Paradise Threatened: Land Use and Erosion on St. John, US Virgin Islands

2 /yr, respectively. Geomorphic evidence indicates that plantation agriculture during the 18th and 19th centuries did not cause severe erosion. Since about 1950 there has been rapid growth in roads and development due to increasing tourism and second-home development. Our field investigations identified the approximately 50 km of unpaved roads as the primary source of anthropogenic sediment. Field measurements of the road network in two catchments led to the development of a vector-based GIS model to predict road surface erosion and sediment delivery. We estimate that road erosion has caused at least a fourfold increase in island-wide sediment yields and that current sedimentation rates are unprecedented. Paving the dirt roads and implementing standard sediment control practices can greatly reduce current sediment yields and possible adverse effects on the marine ecosystems surrounding St. John.     

Effectiveness of BMPS in Reducing Sediment From Unpaved Roads in the Stillwater Creek,Oklahoma Watershed1

Turton, Donald J., Michael D. Smolen, and Elaine Stebler, 2009. Effectiveness of BMPs in Reducing Sediment From Unpaved Roads in the Stillwater Creek, Oklahoma Watershed. Journal of the American Water Resources Association (JAWRA) 45(6):1343‐1351. Abstract: Erosion from rural unpaved roads is thought to be an important source of sediment in sediment‐impaired streams in Oklahoma and other locations. However, no direct measurements of sediment yields from rural unpaved roads were previously available for Oklahoma. Four rural unpaved road segments in the Stillwater Creek Watershed were instrumented in a paired watershed design to measure sediment yields to streams before and after the installation of Best Management Practices (BMPs). One segment of each pair remained under current management to serve as a control. The second segment received BMPs after a 1‐year calibration period. One BMP consisted of widening the ditches, re‐shaping ditches and cutslopes, putting a proper crown on the road surface, and vegetating disturbed areas with grass. The other BMP consisted of creating a proper crown on the road bed, applying a geo‐synthetic fabric to the road bed and surfacing with 127 mm of crusher run gravel containing 12‐15% fines to serve as a binder. Road segment sediment yields for individual storms varied, depending on factors such as rainfall amount and intensity. During the pre‐BMP year, storm sediment yields ranged from 0 to 4.3 Mg on one pair of segments and from 0 to 2.8 Mg on the other. The storm sediment yields and annual yields were in the same order of magnitude as sediment yields from unpaved rural or forest roads reported in other studies. Sediment yields were significantly reduced on both segments by the installation of BMPs, approximately 80% on one segment pair and 20% on the other. The average sediment yield (across the four segments) for the pre‐BMP year was 138 Mg/ha or 120 Mg/km of road. By extrapolating these average yields across the 479 km of unpaved roads in the Stillwater Creek Watershed and comparing it to estimated sediment yields for other land uses obtained from other sources, we conclude that unpaved roads may contribute up to 35% of the total sediment load to Stillwater Creek.     

“R1-R4” AND “BOISED” SEDIMENT PREDICTION MODEL TESTS USING FOREST ROADS IN GRANITICS1

ABSTRACT: Erosion and sedimentation data from research watersheds in the Silver Creek Study Area in central Idaho were used to test the prediction of logging road erosion using the R1-R4 sediment yield model, and sediment delivery using the “BOISED” sediment yield prediction model. Three small watersheds were instrumented and monitored such that erosion from newly constructed roads and sediment delivery to the mouths of the watersheds could be measured for four years following road construction. The errors for annual surface erosion predictions for the two standard road tests ranged from +31.2 t/ha/yr (+15 percent) to -30.3 t/ha/yr (-63 percent) with an average of zero t/ha/yr and a standard deviation of the differences of 18.7 t/ha/yr. The annual prediction errors for the three watershed scale tests had a greater range from -40.8 t/ha/yr (-70 percent) to +65.3 t/ha/yr (+38 percent) with a mean of -1.9 t/ha/yr and a standard deviation of the differences of 25.2 t/ha/yr. Sediment yields predicted by BOISED (watershed scale tests) were consistently greater (average of 2.5 times) than measured sediment yields. Hillslope sediment delivery coefficients in BOISED appear to be overly conservative to account for average site conditions and road locations, and thus over-predict sediment delivery. Mass erosion predictions from BOISED appear to predict volume well (465 tonnes actual versus 710 tonnes predicted, or a 35 percent difference) over 15 to 20 years, however mass wasting is more episodic than the model predicts.     

