Paper: 'Diverse Paradigms of Residential Development Inform Water Use and Drought-related Conservation Behavior'

I just caught this recent (18 November 2020) open-access paper by Kimberly J. Quesnel, Saahil Agrawal and Newsha K Ajami: 'Diverse paradigms of residential development inform water use and drought-related conservation behavior', Environmental Research Letters 15(12) (2020) 124009.

Read the article online by clicking here. The links below may not take you to the reference but to the original Quesnel et al. paper.

Download Quesnel_2020_Environ._Res._Lett._15_124009

Download ERL_15_12_124009_suppdata

Abstract
Widespread urbanization has led to diverse patterns of residential development, which are linked to different resource consumption patterns, including water demand. Classifying neighborhoods based on urban form and sociodemographic features can provide an avenue for understanding community water use behaviors associated with housing alternatives and different residential populations. In this study, we leveraged built environment data from the online real estate aggregator Zillow to develop neighborhood typologies and community clusters via a sequence of unsupervised learning methods. Five distinct clusters, spatially segregated despite no geospatial inputs, were associated with unique single-family residential water use and conservation patterns and trends. The two highest-income clusters had divergent behavior, especially during and after a historic drought, thus unraveling conventional income–water use and income–water conservation relationships. These clustering results highlight evolving water use regimes as traditional patterns of development are replaced with compact, water-efficient urban form. Defining communities based on built environment and sociodemographic characteristics, instead of sociodemographic features alone, led to 3% to 30% improvements in cluster water use and conservation cohesion. These analyses demonstrate the importance of smart development across rapidly urbanizing areas in water-scarce regions across the globe.

Introduction
Today, 55% of the world's population lives in urban areas with the percentage expected to grow to 68% by 2050 (United Nations Department of Economic and Social Affairs Population Division 2019). As residents move away from rural areas and into cities, suburbs, and towns, many different patterns of residential development and urban form are emerging, in turn affecting urban system sustainability (Jabareen 2006, Pandit et al 2017, Pickett and Zhou 2015). Variations in population density, housing structures, and neighborhood configurations are linked to different energy use patterns, transportation behaviors, and public health outcomes (Berrigan and Troiano 2002, Ewing and Rong 2008, Güneralp et al 2017, Stokes and Seto 2019). Housing features including urban versus suburban communities (Breyer and Chang 2014), infill development (Sanchez et al 2020), changing building and landscaping codes (Brelsford and Abbott 2017, Garcia and Islam 2019), and nontraditional housing arrangements (Barnett et al 2019) can also lead to heterogeneous residential water use behavior, which has direct implications for water resources management.

Simultaneously, sociodemographic characteristics like income and education levels often explain variations in residential water use (Schleich and Hillenbrand 2009, Shandas and Parandvash 2010, House-Peters and Chang 2011, Brelsford and Abbott 2017, Quesnel and Ajami 2017, Fan et al 2017). Together, built environment and social features provide the building blocks for defining neighborhoods and communities for urban water demand assessment and planning (Jackson-Smith et al 2016, Stoker et al 2019), which in turn dictate water supply and infrastructure investment decisions at city to neighborhood scales (House-Peters and Chang 2011, Stoker and Rothfeder 2014, Gurung et al 2016). These characteristics are also critical for understanding water conservation behavior during drought (Fielding et al 2012, Polebitski and Palmer 2013, Mini et al 2015), when strategic resource management is particularly critical.

Grouping residential customers for water resources planning and management requires linking bottom–up, household-level data about the built environment with top–down Census block-group or tract scale features. Built environment and social data can also be acquired through customer surveys (Randolph and Troy 2008, Harlan et al 2009, Willis et al 2013, Hannibal et al 2018), but these can be expensive, time-intensive, and limited in scope. Some researchers have successfully accessed assessor records on homes and parcels (for example: Chang et al 2017, Brelsford and Abbott 2017), but this data is generally challenging to obtain in bulk, digitized formats and must be acquired on a city-by-city or county-by-county basis. Around the US, a few agencies have moved to publicly available digital records like New York's Primary Land Use Tax Lot Output (PLUTO) database, which can be used for building-level water demand studies (Kontokosta and Jain 2015), although these are not yet common.

New websites like Zillow, Redfin, or Trulia that aggregate and digitize records from multiple sources, including public agencies, offer the possibility to overcome this historic obstacle, but these sources have yet to be leveraged as a tool for analyzing urban water use. Together with Census information, this housing data can be used as an alternative to traditional sources to develop a holistic depiction of communities within a city, a spatial-scale which provides the granularity of within-city information while being more practical than customer-level analyses. This research aims to demonstrate the value of combining data on single-family residential housing features and urban form from Zillow with Census data to identify residential community groupings. In turn, these classifications can help water resource decision-makers better understand their customers' water use behavior, design optimal conservation policies, and plan for future resources needs and allocation.

