This data release contains water level observations from a sensor grid deployed in the tidal marshes of the Charles H. Wheeler Wildlife Management Area in the central Long Island Sound (Connecticut, USA). Water level sensors were deployed between August 2018 and December 2018 to serve as a calibration/validation (cal/val) dataset for the development, refinement, and evaluation of remote sensing inundation products. These data support several NASA projects seeking to improve remote sensing algorithms for characterizing inundation state and water depth in wetland systems, with a focus on radar and optical imagery. To process these data, water level observations were quality inspected and values below an empirically identified sensor noise floor were converted to NaN values representing a non-inundated state. These in-marsh water levels were then compared to water level observations from a nearby NOAA tide gauge to establish linear regression relationships accounting for time and elevation offsets between these datasets, culminating in the development of in-marsh synthetic tide gauge models with high accuracies (R² > 0.94). In-marsh water level data (HOBO_grids_00_to_33_one_minute_resolution_v1_meters.csv) can be used for direct cal/val of remote sensing observations within the four month collection period, with additional applications to marsh hydrology and ecology studies. Synthetic tide gauge models, described further in the accompanying journal publication, can be used for remote sensing cal/val for the multidecadal NOAA tide gauge record. Additional datasets supporting this analysis include shapefiles of in-marsh water level sensor locations (wheeler_marsh_2018_hobo_sensor_locations_utm18n_v1.shp), the tidal marsh study site extent (wheeler_marsh_extent_utm18n_v1.shp), and NOAA tide gauge water levels (NOAA_Bridgeport_tides_one_minute_resolution_v1_meters.csv).
