Attribute_Accuracy_Report:
No formal attribute accuracy tests were conducted for the weather station sensors. All other continuous monitoring sensors were bench tested for accuracy in the laboratory at the beginning and end of the field season; salinity sensors were also field checked periodically. At the beginning of the field season, data from temperature sensors logging in air were compared to a National Institute of Standards and Technology (NIST) Traceable Digital Thermometer (Fisher brand) co-deployed on the bench. Logged temperature readings that were within manufacturer’s accuracy specification of temperature measured by the NIST Traceable Thermometer were deemed acceptable. Draft shields were placed around air temperature sensors to protect the sensor from unknown influences such as direct exposure to sunlight and wind gusts. Water level sensors were tested in the laboratory for accuracy by comparing logged pressure readings in air to the barometers used for barometric compensation; logged pressure readings that were within manufacturer’s accuracy specification of the barometer’s logged pressure in air were deemed acceptable. Salinity sensors were checked with a purchased calibration standard (Ricca 50.000 millisiemens per centimeter, product number 2248-1) and re-calibrated as needed. These procedures were repeated at the end of the field season. If sensor accuracy was found to have drifted over the field season, data corrections would be applied as described in the Process Steps portion of this metadata record.
Each logged parameter set was queried for maximum and minimum values to be sure logged data fell within expected ranges for the environmental conditions. Data were plotted to look for any obvious instrument errors (data jumps or gaps or noisy data). Erroneous data in the logged data files may be the result of one of several sources: vandalism of an instrument, removal of an instrument for maintenance, corrupt data file due to malfunctioning instrument, or wells with insufficient water level to produce reliable data. These data were removed from the final data report. Draft shields were not used on the dual channel soil/air temperature logger in 2017, although it was used in following years (the draft shield was used on the meteorological station's air temperature sensor all years).
Gaps in the continuous data record for a given parameter may have been the result of one of several sources: sensor not deployed during that time period, vandalism of an instrument, removal of an instrument for maintenance, corrupt data file due to malfunctioning instrument, wells with insufficient water level to produce reliable data, or logged data outside of the manufacturer's accuracy specifications. These data were removed from the final data report. As noted in the positional accuracy report, land surface elevation data was not available for the 2017 well at the eddy co-variance flux tower site. Water depth below land surface (dbs) is reported for this well, however the water elevation (WL_NAVD88) is not reported. Salinity data from the 2017 well deployment at the eddy co-variance flux tower site were excluded from the final dataset because the logged salinity values were greater than manufacturer's accuracy specifications for converting specific conductance to salinity units. The logger was within specification for specific conductance readings, calibration checks fell within acceptable ranges, and discrete checks of well water salinity using a refractometer were comparable to logged salinity (converted from specific conductance). However, because the salinity was outside of manufacturer's specification, it is not reported. Additionally, we have had concerns that wells at the eddy co-variance flux site may become stratified in the summer. Logged salinity in all three years is hypersaline and in 2019 the logged salinity data indicated that there may have been mixing within the well during routine well maintenance as loggers were removed and redeployed (characterized by lower salinity after a redeployment followed by a gradual return to hypersaline conditions). Salinity data from these periods of re-equilibration/re-stratification following routine maintenance in 2019 were excluded from the final dataset. Note that data from the first “Natural Creek Well”, beginning 04/17/19, are not reported; details are in a following process step. The anemometer on the weather station was replaced several times. This model anemometer has separate moving parts for measuring wind speed and wind direction. In some instances the wind speed sensor appeared to be collecting reliable data when the wind direction sensor was rusted in one position. In those cases only the wind speed data was reported.
Horizontal_Positional_Accuracy:
Horizontal_Positional_Accuracy_Report:
Horizontal positions for well locations were determined with a Trimble Real-Time Kinematic (RTK) GPS with Positional Dilution of Precision (PDOP) and Horizontal Dilution of Precision (HDOP) values less than or equal to 1.000 and 0.600 respectively. The latitude and longitude reported for 2018 and 2019 are averages of several RTK measurements taken on at least two different days. Based on the range in these measurements, the horizontal accuracy is estimated to be +/- 10 centimeters. The 2017 well and weather station location were estimated from Google Earth imagery of the study site. No formal positional accuracy tests were conducted.
Vertical_Positional_Accuracy:
Vertical_Positional_Accuracy_Report:
Water level loggers were deployed in wells at different stations within the Great Barnstable Marsh; sensor depth data from these instruments were converted to water elevation relative to NAVD88 using land surface elevations measured by Trimble Real-Time Kinematic (RTK) GPS. An average of elevation measurements made around the 2018 and 2019 well locations was used for the water level calculation. Details on this calculation are described in the process steps. Vertical Dilution of Precision (VDOP) for RTK GPS measurements were less than 0.800; average vertical accuracy of this method is +/- 5 centimeters. The 2017 well location did not have an RTK elevation measurement; water depth below land surface is the only water depth measurement reported for this year. No formal positional accuracy tests were conducted.
