This readme.txt file was generated on <2023/03/25> by Patrick Brezonik, Leif Olmanson, and Martha Burket ------------------- GENERAL INFORMATION ------------------- Olmanson, L. G., Burket , M., & Brezonik, P. L. (2023). Reflectance spectra and related water quality and color data from 325 Minnesota and Wisconsin lakes and rivers [Data set]. Data Repository for the University of Minnesota (DRUM). https://doi.org/10.13020/1C2H-TE39 1. Title of Dataset: Data for 325 Minnesota and Wisconsin water bodies from 2013-2019: in-situ reflectance spectra, associated water quality, in situ and satellite-derived dominant wavelength and hue angle values, and MATLAB codes for spectral data analysis. 2. Author Information Name: Patrick L. Brezonik, Professor Emeritus Institution: Dept. of Civil, Environmental, and Geo- Engineering, University of Minnesota Address: Minneapolis, MN 55455 Email: brezonik@umn.edu Name: Leif G. Olmanson, Research Associate Institution: Dept. of Forest Resources, University of Minnesota Address: St. Paul, MN 55108 Email: olman002@umn.edu Name: Martha Otte Burket Institution: Dept. of Civil, Environmental, and Geo- Engineering, University of Minnesota Address: Minneapolis, MN 55455 Email: Martha.burket@gmail.com 3. Time period of content: 2013-08-08 to 2019-07-24 4. Geographic location of data collection: Lakes and some large rivers in Minnesota and in Wisconsin -------------------------- SHARING/ACCESS INFORMATION -------------------------- 1. Licenses/restrictions placed on the data: CC0 1.0 Universal (CC0 1.0) Public Domain Dedication 2. Links to publications that cite or use the data: Burket, M. O., L. G. Olmanson, and P. L. Brezonik. 2023. Comparison of two water color algorithms: Implications for the remote sensing of water bodies with moderate to high CDOM or chlorophyll levels. Sensors 23: 1071. -------------------------- METHODOLOGICAL INFORMATION -------------------------- 1. Description of methods used for collection/generation of data: Sampling and analytical methods for a(440), chl-a, SD, and TSS are described in Brezonik et al. Ecological Applications, 2019, 29: e01871. In-situ reflectance spectra were measured using methods described in Burket et al. Sensors, 2023, 23: 1071. Values of dominant wavelength and hue angle were computed from in-situ reflectance spectra measured at 1-nm intervals, using previously published equations and methods described in Burket et al. 2023. These values were assumed to be the “true” values of dominant wavelength and hue angle for each water body on the date of sampling. Simulated satellite-based values of dominant wavelength and hue angle for three satellite sensors, Snetinel-2 MSI, Sentinel-3 OLCI, and Landsat 8 OLI, were calculated according to two published methods (Wang et al. 2015. 2018, and van der Woerd and Wernand 2015, 2018; full references are given in Burket et al. 2023). 2. Describe any quality-assurance procedures performed on the data: A representative fraction of water quality samples was analyzed in duplicate or triplicate; standard analytical procedures were used. In-situ reflectance spectra were measured multiple times (5-10) at each site/collection date and the individual measurements were averaged. All calculations were double checked. 3. People involved with sample collection, processing, analysis and/or submission: L. Olmanson, P. Brezonik, and B. Page collected the in-situ reflectance spectra. The data were processed using standard protocols by P. Brezonik and/or L. Olmanson. Water quality samples were processed in the laboratory of Prof. Jacques Finlay, Dept. of Ecology, Evolution, and Behavior, University of Minnesota, St. Paul, and individuals associated with sample collection and analysis are listed in the DRUM file found at https://doi.org/10.13020/01WT-JG66. M. Burket did all the calculations for satellite-simulated and in-situ measured dominant wavelength and hue angle. ----------------------------------------- DATA-SPECIFIC INFORMATION FOR: MNWI_wq_data_and_in_situ_and_satellite_calc_dom_wavelength_and_hue_angle_data ----------------------------------------- 1. Number of variables: 69 2. Number of cases/rows: 325 3. Variable descriptions by column A Station ID: arbitrary name to identify Minnesota sites by state, year of measurement and order of measurement within year. For Wisconsin sites, the number indicates the Wisconsin lake id. B-Q Name of the variable and units of expression are self-explanatory in the Spreadsheet. R Chl_a_corr_mg/m3 is chlorophyll-a concentration corrected for phaeophytin in mg/m3. S TSS mg/L is total suspended solids concentration retained by 0.45 μm filter in mg/L. T a(440)_m^-1 is the Napierian absorption coefficient at 440 nm in reciprocal meters U Turbidity NTU is the water turbidity measured by nephelometer calibrated with formazin standards and reported as nephelometric turbidity units (NTU). V Secchi Depth,_m is the depth at which a Secchi disk was no longer visible in the water column, in meters. W dominant wavelength (hyperspec)nm is the dominant wavelength variable calculated from the hyperspectral data and expressed in nm. X dominant wavelength (vdWW, S2) nm is the dominant wavelength calculated by the van der Woerd and Wernand method from simulated satellite sensor data for the Sentinel-2 MSI sensor, expressed in nm. Y dominant wavelength (vdWW, S3) nm is the dominant wavelength calculated by the van der Woerd and Wernand method from simulated satellite sensor data for the Sentinel-3 OLCI sensor expressed in nm. Z dominant wavelength (vdWW, L8) nm is the dominant wavelength calculated by the van der Woerd and Wernand method from simulated satellite sensor data for the Landsat 8 OLI sensor, expressed in nm. AA Forel Ule index is an index with integer values ranging from 1 to 21 describing the hue of a water body, here determined from the hue angle