Peter Whiting, PhD

Non-point source sediment and pollutant loading to the Great Lakes

The improvement in the water quality of Lake Erie has been trumpeted as one of the great environmental success stories. Since the 1960s, the lake has become less prone to eutrophication, water clarity has improved dramatically, and sensitive species are returning. While much of the improvement has been linked to reducing point and non-point source loads (particularly phosphorus), other important changes have occurred over the same time frame including climate, landuse and agricultural practices.

Several projects have examined the sediment loading to Lake Erie and other Great Lakes:

Climatic and Agricultural Factors in Nutrient Exports from Two Watersheds in Ohio

Abstract Export of agricultural nutrients and sediment to lakes and oceans is of great environmental concern in many agricultural watersheds. Recent years have seen efforts to reduce loads through agricultural practices such as conservation tillage, efficient fertilization, and reservation of erodible areas. Monitoring the efficacy of such efforts is complicated by the fact they take place against a varying climatic and hydrologic background. In this study, statistical analysis was used to identify those climatic, hydrologic, and agricultural variables that best explained variations in nitrate, phosphorus, and total suspended solids over the period 1976-95 in two large agricultural watersheds which feed Lake Erie, those of the Maumee and Sandusky Rivers. The dominant variable was stream discharge; after curvefits to remove its influence, the residual loads were tested via stepwise linear regression to reveal the most significant explanatory variables. Loads of nitrate, total suspended solids, and total phosphorus tended to decrease when previous months were wet, except in the summer, and to decrease when snow cover was extensive. It is speculated that stores of nitrate in the soil were lost during wet periods through increased crop uptake and/or leaching. Nitrogen fertilizer application in the Maumee watershed decreased following dry periods, but not enough to decrease stream loads. Soluble reactive phosphorus loads were negatively correlated to conservation tillage and reserves, and positively correlated to fertilizer and manure sources. Results for total phosphorus were similar to those for total suspended solids, on which most transported phosphorus is adsorbed.

Moog, D.B., and P.J. Whiting, 2002, Climatic and agricultural factors in export of nutrient loads from two watersheds in Ohio: Journal of Environmental Quality, v. 31, p. 72-83.

Climatic and Agricultural Contributions to Changing Loads in Two Watersheds in Ohio

Abstract Trends in climatic variables, streamflow, agricultural practices, and loads of nutrients and suspended solids were estimated for 1976-95 in the Maumee and Sandusky watersheds, two large agricultural basins draining to Lake Erie. To understand the contributions that various factors may have made to the trends in loads, earlier results of models linking loads to explanatory variables were combined with estimated trends in those variables. The study period was characterized by increases in temperature, wintertime precipitation and streamflow, conservation farming, and loads of nitrate and total suspended solids; decreases in snowfall and snow cover, fertilizer, manure from livestock, and loads of soluble reactive phosphorus; and relatively steady exports of total phosphorus. After removing the effects of trends in streamflow, nitrate loads increased much less while total suspended solids and total phosphorus loads declined. The analysis suggests that the nitrate increases were due largely to climatic factors, particularly increases in winter streamflow, decreases in snowfall and snow cover, and declining annual precipitation. Decreases in soluble reactive phosphorus were associated with changes in agricultural practices, particularly declines in fertilizer deliveries and head of livestock.

Moog, D.B., and P.J. Whiting, 2002, Climatic and agricultural contributions to changing loads in two watersheds in Ohio: Journal of Environmental Quality, v. 31, p.83-89.

Maumee-Sandusky ADAPT Implementation: A Distributed Model of Nitrate Export from Farm Fields to Lake Erie

