5. Predicting Urban Irrigation Rates for Silicon Valley Using Satellite Remote Sensing
Urban irrigation can account for more than half of all municipal and industrial water demand in cities of the western United States (Western Resource Advocates, 2010). Consequently, many municipalities have started to regulate the amount of irrigation water that can be applied to outdoor landscapes, due to rapid urban population growth, droughts, and inefficient application practices (Salamone, 2002).
Because water meters in urban areas do not generally track indoor and outdoor uses separately, the share of water used outdoors for irrigation can only be roughly approximated. The 2005 California Water Plan estimated that the residential sector used roughly 2.3 million acre feet (maf) outdoors in 2000, or 42 percent of total residential demand. Parks, golf courses, and other “large landscapes” used another 0.7 maf.
There can be significant year-to-year variation in urban water demand driven by irrigation needs for outdoor landscaping. In hot, dry years, evapotranspiration increases, and fewer water needs are fulfilled by rain and soil moisture (Christian-Smith et al., 2012). Climate warming will likely increase future urban water demands in California. Under the most extreme climate change scenario modeled by Christian-Smith et al. (2012), warming may cause an increase in California’s urban water demand by more than 1 maf annually by 2100.
To support local planning to meet (or mitigate) urban water demands, both now and in the future for Silicon Valley (SV), we propose to implement a technology solution that relies on publicly available NASA satellite imagery to map out irrigation rates for all individual properties in SV. NASA’s Landsat Thematic Mapper satellite imagery, aerial images, weather records, and multiple land-use maps will be used to identify the fraction of irrigated landscaping in all SV urban areas, and to estimate the monthly rate of irrigation being applied to those areas. Following the proven approach described by Johnson and Beliz (2012), we will used SV airport buildings and paved runways to represent the impervious surface water demand, and golf courses to represent the fully irrigated surface water demand.
We will identify areas of urban vegetation cover that are not irrigated using aerial imagery and compute the Landsat Normalized Difference Vegetation Index (NDVI) surplus, defined as the difference between the NDVI signals of the irrigated and non-irrigated vegetation cover areas. Areas between airport runways will be used as training zones for NDVI in non-irrigated vegetation cover. Since NDVI values from irrigated areas remains relatively constant throughout the year, while the NDVI from non-irrigated areas shows a distinct seasonally due to precipitation availability alone, we will compute Landsat NDVI ratios for two time periods each year for the past three years -- the wet, cool period of March-April and the hotter, drier period of August-September (see SV example figure uploaded as a supplement).
Irrigation water delivery records are admittedly difficult to obtain everywhere at equivalent quality and reliability. Therefore, we will use a surrogate for these records that correlates closely with potential irrigation rates, namely the daily evapotranspiration rate (ET) (Mayer et al., 1999; Keith et al., 2002). Daily ET of irrigated landscapes will be computed from real-time local weather station records and used to scale NDVI surplus ratio values (which are dimensionless) using an exponential scaling function (Johnson and Beliz, 2012). This technique has been shown to generate accurate and timely estimates of actual urban outdoor water demands everywhere in the area.
Our team has extensive experience in transitioning remotely sensed data and image products to decision support systems or tools for societal benefit. This experience comes from the Planetary Skin Institute partnership with Cisco, Inc. (www.planetaryskin.org), which provides connections to water and carbon policy and planning projects worldwide.
Christian-Smith, J., M. Heberger, and L. Allen, 2012, Urban Water Demand in California to 2100: Incorporating Climate Change, Pacific Institute, Oakland, CA, 60 pp.
Department of Water Resources (DWR), 2005, California Water Plan Update, Bulletin 160-05, Sacramento, California, December 2005.
Johnson, T D., and K. Belitz, 2012, A remote sensing approach for estimating the location and rate of urban irrigation in semi-arid climates, Journal of Hydrology, 414–415, 86–98.
Keith, D.J., Walker, H.A., Paul, J.F., 2002. Terrestrial vegetation greenness of the lower Galveston Bay watershed from satellite remote sensing and its relation to water use and the salinity regime of the Galveston Bay Estuary (USA). Int. J. Remote Sens. 23 (5), 905–916.
Mayer, P.W., DeOreo, W.B., Opitz, E.M., Kiefer, J.C., Davis, W.Y., Dziegielewski, B., Nelson, J.O., 1999. Residential End Uses of Water. American Water Works Association Research Foundation, Denver, CO. 109 p.
Salamone, D., 2002, A Drying Oasis Series: Florida Water Crisis Chapter 1. Orlando Sentinel. 03 Mar: A1.
Western Resource Advocates, 2010, Urban Sprawl: Impacts on Urban Water Use Smart Water: A Comparative Study of Urban Water Use Across the Southwest. Available from: www.westernresourceadvocates.org/media/pdf/SWChapter4.pdf
Why it should be recognized:
Present and future requirements for SV urban irrigation rates will be made available to local (district) water managers within three months of the start date of funding for this proposed project. Display and download of all urban irrigation map products from this SV Solutions project will be readily facilitated by transfer from www.casa2100.com in any standard (Open Geospatial Consortium compliant) file format required, such as GeoTIFF, ERDAS Imagine (.img), hierarchical data format (.hdf), or Google Maps (.kml). This internet map server with multiple data format delivery capability will ensure that urban irrigation maps can be analyzed by any user that requires zonal totals or averages and break-downs by parcel, city, township, or county boundaries.
The use of publicly available satellite data and climate station records to estimate present and future requirements for urban irrigation will make this technique readily scalable to any other (semi-) arid urban in the world. Local airports near urban centers commonly offer daily weather station records for use in water demand algorithms, which will support this technique even in developing nations overseas.