizing HRU
distributions for the representative model area. The first boundary was the intersection of the MS4 and
Bernalillo County boundaries. The second was the Albuquerque jurisdictional boundary. Figure 21 is a
map showing the extent of the various data layers.
Legend
J City Jurisdiction
| MS4 Boundaries
Land Use Data
Land Use Data
Albuquerque Land Use and Boundaries
NAD
-------
Figure 22 presents a comparison of each HRU distribution summarized using the boundaries from the
MS4-Bernalillo County overlap and Albuquerque. Because the MS4 boundary extended well outside the
urban core, it encompasses more of the vacant or undeveloped areas on the northern, eastern, and western
perimeter of Albuquerque. Limiting the boundary to the intersection of the MS4 area and Bernalillo
County helped emphasize the urban core; however, the overall HRU distribution was still dominated by
vacant land. In contrast, the city's jurisdictional boundary by definition encompasses more urban areas.
Table 16 presents a comparison of the HRU distributions in the two boundaries. Because the objective of
this SUSTAIN case study is to assess placement of urban BMPs for a watershed-based permit, the HRU
distribution developed using the city's jurisdictional boundary was chosen for the representative 100-acre
model area.
Agriculture
Commercial Retail
Commercial Service
Drainage/Flood Control
Industrial/Manufacturing
Multi-family
Parking Lots/Structures
Parks/Recreation
Public/Institutional
Roads
Single Family
Transportation/Utilities
Vacant/Other
Wholesale/Warehousing
MS4-Area/Bemalillo County Intersect
City of Albuquerque Jurisdiction
5%
10%
15% 20% 25%
Percent Area
30%
35%
40%
Figure 22. Comparison of HRU distributions summarized using the MS4 and Albuquerque's
boundaries.
41
-------
Table 16. HRU distribution for the 100-acre SUSTAIN conceptual model
Category
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Name
Agriculture
Commercial Retail
Commercial Service
Drainage/Flood Control
Industrial/Manufacturing
Multifamily
Parking Lots/Structures
Parks/Recreation
Public/Institutional
Single Family
Transportation/Utilities
Vacant/Other
Wholesale/Warehousing
Roads
HRU distribution
MS4area
Bernalillo County
1.00%
1.06%
1.46%
1.40%
1.60%
0.86%
0.14%
20.12%
11.51%
14.54%
1.47%
36.40%
0.51%
7.92%
City of Albuquerque
0.15%
2.86%
3.33%
2.13%
1.47%
2.54%
0.42%
14.67%
3.67%
23.92%
4.19%
25.16%
1.15%
14.35%
100-Acre model
(acres)
0.15
2.86
3.33
2.13
1.47
2.54
0.42
14.67
3.67
23.92
4.19
25.16
1.15
14.35
5.1. BMP Scenarios
Management practices in the 100-acre representative model included a combination of both non-structural
and structural practices. The nonstructural BMPs were represented as changes in the modeled runoff
boundary conditions while the structural practices were modeled explicitly using BMPs commonly
available in SUSTAIN. The nonstructural BMPs scenarios included varying levels of street sweeping
and/or pet waste management. Overall, 11 managed scenarios and one baseline scenario (no street
sweeping and no pet waste management) were constructed as summarized in Table 17. The levels of the
nonstructural BMPs are indicated by corresponding pollutant percent removals.
Table 17. Summary of nonstructural BMP scenarios
Scenario ID
Baseline
1
2
3
4
5
6
7
8
9
10
11
Pollutant percent removal
Street sweeping
-
5%
35%
70%
-
5%
35%
70%
—
5%
35%
70%
Pet waste management
-
—
-
—
10%
10%
10%
10%
20%
20%
20%
20%
42
-------
These combinations of scenarios evaluate performance of the varying levels of both street sweeping and
pet waste management also reflects the interactions between the two nonstructural practices considered
here. For example, for the roads where both street sweeping and pet waste management applies, the
combination of street sweeping with 35 percent pollutant removal and pet waste management with 10
percent pollutant removal will not result in an overall removal efficiency of 45 percent; instead the total
removal efficiency is 41.5 percent, calculated as
41.5% = 1 00% - [(l 00% - 1 0%) x (l 00% - 3 5%)]
As described previously in this section, three types of structural BMPs, i.e., rainwater collection system,
xeriscape, and detention basin, are considered. Figure 23 illustrates the BMP routing configuration.
Runoff from road, parking lot, commercial, industrial, public/institutional, multi -family residential, and
single-family residential areas are collected by rainwater collection system, the overflow from the
collection system during storm event drains to downstream xeriscape; during dry weather, the collected
rainwater will be used to irrigate the downstream xeriscape. The maximum collection system storage
volume is assumed to capture 0.5 inch runoff from the impervious area. Because the pollutant loading
characteristics vary among the different land use categories, the performance of rainwater collection
systems and xeriscape are expected to differ when treating runoff from different land uses. To compare
performance of BMPs for various land use categories, each land use was assigned a unique representation
of collection system and xeriscape. In addition to distributed BMPs, a regional detention basin designed
to treat the entire 100-acre area was also included in the potential BMP network.
Treated Drainage Area Land Distribution
Commercial
Road
^^^^ ^^^^^
Rainwater
Collection
System
Xeriscape
I
1
Industrial
I Road
11
^^^^ ^^^^^
Rainwater
Collection
System
Xeriscape
I
Public-
Institutional
Road
Rainwater
Collection
System
Xeriscape
I
Parking
Road
Rainwater
Collection
System
Xeriscape
\
Multi-
Family
Residential
Road
^^^^ ^^^^^
Rainwater
Collection
System
i — ^n
Xeriscape
I
Single-
Family
Residential
Other
Land Use
Road | Road
^^^^ ^^^^^
Rainwater
Collection
System
| /—
Xeriscape
1.1
I
)
s
Detention
Basin
Outlet
Figure 23. Structural BMP routing network configuration.
Xeriscape
The potential xeriscape area is estimated on the basis of GIS analysis. The first step is estimating the
impervious surface area percentages of the selected land use categories. The city's land use raster
geospatial layer was converted to a polygon file. An aerial image available from ESRI was used as a
43
-------
background to identify impervious surfaces. In each selected land use plot, polygons were drawn to
represent the impervious surfaces (roofs, pavements, and the like). In developing impervious surface
polygons, the focus was on the larger areas, and some of the smaller areas, such as sidewalks, might have
been excluded for simplicity. These smaller areas, however, are believed to be a relatively minor area
compared to the total impervious area and most likely would not significantly influence the overall
calculation of average impervious areas.
For some areas, the aerial image is difficult to interpret because of resolution and other issues in the
image. The resolution quality of the image is a factor for distinguishing smaller details and thus
classifying them as impervious. For example, in housing land uses, smaller details such as
porches/patios, sheds, and the like might have been difficult to distinguish. Because of the nature of the
image, shadows from buildings and large trees could have interfered with classifying some areas as
impervious or pervious. Another issue related to the nature of the image and image resolution is being
able to distinguish between impervious surfaces, dead/dry grass, bare ground/dirt, or xeriscape in some
areas. Figure 24 through Figure 29 show an example plot of each land use type with the impervious
surface polygons. To calculate the percent impervious area, the area of impervious surface in a plot was
compared to the total area of the plot. To generate the average impervious surface area by land use type,
individual plot percentages were averaged by category. After the impervious area percentages were
determined, the approximate percentages of pervious areas available for xeriscape were estimated. This
was done by a visual inspection of an aerial image of the selected land use plots and their pervious areas.
Table 18 summarizes the area percentages that could be converted to xeriscape.
44
-------
Legend
Land use boundary
| | Impervious surface
Albuquerque, New Mexico
Commercial Retail Land Use
Figure 24. Commercial retail land use (Category 2) example plot with impervious surfaces highlighted.
45
-------
Figure 25. Commercial service land use (Category 3) example plot with impervious surfaces
highlighted.
46
-------
Figure 26. Industrial/manufacturing land use (Category 5) example plot with impervious surfaces
highlighted.
47
-------
Legend
Land use boundary
| | Impervious surface
Albuquerque, New Mexico
Multi-family Land Use
Figure 27. Multifamily land use (category 6) example plot with impervious surfaces highlighted.
48
-------
Figure 28. Public/institutional land use (category 9) example plot with impervious surfaces
highlighted.
49
-------
Legend
Land use boundary
| | Impervious surface
Albuquerque, New Mexico
Single Family Land Use
Figure 29. Single family land use (category 10) example plot with impervious surfaces highlighted.
50
-------
Table 18. Estimated potential area that can be converted to xeriscape
Land use category
Commercial (including
retail and service)
Industrial
Multi-family Residential
Parking Lot
Public/Institutional
Single-family Residential
Impervious area
including roads
(acres)
6.72
1.48
1.87
0.55
3.65
20.46
Pervious
area
(acres)
0.90
0.16
1.14
0.02
0.44
11.24
Maximum potential xeriscape area
Potential
xeriscape area
(% of per.
area)
30%
50%
70%
100%
60%
70%
Total area
(acre)
0.27
0.08
0.80
0.02
0.26
7.87
Normalized area
(ac. perac. imp
drainage area)
0.04
0.05
0.43
0.04
0.07
0.38
Detention Basin
One detention basin facility is assumed for the 100-acre representative area. The detention basin is
designed to retain the water quality capture volume (WQCV). This volume was calculated using the
following relationship:
WQCV = ax(o.91/3 -1.19/2 + 0.78/)
where WQCV is the water quality capture volume in inches, a is a coefficient corresponding to WQCV
drain time (defined in Table 19), and / is the percent imperviousness.