Abstract The Maumee-Sandusky ADAPT Implementation (MSAI) was developed as a model of nitrogen export from farm fields in the Maumee River and Sandusky River watersheds, located mostly in northwestern Ohio. MSAI provides a tool for examining the impact of climate and farming practices – including tile drainage – on the amount of nitrate reaching Lake Erie that begins as nitrogen fertilizer. It is highly distributed in that it models over 8500 individual fields that together cover most of the cropped area in these watersheds. This fine discretization of farm data, made possible by a series of roadside transect surveys, adds accuracy and permits the model to take full advantage of increasingly precise geospatial data in categories such as soil type, land use, and weather. It also permits an examination of the spatial distribution of source areas for nitrate loading to Lake Erie, including the sensitivity of these areas to changes in climate and farming practice. Scenarios investigated for this report include changes in tile drainage, temperature, and precipitation. Overall, potential changes are modest. A value of less than 10% is predicted for: the decrease in nitrate export upon removal of all tile drains, the increase in export resulting from draining all fields, the decrease following a 2 °C temperature rise, and the increase accompanying a 30% increase in precipitation. The predicted decrease associated with a 20% precipitation decrease is greater, about 25%. Though overall sensitivities are modest, some subwatersheds are found to be more sensitive than the average. There is a tendency for downstream subwatersheds to be more sensitive and responsible for more nitrate export, largely because that exported nitrate has less time to decay during transport to Lake Erie, and also because the downstream soils tend to be less well drained naturally.

Report to the Lake Erie Commission, January 2004

Estimating TMDL background sediment loading to Great Lakes tributaries from existing data

Abstract The Total Maximum Daily Load (TMDL) for sediment is the maximum quantity of suspended sediment than can enter a waterway without affecting the beneficial uses of the waterway. It is calculated as the sum of all allotments of point sources of suspended sediment, all allotments of non-point sources of sediment, background (natural) loading of sediment, and a margin of safety. The goal of this project was to develop an estimate of background levels of sediment loads in tributaries of the Great Lakes. Such quantification is key to determining permissible TMDL in waters not meeting water quality standards under the Clean Water Act of 1972. We estimated sediment loading for 46 rivers from data collected at USGS gages. Landuse and physiographic attributes were estimated for the gaged basins with a GIS. Basin attributes and sediment yield data are the basis for approaches to estimating background sediment loads. While sediment yield was correlated to key basin attributes, these relationships wee not suitable for developing estimates of background loading. As an alternative, an envelope curve approach was developed.

Whiting, PJ., 2006, Estimating background sediment loading to Great Lakes tributaries from existing data: Journal of the American Water Resources Association, 42 (3), 69-776.

Using radionuclides to study surface processes

Radionuclides are powerful tools for determining rates of processes and illuminating the processes themselves.

We have used a very short-lived radionuclide (Be-7, 53 day half-life) to look at erosion from agricultural fields during single storms.

Depth and areal extent of sheet and rill erosion based on radionuclides in soils and suspended sediment

Abstract Sheetwash and rilling are two important mechanisms of soil erosion by runoff. The relative contribution of each mechanism has been a vexing question because measuring thin sheet erosion is difficult. Fortuitously, various fallout radionuclides have distinct distributions in the soil column thus different depths of erosion produce suspended sediment with unique radionuclide signatures. Those signatures can be used to estimate the depth and areal extent of sheet and rill erosion. We developed a model to execute multiple mass balances on soil and radionuclides to quantify these erosion mechanisms. Radionuclide activities (7Be, 137Cs, 210Pb) in the soil of a 6.03 ha agricultural field near Treynor, Iowa and in suspended sediment washed off the field during thunderstorm runoff were determined by gamma spectroscopy. Using the model, we examined 15.5 million possible combinations of the depth and areal extent of rill and sheet erosion. The best solution to the mass balances corresponded to rills eroding 0.38% of the basin to a depth of 35 mm and sheetwash eroding 37% of the basin to a depth of 0.012 mm. Rill erosion produced 29 times more sediment than sheet erosion.

Whiting, P.J., E.C. Bonniwell, and G. Matisoff, 2001, Depth and areal extent of sheet wash and rill erosion from radionuclides in soils and suspended sediment; Geology, v. 29, p. 1131-1134.

