Polly use of remote sensing products for local water
- 1. Use of Remote Sensing Products for Local Water Supply and Use Applications Jason Polly, GIS Group Leader September 2014
- 2. Introduction Water related data-driven decisions require sufficient spatial and temporal coverage for appropriate implementation. Available supply is directly related to use applications. 2 Consumptive Needs - Projected Water Use Content and imagery courtesy of the report “SWSI 2010” by Colorado Water Conservation Board (CWCB)
- 3. Traditional Methods of Estimating Water Supply 3 Technology used Traditional methods Municipal Meters Gauging Stations Weather Stations Ground Water Monitoring Wells Irrigation Flow Meters
- 4. Remote Sensing (RS) Overview Remote Sensing The science and art of obtaining information about an object area or phenomenon through the analysis of data acquired by a device that is not in contact wit the object, area or phenomenon under investigation. 4
- 5. Comparison – Traditional and RS Techniques 5 Criteria Traditional Remote Sensing Spatial Coverage Single location Can cover large scale applications Temporal Coverage Limited by date of instillation Limited by sensor repeat cycle Precision Limited by data logging capabilities Limited by sensor resolution Cost Maintenance (often yearly) Free for large government sensors and cost based for commercial collection
- 6. Case Studies • Urban Irrigation Monitoring (South Adams County Water and Sanitation District, SACWSD) • Snow Pack (Dolores Water Conservancy District, DWCD) • Crop Consumptive Use (Colorado Water Conservation Board, CWCB. Wyoming State Engineer's Office,. ect) 6
- 7. Urban Irrigation Monitoring Reusable Water Water reuse is any arrangement that utilizes legally reusable municipal return flows to increase municipal water supplies. Return flows are water that returns to a river after being treated at a wastewater treatment plant or to alluvial aquifers via percolation. Reuse can be accomplished in at least two ways: 1) return flows can be physically reused for non-potable and potable purposes. 2) return flows can be reused under various substitution or exchange arrangements. To increase water supply through reuse, municipal return flows must be legally reusable. Under Colorado water law, reusable Water available to Front Range water utilities can generally come from the following sources: 1. Water imported to the South Platte or its tributaries from another river basin 2. Nontributary groundwater from Denver Basin aquifers 3. The historically consumed portion of water rights changed from oneuse to another, such as from irrigation to municipal use 4. Water diverted under a water right that has been decreed to allow for reuse 7 Content and imagery courtesy of the report “Filling the Gap” by Western Resource Advocates (WRA), Trout Unlimited (TU), and the Colorado Environmental Coalition (CEC)
- 8. Urban Irrigation Monitoring Lawn Irrigation Return Flow (LIRF) Major irrigated areas were identified in residential, commercial, industrial, and remaining urban zones District (SACWSD) for the year 2013. High-resolution WorldView-2 satellite images, acquired in May-June produced the early season image. The late season image covering the entire district was generated using data acquired in July- September. Using semi-automated remote sensing classification techniques, the early and late season images were used in combination to produce the 2013 irrigated acreage estimation. Results were summarized by parcel, and irrigated acreage estimates are reported for each Return Flow Plot (RFP’s) serviced by the district. High-resolution satellite imagery for the following 2013 approximate dates: Mid May- early June, 2013 (Early Season Image) July-September, 2013 (Late Season Image) 8 Methodology Flow Chart
- 9. Urban Irrigation Monitoring Lawn Irrigation Return Flow (LIRF) 9 The technical specifications of the WorldView-2 products. The high-resolution panchromatic band provides a very detailed spatial representation of urban features, while the infrared band-4 capability of the multi-spectral bands allows for a better discrimination of irrigated vegetation as compared to natural color imagery. In addition, 11-bit WorldView-2 imagery provides excellent radiometric resolution. World View 2 – Sensor Technical Specifications
- 10. Urban Irrigation Monitoring Lawn Irrigation Return Flow (LIRF) 10