Table 19. Drain time coefficients for WQCV calculation
Drain time
(hours)
12
24
40
Coefficient, a
0.8
0.9
1.0
Percent imperviousness / was estimated for each land use type in the Albuquerque watershed and
tabulated in Table 20. From this information, an area-weighted average was calculated to get the average
percent imperviousness in the watershed.
Table 20. Summary of percent imperviousness in Albuquerque by land use category
Category
1
2
3
4
5
6
7
8
Name
Agriculture
Commercial Retail
Commercial Service
Drainage/Flood Control
Industrial/Manufacturing
Multifamily
Parking Lots/Structures
Parks/Recreation
Percentage
0.15%
2.86%
3.33%
2.13%
1.47%
2.54%
0.42%
14.67%
Area
(acres)
0.15
2.86
3.33
2.13
1.47
2.54
0.42
14.67
Imperviousness
15%
85%
86%
100%
89%
55%
100%
20%
51
-------
Category
9
10
11
12
13
14
Name
Public/Institutional
Single Family
Transportation/Utilities
Vacant/Other
Wholesale/Warehousing
Roads
Percentage
3.67%
23.92%
4.19%
25.16%
1.15%
14.35%
Area
(acres)
3.67
23.92
4.19
25.16
1.15
14.35
Area weighted %
imperviousness
Imperviousness
88%
53%
90%
20%
85%
90%
52.1%
Using an average imperviousness of 52 percent and assuming a drainage time of 40 hours, WQCV was
calculated to be 0.21 inch.
)3 -1.19(0.52)2 +0.78(0.52)]= 0.21
WQCV outside Albuquerque can be estimated using the following relationship:
(WQCV
\ 0.43
\i^-r Other
where WQCVother is the WQCV outside the watershed in inches and d6 is the average depth of runoff-
producing storms from Figure 30.
0.50
0.40
0.60
0.70
0.60
0.50
0.50
0.70
0.80
Source: Guo and Urbonas 1996
Figure 30. Distribution of mean precipitation, de, in inches for the United States.
52
-------
Using the WQCVother equation and assuming a d6 of 0.41 inch yields a WQCVother of 0.202 inch, or 1.68
acre-feet. Considering the implementation of rainwater harvest and xeriscape throughout the drainage
area, the WQCV calculated on the basis of the existing condition might be over designed. The smaller
volume of 25, 50, and 75 percent of the calculated WQCV are also included as options.
53
-------
Chapter 6. Evaluation of BMP Performance
Evaluation of BMP performance provides valuable information regarding the magnitude and trend of
pollutant removal efficiencies under various conditions and helps determine the achievable pollutant
removal effectiveness.
Nonstructural BMPs
Pollutant removal rates of nonstructural BMPs were estimated for each scenario given in Table 17.
Comparing nonstructural BMP reductions to the baseline scenario shows both the importance of
nonstructural BMPs in reducing TSS loads (Figure 31) and E. coll loads (Figure 32).
lP = 0.27"
IP = 0.47" DP = 0.78"
I *
I 5
4
3
2
1
0
10 100
Total TSS Load (Ibs)
1000
10000
Figure 31. Comparison of nonstructural BMP scenarios and their effect on TSS load reductions for
various storm event sizes.
54
-------
IP = 0.27" •P = 0.47" DP = 0.78"
3
2
1
0
1.0E+06 1.0E+07 1.0E+08 1.0E+09 1.0E+10
n-Escherichia Coli
1.0E+11
Figure 32. Comparison of nonstructural BMP scenarios and their effect on E. coli load reductions for
various storm event sizes.
While cost estimates for street sweeping were made, costs for pet waste management programs were
difficult to estimate. For the purposes of comparing scenarios, reduction and associated costs, a cost was
assumed for pet waste management programs. For the case of 10 percent E. coli removal by the
programs, an annual cost of $1,000 was assumed for a 100-acre area. Doubling this estimate for the 20
percent case yields $2,000. It is important to note these assumptions and the current limitations of cost
estimates for pet waste management programs.
Cost versus performance was examined for both TSS and E. coli loads for the three storm scenarios.
Because pet waste management programs would not reduce TSS loads, Figure 33 shows discrete levels of
cost versus reduction as street sweeping frequency increases and pet waste management programs are
considered. Figure 34 shows an almost linear cost to reduction relationship for E. coli, even for larger
storm events. Note, however, that there is a high degree of uncertainty in the cost estimates for pet waste
management programs.
55
-------
eduction
/ U 70
60%
50%
40%
• P = 0
P = 0
AP = 0
27"
47"
78"
• • •
-• • •-
30%
0 20%
0
10%
0% A
$0
AA AA
$2,000 $4,000 $6,000
Cost per 100 acres per year
$8,000
Figure 33. Comparison of nonstructural BMP scenario costs and their reduction of TSS loads for
various storm event sizes.
80%
70%
•2 60%
o
| 50%
=5 40%
O
,,i 30%
• P = 0.27"
P = 0.47"
AP = 0.78"
CL 10%
0%
I
$0
$2,000 $4,000 $6,000
Cost per 100 acres per year
$8,000
Figure 34. Comparison of nonstructural BMP scenario costs and their reduction of E. coli loads for
various storm event sizes.
56
-------
Structural BMPs
The pollutant removal performance of structural BMPs is affected by several factors, including hydrology
and water quality characteristics of the drainage area, storm size, and size of the BMP relative to its
drainage area. In this section, performance of the selected structural BMPs, including the distributed
BMPs (i.e., the combination of rainwater collection system and xeriscape), and regional extended
detention basin, are evaluated. With BMP performance measured as the pollutant load reduction
percentages decrease with the increasing storm size, the evaluation of the BMPs was focused on the
largest design storm (0.78-inches), which has the lowest load reduction percentage.
The performance of the rainwater collection system and xeriscape on E. coll removal was presented in
Figure 35 through Figure 40 for each land use category. For these figures, the charts on left show load
reduction versus dollar spent per acre impervious drainage area while the charts on right plot show the E.
coll load reduction percentage with the near-optimal (i.e., least cost) combination of rainwater capture
storage depth and xeriscape area as square feet per acre impervious area. It can be seen that within the
maximum potential xeriscape area and 0.5-inch maximum rainwater collection storage capture depth, all
land uses, with the exception of parking lots, can achieve 100 percent load reduction for the 0.78-inch
storm; while parking lots have a maximum load reduction of 95 percent.
Commercial
1 UU
90
c 80
•° 70
1 6°
o: 50
1 40
fc 30
°- 20
10
n
~7
/
)
1
I
^J
.>
/
/
z
j
Commercial
BMP Efficiency Curve
aj 80
o
| 60
I
E 40
20
0
100 200 300
$/Acre Impervious
400
Xeriscape Area
(ac./ac. impervious)
Rainwater Collection
Capture Depth
(in)
Figure 35. Commercial land use—rainwater collection system and xeriscape performance and cost-
effectiveness curve.
57
-------
Industrial
100
90
80
70
60
CfL 50
o
-g
0
o
0
CL
40
30
20
10
Industrial
BMP Efficiency Curve
100
1
uj 80
o
| 60
i
I 40
I 20
0
100
200
300
400
Xeriscape Area
(ac/ac. impervious)
Rainwater Collection
Capture Depth
(in.)
$/Acre Impervious
Figure 36. Industrial land use—rainwater collection system and xeriscape performance and cost-
effectiveness curve.
MFR
100 200 300
$/Acre Impervious
400
100
8 90
LU 80
1 70
§ 60
1 50
£ 40
§ 30
1 20
10
°°^
/-
.1
•
%
Mutli-fam ly Residential
BMP Efficiency Curve
$
'0,
\
'"-
~~~~~—^^^
'""SMS
BJ$S
0:025
002
0:015
(TSrPte °01
' Ofefe M05 Xeriscape Area
0^3 (ac,/ac. impervious)
Rainwater Collection
Capture Depth
(in.)
Figure 37. Multi-family residential land use—rainwater collection system and xeriscape performance
and cost-effectiveness curve.
58
-------
Public Institution
Public Institution
BMP Efficiency Curve
1 UU
90
c 80
•2 70
I 6°
o: 50
I 40
g 30
°- 20
10
0
(
/
s
I
1
I
J^
)
x"
Percent Reduction of E Coli
c, 8 S S 8 S
£E^
=0
'
3.15
100 200 300 400 RacaplufeCDe
(In.)
$/Acre Impervious
0.045
0.04
0.035
0.03
0.025
0,02
>-2n« °015
0-25Q g Q QI
0350.4n,,. 0005 XeriscapeArea
0,# (ac./ac. imperviou
action
pth
Figure 38. Public/Institutional land use—rainwater collection system and xeriscape performance and
cost-effectiveness curve.
Parking Lot
100
90
c 80
•B 70
o
-g 60
£ 50
1 40
I 30
^ 20
10
0
100 200 300
$/Acre Impervious
400
Parking Lot
100
^
a
uJ 80
| 60
1
£ 40
N
i
A^ 20
°0
jf
*m~
x:
.---
s
-^~
,,--•
BMP Efficiency Curve —
i
/
"~^l
J15S7&50-3o:3ab
Rainwater Collection
Capture Depth
(in.)