Sediment Production From Forest Roads in Western Montana1

Abstract: Unpaved roads are a primary sediment source in forested watersheds. Validation of erosion models and improvements to road management require information on road erosion rates and the factors controlling erosion. This study measured sediment yields from twenty ～0.05 ha unsurfaced (native) road plots in Belt Supergroup and glacial till parent materials of western Montana, and investigated the factors controlling erosion. Annual sediment yields for individual plots ranged from 0 to 96.9 Mg/ha/yr over 3 years (2002‐2004). Annual mean sediment yield ranged from 2.1 Mg/ha in 2003 to 9.9 Mg/ha in 2004 with an overall mean of 5.4 Mg/ha/yr. The mean of log‐transformed sediment yields for sites in glacial till parent materials was higher than Belt Supergroup parent materials (p = 0.063). A regression model with road slope, time since last grading, roadbed gravel content, and precipitation as predictive variables explained 68% of the variability in sediment yield (F = 28.2; p < 0.0001). Road erosion in western Montana is limited by low erodibility of the dominant parent materials and low rainfall. Management procedures such as reducing the frequency of grading can significantly reduce sediment yields from forest roads.     

Effectiveness of a Coral-Derived Surfacing Material for Reducing Sediment Production on Unpaved Roads, Schoffield Barracks, Oahu, Hawaii

This study evaluated the effectiveness of two application rates of a coral-derived surfacing material for both traffic and nontraffic road conditions using simulated rainfall (110–120 mm h⁻¹ for 30–90 min) on 0.75-m (wide) × 5.0-m (long) plots of similar slope (roughly 0.1 m m⁻¹). The coral is a locally available material that has been applied to unpaved roads surfaces on Schoffield Barracks, Oahu, Hawaii (USA), where this experiment was conducted. The simulations show that compared with a bare control plot, the coral-based surface application rates of 80 and 160 kg m⁻² (equivalent to only 10- and 20-mm thicknesses) reduced road sediment production by 75% and 95%, respectively, for nontraffic conditions. However, after two passes of the research vehicle during wet conditions, sediment production rates for the two coral treatments were not significantly different from those on the bare road plots. The overall effectiveness of the coral-derived surfacing material is unsatisfactory, primarily because the on-road surface thickness associated with the application rates tested was too small. These rates were selected to bracket those applied to training roads in the study area. Furthermore, the composition of the coral-based material does not facilitate the development of a sealed, erosion-resistant surface. When applied at the low rates tested, the coral material breaks down under normal traffic conditions, thereby losing its ability to counter shearing forces exerted by overland flow on long hillslopes where erosion measures are most needed. These simulations, combined with observations on roads in the study area, indicate that this material is not an appropriate road surfacing material for the site—at least for the low application rates examined. These results are preliminary; extended testing of higher applications rates at the hillslope scale under natural climate and traffic conditions is needed to better judge the effectiveness of this material over time.     

Derivation of a GIS-based watershed-scale conceptual model for the St. Jones River Delaware from habitat-scale conceptual models

Conceptual modeling is a useful tool for identifying pathways between drivers, stressors, Valued Ecosystem Components (VECs), and services that are central to understanding how an ecosystem operates. The St. Jones River watershed, DE is a complex ecosystem, and because management decisions must include ecological, social, political, and economic considerations, a conceptual model is a good tool for accommodating the full range of inputs. In 2002, a Four-Component, Level 1 conceptual model was formed for the key habitats of the St. Jones River watershed, but since the habitat level of resolution is too fine for some important watershed-scale issues we developed a functional watershed-scale model using the existing narrowed habitat-scale models. The narrowed habitat-scale conceptual models and associated matrices developed by Reiter et al. (2006) were combined with data from the 2002 land use/land cover (LULC) GIS-based maps of Kent County in Delaware to assemble a diagrammatic and numerical watershed-scale conceptual model incorporating the calculated weight of each habitat within the watershed. The numerical component of the assembled watershed model was subsequently subjected to the same Monte Carlo narrowing methodology used for the habitat versions to refine the diagrammatic component of the watershed-scale model. The narrowed numerical representation of the model was used to generate forecasts for changes in the parameters “Agriculture” and “Forest”, showing that land use changes in these habitats propagated through the results of the model by the weighting factor. Also, the narrowed watershed-scale conceptual model identified some key parameters upon which to focus research attention and management decisions at the watershed scale. The forecast and simulation results seemed to indicate that the watershed-scale conceptual model does lead to different conclusions than the habitat-scale conceptual models for some issues at the larger watershed scale.     