Process_Step:
Process_Description:
Great Barnstable Marsh (part of Sandy Neck Park, a municipal park in the town of Barnstable) was selected in 2017 as study location because of it's vast expanse of undeveloped salt marsh. It serves as natural control site to a number of investigations by study collaborators researching the effects of hydrologic alterations on salt marsh ecosystems and other tidally restricted wetlands. Upon initial site selection in 2017 a station was established with various continuous monitoring sensors (eddy co-variance carbon dioxide flux tower, meterological tower, water level sensor, soil and air temperature sensors). These sensors were removed in the winters and redeployed each spring. In 2018, an additional investigation began in a portion of Great Barnstable Marsh with specific goals of investigating differences between ditched and natural (unditched) salt marsh areas of Great Barnstable Marsh; well water level and temperature data were collected at different dtiched and natural transects in this portion of the marsh. This data release groups continuous monitoring data collected by USGS at all of our Great Barnstable Marsh study locations. The eddy co-variance carbon dioxide data is not reported here because that data was collected by non-USGS study collaborators.
Process_Date: 2020
Process_Step:
Process_Description:
Between 2017 and 2019, water level and conductivity were measured in a PVC well that was installed near to the eddy co-variance flux tower in the spring of each year; the well and logger were removed in the winter. Wells consisted of two pre-manufactured sections of 1 inch (2.54 cm) diameter well casing that were threaded together: a 2.5 inch (~0.762 meter) length of slotted screened casing (slot size 0.01 inch; ~ 3 mm) and a solid well casing that stuck above the sediment surface. An eyebolt was drilled into the inward facing edge of the threaded well cap; water level loggers were hung by aircraft cable threaded through the eyebolt and through the logger cap and secured on both ends using crimped stainless-steel oval sleeves. PVC wells were installed by pushing the well section into the sediment by hand or lightly tapping the top of the well with a mallet. The tip of the slotted screened well section was pushed about 100 cm below the land surface; the threaded portion (connecting the slotted length with the solid length) was below the sediment surface. At least 1 meter of well casing remained above the sediment surface to ensure that the groundwater well was not over-topped with surface water during the high tide. A vent hole was drilled near the top of the solid well casing. The well was destroyed by ice during the first winter deployment and was removed prior to ice formation in the following field seasons. Each spring, wells were reinstalled using the protocol previously described. Exact well heights and sensor deployment lengths are summarized in a separate entity and attribute section. The process date represents the latest process date.
Process_Date: 2019
Process_Step:
Process_Description:
Additional sensors were co-deployed at the long-term monitoring site, including dual channel (soil and air) temperature loggers, a barometer, and a weather station (photosynthethically active radiation, anemometer, rain gauge, air temperature and relative humidity). Soil temperature sensors were attached to a wooden stick and pushed by hand so that the bottom of the temperature sensor was 10 cm below the land surface. The air temperature sensor was attached the groundwater well ~1 m above the land surface. Note that draft shields were not used on the dual channel air temperature logger in 2017; however draft shields were used in subsequent years and draft shields were used on the weather station temperature/relative humidity sensor all years. The barometer was deployed in air near to the well location on the weather station mast. The PAR sensor was deployed above the height of the nearby Eddy Co-Variance CO2 flux tower in a location that would not be shaded by vegetation or other structures; other meteorological sensors were deployed on the same mast with the PAR sensor. Deployments took place between 2017 and 2019, with removal of instruments over the winter season. The process date represents the latest process date.
Process_Date: 2019
Process_Step:
Process_Description:
Well sensors were periodically downloaded in the field using instrument manufacturers' software (HOBOPro or WinSitu); data were inspected for general accuracy. Salinity sensors were periodically cleaned and checked in the field for fouling and calibration drift. Listed below are dates of visits to the eddy co-variance flux tower site.
04/12/2017; 05/24/2017; 05/26/2017; 06/07/2017; 08/09/2017; 08/17/2017; 08/24/2017; 09/12/2017; 11/14/2017; 01/10/2018
05/21/2018; 06/21/2018; 08/06/2018; 08/08/2018; 10/17/2018; 12/3/2018
04/17/2019; 05/01/2019; 05/22/2019; 06/12/2019; 08/22/2019; 09/05/2019; 10/25/2019; 12/23/2019
The process date represents the latest process date.
Process_Date: 2019
Process_Step:
Process_Description:
In 2018 and 2019, in addition to the eddy co-variance flux tower site, water level was also monitored in wells in a portion of the salt marsh with altered hydrology (“ditched”) and in an adjacent portion of the salt marsh with unaltered hydrology (“natural” or “un-ditched”). PVC wells were installed at each plot as described in the previous process step for wells installed at the eddy co-variance flux site. However, wells for plots in the hydrology study had an additional “fill” hole drilled about 10 cm above the sediment surface to allow the well to fill with surface water on the flood tide. Wells were installed along one ditched transect in 2018 and removed in December 2018 for the winter season (to prevent destruction by ice). Wells were reinstalled in 2019 along a different transect which included both ditched and natural salt marsh areas. Exact well heights and sensor deployment lengths are summarized in a separate entity and attribute section.
