U.S. Geological Survey 2012 raster digital data Reston, VA U.S. Geological Survey http://water.usgs.gov/lookup/getspatial?mosaic_et_august2006_kl_NAD83 Evapotranspiration, Plus, LLC 20100209 3496 N. 2500 E., Twin Falls, Idaho 83301 Evapotranspiration, Plus, LLC of Twin Falls, Idaho under contract to the U.S. Geological Survey should be acknowledged as the data source in products derived from these data Hydrological Information Products for the Off-Project Water Program of the Klamath Basin Restoration Agreement U.S. Geological Survey Open-File Report 2012-1199 U.S. Department of the Interior By Daniel T. Snyder, John C. Risley, and Jonathan V. Haynes Prepared in cooperation with The Klamath Tribes Access complete report at: http://pubs.usgs.gov/of/2012/1199 Suggested citation: Snyder, D.T., Risley, J.C., and Haynes, J.V., 2012, Hydrological information products for the Off-Project Water Program of the Klamath Basin Restoration Agreement: U.S. Geological Survey Open-File Report 2012–1199, 17 p., http://pubs.usgs.gov/of/2012/1199 Summary The Klamath Basin Restoration Agreement (KBRA) was developed by a diverse group of stakeholders, Federal and State resource management agencies, Tribal representatives, and interest groups to provide a comprehensive solution to ecological and water-supply issues in the Klamath Basin. The Off-Project Water Program (OPWP), one component of the KBRA, has as one of its purposes to permanently provide an additional 30,000 acre-feet of water per year on an average annual basis to Upper Klamath Lake through “voluntary retirement of water rights or water uses or other means as agreed to by the Klamath Tribes, to improve fisheries habitat and also provide for stability of irrigation water deliveries.” The geographic area where the water rights could be retired encompasses approximately 1,900 square miles. The OPWP area is defined as including the Sprague River drainage, the Sycan River drainage downstream of Sycan Marsh, the Wood River drainage, and the Williamson River drainage from Kirk Reef at the southern end of Klamath Marsh downstream to the confluence with the Sprague River. Extensive, broad, flat, poorly drained uplands, valleys, and wetlands characterize much of the study area. Irrigation is almost entirely used for pasture. To assist parties involved with decisionmaking and implementation of the OPWP, the U.S. Geological Survey (USGS), in cooperation with the Klamath Tribes and other stakeholders, created five hydrological information products. These products include GIS digital maps and datasets containing spatial information on evapotranspiration, subirrigation indicators, water rights, subbasin streamflow statistics, and return-flow indicators. The evapotranspiration (ET) datasets were created under contract for this study by Evapotranspiration, Plus, LLC, of Twin Falls, Idaho. A high-resolution remote sensing technique known as Mapping Evapotranspiration at High Resolution and Internalized Calibration (METRIC) was used to create estimates of the spatial distribution of ET. The METRIC technique uses thermal infrared Landsat imagery to quantify actual evapotranspiration at a 30-meter resolution that can be related to individual irrigated fields. Because evaporation uses heat energy, ground surfaces with large ET rates are left cooler as a result of ET than ground surfaces that have less ET. As a consequence, irrigated fields appear in the Landsat images as cooler than nonirrigated fields. Products produced from this study include total seasonal and total monthly (April–October) actual evapotranspiration maps for 2004 (a dry year) and 2006 (a wet year). Maps showing indicators of natural subirrigation were also provided by this study. “Subirrigation” as used here is the evapotranspiration of shallow groundwater by plants with roots that penetrate to or near the water table. Subirrigation often occurs at locations where the water table is at or above the plant rooting depth. Natural consumptive use by plants diminishes the benefit of retiring water rights in subirrigated areas. Some agricultural production may be possible, however, on subirrigated lands for which water rights are retired. Because of the difficulty in precisely mapping and quantifying subirrigation, this study presents several sources of spatially mapped data that can be used as indicators of higher subirrigation probability. These include the floodplain boundaries defined by stream geomorphology, water-table depth defined in Natural Resources Conservation Service (NRCS) soil surveys, and soil rooting depth defined in NRCS soil surveys. The two water-rights mapping products created in the study were “points of diversion” (POD) and “place of use” (POU) for surface-water irrigation rights. To create these maps, all surface-water rights data, decrees, certificates, permits, and unadjudicated claims within the entire 1,900 square mile study area were aggregated into a common GIS geodatabase. Surface-water irrigation rights within a 5-mile buffer of the study area were then selected and identified. The POU area was then totaled by water right for primary and supplemental water rights. The maximum annual volume (acre-feet) allowed under each water right also was calculated using the POU area and duty (allowable annual irrigation application in feet). In cases where a water right has more than one designated POD, the total volume for the water right was equally distributed to each POD listed for the water right. Because of this, mapped distribution of diversion rates for some rights may differ from actual practice. Water-right information in the map products was from digital datasets obtained from the Oregon Water Resources Department and was, at the time acquired, the best available compilation of water-right information available. Because the completeness and accuracy of the water-right data could not be verified, users are encouraged to check directly with the Oregon Water Resources Department where specific information on individual rights or locations is essential. A dataset containing streamflow statistics for 72 subbasins in the study area was created for the study area. The statistics include annual flow durations (5-, 10-, 25-, 50-, and 95-percent exceedances) and 7-day, 10-year (7Q10) and 7-day, 2-year (7Q2) low flows, and were computed using regional regression equations based on measured streamflow records in the region. Daily streamflow records used were adjusted as needed for crop consumptive use; therefore the statistics represent streamflow under more natural conditions as though irrigation diversions did not exist. Statistics are provided for flow rates resulting from streamflow originating from within the entire drainage area upstream of the subbasin pour point (referring to the outlet of the contributing drainage basin). The statistics were computed for the purpose of providing decision makers with the ability to estimate streamflow that would be expected after water conservation techniques have been implemented or a water right has been retired. A final product from the study are datasets of indicators of the potential for subsurface return flow of irrigation water from agricultural areas to nearby streams. The datasets contain information on factors such as proximity to surface-water features, geomorphic floodplain characteristics, and depth to water. The digital data, metadata, and example illustrations for the datasets described in this report are available on-line from the USGS Water Resources National Spatial Data Infrastructure (NSDI) Node Website http://water.usgs.gov/lookup/getgislist or from the U.S. Government website DATA.gov at http://www.data.gov with links provided in a Microsoft® Excel® workbook in appendix A. Introduction Program Background The Klamath Basin Restoration Agreement (KBRA) was developed by a diverse group of stakeholders, Federal and State resource management agencies, Tribal representatives, and interest groups to provide a comprehensive solution to ecological and water-supply issues in the basin. The KBRA covers the entire Klamath Basin, from headwater areas in southern Oregon and northern California to the Pacific Ocean, and addresses a wide range of issues that include hydropower, fisheries, and water resources. The Water Resources Program (Part IV of the KBRA) includes a section (16) known as the Off-Project Water Program (OPWP) (Klamath Basin Restoration Agreement, 2010, p. 105). Program Goals The primary goals of the OPWP include developing an Off-Project Water Settlement to resolve upper basin water issues, improve fish habitat, and provide for stability in irrigation deliveries (Klamath Basin Restoration Agreement, 2010, p. 105). One of the approaches to achieving these objectives is a water-use retirement program. The water-use retirement program is an effort to permanently provide an additional 30,000 acre-ft of water per year on an average annual basis to Upper Klamath Lake through “voluntary retirement of water rights or water uses, or other means as agreed to by the Klamath Tribes, to improve fisheries habitat and also provide for stability of irrigation water deliveries” (Klamath Basin Restoration Agreement, 2010, p. 105–111). The KBRA sets a 24-month window after the “effective date” for development of a proposal for the Off-Project Water Settlement. There is interest on the part of the Klamath Watershed Partnership (and others) in having a decisionmaking process in place before this time line. To assist parties in the OPWP involved with decisionmaking and implementation, the USGS proposed a two-phase approach. The first phase, which is described in this report, includes compilation and evaluation of relevant existing work and data in the upper basin, and synthesizing that information into a set of five hydrological information products. These products include GIS digital maps and datasets containing spatial information on evapotranspiration, subirrigation indicators, water rights, subbasin streamflow statistics, and return-flow indicators. Should efforts continue, a second phase could be developed to implement a monitoring program to evaluate the level of success of the first phase and to address additional information needs. Understanding the response of streams and groundwater to various land-use changes (such as reduction of irrigation or changes in land management) in particular areas is important to maximizing the benefits to streams and to Upper Klamath Lake while minimizing the impacts to the agricultural community. The hydrology of the region is such that the response to changes in land use will vary from place to place. Because of this, the benefit to the stream from a particular change in land or water use may be greater in one area than another. Description of Project Area The OPWP area is defined in the KBRA as including the Sprague River drainage, the Sycan River drainage downstream of Sycan Marsh, the Wood River drainage, and the Williamson River drainage from Kirk Reef at the southern end of Klamath Marsh downstream to the confluence with the Sprague River, encompassing a total area of approximately 1,900 mi2. Individually, the Sprague, Williamson, and Wood Rivers provide about 33, 18, and 16 percent, respectively, of the total inflow to Upper Klamath Lake and together account for two-thirds of the total inflow (Hubbard, 1970; Kann and Walker, 1999, table 3). Extensive, broad, flat, poorly drained uplands, valleys, and wetlands characterize much of the study area. Elevations in the study area range from about 4,100 ft at Upper Klamath Lake to greater than 9,000 ft in the Cascade Range. In general, land use in the Williamson River, Sprague River, and Wood River basins varies with elevation. At the lowest elevations, adjacent to the major rivers, agricultural lands (primarily irrigated pasture) predominate. Rangelands primarily are on the tablelands, benches, and terraces, and forest is predominant on the slopes of buttes and mountains. Livestock grazing can occur on irrigated pastureland, rangeland, and forestland throughout the study area. Average annual precipitation in the area ranges from as low as about 15 in. near Upper Klamath Lake to about 65 in. at Crater Lake with most precipitation occurring largely as snow in the fall and winter (Western Regional Climate Center, 2012). Previous Studies and Water Conservation Programs Recent studies in the Upper Klamath, Wood River, and Sprague River basins provided a foundation for many of the analyses made for this current study. A study of the regional groundwater hydrology of the Upper Klamath Basin is presented in Gannett and others (2007) and includes discussions of the hydrogeologic units, hydrologic budget, and configuration of the groundwater-flow system. Although the scale of this study is less useful for site-specific analysis, it provides a framework for analysis of the hydrology of the OPWP area. Carpenter and others (2009) provided a comprehensive analysis of hydrologic and water-quality conditions during restoration of the Wood River wetland for 2003–05. In their study, they developed a water budget for the wetland in addition to analyzing the mechanics of groundwater and soil moisture storage. Risley and others (2008) developed streamflow regression models used in this study to estimate a suite of streamflow statistics in study area subbasins. The Natural Resources Conservation Service (2009) presented findings from the Sprague River Conservation Effects Assessment Project (CEAP). Their report documented the effects of water conservation practices on private irrigated lowlands and uplands using field monitoring and hydrologic computer model simulations. Watershed Sciences LCC (2000) conducted a Forward-Looking Infrared (FLIR) survey flown in August 1999 for parts of the Upper Klamath Basin that collected both thermal infrared and color videography to map stream temperatures that can be used to identify point locations where return flows enter streams. Purpose of This Report This report summarizes and provides details on information products created by the USGS for the OPWP and its implementation. These products include a set of digital maps in GIS (ArcMap) format that can be used together as overlays to help evaluate the relative benefits of reducing or curtailing water use in various areas. The maps are not intended to drive the decisionmaking process, but to inform the process. It is envisioned that there will be many additional considerations affecting decisions. The digital maps created for this study, and described below in more detail, are (1) evapotranspiration, (2) subirrigation indicators, (3) water rights, (4) subbasin streamflow statistics, and (5) irrigation return-flow indicators. Access to Data, Metadata, and Example Illustrations The digital data, metadata, and example illustrations for the datasets described in this report are available on-line from the USGS Water Resources National Spatial Data Infrastructure (NSDI) Node Website (U.S. Geological Survey, 2010c) or from the U.S. Government Website DATA.gov (2012). Appendix A consists of a Microsoft® Excel® workbook listing each dataset and URL links to the website for the dataset, metadata, and example illustrations. Evapotranspiration Mapping Development Maps quantifying evapotranspiration (ET) over the entire landscape included in the OPWP were produced under contract for this study by Evapotranspiration, Plus, LLC, of Twin Falls, Idaho. The maps were created using a high-resolution remote sensing technique first developed by the University of Idaho (Allen and others, 2007a, 2007b). The technique known as “Mapping EvapoTranspiration at High Resolution and Internalized Calibration” (METRIC) uses Landsat imagery to estimate monthly actual evapotranspiration at 30-m resolution that can be related to individual irrigated fields. For the KBRA OPWP study, METRIC was applied to 2 separate years of growing season data for which suitable Landsat imagery was available, representing wet (2006) and dry (2004) years. By using these 2 years, it was possible to develop a range of likely actual ET over varied climate conditions. A small number of irrigated areas in the extreme eastern part of the Sprague River basin were not covered by the selected Landsat images used in the METRIC analysis. For these areas, ET was estimated using more traditional approaches that used standard ET models and crop coefficients combined with knowledge of crop and vegetation types. The METRIC procedure uses thermal infrared images from Landsat satellites to quantify ET. Because evaporation uses heat energy, ground surfaces with large ET rates are left cooler than ground surfaces that have less ET. As a consequence, irrigated fields appear on the images as being cooler than nonirrigated fields. The METRIC model is internally calibrated using ground-based reference ET. Both the rate and spatial distribution of ET can be efficiently and accurately quantified. A major advantage of using METRIC over conventional methods of estimating ET that use crop coefficient curves is that neither the crop development stages nor the specific crop type need to be known. In addition to ET, the fraction of reference crop evapotranspiration (ETrF) also is computed by METRIC. The alfalfa reference evapotranspiration (ETr), computed using local weather station meteorological data, is needed in calibrating METRIC to a specific study area. Previous studies have shown that the error between ET estimated from METRIC and measured from lysimeters daily and monthly for various crops and land uses in other areas has been from 1 to 4 percent (Allen and others, 2007b). For the current study, the accuracy of the METRIC ET values for irrigated areas was estimated as 10 percent for seasonal total ET values and 20 percent for monthly ET values (R.G. Allen, Evapotranspiration, Plus, LLC, written commun., 2011). The accuracy of the METRIC ET values for nonirrigated areas was estimated as 20 percent for seasonal total ET values and 40 percent for monthly ET values (R.G. Allen, Evapotranspiration, Plus, LLC, written