calculated from the reflectance hyperspectra by a look-up table, as described in Burket et al. 2023. AB, AC, AD are columns of dominant wavelength data calculated by the Wang method using uncorrected hue angle data, respectively from simulated satellite sensor data for the S2 (MSI), S3 (OLCI), and L8 (OLI) sensors, expressed in nm. AE, Af, AG are columns of dominant wavelength data calculated by the Wang method using hue angle data corrected as described in Burket et al. 2023 from simulated satellite sensor data, respectively for the S2 (MSI), S3 (OLCI), and L8 (OLI) sensors, expressed in nm. AH Hue angle (hyperspec) degrees is the hue angle calculated for each site/date from the reflectrance hyperspectra and reported in degrees. AI, AK, AM are columns of hue angle data in degrees, calculated by the vdWW method using hue angle data corrected as per the vdWW method. AJ, AL, AN are columns of hue angle data in degrees calculated by the vdWW method using uncorrected hue angle data, which applies only to sites where the hue angle was < 30 degrees, the limit for which the vdWW method works for hue angle correction. AO, AQ AS are columns of hue angle data in degrees calculated by the Wang method using hue angle data corrected as per the Wang 2018 method, which was done only for hue angles > 70 degrees. AP, AR, AT are columns of hue angle data in degrees calculated by the Wang method using hue angle data uncorrected as per the Wang 2018 method, which applies to hue angles < 70 degrees because the correction algorithm did not work properly for these low hue angle values. ----------------------------------------- DATA-SPECIFIC INFORMATION FOR: Radiometric_data_Minn_2013-2018 ----------------------------------------- 1. Number of variables: 516 2. Number of cases/rows: 120 3. Variable descriptions by column A Station ID: arbitrary name to identify Minnesota sites by state, year of measurement and order of measurement within year. B-O Name of the variable and units of expression are self-explanatory in the spreadsheet. P-SV Reflectance in Rrs (sr-1) from 400 to 900 nm in 1-nm intervals. ----------------------------------------- DATA-SPECIFIC INFORMATION FOR: Radiometric_data_Wis_2014-2019 ----------------------------------------- 1. Number of variables: 522 2. Number of cases/rows: 206 3. Variable descriptions by column A Station ID: The number indicates Wisconsin DNR Lake ID. B-U Name of the variable and units of expression are self-explanatory in the spreadsheet. V-TB Reflectance in Rrs (sr-1) from 400 to 900 nm in 1-nm intervals. ----------------------------------------- SPECIFIC INFORMATION FOR: Sub-folder labeled MATLAB codes ----------------------------------------- This sub-folder contains ten files of MATLAB codes used in processing the hyperspectral data to calculate dominant wavelength and hue angle data. The files are described as follows: - dataLookup.m: This function imports the spectral response functions, vdWW calculation coefficients, and vdWW correction coefficients for three satellites (Sentinel-2, Sentinel-3, Landsat 8). - alfa2fui.m: This function converts hue angles to FUI values using the cutoffs defined by Novoa et al. 2013. - deriveWangAlphaCorrect.m: This script was used to derive the hue angle correctiosn for the Wang method. This script must be run while driverHyp.m is importing the IOCCG dataset (instead of the in situ data). - driverHyp.m: This script organizes many functions/scripts in order to calculate several chromaticity-related values using hyperspectral data. Calculated values are outputted into the "myAns" matrix. - genAlfaCorrect.m: This function applies hue angle corrections for both the Wang and vdWW method. - genBands2xyz.m: This function uses the vdWW method to convert satellite reflectances to chromaticity coordinates (x, y, and z) - genHyp2Bands.m: This function converts hyperspectral reflectances to simulated satellite band reflectances for a given satellite. - hyp2xyz.m: This function converts hyperspectral reflectances directly to chromaticity coordinates. - wangBands2xyz.m: This function uses the Wang method to convert band reflectances to chromaticity coordinates. - xyz2colorVals.m: This function converts chromaticity coordinates to dominant wavelength, purity, and hue angle. The function can apply hue angle corrections if specified by the user. In addition, several auxiliary data sources and files were used for the calculations. Descriptions/links to these data sources are provided below: - In Situ data: Referenced in driverHyp.m (combined_MN_WI_data_050621.xlsx). This dataset is available in the DRUM repository. - Landsat 8 SRFs: Referenced in dataLookup.m (landsat8_SRFs_v1.2.xlsx). Downloaded from https://landsat.gsfc.nasa.gov/satellites/landsat-8/spacecraft-instruments/operational-land-imager/spectral-response-of-the-operational-land-imager-in-band-band-average-relative-spectral-response/ - Sentinel-3 SRFs: Referenced in dataLookup.m (S3A_OL_SRF_20160713_mean_rsr.nc4). Downloaded from https://sentinels.copernicus.eu/web/sentinel/technical-guides/sentinel-3-olci/olci-instrument/spectral-characterisation-data - Sentinel-2 SRFs: Referenced in dataLookup.m (sentinel2-SRFs.xlsx). Downloaded from https://sentinels.copernicus.eu/web/sentinel/user-guides/sentinel-2-msi/document-library/-/asset_publisher/Wk0TKajiISaR/content/sentinel-2a-spectral-responses - CIE CMFs: Referenced in driverHyp.m (cie1931_1nm.xlsx, cie1931_5nm.xlsx). Downloaded from http://www.cvrl.org/ - IOCCG Synthetic Data: Referenced in driverHyp.m (IOP_AOP_Sun60.xlsx). Downloaded from https://www.ioccg.org/groups/OCAG_data.html - dominant_wavlength_noPlot.m: Referenced in xyz2colorVals.m. This script calculates dominant wavelength and purity from chromaticity coordinates. This function was adapted from the work of other researchers (Lehmann et al. 2018, Giardino et al. 2019).