WRF Webcast Water Use in the Multi-family Housing Sector February 1, 2018 No part of this presentation may be copied, reproduced, or otherwise utilized without permission.
Webcast Agenda Introductions & background Maureen Hodgins 5 min Presentation Jack Kiefer 45 min Question & Answers 10 min Survey 4 questions for audience
WRF Research Demand Forecasting & Management 2009-2017 Goal Help improve utility plans Drivers changing water use trends, population changes, limited supplies, and/or increasing costs $3.5 M 16 projects, 11 published, 5 ongoing
Check out later Published CII, 4375, 2015 Residential, 4309, 2016 Multi-family, 4554, 2018 To be published 2018 Demand patterns f/ sizing meters & service Lines, 4689, est 2018 Commercial, 4619, est 2018 Customer water use Factors impacting demand Published Climate change, 4263, 2013 Recession, 4458, 2016 Passive efficiency, 4495, est 2018 Irrigation controllers, 4227, 2016 To be published Urban landscape research needs, 4633, est 2018 Published Uncertainty & long term forecasts, 4558, 2016 Uncertainty in forecasts & planning Forecasting approaches Published Customer data, 4527, 2016 Short term, 4501, 2017 To be published Long term, 4667, est 2020 Probability mgmt, 4742, est 2020
Summary of demand forecasting & management research
Downloading the 4554 report waterrf.org search 4554 download PDF
FYI
Input your webinar questions here Q&A at end of webinar
Download presentation Slides and recording will be available to WRF subscribers WITHIN 24 hours after the webcast
Water Use in the Multi-family Housing Sector Jack C. Kiefer, Ph.D. Lisa R. Krentz No part of this presentation may be copied, reproduced, or otherwise utilized without permission.
Presentation Overview Background on WRF 4554 Data sources and collection methods Metrics and comparisons Examples and results of modeling variability in water use Summary and conclusions Recommendations
The Multi-family Housing Sector About 25% percent of housing (or about 33 million residences) in the U.S Share of multi-family dwellings increasing in some areas Multi-family housing dominant residential sector in some denser urban areas Many areas plan to direct future development or densify
Water Research Foundation Project 4554 Water Use in the Multi-family Housing Sector Objectives Narrow knowledge gaps Develop, demonstrate, and recommend analytical strategies for: Estimating multi-family water use Categorizing multi-family properties Forecasting water use for water use categories
Defining what is Multi-family Everything other than traditional single-family detached homes Any residential property w/2+ units Master-metered residential properties General practice to lump MF into general residential or commercial customer classes detached, semidetached, row house, or multi-family structures with 5 or more units. Ownership/Tenure Rental Apartments Duplex Multiplex Individually Owned Condominiums Townhouses Jointly Owned Cooperatives
What multi-family looks like
What multi-family looks like
What multi-family looks like
What multi-family looks like
What multi-family looks like
What multi-family looks like
Selected Research Questions To what extent does multi-family water use differ from single-family use? How does water use differ among subclasses of the multi-family sector? What factors influence water use in the multifamily sector and major sub-classes? Does greater development density lead to less use? What are the effects of given property features?
