The magnitude and variability of floods have increased for many nontidal streams on Long Island (LI), NY since the mid-20th century. One of the most densely populated regions of the United States, LI has experienced amplified floods in step with increases in impervious land cover, storm, and sanitary sewers that have accompanied urban development. To better understand the drivers of observed flood trends and effects of urbanization, a nonstationary flood frequency analysis is conducted, using historical annual peak flow records from 17 gaged watersheds on LI using conditional moments based on physical covariates from a two-stage sequential robust linear regression procedure. Regression results indicate that urban development and precipitation are significant co-predictors of peak flows for LI watersheds that have undergone rapid development during the available peak flow record. In watersheds with less intense urbanization or that were fully developed before the peak flow record began, precipitation alone was a significant explanatory variable. Long-term baseflow patterns identified using a nonparametric smoother explained some patterns of decreasing peak flows and heteroskedasticity in the peak flow records. Fitting a log-Pearson III distribution with these conditional moments, floods corresponding to a 20% annual exceedance probability (AEP) are up to 80% higher under a nonstationary framework compared with stationary under current watershed conditions, and differ significantly (95% confidence) from stationary estimates for 6 out of 17 watersheds. Larger floods corresponding to 1% AEPs do not differ significantly between nonstationary and stationary estimates at a 95% confidence level. Nonmonotonic trends observed in two watersheds indicate that recent stormwater management practices, such as rerouting stormwater outfalls away from the channel, substantially reduce flood frequency. Reduced nonstationary flood quantile estimates at these two watersheds are 20 to 40% lower than stationary estimates when accounting for changing watershed conditions over time. Across LI, stormwater management and water-table fluctuations have increased peak flow variability, characteristic of a late phase urban adjustment period on LI. Results of this study demonstrate that a nonstationary framework is a necessary step forward toward a regional flood-frequency analysis for LI. This nonstationary framework will allow flood managers to update flood discharge estimates to current conditions that reflect altered relationships between urban cover and climate for more targeted planning of flood control, transportation infrastructure, and management of floodplain ecosystems.

ABSTRACT Multiple instruments and methods exist for collecting discrete streamflow measurements in small streams with low flows, defined here as less than 5.7 m 3 /s (200 ft3/s). Included in the available methods are low‐cost approaches that are infrequently used, in part, because their uncertainty is not well known. In this work, we evaluated the accuracy and suitability of three low‐cost velocity measurement methods (surface float [SF], velocity head rod [VR], and rising body [RB]) and three conventional current meters (acoustic Doppler velocimeter, and mechanical Price type AA and Price Pygmy meters) relative to discharge calculated from stable artificial hydraulic controls. A total of 231 measurements were made by 20 individuals during 88 site visits to 24 sites in eight states. Accuracies were assessed for all methods and precision was evaluated for the low‐cost methods. The median percent error was below 5% for conventional methods, and below 20% for the low‐cost methods. The SF was the most accurate (median absolute percent error 14%) and precise (mean percent precision of 11%) low‐cost method. The RB and VR, respectively, had 15% and 20% median absolute percent error and 29% and 12% mean percent precision. Results suggest that low‐cost methods, when used appropriately, can be used to estimate discharge data under low flow conditions when measurements with conventional methods are not feasible and the associated accuracies meet end‐user measurement objectives.

External sediment supply is an important control on wetland morphology and vulnerability to storms, sea-level rise, and land use change. Constraining sediment supply and net budgets is difficult due to multiple timescales of variability in hydrodynamic forcing and suspended sediment concentrations, as well as the fundamental limitations of measurement and modeling technologies. We used two independent observational campaigns and one hydrodynamic modeling effort to estimate the sediment supply to Jamaica Bay, New York, USA, an urbanized embayment with a history of extensive wetland loss. We found that all three estimates indicate a net import to the system, ranging from 36 to 74 kt/year, with a mean estimate of 55 kt/year ± 31 kt/year, which is consistent with a prior estimate derived from radionuclide tracers. Net sediment import is controlled by flood-ebb asymmetry in bed shear stress, which results in higher suspended sediment concentrations on flood tide relative to ebb. This indicates a seaward source of sediment that is resuspended by waves in the coastal ocean, likely offshore marine deposits or potentially from the adjacent Hudson River estuary. Despite the net sediment import, a simple sediment budget suggests that the rate of supply is not sufficient to maintain the present geomorphic planform of the system relative to sea-level rise. The convergent estimates from independent methods provide reasonable guidance as context for sediment-based restoration efforts.

