Laserfiche WebLink
event that water tables do change in a location, provisions are made in Section 8 of this <br />memorandum for projecting the resulting change in ETg. <br />If actual ET occurring within any pixel (02 acre area) is of interest for any year, ET, can <br />be calculated using a raster providing spatially-interpolated precipitation for that year that <br />can be added to the basic raster values for ETg to yield spatially-correct ET, estimates <br />throughout the study area. Files of ETg that enable these manipulations are provided <br />digitally with this memorandum in addition to the base-line year ET, for 2001. <br />Precipitation is subtracted in the calculation of ETg and in Equation 5 ETg does not <br />consider a proportion of precipitation as "effective" as is often evaluated in rangeland <br />hydrology. Instead it simply subtracts precipitation because it is over and above the <br />amount supplied by the groundwater. Viewed as aone-dimensional problem, <br />precipitation combines with shallow groundwater to provide for the overall water balance <br />of the vegetation cover, and is used directly or reaches the water table to be consumed at <br />some later time. <br />5. Evaporation from Small Water Bodies <br />The calculation of evaporation (E) from large water bodies (> 3 acres) has been handled <br />elsewhere in SPDSS calculations. However, there are many smaller water bodies, <br />including the water surface of streams, themselves, that represent significant E from the <br />system not captured because of small size and diffuse location. Though these small water <br />bodies are frequently vegetated, ET estimates based upon NDVI* would be artificially <br />low, especially if plant cover is partial. We, therefore, identified small areas of water <br />surface (< 3 acres) within the phreatophyte boundary, assigned a separate rate to account <br />for evaporative losses, and then assured that this step did not double count E from the <br />water surface and ET from vegetation. <br />Small water bodies were identified using TM Band 5 data from the 2001 TM mosaic <br />described in Section 4.1. Water of greater than one centimeter deep absorbs all solar <br />radiation in the wavelengths measured by the TM Band 5 sensor (1.55 to 1.75 um). In <br />order to delineate open water areas from the rest of the scene, a suitable cutoff value was <br />chosen such that all pixels above this value were classified as dry, and all pixels below <br />this as wet. This cutoff value was chosen iteratively by visually observing the extent of <br />the classification over the DOQQ as the cutoff value was raised incrementally. At some <br />point during this process, known dry areas, especially those that were heavily vegetated, <br />began to be classified as wet, and it was then necessary to lower the cutoff value to a <br />level where this did not occur. For the purposes of this classification, this iteratively- <br />determined cutoff value was 0.08 TM Band 5 reflectance. <br />Once open water areas were accurately identified, they were compared with the water <br />body mask for >3 acres developed by RTi and provided by LRE. All overlapping areas <br />were removed from the Band 5 masks in order to avoid double counting. This process <br />resulted in a mask of small water bodies (< 3 acres) including portions of the stream <br />15 <br />