^
-
-
/
004
B;g3035
0,025
n'n?
»6T5"
'o45- 0:005 XeriscapeArea
07BO (ac/ac. impervious)
Figure 39. Parking lot land use—rainwater collection system and xeriscape performance and cost-
effectiveness curve.
59
-------
SFR
100
90
80
I 70
o
"§ 6°
^ 50
1 40
fc 30
^ 20
10
100 200 300
$/Acre Impervious
400
100 -
a
lii 80
a
1 60
^
1 40
S. 20
°0
\
U.WJQ
Single-family Residential
BMP Efficiency Curve
A
fi
^
'or-
5
'~J
,
.
/
^—
8,04 "
0:035
aos
0:025
002
"•"O STr-^R^. Jl-W
O.~fc"xg~ ,&005 Xeriscape Area
07=M (ac./ac. impervious)
Rainwater Collection
Capture Depth
(in
Figure 40. Single-family residential land use—rainwater collection system and xeriscape performance
and cost-effectiveness curve.
BMP performance variations among different land uses are dependent upon hydrology and pollutant
concentration. Table 21 lists the CN and E. coll EMC of each land use category, and summarizes the
maximum BMP performances, unit impervious drainage area cost, and the BMP configurations achieving
maximum E. coll load reduction for the 0.78-inch storm. It was observed that parking lot is the land use
category with the highest CN and E. coll EMC, and it has the highest unit impervious drainage area
annual treatment cost ($310 per acre impervious drainage area); while public/institutional has the lowest
CN and E. coll EMC and, consequently, the lowest unit area annual treatment cost at $103 per acre
impervious drainage area.
Table 21. Maximum distributed BMP performances and composition by land use categories
Land use category
Commercial (including retail
and service)
Industrial
Multi-family Residential
Parking Lot
Public/Institutional
Single-family Residential
CN
94
91
90
98
86
87
E. co// EMC
(#/100 mL)
1,200
450
900
1,500
450
600
Max. E. co//
load reduction
(%)
99%
100%
100%
96%
100%
100%
BMP configurations
Rainwater
storage
(in.)
0.18
0.10
0.13
0.30
0.06
0.10
Xeriscape
(ac./ac. Imp DA)
0.040
0.024
0.043
0.040
0.016
0.039
A detention basin treating the 100-acre composite area is evaluated. The detention basin is sized with the
WQCV, as calculated in the previous section, and has a drain time of 40 hours. Figure 41 plots the TSS
and E. coll load removal efficiencies for the various storm sizes. With the 0.27-inch storm, both TSS and
E. coll have 100 percent removal because of the complete runoff volume reduction through infiltration at
the site. With the 0.47-inch storm, TSS has a 74 percent removal and E. coll a 73 percent removal. With
60
-------
the 0.78-inch storm, percent removal of TSS is 43 percent and that of E. coll is 38 percent. These percent
removal estimates are comparable with pollutant removal rates based on the Sanchez Farm monitoring
data (Radian 2011), which suggested 81 percent removal of TSS, and between 60 percent (under high-
flow conditions) and 73 percent (under low-flow condition) removal of E. coll.
100% 100%
S. 10% -
0%
0.27" Storm
0.47" Storm
0.78" Storm
Figure 41. Detention basin pollutant removal efficiency (100-acre composite drainage area).
This section presents the BMP performance evaluation results. It provides valuable information on the
maximum achievable E. coll load reduction by land use categories, and the optimal distributed BMP
compositions at various treatment levels, and corresponding costs. The information derived on the
maximum effective distributed BMP sizes are used to define the optimization search boundaries as
described in the next section.
61
-------
Chapter?. BMP Optimization
This section discusses selection of a numeric management target for optimization and outlines a series of
11 scenarios which evaluate management strategies using the previously discussed structural and
nonstructural BMPs. The cost-effectiveness curves from each of the 11 scenarios are discussed
independently to show the progression of costs and percent reduction resulting from increased
management. Different management practices may perform better at various points along the spectrum.
To capture the variability of all 11 scenarios, a composite cost-effectiveness curve is developed to
highlight the optimal management "frontier" along the boundary of all curves.
7.1. Optimization Target
The Middle Rio Grande E. coll TMDL calculates bacteria load targets as a function of flow and proposed
E. coli water quality standards. A conversion is also applied to express bacteria loads in colony forming
units per day (cfu/day) (NMED-SWQB, 2010). New Mexico regulates E. coli using two standards on the
basis of a (1) maximum concentration for single samples of 410 cfu/100 mL, and (2) a maximum
concentration of the 30-day geometric mean of 126 cfu/100 mL. These regulatory criteria were used to
develop E. coli load targets for each of the three selected precipitation depths modeled in SUSTAIN.
Table 22 presents load targets and the load reduction required according to SUSTAIN model output.
Table 22. E. coli Load targets by storm scenario depth for single sample and geometric mean
Rainfall
(in.)
0.27
0.47
0.78
Storm event
flow volume
(ft3/event)
5,435
20,581
66,850
Storm event £.
coli load
(W/event)
2.8E+08
3.9E+09
2.3E+10
Geometric
mean
target
(W/event)
1.9E+08
7.3E+08
2.4E+09
Single sample
target
(W/event)
6.3E+08
2.4E+09
7.8E+09
Geometric
mean
required
reduction
31%
81%
89%
Single
sample
required
reduction
0%
39%
66%
On the basis of Table 22, the 0.78-inch storm represents the limiting event for meeting any percent
reduction target calculated using the TMDL criteria. Management of this storm event is inclusive of
managing the 0.27- and 0.47-inch storm events. Therefore, the 0.78-inch storm will be used as the critical
condition for running the optimization scenarios targeting E. coli load reduction.
7.2. Optimization Scenarios and Results
The ultimate optimization objective is to identify the most cost effective stormwater management
solutions that meet the TMDL target. The overall optimization problem formulation can be expressed as:
Objective: Minimize Cost (BMP options, BMP size);
Minimize E. coli load generated from 0.78 inch storm
Subject to: Nonstructural BMP scenarios;
Extent of structural BMP types & sizes
As presented in Chapter 6, both nonstructural and structural BMPs are effective means to reduce E. coli
loading. When they are implemented simultaneously, because the nonstructural BMPs change the inflow
pollutant loading to the structural BMPs, the effectiveness of structural BMPs will be different from that
62
-------
under the condition without nonstructural BMPs. To test the impact of the street sweeping cost values on
recommended management strategies, two independent optimization scenarios were formulated and ran
using (1) literature values from Southern California, and (2) local values from Albuquerque. Considering
the interaction between nonstructural and structural BMPs, 12 independent optimization runs were
performed for each cost scenario (24 total runs) to derive cost-effectiveness curves for various
nonstructural base conditions, i.e., various combination of pet waste management and street sweeping, as
documented in Section 5.1. The decision variables of the optimization runs were the sizes of structural
BMPs, including rainwater collection storage, xeriscape, and regional detention basin. Nonstructural
BMPs controls were applied as part of the boundary condition for each run, with their respective costs
added afterwards to the resulting cost-effectiveness curves. Upper bounds of the decision variables were
set at the maximum values determined on the basis of the land use based BMP performance evaluation
results. Table 23 summarizes the 12 optimization run scenarios.
Table 23. Optimization run scenarios
ID
0
1
2
3
4
5
6
7
8
9
10
11
Nonstructural BMPs
Street sweeping
None
Monthly
Biweekly
Weekly
None
Monthly
Biweekly
Weekly
None
Monthly
Biweekly
Weekly
Pet waste management
None
None
None
None
10%
10%
10%
10%
20%
20%
20%
20%
Structural BMPs
BMP Types:
Rainwater collection system
Xeriscape
Detention basin
Maximum BMP Sizes:
See Table 21 for the sizes of rainwater
collection system and xeriscape.
See Section 4. 1 for the sizes of detention
basin
The cost-effectiveness curves for the 12 optimization runs are plotted in Figure 42 to Figure 48. Figure
42 shows the optimization scenarios without street sweeping. The curves in this figure are common to
both cost scenarios because street sweeping is not considered. Figure 43 shows the scenarios with
monthly street sweeping for literature-based costs, whereas Figure 44 uses local street sweeping costs.
Figure 45 shows the scenarios with biweekly sweeping for literature-based costs, whereas Figure 46 uses
local street sweeping costs. Finally, Figure 47 shows the scenarios with weekly sweeping for literature-
based costs, whereas Figure 48 uses local street sweeping costs. In each figure, three scenarios with
various levels of pet waste management are plotted. As described previously, there is uncertainty
associated with the cost of pet waste management. The costs for implementing and enforcing the
program might be offset by tax revenues or fines from violating pet owners, resulting in a negative cost in
certain cases. For this analysis, pet waste management cost was not represented in the optimization
analysis. The three lines in each plot show the total cost and E. coll load reduction generated with the
0.78-inch storm under three levels of pet waste management options.
63
-------
No Street Sweeping (0%)
100%
80%
o
T3
0)
cr
•E
0)
a
0)
Q.
60%
No Pet Waste Management (0%)
Medium Pet Waste Management (10%)
High Pet Waste Management (20%)
40%
20%
$- $5,000 $10,000 $15,000 $20,000
Cost/100 Acres/Year
$25,000 $30,000
Figure 42. Cost-effectiveness curves of optimization scenarios without street sweeping.