Sediment Availability and Erosion Rates on Off‐Highway Vehicle Trails in the Ouachita Mountains,USA

Factors influencing sediment availability are assessed and erosion rates are quantified for an off‐highway vehicle (OHV) trail system in the Ouachita Mountains of Arkansas. As of May 2012, the Wolf Pen Gap trail system included 77.0 km of "trails" which consist of county roads; open and closed Forest Service roads; and open and closed OHV trails. For a given trail length, the sediment volume available to be eroded is determined by bare trail width and sediment depth. Four condition types are defined that group trail sections based on statistically different trail widths or depths. Trail construction method appears to influence sediment availability differences more than erosion potential (as indexed by trail slope gradient and length). The range for annual trail erosion rates is estimated as 75 and 210 tonne/ha/yr. The high and low rates are obtained using two independent methods. The 210 tonne/ha/yr rate is computed from mean sediment capture at 30 sediment traps installed for 0.5–1.0 year. The 75 tonne/ha/yr rate is computed assuming all available trail sediment measured in a one‐time sampling is eroded over the next year. We argue in support of this assumption and suggest both rate values may be conservative. Trail erosion rates and sediment trap observations indicate frequent trap cleanout will be needed to continue sediment capture from All Terrain Vehicle trails.     

A Method for Spatially Explicit Representation of Sub-watershed Sediment Yield,Southern California,USA

We present here a method to integrate geologic, topographic, and land-cover data in a geographic information system to provide a fine-scale, spatially explicit prediction of sediment yield to support management applications. The method is fundamentally qualitative but can be quantified using preexisting sediment-yield data, where available, to verify predictions using other independent data sets. In the 674-km² Sespe Creek watershed of southern California, 30 unique “geomorphic landscape units” (GLUs, defined by relatively homogenous areas of geology, hillslope gradient, and land cover) provide a framework for discriminating relative rates of sediment yield across this landscape. Field observations define three broad groupings of GLUs that are well-associated with types, relative magnitudes, and rates of erosion processes. These relative rates were then quantified using sediment-removal data from nearby debris basins, which allow relatively low-precision but robust calculations of both local and whole-watershed sediment yields, based on the key assumption that minimal sediment storage throughout most of the watershed supports near-equivalency of long-term rates of hillslope sediment production and watershed sediment yield. The accuracy of these calculations can be independently assessed using geologically inferred uplift rates and integrated suspended sediment measurements from mainstem Sespe Creek, which indicate watershed-averaged erosion rates between about 0.6–1.0 mm year^?1 and corresponding sediment yields of about 2 × 10³ t km^?2 year^?1. A spatially explicit representation of sediment production is particularly useful in a region where wildfires, rapid urban development, and the downstream delivery of upstream sediment loads are critical drivers of both geomorphic processes and land-use management.     

IMPACT OF COAL SURFACE MINING AND RECLAMATION ON SUSPENDED SEDIMENT IN THREE OHIO WATERSHEDS1

ABSTRACT: Prior to PL95–87 little research had been conducted to determine the impacts of mining and reclamation practices on sediment concentrations and yields on a watershed scale. Furthermore, it was unknown whether sediment yield and other variables would return to undisturbed levels after reclamation. Therefore, three small watersheds, with differing lithologies and soils, were monitored for runoff and suspended sediment concentrations during three phases of watershed disturbances: undisturbed watershed condition, mining and reclamation disturbances, and post‐reclaimed condition. Profound increases in suspended‐sediment concentrations, load rates, and yields due to mining and reclamation activities, and subsequent drastic decreases after reclamation were documented. Even with increases in runoff potential, reductions in suspended‐sediment concentrations and load rates to below or near undisturbed‐watershed levels is possible by using the mulch‐crimping technique and by removing diversions. Maximum concentrations and load rates occurred during times of active disturbances that exposed loose soil and spoil to high‐intensity rains. Sediment concentrations remained elevated compared with the undisturbed watershed when diversions were not well maintained and overtopped, and when they were not removed for final reclamation. Diversions are useful for vegetation establishment, but should be maintained until they are removed for final reclamation after good vegetative cover is established.     