commun., 2011). These larger values for estimated accuracy relative to other studies are a result of a number of factors including the limited availability of Landsat images not impeded by cloud cover or sensor failure during the period of interest and the heterogeneity of the study area with regard to vegetation, terrain, and soils. When making comparisons between individual areas of actual evapotranspiration, the relative difference between the areas likely has a much better accuracy than the accuracy of the absolute values of actual evapotranspiration for the individual areas. Products produced from this study include total seasonal and total monthly (April–October) actual evapotranspiration maps, in millimeters, for 2004 (dry year) and 2006 (wet year) and Landsat image maps for April–November 2004 and April–November 2006. Full details regarding Landsat image processing, METRIC calibration, and map production for this study are provided in separate reports written by the contractor and included in the GIS metadata (Evapotranspiration, Plus, LLC, 2011a, 2011b, 2011c). Subirrigation Indicators Definition “Subirrigation” as used here is the evapotranspiration of shallow groundwater by plants with roots that penetrate to or near the water table. Subirrigation often occurs in locations where the water table is at or above the plant rooting depth. It can occur where the water table is naturally high or where it is artificially elevated from irrigation. Certain settings, such as lowland areas along present flood plains, are more likely to naturally subirrigate than areas more distant or elevated above surface-water features. This study deals primarily with natural subirrigation occurrence. Because of the difficulty in defining the exact occurrence of subirrigation, this study presents several sources of spatially mapped data that can be used as indicators of higher subirrigation probability. These include (1) the floodplain boundaries and features reflecting stream geomorphology, (2) the water-table depth defined in NRCS soil surveys and by topographic analysis, and (3) the rooting depth defined in NRCS soil surveys. The indicators may be used separately or together, such as depth to water and plant rooting depth, to determine the overall likelihood that subirrigation may take place. Map Descriptions Floodplain Boundaries and Features Floodplains boundaries and features were delineated in a study of Sprague River basin geomorphology conducted by the USGS and the University of Oregon (J.E. O’Connor, U.S. Geological Survey, written commun., 2011). In the study, channel and floodplain processes were evaluated for 81 mi of the Sprague River, including the lower 12 mi of the South Fork Sprague River, the lower 10 mi of the North Fork Sprague River, and the lower 39 mi of the Sycan River. In addition to floodplain boundaries, other GIS layers created for the USGS Sprague River basin geomorphology study are channel centerlines, fluvial bars, vegetation, water features, and built features such as irrigation canals, levees and dikes, and roads that were created from aerial photographs taken from 1940 through 2005, 7.5-minute USGS topographic maps, digital orthophoto quadrangles, and LiDAR (Light Detection and Ranging) images (Watershed Sciences, LCC, 2000). Additional details on the USGS Sprague River basin geomorphology study that developed the floodplain boundary GIS layer can be found at the project website (U.S. Geological Survey, 2011a) or by viewing the metadata for the study (U.S. Geological Survey, 2011b). . The geomorphic unit categories for the areas in and adjacent to floodplains from the Sprague River Oregon Geomorphology dataset (U.S. Geological Survey, 2011b) were assigned qualitative values for subirrigation potential (J.E. O’Connor, U.S. Geological Survey, written commun., 2011). Determination of low, medium, or high subirrigation potential was made on the basis of the characteristics of areas from existing datasets and field observations of soils, vegetation, topography, and hydrology. However, some areas, including wetlands, springs, and ponds, were not mapped with the geomorphic floodplain and are not represented. Soil Rooting Depth The soil rooting depth map is based on data from the USDA NRCS Klamath County soil survey (Cahoon, 1985, p. 13–96) and supplemented by the Soil Survey Geographic (SSURGO) Database (Soil Survey Staff, 2010). The area of the soil survey excludes most public lands, such as National Forest or National Park areas or small private inholdings with these areas. Values of rooting depths typically are presented as either a range between 10 and 60 in. or as being greater than 60 in. For the purposes of this study, minimum, mean, and maximum rooting depths were calculated using the minimum and maximum rooting depth values. For calculation purposes, rooting depths greater than 60 in. are reported as equal to 60 in. Areas where the rooting depth is greater than the depth to water might support subirrigation. Depth to Water The depth-to-water map is based on data for the seasonal high water-table depth presented in the Natural Resources Conservation Service soil survey for southern Klamath County, Oregon (Cahoon, 1985, table 18, p. 258–263) and supplemented by the Soil Survey Geographic (SSURGO) Database (Soil Survey Staff, 2010). As noted above, the area of the soil survey excludes most public lands. Values of seasonal high water-table depth in Cahoon (1985, table 18) or the SSURGO dataset are typically presented as a range between minimum and maximum values. For the purposes of this study, a mean water-table depth was calculated using the minimum and maximum depth to water values. Maps of areas where the depth to water is less than the plant rooting depth provide insight into the likelihood that subirrigation may take place. Water-Rights Mapping Description of Mapping Water-right information in the map products is from digital datasets obtained on July 18, 2011, from the Oregon Water Resources Department (OWRD) and was, at the time acquired, the best available compilation of water-right information. Because the completeness and accuracy of the water-right data could not be verified, users are encouraged to check directly with the OWRD for situations where specific information on individual rights or locations is essential. The two water-right maps produced for the study were a “point of diversion” (POD) map that shows locations of diversion from streams, and a “place of use” (POU) map that shows irrigated areas. Only surface-water rights are included on the maps; groundwater rights are not included. In compiling the surface-water rights data, all decrees, certificates, permits, and unadjudicated claims in the study area were aggregated. The objective was to assemble all known water rights and claims into a common GIS geodatabase consisting of one POU polygon feature class and one relating POD point feature class. For both maps, related POUs and PODs share the same “snp_id” value. All other fields whenever possible were carried through the process to preserve as many original POU and POD attributes as possible. Note that POU polygons may overlap adjacent POU polygons and care is advised to ensure that the correct polygon(s) are selected or used in analyses, such as summation of attributes, to meet the intended purposes of the user. All Oregon surface-water rights, including decrees, certificates, and permits (http://gis.wrd.state.or.us/data/wr_state.zip), were downloaded from the OWRD GIS water-right website (Oregon Water Resources Department, 2012a). Surface-water irrigation water rights for the study area and within a 5-mi buffer of the study area were then selected. The POU area was totaled by water right for primary and supplemental water rights. The maximum annual volume (acre-feet) allowed under each water right was calculated using the POU area and duty (annual irrigation application in feet). In situations where no duty was specified, the maximum annual volume allowed under each water right was estimated assuming a duty of 3 ft/yr (82 percent of surface-water irrigation PODs in the study area had a duty of 3 ft/yr). Often a water right has more than one designated POD. In these cases, the volumes were equally distributed to each POD within the particular water right. The POUs and PODs of Klamath Basin unadjudicated claims were provided in a GIS geodatabase (D. Mortenson, Oregon Water Resources Department, written commun., 2011). To supplement the geodatabase, data (such as priority dates, id numbers, and volumes) for many, although not all, of the claims were downloaded from OWRD’s Water Rights Information System (WRIS) (2012b). Although, the PODs for the claims in the OWRD provided geodatabase did not include a use field, it was assumed that all PODs for each surface-water irrigation claim were used for surface-water irrigation. In cases where claims included multiple PODs, volumes were equally distributed. The maximum annual volume allowed under each claim was either provided or estimated. For approximately 25 percent of the claims, the maximum annual volume for surface-water irrigation was provided by WRIS in acre-feet. For the remaining 75 percent of the claims, volumes were estimated using the POU area and assuming a duty of 3 ft/yr (no claims had assigned duties). Additionally, an annual volume by claim from the adjudication process for the 1864 Walton claims was provided to the study (D. Watson, Ranch and Range Consulting, written commun., 2011). Each of these volumes was a result of proposed order, stipulated agreement, or uncontested agreement and was current as of May 23, 2011. Limitations of Water-Rights Data The information reflected in this dataset is derived by interpretations of paper records by OWRD. The user must refer to the actual water-right records for details on any water right. Care was taken by OWRD in the creation of the dataset but it is provided "as is." The USGS and the OWRD can not accept any responsibility for errors, omission, or accuracy of the information. There are no warranties, expressed or implied, including the warranty of merchantability or fitness for a particular purpose, accompanying this information (Oregon Water Resources Department (2012b). The data from the OWRD Unadjudicated Claims geodatabase (Oregon Water Resources Department, 2012b; D. Mortenson, Oregon Water Resources Department, written commun., 2011) are based on claims as originally filed by claimants in the Klamath Basin Adjudication. The OWRD provides no warranty or guarantee as to the accuracy of the information presented within these data, and is not intended to express a position on the nature or validity of any claim. Any information contained herein does not reflect any recommendation or final determination by the OWRD of the relative water rights in the Klamath Basin. The OWRD datasets may not reflect actual water use or recent changes in land or water use as can sometimes be observed by comparison with the Landsat images or evapotranspiration mapping. A partial list of the reasons for this include (1) the underlying OWRD dataset needing updating, (2) water-right holders not submitting a change of use or transfer of existing water rights, (3) water-rights data may not reflect land-use changes subsequent to the initiation of the water right, (4) water not being diverted to POUs based on Claims that have not yet been approved, (5) POU in the source OWRD database not reflecting recent findings of the adjudication of water rights in the Upper Klamath basin, (6) claimed POUs that OWRD has denied, (7) possible abandoned water rights, (8) claim/water right overlaps, (9) water rights not being utilized during a particular year, or (10) areas irrigated with groundwater or both surface water and groundwater. In the area of the Wood River Valley, there are a number of irrigation water-rights POU polygons missing from the OWRD dataset because the rights have been leased for instream use. In the past, OWRD has removed irrigation water rights with instream leases from the publicly available GIS water-rights geodatabase. The current practice, however, is to provide information regarding these leased water rights to the public. This practice was in place on July 18, 2011, when the GIS water-rights geodatabase was acquired from OWRD. However, most leased water rights were not included in the July 18, 2011 data acquisition and subsequently are not included in this report and associated maps. OWRD has indicated that the omission of these water rights was unintentional and that they are working to correct the dataset; the updated information was not available at the time this report was prepared. Subbasin Streamflow Statistics Importance and Relevance Streamflow statistics were computed for 72 subbasins in the Off-Project Water Program area and adjacent areas and include annual flow durations (5-, 10-, 25-, 50-, and 95-percent exceedances) and 7-day, 10-year (7Q10) and 7-day, 2-year (7Q2) low flows. Streamflow statistics were computed using regional regression equations based on historical unregulated streamflow data; the statistics represent estimated natural flow conditions in the subbasins as though irrigation diversions did not exist. The statistics were computed for the purpose of providing decisionmakers with the ability to estimate streamflow that would be expected after water conservation techniques have been implemented or a water use has been retired. Data Sources The streamflow statistics were computed using regional regression equations presented in Risley and others (2008). Although that report contains regression equations applicable for all of Oregon, equations used for this study were created from the Region 8 subset of 25 streamflow gaging stations in south-central Oregon. For the regression equations, computed annual flow statistics based on the daily mean streamflow records at the gaging stations were used as the dependent variables. Basin characteristics (such as drainage area and mean annual precipitation) of the drainage areas upstream of the gaging stations were the independent (explanatory) variables in the equations. The equations relating dependent and independent variables were computed using time periods when streamflow was unregulated. For some of the streamflow records, estimated irrigation water use was added to the record so that the record would reflect more natural conditions. Details on the procedure used to adjust the records for irrigation water use are provided in Risley and others (2008, p. 8, 10). A total of 7 equations were used to compute the annual flow statistics: 5-, 10-, 25-, 50-, and 95-percent exceedances, and 7-day, 10-year (7Q10) and 7-day, 2-year (7Q2) low flows. Basin characteristics used to create the equations were computed using a geographic information system (GIS) and various data layers. Descriptions for all data layers are documented in Risley and others (2008, table 5). Methods For this study, the Off-Project Water Program area and adjacent areas were divided into 72 subbasins. Preliminary subbasins were delineated on the basis of the locations of the pour points (referring to the outlet of the contributing drainage basin) for Hydrologic Unit Code (HUC) Level 6 (12-digit) classification of drainage basins from the 1:24,000 Watershed Boundary Dataset from the USDA Geospatial Data Gateway (Natural Resources Conservation Service, 2010). However, locations of the pour points for some subbasins were manually delineated on the basis of their proximity to streamflow gages or other criteria thought to be useful for the study. Final delineation of the subbasins was accomplished for each of the 72 pour points using StreamStats for Oregon (U.S. Geological Survey, 2010a), a Web-based GIS tool developed by the USGS (Ries and others, 2008). StreamStats also calculates the basin characteristics required to estimate the streamflow statistics using the Region 8 regression equations from Risley and others (2008, table 5). The calculation of the streamflow statistics using the Region 8 regression equations from Risley and others (2008, table 14) were performed in a Microsoft Excel spreadsheet. The calculations also can be performed using the USGS National Streamflow Statistics (NSS) Program (U.S. Geological Survey, 2012). For the NSS Program, the following settings must be used: Options / Analysis Type / Other; State / Oregon; Rural / New / LowFlow_Ann_Region08_2008_5126. The basin characteristics that are used as the independent variables in the regression equations to compute each of the 7 annual statistics: 5-, 10-, 25-, 50-, and 95-percent exceedances, and 7-day, 10-year (7Q10) and 7-day, 2-year (7Q2) low flows, consist of drainage area (in square miles) and mean annual precipitation (in inches) (Risley and others, 2008, table 5). Details about and the regression equations used to compute the annual flow statistics are provided in Risley and others (2008, table 14). As discussed in Risley and others (2008), to expand the number of available unregulated streamflow-gaging stations needed to create the regression equations, it was necessary to augment the daily-mean streamflow records for some stations with estimated monthly crop consumptive use. This procedure created records that were more representative of natural streamflow conditions. The procedure that was used to estimate consumptive use was developed by the Oregon Water Resources Department (Cooper, 2002). A discussion describing this procedure used also is provided in Risley and others (2008, p. 10). Upper and lower prediction intervals at the 90-percent confidence level for all 7 streamflow statistics (5-, 10-, 25-, 50-, and 95-percent exceedances, and 7Q2 and 7Q10 low flows) for the 72 basins included in the study were computed using the NSS Program (U.S. Geological Survey, 2012). Prediction intervals represent the probability that the true value of the characteristic will fall within the margin of error. For example, a prediction error at the 90-percent confidence level means there is a 90-percent chance the true value of