Headwaters to the Sea—EMAP’s Estimate of Nutrient Transport
in the Mississippi River Basin
Brian H HillUS Environmental Protection AgencyOffice of Research and Development
National Health and Environmental Effects LaboratoryMid-Continent Ecology Division
Duluth, MN
•Wadeable Stream Assessment(WSA, 510 sites)
•Great Rivers Ecosystems(GRE, 337 sites)
WSA
EMAP-GRE
Mainstem Rivers
0
1000
2000
3000
4000
5000
6000
Tota
l P, μ
g L-
1
0
50
100
150
200
250
Tributaries
Tota
l N, μ
g L-
1
0
1000
2000
3000
4000
5000
6000
Upper Mississippi RiverMissouri RiverOhio RiverLower Mississippi River
0
50
100
150
200
250
N &P in basin tributaries vs. rivers
Total N, mg L-10 1 2 3 4 5 6 7
Dis
solv
ed in
orga
nic
N, m
g L-
1
0
1
2
3
4
5
6
7
Mississippi River (r2=0.93)Missouri River (r2=0.87)Ohio River (r2=0.83)
Total P, mg L-10.0 0.1 0.2 0.3 0.4 0.5
Dis
solv
ed in
orga
nic
P, m
g L-
1
0.0
0.1
0.2
0.3
0.4
0.5
Mississippi River (r2=0.44)Missouri River (r2=0.69)Ohio River (r2=0.31)
Dissolved vs. total nutrients
TN and DIN areinterchangeable
DIP grossly underestimates
available P
N:P0.001 0.01 0.1 1 10 100 1000
C:N
0.1
1
10
100
1000
N-limited
C-limited
P-limited
Great River mainstems
Nutrient stoichiometry—tributaries vs. mainstem rivers
N:P0.001 0.01 0.1 1 10 100 1000
C:N
0.1
1
10
100
1000
Upper Mississippi River tributariesMissouri River tributariesOhio River tributariesLower Mississippi River tributaries
N-limited
C-limited
P-limited
Tributaries
Redfield ratio—N:P 16:1C:N 8:1
N vs. land use for small streams
Agr
icul
tura
l lan
d us
e, %
of w
ater
shed
are
a
0
10
20
30
40
50 Upper Mississippi River tributariesMissouri River tributariesOhio River tributariesLower Mississippi River tributaries
Agricultural land use in the basin
Agricultural land use, % of watershed area0 20 40 60 80 100
Tota
l N, μ
g L-
1
1e+1
1e+2
1e+3
1e+4
1e+5
Uppper Mississippi River basin, r=0.59Missouri River basin, r=0.73Ohio River basin, r=0.57Lower Mississippi River basin, r=0.49
Stream N as a function of land use
Agricultural land use, % of watershed area0 20 40 60 80 100
Tota
l P, μ
g L-
1
1
10
100
1000
10000
Upper Mississippi River basin, r=-0.10Missouri River basin, r=0.63Ohio River basin, r=0.55Lower Mississippi River basin, r=0.48
Stream P as a function of land use
NHDGreatly improved 1:100K National
Hydrography Dataset
PLUS (9 more components):• Value Added Attributes• Elevation-based Catchments• Catchment Characteristics (NLCD)• Headwater Node Areas• Cumulative Drainage Area Characteristics• Flow Direction, Flow Accumulation, and
Elevation Grids• Min/Max Elevations and Slopes• Flow Volume & Velocity Estimates• Flow Gages with Network Locations
NHDPlus—linking the stream network to the landscape
National Hydrography
Dataset (NHD)
National Elevation Dataset (NED)
Watershed Boundary Dataset (WBD)
A joint USGS-USEPA venture
Watershed area, log10 km21e-1 1e+0 1e+1 1e+2 1e+3 1e+4 1e+5 1e+6 1e+7
Mea
n an
nual
dis
char
ge, l
og10
m3 s
-1
1e-5
1e-4
1e-3
1e-2
1e-1
1e+0
1e+1
1e+2
1e+3
1e+4
1e+5
1e+6
r2=0.84
Runoff as a function of watershed area
Watershed area, log10
km20 1 2 3 4 5 6 7
N lo
ad, l
og10
T y
-1
0
1
2
3
4
Upper Mississippi River tributarieslog10 N load=0.96*log10 WA, r2=0.983Missouri River tributarieslog10 N load=0.40*log10 WA, r2=0.796Ohio River tributarieslog10 N load=0.850*log10 WA, r2=0.92Lower Mississippi River tributarieslog10 N load=0.63*log10 WA, r2=0.91
Load=mean annual discharge x mean nutrient concentration
Watershed area, log10 km20 1 2 3 4 5 6 7
P lo
ad, l
og10
T y
-1
0
1
2
3
4
Upper Mississippi River tributarieslog10 P load=0.41*log10 WA, r2=0.78Missouri River tributarieslog10 P load=0.22*log10 WA, r2=0.65Ohio River tributarieslog10 P load=0.36*log10 WA, r2=0.65Lower Mississippi River tributarieslog10 P load=0.40*log10 WA, r2=0.70
Load=mean annual discharge x mean nutrient concentration
Watershed area, log10 km20 1 2 3 4 5 6 7
N lo
ad, l
og10
T y
-1
0
1
2
3
4
5
6
7
P lo
ad, l
og10
T y
-1
0
1
2
3
4
5
6
log10 N load=0.94*log10 WA, r2=0.92