Empirical models based on the universal soil loss equation fail to predict sediment discharges from Chesapeake Bay catchments

The Universal Soil Loss Equation (USLE) and its derivatives are widely used for identifying watersheds with a high potential for degrading stream water quality. We compared sediment yields estimated from regional application of the USLE, the automated revised RUSLE2, and five sediment delivery ratio algorithms to measured annual average sediment delivery in 78 catchments of the Chesapeake Bay watershed. We did the same comparisons for another 23 catchments monitored by the USGS. Predictions exceeded observed sediment yields by more than 100% and were highly correlated with USLE erosion predictions (Pearson r range, 0.73-0.92; p < 0.001). RUSLE2-erosion estimates were highly correlated with USLE estimates (r = 0.87; p < 001), so the method of implementing the USLE model did not change the results. In ranked comparisons between observed and predicted sediment yields, the models failed to identify catchments with higher yields (r range, -0.28-0.00; p > 0.14). In a multiple regression analysis, soil erodibility, log (stream flow), basin shape (topographic relief ratio), the square-root transformed proportion of forest, and occurrence in the Appalachian Plateau province explained 55% of the observed variance in measured suspended sediment loads, but the model performed poorly (r(2) = 0.06) at predicting loads in the 23 USGS watersheds not used in fitting the model. The use of USLE or multiple regression models to predict sediment yields is not advisable despite their present widespread application. Integrated watershed models based on the USLE may also be unsuitable for making management decisions.     

Effects of sedimentation on soil nutrient dynamics in riparian forests

The influence of sedimentation rates on biogeochemistry of riparian forests was studied near ephemeral streams at Fort Benning, GA. Upper reaches of seven ephemeral streams had received varying rates of sedimentation stemming from erosion along unpaved roadways at the military installation. Two reference catchments were also included in the study. Decomposition of foliar litter, microbial C and N, N mineralization, and arthropod populations were compared within and among catchments. Rates of sedimentation over the past 25 yr ranged from 0 in references to 4.0 cm yr(-1). Decomposition rates declined exponentially with sedimentation rates as low as 0.20 to 0.32 cm yr(-1) and appeared to reach an equilibrium at a sedimentation rate of 0.5 cm yr(-1). Nitrogen mineralization and microbial C and N followed the same trend. Sedimentation had no discernible effect on arthropod populations. These data suggest that biogeochemical cycles may be altered by sedimentation rates that commonly occur in some floodplain forests.     

RESERVOIR SEDIMENTATION RATES LINKED TO LONG-TERM CHANGES IN AGRICULTURAL LANI) USE1

ABSTRACT: Long-term land use and reservoir sedimentation were quantified and linked in a small agricultural reservoir-watershed system without having historical data. Land use was determined from a time sequence of aerial photographs, and reservoir sedimentation was determined from cores with ¹³⁷Cs dating techniques. They were linked by relating sediment deposition to potential sediment production which was determined by the Universal Soil Loss Equation and by SCS estimates for gullied land. Sediment cores were collected from Tecumseh Lake, a 55-ha reservoir with a 1,189-ha agricultural watershed, constructed in 1934 in central Oklahoma. Reservoir sediment deposition decreased from an average of 5,933 Mg/yr from 1934 to 1954, to 3,179 Mg/yr from 1954 to 1962, and finally to 1,017 Mg/yr from 1962 to 1987. Potential sediment production decreased from an average of 29,892 to 11,122 and then to 3,589 Mg/yr for the same time periods as above, respectively. Reductions in deposition and sediment production corresponded to reductions in cultivated and abandoned cropland which became perennial pasture. Together, cultivated and abandoned cropland accounted for 59 percent of the watershed in 1937, 24 percent in 1954, and 10 percent in 1962. Roadway erosion, stream bank erosion, stored stream channel sediment, and long-term precipitation were considered, but none seemed to play a significant role in changing sediment deposition rates. Instead, the dominant factor was the conversion of fields to perennial pastures. The effect of conservation measures on reservoir sedimentation can now be quantified for many reservoirs where historical data is not available.     

PERSISTING SEDIMENT YIELDS AND SEDIMENT DELWERY CHANGES1

ABSTRACT: The Buffalo River is a tributary to the Mississippi River in west-central Wisconsin that drains a watershed dominated by agricultural land uses. Since 1935, backwater from Lock and Dam 4 on the Mississippi River has inundated the mouth of the Buffalo's valley. Resurveys of a transect first surveyed across the lake in 1935 and cesium-137 dating of backwater sediments reveal that sedimentation rates at the Buffalo's mouth have remained unchanged since the mid-1940s. Study results indicate that sediment yields from the watershed have persisted at relatively high levels over a period of several decades despite pronounced trends toward less cultivated land and major efforts to control soil erosion from agricultural land. The maintenance of sediment yields is probably due to increased channel conveyance capacities resulting from incision along some tributary streams since the early 1950s. Post-1950 incision extended the network of historical incised tributary channels, enhancing the efficient delivery of sediment from upland sources to downstream sites.     