64
-------
Monthly Street Sweeping (5%)
100%
No Pet Waste Management (0%)
Medium Pet Waste Management (10%)
High Pet Waste Management (20%)
$5,000
$10,000 $15,000 $20,000 $25,000
Cost/100 Acres/Year
$30,000
Figure 43. Literature-based cost optimization with monthly street sweeping (5% load removal).
Monthly Street Sweeping (5%)
100%
80%
60%
•a
-------
Biweekly Street Sweeping (35%)
100%
20%
• No Pet Waste Management (0%)
• Medium Pet Waste Management (10%)
• High Pet Waste Management (20%)
$5,000
$10,000 $15,000 $20,000 $25,000
Cost/100 Acres/Year
$30,000
Figure 45. Literature-based cost optimization with biweekly street sweeping (35% load removal).
Biweekly Street Sweeping (35%)
100%
80%
60%
•a
-------
Weekly Street Sweeping (70%)
100%
No Pet Waste Management (0%)
Medium Pet Waste Management (10%)
High Pet Waste Management (20%)
$5,000 $10,000 $15,000 $20,000 $25,000 $30,000
Cost/100 Acres/Year
Figure 47. Literature-based cost optimization with weekly street sweeping (70% load removal).
Weekly Street Sweeping (70%)
100%
80%
60%
•a
-------
The composite optimal cost-effectiveness curve shown in Figure 49 was developed by identifying the
most cost-effective solutions from among all model runs for the various sets of cost assumptions. Figure
50 and Figure 51 plot the cost distributions for the literature and local composite curves, respectively,
shown in Figure 49. The lower panel of the figure zooms into the region associated with the TMDL
target and highlights the required 66 percent solutions.
100%
o
80%
60%
I
y 40%
S.
20%
0%
75%
70%
I
/
/
1
>
{
r
1
/
,
I
—
-
-
—
—
• • •
/•
X"1
--
.
--
^^
•
^B
""
^^m
— ^
—
I^^H
X"
^B
^•^M
M^
tm^
a^^
—X-
m
• Medium pet waste mgmt only
• High pet waste mgmt only
^^-> No street sweeping, high pet waste mgmt
^^—Monthly street sweeping, high pet waste mgmt
^^— Biweekly street sweeping, high pet waste mgmt
^^— Weekly street sweeping, high pet waste mgmt
• • • • Literature costs: Biweekly street sweeping, high pet waste mgmt.
• • Literature costs: Weekly street sweeping, high pet waste mgmt.
$10,000 S20.000 $30,000
Cost/100 Acres/Year
$40,000
$50,000
$60,000
o
65%
1
ID
O
I
60%
55%
50%
Local costs: No street sweeping, high pet waste mgmt
Literature costs: Biweekly street sweeping, high pet waste mgmt.
Literature costs: Weekly street sweeping, high pet waste mgmt.
$5,000
$8,000
$11,000 $14,000
Cost/100 Acres/Year
$17,000
$20,000
Figure 49. Composite optimal cost-effectiveness curve and selected solution meeting target.
68
-------
Figure 49 illustrates the sensitivity of street sweeping costs for influencing the recommended
management strategies. Note that because they have no assigned cost, medium and high pet waste
management are shown as discrete points at $0. Their respective performance values reflect the net
benefit of each level in the representative 100-acre watershed. Two notable inflection points are in the
lower panel of Figure 49—one at 59 percent along the green curve, and another at 70 percent along the
orange dotted curve. A look at the BMP cost distributions for both curves (Figure 50 and Figure 51)
reveals the preferred selection order for the various BMPs.
I
o
o
c
CO
(A
O
^
+rt
1/3
o
O
$30.0
$25.0
$20.0
$15.0
$10.0
$5.0
$0.0
DETENT] ON POND
I STREET SWEEPING
I XERISCAPE
RAINWATER COLLECTION SYSTEM
REDUCTION TARGET
0% ,2%
<£
&
CM
CO
iS^
in
TT
O
LO
LO
LO
o
CD
LO
CD
^o ^o
o^ o^
o in
r^- h*
o
Effectiveness (% Reduction)
Figure 50. Literature-based street sweeping cost data: BMP cost composition of the optimal solutions.
$60.0
•o
I
>• $50.0
£ $40.0
|
^ $30.0
c
co
o $20.0
^ $10.0
o
O
$0.0
• STREET SWEEPING
• DETENTION POND
• XERISCAPE
- RAINWATER COLLECTION SYSTEM
-REDUCTION TARGET
26%
0%
60%
LO
o m o m
CN CM CO CO
to
o to o 10
in in CD CD
o
r--
o
oo
Effectiveness (% Reduction)
Figure 51. Local street sweeping cost data: BMP cost composition of the optimal solutions.
69
-------
Note that at the lower reduction range from 0 to approximately 57 percent, structural practices like
rainwater collection and xeriscaping are more cost-effective for controlling runoff and bacteria
mobilization than street sweeping. At the reduction range between 57 to about 63 percent (on the basis of
literature street sweeping costs), biweekly sweeping is optimal, and weekly sweeping becomes optimal
when the reduction requirement exceeds 63 percent. Figure 51 shows that 59 percent is the point where
using xeriscaping comes into play. Without street sweeping, xeriscaping is able to yield a higher percent
removal because runoff carries more pollutants. When the detention pond enters the picture, there is an
initial mobilization cost with the benefit it provides, which when optimized, reduces the overall need for
xeriscaping. Beyond that point, the required size of the detention pond gradually increases. From 57
percent reduction upward in the lower cost scenario, those cost assumptions begin to have a more notable
influence on the results; biweekly street sweeping is chosen next, followed by weekly street sweeping at
63 percent and beyond. Similar to what occurred in Figure 51, Figure 50 also shows a brief dip in the use
of rainwater collection systems and xeriscaping because of the added benefit of pollutant removal that
more frequent street sweeping provides; however, as the reduction requirement increases, using those
distributed BMPs recovers. Under the literature-based street sweeping cost scenario, the detention basin
is never selected until the reduction requirement reaches 71 percent, suggesting that regional detention
basin is not preferred over those other practices until a high load reduction is required.
Another very interesting finding between these comparisons is the order in which the detention pond is
preferred as a function of street sweeping cost. In Figure 50, which assumes the lower literature-based
street sweeping costs, the detention pond is chosen after all other options have been exhausted. On the
other hand, Figure 51 shows that when street sweeping is more expensive, the detention pond is chosen
before street sweeping. It is interesting to note that monthly street sweeping was never selected among
the optimal solutions under the literature-based costs scenario, implying that waiting such a long time
between street sweeping events is not cost-effective for managing bacteria load in runoff. The use of
xeriscaping also moves from third second place to third as street sweeping cost increases, suggesting that
when biweekly street sweeping is cheaper, it initially reduces the demand for xeriscaping as a means of
controlling bacteria loading in runoff. Also, when street sweeping is more expensive, xeriscaping
supplements detention ponds and on-site rainwater collection as means for controlling bacteria loads
associated with runoff.
The TMDL target established in Section 7.1 can be translated into a 66 percent E. coll load reduction
during the 0.78-inch storm. As indicated in Figure 49, Figure 50, and Figure 51, the optimal solutions
that achieve a 66 percent E. coll load reduction have total costs ranging between $8,736 and $12,955 per
year for the 100-acre representative study area, depending on the assumed costs for street sweeping. For
the literature-based scenario, the solution selected high pet waste management, rainwater collection
systems, weekly street sweeping, and xeriscaping. Assuming literature-based street sweeping values,
Table 24 lists component costs of the nonstructural and structural BMPs, and the corresponding load
reduction of both E. coll and TSS. The table shows that the nonstructural BMP components (pet waste
management and street sweeping) cost $6,197 per year and achieve 51.7 percent E. coll and 34.5 percent
TSS loading reduction, but the structural BMP components (rainwater collection and xeriscaping) cost
$2,539 per year and generate an additional 14.3 percent load reduction of E. coll and 18.0 percent load
reduction of TSS. With literature-based street sweeping costs, on a unit cost per percent E. coll load
reduction, nonstructural BMP (about $120/percent reduction) is more cost-effective than structural BMPs
($177.5/percent reduction).
70
-------
Table 24. Optimal literature-based costs solution achieving 66% E. coli load reduction (0.78-inch storm)
Solution component
Cost ($/year per 100 ac.)
E. coli load reduction (0.78-inch storm)
TSS load reduction (0.78-inch storm)
BMPs
High pet waste mgmt.,
weekly street sweeping
$6,197
51.7%
34.5%
Rainwater collection
and xeriscape
$2,539
14.3%
18.0%
Total
$8,736
66.0%
52.6%
On the other hand, Table 25 lists component costs of the nonstructural, distributed, and centralized BMPs,
and the corresponding load reduction of both E. coli and TSS assuming local street sweeping costs. The
table shows that the nonstructural BMP component (high pet waste management) cost $0 per year and
achieves 12 percent E. coli (TSS loading reduction is not applicable), while distributed BMP components
(rainwater collection and xeriscaping) cost $4,477 per year and generate an additional 43.5 percent load
reduction of E. coli and 50.0 percent load reduction of TSS. The centralized BMP (wet pond) cost is
$8,478 per year and generates an additional 22.6 percent reduction of E. coli and 13.6 percent load
reduction of TSS. Using local street sweeping costs, the unit cost per percent E. coli load reduction for
distributed BMPs ($103/percent reduction) is more cost-effective than the wet pond ($375/percent
reduction).
