Retention of Riverine Sediment
and Nutrient Loads by Coastal Plain
Floodplains
Gregory B. Noe* and Cliff R. Hupp
U.S. Geological Survey, 430 National Center, Reston, Virginia 20192, USA
ABSTRACT
Despite the frequent citation of wetlands as effec-
tive regulators of water quality, few quantitative
estimates exist for their cumulative retention of the
annual river loads of nutrients or sediments. Here
we report measurements of sediment accretion and
associated carbon, nitrogen, and phosphorus accu-
mulation as sedimentation over feldspar marker
horizons placed on floodplains of the non-tidal,
freshwater Coastal Plain reaches of seven rivers in
the Chesapeake Bay watershed, USA. We then
scale these accumulation rates to the entire extent
of non-tidal floodplain in the Coastal Plain of each
river, defined as riparian area extending from the
Fall Line to the upper limit of tidal influence, and
compare them to annual river loads. Floodplains
accumulated a very large amount of material
compared to their annual river loads of sediment
(median among rivers = 119%), nitrogen (24%),
and phosphorus (59%). Systems with larger
floodplain areas and longer floodplain inundation
retained greater proportions of riverine loads of
nitrogen and phosphorus, but systems with larger
riverine loads retained a smaller proportion of that
load on floodplains. Although the source and long-
term fate of deposited sediment and associated
nutrients are uncertain, these fluxes represent the
interception of large amounts of material that
otherwise could have been exported downstream.
Coastal Plain floodplain ecosystems are important
regulators of sediment, carbon, and nutrient
transport in watersheds of the Chesapeake Bay.
Key words: floodplain; sediment; nitrogen;
phosphorus; retention; wetland; river.
INTRODUCTION
Floodplains and other riparian features are known
to be important locations for sediment storage in
fluvial systems (Phillips 1989; Steiger and others
2003). Large amounts of sediment eroded from
post-colonial agriculture in the US, and its associ-
ated nutrients, are thought to be stored in stream
valleys (Meade 1982; Walters and Merritts 2008).
The in-stream processes that entrain, transport, and
store the sediment retained in stream valleys are
important for understanding material transport in
watersheds, as suggested by the inability of upland
erosion rate estimates to predict current sediment
yields in the mid-Atlantic region of the US (Boomer
and others 2008). Although the literature contains
abundant references to the buffering function of
riparian areas (sediment and nutrient trapping
from upland flow), particularly in headwaters (for
example, National Research Council 2002), much
less literature has been devoted to the retention
function of floodplain ecosystems (trapping from
flood flow).
Received 8 December 2008; accepted 6 April 2009;
published online 22 May 2009
Authors Contributions C.H. designed and implemented the sampling
network. G.N. performed the chemical analyses of sediment, statistical
analyses, and the scaling and retention analyses. G.N. wrote the article
and both authors discussed the results and edited the manuscript.
*Corresponding author; e-mail: gnoe@usgs.gov
Ecosystems (2009) 12: 728–746
DOI: 10.1007/s10021-009-9253-5
Ó 2009 Government Employee: United States Geological Survey, Department of the Interior
728
The storage of sediment in floodplains has con-
sequences for the retention and cycling of nutrients
and contaminants. Sedimentation has been appre-
ciated as an important flux for phosphorus (P) in
wetlands (Johnston 1991). Although denitrifica-
tion has received considerable attention among
nitrogen (N) fluxes in floodplains, there is less
recognition of the potentially large magnitude of N
sedimentation fluxes (Olde Venterink and others
2006). Noe and Hupp (2005) reported high rates of
P as well as carbon (C) and N accumulation asso-
ciated with floodplain sedimentation along Coastal
Plain rivers. Large amounts of radioisotopes and
trace metals also accumulate in association with
sedimentation in floodplains (Hupp and others
1993; Malmon and others 2002).
Few quantitative estimates exist for the retention
of river loads by floodplains at the scale of catch-
ments or large river reaches. Here we provide
estimates for plot-scale sediment and sediment-
associated C, N, and P accumulation rates in
Coastal Plain floodplains located in the Chesapeake
Bay watershed, USA. Plot-scale accumulation rates
were first reported for three rivers in Noe and Hupp
(2005); in this study we report rates for an addi-
tional four rivers. For these seven rivers, we then
scale rates from individual plots to the entire extent
of floodplain along their freshwater, non-tidal
Coastal Plain reaches. Finally, we compare flood-
plain trapping to river loads to calculate the percent
retention of river sediment, N, and P loads by
floodplains. We expected that rivers with larger
floodplain area (size of sink), longer floodplain
inundation (river to floodplain exchange), and
larger river loads (size of source) had greater pro-
portions of their loads retained by floodplains.