Applications of Turbidity Monitoring to Forest Management in California

Many California streams have been adversely affected by sedimentation caused by historic and current land uses, including timber harvesting. The impacts of timber harvesting and logging transportation systems on erosion and sediment delivery can be directly measured, modeled, or inferred from water quality measurements. California regulatory agencies, researchers, and land owners have adopted turbidity monitoring to determine effects of forest management practices on suspended sediment loads and water quality at watershed, project, and site scales. Watershed-scale trends in sediment discharge and responses to current forest practices may be estimated from data collected at automated sampling stations that measure turbidity, stream flow, suspended sediment concentrations, and other water quality parameters. Future results from these studies will provide a basis for assessing the effectiveness of modern forest practice regulations in protecting water quality. At the project scale, manual sampling of water column turbidity during high stream flow events within and downstream from active timber harvest plans can identify emerging sediment sources. Remedial actions can then be taken by managers to prevent or mitigate water quality impacts. At the site scale, manual turbidity sampling during storms or high stream flow events at sites located upstream and downstream from new, upgraded, or decommissioned stream crossings has proven to be a valuable way to determine whether measures taken to prevent post-construction erosion and sediment production are effective. Turbidity monitoring at the project and site scales is therefore an important tool for adaptive management. Uncertainty regarding the effects of current forest practices must be resolved through watershed-scale experiments. In the short term, this uncertainty will stimulate increased use of project and site-scale monitoring.     

Road Erosion and Delivery Index (READI): A Model for Evaluating Unpaved Road Erosion and Stream Sediment Delivery

The Road Erosion and Delivery Index (READI) is a new geographic information system–based model to assess erosion and delivery of water and sediment from unpaved road networks to streams. READI quantifies the effectiveness of existing road surfacing and drain placements in reducing road sediment delivery and guides upgrades to optimize future reductions. Roads are draped on a digital elevation model and parsed into hydrologically distinct segments. Segments are further divided by engineered drainage structures. For each segment, a kinematic wave approximation generates runoff hydrographs for specified storms, with discharge directly to streams at road–stream crossings and onto overland‐flow plumes at other discharge points. Plumes are attenuated by soil infiltration, which limits their length, with delivery occurring if plumes intersect streams. Sediment production and sediment delivery can be calculated as a relative dimensionless index. READI predicts only a small proportion of new drains and new surfacing results in the majority of sediment delivery reductions. The model illustrates how the spatial relationships between road and stream networks, controlled by topography and network geometries, influence patterns of road–stream connectivity. READI was applied in seven northern California basins. The model was also applied in a recent burn area to examine how reduced hillslope infiltration can result in increased hydrologic connectivity and sediment delivery.     

Regional scale variability in sediment and nutrient delivery from small agricultural watersheds

Although many studies have pointed out the various controlling factors of sediment and nutrient delivery on a plot or watershed scale, little is known on the spatial variability of sediment and nutrient delivery on a regional scale. This study was conducted to reveal regional variations in sediment-associated nutrient delivery in central Belgium. Sediment deposited in 13 small retention ponds was sampled and analyzed for total phosphorus (TP), K, Mg, and Ca content. The TP content of the sediment deposits varied from 510 to 2001 mg P per kg sediment. Nutrients are predominantly fixed on the very fine sediment fraction (<16 microm), which is the reason why the nutrient trap efficiency of the ponds is only a fraction of the sediment trap efficiency. Average nutrient trap efficiency of the studied ponds varies between 4 and 31%, whereas sediment trap efficiency varies between 10 and 72%. For watersheds ranging from 7 to 4873 ha, sediment yield ranged between 1.2 and 20.6 Mg ha(-1) yr(-1), whereas TP export varied from 1.8 to 39.7 kg ha(-1) yr(-1). The observed spatial variability in nutrient losses is primarily attributed to regional variations in erosion and sediment yield values and to a far lesser degree to the spatial variations in fertilizer application. Redistribution of manure in the framework of an agricultural policy may increase the rate of nutrient delivery by ways of erosion and sediment transport.     