First posted December 30, 2019 For additional information, contact: Director, New York Water Science CenterU.S. Geological Survey425 Jordan RoadTroy, NY 12180–8349 Tidal wetland loss in Jamaica Bay, New York, is well documented. Maintaining wetlands is important from an environmental and ecological perspective and because wetlands buffer coastal communities from storm damage. An estimate of suspended-sediment flux through Rockaway Inlet is needed to improve understanding of sediment dynamics in Jamaica Bay and could be used in salt marsh restoration efforts. To estimate sediment flux, an index-velocity station and turbidity sensor were installed and operated in Rockaway Inlet near the mouth of Jamaica Bay from November 2014 to December 2016 and point and cross-sectional suspended-sediment samples were collected and analyzed. Index-velocity data coupled with cross-sectional acoustic Doppler current profiler measurements were used to develop an index-velocity rating. A simple linear regression rating with a strong coefficient of determination (R2 of 0.981) was developed. Discharge was computed from the stage-area and index-velocity relations, and a low-pass Godin filter was used to remove the tidal aliasing. A second simple linear regression (R2 of 0.75) between fixed-point suspended-sediment concentration (SSC) samples and turbidity allowed for the calculation of SSC through Rockaway Inlet, and then sediment flux was found by multiplying SSC and discharge for continuous (6-minute) data. Turbidity values were low in the near-ocean conditions at Rockaway Inlet, with daily means ranging from 0.6 to 8.2 formazin nephelometric units during the period of November 2014 through December 2016. During this time, computed daily mean suspended-sediment concentrations ranged from 3 to 13 milligrams per liter. High sediment loads generally occurred during incoming tides, during both storm and nonstorm conditions, suggesting a net inward sediment flux into Jamaica Bay. The fate of sediment after it enters Jamaica Bay was not investigated. Trends in sediment flux during major storms could not be evaluated because no major storms occurred during this investigation.

Urbanization affects streamflow characteristics, coastal flooding, and groundwater recharge. Increasing impervious areas, streamflow diversions, and groundwater pumpage are some of the ways that the natural water cycle is affected by urbanization. Assessment of the relations among these factors and changes in land use helps water-resource managers with issues such as stormwater management and vulnerability to flood and drought. Scientists with the U.S. Geological Survey (USGS) have the expertise to monitor and model urban hydrologic systems. Streamflow and groundwater data are available in national databases, and analyses of these data, including identification of long-term streamflow trends and the efficacy of management practices, are published in USGS reports.

In response to Hurricane Sandy, personnel from the U.S. Geological Survey (USGS) deployed a temporary network of storm-tide sensors from Virginia to Maine. During the storm, real-time water levels were available from tide gages and rapid-deployment gages (RDGs). After the storm, USGS scientists retrieved the storm-tide sensors and RDGs and surveyed high-water marks. These data demonstrate that the timing of peak storm surge relative to astronomical tide was extremely important in southeastern New York. For example, along the south shores of New York City and western Suffolk County, the peak storm surge of 6–9 ft generally coincided with the astronomical high tide, which resulted in substantial coastal flooding. In the Peconic Estuary and northern Nassau County, however, the peak storm surge of 9 ft and nearly 12 ft, respectively, nearly coincided with normal low tide, which helped spare these communities from more severe coastal flooding.