We performed our analysis within a single utility—exploring high-resolution data within a small area to gain insights into behavior not possible at more aggregated scales while discovering information that can be applied in a broader context. Our study area, the City of Redwood City (Redwood City) is situated on the San Francisco Bay peninsula in California and represents a microcosm of diversity—in 2017, block-group level median household income within the service area ranged from less than $40 000 (including several block-groups classified as Disadvantaged Communities) to over $220 000. The San Francisco Bay Area represents a region with varied and evolving water supply and demand regimes, making it a particularly valuable place to study urban water with lessons for other growing, semi-arid and arid regions across the Western U.S. and the world (Gonzales and Ajami 2017b). In our study, we focused specifically on single-family residential water use, which accounts for 65% of California's urban water use (California Department of Water Resources 2016) and about half of Redwood City's potable use (City of Redwood City 2015). Redwood City demand is seasonal, with higher use in the summer due to landscape irrigation.

Our research takes place over a 10-year period from 2008–2017 which included two     historic droughts and an economic recession. In particular, the 2012–2016 drought was one of the most severe in California's history (U.S. Geological Survey: California Water Science Center 2018). The drought was not only exceptional hydrologically, but also in terms of state and local political actions, public awareness, and news media coverage that led to high drought saliency and has been associated with high conservation rates (Quesnel and Ajami 2017, Gonzales and Ajami 2017a, Bolorinos et al 2020). This historic drought provides an important setting for examining not only the drivers of water conservation, but also rebound once mandatory restrictions were lifted and the drought was declared over (Gonzales and Ajami 2017a, Bolorinos et al 2020). Evaluating water conservation during the 2012–2016 drought, an extreme event more likely to occur in the future, provides an opportunity to evaluate changing residential water use behavior across customers under escalating climatic and policy regimes.

Cutting to the chase...

Discussion and conclusion
This research demonstrates the coupled role of urban form and sociodemographic characteristics in water use and drought-related conservation behavior. A crucial finding of this study is the importance of the built environment when classifying urban water customers and creating residential neighborhood typologies useful for water use and conservation analyses and developing demand management strategies. Traditionally, many researchers have stratified customers by income, finding that affluence is associated with high water use, high levels of drought-related conservation, and subsequently high conservation backslide post-drought. These water use characteristics are typically linked to more wealthy customers having bigger lots and therefore using more water outdoors. However, this study proves that income does not always correlate to high water use, and that affluent residents choosing to live in dense, new developments have very different water use patterns and water use behavior.

We bypassed previous data constraints faced by researchers, utilities, municipalities and other public agencies trying to better understand how built environment factors are linked to water use by utilizing data from an online real estate aggregator. By obtaining customer-level information from the Zillow ZTRAX database, this research for the first time demonstrates the possibility for this kind of platform to serve as a new tool for incorporating housing features into water use analyses. City and county assessors are legally required to fulfill public record requests and provide data in response to inquiries, but acquiring these individual datasets for comparative regional, state, or national assessments can be prohibitive. Thus, using aggregated data generated by Zillow and similar websites can provide a gateway for widespread analyses, for example opening the door for multi-city or even multi-state studies. These new websites offer a way to access built environment data uniformly in one place instead of filing separate records request for each service area of interest. While this research presents an alternative to traditional methods for obtaining information about the built environment, water use data remains sparse and challenging to acquire (Chini and Stillwell 2016, Josset et al 2019).

As this is the first research to use Zillow in water demand research, there are a host of avenues for future research. For example, our study was set within one utility, but the potential for Zillow truly lies in its ability to cross administrative borders. For example, where cross-city and multi-city comparisons in different counties previously required obtaining data from each individual jurisdiction, Zillow or other online aggregators store this data in one central location. Having one central database also provides the benefit of data consistency.

Additionally, here we focused primarily on neighborhood clustering to inform residential water use modeling at the monthly scale, but future studies could explore how emerging databases, online platforms, and data aggregators with high spatial and/or temporal resolution can be coupled with higher temporal resolution water use data from smart meters to better understand and predict water use and conservation. These insights could then be used to develop short-term and long-term demand management strategies. Another avenue for future research would be to investigate how different emerging urban form paradigms including infill development, peri-urban, and suburban designs shift both short- and long-term water use behavior. Land managers and water managers often do not coordinate or even talk to each other (Gober et al 2013), but with these new developments and unprecedented urbanization, there is increasing pressure for integrated planning and management.

This research lays the framework for future big data-driven urban water research and provides evidence for the implications evolving urban form on water use. Our results also point out future paths of more water-efficient urban development. Cities all over the world are expanding at rapid rates, and there are many different ways for them to develop. Here, we showed that dense housing patterns and new houses, regardless of the size, can result in lower water use than traditional sprawl. However, these low water use communities also have lower conservation rates and sometimes faster post-drought water demand rebound, and actions must be taken to account for this reduced water system flexibility. City planners and water managers must work together to develop cities of the future that house an increasingly large portion of the population and meet the wide-range of sustainability-oriented goals critical to addressing 21st century urban challenges.

Good stuff!

Enjoy!

"You can't teach an old dogma new tricks." - Dorothy Parker (quoted in @TheMVTimes via @TheWeek)