the characteristic will fall within the margin of error. Details about and the equations used to compute the prediction intervals are provided in Risley and others (2008, p. 16). Prediction intervals are not calculated for basins if the value of one or both of the basin characteristic values (drainage area and mean annual precipitation) for that basin is outside the range of the basin characteristic values from the set of gaging stations used to create the regression equations. For Region 8 regression equations, prediction intervals are not calculated for values of drainage area or mean annual precipitation outside the range of 18.32 to 1,591.12 mi2 or 13.9 to 80.2 in., respectively (Risley and others, 2008, table 17). Very few gaging stations with sufficient record were available in Region 8 for use in the regression analyses by Risley and others (2008, p. 17) for estimating streamflow statistics. As a result, for some of the 72 subbasins, the basin characteristics used in the regression equations had values of some variables outside of the range of values used in the development of the regression equations by Risley and others (2008). Typically if one or more of the independent variables in a multiple regression are outside the range of the dataset used to develop the regression equations, increased prediction error can be expected. Additionally, streams with substantial groundwater inflows or streams heavily influenced by wetland areas, such as occurs in some parts of the study area, may not be well represented in the analysis. These factors may contribute to increased uncertainty in the estimates of the streamflow statistics for the 72 subbasins presented in this study. Of the 10 sets of regional regression equations presented in Risley and others (2008) that cover Oregon, the Region 8 regression equations, which include the Upper Klamath Basin and south-central Oregon, have the highest prediction errors. The cause of the errors can be related to two main factors—limited unregulated daily-mean streamflow data and a complex groundwater system. For Region 8, records for only 15 gaging stations with a minimum of 10 years of unregulated streamflow data were available for creating regression equations for the 7 annual streamflow statistics (flow durations [5-, 10-, 25-, 50-, and 95-percent exceedances] and 7-day, 10-year [7Q10] and 7-day, 2-year [7Q2] low flows). Other regions of the State have a greater number of available unregulated streamflow records available for creating regression equations. For example, unregulated streamflow records for 59 gaging stations were available for creating regression equations in Region 3, in the Willamette River basin. As described in Gannett and others (2007), the regional groundwater-flow system in the Upper Klamath Basin is complex, substantial, and variable. “Transmissivity estimates range from 1,000 to 100,000 feet squared per day and compose a system of interconnected aquifers.” “Groundwater discharges to streams throughout the basin, and most streams have some component of groundwater (baseflow). Some streams [such as Wood River and Spring Creek] however, are predominately groundwater fed and have relatively constant flows throughout the year.” If a greater density and number of unregulated streamflow records for gaging stations were available for creating the Region 8 regression equations, the groundwater component of the region’s streamflow could have been more accurately modeled in the regression equations. That in turn would have reduced some of the uncertainty in the estimates of streamflow statistics for the 72 subbasins in the study area. Irrigation Return-Flow Indicators Description Irrigation-return flow is defined herein as unconsumed irrigation water that returns to streams through subsurface flow. Often irrigation-return flow recharges the groundwater system, follows shallow flow paths, and discharges to an adjacent downgradient stream. However, depending on location and the groundwater hydrology, the irrigation-return flow may instead enter and flow through intermediate or even regional groundwater-flow paths bypassing adjacent streams and discharging to distant downgradient rivers or regional discharge areas. The travel time of irrigation-return flow from infiltration point to discharge point may be on the order of days to months for local groundwater-flow systems or from years to decades for intermediate and regional groundwater-flow systems. The greater the distance traveled by the irrigation-return flow, the more likely the discharge will be distributed more broadly spatially and temporally. Irrigation-return flow may result in higher water tables at the place of application or downgradient near discharge areas making it vulnerable to loss by subirrigation, which diminishes the potential return flow. Irrigation-return flow also is subject to loss due to groundwater pumping. The potential for, location, and timing of subsurface return flow of irrigation water for an agricultural area is typically best determined using a numerical flow model. The scale of modeling necessary to evaluate the OPWP, however, exceeded the resolution of the present regional flow model developed by the USGS for the Upper Klamath Basin (Gannett and others, 2012). As a consequence, it was not possible to make the necessary refinements to that model in the time allotted for this study. Instead, a more qualitative approach was used. Maps were developed using available information to show the relative potential for return flow in the study area. Data used as indicators for return-flow potential included depth to water, floodplain boundaries and features defined by stream geomorphology, and distance to surface-water features. Shallow depths to water are often indicative of proximity to a discharge area; infiltration of irrigation water in these areas may be expected to discharge to adjacent streams and to have short travel times. Geomorphic features of floodplains can be used to identify areas that are in close proximity of streams and that have soils conducive to the rapid infiltration of excess irrigation. The distance to the nearest surface-water feature can be used as a surrogate for travel time between infiltration of excess irrigation and discharge to a surface-water feature. Large distances can increase the likelihood that irrigation-return flow will enter intermediate or regional groundwater-flow systems, bypassing adjacent streams and not contributing to their flow. Large lakes, perennial streams, and streams known to be gaining flow from groundwater indicate interaction with the groundwater-flow system, as opposed to intermittent streams, which may only exist as a result of surface runoff. Map Descriptions Datasets for depth to water are described in the section, “Subirrigation Indicators.” Floodplain Boundaries and Features The dataset delineating floodplain boundaries and features for the Sprague River basin previously described in section, “Subirrigation Indicators,” also can be used as an indicator of irrigation-return flow. The geomorphic unit categories for the areas in and adjacent to floodplains from the Sprague River Oregon Geomorphology dataset (U.S. Geological Survey, 2011b) were assigned qualitative values for return flow potential (J.E. O’Connor, U.S. Geological Survey, written commun., 2011). Determination of low, medium, or high return-flow potential was made on the basis of the characteristics of areas from existing datasets and field observations of soils, vegetation, topography, and hydrology. As previously noted, some areas, including wetlands, springs, and ponds, were not mapped with the geomorphic floodplain and are not represented in the dataset. Distance to Surface-Water Features In this study, a GIS analysis was done to compute the distance between the point of interest and the nearest surface-water features. The assumption made is that the greater the distance from the surface-water feature, the lower the likelihood that applied irrigation will appear as return flow at the stream or river in useful spatial and temporal scales. Two analyses were made using different sets of surface-water features. The first analysis calculated the distance from each point in the study area to the nearest perennial stream or perennial large lake or pond. The second analysis calculated the distance from each point in the study area to the nearest gaining (receiving groundwater discharge) stream (and downstream reaches) or perennial large lake or pond. Distance to Perennial Streams and Lakes Perennial streams, lakes, and ponds were selected from the National Hydrography Dataset (U.S. Geological Survey, 2010b). The dataset was further restricted to lakes and ponds greater than 1 km2 in area. The horizontal distance between each point in the study area and the nearest surface-water feature was then calculated using a GIS. Distance to Gaining Streams and Lakes Gaining stream reaches were identified in the regional study of groundwater hydrology of the Upper Klamath Basin by Gannett and others (2007, p. 22–37; figure 7, p. 24; and table 6, p. 72–84). Stream reaches downstream of the gaining stream segments and large (greater than 1 km2) perennial lakes and ponds from the National Hydrography Dataset also were included. The horizontal distance between each point in the study area and the nearest of these surface-water features was then calculated using a GIS. Acknowledgments The authors thank the many people that contributed their time and knowledge to help complete this study. Dorothy Mortenson and Bob Harmon, Oregon Water Resources Department, provided water-rights data. Dani Watson, Ranch and Range Consulting, provided updates to some of the water-rights information. Chrysten Lambert and Shannon Peterson, Klamath Basin Rangeland Trust, assisted in defining and identifying instream leases in the Wood River basin. USGS employees whose efforts contributed to the study include: Esther Duggan, Charlie Cannon, Tess Harden, and Tana Haluska for their assistance with processing of the data; Jim O’Connor for his analysis of the geomorphology of the Sprague River basin; and Marshall Gannett for insights on the hydrology of the Upper Klamath Basin. References Cited Allen, R.G., Tasumi, Masahiro, and Trezza, Ricardo, 2007a, Satellite-based energy balance for mapping evapotranspiration with internalized calibration (METRIC)—Model: Journal of Irrigation and Drainage Engineering, v. 133, no. 4, p. 380–394, accessed June 27, 2012, at http://www.kimberly.uidaho.edu/water/papers/remote/ASCE_JIDE_Allen_et_al_METRIC_model_2007_QIR000380.pdf. Allen, R.G., Tasumi, Masahiro, Morse, A.T., Trezza, Ricardo, Wright, J.L., Bastiaanssen, Wim, Kramber, William, Lorite, Ignacio, and Robison, C.W., 2007b, Satellite-based energy balance for mapping evapotranspiration with internalized calibration (METRIC)—Applications: Journal of Irrigation and Drainage Engineering, v. 133, no. 4, p. 395–406, accessed June 27, 2012, at http://www.kimberly.uidaho.edu/water/papers/remote/ASCE_JIDE_Allen_et_al_METRIC_application2007_QIR000395.pdf. Cahoon, J.S., 1985, Soil survey of Klamath County, Oregon, southern part: U.S. Department of Agriculture Soil Conservation Service, 269 p., 106 soil map sheets, accessed June 27, 2012, at http://soildatamart.nrcs.usda.gov/Manuscripts/OR640/0/or640_text.pdf. Carpenter, K.D., Snyder, D.T., Duff, J.H., Triska, F.J., Lee, K.K., Avanzino, R.J., and Sobieszczyk, Steven, 2009, Hydrologic and water-quality conditions during restoration of the Wood River Wetland, Upper Klamath River Basin, Oregon, 2003–05: U.S. Geological Survey Scientific Investigations Report 2009–5004, 66 p. (Also available at http://pubs.usgs.gov/sir/2009/5004.) Cooper, R.M., 2002, Determining surface-water availability in Oregon: Oregon Water Resources Department Open-File Report SW 02-002, 157 p., accessed August 6, 2012, at http://cms.oregon.gov/owrd/SW/docs/SW02_002.pdf. Evapotranspiration, Plus, LLC, 2011a, Completion report on the production of evapotranspiration maps for year 2004 for the Upper Klamath and Sprague area of Oregon using Landsat Images and the METRIC model: Twin Falls, Idaho, March 2011, Revised March 28, 2011, 55 p., accessed June 27, 2012, at http://water.usgs.gov/GIS/dsdl/Report_KBRA_OPWP_ET_2004_ETplus.pdf. Evapotranspiration, Plus, LLC, 2011b, Completion report on the production of evapotranspiration maps for year 2006, Landsat path 45 covering the Upper Klamath and Sprague area of Oregon using Landsat Images and the METRIC model: Twin Falls, Idaho, May 2011, 64 p., accessed June 27, 2012, at http://water.usgs.gov/GIS/dsdl/Report_KBRA_OPWP_ET_2006_ETplus.pdf. Evapotranspiration, Plus, LLC, 2011c, Production of evapotranspiration maps for years 2004 and 2006 for Landsat Path 44 covering the Upper Sprague River area of Oregon using Landsat images and vegetation indices: Twin Falls, Idaho, May 2011, revised September 8, 2011, 7 p., accessed June 27, 2012, at http://water.usgs.gov/GIS/dsdl/Report_KBRA_OPWP_ET_path44_2004_2006_ETplus.pdf. Gannett, M.W., Lite, K.E., Jr., La Marche, J.L., Fisher, B.J., and Polette, D.J., 2007, Ground-water hydrology of the upper Klamath Basin, Oregon and California: U.S. Geological Survey Scientific Investigations Report 2007–5050, 84 p. (Also available at: http://pubs.usgs.gov/sir/2007/5050/.) Gannett, M.W., Wagner, B.J., and Lite, K.E., Jr., 2012, Groundwater simulation and management models for the upper Klamath Basin, Oregon and California: U.S. Geological Survey Scientific Investigations Report 2012–5062, 92 p. (Also available at: http://pubs.usgs.gov/sir/2012/5062/.) Hubbard, L.L., 1970, Water budget of Upper Klamath Lake southwestern Oregon: U.S. Geological Survey Hydrologic Atlas HA–351, 1 sheet. (Also available at: http://pubs.er.usgs.gov/publication/ha351.) Kann, Jacob, and Walker, W.W., Jr., 1999, Nutrient and hydrologic loading to Upper Klamath Lake, Oregon, 1991–1998: Prepared for Klamath Tribes Natural Resources Department and Bureau of Reclamation Cooperative Studies, Ashland, Oregon, Aquatic Ecosystem Sciences LLC, November 1999, 39 p. plus appendices, accessed June 27, 2012, at http://www.wwwalker.net/pdf/ulk_data_jk_ww_1999.pdf. Klamath Basin Restoration Agreement, 2010, Klamath basin restoration agreement for the sustainability of public and trust resources and affected communities: Yreka, California, KlamathRestoration.gov, February 18, 2010, 371 p., accessed June 27, 2012, at http://klamathrestoration.gov/sites/klamathrestoration.gov/files/Klamath-Agreements/Klamath-Basin-Restoration-Agreement-2-18-10signed.pdf. Natural Resources Conservation Service, 2009, Sprague River CEAP study report: USDA Natural Resources Conservation Service, Portland, Oregon, 100 p. Natural Resources Conservation Service, 2010, Geospatial Data Gateway: Website, accessed August 20, 2010, at http://datagateway.nrcs.usda.gov/. Oregon Water Resources Department, 2012a, GIS water right website, accessed August 20, 2012, at http://www.oregon.gov/owrd/Pages/maps/index.aspx. Oregon Water Resources Department, 2012b, Water Rights Information System (WRIS): Website, accessed September 3, 2012, at http://cms.oregon.gov/owrd/pages/wr/wris.aspx . Ries, K.G., III, Guthrie, J.G., Rea, A.H., Steeves, P.A., and Stewart, D.W., 2008, StreamStats—A water resources web application: U.S. Geological Survey Fact Sheet 2008–3067, 6 p. (Also available at http://pubs.er.usgs.gov/usgspubs/fs/fs20083067.) Risley, J.R., Stonewall, Adam, and Haluska, T.L., 2008, Estimating flow-duration and low-flow frequency statistics for unregulated streams in Oregon: U.S. Geological Survey Scientific Investigations Report 2008–5126, 22 p. (Also available at: http://pubs.usgs.gov/sir/2008/5126.) Soil Survey Staff, 2010, Soil survey geographic (SSURGO) database for Klamath County, Oregon, Survey area symbol–OR640, Survey area name-Klamath County, Oregon, southern part: United States Department of Agriculture, Natural Resources Conservation Service, accessed October 25, 2010, at http://soildatamart.nrcs.usda.gov. U.S. Geological Survey, 2010a, StreamStats for Oregon: accessed June 27, 2012, at http://water.usgs.gov/osw/streamstats/oregon.html. U.S. Geological Survey, 2010b, National hydrography dataset: accessed August 20, 2010, at http://nhd.usgs.gov. U.S. Geological Survey, 2010c, Water resources NSDI node: Website, accessed August 20, 2012, at http://water.usgs.gov/lookup/getgislist. U.S. Geological Survey, 2011a, Sprague River basin geomorphology: Website, accessed July 16, 2012, at http://or.water.usgs.gov/proj/Sprague/. U.S. Geological Survey, 2011b, Sprague River Oregon geomorphology—Metadata: accessed May 30, 2012, at http://water.usgs.gov/lookup/getspatial?sprague_river_oregon_geomorphology. U.S. Geological Survey, 2012, National Streamflow Statistics Program: Website, accessed August 20, 2012, at http://water.usgs.gov/osw/programs/nss/index.html. U.S. Government, 2012, Data.gov: Website, accessed August 20, 2012, at http://www.data.gov/. Watershed Sciences, LLC, 2000, Remote sensing survey of the Upper Klamath River basin—Thermal infrared and color videography, Final report prepared for the Oregon Department of Environmental Quality: Corvallis, Oregon, 387 p. plus 30 p. plus appendix, accessed June 27, 2012, at http://www.deq.state.or.us/wq/tmdls/docs/klamathbasin/flir/upklamath.pdf. Western Regional Climate Center, 2012, Cooperative climatological data summaries, NOAA cooperative stations—Temperature and precipitation, Oregon: accessed July, 15, 2012, at http://www.wrcc.dri.edu/summary/Climsmor.html. Appendix A. Access to Data, Metadata, and Example Illustrations The digital data, metadata, and example illustrations for the datasets described in this report are available on-line from the USGS Water Resources National Spatial Data Infrastructure (NSDI) Node Website (U.S. Geological Survey, 2010c), or from the U.S. Government website DATA.gov (2012). A Microsoft Excel workbook, listing each dataset and URL links to the website for the dataset, metadata, and example illustrations, is available at: http://pubs.usgs.gov/of/2012/1199/KBRA_OPWP_Appendix_A_datasets_v2.xlsx. The