First posted July 21, 2015 For additional information, contact: Director, New York Water Science Center U.S. Geological Survey 2045 Route 112, Building 4 Coram, NY 11727 http://ny.water.usgs.gov The hybrid cyclone-nor’easter known as Hurricane Sandy affected the mid-Atlantic and northeastern United States during October 28-30, 2012, causing extensive coastal flooding. Prior to storm landfall, the U.S. Geological Survey (USGS) deployed a temporary monitoring network from Virginia to Maine to record the storm tide and coastal flooding generated by Hurricane Sandy. This sensor network augmented USGS and National Oceanic and Atmospheric Administration (NOAA) networks of permanent monitoring sites that also documented storm surge. Continuous data from these networks were supplemented by an extensive post-storm high-water-mark (HWM) flagging and surveying campaign. The sensor deployment and HWM campaign were conducted under a directed mission assignment by the Federal Emergency Management Agency (FEMA). The need for hydrologic interpretation of monitoring data to assist in flood-damage analysis and future flood mitigation prompted the current analysis of Hurricane Sandy by the USGS under this FEMA mission assignment. The analysis of storm-tide impacts focused on three distinct but related aspects of coastal flooding from Hurricane Sandy, including flooding inland along the tidal reach of the Hudson River. These aspects are (1) comparisons of peak storm-tide elevations to those of historical storms and to annual exceedance probabilities, (2) assessments of storm-surge characteristics, and (3) comparisons of maps of inundation extent that were derived from differing amounts of available storm-tide data. Most peak storm-tide elevations from Hurricane Sandy were greater than about 9.5 feet (ft) above North American Vertical Datum of 1988. Peak storm-tide elevations from Hurricane Sandy were compared with data for the intense nor’easter of December 11–13, 1992, and Hurricane Irene (August 27–28, 2011), which weakened to a tropical storm before arriving in New York. Peak storm-tide elevations from Hurricane Sandy were higher than those from the December 1992 nor’easter at 24 of 27 sites; most differences were greater than about 0.7 ft or 9 percent (above the historical storm tide). Peak storm-tide elevations from Hurricane Sandy were higher than those from Tropical Storm Irene at all sites; most differences were greater than about 2.5 ft or 48 percent. Data from permanent and temporary monitoring sites and HWM sites were compared with corresponding FEMA flood elevations for the 10-, 2-, 1-, and 0.2-percent annual exceedance probabilities in New York. Peak storm-tide elevations from Hurricane Sandy had annual exceedance probabilities less than or equal to 1 percent and (or) greater than 0.2 percent at a plurality of sites—184 of 413. Peak storm-tide elevations greater than or equal to the 0.2-percent flood elevation accounted for 81 of 413 sites. Peak storm-tide elevations less than the 10-percent flood elevation accounted for only 10 of 413 sites. Data from selected permanent monitoring sites in the USGS and NOAA networks were used to assess storm-surge magnitude associated with the peak storm tide, and magnitude and timing of the peak storm surge. Most magnitudes of the peak storm surge were greater than about 8.3 ft, and most magnitudes of the storm surge component of the peak storm tide were greater than about 7.8 ft. Timing of peak storm surge arrival with respect to local phase of tide controlled where the most extreme peak storm-tide levels and coastal flooding occurred. This finding has bearing not only for locations impacted by the highest storm tides from Hurricane Sandy, but also for those that had the greatest storm surges yet were spared the worst flooding because of fortuitous timing during this storm. Results of FEMA Hazus Program (HAZUS) flood loss analyses performed for New York counties were compared for extents of storm-tide inundation from Hurricane Sandy mapped (1) pre-storm, (2) on November 11, 2012, and (3) on February 14, 2013. The resulting depictions of estimated total building stock losses document how differing amounts of available USGS data affect the resolution and accuracy of storm-tide inundation extents. Using the most accurate results from the final (February 14, 2013) inundation extent, estimated losses range from $380 million to $5.9 billion for individual New York counties; total estimated aggregate losses are about $23 billion for all New York counties. Quality of the inundation extents used in HAZUS analyses has a substantial effect on final results. These findings can be used to inform future post-storm reconstruction planning and estimation of insurance claims.

A radiocarbon-calibrated aminochronology, based on the bivalve Mulinia lateralis, is presented for Chesapeake Bay core MD03-2661, a 25 m piston core drilled near Kent Island (38 degrees 53.21'N; 76 degrees 23.89'W) during the 2003 USGS Marion-Dufresne cruise. Three separate approaches were used to calibrate amino acid racemization (AAR) data for aspartic acid with radiocarbon data. For the first approach, a direct or paired analysis calibration incorporated eight articulated specimens, thereby allowing for the application of AAR and radiocarbon analysis of the same specimen and effectively eliminating both intrashell variability and time averaging as factors in the calibration. A second direct approach relied on valves that were bilaterally split to facilitate both AAR and radiocarbon dating, thus effectively eliminating time averaging effects from this calibration. For the third indirect approach, nine independent radiocarbon dates were combined with 129 Asx D/L ratios from the same core depths to produce an indirect calibration model, from which intershell variability and time averaging could be estimated. Variability in AAR ratios was recognized from a myriad of sources, including analytical error, intrashell variability, inherent variability, time averaging, and contamination. The majority of this variability was controlled for through experimental design or by the application of these three independent calibration approaches. The direct calibration of articulated shells and the indirect calibration yielded virtually identical age models, well within their respective 95% confidence intervals. This study establishes an aminostratigraphic reference section for the Holocene record of the Chesapeake Bay and demonstrates the usefulness of multiple calibration approaches and the potential utility of AAR for future studies of sedimentary processes and chronologies in the bay. (C) 2012 Elsevier B.V. All rights reserved.