Pamela L Nagler Armando Barreto-Muñoz Chritopher J Jarchow Kamel Didan 20200806 comma-separated values (csv) and tagged image file format (tif) Flagstaff, AZ U.S. Geological Survey https://doi.org/10.5066/P98PGDJ1 Pamela L. Nagler Armando Barreto-Muñoz Sattar Chavoshi Borujeni Christopher J. Jarchow Martha M. Gómez-Sapiens Hamideh Nouri Stefanie M. Herrmann Kamel Didan 2020 journal manuscript Wiley Online Library Hydrological Processes https://doi.org/10.1002/hyp.13911 These operational land imager (OLI) value added data sets, maps, and associated ancillary data were compiled as part of an ongoing research aimed at quantifying the riparian vegetation greenness and water use in the lower Colorado River Delta in Mexico. In order to create trend and anomaly maps that characterize these ecosystems Vegetation Index (NDVI) time series imagery from Landsat OLI were acquired and processed over time and space along seven predefined reaches that capture different natural states and management conditions. We used Landsat OLI 30m data as an improvement upon past studies that were based on coarser remote sensing data from the NASA MODIS sensor (250 m). The OLI 30m images provide better characterization and performance over these rather narrow riparian corridors. To capture the change over time we used a simple differencing technique that compares two annual average growing season VI cycles (limited to May-October). These anomaly maps capture how the corridor vegetation health responds to both natural and anthropogenic changes. We limited this study to the full OLI record (2013-2019) since we were interested in understanding the response to Minute 319 pulse flow of 2014. The difference maps are an ideal tool for capturing how the released water impacted vegetation immediately and over long time. The Minute 319 pulse flow science team in collaboration with the University of Arizona have developed a data processing system to support this effort with focus on understanding how the riparian corridor is responding to these natural and anthropogenic stressors. All data associated with this project were acquired from the LP-DAAC and pre-processed to remove and capture issues prior to further processing (see below) which involved reprojection to a common projection, masking to only retain the area of interest, quality analysis to discard poor data, and then value addition to generate the NDVI and the difference maps. The data acquisition and analysis were performed at the University of Arizona VIP lab (vip.arizona.edu) using their large Linux cluster of computing and storage resources. A mix of off the shelf software and specialized in-house tools were used to carry the different steps and analyses. These datasets were created to continue monitoring riparian zone trends and changes in the Lower Colorado Delta as part of the Minute 139 of the 1944 Water Treaty between the United States and Mexico. These datasets, while specific to the research questions addressed by this research (see Larger Work Citation), were designed to be accessible and used by others involved in research efforts on the lower Colorado River Delta. These maps capture a critical period of the Lower Colorado River riparian zone; in particular, as a pre- and post- pulse flow water release and provide a first synoptic view of how the vegetation responded and will continue to respond, and thus are very useful for follow up studies that may compare trends spatially to restoration areas, or temporally, either by comparing to prior years or more likely, by extending the performance period beyond the year 2019. These data sets are highly dependent on the input data surface reflectance correction and calibration which is handled at the LP-DAAC where OLI data is distributed. Our processing pipeline assumes no disturbance when filling the gaps which may result in disturbance (fire, land cover change, etc.) removal. Data users should read the entire metadata record and acquire the manuscript identified as the ‘Larger Work Citation’ and any manuscripts identified as ‘Cross Reference' to have a complete understanding of how these data were created and used. The data are specific to the uses identified above, as described in the ‘Larger Work Citation’, and any other use of these data would be inappropriate. See 'Distribution liability' statements for more information. 2003 2019 ground condition Complete None planned -115.32 -114.69 32.75 31.925 USGS Thesaurus ecosystem monitoring field inventory and monitoring Landsat images remote sensing vegetation wetland ecosystems USGS Biocomplexity Thesaurus ecosystem sustainability estuarine ecosystems deltas riparian habitats terrestrial ecosystems vegetation change None Minute 319 Agreement NDVI normalized difference vegetation index riparian ecosystem USGS Metadata Identifier USGS:5f2ae9fc82ceae4cb3c09329 Geographic Names Information System (GNIS) Arizona Yuma Morelos Dam Mexico San Luis None Gulf of California Getty Thesaurus of Geographic Names Northern International Boundary Southern International Boundary United States International Boundary none none Pamela L Nagler U.S. Geological Survey Research Physical Scientist mailing and physical