datasets are provided as Environmental Systems Research Institute, Inc. (ESRI) ArcMap file geodatabases or shapefiles or as ERDAS IMAGINE .IMG files. All data files have been compressed as .ZIP files. The metadata are provided as .XML (Extensible Markup Language) files. Instructions for accessing the metadata are provided in the section “Viewing Metadata” below. The example illustrations are in the form of Adobe® Systems .PDF (Portable Document Format) files. Viewing Metadata The metadata prepared for the datasets uses the FGDC XML (Federal Geographic Data Committee Extensible Markup Language) format. Suggestions for viewing metadata in FGDC XML format using ArcCatalog: For ArcGIS 10: 1. Navigate to the XML file in the catalog tree 2. Click on the “Description” tab 3. Scroll to the bottom and click “FGDC Metadata”. If this option is not present, change the metadata style (in Customize - ArcCatalog Options – Metadata) to “FGDC CSDGM Metadata” (where CSDGM stands for Content Standard for Digital Geospatial Metadata). For ArcGIS 9 1. Navigate to the XML file in the catalog tree 2. Click on the “Metadata” tab 3. Click “FGDC Metadata.” If this option is not present, change the metadata style (in Customize - ArcCatalog Options – Metadata) to “FGDC CSDGM Metadata.” It is also possible to view FGDC XML metadata using a web browser. Navigate to http://geo-nsdi.er.usgs.gov/validation/. After validation, the metadata may be viewed in a variety of formats. The “Questions and Answers” Output uses a “Plain Language” format that may be helpful to those unfamiliar with metadata. Alternatively, FGDC XML metadata may also be viewed using a web browser if the stylesheet “fgdc_classic.xsl” is present in the same directory as the XML file. The stylesheet is available from http://water.usgs.gov/GIS/metadata/usgswrd/XML/fgdc_classic.xsl. To download the file from the web browser use the File command and “Save As” with the filename “fgdc_classic.xsl” and place the file in the directory with the XML file. These products include a set of digital maps in GIS (ArcMap) format that can be used together (as overlays) to help evaluate the relative benefits of reducing or curtailing water use in various areas. The maps are not intended to drive the decision making process, but to inform it. It is envisioned that there will be many additional considerations affecting decisions. This dataset was developed as part of the study described in the following report: Snyder, D.T., Risley, J.C., and Haynes, J.V., 2012, Hydrological information products for the Off-Project Water Program of the Klamath Basin Restoration Agreement: U.S. Geological Survey Open-File Report 2012–1199, 17 p., http://pubs.usgs.gov/of/2012/1199. Background Information: Note that the report below is a text version of the complete report and is missing figures and equations. The complete report, which contains figures and equations, is distributed with this dataset as http://water.usgs.gov/GIS/dsdl/Report_KBRA_OPWP_ET_2006_ETplus.pdf. Completion Report on the Production of Evapotranspiration Maps for Year 2006, Landsat Path 45 Covering the Upper Klamath and Sprague area of Oregon using Landsat Images and the METRIC (TM) Model Report by Evapotranspiration, Plus 3496 N. 2500 E. Twin Falls, ID 83301 Growing Season ET for Irrigated Areas along the Upper Sprague River, 2006 Submitted to US Geological Survey Oregon Water Science Center Portland, OR May 2011 1. Introduction This report describes the procedures for and products from processing satellite, weather and land-use data for the Landsat World Reference System (WRS) Path 45 covering portions of south-central Oregon containing agricultural and mountain areas from near Crescent, Oregon south to the Oregon-California border and containing land areas in the upper Klamath and the Sprague River basins. The purpose of the application was the production of spatial and temporal maps of monthly and growing season evapotranspiration (ET) for the region for the years 2004 and 2006. In this second report, products for 2006 are reported. The first report, submitted in March 2011, described similar production for year 2004. The final products include 30 m resolution images of Actual Evapotranspiration (ET) and also images showing ET expressed as a fraction of Reference Crop ET (ETrF). ET was calculated at the same 30 m spatial resolution as the Landsat satellite images. Nine Landsat images were processed along Landsat WRS path 45 by combining portions of WRS rows 30 and 31 to produce estimates of monthly and growing season (April - October) ET. ET was obtained using the METRIC model developed by the University of Idaho. The METRIC procedure utilizes the visible, near-infrared and thermal infrared energy spectrum bands from Landsat satellite images and weather data to calculate ET on a pixel by pixel basis. Energy is partitioned into net incoming radiation (both solar and thermal), ground heat flux, sensible heat flux to the air and latent heat flux. The latent heat flux is calculated as the residual of the energy balance and represents the energy consumed by ET. The topography of the region was incorporated into METRIC via a digital elevation model (DEM), and used to account for impacts of slope and aspect on solar radiation absorption. METRIC was calibrated for each image using ground based meteorological information and identified 'anchor' conditions (the cold and hot pixels of METRIC) present in each image. A detailed description of METRIC can be found in Allen et al. (2007a,b; 2010). Work by the University of Idaho (UI) during this project included further development of the METRIC model to perform more accurately under the specific conditions of the study area. Specific enhancements included a new cloud gap filling procedure for ETrF (see footnote 1) images that allows the operator to adjust for background evaporation occurring from recent precipitation to better reflect total evaporation over longer (monthly) periods, the generation of gridded ETr maps used to estimate monthly and seasonal ET, improved computation of surface reflectance, albedo, terrain roughness and windspeed in mountainous areas to improve estimation of ET on sloped terrain. A description of the aerodynamic functions used in mountainous terrain is given in Appendix D. For Landsat 5 images, sharpening of the thermal band provided spatial refinement to the final ET products. FOOTNOTE: 1 ETrF is the fraction of alfalfa reference ETr as calculated by the standardized ASCE-EWRI Penman-Monteith equation (ASCE-EWRI, 2005) and represents the relative amount of reference ET occurring on any particular pixel of an image. ETrF is a direct product from METRIC. ETr is also used to calibrate the METRIC process and is calculated using hourly meteorological information from a weather station. Typical ranges for ETrF are 0 to about 1.1. ETrF is synonymous with the crop coefficient. Figure 1a shows the domain of the Landsat images processed by METRIC for years 2006 and 2006. The image is a 'false composite' of bands 2, 3 and 4, where 'green vegetation' shows as a red color. Forest vegetation in mountainous areas show as dark red and broadleaf vegetation, including agricultural crops generally shows as a lighter red color. Figure 1a. ##Not Shown## False color composite Landsat image of path 45, rows 30 and 31 corresponding to 07/25/2006 showing the study area processed by METRIC. Figure 1b shows an overlay on Landsat path 45, row 30 (southern portion) for the water basins of Williamson, Wood River/Upper Klamath Lake and Sprague, which are of interest to the USGS studies. The portion of row 31 of path 45 lying south of row 30, to the California state line, was added to the total area processed. That additional area, shown in Figure 1a, covers nearly all of the river basin domains shown in Figure 1b. The exception is a portion of the upper Sprague system that lies in path 44, east of path 45. That portion was estimated separately from METRIC using more simple vegetation index-based ET relationships that were derived from sampling of METRIC products from path 45. Landsat imagery was used in all cases. Figure 1b. ##Not Shown## Overlay of area of interest (purple lines) for ET processing and southern half of Landsat path 45, row 30 (courtesy of Daniel Snyder, USGS). 2. Image Selection and pre-processing For this application, images from Landsat 5 and Landsat 7 satellites were utilized due to their high resolution and presence of a thermal band. The image archive for Landsat 5 dates back to 1984 and the satellite is still in operation. Landsat 7 was launched in 1999. Landsat 7 images acquired after May 2003, although from a newer satellite than Landsat 5, are less preferred than Landsat 5, due to an anomaly with the Landsat 7 satellite caused by the malfunction of the scan line corrector (SLC). As a result, Landsat 7 images processed for years 2004 and 2006 are "SLC-off" images containing wedge shaped gaps extending from the edges of the image and stretching towards the centers. To obtain as complete coverage as possible, the gaps in ETrF maps produced by METRIC are generally filled in during post processing using the natural neighbor tool of Arc-GIS. The Landsat 7 images were only used during periods when Landsat 5 images were not available due to clouds. The most important criteria for the image selection is an assessment of cloud conditions at the time of the satellite overpass. The occurrence of conditions impeding the clearness of the atmosphere, such as clouds (including thin cirrus clouds and jet contrails), smoke, haze and similar over the study area may render parts of an image unusable for processing in METRIC. Even very thin cirrus clouds have a much lower surface temperature than the ground surface and because METRIC needs surface temperature estimates to solve the energy balance, areas with cloud cover cannot be used in the surface energy balance estimations. In addition, in cases of partial cloud cover, land areas recently shaded by clouds may be cooler as they have not yet reached a thermal equilibrium corresponding to the clear sky energy loading, and will also have to be masked out. A total of 9 Landsat image dates were selected for METRIC processing for year 2006. These dates are shown in Table 1. Table 1 - Dates of the Landsat satellite images used for METRIC processing in 2006. # Date Image Type 1 04/28/2006 Landsat 7 ETM+ 2 05/06/2006 Landsat 5 TM* 3 05/30/2006 Landsat 7 ETM+ 4 06/23/2006 Landsat 5 TM 5 07/09/2006 Landsat 5 TM 6 07/25/2006 Landsat 5 TM 7 08/26/2006 Landsat 5 TM 8 09/27/2006 Landsat 5 TM 9 10/29/2006 Landsat 5 TM Dem and Land Use maps used for METRIC processing To enable processing with METRIC, other basic input files are needed besides the satellite images. METRIC requires the use of DEM (Digital Elevation Model) and LU (Land Use) files as inputs. A digital elevation map (DEM) is used during METRIC processing to adjust surface temperatures for lapse effects caused by elevation variation. Maps of slope and aspect (aspect is the cardinal direction of an inclined surface) are also derived from the DEM at 30 m resolution and are used in estimating solar radiation on slopes. These images were created using the tools of the ERDAS Imagine processing system based on the DEM. A land use (LU) map was used to support the estimation of aerodynamic roughness and soil heat flux during METRIC processing. The NLCD (National Land Cover Database) Land Use map was obtained from the USGS-seamless webpage (http://seamless.usgs.gov/). The 30 m DEM was downloaded from the same website. 3. The METRIC Model METRICT (Mapping Evapotranspiration with high Resolution and Internalized Calibration) is an ERDAS coded model that bases the ET estimate on the evaluation of the energy balance at the earth's surface. METRICT processes instantaneous remotely-sensed digital and weather data and estimates the partitioning of energy into net incoming radiation, heat flux into the ground, sensible heat flux to the air, and latent heat flux. The latent heat flux, which is computed as a residual in the energy balance, represents the energy consumed by ET: LE = Rn - G - H where LE=latent energy consumed by ET; Rn=net radiation; G=sensible heat flux conducted into the ground; and H=sensible heat flux convected to the air. One very strong advantage of using energy balance is that actual ET rather than potential ET based on amount of vegetation is computed so that reductions in ET caused by a shortage of soil moisture are captured. A disadvantage of the energy balance approach is in the complexity of calculations. In traditional applications of energy balance, the computation of LE is only as accurate as the summed estimates for Rn, G, and H. METRIC attempts to overcome this disadvantage by focusing the internal calibration on LE and with H used to absorb all intermediate estimation errors and biases. METRICT utilizes spectral raster images from the visible, near infrared, and thermal infrared energy spectrum to compute the energy balance on a pixel-by-pixel basis. In METRIC, Rn is computed from the satellite-measured narrow-band reflectance and surface temperature; G is estimated from Rn, surface temperature, sensible heat flux and vegetation indices; and H is estimated from surface temperature ranges, surface roughness, and wind speed using buoyancy corrections. Figure 2 shows a general schematic of the METRIC process. Figure 2. ##Not Shown## General schematics of the METRIC process. Calibration of METRIC METRIC version 2.0.5 was used for the UI processing, but with some modifications during 2010 and early 2011. The 2.0.5 version was released by the University of Idaho in January 2010. A detailed description of METRIC can be found in Allen et al. (2007a,b) and Allen (2008). The main focus for the processing was to generate estimates of ET from lands having agricultural production, so that METRIC was calibrated with primary focus on accurate estimation of ET from the agricultural areas. However, because the full Landsat images were processed, efforts were made to minimize uncertainty in ET estimates from other land cover types present within the image, including forests, riparian vegetation and rangeland. Calibration Philosophy. METRIC uses a vertical near surface-to-air difference, dT, to estimate sensible heat flux. Sensible heat flux (H) is the amount of heat that is convected from a surface into the air, thereby reducing the amount of available energy for evaporation. The dT function is modeled as linearly proportional to surface temperature and is defined using the properties of two user selected anchor pixels, the "cold" and the "hot" pixels, that represent the extreme conditions encountered within the image (a condition having nearly complete conversion of available energy into evapotranspiration and a condition having nearly zero conversion of available energy into evapotranspiration). The cold anchor pixel generally represents a fully vegetated and actively transpiring vegetation, while the hot anchor pixel represents a bare and dry or nearly dry agricultural soil with little or no vegetation. The selection of cold and hot anchor pixels by the user is described by Allen et al., (2007b) and Allen (2008). These pixels are generally selected from agricultural fields for consistency and to match assumptions made in the estimation of soil heat flux, for example, where that algorithm was developed for agricultural soils. The surface temperature used to estimate dT was 'delapsed' to account for differences in surface temperature occurring as a result of elevation differences. During the internal calibration of sensible heat flux in METRIC, a fraction of ETr, ETrF, is assigned to the hot and cold conditions. ETrF is equivalent to the crop coefficient (Kc) based on full-cover alfalfa as the reference crop. ETrF at the cold pixel is normally assigned a value of 1.05 (Allen et al., 2007a,b) unless vegetation cover is insufficient to support this assumption (for example, early in spring and during winter when full, robust vegetation cover is rare). The 1.05 assignment to ETrF is used to account for the variation in ET inherent within a large population of fully vegetated fields. Previous applications of METRIC and comparisons against lysimeter measurements of ET at Kimberly, Idaho show that the "nearly coldest", or wettest, agricultural fields having full vegetation cover tend have ET rates that are typically 5% higher than that of the alfalfa reference ETr. This is because, for a large population of fields, some fields may have a wet soil surface beneath the canopy, or the canopy may be wet from recent (sprinkler) irrigation or precipitation, that tend to increase the total ET rate to about 5% above ETr. In addition, when viewing a large population of fields containing full cover alfalfa, a specific subpopulation of fields will have somewhat wetter conditions and therefore slightly higher ET and slightly cooler temperature than the "mean" full cover condition represented by the alfalfa reference. When the METRIC image is calibrated using an ETrF of 1.05 at the cold pixel, sampling of ETrF over a large population of full cover, irrigated fields tends to produce, on average, an ETrF value of 1. The cold pixel is selected from a population of fields having full cover and relatively cold temperatures. Ideally, an alfalfa field is preferred for calibration, since the ASCE Penman-Monteith equation is calibrated to an alfalfa reference. However, Wright (1982) has shown that most agricultural crops, when at full cover, transpire at levels very similar to those of alfalfa. Therefore, the selected location for the cold pixel does not need to be alfalfa, but can be any pixel from within the interior of a fully vegetated, cool, field (crop type is generally unknown when applying METRIC). During calibration of METRIC via the assignment of ETrF values for the cold and hot pixel conditions, normally only a single weather station is utilized in the calibration. A single station is used during calibration for several