N export=1,233,673 T y-1
log10 P load=0.78*log10 WA, r2=0.88
P export=113,344 T y-1
Load=mean annual discharge x mean nutrient concentration
Predicted tributary N load, T y-11e+2 1e+3 1e+4 1e+5 1e+6
Mea
sure
d m
ains
tem
N lo
ad, T
y-1
1e+2
1e+3
1e+4
1e+5
1e+6
Predicted tributary P load, T y-11e+1 1e+2 1e+3 1e+4 1e+5
Mea
sure
d m
ains
tem
P lo
ad, T
y-1
1e+2
1e+3
1e+4
1e+5
Observed vs. predicted N & P loads
1:1
1:1
Measured N load is less than predicted, but within the same order of magnitude
Measured P load is significantly less than predicted
BasinUMS MO OH LMS MARB
N lo
ad, T
y-1
1e+2
1e+3
1e+4
1e+5
1e+6
TributaryPoint source
Tributary (non-point source) vs. point source N & P contributions
BasinUMS MO OH LMS MARB
P lo
ad, T
y-1
1e+1
1e+2
1e+3
1e+4
1e+5
Nutrientinflow
Nutrient outflow
Bed sediments
Water column
Sw
Sb
Spiral length (S)=Sw + SbSw is distance transported before uptake by the stream bedSb is the distance transported while in the stream bed
This interface is where most of the uptake and release occurs...
Nutrient spiraling
The more time an atom of nutrient spends in the water column, the slower the uptake rate and the longer the spiral length, the less likely it will be metabolized and/or lost from the system.
2.071.841.231.542.141.69Entire(1300-1500)
1.861.521.842.021.801.620-800
2.031.552.351.891.861.490-400
1.901.841.831.191.681.540-162
2.792.791.952.161.831.360-81
2.792.891.952.161.831.360-73
1.241.241.241.241.231.360-65
1.171.161.161.401.161.360-57
1.171.161.161.401.161.360-40
D-WLn_P
D-WSRP
D-WLn_N
D-WNH3
D-WNOx
CriticalD-W
Interval(km)
Upstream serial autocorrelationHow far upstream can we track nutrient influences?
Based on Durbin-Watson test for 1st order autocorrelation
Microbial Enzyme Stoichiometry
Peptidase : Phosphatase (N : P)0.01 0.1 1 10 100
Gly
cosi
dase
: Pe
ptid
ase
(C :
N)
0.01
0.1
1
10
100 Mississippi RiverMissouri RiverOhio River
N-limited
C-limited
P-limited
Seitzinger et al. 2006. Ecological Applications 16(6):2064-2090
Seitzinger et al. 2006. Ecological Applications 16(6):2064-2090
Seitzinger et al. 2006. Ecological Applications 16(6):2064-2090
1 km2=100 ha
x100
Burgin & Hamilton 2007, Frontiers in Ecology and Environment 5 (2):89-96
0
5000
10000
15000
20000
25000
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
Year
Are
a of
Hyp
oxia
(sq
km)
Hypoxia in the Gulf of Mexico—Linking watersheds, drainage networks and receiving waters
Atchafalaya River plume
Southwest Pass plume
Hypoxia 21-25 July 2004
Atchafalaya River plume
Southwest Pass plume
Hypoxia 21-25 July 2004
Combined Mississippi and Atchafalaya River Flux
0
500000
1000000
1500000
2000000
2500000
3000000
1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004
Year
NO
x an
d TN
Flu
x
0
20
40
60
80
Per
cent
Total NO3+NO2 flux Total TN flux %NOx/TN
1960 1970 1980 1990 2000
N :
P
5
10
15
20
25
Si :
N
0
1
2
3
4
5
1960 1970 1980 1990 2000
DIN
(nm
ol L
-1)
0
100
200
300
DIP
(nm
ol L
-1)
0
5
10
15
20
25
Compared to prior study
MS MO OHThis study DIN 287 155 157
DIP 21 17 4DSi 205 188 39N:P 20 11 76Si:N 3 14 0.4Si:P 14 80 26
Justic et al. (1995)1981-87 data DIN 114
DIP 8DSi 108N:P 15Si:N 1Si:P 14
1960-62 data DIN 36DIP 4DSi 160N:P 9Si:N 4Si:P 40
DIN___DIP___
N:P___Si:N___
Conclusions •N & P were positively correlated with % of watershed in agriculture
•N & P loads fit a simple regression model based on cumulative watershed area
•Disparity between basin-wide projections of N & P and the sum of sub-basin models—suggesting other sources (e.g., point sources) & losses (e.g., N & P sequestration, denitrification)
•Serial correlation of N & P with river distance suggests that spiral lengths (sum of transport in water and bed phases) may be > 65km
•N:P in the tributaries and in the Great Rivers suggest N-limitations relative to available P concentrations
•Microbial enzyme activity associated with the acquisition of C, N & P also suggests N-limitation, along with C-limitation (indicative of nutrient enrichment)