Short-term Erosion Rates from a 7Be Inventory Balance

Abstract Detailed soil erosion studies benefit from the ability to quantify the magnitude of erosion over time scales appropriate to the process. An inventory balance for 7Be was used to calculate sediment erosion in a 30.73 m2 plot during a series of runoff-producing thunderstorms occurring over three days at the Deep Loess Research Station in Treynor, IA. The inventory balance included determination of the pre- and post-storm 7Be inventories in the soil, the atmospheric influx of 7Be during the event, and profiles of the 7Be activity in the soil following the atmospheric deposition. The erosion calculated in the plot using the 7Be-inventory balance was 0.058 g/cm2, which is 23% of the annual average erosion determined using 137Cs inventories. The calculated erosion from the mass balance is similar to the 0.059 g/cm2 of erosion estimated from the amount of sediment collected at the outlet of the 6-ha field during the study period and the delivery ratio (0.64). The inventory balance of 7Be provides a new means for evaluating soil erosion over the time period most relevant to quantifying the prediction of erosion from runoff.

Wilson, C.G., G. Matisoff, and P.J. Whiting, 2003, Short-term erosion rates from a 7Be inventory balance: Earth Surface Processes and Landforms, v. 28(9), p. 967-977.

We have also used Be-7 and other fallout radionuclides (210Pb, and 137Cs ) to trace sediment and establish sediment transport distances in a variety of settings. Whether fine sediment (and adsorbed pollutants) carried into a stream channel moves with the water and to receiving waters in a few hydrographs or whether the fine sediment participates in many deposition and re-suspension episodes that greatly extends basin transit time will determine the area affected by pollutants and the pollutant toxicity.

Fine sediment residence times in rivers determined using fallout radionuclide

Abstract Targeting of erosion and pollution control programs is much more effective if the time for fine particles to be transported through a watershed, their travel distance, the proportions of old and new sediment in suspension, and the landscape erosion rate can be estimated. In this paper we present a novel technique for the tracing of suspended sediment in a mountain stream using fallout radionuclides sorbed to sediment. Atmospherically-delivered 7Be, 210Pb, and 137Cs accumulate in the snowpack, are released with its melting and sorb to fine particulates, a portion of which are carried downstream into stream channels. The half-life of cosmogenic 7Be is short (53.4 d) thus sediment residing on the stream bed should contain little of the radionuclide. The different signatures of newly delivered landscape sediment with its 7Be tag and the older untagged channel sediment is the basis for the tracing. The total flux of such radionuclides compared to the inventory in the soil permits estimates of rates of landscape erosion.

Fine suspended particulates in the Gold Fork River, Idaho, are transported downstream through the drainage in one or more steps having lengths of tens of kilometers. Step length decreases from about 60 km near the peak of the hydrograph to about 12 km near baseflow. The percent of sediment in suspension that is ‘new’ (i.e., recently delivered from the landscape) ranges from 96 to 12 %. The remaining sediment is resuspended older channel sediment. Residence times for particulates range from 1.6 days early in the hydrograph at the upper site to 103 days late in the hydrograph at the lowest elevation location. Erosion rates of fine sediment calculated from the flux of radionuclides average 0.0023 cm/yr. The long distance transport of fine particles suggests that delivery through the Gold Fork drainage to the basin outlet is fairly rapid once particles reach the channel and perhaps is also rapid in similar and smaller basins.

Bonniwell, E.C., G. Matisoff and P.J. Whiting, 1999, Fine sediment residence times in rivers determined using fallout radionuclides (7Be, 137Cs, 210Pb): Geomorphology, v. 27, p. 75-92.

Soil erosion and sediment sources in an Ohio watershed using beryllium-7, cesium-137, and lead-210