Well water level data loggers were deployed from October 2018 – December 2018 and April 2019 – October 2019. Loggers were periodically downloaded in the field using instrument manufacturers' software (HOBOPro, WinSitu); data were inspected for general accuracy. Dates of site visits are listed as follows,
Ditched Transect 2018: 6/21/18, 7/9/18, 7/19/18, 9/19/18, 10/17/18,12/3/18.
Ditched Transect 2019: 4/17/19, 5/1/19, 5/8/19, 6/12/19, 8/22/19, 10/25/19.
Natural Transect 2019: 4/17/19, 5/8/19, 5/22/19, 6/12/19, 8/22/19, 10/25/19.
During the 04/17/19 installation of the “Natural Creek Well”, it was noted that the well was slow to recharge. We had concerns that this well may be clogged, so an additional well was installed nearby on 5/22/19 with an In-Situ AquaTroll 200 deployed until 6/12/19. On 6/12/19 the In-Situ AquaTroll 200 was removed to be used in a different study and the HOBO U20-001-01 from the first “Natural Creek Well” was redeployed into the second “Natural Creek Well”. Deployment string lengths for these two different logger types were significantly different. Although this does not affect the reported water level (deployment string length is factored into water level calculations), it does affect the reported well water temperature. Groundwater temperature changes with depth from the land surface, thus an observed jump in groundwater temperature in the “Natural Creek Well” well on 6/12/19 is due to a difference in the depth at which the logger was measuring well water temperature. Note that data from the first “Natural Creek Well”, beginning 04/17/19, are not reported. The process date representing the latest process date.
Process_Date: 2019
Process_Step:
Process_Description:
Pressure data from non-vented water level sensors were barometrically corrected in manufacturers’ software (HOBOPro or WinSitu BaroMerge); corrected pressure was converted to sensor depth assuming a brackish water density (HOBOPro: 1.010 grams per cubic centimeter; WinSitu: 1.012 grams per cubic centimeter). Barometric pressure and density corrections were processed throughout the deployment period and checked while finalizing the data report. Raw pressure and barometric pressure are reported in this data release; the sensor depth calculated in the manufacturer's software is not reported because the sensor deployment heights must be accounted for to have meaningful water level data. Data from all continuous monitoring sensors were exported from manufacturers' software as a .csv file and compiled using MATLAB. Water elevation calculations were performed in MATLAB as described in the following process steps. Data resulting from erroneous measurements (malfunctioning sensors or insufficient water within a well) were excluded from the final dataset. Data were processed throughout the deployment period and checked while finalizing the data report; the process date represents the latest date of data processing.
Process_Date: 2020
Process_Step:
Process_Description:
Field measured deployment heights for specific time periods are summarized in a txt file (tab-delimited, filename: GBM_WellSensorDeploymentHeight_2017_2019.txt) included with this data release. These deployment heights were used to convert sensor depth (calculated from pressure in the manufacturer's software; described in previous process step) to water depth below land surface (dbs) and water elevation (WL_NAVD88). Calculations were processed throughout the deployment period in MATLAB. Positive dbs indicates water above the land surface; negative dbs indicates water below the land surface. The following parameters are used for these two calculations (units for all measurements are in meters):
a: depth of sensor as calculated in manufacturer's software (see previous process step for description of measured pressure data and assumptions of density constants);
b: land surface elevation (in the NAVD88 datum) at deployment location measured by RTK (land_surface_elevation; positive above (elevation) and negative below the datum);
c: deployment well height above land surface measured periodically in the field using a meter stick (height_above_land_surface; always positive);
d: total length of the deployment cable and water level sensor measured in the field using a meter stick (deployment_length; always positive);
Well water elevation (WL_NAVD88) is calculated as:
WL_NAVD88 = a +b+c-d
Well water depth below land surface (dbs) is calculated as:
dbs = a+c-d
The process date represents the latest date of data processing.
Process_Date: 2020
Process_Step:
Process_Description:
Added keywords section with USGS persistent identifier as theme keyword.
Process_Date: 20200807
Process_Contact:
Contact_Information:
Contact_Organization_Primary:
Contact_Organization: U.S. Geological Survey
Contact_Person: VeeAnn A. Cross
Contact_Position: Marine Geologist
Contact_Address:
Address_Type: Mailing and Physical
Address: 384 Woods Hole Road
City: Woods Hole
State_or_Province: MA
Postal_Code: 02543-1598
Contact_Voice_Telephone: 508-548-8700 x2251
Contact_Facsimile_Telephone: 508-457-2310
Contact_Electronic_Mail_Address: vatnipp@usgs.gov
