Eric W. Sanderson
Wildlife Conservation Society, 2300 Southern Blvd., Bronx NY 10460
19 January 2014
© Wildlife Conservation Society, 2014
Please submit comments or suggestions to m2409@themannahattaproject.org
Water flow calculations are based on a simple, flow-through storm event model, based on methods adapted from Vörösmarty et al. (1989), Vörösmarty et al. (1996) and Mitchell et al. (2001). The water flow calculations depend on the ecosystems, lifestyle, and climate scenarios selected by the user and are calculated across the entire vision extent. Unlike the other models (i.e. geography, population, carbon and biodiversity), which simulate annual estimates, the water model is based on storm events that occur on a single day in June. Therefore all the precipitation falls as a rain within a single day, potential evapotranspiration is fixed by the climatic conditions on that June day, and piped water inputs are for a single day’s consumption (though that day is based on average annual consumption rates pro-rated for a day in June, not necessarily June-specific daily consumption). Future iterations of the model may include longer time series of water flow estimates that incorporate diurnal variation in these parameters, leading eventually to annual estimates. The internal workings of the model are all calculated in water depth in millimeters and then reported to the user in gallons of water through the environmental performance dashboard.
This simple bucket model begins with inputs from the piped water supply and precipitation.
Piped water demand is lifestyle-dependent. It is estimated on a per capita basis for residents using the residential water consumption rate parameter and on a per area basis for other use cases (e.g. office, retail, public assembly, etc.) based on the use water consumption rate density parameter, and adjusted for “gray water recycling” if that modifier applies, based on the proportion of graywater recycled parameter. The piped water demand is then divided between indoor and outdoor water use based on the proportion of piped water used outdoors parameter. We also estimate any water used to generate steam if the cogeneration plant ecosystem occurs within the vision extent. Steam water is adjusted for steam pipe leakage and proportion of steam recycled, both of which are lifestyle-dependent.
Out of doors water inputs (“surface inputs”) come from outdoor water use and precipitation and wastewater, assuming that not all of the wastewater is piped for some lifestyles (e.g. Lenape and average Earthling lifestyles) according to the proportion of wastewater piped parameter. Precipitation is determined by the user based on the precipitation event selection, which in turn is dependent on the climate event selection. These selections are both made through the “lifestyle/climate scenario” tab on the left side of the interface. For each precipitation event for each climate scenario, parameters define the precipitation intensity (mm/hr) and the precipitation duration (hours). The intensity and duration values reflect the "design storm event" (1.75 in/hour for one hour) used by the City of New York’s Department of Environmental Protection (NYC DEP 2008) and variations on it. The user can also select a “clear day” in which case there is no precipitation and all the water flow comes via the piped water supply.
Potential evapotranspiration is estimated using the Hamon (1963) method as described in Vörösmarty et al. (1996). The Hamon (1963) method depends on the average day length (hours of day light parameter) and the saturated vapor pressure, which itself is a function of the mean daily temperature. All storm events are assumed to occur in June; the June temperature parameter varies as a function of the climate scenario.
Actual evapotranspiration is estimated by comparing the evapotranspiration demand to the surface inputs (precipitation plus outdoor water use) plus water availability from open water ecosystems (i.e. estuary, pond, swimming pool), which is reported in stored > open water. In some applications, the evaporative draw from soil water is treated with a soil drying function that limits the amount of water available (c.f. Vörösmarty et al. (1998)), but in testing, we found it makes only a small difference, and so we will neglect the soil drying function for version 1 of Visionmaker NYC. Actual evapotranspiration is reported as water vapor > evapotranspiration metric. We also estimate the amount of steam lost from the steam pipe system, if steam heat is used within the vision extent, based on the lifestyle-dependent parameter steam pipe leakage, reported under water vapor > steam. We also estimate water held on plants according to an area-weighted average of the ecosystem dependent proportion of rainwater intercepted parameter; presumably this water also eventually enters the atmosphere.
Whatever surface inputs remain after evapotranspiration is poured over the pervious (e.g. soil) and impervious surfaces (e.g. concrete, stone or asphalt) and followed separately. Each ecosystem is described by a proportion of impervious surface parameter, which is used to calculate the area of impervious surface within the vision; all remaining areas are treated as pervious. The impervious water storage capacity is currently fixed at 3 mm for buildings, sidewalks and other completely impervious surfaces, but can be increased by painting the cistern, green roof, or bioswale modifers; the pervious water storage capacity depends on the ecosystem and is averaged across the pervious portions of the vision extent. Both impervious and pervious water storage have initial conditions set by parameters. After the storm event, the water held in the impervious storage is shown as "puddles on hardtops". We also report water held by cisterns, green roofs, and bioswales. The water held in the pervious (soil) storage is shown as mud > soil water. We report an empty metric (“Not estimated”) for water in the deep groundwater, recognizing that there is ground water storage and movements under the streets of Manhattan, but these are not estimated by the current version of the Visionmaker NYC water model.
After the pervious and impervious water stores are filled, any remaining water becomes outdoor runoff. Runoff is treated as flowing into streams first, if any exist. Streams have a stream drainage capacity which is estimated as if it were a stormwater capture system with a coefficient of 1.0, assuming the city’s design storm event. Runoff not captured by streams is then routed into the stormwater drainage system, if it exists; and finally to an undifferentiated floodwaters. Remember that the model does not take into account the internal geographic pattern of the vision, which results in the assumption that any portion of the outdoor runoff can reach any length of stream or storm water drain within the vision. Similarly floodwaters only include waters generated within the vision extent and do not include any topographically oriented flooding from adjacent areas outside the vision extent. Stream drainage capacity and stormwater drainage capacity are ecosystem dependent and are averaged across the vision extent. The parameterization for built ecosystems assumes that they have stormwater systems built to current New York City code (i.e. designed for the “design storm event”); as a result, flooding generally occurs only with more severe storm events.
Runoff is reported in two ways, by source and destination. We report estimates of impervious runoff from hard tops, pervious runoff from soil, and runoff from people (as explained below). We report the destinations of the water runoff as runoff into streams, floodwaters, and into the combined sewer system, as stormwater drainage and runoff from people.
Runoff from people is wastewater or sewage. We estimate indoor runoff after adjusting for graywater recycling, based on the indoor water use plus water resulting from steam heat (if applicable.) We assume that all piped water passes through the vision during the storm event day and into the sewage system and further that all of Manhattan has a combined sewer system, which joins the sewage flow with the storm water flow. Steam heat water depends on the amount of heat received from steam within the vision extent (see the carbon model.)
Finally we estimate the water lost from leaky pipes using the lifestyle-dependent parameter water pipe leakage, which is the proportion of flow lost to leaks from all the piped flows: combined sewer and piped water demand. This water is assumed to pass into the groundwater, which is not modelled.
Validation of the water model is ongoing. Check back here for updates.