Abstract Soil cores and suspended sediments were collected within the, Old Woman Creek, Ohio (OWC) watershed following a thunderstorm and analyzed for Be-7, Cs-137, and Pb-210 activities to compare the effects of till vs. no-till management on soil erosion and sediment yield. The upper reaches of the watershed draining tilled agricultural fields were disproportionately responsible for the majority of the suspended sediment load compared with lower in the watershed (2.0-7.0 metric tons/km(2) [Mg/km(2)] vs. 1.2-2.6 Mg/km(2)). About 6 to 10 times more sediment was derived from the subbasins that are predominantly tilled (6.8-12.4 Mg/km(2)) compared with the subbasins undergoing no-till practices (0.5-1.1 Mg/km(2)). In undisturbed soils the Ph-210 activities decreased with movement toward the-bottom of the cores to the constant supported Pb-210 value at a depth of about 10 cm. There was a subsurface maximum in Cs-137 activity within the top 10 cm. In contrast, the Ph-210 and Cs-137 distributions. in soils that are currently or were previously filled were nearly homogeneous with depth, reflecting. continuing or previous mixing by plowing. The activities of Ph-210 and 7Be were linearly correlated and were higher in sus ended sediments derived from no-till subbasins than those derived from tilled subbasins, indicating that the soil surface is the source Of suspended sediment. The study demonstrates that no-till farming results in decreases in soil erosion and decreases in suspended sediment discharges and that those eroded sediments have a radionuclide signature corresponding to the tillage practice and the depth of erosion.

Matisoff, G., E.C. Bonniwell, and P.J. Whiting, 2002, Soil erosion and sediment sources in an Ohio watershed using Beryllium-7, Cesium-137, and Lead-210: Journal of Environmental Quality, v. 31, p. 54-61.

Radionuclides as indicators of sediment transport in agricultural watersheds that drain to Lake Erie

Abstract An issue in evaluating the success of agricultural management practices is the speed that eroded particles make: their way through the downstream waters. In this study at Old Woman Creek (OWC) and Rock Creek (RC), two largely agricultural watersheds in Ohio, the flux of sediment and radionuclides (Be-7, Pb-210, and Cs-137) in thunderstorm runoff was examined to better understand transport of eroded agricultural soils. The hydrograph in an agricultural area under no-till was similar in timing, but of lesser magnitude, than the hydrograph from a similar-sized area under conventional tillage. The activities of Pb-210 and Be-7 are linearly correlated and are higher in suspended sediments derived from no-till subbasins than those derived from conventionally tilled subbasins. A suspended sediment plume, identified by its unique radionuclide signature, was traced through 17 km of OWC stream channel in approximately 13.4 h (0.35 m/s). The downstream exponential decrease of Be-7 activities in suspended sediments 3 to 12 h after passage of the sediment plume was used to estimate transport distances of suspended sediment from 2 to 17 km,. respectively. Transport distances of suspended sediments were also calculated from wave kinematics and indicate that at OWC suspended sediment transport distances were longer in streams draining areas of no-till (19-26 km) than in the streams draining areas of conventional tillage (6-15 km). Suspended sediments travel 7 to 22 km at RC. The transport distances are long relative to the lengths of the stream channel and indicate that erosion control methods implemented in the watershed should be reflected quickly in downstream waters.

Matisoff, G., E.C. Bonniwell, and P.J. Whiting, 2002, Radionuclides as indicators of sediment transport in agricultural watersheds that drain to Lake Erie: Journal of Environmental Quality, v. 31, p. 62-72.

Suspended sediment sources and transport distances in the Yellowstone basin

Abstract The activity of fallout radionuclides (7Be, 210Pb, and 137Cs) was measured on upland and floodplain soils and on suspended sediments to quantify sources of fine sediment and to estimate sediment transport distances in stream channels in the Yellowstone River basin. Samples were collected seven times during snowmelt and runoff at nine locations from the headwaters of Soda Butte Creek to Billings, Montana, a 423-km-long reach of channel. The inventory of radionuclides in soil increases with precipitation and is highest in the headwaters. The activity of radionuclides in suspended sediment decreases downstream and more activity is observed earlier than later in the flood hydrograph.

The radionuclide activity of sediment derived from erosion of upland soils differs from that derived from bank erosion. Fine suspended sediment has an intermediate radionuclide signature that is quantified in terms of the relative contribution of these two sources of fine sediment. At sites high in the drainage, soils contribute 50% to the suspended load and this value decreases to 11-26% downstream. Fine sediment transport distances were calculated from the exponential decrease in radionuclide concentration below a point source. Transport distances increase from a few kilometers in the headwaters to hundreds of kilometers downstream. These estimates are consistent with transport distances estimated from the settling velocity of the particles and from the distribution of mine tailings downstream from a dam failure. This study of a large watershed confirms earlier results from smaller basins and suggests that transport distances increase with basin size.