Data Collection Water use for multi-family properties or classes Number of dwelling units (scale measure) Sub-classifications Property characteristics Other potentially influential variables Price Income Climatic Secondary information only
Utility Partners Denver Water New York City Department of Environmental Protection Phoenix Water Services San Diego County Water Authority Tampa Bay Water
Other Key Data Sources U.S. Census American Housing Survey American Community Survey New York University Furman Center Fannie Mae
What the US Census tells us Increase in share of population living in multiple unit structures Rental tenure dominates Lower incomes Smaller households Younger householders Proportionally fewer (in unit) clothes washers and dish washers (especially renter-occupied)
30 US Counties with Largest Multi-family Housing Share (2015) County Estimated Population in Multiple Unit Structures Percent of Total Population New York County, New York 1,532,721 96.70% Bronx County, New York 1,230,537 87.00% Kings County, New York 2,152,291 82.90% Hudson County, New Jersey 537,363 80.50% Suffolk County, Massachusetts 543,604 74.60% Queens County, New York 1,565,414 67.70% Alexandria city, Virginia 92,090 60.60% San Francisco County, California 469,879 55.60% Essex County, New Jersey 426,869 55.00% Arlington County, Virginia 121,959 54.00% District of Columbia, District of Columbia 319,136 50.50% Passaic County, New Jersey 249,609 49.70% Providence County, Rhode Island 274,416 45.20% Cook County, Illinois 2,312,233 44.90% Westchester County, New York 417,411 43.90% St. Louis city, Missouri 131,575 43.10% Milwaukee County, Wisconsin 372,086 39.80% Union County, New Jersey 217,209 39.60% Albany County, New York 108,305 37.10% Cass County, North Dakota 61,618 37.00% Middlesex County, Massachusetts 564,325 37.00% Fulton County, Georgia 352,530 36.10% Essex County, Massachusetts 273,479 36.10% Richmond city, Virginia 75,409 36.10% Orleans Parish, Louisiana 135,473 35.80% Grand Forks County, North Dakota 23,343 35.40% Denver County, Colorado 234,955 35.20% Los Angeles County, California 3,509,565 35.10% Hampden County, Massachusetts 158,374 34.70% Bristol County, Massachusetts 187,627 34.70%
Selected High Population Counties with Largest Gains in Multi-family (2005-2015) County 2015 Population Change in Population 2005-2015 Change Single Unit 2005-2015 Change Multiple Unit 2005-2015 Gain in Multiple Unit Share (Percentage Points) San Diego County, California 3,213,597 389,338 146,349 237,461 4.1 Queens County, New York 2,310,954 95,615-36,104 128,633 2.9 Cook County, Illinois 5,147,733-58,624-134,641 80,800 2.1 Santa Clara County, California 1,881,601 211,711 120,042 94,484 2.1 Kings County, New York 2,597,054 151,038-27,162 174,025 2.0 Bexar County, Texas 1,868,418 385,820 269,400 107,842 2.0 Hillsborough County, Florida 1,328,563 216,846 153,129 72,462 1.8 King County, Washington 2,081,072 325,254 195,100 125,130 1.7 Alameda County, California 1,607,666 186,358 106,420 77,970 1.6 Suffolk County, New York 1,469,390 24,748-313 23,868 1.5 Orange County, California 3,125,316 180,779 87,164 97,767 1.4 Middlesex County, Massachusetts 1,526,878 121,367 57,803 60,809 1.1 Nassau County, New York 1,341,365 31,289 14,298 18,607 1.1 Orange County, Florida 1,255,930 253,081 166,064 80,651 1.0
Top 30 Counties with Largest Gains in Multi-family (2005-2015) County 2015 Population Gain in Multiple Unit Share (Percentage Points) Arlington County, Virginia 226,054 8.3 Monongalia County, West Virginia 98,487 7.8 Natrona County, Wyoming 80,557 7.4 Harrison County, Mississippi 196,586 7.3 Suffolk city, Virginia 87,689 7.0 Cattaraugus County, New York 75,204 6.9 Leon County, Florida 272,694 6.8 Polk County, Oregon 77,626 6.8 Grand Forks County, North Dakota 66,007 6.7 Vigo County, Indiana 99,184 6.5 Orleans Parish, Louisiana 377,979 6.3 Craighead County, Arkansas 100,642 6.3 Jefferson County, Texas 238,066 6.2 Bradley County, Tennessee 101,200 6.1 Cass County, North Dakota 166,478 6.0 Cumberland County, North Carolina 313,216 6.0 Rockland County, New York 318,982 5.9 Muskingum County, Ohio 84,292 5.8 Dorchester County, South Carolina 150,535 5.7 Union County, New Jersey 548,291 5.7 Forsyth County, North Carolina 358,862 5.7 Houston County, Georgia 148,343 5.6 Portsmouth city, Virginia 92,948 5.5 Erie County, Ohio 74,194 5.5 Adams County, Illinois 65,406 5.5 Elmore County, Alabama 77,387 5.4 Lane County, Oregon 354,895 5.3 Bartholomew County, Indiana 80,243 5.2 Monroe County, Florida 75,588 5.2 Armstrong County, Pennsylvania 66,303 5.1