Table 25. Optimal local costs solution achieving 66% E. coli load reduction (0.78-inch storm)
Solution component
Cost ($/year per 100 ac.)
E. coli load reduction (0.78-inch
storm)
TSS load reduction (0.78-inch
storm)
BMPs
High pet waste
management and
no street sweeping
-
12.0%
—
Rainwater
collection and
xeriscape
$4,477
43.5%
50.0%
Extended
detention
pond
$8,478
22.6%
13.6%
Total
$12,955
66.1%
63.6%
Table 26 shows structural BMP composition for the selected optimal solution with the literature-based
cost solution. Table 27 shows similar information for the local cost solution. They suggests that parking
lots should have the highest treatment level (0.3-inch collection depth, and 0.03-acre xeriscape per acre
impervious drainage area), followed by commercial and multi-family residential (0.2-inch collection
depth, and 0.03-acre xeriscape per acre impervious drainage area). The literature cost solution then adds
industrial area (0.2-inch collection depth, and 0.01-acre xeriscape per acre impervious drainage area), and
public/institutional area (0.1-inch rainwater storage depth), but does not use treatment for single-family
residential.
Table 26. Distributed BMP composition of the optimal solution achieving 66 percent E. coli load
reduction for the 0.78-inch storm
Land use category
Commercial
(including retail and service)
Industrial
Multi-family Residential
Parking Lot
BMP configurations
Rainwater storage
(in.)
0.2
0.2
0.2
0.3
Xeriscape
(ac./ac. impervious
drainage area)
0.03
0.01
0.03
0.03
71
-------
Land use category
Public/Institutional
Single-family Residential
BMP configurations
Rainwater storage
(in.)
0.1
--
Xeriscape
(ac./ac. impervious
drainage area)
-
--
On the other hand, the local cost solution uses single-family residential treatments (0.2-inch rainwater
storage depth) and reduces industrial treatment. This solution also uses 50 percent of the maximum
possible design size for the wet pond. For both of these solutions, the treatment prioritization was
dependent on the runoff and pollutant loading characteristics of the various land use categories, as
previously discussed in Section 5.1.
Table 27. BMP composition of the optimal local cost solution achieving 66 percent E. coli load
reduction for the 0.78 inch storm
Land use category
Commercial
(including retail and service)
Industrial
Multi-family Residential
Parking Lot
Public/Institutional
Single-family Residential
100-acre drainage area
BMP configurations
Rainwater storage
(in.)
0.2
0.1
0.2
0.3
0.1
0.2
n/a
Xeriscape
(ac./ac. impervious
drainage area)
0.03
-
0.03
0.03
-
-
n/a
Extended detention
pond
(%ofWQCV)
-
-
-
-
-
50%
72
-------
Chapter 8. Summary and Conclusions
This case study presented an approach to identify cost-effective stormwater management strategies with
the objective of reducing E. coll loading based on a target consistent with the Middle Rio Grande E. coll
TMDL. The report documents the various steps in developing a SUSTAIN application, including
establishing baseline hydrology and water quality representation, identifying the critical condition for
management, formulating management objectives, and synthesizing model outputs with a focus on the E.
coli TMDL and pending WBP. Through this study, a framework was established that can be applied to
support the development of implementation strategies for meeting water quality targets associated with
both the E. coli TMDL and pending WBP. The study concludes by reflecting on key insights gained in
relation to the original study objectives, which are as follows:
• Develop a boundary condition using a simple hydrology method (such as TR-20/55 or Rational
Method), in conjunction with event-mean concentrations
• Evaluate the trade-offs in cost and management performance of both structural and nonstructural
BMPs during the optimization process
• Estimate water quality performance of proposed management practices based on critical
conditions identified in the Middle Rio Grande E. coli TMDL to support a pilot MS4 watershed-
based permit under development by EPA Region 6.
Before dissecting these three objectives, it is important to reassess the role of models in general and
SUSTAIN specifically in the planning process. The optimization approach used in SUSTAIN offers a
unique evaluation platform for evaluating complex, spatially heterogeneous urban system and associated
cost benefit implications of management decisions. In a classic text on numeric methods and modeling,
Hamming suggests that "the purpose of computing is insight, not numbers" (Hamming 1973). The utility
of this approach lies not in the ability to compute and a single number but to uncover patterns and trends
that provide important insights that would not have otherwise been gained. SUSTAIN is a comprehensive
decision support system meant to enhance the stormwater management planning and implementation
process by allowing stakeholders to evaluate and visualize the tradeoffs between management and cost.
As the planning process moves forward and implementation begins, the framework developed through
this case study offers a methodology that can be used for developing future modeling studies. The
following sections describe lessons learned for each case study objective.
8.1. Baseline Representation
While previous SUSTAIN case studies focused on evaluating management practices intended for specific
locations (sewersheds) in support of narrowly focused management objectives, this study in Albuquerque
focused on broader management strategies at a bigger spatial scale. Even though no existing continuous
simulation model was available as a basis for developing input boundary conditions, it is unlikely that
such a model would have offered plug and play integration into the SUSTAIN system as this study was
spatially broader than the scope of many detailed models. Moreover, the arid climate setting for this
analysis (about 9.5 inches of rainfall per year) explains why continuous simulation models are not widely
applied or available. These challenges presented an opportunity to demonstrate alternative approaches to
model development. The study began by first identifying locally acceptable methods for representing key
spatial and temporal characteristics of the system, specifically as follows:
• Temporal representation of boundary conditions
o Selection of design storms
73
-------
o Development of baseline runoff hydrographs
o Development of baseline water quality load time series
• Spatial representation of the study area
Considering the arid climate and frequency of rainfall events, this study explored an alternative approach
for developing the baseline rainfall/runoff response. Three design storms were selected and linked to
instream conditions using the FDC zones established during E. coli TMDL development. A simple
hydrology method, TR-55, in conjunction with event-mean concentrations, was used to develop the
boundary conditions for each HRU. The method TR-55 approach offered a flexible alternative to a true
continuous simulation boundary condition that was easy to configure and integrated seamlessly with
SUSTAIN via external ASCII time series files. As TR-55 is probably the most widely used representation
of hydrology, its use in a SUSTAIN application creates an opportunity for expansion of the software user
community.
Although the Rio Grande watershed drainage area encompasses approximately 14,000 square miles at the
pour point where the river reaches Albuquerque, New Mexico; the scope of this case study was limited
strictly to stormwater contributions originating from local urban areas in Albuquerque and Bernalillo
County. While this case study also considered the proposed MS4 watershed-based permit area, the entire
extent of this boundary is beyond the practical scale for a SUSTAIN model application. The watershed-
based permit area spans 842.5 square miles (approximately 540,000 acres) and encompasses areas outside
the Bernalillo County land use spatial data set used to develop the SUSTAINHRUs and baseline
condition. As with any numerical construct, networks modeled in SUSTAIN are simplifications of reality.
In lieu of modeling the entire MS4 permit area, a representative watershed model was constructed for this
analysis that (1) represents the urban land use distribution typically found in the MS4 permit area, and (2)
satisfies the requirements of model spatial scale for use in SUSTAIN. This study modeled a representative
100-acre area that had the same relative land use distribution as the entire MS4 permit area. Building the
SUSTAIN analysis on the 100-acre representative site yields model results without the overhead of
simulating the full spatial extent of the MS4 permit area.
8.2. Evaluation of Structural and Nonstructural BMPs
Stormwater BMPs for the arid and semi-arid region are an important tool for preserving and improving
the water quality of the region's precious water resources, and those BMPs should be designed using the
unique climate characteristics to meet the special needs (water conservation, reuse, and sustainable land
use management) of the region. Several references suggest that the most effective BMPs for the arid
Southwest are those that focus on stormwater conservation and water reuse, for example rainwater
harvesting, local groundwater infiltration when feasible, and minimizing evapotranspiration losses. In
this study, rainwater harvesting and xeriscape were selected as site-scale practices because (1) they
conserve water, and (2) they contribute to runoff source control. In addition, regional detention basins
were also modeled because they not only provide water quality management (sediment settling), but also
flood control. In addition to those structural practices two nonstructural practices, street sweeping and pet
waste management, were modeled.
Before the optimization analysis, each individual BMP was evaluated for its effectiveness in reducing
pollutant loading. The evaluation was based on the largest design storm, which represented the limiting
boundary condition. BMP performance evaluation results revealed valuable information on the maximum
achievable E. coli load reduction by land use category and the optimal distributed BMP composition and
cost at various treatment levels. The maximum distributed BMP sizes associated with the largest design
74
-------
storm were used to define the upper bounds for optimization. On the basis of the assumptions outlined in
this study, the following insights were gained with respect to the cost-effectiveness of management
practices:
• The incentive-based rebate programs designed to encourage the use of xeriscape and rainwater
harvesting features were the first to be maximized during optimization because they represented
the most cost-effective component for achieving the management target. Opportunity for those
types of features focused primarily on residential land uses that shared the burden of capital and
O&M costs with the homeowners (through the rebate incentive program) rather than entirely on
the public entity. However, this introduces some added uncertainty on BMP performance in
cases where homeowners do not maintain the practices at designed conditions.