METHODS
Site Description
Sediment deposition rates were measured at mul-
tiple sites along the freshwater Coastal Plain
reaches of seven rivers in the Chesapeake Bay
watershed (Figure 1; see Gellis and others 2008, for
more details). These rivers were chosen because
they have relatively broad, forested floodplains
characteristic of riverine wetlands on the Atlantic
and Gulf Coastal Plains. Other major Chesapeake
Bay tributaries (Susquehanna, Potomac, James,
Rappahannock, and York Rivers) are embayed to
the Fall Line and do not support large areas of
forested bottomlands (Hupp 2000), although their
Figure 1. Watershed
maps of the seven rivers
included in this study
with the locations of
sampling sites and river
input monitoring (RIM)
sites.
Retention of Riverine Sediment and Nutrient Loads 729
tributaries may have well-developed floodplains.
Study sites (1–6 sites per river) were selected to be
relatively easily accessed by road, support contig-
uous mature bottomland forest, generally experi-
ence hydroperiods typical of Coastal Plain rivers,
and have minimal tidal influence at the most
downstream site. Multiple floodplain transects
(typically three) were established at each study site,
oriented perpendicular to the river and separated
by 50–100 m, beginning on the channel edge
(usually a natural levee) and continuing for a few
hundred meters into the low backswamp area.
Each transect typically had 4–6 monitoring plots
where periodic measurements were made of sedi-
ment accretion rate and sediment texture and
composition. This sampling design systematically
encompassed all typical floodplain micro-topo-
graphic and vegetation features present at each
floodplain site.
For each river, the duration of floodplain inun-
dation over the period of sediment deposition
measurements was estimated by calculating the
number of days that river stage at the nearest gage
was above the elevation of the floodplain soil sur-
face (Table 1). The stage threshold for floodplain
inundation was determined from observations of
flooding at sites (Ross and others 2004) or either
cross-sectional surveys of stream and floodplain
topography (E. Doheney, USGS, personal commu-
nication) or the break in slope of the stage-dis-
charge rating curve at the nearby river gage. These
methods provide general estimates for the inun-
dation of floodplain along a river but ignore inter-
and intra-site variation in flooding due to variation
in river stage and floodplain elevation.
Sediment Accretion Rates
An artificial marker horizon was created at each
monitoring plot by placing powdered white feld-
spar clay to a depth of about 20 mm on about
0.5 m
2
of soil surface that had been cleared of
coarse organic detritus. The clay becomes a firm
marker after absorption of soil moisture that per-
mits accurate measurement of short-term net ver-
tical accretion above the clay surface (Baumann
and others 1984; Hupp and others 1993; Kleiss
1996). In general, the marker horizons were
examined for depth of burial annually and at se-
lected times after flooding events. Here we report
accretion rates (change in depth; m y
-1
) for the
latest sampling dates when the largest number of
marker horizons could be sampled for each site
(Table 1). These periods of record were not
concurrent for all rivers but ranged from 1996 to
2003. Depth of burial was measured by coring the
ground surface and measuring the vertical depth of
sediment above the marker horizon when surface
water was not present. Sediment adjacent to the
marker horizon was sampled to a depth corre-
sponding to the depth of deposition and then
analyzed for bulk density and percent organic
matter (loss on ignition).
Sediment and Nutrient Accumulation
Rates
Nutrient concentrations and mass accumulation
rates (change in mass; g m
-2
y
-1
) were directly
measured at a subset of sites where sediment
accretion was measured, along the Chickahominy,
Mattaponi, and Pocomoke Rivers (Noe and Hupp
2005). Two to four replicate cores were collected
above each marker horizon after coarse woody
debris larger than 2 mm in diameter was removed
from the coring location. The bulk accumulated
sediment above each marker horizon was com-
posited, transported to the laboratory, dried,
ground, analyzed for TN and TC by elemental
analysis (Carlo–Erba CHN analyzer), analyzed for
TP by microwave-assisted acid (HNO
3
, HCl, and
then HF) digestion and analysis on an ICP-OES
(Perkin–Elmer), and analyzed for organic content
by combustion at 400°C for 16 h (Noe and Hupp
2005). Preliminary analyses indicated that inor-
ganic C was a negligible proportion of TC. In
addition, we collected soil cores above marker
horizons along the Coastal Plain floodplain of the
Piedmont-dominated, alluvial Roanoke River in
North Carolina (Hupp and others 2009). The data
from the Roanoke River floodplain was included to
develop a more general database of floodplain
sediment characteristics that was used to parame-
terize regressions predicting soil nutrient concen-
trations at additional sites (see below). As with
sediment accretion rates, nutrient and sediment
accumulation rates measured this way represent
net rates over the duration of marker-horizon
deployment (see Noe and Hupp 2005).
