Despite a rapid push to digitize infrastructure data, there remains a dearth of readily available water distribution network (WDN) data to assess resilience challenges such as vulnerability assessment, proactive maintenance, and network reconfiguration. Synthetic models –realistic network models that imitate real-world networks’ appearance and behavior – have been rigorously developed and tested for infrastructures – namely electricity transmission and distribution – but have not yet become pervasive for water systems. We describe the challenges of developing synthetic WDNs using the case studies including Puerto Rico, New York City, Phoenix, Kentucky, and Atlanta. The synthetic WDN models are able to estimate network topography, pipe diameter, pump location and performance, and with EPANET hydraulic performance (see synthetic.resilientinfrastructure.org for additional background). The accuracy of synthetic WDN models is estimated using a locations where true data are available. While realistic networks and hydraulic performance can be synthesized with minimal data (i.e., roadway networks and locations of WTPs), the provision of basic data (i.e., pipe diameters, pump locations and performance, and tank locations) can significantly improve accuracy. The opportunities for synthetic WDNs are described both in characterizing vulnerable assets and network configurations to disrepair and hazards (e.g., climate extreme events) as well as power system outages that cascade to the WDNs.

Access to clean water is critical for drinking, hygiene, and to ensure a stable food supply, yet water insecurity continues to plague billions. Climate change is only expected to exacerbate water stress, leading to less freshwater in the cryosphere, increases in drought, and general water cycle instability. With conventional water supplies at risk, there is a need for innovative technologies capable of generating usable water from atypical sources. Desalination is one such technology already in use in many areas around the world. Seawater desalination via reverse osmosis is a leading desalination technique, but generally recovers up to only half of the inlet water. The rest forms a concentrated brine that requires disposal, most often via discharge into nearby water bodies. Increasingly, there is a push for the adoption of zero/minimal liquid discharge (ZLD/MLD) technologies that recover additional water from this brine while reducing the liquid volumes requiring disposal.

Understanding the implications of ZLD/MLD treatment trains requires a systems approach. Specifically, the analysis presented here consists of technoeconomic and life cycle analysis results of seven overarching treatment trains. These treatment trains consist of different combinations of pre-treatment, concentrator, crystallizer, and disposal technologies, resulting in 75 treatment train configurations. The levelized cost of water (LCOW) and specific energy consumption (SEC) are calculated across a range of potential water recoveries, with life cycle analysis of the ZLD/MLD processes informing emissions and water consumption estimates as well. Combined, the interplay of these metrics provides insight into the tradeoffs at the center of ZLD/MLD processes. The energy-water nexus in particular plays a critical role, with increasing water recovery coinciding with increasing energy demands and increasing energy demands resulting in increasing water consumption. Ultimately, ZLD/MLD treatment trains increase water recovery from desalination brines, but tighten the connections between energy and water systems. The feasibility of such systems will depend greatly on the severity of water insecurity in a location, access to infrastructure, and access to disposal options.

Climate change and sea level rise (SLR) are increasing the risk of tidal and storm flooding in coastal, urban communities. By 2100, sea level is expected to increase at least another three feet on the East Coast of the United States, particularly in Norfolk, VA. Developed watersheds face special challenges in resilience and habitat restoration, requiring deeper involvement with surrounding communities. With a focus on the Southside of Norfolk, this project will study Campostella and Campostella Heights as underserved communities (addressing to UN SDG 16: Peace, Justice, and Strong Institutions) that are vulnerable to rising flood risk. As a community-based research project, the objectives are: (1) To investigate the flood damage, vulnerability, and risk perception in the Southside Norfolk area, (2) To analyze the effect of compound flooding on the Campostella and Campostella Heights neighborhoods at both buildings and urban levels (addressing UN SDG 11: Sustainable Cities and Communities), (3) To propose nature-based solutions (at multiple scales) to improve resilience as well as wildlife habitats in the coastal community case studies (addressing UN SDG 14: Life Below Water)., and (4) To enhance public awareness by full community involvement in design process with a focus on adaptation before significant storm and flooding damage occurs. As for research methodology, a field investigation (observation, interview, and questionnaire) will be performed to collect data related to flooding in the Campostella and Campostella Heights neighborhoods. As a community-based project, civic league members and the residents of the community will be directly involved in data collection, identification of hot spots for recurrent nuisance flooding, evaluation project alternatives, and getting feedback about design solutions to reduce the exposure of the community to existing and future coastal hazards. In conclusion, all survey-based data will be summarized and integrated to develop community-scale adaptive design strategies, utilizing green infrastructure and nature-based solutions, to mitigate flooding and enhance urban resilience in this vulnerable area. Also, we will train the community members on potential interventions that may help in alleviating their flooding problems, while engaging them in planning, data collection related to flooding, and disseminating project results to the larger community to develop community support for the implementation of coastal resilience solutions.