- 11. Urban Irrigation Monitoring Lawn Irrigation Return Flow (LIRF) The Normalized Difference Vegetation Index (NDVI) has been in use for many years to measure and monitor plant growth (vigor), vegetation cover, and biomass production from multi-spectral satellite data (Jackson and Huete 1991, Jensen 1996, Lillesand and Kiefer 2000). NDVI was derived from the difference between the near-infrared region of the electromagnetic spectrum (e.i., WorldView-2 Band 4), and the visible red region (e.i., WorldView- 2 Band 3), using the following equation: 11 NDVI – Irrigation Scaling
- 12. Urban Irrigation Monitoring Lawn Irrigation Return Flow (LIRF) Multi-temporal Image Analysis Approach Since acceptable imagery was obtained for the early and late season periods, a multi-temporal image analysis approach was possible.. This approach was adopted from the 2009 (Riverside, 2009) analysis and greatly increases the overall accuracy on the analysis by capturing irrigated areas on two separate dates. Demonstrating the approach: 12 Early/Late Seasonal Image Comparisons
- 13. Urban Irrigation Monitoring Lawn Irrigation Return Flow (LIRF) 13 Land Cover Classification Augmentation Plan, Case No. 2001CW258, all deep percolation occurring under trees is fully consumed and therefore does not return to the stream system. This results in a lower percentage of return flows being claimed when using the Cottonwood Curve than if trees and shrubs were not present. To produce a more detailed land cover classification, Riverside used an unsupervised classification technique to separate the ‘Trees and Shrubs’ from the irrigated class previously obtained from the NDVI analysis, as well as the ‘Water’ class from the ‘Non-irrigated’ class. The unsupervised classification was performed in ERDAS Imagine using the ISODATA algorithm to iteratively divide the WorldView2 data into clusters or groups of pixels with similar spectral characteristics. The remote sensing analyst then assigned each cluster to its corresponding land cover categories (e.g., irrigated grass, trees and shrubs, non-irrigated, and water).
- 14. Urban Irrigation Monitoring Lawn Irrigation Return Flow (LIRF) The parcels included for LIRF analysis and parcels not included for LIRF analysis were clipped by the RFP zones to summarize the irrigated acreage. The parcels were then overlaid with the NDVI classification and the unsupervised classification raster data. The irrigated areas were tabulated to obtain the irrigated acreage in the included parcels and parcels not included at this time for the summary reports. 14 Methodology Flow Chart Summary Statistics by Return Flow Area
- 15. SnowPack (RS Methods) NOAA National Weather Service's National Operational Hydrologic Remote Sensing Center (NOHRSC) SNOw Data Assimilation System (SNODAS) • SNODAS is a modeling and data assimilation system developed by the NOHRSC to provide the best possible estimates of snow cover and associated variables to support hydrologic modeling and analysis. • The aim of SNODAS is to provide a physically consistent framework to integrate snow data from satellite and airborne platforms, and ground stations with model estimates of snow cover (Carroll et al. 2001). • The snow model has high spatial (1km) and temporal (1 hour) resolutions and is run for the conterminous United States. 15 Image/photo courtesy of Andrew P. Barrett and the National Snow and Ice Data Center, University of Colorado, Boulder
- 16. SnowPack (RS Methods) Process Historical SNODAS Data and Provide SWE Traces for the 2013 Snowmelt Season • Riverside processed the SNODAS Snow Water Equivalent (SWE) grids for the period October 2003-July 2013 to generate daily time series of basin-average SWE. • The McPhee watershed was divided into six sub-basins to show snowpack patterns in different parts of the watershed. The SWE traces for the 2013 snowmelt season were updated weekly from March-May 2013 to help characterize the 2013 snow season in real time. The historical SNODAS SWE time series were also provided for context for the 2013 snow conditions. • The SNODAS SWE time series were plotted using a graph similar to that commonly used for SNOTEL data This type of graph is useful for assessing basic information about the magnitude of the snow accumulation and the timing of the snowmelt. • In WY 2013, the snow accumulation for the McPhee watershed was average to below-average. The snow melted out relatively late, particularly for the modest snow accumulations, due to cool spring temperatures. 16 Image courtesy of Amy Volckens Riverside Technology, inc.