reasons. One, the locations for the cold and hot conditions are selected as close as possible to the single calibration weather station (usually within 20 km) so that wind speed and reference ET from the station can be assumed to closely approximate that for the selected calibration pixels. The internal calibration of the sensible heat flux function within METRIC is tied to the wind speed occurring at the calibration locations. Secondly, the internal calibration of the sensible heat flux function within METRIC generally requires the use of the same wind speed as was used in its determination, throughout the image. Third, the assignment of the ETrF at the hot pixel is closely tied to any recent precipitation occurring at the calibration weather station. Fourth, the assignment of ETrF at the cold and hot pixel conditions and the application of the METRIC process to the image should create (if calibrated and applied correctly) an ETrF surface over the image that has general limits of 0 and 1, and that can be later applied to an ETr surface that may vary over the image. Special Calibration Cases. Table 2 summarizes locations, NDVI and ETrF values assigned for cold and hot pixel conditions. For 4/28/2006, 5/06/2006 and 5/30/2006, the daily surface soil water balance run using data from the Agency Lake weather station indicated residual evaporation from recent rain events. The water balance suggested values of ETrF = 0.22, 0.20 and 0.22, respectively, for the bare soil condition for the three dates. These values were used to represent the driest bare soil conditions in the image area surrounding the calibration weather station (Agency Lake) to adjust the calibration for the presence of the background evaporation. The other image dates were estimated from the daily soil water balance to have reached a relatively dry state, where only residual, diffusive evaporation, estimated at ETrF = 0.10, was occurring. The daily soil water balance is shown later in Figure 6. Table 2. ETrF values assigned to and locations (X, Y coordinates in UTM meters zone 10 WGS1984) for the hot and cold pixels for each image date. Date X Y NDVI ETrF 4/28/2006 cold 594420 4720470 0.705 0.95 hot 589710 4702770 0.124 0.22 5/06/2006 cold 592260 4719300 0.787 1.05 hot 589620 4702740 0.148 0.20 5/30/2006 cold 589740 4707060 0.763 1.05 hot 587070 4702980 0.150 0.22 6/23/2006 cold 585180 4701930 0.786 1.05 hot 590250 4679400 0.126 0.10 7/09/2006 cold 590010 4705650 0.800 1.05 hot 599010 4677600 0.149 0.10 7/25/2006 cold 587610 4706700 0.837 1.05 hot 589920 4711560 0.200 0.10 8/26/2006 cold 587640 4706790 0.836 1.05 hot 587760 4703640 0.139 0.10 9/27/2006 cold 597210 4690110 0.807 1.05 hot 588510 4703370 0.180 0.10 10/29/2006 cold 587190 4707360 0.774 1.05 hot 586830 4705860 0.143 0.10 4. Weather data processing METRIC utilizes alfalfa reference ET (i.e., ETr) as calculated by the American Society of Civil Engineers (ASCE) standardized Penman-Monteith equation (ASCE-EWRI 2005) for calibration of the energy balance process and to establish a daily soil water balance to estimate residual soil evaporation from bare soil following precipitation events (Allen et al., 2007a). The ETr is used as a means to 'anchor' the surface energy balance by representing the ET from locations having high levels of vegetation and cooler surface temperatures. Therefore, high quality estimates of ETr are needed, which, in turn, require high quality weather data. Therefore, before processing the satellite images, the quality and accuracy of the meteorological data were assessed. Hourly weather data time steps are needed to produce ETr for calibration of the METRIC energy balance estimation process at the time of the Landsat overpasses. The hourly ETr values are summed to daily totals to provide a basis for producing daily and monthly ET. ETr was calculated using the RefET software (version 3) of the University of Idaho (Allen, 2008). Quality Assessment and Quality Control of the Weather Data To apply METRIC, reference ET is calculated from weather data sets having the following parameters, plus some of these parameters are used in the METRIC calibration: * Wind speed (hourly average): for computation of sensible heat flux (wind speed at satellite overpass time is required) and reference evapotranspiration (ETr) with the REFET software. * Precipitation (24 hour): to evaluate evaporative soil moisture conditions at the satellite overpass time. * Dew point temperature (hourly average): for calculation of atmospheric transmissivity and instantaneous incident solar radiation (clear sky) at satellite overpass time. Also used for reference ET calculation. * Incident solar radiation (hourly average): for reference ET calculation * Air temperature (hourly average): maximum and minimum temperature for reference ET calculation. Before being used for these calculations, QA (Quality Assessment) and QC (Quality Control) procedures as recommended by ASCE-EWRI (2005) were applied to investigate the general quality of data. In the case of solar radiation, for example, measured values (hourly or daily) were compared to estimated clear sky solar radiation taken as the upper bound for measured. Sensor malfunctioning, calibration problems, low maintenance and other issues can lead measured values to have systematic bias. Such systematic errors can be corrected based on expected clear sky conditions. Adjustments are applied by means of appropriate coefficients. In Figure 4 good agreement between registered solar radiation (Rs) and theoretical clear-sky solar radiation (Rso) indicates appropriate calibration of the sensor at Agency Lake for the date shown. Figure 4. ##Not Shown## Solar radiation (Rs) plotted against theoretical clear-sky solar radiation (Rso). In Figure 5 a plot of hourly mean air temperature and dewpoint is shown for a 24-hour period. In agricultural settings one can expect the recorded minimum temperature to be close to the dewpoint temperature observed at the same time, as in the case in the figure shown for 8/26/2006. Figure 5. ##Not Shown## Air temperature and dew point temperature registered at AGENCY LAKE on 8/26/2006 5. Using a daily soil water balance model for METRIC calibration. A daily soil water balance was applied to the 2006 period using precipitation and ETr data from the Agency Lake weather station. The water balance estimates residual evaporation from a bare soil surface on each image date as shown in Figure 6. The soil water balance is based on the two-stage daily soil evaporation model of the United Nations Food and Agriculture Organization's Irrigation and Drainage Paper 56 (Allen et al., 1998). Fig. 6 shows a simulation of evaporation from the upper 0.125 m of soil at Agency Lake. During the drying cycle after a wetting event, a typical bare agricultural soil can be expected to continue to evaporate at a small rate beyond the first several weeks due to diffusion of liquid water and vapor from beneath the upper soil layer. This evaporation can continue at very low rates for several additional weeks, provided no new wetting events occur, especially from tilled soils that have a moderate amount of water stored within the soil profile. This is typical of agriculture. Figure 6. ##Not Shown## Daily ETrF for bare soil estimated from the soil water balance for 2006 using weather data from the Agency Lake weather station. 6. METRICT processing and results METRIC produces 30x30 m spatial resolution maps of both ETrF (Fraction of Reference Evapotranspiration) and actual ET. The main products produced by METRIC are: - Instantaneous ETrF and ET maps, at satellite time for every image. - Daily ETrF and ET maps, for every image. - Monthly ETrF and ET maps . - Seasonal ETrF and ET maps. Intermediate Products During the METRICT process, dimensionless vegetation indices (NDVI, LAI and NDWI), surface reflectance (albedo), and surface and DEM-delapsed temperature maps are created. NDVI (normalized difference vegetation index) and LAI (leaf area index) maps are used in METRICT as indicators of biomass and aerodynamic roughness, and as predictors of ratios of soil heat flux to net radiation or sensible heat flux. The LAI is defined as the total one-sided green leaf surface area per unit ground surface area. The typical range for LAI is zero to six, where zero represents bare soil and greater than four represents dense vegetation. LAI values above three represent "full cover" conditions, and generally imply maximum ET in well irrigated areas. ##NOTE: Using "p" to represent the letter Rho of the Greek alphabet in the following discussion.## NDVI is calculated as the relative difference in reflectance between the shortest near infrared band (band 4) and the red band (band 3), respectively: NDVI = (p4-p3)/(p4+p3) where p3 and p4 are the at-satellite reflectances in bands 3 and 4 respectively. NDVI is somewhat sensitive to the color of the soil, spectral bandwidth, and atmospheric attenuation. Typically, NDVI varies between 0.1 and 0.8, with the higher value indicating dense vegetation and values less than about 0.2 associated with soil/rocks. Negative NDVI values typically indicate water bodies and snow, which reflect more energy in the red spectrum than in the near infrared. NDWI (normalized difference water index) is calculated as the relative difference at satellite reflectance between bands 5 and 2 NDWI = (p5-p2)/(p5+p2) where p5 and p2 are the at-satellite reflectances in bands 5 and 2 respectively. This is an index defined for the identification of water bodies. A value lower than zero indicates the presence of water bodies. In combination with NDVI, NDWI produces a good map of watery areas. Lapse rate In METRIC, the simulation of DEM delapsed temperature is necessary for estimating the near surface temperature gradient (dT) used to estimate sensible heat flux. This requires the establishment of an atmospheric lapse rate. For the area of study a unique lapse rate was used on each image date for elevations less than 1750 m, to represent lapsing trends along the agricultural valleys inside the image; this lapse rate is called the "flat" lapse rate during METRIC processing. Another lapse rate was used for elevations greater than 1750 m that represents mountainous conditions; this one is called the "mountain" flat rate. Unique values were sometimes required for specific images, determined by operator observation of surface temperature trends. Common (standard) values for the lapse rates are 6.5 K/1000 m for the 'flat' rate and 10 K/1000 m for the 'mountain' rate where K is degrees Kelvin. Refinements to the METRIC Mountain Model The METRIC model gives special treatment to mountainous areas and areas of other steep terrain during the computation of solar radiation inputs, where the influences of slope and aspect on energy inputs are calculated, and during the computation of convective heat exchange (H), where influences of slope and terrain roughness on estimated wind speed and aerodynamic transport are estimated. During the application to the Klamath region, additional refinements were made to both reflectance estimation (that uses solar radiation estimates as inputs) and to aerodynamic components. These refinements improved the behavior of the algorithms for both north and south facing steep slopes along the Cascade Range. The refinements to aerodynamics are described in Appendix D. Refinements to reflectance calculations were to parse solar radiation, by band, into beam, diffuse and terrain reflectance components during estimation of slope and aspect effects, and then reassembling the components prior to calculating reflectances. The result was improved estimation of reflectances on steep, north-facing slopes. Sharpening Although the final products from METRIC are of high spatial quality when produced from Landsat imagery, an even finer resolution for the images is often desirable, especially when ET within individual field parcels is needed. Landsat 5 images have 120 m spatial resolution of longwave (thermal) band that is coarser than the 30 m for coincident shortwave bands, and the 120 m thermal information tends to dominate the resolution of the final ET product. To improve the quality of the results, a procedure known as sharpening was applied to the final individual ETrF images generated with the METRIC code. This procedure is described in the METRIC manual (Allen et al., 2010) and in a paper by Trezza et al. (2008). The basic sharpening philosophy and procedure followed is based on the application of an established Surface Temperature (Ts) vs NDVI relationship to produce a first estimate of Ts at every short wave pixel, assuming a linear relationship and correspondence between NDVI and Ts. Later, to preserve original Ts information, this first estimate of Ts is adjusted so that Ts averaged over all shortwave pixels lying within an original thermal pixel matches the original average Ts of that thermal pixel. In most of the cases the redistribution of the bias between the original thermal Ts and the estimate Ts is an iterative process. Figure 7 shows an example of an ETrF map for a Landsat scene from 2004, before and after sharpening surface temperature. This procedure was applied to all Landsat 5 images to enhance the resolution of the final ETrF product. Landsat 7 images were not sharpened because they are already at 60 m resolution. Figure 7. ##Not Shown## Left: Close-up of ETrF image from path 45 corresponding to June 17th 2004; the area is close to Christmas Valley, OR. Right: The same ETrF map but using sharpened surface temperature. Gapfilling for Landsat 7 images Landsat 7 images acquired after May 2003 have information gaps caused by the malfunction of the scan line corrector. As a result, Landsat 7 images processed for year 2004 and 2006 are "SLC-off" images where wedge shaped gaps exist in the images, extending from the edges of the image and stretching towards the centers. To obtain as complete coverage as possible, the gaps in ETrF maps produced by METRIC were filled in during post processing using the natural neighbor tool of Arc-GIS. Figure 8 shows a close-up of an area along the Sprague River from 2004, where the natural neighbor interpolation procedure was applied. The quality of the interpolation depends on the location of the gap, being better over homogenous landscapes. Figure 8. ##Not Shown## Left: Close-up of ETrF image corresponding to July 11th 2004, showing gaps (stripes) originated from the Landsat 7 image; the area is close to Sprague River. Right: The same ETrF map, after gaps were filled using natural neighbor interpolation. Daily ETrF products METRIC was applied for every image included in Table 1 to obtain instantaneous (at satellite) and daily ETrF maps. As previously described, a total of 9 images were processed (Table 1). Maps of reflectance of short wave radiation, vegetation indices (NDVI and LAI), surface temperature, net radiation and soil heat flux were generated as intermediate products during METRIC processing. The final output from the METRIC energy balance model were images showing instantaneous ETrF (fraction of alfalfa based reference ET, ETr) at the satellite overpass time. For land covers other than rangeland, the estimate of daily ETrF was set equal to the instantaneous at the satellite overpass time, based on extensive ET measurements made using precision weighing lysimeters at Kimberly, Idaho (Allen et al., 2007b; Allen, 2008). The following section presents a view of each instantaneous ETrF image, with some comments in the figure captions (figures 9 - 17). Figure 9. ##Not Shown## ETrF map for 04/28/2006. Masked cloudy areas are identified as black. The image shows the product after filling the gaps in the Landsat 7 image. Snow-covered areas are shown as a turquoise, where an ETrF value of 0.5 was used to approximate the sublimation from snow. Mountainous areas were relatively 'wet,' with lower-lying agricultural areas more dry. Figure 10. ##Not Shown## ETrF map for 05/06/2006. Masked cloudy areas are identified as black. This image was relatively wet in the mountainous regions, with drier conditions in agricultural areas. Figure 11. ##Not Shown## ETrF map for 05/30/2006. Masked cloudy areas are identified as black. The image shows the product after filling the gaps from the Landsat 7 image. Figure 12. ##Not Shown## ETrF map for 06/23/2006. Masked cloudy areas are identified as black. Lower lying areas that are not irrigated have dried considerably. Transpiration is shown to remain strong in mountainous regions. Figure 13. ##Not Shown## ETrF map for 07/09/2006. Masked cloudy areas are identified as black. Figure 14. ##Not Shown## ETrF map for 07/25/2006. Masked cloudy areas are identified as black. Redish spots are negative ETrF values computed during the energy balance process, usually due to complexities in terrain. Negative values are set to 0 during the splining of monthly ET. Figure 15. ##Not Shown## ETrF map for 08/26/2006. Masked cloudy areas are identified as black. Figure 16. ##Not Shown## ETrF map for 09/27/2006. Masked cloudy areas are identified as black. Figure 17 ##Not Shown## ETrF map for 10/29/2006. Masked cloudy areas are identified as black. Mountainous regions showed very high values for ETrF. Some of this was caused by recent rain events in mountain areas. Some of the high values may have been caused by artifacts associated with very cold surface temperatures in mountains due to low sun angles and relatively short time between sunrise and satellite overpass (about 1100 hours) for surface warming. These artifacts are interpreted in the METRIC process as indication of evaporative cooling. The period following the October 29 image was very wet (see Figure 6) so that these high ETrF values in the mountains may be representative of conditions during that period. In addition, total ET, represented by reference ET (and based on weather) is relatively low during October-November, so that error in ETrF during this period has less impact on total growing season ET. The resulting monthly ETrF product (Figure 26) had lower values for ETrF due to