Urban flood is a significant threat to cities worldwide. The rapid nature of flood water accumulation allows little evacuation time causing significant loss of life and damage to property. Issues like global warming, deforestation has been the reason for changing rainfall patterns questioning the resilience of existing flood infrastructure. Most of the existing urban flood models do not capture real time rainfall intensity, flood extents and depth of submergence making the of the model invalidate. Social media platforms (i.e. twitter) generate vast amount of data during a crisis or disaster situation. People tend to share information about on-ground-situation such as affected areas, water levels, experiences, inconveniences, or even immediate community responses. This rich yet underutilized resource has the potential to be used as a real time- user generated data source. The study presents an approach to hybridize social media data particularly tweets with existing flood mitigation measures for improved resilience and capacity building. Firstly, we collected tweets using particular keywords such as ‘flood’, “flooding’, ‘inundation’. The tweets were analyzed to focus on geo-tagged parts that provide specific location-based information. The data is then correlated to existing flood indices such as intense rainfall or response mechanisms like water level sensors and emergency response strategies for validation and acceptability. The study also evaluates the infrastructure in current urban setting to find areas that can be benefitted from the hybridization of social media data. Overall, the model offers a dynamic and more adaptive measure for disasters in terms of improved preparedness, response, and recovery. The framework is expected to enhance the decision-making process including emergency responses to people affected with flooding events. There might be some limitations such as processing the tweets even with sophisticated technologies may not ensure enough accuracy specially when the user profile location and tweet location does not agree. Nevertheless, the research holds significance for urban planners, environmental scientists, hydrologists, policymakers etc. offering a novel approach to integrate digital social media data for an improved model that can be adopted in cities worldwide.

Urbanization and socioeconomic development are occurring rapidly across Sub-Saharan Africa, where the impacts of rapid urbanization compound more frequent and intense extreme weather events driven by climate change. Hazards such as urban flooding are affected by where urban development occurs, increased impervious surface cover, and the construction of inadequate stormwater infrastructure to handle present and future precipitation events. The compound hazards of rapid urbanization and flooding are particularly evident in cities like Kigali, Rwanda, where highly urbanized sub-catchments increase runoff, resulting in frequent and destructive floods. Improving urban resilience and sustainability requires policymakers to develop climate-resilient stormwater systems to better manage urban floods, protect infrastructure, and preserve water resources. However, urban planning and flood management in rapidly growing urban areas face challenges due to a scarcity of data and limited institutional capacity to establish and maintain monitoring networks and integrate data into the decision-making process. Building footprints and data on historic formal and informal urban development are critical to urban stormwater planning, capturing rapid population changes, and describing who is at risk from extreme precipitation and flooding events. Furthermore, building footprints describe high-resolution changes in impervious surface area, which alters runoff and urban flood extents.

Our work aims to develop local-level time-series building footprint datasets for major urban areas in East Africa from 2005-present by leveraging historical high-resolution satellite imagery, which provides access to timely, consistent, multi-decadal data on urban areas globally. Machine learning models like deep convolutional neural networks can extract building footprints from this high-resolution satellite imagery. However, existing building footprint extraction models are predominately trained and validated on ground truth data from high-income countries, where urban areas look and evolve differently than in low-middle income countries. Therefore, these models systematically misclassify and undercount building footprints in dense, often informal, urban settlements whose populations are vulnerable to urban flood hazards. Our work aims to improve the quality of multi-temporal building footprint datasets in developing urban areas without additional, costly ground truth data by pairing existing neural network models for building footprint extraction with a Hidden Markov Model (HMM). A HMM enforces physically consistent probabilistic time-series relationships between extracted building footprints through transition probabilities (how likely a building will be constructed, torn down, or remain the same each period) and emission probabilities (how accurately the neural network model extracts building footprints from a satellite image). The combined model harnesses every year of satellite imagery to improve the consistency and accuracy of past and future building footprints. We present a set of experiments and initial results demonstrating the benefits of the paired method. Furthermore, we compare footprint corrections from HMMs calibrated with modeled data to HMMs calibrated with ground-truth data, highlighting both the potential and associated uncertainty for improving time-series building footprint data without additional ground-truth data collection. We will present the initial results of a comparative study of the spatiotemporal urban development and implications for climate change risk and adaptation opportunities of urban areas in East Africa.