Whiting, P.J., G. Matisoff, W. Fornes, and F. Soster, 2005, Suspended sediment sources and transport distances in the Yellowstone basin: Geological Society of America Bulletin, 515-529.

We have also been working with Mike Ketterer (NAU) to develop the technique for using Pu – another fallout radionuclide – to determine rates of soil erosion. The work is supported by NCASI and is taking place in the Murder Creek watershed in Georgia.

Understanding the flows that move the bulk of the bedload in gravel bed streams

My students and I have studied the sediment transport regime of about 25 gravel-bed streams in Idaho. The research is based upon long-term records of flow and bedload transport collected by the US Forest Service and is supplemented by our own work at these and other sites. In total, the data is perhaps the most extensive set of such information yet collected on gravel-bed streams. I have shown that the discharge moving the most bedload sediment over the long-term is close to the bankfull discharge, confirmed that the load is finer than the bed surface and subsurface, demonstrated clockwise hysteresis in bedload transport indicating sediment supply limitations in these armored gravel systems, and for the first time developed a technique for determining which sizes of sediment in a channel are limited in their transport by deficient supplies and which sizes are limited in their transport by stream hydraulics.

Sediment transporting flows in headwater streams

Abstract The equilibrium alluvial stream channel has a geometry that allows it to pass the water and sediment supplied from the watershed. At the same time, the equilibrium alluvial channel is built and maintained by the flows and sediment delivered to it. A prerequisite for understanding the creation of the equilibrium channel is an understanding of the sediment conveyance and competence of the flows the channel receives.

This study describes the bedload transport regime as it is linked to hydrology and geomorphology in twenty-three headwater gravel-bed streams in snowmelt-dominated parts of central and northern Idaho. At sites, drainage area ranges from 1.29 -381 km2, stream gradient ranges from 0.0043 – 0.0670 and median bed surface particle size ranges from 4 to 207 mm. Stream architecture includes riffle-pool, planar and step-pool beds.

The bedload is much finer than the surface and subsurface material suggesting selective transport of the finer sizes. Nonetheless, the majority of the load is sand at all flow discharges. Progressively coarser sediment was collected as flow discharge increased and painted rock experiments documented the transport of yet coarser particles at higher discharges. The supply of sediment to the streams appears limited as indicated by observed clockwise hysteresis in bedload transport rates during each spring snowmelt and by the coarse surface armor observed at sites.

Flows above bankfull discharge move 37 percent of the bedload whereas flows between mean annual discharge and bankfull move 57 percent of the bedload. The bedload effective discharge has a recurrence interval that averages 1.4 years and the magnitude of effective discharge averages 80 percent of bankfull discharge. The recurrence interval of bankfull discharge averages 2.0 years. The ratio of effective discharge to bankfull discharge is independent of basin size, grain size and gradient, although the ratio increases with the relative magnitude of large infrequent events.

Whiting, P.J., J.F. Stamm, D.B. Moog and R.L. Orndorff, 1999, Sediment transporting flows in headwater streams: Bull. Geol. Soc. America, v. 111, p. 450-466.

Annual Hysteresis In Bedload Rating Curves

Abstract The relationship between flow and bedload transport measured for ten years in six gravel-bed streams in Idaho exhibits annual hysteresis. At a given flow rate, more bedload is carried by discharges preceding the first annual occurrence of a “threshold” rate which is characteristic of each stream. Incorporating the effect of hysteresis leads to a better fit in the bedload-flow regression. As the turning point for hysteresis, a constant threshold discharge is found to work better than the annual peak discharge. This bimodal hysteresis model is also found to outperform one with a more gradual transition, based on cumulative discharge. These results are interpreted to reflect a buildup of readily-moved sediment supplies during the low-flow periods from late summer to early spring – supplies which are then exhausted by rising springtime discharges up to the threshold. The threshold is greater than mean annual discharge and about one-half bankfull discharge.

Moog, D.B., and P.J. Whiting, 1998, Annual hysteresis in bedload rating curves: Water Resources Research, v. 34, p. 2393-2399.