influence of the September 27 image. Monthly ET and ETrF Individual satellite images are processed using METRIC and yield daily maps of ETrF for the image dates only. ETrF changes with time between images as vegetation develops or matures or as surface water availability varies. Because the objective of METRIC applications is to produce monthly and seasonal ET based on the information provided by the individual images, ETrF information from individual satellite image dates is interpolated between image dates to follow the trends caused by vegetation development and evaporation from precipitation. These interpolated, daily ETrF values are then multiplied by daily reference ET for each day to account for impacts of weather on potential ET demand. These products (of ETrF x ETr) are then summed over monthly periods to produce monthly ET. Cubic spline interpolation of ETrF values between satellite dates METRIC uses a cubic spline interpolation method to describe a smoothed variation in ETrF between images. This methodology was found to work better than a simple linear interpolation. For illustration of the cubic spline interpolation method, the figure 18 below shows an example (from another region) of point values of ETrF sampled from a single pixel from multiple images processed using METRIC. In this figure, values for each image date are connected using linear line segments between image dates. Figure 18. ##Not Shown## Interpolated ETrF using linear interpolation between images dates. Relatively abrupt changes in slope occur between dates. Figure 19 shows the application of a spline interpolation method for the same image dates. This smoother interpolation is in most cases a better representation the development of ETrF for vegetation compared to the linear interpolation. Figure 19. ##Not Shown## Interpolated ETrF using cubic spline interpolation between images dates. The application of the cubic spline procedure to derive monthly and seasonal ETrF and ET is applied one month at a time. Once the daily images for ETrF for each day of the month are created, for each day, the ETrF for every pixel in an image is multiplied by the reference ET (ETr), computed for each specific day according to weather data: ETdaily= ETrFdaily x ETr daily Following the computation of daily ET for each day of the month, the ETdaily was summed to produce ETmonth. The average monthly ETrFmonth was then determined by dividing the ETmonth by the summed ETr month: ETrFmonth = ETmonth / ETr month Because ETr can change spatially within an image domain, an inverse distance interpolation procedure of Arc-GIS with standard default parameters was used to produce a daily ETr surface using twelve Agrimet weather stations to create daily maps of ETr. The resolution of the daily ETr images was coarser than that for ETrF, since ETr changes only gradually in space. Location information for the Agrimet stations is listed in Table 3. Description of some stations is provided in Appendix A. Once ET and ETrF images for all months (April through October) were produced, the same concept as above was applied for the generation of the seasonal images, by summing the monthly ET and dividing by summed ETr to generate the average seasonal ETrF. As such the final generated products were the monthly ET and ETrF images from April through October and the seasonal total ET and average ETrF images for both considered paths. All ETrF images were generated as a Float Single Data Type and the ET images were generated as 16-bit Signed data type previously rounded to avoid data truncation. Table 3. Agrimet stations used to calculate daily reference ET during 2004 and 2006, including daily ETr surfaces for use during splining and integrating METRIC ET over monthly periods. Station State Lat.Dec. Long.Dec Lat.DMS Long.DMS Elevation, ft Christmas_Valley OR 43.24139 120.728 43° 14' 29" 120° 43' 41" 4305 Agency_Lake OR 42.56528 121.983 42° 33' 55" 121° 58' 57" 4150 Beatty OR 42.47806 121.274 42° 28' 41" 121° 16' 26" 4320 Lakeview OR 42.12222 120.523 42° 07' 20" 120° 31' 23" 4770 Lorella OR 42.07778 121.224 42° 04' 40" 121° 13' 27" 4160 Klamath_Falls OR 42.16472 121.755 42° 09' 53" 121° 45' 18" 4100 Worden OR 42.0125 121.788 42° 00' 45" 121° 47' 15" 4080 Medford OR 42.33111 122.938 42° 19' 52" 122° 56' 16" 1340 Cedarville CA 41.58528 120.171 41° 35' 07" 120° 10' 17" 4600 Powell_Butte OR 44.24833 120.95 44° 14' 54" 120° 56' 59" 3200 Hills_Creek_Dam OR 43.70972 122.421 43° 42' 35" 122° 25' 17" 1560 Lookout_Point_Dam OR 43.91556 122.752 43° 54' 56" 122° 45' 08" 940 Dealing with clouded parts of images Satellite images often have clouds in portions of the images, and the Path 45 images of Oregon for years 2004 and 2006 were no exception. ETrF cannot be directly estimated for clouded areas using surface energy balance because cloud temperature masks surface temperature and cloud albedo masks surface albedo. ETrF for clouded areas must be filled in before splining of monthly ET. Because clouded (or 'missing') portions of an image generally result in long periods between valid ETrF data (sometimes longer than several months), a special cloud-filling procedure was used. ETrF for cloud masked areas is filled in for individual Landsat dates prior to splining ETrF between images. The ETrF data inserted into masked areas are 'borrowed' from adjacent images in time, but with adjustment for background evaporation occurring from precipitation events, and, in some cases, adjusting total ETrF to account for substantial changes in image-wide vegetation amounts, for example during early spring. The adjustment for background evaporation is made in proportion to the amount of exposed bare soil in a pixel. The latter is estimated in proportion to 1/NDVI, where NDVI is the normalized difference vegetation index. The impact of the adjustment for background evaporation during cloud-filling is more seamless agreement between filled and adjacent nonfilled areas. The cloud mask-gap filling and interpolation of ET between image dates entails interpolating the ETrF for the missing area from the previous and following images. An ERDAS Imagine Modelmaker code was created by the University of Idaho METRIC group to conduct the 'filling' of cloud masked portions of images. The procedure is explained in details in Appendix 19 of the METRIC manual (Allen et al, 2010). Results of monthly ETrF maps Figures 20 to 26 show monthly ETrF maps for the period between April and October 2006 that were produced using cloud-filled images for individual dates and splining ETrF between dates. Figure 20. ##Not Shown## Average ETrF map for April 2006 Figure 21. ##Not Shown## Average ETrF map for May 2006 Figure 22. ##Not Shown## Average ETrF map for June 2006 Figure 23. ##Not Shown## Average ETrF map for July 2006 Figure 24. ##Not Shown## Average ETrF map for August 2006 Figure 25. ##Not Shown## Average ETrF map for September 2006 Figure 26. ##Not Shown## Average ETrF map for October 2006 Seasonal ET and ETrF Seasonal ET from April to October 2006 was calculated by summing ET from each month. Finally an average ETrF map was generated by dividing the seasonal ET by the total ETr for the same period. The average seasonal ETrF map is shown in Figure 27. A close up of seasonal ETrF is shown in Figure 28 for the Sprague River area and in Figure 29 for the Klamath Falls area. Total ET from nonirrigated areas, as computed by the METRIC process, was generally in the range of annual precipitation. Table 4 is a summary of ranges of ET estimated from METRIC for nonirrigated areas near noted weather station locations. April-October 2006 ET is compared to precipitation from January-December, 2006. The annual period was summed to consider winter precipitation that may have been stored in soil and carried into the growing season. ET ranged considerably with land use type and aspect, as well as probable soil types and depths. Some error may exist in both growing season ET and precipitation, with the former occurring during temporal interpolation between satellite images and the latter occurring from spatial interpolation of point measurements. Table 4. Ranges of ET estimated from METRIC for nonirrigated areas near noted locations during 2006, where ET is likely to be from precipitation, only (no shallow ground-water, wetlands, etc) General ET Precipitation, range, mm during mm during January- Location April-October December ------------------------------------------------------------------------ Christmas Valley 100 - 240 170 Mountains in NW Image 600 - 900 --- Timbered areas west of Crater Lake 700 - 1000 --- West of Agency Lake 300 - 600 400 Upper Sprague basin 100 - 300 260 (Lorella) South of Klamath 70 - 400 310 Figure 27. ##Not Shown## Average seasonal ETrF map for the period between April to October 2006 Fig. 28. ##Not Shown## Close up of average seasonal ETrF for the period April-October in the Sprague River area for 2006. Fig. 29. ##Not Shown## Close up of average seasonal ETrF for the period April-October near Klamath Falls, 2006. Fig. 30. ##Not Shown## Number of clouded images as part of the 9-image-based seasonal ET product for 2006. Black = 0, magenta = 1, dark blue = 2, light blue = 3 (see legend below the image). 7. References Allen, R.G., 2008. REF-ET: Reference Evapotranspiration Calculation Software for FAO and ASCE Standardized Equations. University of Idaho, 82 pp. [http://www.kimberly.uidaho.edu/ref-et/index.html]. Contact author for updates. Allen, R.G., Pereira, L.S., Raes, D., Smith, M., 1998. Crop Evapotranspiration. Guidelines for computing crop water requirements. FAO Irrigation and Drainage Paper 56. FAO, Rome, 300 pp. http://www.kimberly.uidaho.edu/water/fao56/ Allen, R.G., Pereira, L.S., Smith, M., Raes, D., Wright, J.L., 2005. FAO-56 Dual Crop Coefficient Method for Estimating Evaporation from Soil and Application Extensions. J. Irrig. Drain. Engr., 131(1), 2-13. http://ddr.nal.usda.gov/bitstream/10113/39505/1/IND43693971.pdf Allen, R.G., Tasumi, M., Morse, A., Trezza, R., Wright, J.L., Bastiaanssen, W., Kramber, W., Lorite, I., Robison, C.W., 2007a. Satellite-Based Energy Balance for Mapping Evapotranspiration with Internalized Calibration (METRIC) - Applications. J. Irrig. Drain Engr., 133(4), 395-406. http://www.kimberly.uidaho.edu/water/papers/remote/ASCE_JIDE_Allen_et_al_METRIC_application2007_QIR000395.pdf Allen, R.G., Tasumi, M., Trezza, R., 2007b. Satellite-Based Energy Balance for Mapping Evapotranspiration with Internalized Calibration (METRIC) - Model. J. Irrig. Drain Engr., 133(4), 380-394. http://www.kimberly.uidaho.edu/water/papers/remote/ASCE_JIDE_Allen_et_al_METRIC_model_2007_application2007_QIR00380.pdf Allen, R.G., Tasumi, M., Trezza, R., Kjaersgaard, J.H., 2010. METRIC. Mapping Evapotranspiration at High Resolution. Applications Manual, V 2.0.4. University of Idaho. 166 pp. ASCE-EWRI, 2005. The ASCE Standardized Reference Evapotranspiration Equation. ASCE, Reston, Virginia. http://irrisoftet/downloads/literature/ASCE Standardized_Equation Jan 2005.pdf Trezza, R., Allen, R.G., Robison, C.W., Kramber, W.J., Kjaersgaard, J., Tasumi, M., Garcia, M., 2008. Enhanced Resolution of Evapotranspiration from Riparian Systems and Field Edges by Sharpening the Landsat Thermal Band. Paper presented at the 2008 World and Environmental Resources Congress of ASCE and EWRI, Honolulu, HI, May 12-16, 2008. Published on CD-ROM, ASCE. 12 p. Wright, J.L., 1982. New evapotranspiration crop coefficients. J. Irrig. Drain. Engr., 108(1), 57-74. http://eprints/nwisrl.ars.usda.gov/382/478.pdf Appendix A. Descriptions of weather stations within the project area On November 13 - 16, 2010, Rick Allen, Ricardo Trezza and Eric Kra toured the upper Klamath and Sprague River basins to review general land-use and agricultural production conditions and to review Agrimet weather stations. Of the 12 Agrimet stations used to estimate reference ETr (Table A-1 below), the Klamath Falls, Beatty, Lakeview, Lorella and Medford stations were visited. We were unable to reach the Agency Lake site due to locked access. Table A-1. Agrimet stations used to calculate daily reference ET during 2004 and 2006, including daily ETr surfaces for use during splining and integrating METRIC ET over monthly periods. Station State Lat.Dec. Long.Dec Lat.DMS Long.DMS Elevation, ft Christmas_Valley OR 43.24139 120.728 43° 14' 29" 120° 43' 41" 4305 Agency_Lake OR 42.56528 121.983 42° 33' 55" 121° 58' 57" 4150 Beatty OR 42.47806 121.274 42° 28' 41" 121° 16' 26" 4320 Lakeview OR 42.12222 120.523 42° 07' 20" 120° 31' 23" 4770 Lorella OR 42.07778 121.224 42° 04' 40" 121° 13' 27" 4160 Klamath_Falls OR 42.16472 121.755 42° 09' 53" 121° 45' 18" 4100 Worden OR 42.0125 121.788 42° 00' 45" 121° 47' 15" 4080 Medford OR 42.33111 122.938 42° 19' 52" 122° 56' 16" 1340 Cedarville CA 41.58528 120.171 41° 35' 07" 120° 10' 17" 4600 Powell_Butte OR 44.24833 120.95 44° 14' 54" 120° 56' 59" 3200 Hills_Creek_Dam OR 43.70972 122.421 43° 42' 35" 122° 25' 17" 1560 Lookout_Point_Dam OR 43.91556 122.752 43° 54' 56" 122° 45' 08" 940 Klamath Falls Agrimet The Klamath Falls Agrimet weather station is located at the Oregon State University Research Center south of Klamath Falls. The area is mostly agricultural with some residential and industrial development. Windbreaks to the south and west of the station may impact air flow at times, as might proximity of research buildings to the weather station. ##Image Not Shown## Klamath Falls Agrimet from the hiway, looking SW ##Image Not Shown## Closeup of Klamath Falls Agrimet looking SW. Lakeview Agrimet ##Image Not Shown## The Lakeview Agrimet station is located near two center pivots and north of an irrigated cemetery. The very local landcover is dry grass, however, fetch is predominately irrigated. ##Image Not Shown## Lakeview Agrimet looking NW ##Image Not Shown## Lakeview Agrimet looking NW, with Allen ##Image Not Shown## Closeup of Lakeview Agrimet station Medford Agrimet Station ##Image Not Shown## The Medford Agrimet station is located at the Oregon State University Research Center west of Medford. The area is partially agricultural with some residential and industrial development. The station itself is located just north of research buildings and just north of a small grapevine study. The area to the north and east is mostly open. The buildings to the south and the grapevines probably impact air flow at times. It would be helpful on all Agrimet stations if anemometers were set at 3 m height above ground rather than the current 2 m height. ##Image Not Shown## Medford Agrimet station with Trezza and Allen ##Image Not Shown## Medford Agrimet Station looking East ##Image Not Shown## Medford Agrimet Station looking North. Appendix B. Generation of Precipitation and Reference ETr surfaces Daily precipitation surfaces were created using precipitation (P) information from 45 COOP stations and 7 Agrimet stations located within and adjacent to the scene processed. Data for COOP stations were downloaded from the NOAA National Climatic Data Center (NCDC) web site. Agrimet data were obtained from the USBR Agrimet web site. A shapefile indicating the COOP and Agrimet station locations was created and is available. The locations of the stations for precipitation are shown in the following figure B1. An Inverse Distance Weighting function was used for the interpolation of P and ETr, which can create some discontinuities and some 'bulls eyes' around stations having higher or lower readings as compared to surrounding stations (figure B2). This is mostly an artifact from the interpolation method. Towards the center of the image is Crater Lake, where there is a COOP station on the mountain (Crater Lake is a lake inside a volcano) and the mountain receives substantially more P as compared to the surrounding areas. The ETr surfaces are based on 9 (2004) or 10 (2006) Agrimet weather stations. We used a spline interpolation for the ETr surfaces where we increased the tension setting in Arc to 10 to prevent the spline from increasing ETr beyond reasonable values for areas in between weather stations. The process created an ERDAS file for each day of 2004 and 2006, and stacks by month and the seasonal sum, all in units of mm. The COOP station P data were adjusted for the time-of-day (usually 7 am) of readings, so that if the precipitation was recorded before noon, we moved the data to the previous day, while if the precip was recorded after noon, we did not move it. For this reason, there is sometimes a one day 'shift' in precip between nearby stations, so that one station may have recorded say 10 mm one day and nothing on the next, while a neighboring station is the opposite. Later, in the process of adjusting the image date ETrF from METRIC for background evaporation, we therefore typically take the average of three days when estimating what the ETrF was at the satellite overpass date. The P surfaces have one file for each day of 2004 and 2006, and stacks by month and the seasonal sum, all in units of mm. At this point, the gridded precipitation data have not been used. They were assembled in case a gridded evaporation process model would have been needed to estimate total evaporation over monthly periods from bare soil conditions, to use to adjust Landsat images for background evaporation differences between image dates and surrounding monthly periods. However, review of ET data from METRIC did not indicate the need to make this adjustment. STATSGO soil maps for Oregon and California, and the derived water content at 15 bars and 1/3 bar, available water capacity and soil texture for Oregon in were assembled, but again, not required. Other input to the soil water balance model, including TEW (total evaporable water), REW (relative extractable water), De_initial and P_eff_to_D_initial, were computed. The following two figures (figs. B2 and B3) show gridded precipitation summed over January - December 2006 for an area slightly larger than the processed image area and gridded reference ET summed for the April - October 2006 period for the nearly the same area. Fig. B2. ##Not Shown## Interpolation of total