Comparison of Unit Usage Rates (gallons per unit per day) Utility Multi-family Sector (Composite) Units Mean (Grand) Median Denver (2014) 192,560 133 127 New York City (2014) 2,143,108* 170 137 Phoenix (2014) 181,101 182 158 San Diego County (2012) Tampa Bay Water (2014) 423,788 164 n/a 280,865 117 97 *Excludes properties in One-Family Dwelling or Mixed Residential/Commercial Land Use Tax Classes Relatively weak association with climate Distributions skewed to the right
Comparison of Unit Usage Rates (gallons per unit per day) Utility Multi-family Sector (Composite) Units Mean (Grand) Single-Family Sector Median Units Mean Median Denver (2014) 192,560 133 127 202,367 271 n/a New York City (2014) 2,143,108* 170 137 315,246 202 170 Phoenix (2014) 181,101 182 158 305,341 331 242 San Diego County (2012) Tampa Bay Water (2014) MF<SF in all cases 423,788 164 n/a 670,692 308 n/a 280,865 117 97 404,903 191 150 *Excludes properties designated as One-Family Dwelling or Mixed Residential/Commercial Stronger association with climate in SF sector
Annual Precipitation and Average Unit Use Single-Family
Annual Precipitation and Average Unit Use
Gallons per Capita per Day (GPCD) Estimated Average Annual Water Use Per Capita for Five Water Systems (gallons per capita per day) Single-Family Multifamily 120 100 100 111 104 83 80 60 73 61 68 71 70 56 40 20 0 Denver (2014) New York City (2014) Phoenix (2014) San Diego County (2012) Tampa Bay Water (2014)
Gallons per Capita per Day (GPCD) Estimated Average "Min-Month" Water Use Per Capita for Five Water Systems (gallons per capita per day) Single-Family Multifamily 120 100 80 60 49 54 55 66 82 70 77 64 60 55 40 20 0 Denver (2014) New York City (2014) Phoenix (2014) San Diego County (2012) Tampa Bay Water (2014)
Evaluation of Sub-Classification Do average usage patterns differ significantly based on definitional groupings? Different utilities define multi-family sector differently Different utilities sub-classify multi-family sector differently Analysis depends on use of external and linkable data Land use codes Property use codes Building types
Sub-classification Example-Tampa Bay Water
Sub-classification Example-Tampa Bay Water
Sub-classification Example-Tampa Bay Water
New York City Classifications Multi-family Two or Three Family Dwellings Multi-family Buildings Mixed Residential & Commercial Two Family Dwellings 4 additional subclasses Condominiums 8 additional subclasses Elevator Apartments 10 additional subclasses Walk-up Apartments 10 additional subclasses Residence Multiple Use 7 additional subclasses
Gallons per Unit per Day (GPUD) 250 New York City Mean 2014 Unit Use by Residential Land Use Tax Classes (GPUD) 200 202 201 171 167 150 100 50 0 One Family Dwellings Mixed Residential & Commercial Buildings Multi-Family Buildings Two-Three Family Dwelling
Gallons per Unit per Day (GPUD) 176 New York City Mean 2014 Unit Use for Building Classes within Multi-family Buildings" Land Use Tax Class (GPUD) 174 174 172 172 170 168 166 166 164 162 ELEVATOR APARTMENTS CONDOMINIUMS WALK UP APARTMENTS Multifamily Buildings Class Average
Gallons per Unit per Day (GPUD) New York City Mean 2014 Unit Use for Building Sub-Classes within "Elevator Apartments" Building and Multi-family Buildings" Land Use Tax Class (GPUD) 250 226 200 200 185 181 150 161 150 139 120 100 50 0 Fireproof (Standard Construction Without Stores) Co-op Conversion From Loft/Warehouse Miscellaneous Semi-fireproof (Without Stores) Converted Cooperatives (Other Than Condominiums) Artists in Residence Luxury Type Multifamily Buildings - Elevator Apartments Subclass Average
