Laserfiche WebLink
Groeneveld et al. (in review) found that NDVI* and ET*, both dimensionless indices, <br />were equivalent using data sets from three widely disparate western North American <br />climates, locations and phreatophyte species. NDVI* in this context functions as a crop <br />coefficient that is mapped onto the ETg distribution that is, itself stretched from 0-1. The <br />equality expressed in Equation 4, formed the basis for making the annualized ET <br />estimates for phreatophyte vegetation: <br />NDVI*; = ET*, (4) <br />TM data processed to NDVI* for each pixel in the riparian habitat provided aspatially- <br />correct estimate for ET*. ET* is a variable that can be decoded to provide estimated <br />annual ET at any pixel "i" (ET,) using raster values of annual total ETo and precipitation <br />for that pixel. <br />4. Calculations to Yield Spatially-Correct Phreatophyte ETg from NDVI* <br />Oblique air photographs were taken during a familiarization flight over the South Platte <br />region as an aid for identification of features on satellite images and for differentiation of <br />phreatophyte vegetation from upland cover. These photos were catalogued and used as a <br />reference for mapping and interpretation of phreatophyte vegetation. They are contained <br />in the GIS files in the Splatte_aerial~hoto folder in the electronic data that accompany <br />this report. <br />The remote-sensing software used for analysis was ENVI (Environment for Visualizing <br />Images) marketed by ITT Visual Information Systems. Microsoft Excel was used for <br />spreadsheet calculations; and ESRI ArcGIS version 9.1 was used for spatial display of the <br />data and for mapping vectors to enclose the riparian vegetation. <br />4.1 The 2001 Mid-Summer TM Mosaic and Delineation of Phreatophyte Vegetation <br />Polygons outlining the extent of riparian phreatophyte vegetation were mapped using <br />ArcMap software by referencing the geocorrect Landsat TM images, 2001 land irrigation <br />maps created by Riverside Technology, Inc. (RTi), and high-resolution digital orthophoto <br />quarter quadrangles (DOQQs). The objective of this process was to accurately identify <br />and separate phreatophyte cover from surrounding lands. High vegetation vigor produces <br />a signal that is directly proportional to its water use and so, an important step was to <br />remove any high-NDVI* edges of the intensively cropped and irrigated lands adjacent to <br />the riparian zone since water use from agricultural irrigation was covered within other <br />calculations external to this effort. Mapped polygons of riparian vegetation also excluded <br />non-phreatophyte cover within phreatophyte ETg estimation to eliminate the potential <br />minor errors that could arise from measurement of vegetation external to groundwater <br />connection. A gross example of such vegetation would be the inclusion of weedy uplands <br />adjacent to irrigated farmland that received irrigation tail water or sprinkler overspray. <br />7 <br />