Evapotranspiration (ET) at Blue Cypress marsh site, daily data, Indian River County, Florida, June 1, 1995 October 20, 2014

The data set consists of evapotranspiration (ET) measurements/estimates based on data collected at the USGS Blue Cypress marsh station 274143080424100 beginning June 1, 1995 and ending October 20, 2014. The station is located at a nearly flat wetlands site (27o 41 43 N 80o 42 41 W) within the Blue Cypress Marsh Conservation Area, Indian River County, Florida. The dominant plant cover at the study site is sawgrass (Cladium jamaicense), with secondary amounts of other wetland plant species. Sawgrass height generally varies from 1.8 to 2.4 meters. The soils at the site are peats. The water-table generally is above land surface but can be greater than a meter below land surface during drought periods. Controlled burns and wildfires occur periodically in the marsh. The eddy-covariance method was used to measure ET for the period January 1, 2000 to September 1, 2005 and the period December 11, 2009 to October 20, 2014. The unique contribution of the present Data Release is dissemination of a dataset of ET estimates for a period prior to the first period of measurements (June 1, 1995 to December 31, 1999) and for the time interval between the two periods of measurement (September 2, 2005 to December 10, 2009). The ET data for the two measurement periods were USGS approved and released prior to the present Data Release and is available at the following two locations: 1) for January 1, 2000 to September 1, 2005, USGS station ST JOHNS RV MARSH AT BLUE CYPRESS NR FELLSMERE FL - U.S. Geological Survey, 2016, National Water Information SystemWeb interface, accessed December 5, 2017, at http://dx.doi.org/10.5066/F7P55KJN and 2) for December 11, 2009 to October 20, 2014, Shoemaker, W. B., 2016, Evapotranspiration and Carbon Cycling Data within Blue Cypress Marsh, Vero Beach, Florida, 2009 to 2014 Web interface, accessed January 4, 2017, at https://sofia.usgs.gov/exchange/blue_cypress_et_c/ - specifically, the variable et_cor2_mm was utilized as measured actual ET. Water level data at the site is also available within these previous data releases for the site during these two time periods. The USGS report Sumner, D. M., 2001. Evapotranspiration from a cypress and pine forest subjected to natural fires in Volusia County, Florida, 1998-99, U. S. Geological Survey Water-Resources Investigations Report 01-4245, 56 pp.) provides information on the background, theory, and methodology of the eddy covariance method of measuring ET. Generally, eddy-covariance measurements result in missing data (for reasons discussed in Sumner (2001)) and gap-filling methods are traditionally used to interpolate missing data the measured data in the two ET datasets are composed of both directly measured and inferred measurements of ET within these periods of record. Estimates of ET for the two large unmeasured periods (June 1, 1995 to December 31, 1999; and September 2, 2005 to December 10, 2009) were based on the methodology outlined in Sumner, David M., 2006, Adequacy of selected evapotranspiration approximations for hydrologic simulation, Journal of American Water Resources Association (JAWRA), 42(3): 699-711. In this methodology, an annually-invariant pattern of monthly vegetation coefficients are applied as multipliers to independent estimates of daily potential ET to generate estimates of actual ET. The monthly vegetation coefficients are developed as regression-derived estimates based on the actual ET measurements. The source of the potential and reference ET data was a Statewide-wide database of daily values at 2-kilometer resolution and extending from June 1, 1995 to December 31, 2015. This database is described in John R. Mecikalski, David M. Sumner, Jennifer M. Jacobs, Chandra S. Pathak, Simon J. Paech, and Ellen M. Douglas. 2011. Use of Visible Geostationary Operational Meteorological Satellite Imagery in Mapping Reference and Potential Evapotranspiration over Floridain Evapotranspiration. Leszek Labedzki (ed), InTech Publishers, Vienna, Austria, 446 p. and is accessible at: https://fl.water.usgs.gov/et/. Daily potential ET data were obtained for the pixel (number 91413) including the location of the ET station for the period June 1, 1995 to October 20, 2014 (including both ET measurement periods and both periods of unmeasured ET for which ET estimates were desired). The actual ET method based on annually-invariant monthly vegetation coefficients applied to potential ET estimated the 3839 measured values of daily actual ET with an r-squared value of 0.85, a standard error (standard deviation of estimation residuals) of 0.58 millimeters per day and a coefficient of variation (standard error divided by mean ET) of 17 percent; monthly averages of actual ET were estimated with an r-squared value of 0.95, a standard error of 0.27 millimeters per day and a coefficient of variation of 8 percent. Water level was not a substantive explanatory variable for actual ET the correlation between regression residuals of actual ET and water level was small (r-squared = 0.002). Fires removed most of the sawgrass canopy in February 2001, March 2008, and March 2014, but re-growth is rapid (e.g., canopy had recovered to a height of 1.2 meters 67 days after February 2001 fire). Although the impacts of fire on ET are not explicitly incorporated into the annually-invariant monthly vegetation coefficients, the effects of this simplification are likely of secondary importance based on the model error statistics cited above. The regression-inferred, annually-invariant monthly vegetation coefficients were as follows: 1.11 (January), 0.95 (February), 0.93 (March), 0.94 (April), 0.92 (May), 0.92 (June), 0.87 (July), 0.88 (August), 0.88 (September), 0.99 (October), 1.12 (November) and 1.20 (December). Although measured actual ET and database-obtained potential ET data are previously released as noted above, these data are reproduced in the present release for convenience and comparison to simulated actual ET; in addition, a seamless time series of measured and estimated actual ET for the period June 1, 1999 to December 10, 2014 is provided; estimates of actual ET, derived from the methods described above, were used to gap-fill missing measured data. Although the actual ET estimator performs relatively well in reproducing measured data based on the error statistics noted above, it should be noted that this degree of accuracy is not assured for non-measured time periods.