Mail Stop 9396, 520 North Park Avenue

Tucson AZ 85719 US 520-670-3357 pnagler@usgs.gov The data authours would like their gratitude to the National Aeronautics and Space Administration (NASA) for providing funds for work on the Landsat OLI and MODIS data fusion. Pamela L. Nagler Edward P. Glenn Uyen Nguyen Russell L. Scott Tanya Doody 2013 MDPI (online) Remote Sensing https://doi.org/10.3390/rs5083849 Christopher J. Jarchow Pamela L. Nagler Edward P. Glenn Jorge Ramírez-Hernández J. Eliana Rodríguez-Burgueño 2016 Elsevier, ScienceDirect (online) Ecological Engineering https://doi.org/10.1016/j.ecoleng.2016.10.056 Christopher J. Jarchow Pamela L. Nagler Edward P. Glenn 2017 Elsevier, ScienceDirect (online) Ecological Engineering https://doi.org/10.1016/j.ecoleng.2016.08.007 David P. Groeneveld William M. Baugh 2007 Elsevier, ScienceDirect (online) Journal of Hydrology https://doi.org/10.1016/j.jhydrol.2007.07.001 No formal attribute accuracy tests were conducted No formal logical accuracy tests were conducted Data set is considered complete for the information presented, as described in the abstract. Users are advised to read the rest of the metadata record carefully for additional details. No formal positional accuracy tests were conducted No formal positional accuracy tests were conducted U.S. Geological Survey, EarthExplorer 2019 Geotiff raster data Sioux Falls, SD U.S. Geological Survey https://earthexplorer.usgs.gov/ digital raster data 20130101 20191030 publication date 30-meter NDVI, EVI, Surface Reflectances and Quality Assessment These 2013, 2014, 2015, 2016, 2017, 2018 and 2019 data (VI and QA information) were used to create 6 Growing Season Normalized Difference Vegetation Index (NDVI) Difference Maps PROCESS STEP 1, INPUTS - SOURCE DATA: Landsat 8 OLI/TIRS C1 Level-2 on demand Tier1 scenes 038038, 038037, 039038 and 039037 were acquired for years 2013, 2014, 2015, 2016,2017, 2018 and 2019 from the Landsat USGS data distribution portal EarthExplorer from https://earthexplorer.usgs.gov/. All data available for each year were acquired and consisted of Geotif raster images in the UTM projection Zone 11 with a spatial resolution of 30 meters per pixel. Each data set contained the vegetation indices NDVI and EVI and the surface reflectance bands as well as the per pixel quality flag layer. 2019 PROCESS STEP 2, DATA PRE-PROCESSING: Data is acquired in a granule/scene format that is not ready for analysis nor cover the full region of interest (ROI). Our processing pipeline consisted of few small analysis steps, starting with mosaicking (Stitching ) the scenes to construct a larger area. Then subsetting to only retain the seven reaches that cover the Lower Colorado riparian zone in Mexico. Four Landsat scenes were geographically stitched into a single larger mosaic for each data band (VIs, surface reflectance, and QA). The larger image was then cropped to create a single long and narrow map covering the 7 reaches of the Lower Colorado Delta. The extent of the ROI in UTM coordinates is (in meters): North = 3623850, West = 654120, South = 3529890, and East = 720930. The resulting data is then filtered using the Landsat provided per pixel Quality Assurance (QA) information (data layers: ‘sr_aerosol’ and ‘pixel_qa’) using a conservative approach that attempts to retain the highest possible quality only. Only pixel that are cloud free with low or average aerosol loads (they are more effectively corrected) are retained. Cloud shadow pixels are also removed. In addition, a buffer of 8 pixels (240m) around any cloudy pixel was also removed to minimize residual and sub pixel clouds as well as cloud shadow. The following QA field were used for this process: Cloud shadow: Bit 3 from band ‘sr_pixelQA’, Cloud flag: Bit 5 from band ‘sr_pixelQA’, and Aerosol: Bit 6 & 7 from band ‘sr_aerosol’. The NDVI is then scaled and stretched to address any remaining differences in the dynamic range of the data acquired at different times and under different atmosphere conditions. This makes comparing scene-to-scene over time less prone to illumination conditions and biases and assures the accuracy of changes in the VI or ET images. This method was proposed and employed by Jarchow, Nagler & Glenn (2017) and implements a simple empirical scaling algorithm following methods in Groeneveld & Baugh (2007) that minimize the dynamic range bias. The NDVI scaling process assumes that there is always a maximum saturation NDVI signals (NDVIs) and an absolute bare soil minimum NDVI signal (NDVIo). This scaling process standardizes NDVI across all time periods by stretching all NDVI values in the scene between the two NDVIo and NDVIs. NDVI* = (NDVI – NDVI0)/(NDVIs – NDVI0) Using the above equation, a NDVI* scaled raster was computed for each scene in the time period (2013-2019). The actual processing was performed using in -house custom tools developed at the Vegetation Index and Phenology Lab at the University of Arizona. However, these processing steps can be easily reproduced with any Image processing or GIS software since most of these steps are quite standard. 