• Nonstructural BMPs are effective and economical measures for reducing nonpoint source
pollution. Because pet waste management had no assigned cost (except for possible public
education costs), it was maximized in every optimized solution—this type of source control
directly addresses the E. coll pollutant problem. When the lower literature costs were used for
street sweeping, it was selected early in the trajectory of near-optimal solutions; however, when
the higher local costs were used, it was not selected as part of the required treatment level. Street
sweeping cost assumptions also affected the order in which structural practices were selected.
When street sweeping costs are lower, the practice reduces the overall need for structural
practices by controlling loading at the source. The results also suggested that more infrequent
sweeping (i.e., monthly) was not cost-effective compared to what can be achieved with
distributed practices. When street sweeping costs are higher, centralized wet ponds become an
attractive and necessary supplement to rainwater conservation practices. Street sweeping was
considered only after all structural practices were exhausted. Those outcomes should be
interpreted with the following caveat:
o The large range of street sweeping costs (literature versus local sources) suggests cost
estimates might not be directly comparable. For example, if the higher cost estimates
include equipment purchase, storage, and maintenance, a more cost-effective solution
might be to contract the street sweeping work to a private entity that is able to reduce
overhead costs through economies of scale. For instance, a street sweeping contractor
will share overhead costs among its clients through work scheduling, because their
equipment will not be sitting idle during off weeks.
• For the scenarios where street sweeping costs were low, regional detention basins were not
selected for achieving a management target of a 66 percent E. coll reduction. That outcome
should be interpreted with the following caveat:
o The study did not consider flood control as a management objective. The objective
function and constraints that guide the optimization process were focused solely on
minimizing cost to achieve reductions in E. coll load during a 0.78-inch design storm.
Flood control is a critical component of stormwater management in Albuquerque and
must be properly evaluated against an independent set of management objectives and
constraints. The results of this study were not intended to evaluate BMPs against flood
control criteria.
• The most cost-effective solution is dependent on the selected management (load reduction) target.
As shown in Figure 49, the composition of any best solution (cost and practices) is dependent on
the management target selected on the y-axis. In this case study, a load percent reduction target
of 66 percent was calculated using the runoff volume of the limiting design storm and E. coli
concentration specified in the TMDL. Selecting a different load reduction target would result in a
different mixture of BMPs.
75
-------
It should be emphasized that the conclusions of this study are highly sensitive to the cost assumed for the
various management practices and programs. BMP costs are critical to the optimization goals of meeting
management objectives while minimizing the total solution cost. The cost of the site-scale, distributed
BMPs were determined using information about the local rebate program(s) that promote xeriscape and
rainwater harvesting features. Regional detention basin costs in this study were based on capital and
maintenance costs estimated by Bernalillo County. This study also does not explicitly consider costs
associated with a comprehensive pet waste management initiative. It is reasonable to assume that in
practice, this type of program would incur some costs associated with administration, public outreach and
education, and production costs associated with literature, signage, or other outreach materials. Some of
those costs might be offset by the institution and enforcement of fines collected from pet waste violators.
Although it appears that the rebate program cost is significantly less than that of the capital costs, rebates
were considered a reasonable way to reflect the economic value of rainwater harvesting and water
conservation. Implementation of those distributed BMPs is inherently more complex than that of
centralized practices and subject to a higher degree of uncertainty because these practices are generally
distributed throughout a catchment area. The costs of O&M, while more focused and predictable for
centralized practices, could become more burdensome and unpredictable when BMPs are scattered
throughout a watershed.
8.3. Applicability to Watershed Based MS4 Permit
The Middle Rio Grande watershed, and specifically Albuquerque, was selected for this case study
because of the recent E. coli TMDL and ongoing development of an MS4 WBP. According to EPA
Region 6, this permit will identify a number of implementation options for regulated MS4 operators. The
SUSTAIN application framework established in this study serves as a quantitative tool to evaluate and
select management practices for fulfillment of the WBP. Specifically, the findings related to (1) the cost-
effectiveness of specific BMPs, (2) conditions under which BMPs are and are not effective, and
(3) tradeoffs between structural and nonstructural BMPs provide meaningful guidance to stakeholders
(both implementers and regulators) for selecting cost-effective measures.
While it was not practically feasible to model the full permit area or contributing TMDL boundaries in
detail, relevance was maintained by using a locally representative land use distribution scaled to a
manageable size (approximately 100 acres) for use as a single subwatershed in SUSTAIN. The
configuration of this scaled-down model accounted for the spatial distribution and relative magnitude of
runoff and pollutant loadings from individual land uses in the Albuquerque metropolitan area.
Nevertheless, specific implementation strategies for compliance under the MS4 permit will need to be
further tailored to conditions and constraints at specific locations. It is recognized that any location in a
watershed will have unique characteristics that will influence what can and cannot be done there. For this
reason, no model can replace the need for on-the-ground engineering and planning. Nevertheless, given
the range of conditions and constraints reflected in this modeling study, it is likely that a cost-effective
solution for achieving a 66 percent E. coli reduction would cost between $8,736 and $12,955 per 100
acres per year (assuming a 20-year BMP life cycle). These outcomes offer a scalable benchmark on
which to compare future monitoring, modeling, and implementation outcomes for proposed management
strategies in the Albuquerque metropolitan region.
76
-------
Chapter 9. References
ABCWUA (Albuquerque Bernalillo County Water Utility Authority). 2012. Xeriscape and Rainwater
Harvesting Landscape Rebate Program, .
Accessed May, 2012
Ahn, J.H., S.B. Grant, C.Q. Surbeck, P.M. DiGiacomo, N.P Nezlin, and S. Jiang. 2005. Coastal water
quality impact of stormwater runoff from an urban watershed in Southern California.
Environmental Science & Technology 3 9(16): 5 940-5 95 3
Albuquerque GIS (Albuquerque geographic information systems). 2012. Data downloads.
. Accessed January 14,2012.
City of Costa Mesa. 2009. Evaluation of 'Current and Proposed Street Sweeping Service Levels. City
Council Study Session Agenda Report. Prepared by Public Services Department - Maintenance
Services Division for City of Costa Mesa, California. October 13, 2009.
Gautam, M., K. Acharya., and M. Stone. 2010. Best management practices for stormwater management
in the Desert Southwest. Journal of Contemporary Water Research & Education 146:39-49.
Glass, S. 2012, January 6. Email from Steve Glass, Bernalillo County, to Ryan Murphy, Tetra Tech,
Inc., regarding Bernalillo County water quality monitoring data.
Guo, J.C.Y., and B. Urbonas. 1996. Maximized Detention Volume Determined by Runoff Capture
Ratio, Journal of Water Resource Planning and Management 122(1):33.
doi:10.1061/(ASCE)0733-9496(1996)122:l(33)
Hamming, R.W. 1973. Numerical Methods for Scientists and Engineers. Dover Publications, Inc: New
York, NY.
LaBadie, K.T. 2010. Identifying Barriers to LID and GI in the Albuquerque Area.
. Accessed August 23, 2011.
Lin, J.P. 2004. Review of Published Export Coefficient and Event Mean Concentration Data. U.S.
Army Corps of Engineers, Wetlands Regulatory Assistance Program, Vicksburg, MS.
MRGARWG (Middle Rio Grande-Albuquerque Reach Watershed Group). 2008. Middle Rio Grande-
Albuquerque Reach Watershed Restoration Action Strategy. Prepared for the New Mexico
Environmental Department, Santa Fe, NM.
NMED-SWQB (New Mexico Environment Department-Surface Water Quality Bureau). 2010. Total
Maximum Daily Load (TMDL)for the Middle Rio Grande Watershed. New Mexico Environment
Department, Santa Fe, NM.
77
-------
NRC (National Research Council). 2008. Urban Stormwater Management in the United States. The
National Academic Press, Washington, DC.
NRCS (Natural Resources Conservation Service). 2012. Win TR-55 Watershed Hydrology.
.
Accessed April 23, 2012.
Penttila, R. 2011. Albuquerque's Move Toward a Low Impact Development Future. Presentation to the
2011 MS4 Region 6 Conference, .
Accessed October 2012.
Surbeck, C.Q., S.C. Jiang, J.H. Ahn, and S.B. Grant. 2006. Flow Fingerprinting Fecal Pollution and
Suspended Solids in Stormwater Runoff from an Urban Coastal Watershed. Environmental
Science & Technology 40(14):4435-4441
USEPA (U.S. Environmental Protection Agency). 1999. Preliminary Data Summary of Urban Storm
Water Best Management Practices. EPA-821-R-99-012. Office of Water, Washington, DC
20460. . Accessed October 2012.
USEPA (U.S. Environmental Protection Agency). 2009. SUSTAIN—A Framework for Placement of Best
Management Practices in Urban Watersheds to Protect Water Quality. EPA/600/R-09/095. U.S.
Environmental Protection Agency, Water Supply and Water Resources Division, National Risk
Management Research Laboratory, Cincinnati, OH.
. Accessed August 13, 2011.
USEPA (U.S. Environmental Protection Agency). 2012. Report on Enhanced Framework (SUSTAIN)
and Field Applications to Placement of BMPs in Urban Watersheds. EPA/600/R-11/144. U.S.
Environmental Protection Agency, Water Supply and Water Resources Division, National Risk
Management Research Laboratory, Cincinnati, OH.
USGS (U.S. Geological Survey). 1996. HYSEP: A Computer Program for Stream/low Hydrograph
Separation and Analysis. Water Resources Investigation Report 96-4040. U.S. Geological
Survey, Reston, VA.