Gallons per Unit per Day (GPUD) 250 Phoenix 2014 Multi-family Annual Average Unit Usage Rates by Major Subclass (gallons per unit per day) 200 180 206 150 120 100 50 0 Apartment Condominium Multiplex Multifamily Class Average
Median Unit Use and Housing Density (Phoenix)
Seasonal Peaking and Housing Density (Phoenix)
Median Unit Use and Housing Density (Tampa Bay Water)
Median Unit Use and Housing Density (New York City)
Assessment of Multi-family Demand Determinants Regional water use and socioeconomic databases for San Diego County and Tampa Bay Water NYU Furman Center housing and demographic database Property-level water use and survey data from 2012 Fannie Mae Multi-family Market Research Energy and Water Survey Utility-provided databases with selected property attribute and geographic assignments Tampa Bay Water Phoenix New York City Denver
Multi-family Class-level Forecast Model (San Diego County) Balanced panel model of water use, socioeconomics, weather, and climate 22 agencies, 120 months each Variable Estimated Multifamily Elasticity Marginal Price for Water (inflation-adjusted) -0.14 Median Household Income (inflationadjusted) +0.07 Housing Density (housing units per acre) -0.30 Household Size +0.56
Multi-family Class-level Forecast Model (San Diego County) Balanced panel model of water use, socioeconomics, weather, and climate 22 agencies, 120 months each Variable Estimated Multifamily Elasticity Estimated Single- Family Elasticity Marginal Price for Water (inflation-adjusted) -0.14 < -0.23 Median Household Income (inflationadjusted) +0.07 < +0.54 Housing Density (housing units per acre) -0.30-0.31 Household Size +0.56 +0.44 <
Factor Analysis of NYC Community Districts Socioeconomic data from NYU Furman Center available for 57 metrics across 59 Community Districts (64 metrics total) 1. Aggregate premise level water use data to community district level 2. Condense demographic metrics into fewer thematic factors 3. Evaluate median water use per dwelling unit with respect to factor scores
Factor Analysis of NYC Community Districts Thematic Factor Affluence Poverty Property Value/Cooling Towers Severe Crowding Estimated Effect of 1 unit change in factor score -18 gpud +23 gpud +8 gpud +15 gpud
Property-level Assessments (Fannie Mae Survey) Sample of 955 multi-family properties for 2012 categorized by region
Property-level Assessments (Fannie Mae Survey) Regression analysis of cross-sectional data on multi-family property features, accounting for region (n=323) Variable Estimated Elasticity or Effect Average Cost of Water -0.26 Presence of Pool +10% Tenant Pays for Water -17% Level + Mechanism Property Receives Govt Subsidy -12% Property Built >=2001-16% Senior Living Facility -18%
Data Available from Utility Partners for Property-level Assessments System Units Lot Size Year Built Assessed Value Presence of Pool(s) Presence of Reclaimed Water Presence of Cooling Tower(s) Denver X X (pervious area) X New York City X X X X X Phoenix X X X Tampa Bay Water X (Apt only X X X X X X X (Apt only)
System Denver (2015) Tampa Bay Water (2010-2014 average) Phoenix (2010-2014 average) New York City (2014) Class Density (Units/Acre) Elasticity Assessed Value Elasticity Effect of Pool(s) Effect of Reclaimed Water Effect of Cooling Tower(s) Age Profile Multi-family Total -0.18 +-+- Multi-family Total -0.14 0.10 +9.4% -20.5% +-+- Condo -0.14 0.09 +7.8% -18.2% +-+ Townhouse -0.06 0.21 +13.0% -18.4% +- Less than 10 Units -0.05 0.02 +20.5% -17.4% +-+- 10 or More Units -0.21-0.05 +9.9% -4.7% Access to alternative supplies has significant Pools increase effect on TBW water use All density estimates Effects of income demands negative and proxy are mixed statistically significant Multi-family Total -0.44 +-+ Apartment -0.45 +20.3% +28.7% +-+ Condo -0.54 +- Impact of cooling towers proportionally larger in the desert Multiplex -0.23 +-+- Condominiums -0.14 0.18 +18.4% +-+- Elevator Apartments -0.20-0.04 +15.7% +-+- Walk-up Apartments -0.08-0.01 +5.8% -+-+
Summary & Conclusions Water use per unit in the Multi-family sector is generally lower than water use per unit in the Single-family sector The gap between SF and MF unit usage rates narrows when accounting for household size and seasonality For a given climate, generally lower seasonal use in the MF sector relative to SF