2019 PROCESS STEP 3, FINAL OUTPUT(S): To create the seasonal NDVI* data, the NDVI scaled rasters were averaged between May 1st and October 30th (predefined growing season). This process was implemented at the pixel level and for years 2013, 2014, 2015, 2016, 2017, 2018, 2019, 2014-2019, and 2003-2019. The final output maps were obtained by calculating the difference between two consecutive years (ex: 2014 and 2013, the seasonal NDVI* from 2013 and 2014 were used). Only data within the ROI was retained and everything else was masked out to highlight the riparian corridor seven reaches only. Geotif files were generated for all years 2013-2014, 2014-2015, 2015-2016, 2016-2017, 2017-2018, 2018-2019, 2014-2019, and 2003-2019 difference maps. 2019 Data Quality Assessment and Quality Control (QAQC): As with all space borne imagery, these data sets are highly dependent on the surface reflectance calibration and atmosphere correction. The LP-DAAC provides and distributes calibrated and atmospherically corrected data and further tags each pixel with additional QA information to support user specific data filtering. To remove clouds, high aerosol, shadow and otherwise any poor-quality data we employ the QA pixel information provided with the data. While generally effective filtering data based on per-pixel QA will usually depend on the accuracy of these QA flags, nonetheless, commission error are usually tolerable as opposed to omission errors. In addition, we have implemented an 8 pixels buffer around clouds to remove any residual clouds and shadows. Throughout the full processing pipeline we retained the same spatial domain and projection as the input to eliminate geometric transformation and pixel aggregation biases. QA filtering consisted of screenig for clouds, cirrus, cloud shadow and aerosols. Only cloud free pixels and low or average aerosol conditions were retained. A second step for filtering is based on using the NDVI Long term average value to remove outliers that are outside the standard deviation. Any resulting gaps were then filled using current year trends and long-term average behavior. 2019 Finalize Spatial Data for Dissemination: Data sent to the Southwest Biological Science Center Data Steward for dissemination and preservation per USGS Data Management Policies SM 502.6, SM 502.7, SM 502.8 & SM 502.9 (1 October 2016). 2020 Raster Universal Transverse Mercator 11N 0.9996 -117.0 0.0 500000.0 0.0 row and column 30.0 30.0 meters D_WGS_1984 WGS_1984 6378137.0 298.257223563 GrowingSeasonNDVI_2013-2014.tif to GrowingSeasonNDVI_2018-2019.tif These data represent the NDVI difference between two consecutive years. It shows the change in vegetation as a result of riparian ecosystem interactions within these particular years. Landsat scenes were averaged between May 1st and October 30th from each year. Only the best quality pixels (cloud free) were used. Data are stored in GeoTIF images/maps and capture the interannual change in NDVI, which is a proxy measure of Vegetation health and productivity. Producer defined Value Average yearly seasonal normalized difference vegetation index (NDVI) difference Producer defined -0.41949999332428 0.99989998340607 VI units GrowingSeasonNDVI_2003-2019.tif These data represent all years normalized difference vegetation index (NDVI) difference between 2003 to 2019. It shows the change in vegetation as a result of riparian ecosystem interactions within these particular years, 2003-2019. Landsat scenes were averaged between May 1st and October 30th from each year. Only the best quality pixels (cloud free) were used. Data are stored in GeoTIF images/maps and capture the interannual change in NDVI, which is a proxy measure of Vegetation health and productivity. Producer defined Value All years (2003-2019) change in seasonal normalized difference vegetation index (NDVI) difference Producer defined -0.68489998579025 0.47889998555183 millimeters GrowingSeasonNDVI_2014-2019.tif These data represent a post-pulse five-year change in normalized difference vegetation index (NDVI) difference between 2014 to 2019. It shows the change in vegetation as a result of riparian ecosystem interactions following a post-pulse flow period from 2014 to 2019. Landsat scenes were averaged between May 1st and October 30th from each year. Only the best quality pixels (cloud free) were used. Data are stored in GeoTIF images/maps and capture the interannual change in NDVI, which is a proxy measure of Vegetation health and productivity. Producer defined Value Post-pulse five-year change in seasonal normalized difference vegetation index (NDVI) difference Producer defined -0.46729999780655 0.40669998526573 millimeters U.S. Geological Survey - ScienceBase U.S. Geological Survey mailing and physical

Denver Federal Center, Building 810, Mail Stop 302

Denver CO 80225 United States 1-888-275-8747 sciencebase@usgs.gov The author(s) of these data request that data users contact them regarding intended use and to assist with understanding limitations and interpretation. Unless otherwise stated, all data, metadata and related materials are considered to satisfy the quality standards relative to the purpose for which the data were collected. Although these data and associated metadata have been reviewed for accuracy and completeness and approved for release by the U.S. Geological Survey (USGS), no warranty expressed or implied is made regarding the display or utility of the data for other purposes, nor on all computer systems, nor shall the act of distribution constitute any such warranty. This zip file contains data available in tabular comma-separated values and raster tif formats. The user must have software capable of uncompressing the zip file, and displaying the tabular and raster data sets. 20200827 Pamela L Nagler U.S. Geological Survey Research Physical Scientist mailing and physical

Mail Stop 9396, 520 North Park Avenue

Tucson AZ 85719 US 520-670-3357 pnagler@usgs.gov Content Standard for Digital Geospatial Metadata FGDC-STD-001-1998