78
-------
Appendix: HRU Runoff Hydrographs
79
-------
— Mid-range Flows (P-0.27 in)
— Moist Conditions (P=0.47 in)
High Flow (P=0.78 in)
0 200 400 600 800 1,000 1,200 1,400 1,600
Minute
Figure 52. Unit-area (1 acre) runoff hydrographs representing the Parking Lot.
1.4
—Mid-range Flows (P=0.27 in)
— Moist Conditions (P=0.47 in)
High Flow(P=0.78in)
0 200 400 600 800 1,000 1,200 1,400 1,600
Minute
Figure 53. Unit-area (1 acre) runoff hydrographs representing the Drainage/Flood Control.
80
-------
—Mid-range Flows (P=0.27 in)
— Moist Conditions (P=0.47 in)
High Flow(P=0.78in)
200
400
600 800 1,000 1,200 1,400 1,600
Minute
Figure 54. Unit-area (1 acre) runoff hydrographs representing the Road.
1.4
1.2
1.0
I
3
0.6
TO
£
•c 0.4
0.2
0.0
— Mid-range Flows (P=0.27 in)
—Moist Conditions (P=0.47 in)
High Flow(P=0.78in)
0 200 400 600 800 1,000 1,200 1,400 1,600
Minute
Figure 55. Unit-area (1 acre) runoff hydrographs representing the Single Family Residential.
81
-------
I
cc
03
01
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
—Mid-range Flows (P=0.27 in)
— Moist Conditions (P=0.47 in)
High Flow(P=0.78in)
200 400 600 800 1,000 1,200 1,400 1,600
Minute
Figure 56. Unit-area (1 acre) runoff hydrographs representing
1 A
1 7
In
.u
1M
1 °-8
3
OC 06
TO U'b
V.
•;= r\ A
c u.^
z>
0-1
./
n n .
I
the Multi-Family Residential.
— Mid-range Flows (P=0.27 in)
— Moist Conditions (P=0.47 in)
High Flow(P=0.78in)
0 200 400 600 800 1,000 1,200 1,400 1,600
Minute
Figure 57. Unit-area (1 acre) runoff hydrographs representing the Industrial.
82
-------
—Mid-range Flows (P=0.27 in)
— Moist Conditions (P=0.47 in)
High Flow(P=0.78in)
cc.
03
01
200
400
600 800 1,000 1,200 1,400 1,600
Minute
Figure 58. Unit-area (1 acre) runoff hydrographs representing the Commercial.
1 ~)
l.Z
1 n
l.U
I
1 — p n s
1
s
1
4-»
07 _
n n
i
•—
— Mid-range Flows (P=0.27 in)
— Moist Conditions (P=0.47 in)
High Flow(P=0.78in)
0 200 400 600 800 1,000 1,200 1,400 1,600
Minute
Figure 59. Unit-area (1 acre) runoff hydrographs representing the Public-Institutional.
83
-------
€
!
O)
1.2
In
.U
Oc
0.4
O-i
./
n n
1
L
— Mid-range Flows (P=0.27 in)
— Moist Conditions (P=0.47 in)
High Flow(P=0.78in)
200 400 600 800 1,000 1,200 1,400 1,600
Minute
Figure 60. Unit-area (1 acre) runoff hydrographs representing the Transportation.
1.4
1.2
1.0
— Mid-range Flows (P=0.27 in)
—Moist Conditions (P=0.47 in)
High Flow(P=0.78in)
0 200 400 600 800 1,000 1,200 1,400 1,600
Minute
Figure 61. Unit-area (1 acre) runoff hydrographs representing the Vacant.
84
-------
1.4
1.2
1.0
!°8
cc
1
<
c 0.4
0.6
0.2
*•- in n
— Mid-range Flows (P=0.27 in)
— Moist Conditions (P-0.47 in)
High Flow(P=0.78 in)
«.,.,. — ££
0.0
0 200 400 600 800 1,000 1,200 1,400 1,600
Minute
Figure 62. Unit-area (1 acre) runoff hydrographs representing the Parks.
IT
./
1A
.u
1
ro,
! 0.6
f
•E 0.4
z>
0.2
n n
NoP
Gene
— Mid-range Flows (P=0.27 in)
— Moist Conditions (P-0.47 in)
High Flow(P=0.78 in)
tuno
?rate
kH
.u
0 200 400 600 800 1,000 1,200 1,400 1,600
Minute
Figure 63. Unit-area (1 acre) runoff hydrographs representing the Agricultural.
85
-------
Table 28. Summary of rainfall events at the Albuquerque International Airport (2000-2009)
Interval
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
Start
Time
1/1/004:00
1/31/006:00
2/22/00 7:00
3/7/00 9:00
3/21/004:00
3/28/00 1 1 :00
6/2/0016:00
6/28/00 15:00
7/11/00 17:00
7/21/00 16:00
8/17/0023:00
9/21/0021:00
10/4/00 15:00
10/21/00 11:00
10/28/000:00
11/3/006:00
11/23/00 1:00
12/26/00 8:00
1/27/01 10:00
2/26/01 2:00
3/7/01 14:00
4/5/01 18:00
4/27/01 15:00
5/13/01 5:00
5/19/01 8:00
6/29/01 21 :00
7/17/01 6:00
7/26/01 0:00
7/31/01 17:00
8/5/01 15:00
8/29/01 19:00
9/12/01 18:00
10/8/01 13:00
11/14/01 13:00
11/23/01 3:00
12/29/01 20:00
1/30/02 13:00
4/6/0217:00
6/22/02 16:00
7/7/02 5:00
7/17/02 14:00
8/2/02 20:00
8/19/02 15:00
9/7/02 23:00
9/18/02 12:00
10/23/02 13:00
11/4/026:00
11/9/0221:00
12/3/02 0:00
End
Time
1/4/0010:00
2/4/00 6:00
2/25/00 8:00
3/10/00 14:00
3/26/00 3:00
4/3/0017:00
6/5/0017:00
7/3/00 23:00
7/16/0023:00
7/24/00 17:00
8/21/0021:00
9/24/00 21 :00
10/16/004:00
10/27/0021:00
10/31/00 14:00
11/10/000:00
11/26/009:00
12/29/00 17:00
2/2/01 16:00
3/3/01 11:00
3/11/01 14:00
4/9/01 11:00
4/30/01 23:00
5/17/01 14:00
5/22/01 15:00
7/5/01 22:00
7/24/01 1:00
7/29/01 18:00
8/5/01 9:00
8/19/01 21:00
9/1/01 20:00
9/19/01 16:00
10/13/01 14:00
11/19/01 13:00
11/26/01 4:00
1/2/02 4:00
2/3/02 0:00
4/10/02 19:00
6/25/02 16:00
7/13/02 17:00
7/25/02 0:00
8/10/027:00
8/23/02 18:00
9/16/02 12:00
9/21/02 15:00
10/30/027:00
11/7/027:00
11/13/02 1:00
12/6/02 18:00
Total
Rainfall
(in.)
0.16
0.22
0.20
0.24
0.83
0.20
0.16
0.48
0.30
0.42
0.42
0.27
1.42
0.88
0.34
0.37
0.49
0.16
0.16
0.20
0.21
0.22
0.29
0.18
0.17
0.48
0.36
0.46
0.21
1.27
0.11
0.51
0.14
0.50
0.18
0.17
0.26
0.39
0.11
0.45
0.43
0.93
0.64
1.18
0.26
0.49
0.11
0.38
0.36
Wet
Count
(hr)
5
8
2
3
31
6
2
6
7
2
10
1
38
13
9
9
9
8
7
13
6
9
4
4
4
7
9
5
3
16
2
10
4
16
2
7
7
11
1
7
6
10
3
21
4
9
2
5
14
Peak
Intensity
(in./hr)
0.08
0.07
0.18
0.16
0.07
0.07
0.10
0.22
0.14
0.38
0.17
0.27
0.15
0.42
0.15
0.11
0.13
0.04
0.05
0.05
0.12
0.10
0.12
0.09
0.13
0.20
0.13
0.37
0.12
0.24
0.09
0.19
0.06
0.07
0.09
0.05
0.07
0.15
0.11
0.33
0.19
0.32
0.61
0.20
0.17
0.22
0.06
0.16
0.07
Average
Peak
(in./hr)
0.03
0.03
0.10
0.08
0.03
0.03
0.08
0.08
0.04
0.21
0.04
0.27
0.04
0.07
0.04
0.04
0.05
0.02
0.02
0.02
0.04
0.02
0.07
0.05
0.04
0.07
0.04
0.09
0.07
0.08
0.05
0.05
0.03
0.03
0.09
0.02
0.04
0.04
0.11
0.06
0.07
0.09
0.21
0.06
0.07
0.05
0.05
0.08
0.03
86
-------
Interval
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
Start
Time
2/8/0319:00
2/13/03 1:00
2/18/03 12:00
2/24/03 17:00
3/16/03 19:00
6/1/0314:00
7/20/03 17:00
8/9/03 0:00
8/24/03 20:00
8/28/03 17:00
9/7/0314:00
10/7/03 10:00
11/12/03 15:00
12/12/033:00
2/21/04 11:00
3/2/0413:00
4/2/04 8:00
4/8/04 0:00
6/29/04 1 :00
7/11/04 16:00
8/10/04 15:00
9/18/0421:00
10/5/044:00
10/11/044:00
10/17/04 14:00
10/26/0420:00
11/19/040:00
11/28/04 13:00
12/29/0421:00
1/2/0512:00
1/26/05 23:00
2/11/053:00
2/15/0521:00
3/5/05 9:00
3/13/0521:00
3/25/05 4:00
4/10/055:00
4/16/05 14:00
4/24/05 4:00
5/3/0515:00
7/17/05 18:00
7/22/05 14:00
7/28/05 17:00
8/6/0521:00
8/12/052:00
9/2/0519:00
9/28/05 4:00
10/9/055:00
10/15/05 11:00
4/29/06 2:00
6/7/06 22:00
End
Time
2/11/0323:00
2/16/03 16:00
2/23/03 12:00
3/3/0321:00
3/24/03 1 1 :00
6/4/0317:00
7/25/03 15:00
8/12/03 1:00
8/28/03 10:00
9/2/0318:00
9/13/03 18:00
10/13/0322:00
11/16/038:00
12/15/039:00
2/27/04 2:00
3/8/04 8:00
4/7/0415:00
4/13/04 19:00
7/2/0419:00
7/30/04 19:00
8/14/04 16:00
9/23/04 0:00
10/9/04 19:00
10/16/04 15:00
10/20/04 15:00
10/30/045:00
11/26/04 10:00
12/2/04 10:00
1/1/0523:00
1/7/0521:00
2/2/05 7:00
2/15/05 17:00
2/26/05 23:00
3/9/0512:00
3/18/05 1:00
3/29/05 15:00
4/13/05 19:00
4/19/05 17:00
4/28/05 20:00
5/6/05 20:00
7/20/05 18:00
7/25/05 14:00
7/31/05 17:00
8/10/050:00
8/18/05 16:00
9/14/05 13:00
10/2/05 17:00
10/13/0521:00
10/21/05 16:00
5/2/06 5:00
6/12/06 15:00
Total
Rainfall
(in.)