Surface Particle Sizes on Armored Gravel Streambeds: Effects of Supply and Hydraulics

Abstract Most gravel bed streams exhibit a surface armor in which the median grain size of the surface particles is coarser than that of the subsurface particles. This armor has been interpreted to result when the supply of sediment is less than the ability of the stream to move sediment. While there may be certain sizes in the bed for which the supply is less than the ability of the stream to transport these sizes, for other sizes of particles the supply may match or even exceed the ability of the channel to transport these particles. These sizes of particles are called ‘supply-limited’ and ‘hydraulically-limited’ in their transport, respectively, and can be differentiated in dimensionless sediment transport rating curves by size fractions. The supply- and hydraulically-limited sizes can be distinguished also by comparing the size of particles of the surface and subsurface. Those sizes that are supply-limited are winnowed from the bed and are under-represented in the surface layer. Progressive truncation of the surface and subsurface size distributions from the fine end and recalculation until the size distributions are similar (collapse), establishes the break between supply- and hydraulically-limited sizes. At sites along 12 streams in Idaho ranging in drainage area from about 100 to 4900 km2, sediment transport rating curves by size class and surface and subsurface size distributions were examined. The break between sizes that were supply- and hydraulically-limited as determined by examination of the transport rate and surface and subsurface size distributions was similar. The collapse size as described by its percentile in the cumulative size distribution averaged D36 of the surface and D73 of the subsurface. The discharge at which the collapse size began to move averaged 88 percent of bankfull discharge. The collapse size decreased as bed load yield increased and increased with the degree of selective transport

Whiting, P.J., and J.G. King, 2003, Surface particle sizes on armored gravel streambeds: effects of supply and hydraulics: Earth Surface Processes and Landforms, v. 28, p. 1459-1471.

How much water does a river need?

Much of my research on gravel bed streams was in support of federal water rights claims in the state of Idaho. In particular, my colleagues and I have answered the question – How much water is necessary to preserve a channel’s ability to transport all the sediment delivered to it which is ultimately the basis for maintaining a discrete channel? We have concluded that because of supply limitations in these armored gravel bed streams, the channel can be maintained with less than all the water. Any sediment not carried by unclaimed low flows will be moved by claimed higher flows that are capable of moving additional sediment if available. In a related effort, I have suggested an approach for estimating the flows that maintain the floodplain.

Geomorphic and hydrologic aspects of streamflow maintenance

Abstract In the last decades, insight from the fields of ecology, geomorphology and hydrology has been applied to the question of the streamflows necessary for environmental maintenance. For instance, spawning of salmon or trout requires how much water, for how long, at what time? Or what flows are necessary for the sustenance of streamside vegetation? Answers to these and similar questions have been sought to minimize environmental degradation in the development or re-licensing of water projects, in restoring riverine ecosystems and in balancing multiple uses for limited water resources.

In this contribution, the variety of environmental maintenance flows applied to rivers are described as are fundamental principles of environmental maintenance flows. These environmental maintenance flows include flows to maintain aesthetics and recreation

Whiting, P.J., 2002, Geomorphic and hydrologic aspects of streamflow maintenance: in Annual Review of Earth and Planetary Sciences, v. 30, 181-206.

Floodplain maintenance flows

Abstract A general understanding of river and floodplain geomorphology provides a foundation on which to build a description of the flows needed to maintain the floodplain. Floodplain maintenance concepts can be applied to the quantification of water-rights, the preservation of natural floodplain and their ecosystems, and the restoration of altered streams. An example of how to make an estimate of floodplain maintenance flows is provided. Along the East Branch of the Chagrin River in Ohio, floods with a recurrence interval of 4 years or greater are required to inundate the floodplain to a depth sufficient to transport onto the floodplain the fine sediment making up the floodplain.

Whiting, P.J., 1998, Floodplain maintenance flows: Rivers, v. 6, p. 160-170.

Other research interests/projects

A variety of other projects have interested me over the years

• Channel topography
• Flow over channel bars
• Storage and release of water from stream banks
• Extending flow records in time
• Spring-fed streams
• Channel restoration in urban streams
• Scientific communication