Investigating Alternate Methods to Automate Irrigated and Natural Vegetation Classifications for Water Accounting

Project ID: 6676
Principal Investigator: Ronald Miller
Research Topic: Water Supply Forecasting
Funded Fiscal Years: 2005, 2006, 2007 and 2008
Keywords: None

Research Question

* Can new data sources and classification methods be developed that will allow faster, cheaper and more detailed vegetation monitoring for evapo-transpiration (ET) and consumptive use estimation?

Need and Benefit

D-8260 has worked with the regions over the years to support and produce irrigated and riparian vegetation classifications. These are needed most critically as input to water accounting formulations for consumptive use estimations and reporting. Dynamic updates of vegetation statistics are required as input to ET and water accounting models.

Vegetation classifications can typically be performed spatially by image interpretation or spectrally which uses the computer to identify unique areas in spectral space that represent the target vegetation types. Most classifications to date have used either color infrared (CIR) photography for detailed visual mapping or Landsat multi-spectral scanner satellite imagery for automated classification of large areas. D-8260 has worked with Lower Colorado (LC) and Upper Colorado (UC) regions to produce vegetation statistics for five-year Colorado River consumptive use reports to Congress. D-8260 has provided similar vegetation mapping for the Middle Rio Grande (MRG) to support water accounting. The Albuquerque Area Office has recently indicated that specialized vegetation inventory methods need to be developed for MRG to meet their special water accounting needs.

MRG requirements include crop and riparian species classifications in addition to irrigation status. The Rio Grande agriculture and riparian zones contain a wide variety of vegetation types that occur in relatively small areas. Quarterly statistics are needed from single date classifications for input into ET models. More frequent ET estimates are also needed to identify shorted fields and riparian depletions are needed to be estimated daily. The MRG processing requires a 30-day processing turn-around for each quarter when an approved method becomes operational. Conventional Landsat classifications are not considered sufficient in resolution and detailed vegetation mapping using aerial photography is too costly and time consuming.

The UC Region/Albuquerque Area Office have requested that two levels of research be proposed. This proposal addresses development of an operational plan within the next two years. This plan may not sufficiently address all MRG needs, especially the ET, riparian depletion statistics, and high automation level requirements. A separate scoping request will be submitted to evaluate new technologies that could meet these goals. They are willing to provide supplemental funding ($25, 000 /year for 2 years) if Science and Technology (S&T) funds can be allocated to develop an optimum process for MRG water accounting. The annual costs for the new methods must fit within the operational budget allowed for the MRG. (approximately $50, 000 per year). If vegetation inventories can be developed and optimized for MRG, better consumptive use estimates and water accounting can be produced for UC and Reclamation as a whole. Findings and techniques may also be able to be applied to other Reclamation districts containing similar vegetation patterns.

Contributing Partners

Contact the Principal Investigator for information about partners.

Research Products

Please contact about research products related to this project.

Return to Research Projects

Last Updated: 6/22/20