Here we extend the accumulation rate estimates
to more sites along the seven rivers where only
sediment accretion was measured. This extension
was accomplished for sediment accumulation rates
using measured sediment bulk density at each site,
measured vertical accretion rates at each plot, and
known ages of marker-horizon deployment
(Equation 1), and for nutrient accumulation rates
730 G. B. Noe and C. R. Hupp
Table 1. Site Attributes and Accumulation Rates
River Site Floodplain
area (ha)
Duration
(year)
# of marker
horizons
Variable Accumulation rate:
Geometric mean
(g m
-2
y
-1
)
Accumulation
rate: 90% confidence
interval (g m
-2
y
-1
)
Chickahominy Upham Brook (UPH) 1259 3.1 9 Accretion (mm y
-1
) 4.8 1.7
Sediment 4600 2224
Mineral sediment 4152 2059
Organic sediment 424 178
Carbon 166 79
Nitrogen 11.5 4.5
Phosphorus 3.48 1.53
Bottoms Bridge (BBR) 2479 3.0 3 Accretion (mm y
-1
) 0.4 2.1
Sediment 1947 2219
Mineral sediment 1638 1454
Organic sediment 270 876
Carbon 121 426
Nitrogen 7.4 17.4
Phosphorus 1.15 4.48
Providence Forge (PRF) 1012 1.6 13 Accretion (mm y
-1
) 2.9 1.4
Sediment 2983 1433
Mineral sediment 2607 1252
Organic sediment 376 181
Carbon 164 79
Nitrogen 11.2 5.4
Phosphorus 2.46 1.18
Choptank Holiday Park (HOP) 163 1.0 11 Accretion (mm y
-1
) 3.0 1.6
Sediment 979 479
Mineral sediment 744 369
Organic sediment 232 111
Carbon 102 49
Nitrogen 5.2 2.5
Phosphorus 1.02 0.50
Red Bridges (REB) 144 0.9 16 Accretion (mm y
-1
) 1.1 0.3
Sediment 418 127
Mineral sediment 316 95
Organic sediment 91 37
Carbon 40 16
Nitrogen 2.2 0.8
Phosphorus 0.40 0.13
Retention of Riverine Sediment and Nutrient Loads 731
Table 1. Continued
River Site Floodplain
area (ha)
Duration
(year)
# of marker
horizons
Variable Accumulation rate:
Geometric mean
(g m
-2
y
-1
)
Accumulation
rate: 90% confidence
interval (g m
-2
y
-1
)
Dragon Run Mascot (MST) 1707 2.3 13 Accretion (mm y
-1
) 3.0 1.8
Sediment 2069 1258
Mineral sediment 1861 1135
Organic sediment 206 123
Carbon 89 53
Nitrogen 7.0 4.2
Phosphorus 1.43 0.86
Mattaponi Burkes Bridge (BRK) 1154 6.3 9 Accretion (mm y
-1
) 1.8 1.1
Sediment 705 292
Mineral sediment 542 280
Organic sediment 124 52
Carbon 59 27
Nitrogen 3.4 1.4
Phosphorus 0.37 0.14
Aylett (AYL) 977 2.3 11 Accretion (mm y
-1
) 3.7 0.9
Sediment 2409 581
Mineral sediment 1854 447
Organic sediment 555 134
Carbon 244 59
Nitrogen 12.6 3.0
Phosphorus 1.39 0.34
Pamunkey Engel Farm (ENF) 843 1.0 15 Accretion (mm y
-1
) 1.5 0.5
Sediment 698 226
Mineral sediment 592 191
Organic sediment 105 34
Carbon 46 15
Nitrogen 2.8 0.9
Phosphorus 0.64 0.21
Pampatike (PAM) 825 2.4 8 Accretion (mm y
-1
) 4.0 2.0
Sediment 1511 754
Mineral sediment 1289 681
Organic sediment 200 83
Carbon 87 36
Nitrogen 5.9 2.6
Phosphorus 1.20 0.49
732 G. B. Noe and C. R. Hupp