Summary & Conclusions Development density is statistically important More units per acre (i.e., higher unit density) - lower unit usage rates Results are consistent with Notion of shared outdoor (and other) uses Less irrigated area
Summary & Conclusions Two main effects from densification Between class: MF is denser than SF Within class: more dense MF, generally less water use (per dwelling unit)
Summary & Conclusions Water use in the MF sector is also influenced by Climate and weather Property features (water end uses, age) Socioeconomics Price The estimated effects of water use determinants tend to vary by geographic area Underlying climate Sector/subclass structure
Summary & Conclusions The ability to obtain information on housing units is essential for: Accounting for scale differences Calculating densities Joins to tax appraiser data typically permits estimation of units and additional classification options Sub-classification of the MF sector matters Can affect sample statistics and modeling relationships Can t easily infer reasons for differences based on segmentation alone
Recommendations Patterns of urban development and re-development is a key area of future uncertainty Keep a watchful eye on housing trends Recognize potential impacts of densification and other development policies Investigate the underlying structure of the overall multifamily sector Test for statistical associations with key variables for the purposes of forecasting and profiling Seek more uniformity and consistency in classifying and sub-classifying multi-family Better metrics More robust comparisons
Q&A No part of this presentation may be copied, reproduced, or otherwise utilized without permission.
Thank you Comments or questions, please contact: mhodgins@waterrf.org jkiefer@hazenandsawyer.com For more information visit: www.waterrf.org No part of this presentation may be copied, reproduced, or otherwise utilized without permission.
Extra Support Slides No part of this presentation may be copied, reproduced, or otherwise utilized without permission.
SF Metric Value/MF Metric Value Ratio of Single-Family to Multifamily Usage Rate Estimates for Five Water Systems Ratio Annual Use per Unit Ratio Annual Use per Capita Ratio "Min-Month" Use per Capita 2.5 2.0 2 1.8 1.9 1.6 1.5 1.4 1.2 1.3 1.5 1.2 1.2 1.2 1.1 1 0.9 0.9 0.8 0.5 0 Denver (2014) New York City (2014) Phoenix (2014) San Diego County (2012) Tampa Bay Water (2014)
Denver (2011-2015) Average Seasonal Patterns Avg Peak Avg Trough = 1.89 San Diego County (2000-2012) Phoenix (2002-2014) Avg Peak Avg Trough = 1.23 Avg Peak Avg Trough = 1.48
New York City (2011-AMR Sample) Average Seasonal Patterns Avg Peak Avg Trough = 1.10 Tampa Bay Water (2002-2014) Unique Seasonal Pattern Spring is dry season Snowbirds /Spring Break Avg Peak Avg Trough = 1.06
Estimated Socioeconomic Effects for Tampa Bay Water Project data supporting Tampa Bay Water s forecast model re-development Multi-family Class Price Elasticity Income Elasticity All Multi-family -0.13 +0.29 Condo -0.20 +0.32 Townhouse -0.41 +0.88 Less than 10 Units -0.01 +0.29 10 or more Units -0.14 +0.31 Mobile Home (single unit) Mobile Home (multiple unit/parks) -0.01 +0.34-0.16 +0.30 Single-Family -0.37 to -0.73 +0.37 to +0.45
2015 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 1992 1991 1990 1989 1988 1987 1986 1985 1984 1983 1982 1981 1980 1979 1978 1977 1976 1975 1974 1973 1972 1971 1970 Calculated Value 25 Plot of Building Age Portion of Regression Equation = Exp(1.9247 + 0.0789*AGE - 0.0021*AGE^2 + 1.9756e-05*AGE^3-6.3819e-08*AGE^4) Denver Age profile: +-+- 20 15 10 Accounting for number of units and pervious area, housing stock built in 1986 would be simulated as using the most water per premise in 2015 5 0 Older Implied Calendar Year
2015 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 1992 1991 1990 1989 1988 1987 1986 1985 1984 1983 1982 1981 1980 1979 1978 1977 1976 1975 1974 1973 1972 1971 1970 Calculated Value Plot of Building Age Portion of Regression Equation 25 20 15 10 Annual March September 5 0 Older Implied Calendar Year