0.25
0.39
0.17
0.21
1.44
0.15
0.41
0.13
0.20
0.37
0.29
1.50
0.48
0.11
1.08
0.63
2.47
0.53
0.60
2.25
0.12
0.88
0.34
0.39
0.11
0.23
1.26
0.11
0.25
0.76
0.62
0.81
0.94
0.31
0.69
0.11
0.16
0.61
0.40
0.29
0.30
0.57
0.16
0.15
0.33
1.27
1.56
0.39
0.55
0.13
0.11
Wet
Count
(hr)
5
8
5
10
28
3
4
2
5
7
8
19
15
5
14
16
24
11
9
24
3
14
5
8
2
6
18
3
3
17
17
26
24
8
16
6
6
4
10
3
1
1
1
4
6
11
17
12
8
4
4
Peak
Intensity
(in./hr)
0.13
0.26
0.07
0.04
0.23
0.11
0.32
0.07
0.10
0.14
0.19
0.41
0.09
0.04
0.20
0.18
0.57
0.12
0.15
0.55
0.08
0.21
0.22
0.17
0.09
0.06
0.29
0.07
0.12
0.21
0.17
0.08
0.11
0.15
0.10
0.03
0.07
0.24
0.10
0.20
0.30
0.57
0.16
0.12
0.12
0.52
0.32
0.10
0.17
0.06
0.07
Average
Peak
(in./hr)
0.05
0.05
0.03
0.02
0.05
0.05
0.10
0.06
0.04
0.05
0.04
0.08
0.03
0.02
0.08
0.04
0.10
0.05
0.07
0.09
0.04
0.06
0.07
0.05
0.05
0.04
0.07
0.04
0.08
0.04
0.04
0.03
0.04
0.04
0.04
0.02
0.03
0.15
0.04
0.10
0.30
0.57
0.16
0.04
0.05
0.12
0.09
0.03
0.07
0.03
0.03
87
-------
Interval
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
Start
Time
6/26/06 19:00
7/6/06 0:00
7/18/066:00
7/24/06 19:00
8/5/06 3:00
9/1/068:00
9/6/0615:00
9/20/06 13:00
10/8/060:00
12/19/06 10:00
12/28/06 14:00
1/30/07 19:00
2/11/075:00
3/21/07 19:00
4/6/07 20:00
4/12/07 14:00
4/23/07 20:00
5/1/0713:00
5/6/07 1 1 :00
5/14/07 16:00
5/25/07 14:00
6/9/0716:00
7/4/0717:00
7/19/07 17:00
7/29/07 14:00
8/4/0719:00
8/23/07 19:00
9/17/07 15:00
10/1/07 19:00
11/22/0721:00
11/29/07 13:00
12/8/07 15:00
1/6/0812:00
1/27/08 20:00
2/21/08 3:00
4/9/0815:00
6/29/08 16:00
7/3/0816:00
7/26/08 14:00
8/3/0819:00
8/15/08 15:00
8/30/08 13:00
10/4/08 18:00
10/14/085:00
11/27/088:00
12/13/080:00
12/26/08 10:00
3/9/09 4:00
4/11/093:00
4/17/09 17:00
5/18/0920:00
End
Time
7/2/0618:00
7/11/06 18:00
7/21/0623:00
8/4/0621:00
8/27/06 22:00
9/4/06 22:00
9/12/06 16:00
9/23/06 21 :00
10/12/0621:00
12/23/06 1:00
1/2/0713:00
2/2/07 22:00
2/17/078:00
3/27/07 7:00
4/12/07 11:00
4/16/07 17:00
4/26/07 21 :00
5/5/0718:00
5/11/07 19:00
5/21/0721:00
5/28/07 14:00
6/14/07 19:00
7/7/07 17:00
7/27/07 14:00
8/3/07 22:00
8/10/07 17:00
9/1/0719:00
9/26/07 12:00
10/7/0721:00
11/26/07 17:00
12/4/074:00
12/14/07 15:00
1/10/08 16:00
2/1/080:00
2/26/08 4:00
4/12/08 16:00
7/2/0817:00
7/25/08 19:00
7/31/08 14:00
8/12/08 13:00
8/19/0823:00
9/3/0810:00
10/8/086:00
10/17/089:00
11/30/08 15:00
12/21/08 11:00
12/29/08 14:00
3/12/09 10:00
4/15/093:00
4/21/09 1:00
5/28/09 12:00
Total
Rainfall
(in.)
1.03
1.68
0.15
2.06
3.40
0.25
0.54
0.31
1.60
0.37
1.13
0.12
0.70
0.64
0.27
0.68
0.11
0.74
0.13
0.81
0.32
0.66
0.33
0.93
0.32
0.92
0.13
0.72
0.17
0.11
0.48
0.76
0.19
0.20
0.24
0.11
0.50
3.27
0.11
0.40
0.38
0.26
1.25
0.13
0.21
0.39
0.17
0.24
0.20
0.14
0.31
Wet
Count
(hr)
5
10
4
25
30
4
10
3
17
10
28
4
24
15
6
10
2
9
3
6
1
5
1
7
7
4
5
7
4
4
6
19
7
3
9
2
2
23
3
10
7
4
13
5
5
15
4
7
6
6
12
Peak
Intensity
(in./hr)
0.67
0.51
0.06
0.52
0.49
0.12
0.14
0.20
0.42
0.10
0.08
0.05
0.09
0.15
0.10
0.28
0.10
0.23
0.09
0.69
0.32
0.43
0.33
0.77
0.15
0.85
0.07
0.47
0.13
0.06
0.34
0.11
0.06
0.17
0.09
0.07
0.42
1.18
0.05
0.18
0.13
0.14
0.19
0.05
0.14
0.10
0.09
0.08
0.08
0.07
0.08
Average
Peak
(in./hr)
0.21
0.17
0.04
0.08
0.11
0.06
0.05
0.10
0.09
0.04
0.04
0.03
0.03
0.04
0.05
0.07
0.05
0.08
0.04
0.13
0.32
0.13
0.33
0.13
0.05
0.23
0.03
0.10
0.04
0.03
0.08
0.04
0.03
0.07
0.03
0.05
0.25
0.14
0.04
0.04
0.05
0.06
0.10
0.03
0.04
0.03
0.04
0.03
0.03
0.02
0.03
88
-------
Interval
152
153
154
155
156
157
158
159
160
161
162
Start
Time
6/10/09 1:00
6/24/09 15:00
7/3/09 4:00
7/21/09 18:00
8/13/09 19:00
8/22/09 21 :00
9/9/0917:00
9/16/090:00
10/7/094:00
10/20/09 14:00
10/28/094:00
End
Time
6/13/096:00
6/29/09 18:00
7/6/09 7:00
7/24/09 20:00
8/17/097:00
8/28/09 13:00
9/14/09 17:00
9/20/09 15:00
10/11/09 11:00
10/24/09 10:00
11/1/09 13:00
Total
Rainfall
(in.)
0.23
0.47
0.51
0.19
0.33
0.60
0.16
1.25
0.21
1.00
0.30
Wet
Count
(hr)
6
9
4
3
8
11
7
17
5
14
6
Peak
Intensity
(in./hr)
0.06
0.17
0.18
0.11
0.10
0.27
0.05
0.17
0.10
0.28
0.11
Average
Peak
(in./hr)
0.04
0.05
0.13
0.06
0.04
0.05
0.02
0.07
0.04
0.07
0.05
89
------- |