January - December, 2006 precipitation over the study area, with orange at 200 mm to blue at 1800 mm. The cross hair is centered over Crater Lake. Fig. B3. ##Not Shown## Interpolation of total Alfalfa reference ETr from April to October, 2006 over the study area, with orange at 980 mm to blue at 1250 mm. The cross hair is centered over Crater Lake. Appendix C. Description of Products contained in this Drive for METRIC processing for Klamath, Oregon, Year 2006 All spatial data files are presented in ERDAS Imagine data format (*.img). This format is a raster format that can contain multiple layers. Data are generally in 'float' (real) value expressions (including all images using mm depths and all ETrF), but some are expressed in integer form (DEM). The Imagine formatted files are readily read by all modern Arc-GIS systems. For some images, including the monthly ET and ETrF files, the data have been 'colorized' to display in ERDAS imagine in color. This helps with visualization to the person viewing the data, but do not impact the data themselves. The colorization may not transfer into the Arc-GIS system. The colorization is viewed in ERDAS by opening the files in 'pseudo color' mode. Each "img" file is accompanied by an 'rrd' file that is generated by ERDAS to facilitate rapid zooming and statistics. The rrd files are not important to Arc-GIS usage. Primary Folder Klamath_2006_path_45: SubFolder: 2006_path_45_original_Landsat_images_trimmed_to_Klamath This folder contains the original Landsat images(8 bit digital numbers (0-255)) used during METRIC processing. Each image is comprised of seven layers, where Layer 1 = Landsat Band 1; Layer 2 = Landsat Band 2; Layer 3 = Landsat Band 3; Layer 4 = Landsat Band 4; Layer 5 = Landsat Band 5; Layer 6 = Landsat Band 6 (thermal band); Layer 7 = Landsat Band 7. Image Name Units Description l71045030_03120060428_klamath.img DN Landsat 7 image corresponding to 04/28/2006 l5045030_03120060506_klamath.img DN Landsat 5 image corresponding to 05/06/2006 l71045030_03120060530_klamath.img DN Landsat 7 image corresponding to 05/30/2006 l5045030_03120060623_klamath.img DN Landsat 5 image corresponding to 06/23/2006 l5045030_03120060709_klamath.img DN Landsat 5 image corresponding to 07/09/2006 l5045030_03120060725_klamath.img DN Landsat 5 image corresponding to 07/25/2006 l5045030_03120060826_klamath.img DN Landsat 5 image corresponding to 08/26/2006 l5045030_03120060927_klamath.img DN Landsat 5 image corresponding to 09/27/2006 l5045030_03120061029_klamath.img DN Landsat 5 image corresponding to 10/29/2006 The naming convention is "L1t" = level 1, terrain corrected, followed by MMDDYYYY for the date, followed by the path and center row, followed by the satellite type (Landsat 5 or 7). All of the images correspond to Landsat WRS path 45, comprised mainly of row 30, plus some portions of row 31 residing north of the Oregon-California state line, as shown in the following figure. Spatial resolution of pixels is 30 m for all bands. However, original resolution of Landsat 5 band 6 (the thermal band) was 120 m and of Landsat 7 was 60 m. These bands were resampled, however, using cubic convolution, by the USGS EROS data center prior to dissemination. ##Image Not Shown## The following table shows the seven bands and their wavelength range from each satellite. Band Original Landsat 5 Landsat 7 Number resolution (m) wavelength (µm) wavelength (µm) ------------------------------------------------------------------------------ B1 (blue) 30 0.452 - 0.518 0.452 - 0.514 B2 (green) 30 0.528 - 0.609 0.519 - 0.601 B3 (red) 30 0.626 - 0.693 0.631 - 0.692 B4 (NIR) 30 0.776 - 0.904 0.772 - 0.898 B5 (MIR) 30 1.567 - 1.784 1.547 - 1.748 B6 (thermal,TIR) 60 (LS7), 120 (LS5) 10.45 - 12.42 10.31 - 12.36 B7 (MIR) 30 2.097 - 2.349 2.065 - 2.346 B8 (panchromatic)* 10 (LS7 only) NA 0.515 - 0.896 *Not used for ET and ETrF map generation. SubFolder: 2006_path_45_landuse map This folder contains the landuse map used for METRIC processing. The map was derived from the USGS NLCD (National Land Cover Database) Land Use map, and it was downloaded from the USGS-seamless webpage (http://seamless.usgs.gov/). The NLCD map is primarily used during determination of aerodynamic roughness values. Image Name Description landuse_p45r30_31.img Land use map SubFolder: 2006_path_45_DEM This folder contains the DEM map used for METRIC processing. The map was downloaded from the USGS-seamless webpage (http://seamless.usgs.gov/) and has 30 m resolution. Image Name Description dem_combined_reproj.img Map of pixel elevation, in meters, combined for paths 44 and 45. Folder: 2006_path_45_cloudmasked_ETrF_on_image_date This folder contains the daily images produced from METRIC and represents the ET estimate for each Landsat image date. The pixel values represent the ratio (ETrF) between actual evapotranspiration (ET) and alfalfa-reference evapotranspiration (ETr). A value of ETrF =0.6 means that ET is 60% of ETr. "Black" areas in these images are areas that were 'cloud masked' to delete those areas that were impacted by cloud cover. Image Name Units Description: ETrF = ET/ ETr --------------------------------------------------------------------------------------------------------------- etrf24_cloudmasked_04282006_p45r30_l7_klamath.img fraction ETrF image corresponding to 04/28/2006 etrf24_cloudmasked_05062006_p45r30_l5_klamath_color.img fraction ETrF image corresponding to 05/06/2006 etrf24_cloudmasked_05302006_p45r30_l7_klamath.img fraction ETrF image corresponding to 05/30/2006 etrf24_cloudmasked_06232006_p45r30_l5_klamath_color.img fraction ETrF image corresponding to 06/23/2006 etrf24_cloudmasked_07092006_p45r30_l5_klamath_color.img fraction ETrF image corresponding to 07/09/2006 etrf24_cloudmasked_07252006_p45r30_l5_klamath_color.img fraction ETrF image corresponding to 07/25/2006 etrf24_cloudmasked_08262006_p45r30_l5_klamath_color.img fraction ETrF image corresponding to 08/26/2006 etrf24_cloudmasked_09272006_p45r30_l5_klamath_color.img fraction ETrF image corresponding to 09/27/2006 etrf24_cloudmasked_10292006_p45r30_l5_klamath_color.img fraction ETrF image corresponding to 10/29/2006 The naming convention is ETrF24 followed by the date expressed as MMDDYYYY, followed by the path row and satellite number. The image named number_cloud_masked_images_p45r30_klamath_2006.img shows the number of images that were clouded for any particular location during the 2006 growing season (see Figure 30 in the main text). Folder: 2006_path_45_monthly_et_maps This folder contains the monthly ET maps (in millimeters) for every month, from April to October 2006. For example, a pixel value = 120 means that a total 120 mm of ET was calculated for that particular pixel for that particular month. Image Name Units Description ------------------------------------------------------------------------------------ et_april2006_kl.img millimeters Total ET in millimeters for April 2006 et_may2006_kl.img millimeters Total ET in millimeters for May 2006 et_june2006_kl.img millimeters Total ET in millimeters for June 2006 et_july2006_kl.img millimeters Total ET in millimeters for July 2006 et_august2006_kl.img millimeters Total ET in millimeters for August 2006 et_september2006_kl.img millimeters Total ET in millimeters for September 2006 et_october2006_kl.img millimeters Total ET in millimeters for October 2006 Folder: 2006_path_45_monthly_etrf_maps This folder contains the average ETrF for every month, from April to October 2006. This average ETrF was obtained dividing the total ET by the total ETr for a particular month. The "monthly" average ETrF" was produced from the image date-specific ETrF by splining ETrF between image dates for each day between the images, multiplying by ETr of each day, summing the ET product over a month to produce the monthly ET in the previous table, and then dividing by monthly summed ETr to obtain a monthly average ETrF. ETr data were calculated using the University of Idaho REF-ET software using meteorological weather parameters from six Agrimet weather stations in the study area following extensive quality assessment/quality control. The standardized ASCE-EWRI (2005) Penman-Monteith reference ET equation for the tall alfalfa reference type was used. Image Name Units Description ----------------------------------------------------------------------------- etrf_april2006_kl_color.img fraction Average ETrF for April 2006 etrf_may2006_kl_color.img fraction Average ETrF for May 2006 etrf_june2006_kl_color.img fraction Average ETrF for June 2006 etrf_july2006_kl_color.img fraction Average ETrF for July 2006 etrf_august2006_kl_color.img fraction Average ETrF for August 2006 etrf_september2006_kl_color.img fraction Average ETrF for September 2006 etrf_october2006_kl_1.img fraction Average ETrF for October 2006 Folder: 2006_path_45_seasonal_et_maps This folder contains the total ET maps (in millimeters) for the period between April to October 2006 for path 45 (that are based on METRIC energy balance). Image Name Units Description total_et_april_october_2006_kl_color.img mm Total ET in millimeters for the period between April to October 2006 SubFolder: 2006_path_45_seasonal_etrf_maps This folder contains the average ETrF from April to October 2006 for path 45. This average ETrF was obtained by dividing the total seasonal ET (from April to October) by the total seasonal ETr (from April to October). Image Name Units Description average_etrf_april_october_2006_kl_color.img millimeters Total ET in millimeters for the period between April to October 2006 SubFolder: 2006_path_45_Cloud_filling_sources This folder contains on image for each ETrF date showing which image date (later in time) that ETrF data were taken from to fill gaps for the image being filled. Image Name Units Description etrf24_cloudfill_source_04282006_p45r30_l7_klamath.img --- Image no. later in time for data borrowing Folder: 2006_path_45_monthly_reference_etr_maps This folder contains the total alfalfa-refence evapotranspiration (ETr) maps for every month in 2006. ETr is expressed in millimeters. Reference ET results from REF-ET for 2006 were previously provided with the 2004 data sets. The standardized ASCE-EWRI (2005) Penman-Monteith reference ET equation for the tall alfalfa reference type was used. Appendix D. Algorithms for estimating aerodynamic roughness and wind speed in mountains Richard G. Allen and Ricardo Trezza This appendix describes algorithms developed during early 2011 for the University of Idaho METRIC Mountain model to improve estimation of sensible heat transfer, via the METRIC 'dT function' for mountainous regions. Form drag increases in mountains as a result of impacts of bluff body effects of hills and mountains and separation of flow. Usually, a wake region of chaotic flow where pressure deficiency exists occurs on the lees of mountains. Roughness lengths for such complex terrains are thus much larger than the typical vegetative canopy (Hansen 1993). Hansen summarized roughness lengths for steep mountainous areas to range from 0.7 to 3.0 m, whereas typical roughness for forest ranged from about 0.3 to 1.0 m. Garratt (1977) summarized data and modeling by Fiedler and Panofsky (1972) that suggested roughness of mountainous areas to be 2 to 3 times larger than for 'flat' systems. The literature does not have many suggestions on estimating aerodynamics over large areas of mountainous terrain. The following algorithms are relatively simple functions to increase or decrease wind speed according to noted, physical trends experienced on slopes and to increase aerodynamic roughness in complex terrain to follow general observations from literature. The basic forms are based on intuition of the writers and are intended to provide approximate adjustments. The nature and shape of the functions and the values for coefficients were determined by making multiple runs of METRIC over mountainous areas of the Klamath watershed in Landsat path 45, row 30 over various times of the year and noting the formulations that produced estimates of ET on a variety of slopes and aspects and for a range of vegetation amounts that were in line with precipitation inputs. Definitions: SDE3 standard deviation of elevation within a 3 km diameter circle containing the pixel at the center (m) RE3 relative elevation (0 - 1) of a pixel within a 3 km diameter circle S slope, degrees zom_flat zom for surface in flat terrain (standard calculation), m zom_terrain additional roughness caused by terrain, m zom_terrain_max maximum terrain roughness for SDE3 >= 200 m. aspect aspect of slope (0 N, 180 deg. South) aspectwind aspect of wind (direction that wind is coming from (180 = S)) Terrain Roughness Convective transport is increased with terrain roughness due to impacts of large scale mixing. The standard deviation of elevation (SDE) in a locality provides a good indication of the relative change in terrain elevation with distance and the associated increase in roughness and form drag. The function derived combines, additively, roughness due to vegetation and surface features and that of larger DEM scale roughness represented by the SDE. When not water and not agriculture, effective (adjusted) roughness is estimated as: ##Equation Not Shown## where ##Equation Not Shown## with zom_terrain_max = 3 m, and with Cz = 1.0 for full effect and = 0.5 for reduced effect (used to scale impact). The above equation for zom applies the terrain roughness increase with a baseline of 30% implementation if smooth terrain and then increasing the application according to the background (flat) roughness. The SDE is applied to a 3 km diameter circle of the 30 m DEM, with the pixel of interest at the center. The SDE3 is derived using a standard deviation tool in ERDAS. The sine function creates an "S-curve" shape for the function in terms of SDE, with maximum value for an SDE3 of 200 m and largest increase in the function at an SDE3 of 100 m. Wind Speed increase with elevation Wind speed is known to increase with elevated position on a slope due to convergence of flow lines and subsequent acceleration. Therefore, in sloped areas, and where SDE is significant to indicate the possibility of flow convergence, wind speed is increased as: IF SDE3 > 30 m then IF SDE3 < 50 m then ##Equation Not Shown## else ##Equation Not Shown## endif endif where Cu = 1.0 (coefficient). No adjustment is made when the SDE3 is less than 30 m, and is maximum when SDE3 is 50 m or more, in proportion to the relative elevation of the pixel on a slope. The relative position, RE3 is represented by the relative elevation of the point within a 3 km diameter circle. The adjustment doubles wind speed when at the highest point within a 3 km diameter when SDE3 is 50 m or greater. Wind Speed decrease from shielding In sloped areas, and where SDE is significant, wind speed is reduced on leeward sides of terrain due to sheltering of wind. We model this as: IF S > 5 degrees: (i.e., if on a significant slope) IF SDE3 > 30 m then (if in rough terrain) IF 0.1 < RE3 < 0.95 then (if on the midslope portion of a slope) ##Equation Not Shown## where ##Equation Not Shown## else ##Equation Not Shown## endif else ##Equation Not Shown## endif else ##Equation Not Shown## endif where aspect = aspect of slope (0 N, 180 deg. South) aspectwind = aspect of wind (direction that wind is coming from (180 = S)) Cs = scaling multiplier (= 4) Ca has a range of (-1 <= Ca <= 1). This adjustment only occurs on leeward slopes. No adjustment occurs on windward slopes. In application, the aspect of wind direction at Agency Lake was used as a starting point for aspectwind. However, because the Agrimet weather stations are in valleys, where wind direction may be shaped somewhat by valley orientation. In addition, wind over mountains may have a regionally influenced direction that is somewhat different from that of a local weather station. We recognize that wind direction in mountains can have substantial local influence as well, according to mountain shape and orientation, up and downslope flow of air due to density differences, and valley depths and orientations. The wind aspect can, however, be used as an effective tuning parameter where the values are varied to help ET estimations on various sloped aspects to conform with expected values. This was the case for the Klamath area applications. In most cases, the apparent air flow in mountainous areas was from south to north (180 degree aspect). Values are shown in Table D-1 and D-2 for years 2004 and 2006. The impact of the reduction in wind speed can be quite strong, reducing wind speed by about 4% per 1 degree of slope on leeward faces. Table D-1 - Values for wind aspect used to estimate areas of wind speed reduction due to shielding on leeward mountain slopes in 2004. Date Approximate Wind Aspect (direction) during the Satellite Wind Aspect used in Overpass Time, deg. METRIC, degrees Image no. Date from North from North --------------------------------------------------------------------------- 1 04/30/2004 150 150 2 06/01/2004 180 180 3 06/17/2004 shifting 180 4 07/11/2004 150 150 5 08/04/2004 150 150 6 08/20/2004 150 150 7 09/21/2004 100 180 8 10/07/2004 100 100 9 11/08/2004 100 180 Table D-2 - Values for wind aspect used to estimate areas of wind speed reduction due to shielding on leeward mountain slopes in 2006. Wind Aspect # Date (0 = N), deg. 1 04/28/2006 180 2 05/06/2006 180 3 05/30/2006 180 4 06/23/2006 180 5 07/09/2006 315 6 07/25/2006 180 7 08/26/2006 180 8 09/27/2006 180 9 10/29/2006 180 Monin-Obukov boost in Aerodynamic resistance calculation on windward slopes In sloped areas, and where SDE is significant, aerodynamic resistance, rah, is decreased on windward sides of terrain due to a boost to buoyancy-induced instability caused by the vertical component of the wind. To account for this boost: IF S > 5 degrees then (on a slope) IF SDE3 > 30 m then (in rough terrain): ##NOTE: Using "Y" to represent the letter Psi of the Greek alphabet for stability parameter in the following discussion.