- 17. SnowPack (RS Methods) Prepare Operational SNODAS maps for the 2013 Snowmelt Season • In addition to the basin-average time series, Riverside prepared several maps showing the SWE conditions being modeled by SNODAS. The maps included labels with the current snowpack volume in each sub-basin as well as the seven-day change in the snowpack volume. 17
- 18. SnowPack (RS Methods) Prepare Operational SNODAS maps for the 2013 Snowmelt Season SNODAS SWE on February 7, 2012 SNODAS SWE on February 7, 2013 18
- 19. SnowPack (RS Methods) Prepare Operational SNODAS maps for the 2013 Snowmelt Season SNODAS SWE on April 3, 2012 SNODAS SWE on April 2, 2013 19
- 20. Remote Sensing-Based ET Estimation METRIC method: Mapping EvapoTranspiration with high Resolution and Internalized Calibration (developed by Dr. Rick Allen, University of Idaho) METRIC is a sort of “hybrid” between pure remotely-sensed energy balance and weather-based ET • Advantages • Can acquire data rapidly over large regions • Do not require irrigation diversion and pumping well records • Can detect use of subsurface supplies • Do not require crop classification • Can detect actual field conditions 20 methods Combines the strengths of energy balance from satellite and accuracy of ground-based reference ET calculation: satellite-based energy balance provides the spatial information and distribution of available energy and sensible heat fluxes over a large area (and does most of the “heavy lifting”) reference ET calculation “anchors” the energy balance surface and provides “reality” to the product.
- 21. • New Mexico – Water consumption by invasive vegetation along the Rio Grande • Colorado – Conjunctive management of ground-water and surface water by State Engineer along the South Platte – Assessment of water shortage and salinity impacts along the Arkansas River • Nebraska – Ground-water management and mitigation in the Ogallala Aquifer in western Nebraska – Testing against measured ET in central NE • Wyoming – Green River Basin crop consumptive use estimates • Morocco – Used in providing a complete water budget where no ground water records exist. 21 Remote Sensing-Based ET Estimation
- 22. Crop Consumptive Use (METRIC method) 22
- 23. Remote Sensing-Based ET Estimation METRIC method: 23 The satellite can not “see” ET therefore ET is calculated as a “residual” of the energy balance: ET = Rn – G - H R n Net radiation H Heating of air ET Evapotranspiration G Soil heat flux Basic Truth: Evaporation consumes Energy
- 24. Remote Sensing-Based ET • Net Radiation (Rn), calculated using 24 – Sun-earth geometry – Spectral reflectance from the surface – Thermal radiance from the surface – Transmissivity of Atmosphere – Ground Heat Flux (G), Calculated using – Vegetation Amount – Net radiation – Thermal radiance – Sensible Heat Flux (H), Calculated using – Thermal radiance – Wind speed – Surface cover type and roughness – Surface to air temperature difference, dT Rn H ET G underlined terms are obtained from the satellite data Estimation
- 25. 25 Remote Sensing-Based ET Estimation METRIC Requirements: •Satellite images with Thermal Band High resolution (Landsat 5, 7 and now 8) is needed for field scale maps •Good quality weather data for best calibration
- 26. 26 Future Implications Criteria Remote Sensing Perceived Future Implications Spatial Coverage Can cover large scale applications Greater cloud free repeat cycles with increased satellite constellations Temporal Coverage Limited by sensor repeat cycle Greater cloud free repeat cycles with increased satellite constellations Precision Limited by sensor resolution Recent lift on U.S satellite resolution restrictions .25m panchromatic. Cost Free for large government sensors and cost based for commercial collection 2007 Landsat archive open to public.
- 27. 27 Questions Jason.Polly@Riverside.com 2950 E. Harmony Road Suite 390 Fort Collins, CO 80528