Evaluating agricultural best management practices in tile-drained subwatersheds of the Mackinaw River, Illinois

Best management practices (BMPs) are widely promoted in agricultural watersheds as a means of improving water quality and ameliorating altered hydrology. We used a paired watershed approach to evaluate whether focused outreach could increase BMP implementation rates and whether BMPs could induce watershed-scale (4000 ha) changes in nutrients, suspended sediment concentrations, or hydrology in an agricultural watershed in central Illinois. Land use was >90% row crop agriculture with extensive subsurface tile drainage. Outreach successfully increased BMP implementation rates for grassed waterways, stream buffers, and strip-tillage within the treatment watershed, which are designed to reduce surface runoff and soil erosion. No significant changes in nitrate-nitrogen (NO-N), total phosphorus (TP), dissolved reactive phosphorus, total suspended sediment (TSS), or hydrology were observed after implementation of these BMPs over 7 yr of monitoring. Annual NO-N export (39-299 Mg) in the two watersheds was equally exported during baseflow and stormflow. Mean annual TP export was similar between the watersheds (3.8 Mg) and was greater for TSS in the treatment (1626 ± 497 Mg) than in the reference (940 ± 327 Mg) watershed. Export of TP and TSS was primarily due to stormflow (>85%). Results suggest that the BMPs established during this study were not adequate to override nutrient export from subsurface drainage tiles. Conservation planning in tile-drained agricultural watersheds will require a combination of surface-water BMPs and conservation practices that intercept and retain subsurface agricultural runoff. Our study emphasizes the need to measure conservation outcomes and not just implementation rates of conservation practices.     

Patterns of short-term sedimentation in a freshwater created marsh

This study investigated different sedimentation measurement techniques and examined patterns of short-term sedimentation in two 1-ha replicate created freshwater marshes in central Ohio, USA. Short-term (one-year) sediment accumulation above feldspar, clay, glitter, and sand artificial marker horizons was compared at different water depths and distances from wetland inflow. A sediment budget was also constructed from turbidity and suspended sediment data for comparison with marker horizons. Glitter and sand marker horizons were the most successful for measuring sediment accumulation (81-100% marker recovery), while clay markers were completely unsuccessful. The sedimentation rate for both wetlands averaged 4.9 cm yr(-1) (36 kg m(-2) yr(-1)), and ranged from 1.82 to 9.23 cm yr(-1) (12.4 to 69.7 kg m(-2) yr(-1)). Sedimentation rates in deep, open water areas were significantly higher than in shallow, vegetated areas for both wetlands (t test, p < 0.05). However, observed sedimentation patterns may be attributed more to preferential flow through open water areas than to water depth or presence of vegetation. Contrary to the expected spatial distribution, sedimentation was highly variable within the wetlands, suggesting that bioturbation and turbulence may cause significant resuspension or that high hydrologic loads may distribute sediments throughout the basins. A sediment budget estimated sediment retention of approximately 740 g m(-2) yr(-1) per wetland (43% removal rate), yet gross sediment accumulation was 36,000 g m(-2) yr(-1) measured by marker horizons. These results suggest that erosive forces may have influenced sedimentation, but also may indicate problems with the sediment budget calculation methodology.     

Nitrogen loads through baseflow, stormflow, and underflow to Rehoboth Bay, Delaware

A detailed study of water and nitrogen (N) discharge from a small, representative subwatershed of Rehoboth Bay, Delaware, was conducted to determine total N loads to the bay. The concentrations of ammonium (NH4(+)), nitrate + nitrite (NO3(-) + NO2(-)), and dissolved and particulate organic N were determined in baseflow and storm waters discharging from Bundicks Branch from October 1998 to April 2002. A novel hydrographic separation model that accounts for significant decreases in baseflow during storm events was developed to estimate N loads during unsampled storms. Nitrogen loads based on gauged flows alone (7100-19,100 kg/yr) significantly underestimated those based on land use-land cover (LULC) and estimated N export factors from different classes of LULC (32,000-40,600 kg/yr). However, when ungauged underflow and associated N loads were included in the total loads (25,500-33,800 kg/yr), there was much better agreement with LULC export models. This suggests that in permeable coastal plain sediments, underflow contributes significantly to N fluxes to estuarine receiving waters, particularly in drier years. Based on the similarity in LULC, N loads from the Bundicks Branch subwatershed were used to estimate upland loads to the entire Rehoboth Bay Watershed (259,000-316,000 kg/yr). These N loads from the watershed were much greater than those from direct atmospheric deposition (49,000-64,500 kg/yr) and from a local wastewater treatment plant (9700-13,700 kg/yr). While the watershed was the principal source of N at all times during the year, the relative contributions from the watershed, wastewater, and direct atmospheric deposition varied predictably with season.