## IF stability parameter, Yz1, Yz2 or Y200 < 0 (unstable condition) then adjust each one as ##Equation Not Shown## (increase the instability boost) else IF stability parameter, Yz1, Yz2 or Y200 > 0 (stable condition) then adjust each one as ##Equation Not Shown## (decrease the stability retardation) endif endif endif endif Cpsi is a scaling coefficient = 1.0. Ca is the same calculation as for wind sheltering (see previous section) and is a function of aspect and wind direction (-1 <= Ca <= 1). This adjustment only occurs on windward slopes. No adjustment is made for leeward slopes. The adjustment is made for both momentum and heat transfer stability parameters and for both stable and unstable conditions. Other adjustments Roughness of snow zom for snow is computed by averaging zom for pure snow and zom for the terrain. This accounts for the impact of trees, etc. protruding from the snow. zom_snow_cover = (zom_snow + zom_LU)/2 ETrF limit for snow The METRIC energy balance uses surface temperature to estimate the near surface gradient, dT, used in the calculation of sensible heat flux, and that function implies a history of warming of the surface during early and mid-morning hours, as is characteristic of dry and evaporating surfaces. That characteristic warming is not present with snow and ice-covered surfaces, however, where surface temperature is constrained to less than or equal to 273 K. Therefore, low surface temperature of snow and ice, as opposed to nonfrozen, evaporating surfaces can cause the dT to be understated, with understatement of sensible heat flux, and therefore overstatement of ET. Therefore, in METRIC, if the estimated ETrFsnow > 0.5, then ETrFsnow is constrained to 0.5. In all likelihood, the ET is lower than this. Appendix E. Domain of the three river basins of interest to USGS (Williamson, Wood River/Upper Klamath, and Sprague) overlaying Landsat WRS path 45/30, 44/30, and 44/31 scenes. ##Image Not Shown## The three scene areas overlapped, as shown by the colors (grey, purple and tan). Green and yellow speckles on the overlay represent agricultural and other landuses of interest. Graphic is courtesy of Daniel Snyder, USGS, Portland. Production of Evapotranspiration Maps for Years 2004 and 2006 for Landsat Path 44 Covering the Upper Sprague River area of Oregon using Landsat Images and Vegetation Indices by Evapotranspiration, Plus 3496 N. 2500 E. Twin Falls, ID 83301 ##Cover Image Not Shown## Submitted to US Geological Survey Oregon Water Science Center Portland, OR May 2011 Revised September 8, 2011 by D.T. Snyder USGS Background The area of interest to the Evapotranspiration (ET) study of the Klamath and Sprague River systems of Oregon is outlined in the figure on the cover of this report (courtesy of Dan Snyder, USGS). A general outline of two of the areas of interest is shown in Figure 1. Most of the areas of interest lie in path 45 of the Landsat WRS coverage, as shown in the cover figure. A very small portion of the upper Sprague River basin lies to the east of path 45 and is covered by path 44, only. Portions of the Sprague basin lie in an overlap of both paths 45 and 44. The areas lying in path 45 were processed using the University of Idaho METRIC energy balance based procedure to produce actual ET. Parts of rows 30 and 31 were processed. Details for years 2004 and 2006 are provided in two other reports. Each Landsat path requires its own METRIC application and calibration because the images occur on different dates. Full METRIC applications for path 44 were not considered to be economical due to the small area. Instead, a relatively rapid, vegetation-index-based method for estimating fraction of reference ET, ETrF, was applied, where the coefficients for the method were based on data derived from the full METRIC application for path 45. Some adjustment to the general coefficients was made to account for background evaporation stemming from antecedent precipitation. Figure 1: ##Image Not Shown## Approximate area of interest in Klamath and Sprague River basins of south-central Oregon. ##NOTE: Using "p" to represent the letter Rho of the Greek alphabet for "at-satellite reflectances" in the following discussion.## Procedure The normalized difference vegetation index, NDVI, is computed as: NDVI = (pt,4 - pt,3) / (pt,4 + pt,3) (1) where: pt,3 and pt,4 are at-satellite reflectances for bands 3 (red) and 4 (near infrared) of Landsat 5 or 7. The NDVI is a sensitive indicator of the amount and condition of green vegetation. Theoretically, values for NDVI can range between -1 and +1. Green surfaces typically have NDVI values between about 0.2 and 0.85 for Landsat data, and water, snow and clouds usually have values less than zero. The NDVI determined from Eq. 1 used reflectance values that were based on 'top of atmosphere' radiance, for consistency with common practice. The computational steps were as follow: 1. METRIC submodels m001 and m01 were used to generate top of the atmosphere NDVI 2. From NDVI, baseline estimates for ETrF were calculated as: ETrF = 1.2 * NDVI when NDVI > 0 (2a) ETrF = 0.7 when NDVI < 0 (representing water bodies) (2b) Figure 2. ##Image Not Shown## Left: NDVI image for 07/12/2004. Right: ETrF image estimated as ETrF=1.2*NDVI for the same date. Water bodies (black spots in the NDVI image) were assigned a fixed value of ETrF=0.7. The 1.2 coefficient in Eq. 2a was based on ETrF vs. NDVI relationships observed from the path 45 images processed for years 2004 and 2006. The 1.2 multiplier, with 0.0 offset, represents ET conditions where the soil surface is relatively dry, so that the majority of ET is due to the presence of vegetation and transpiration. 3) ETrF images were cloudmasked. A value of -2 was assigned for clouds and Landsat 7 gaps. 4) ETrF images were cloudfilled. Figure 3. ##Image Not Shown## Left: masked ETrF image for 05/25/2004; black areas are clouds masked with a value of -2. Right image: The same image after ETrF in the cloudy areas was estimated using the METRIC cloud filling algorithm. 5). The METRIC splining model was used to generate monthly ETrF images for path 44. Figure 4. ##Image Not Shown## Mosaiced Image for August 2004. 6) Depending on the match of monthly ETrF for Path 44 that was derived from NDVI to monthly ETrF for path 45 that was derived from energy balance, ETrF for path 44 was adjusted to account for higher or lower background evaporation due to specific moisture conditions observed for the particular month and images on which the monthly ETrF were based. This was done as: ETrF_adjusted = a*(ETrF/1.2) + b (3) where a and b are coefficients determine by iterative, visual comparisons between the path 44 and path 45 products. Only classes 52 (shrub) and 71 (grassland) were adjusted, since these classes are the most subject to precipitation induced evaporation. The adjustment was most pronounced at pixels having low NDVI. A comparison between pre and post adjustment for monthly ET in April 2004 is shown in Figure 5, where the overlap between the two paths is quite pronounced before adjustment and less pronounced following adjustment. Figure 5. ##Image Not Shown## Close-up of the mosaiced ETrF image for April 2004. Left: Before adjustment; Right: After adjustment. 7) Final Path 44 monthly and growing season ETrF and ET images were mosaiced with Path 45 images using Arc-GIS. Moasiced images can be considered to be the final products and used as inputs to water balances. These image files are housed in the folder named Klamath_2004_2006_paths_44_45_Final_Mosaiced_Products that was delivered to the USGS. Figure 6. Mosaiced Path 45 and 44 areas processed for seasonal ETrF for 2004. Figure 7. ##Image Not Shown## Close-up of the path overlap area for the seasonal ETrF map for 2004. 2006 ground condition Complete None planned -123.425204 -120.483416 43.490807 41.972926 None inlandWaters Klamath Basin Restoration Agreement Evapotranspiration METRIC None Upper Klamath Basin Sprague River Basin Wood River Basin Williamson River Basin Oregon None Evapotranspiration, Plus, LLC of Twin Falls, Idaho under contract to the U.S. Geological Survey should be acknowledged as the data source in products derived from these data. Daniel T. Snyder U.S. Geological Survey Hydrologist mailing address

2130 SW 5th Ave

Portland OR 97201-4976 USA 503-251-3287 N/A 503-251-3470 dtsnyder@usgs.gov Monday - Friday 9 a.m. to 5 p.m. PDT (Warning: Although accurate at the time of production, this information may have become obsolete. See the Metadata_Reference_Information section for a current contact.) http://water.usgs.gov/GIS/browse/mosaic_et_august2006_kl_NAD83.pdf Illustration of data set Portable Document Format (PDF) This data set was created by Evapotranspiration, Plus using the METRIC model (Mapping Evapotranspiration with high Resolution and Internalized Calibration) developed by the University of Idaho. None Unclassified None Microsoft Windows XP Version 5.1 (Build 2600) Service Pack 3; ESRI ArcCatalog 9.3.1.4000 The METRIC model, used to create the maps of monthly and growing season evapotranspiration (ET), was calibrated for each image using ground based meteorological information and identified 'anchor' conditions (the cold and hot pixels of METRIC) present in each image. Work by the University of Idaho (UI) during this project included further development of the METRIC model to perform more accurately under the specific conditions of the study area. Specific enhancements included a new cloud gap filling procedure for ETrF1 images that allows the operator to adjust for background evaporation occurring from recent precipitation to better reflect total evaporation over longer (monthly) periods, the generation of gridded ETr maps used to estimate monthly and seasonal ET, improved computation of surface reflectance and albedo in mountainous areas to improve estimations of ET on sloped terrain. For Landsat 5 images, sharpening of the thermal band provided spatial refinement to the final ET products. Some potential limitations: All of the maps in this series are based on Landsat images. Most of the ET maps in this series are based on Landsat 5 images. However, some are based on Landsat 7 images. Landsat 7 images acquired after May 2003, although from a newer satellite than Landsat 5, are less preferred than Landsat 5, due to an anomaly with the Landsat 7 satellite caused by the malfunction of the scan line corrector (SLC). As a result, Landsat 7 images processed for years 2004 and 2006 are "SLC-off" images containing wedge shaped gaps extending from the edges of the image and stretching towards the centers. To obtain as complete coverage as possible, the gaps in ETrF maps produced by METRIC are generally filled in during post processing using the natural neighbor tool of Arc-GIS. The quality of the interpolation depends on the location of the gap, being better over homogenous landscapes. The Landsat 7 images were only used during periods when Landsat 5 images were not available due to clouds. A very small portion of the upper Sprague River basin lies to the east of path 45 and is covered by path 44, only. Portions of the Sprague basin lie in an overlap of both paths 45 and 44. Full METRIC applications for path 44 were not considered to be economical due to the small area. Instead, a relatively rapid, vegetation-index-based method for estimating fraction of reference ET, ETrF, was applied, where the coefficients for the method were based on data derived from the full METRIC application for path 45. Some adjustment to the general coefficients was made to account for background evaporation stemming from antecedent precipitation. More information on the accuracy of this dataset can be found in the documents http://water.usgs.gov/GIS/dsdl/Report_KBRA_OPWP_ET_2006_ETplus.pdf and http://water.usgs.gov/GIS/dsdl/Report_KBRA_OPWP_ET_path44_2004_2006_ETplus.pdf that were prepared by Evapotranspiration, Plus and are included with this dataset. Unknown See Accuracy Report Not applicable for raster data Data are complete These data were derived from 30-meter resolution Landsat imagery. The thermal infrared band was re-sampled from data acquired at 120-meter resolution for Landsat 5 and 60-meter resolution for Landsat 7. The re-sampling was done by the U.S. Geological Survey Earth Resources Observation and Science (EROS) Center prior to distribution. Additionally, for Landsat 5, the thermal infrared band was sharpened by Evapotranspiration, Plus as part of this project. More information on the accuracy of this dataset can be found in the documents http://water.usgs.gov/GIS/dsdl/Report_KBRA_OPWP_ET_2006_ETplus.pdf and http://water.usgs.gov/GIS/dsdl/Report_KBRA_OPWP_ET_path44_2004_2006_ETplus.pdf that were prepared by Evapotranspiration, Plus and are included with this dataset. U.S. Geological Survey Center for Earth Resources Observation and Science (EROS) 2006 http://glovis.usgs.gov/ online 2006 ground condition Landsat The METRIC procedure used to generate this dataset utilizes the visible, near-infrared and thermal infrared energy spectrum bands from Landsat satellite images and weather data to calculate ET on a pixel by pixel basis. U.S. Geological Survey Unknown http://seamless.usgs.gov/nlcd.php online Unknown ground condition NLCD A land use (LU) map was used to support the estimation of aerodynamic roughness and soil heat flux during METRIC processing. U.S. Geological Survey Unknown http://seamless.usgs.gov online unknown ground condition DEM A digital elevation map (DEM) is used during METRIC processing to adjust surface temperatures for lapse effects caused by elevation variation. Maps of slope and aspect (aspect is the cardinal direction of an inclined surface) are also derived from the DEM at 30 m resolution and are used in estimating solar radiation on slopes. These images were created using the tools of the ERDAS Imagine processing system based on the DEM. Evapotranspiration (ET) was obtained using the Mapping EvapoTranspiration at High Resolution and Internalized Calibration (METRIC) model developed by the University of Idaho. The METRIC procedure utilizes the visible, near-infrared and thermal infrared energy spectrum bands from Landsat satellite images and weather data to calculate ET on a pixel by pixel basis. Energy is partitioned into net incoming radiation (both solar and thermal), ground heat flux, sensible heat flux to the air and latent heat flux. The latent heat flux is calculated as the residual of the energy balance and represents the energy consumed by ET. The topography of the region was incorporated into METRIC via a digital elevation model (DEM), and used to account for impacts of slope and aspect on solar radiation absorption. METRIC was calibrated for each image using ground based meteorological information and identified 'anchor' conditions (the cold and hot pixels of METRIC) present in each image. A more detailed description of the process used to create this dataset can be found in the documents http://water.usgs.gov/GIS/dsdl/Report_KBRA_OPWP_ET_2006_ETplus.pdf and http://water.usgs.gov/GIS/dsdl/Report_KBRA_OPWP_ET_path44_2004_2006_ETplus.pdf that were prepared by Evapotranspiration, Plus and are included with this dataset. Landsat NLCD DEM Unknown Dr. Richard G. Allen, PE Evapotranspiration, Plus mailing address

3496 N. 2500 E.

Twin Falls ID 83301 USA 208-423-6601 Contact for technical questions related to this dataset. (Warning: Although accurate at the time of production, this information may have become obsolete. See the Metadata_Reference_Information section for a current contact.) This dataset was projected from WGS84 UTM Zone 10N to NAD83 UTM Zone 10N using ArcGIS 10.0 Service Pack 2, Project Raster Tool by U.S. Geological Survey personnel. The parameters used: Resampling Technique: Nearest (Per Alan Rea, USGS Hydrologist, Pers. Comm.) Output Cell Size: 30 Geographic Transformation: WGS_1984_(ITRF00)_To_NAD_1983 20110907 Raster Pixel 5514 7815 1 Universal Transverse Mercator 10 0.999600 -123.000000 0.000000 500000.000000 0.000000 row and column 30.000000 30.000000 meters North American Datum of 1983 Geodetic Reference System 80 6378137.000000 298.257222 Grid cell Grid cell used for class value ESRI VALUE Evapotranspiration (ET), in mm, for the month of August, 2006. ET values less than zero should be considered equal to zero, except along image boundaries where they should be considered “no data”. Evapotranspiration, Plus -348.9944153 383.0260315 mm, precision of the monthly actual evapotranspiration values is estimated as 200 mm (R.G. Allen, Evapotranspiration, Plus, LLC, Twin Falls, ID, written commun., 2011). 200 .2 Accuracy of the monthly actual evapotranspiration values is estimated as 20% of the value (R.G. Allen, Evapotranspiration, Plus, LLC, Twin Falls, ID, written commun., 2011). U.S. Geological Survey Ask USGS -- Water Webserver Team mailing address

445 National Center

Reston VA 20192 USA 1-888-275-8747 (1-888-ASK-USGS) http://water.usgs.gov/user_feedback_form.html Downloadable Data Although these data have been used by the U.S. Geological Survey, U.S. Department of the Interior, no warranty expressed or implied is made by the U.S. Geological Survey as to the accuracy of the data. The act of distribution shall not constitute any such warranty, and no responsibility is assumed by the U.S. Geological Survey in the use of these data, software, or related materials. The use of firm, trade, or brand names in this report is for identification purposes only and does not constitute endorsement by the U.S. Geological Survey. The names mentioned in this document may be trademarks or registered trademarks of their respective trademark owners. IMAGINE Image Unknown PKZIP compression Winzip http://water.usgs.gov/GIS/dsdl/mosaic_et_august2006_kl_NAD83.zip None. This dataset is provided by USGS as a public service. 20110930 U.S. Geological Survey Ask USGS -- Water Webserver Team mailing address

445 National Center

Reston VA 20192 USA 1-888-275-8747 (1-888-ASK-USGS) http://answers.usgs.gov/cgi-bin/gsanswers?pemail=h2oteam&subject=GIS+Dataset+mosaic_et_august2006_kl_NAD83 FGDC Content Standards for Digital Geospatial Metadata FGDC-STD-001-1998