Streamflow Changes in the
Upper Mississippi River Basin
Why is this important?
• MR is the longest and largest river in N. America covering all or
parts of 31 states and 2 provinces
• MR and its basin provide important habitat for fish, wildlife and
living ground for American people
• Human activities have greatly altered this river ecosystem
• Much of the basin is intensively cultivated and cultivation has
Clayton County
Fall view of Mississippi River,
photo by Gary Hightshoe
Keith Schilling, Iowa DNR Geological Survey, Iowa City, Iowa
reduced biodiversity, altered biogeochemistry, and impacted
regional climates and basin hydrology;
• Many tributaries deliver substantial amounts of sediment,
nutrients, and contaminants into the river contributing to many
problems, including Hypoxia of the Gulf of Mexico.
• It is important to know the difference in amounts of nutrients
and contaminants carried by surface runoff and baseflow order
to effectively and efficiently deal with the problems
Is the MR flow increasing ?
We selected 8 USGS gauging stations and
checked their Q
Station 1 – 4: Along the MR
Station 5 – 8: Outside the MR basin
Q at 4 USGS
stations
outside the
MR basin
since 1940s:
Q at 4 USGS
stations along
MR since
1940s:
An increasing
trend
Constant
or decreasing
1. Clinton, IA
2. St. Luis, MO
45%
31%
3. Memphis, TN
35%
4. Vicksburg, MS
5.Columbia River at
Dalles, WA
6. Rio Grande River at
San Felipe, NM
7. Savannah River at
Augusta, GA
8. Sesquehanna River at
Harrisburg, PA
38%
1
Is Streamflow Increasing in Iowa?
What we observed…
What has caused increased
streamflow?
• Precipitation has increased - P increased
about 7% during the last 60 years
• However, increasing P alone not sufficient
to explain magnitude of increase
• Fundamental change in rainfall/runoff
relationship – Q increasing at greater rate
than increasing P alone can explain
• Changes in land use/land cover
1940
1900 10
Streamflow (in)
Streamflow (in)
1950 1960
20 30 40
1970 1980 1990 2000
60 70 80 90 2000
50
Streamflow (in)
Streamflow (in)
Streamflow (in)
45
Wapsipinicon
River16
East
Nishnabotna
River
25
50
Cedar River at Cedar
40
3530 14Rapids
70
55
45
2012
35
60
35
3025
50
60
10
40
30
55
45
30
252015 8
50 35
25
50
40
25
20 10 6
15
40 30
20
45
4
35
20
15
40 25
15 10 5 2
30
30
10
15
35
10 5 0 0
25 20
20
10
30
55
0
20 15
1940 1950 1960 1970 1980 1990 2000
1940 1950 1960 1970 1980 1990 20000 0
25
10
1940 1950 1960 1970 1980 1990 2000
1940 1950 1960 1970 1980 1990 2000
15
20 1940 1950 1960 1970 1980 1990 2000
1940 1950 1960 1970 1980 1990 2000
Precipitation (in)
Precipitation (in)
since 1940s around the red area;
• these increases are statistically significant (p < 0.01)
• But increasing Q is not observed in some stations,
i.e., Arkansas River.
• Increasing Q is not observed in some major rivers
outside the MR basin;
• The finding of increasing Q is consistent with other
studies
• Lins and Slack (1999)
• Schilling and Libra (2003)
Precipitation
(in)(in)
Precipitation
(in)
Precipitation
Floyd River
North Raccoon River
• Large increases in Q were observed in most stations
1940
1900 10
1950
30
20
1960 1970 1980 1990 2000
50 60 70 80 90 2000
40
Land Cover in Iowa around 1850
CULTIVATION IN THE MRB
Note: there was little cropland
Grassland
Current Land Cover in Iowa
Note: > 80% is cropland
Forest
How has agriculture changed?
• In Iowa, changes in
agricultural land use began
around 1940
• Soybean acreage increased
from 1 to 11 million acres
from 1940 to 2000
• With corresponding
increase in corn acreage,
row crop acreage in Iowa
increased approximately
30-40% from 1940-2000
2
How has the agriculture changed?
• Regional scale - Iowa
Fraction of Watershed in Row Crop
• Watershed scale
Percentage of Land in Corn and Soybeans
(by County)
100.0%
90.0%
80.0%
Buena Vista
Calhoun
Carroll
Dallas
Greene
Guthrie
Pocahontas
Sac
Webster
70.0%
60.0%
50.0%
40.0%
30.0%
20.0%
0.9
NRac
Winn
Floyd
0.8
ENish
Waps
0.7
0.6
Turk
Maq
0.5
0.4
Engl
Thom
0.3
0.2
0.1
1920
1940
1960
1980
2000
2020
Year
10.0%
0.0%
1920
1940
1960
1980
2000
2020
LAND
COVER
Soybean acres (x 1000)
Continental Scale: Mississippi River Basin
20000
a.
15000
10000
5000
0
1940
1960
1980
2000
STREAM
NITRATE
STREAMFLOW
(BASEFLOW)
Year
What has changing land cover
done to watershed hydrology?
Difference in ET between perennial vs.
annual crops
LAND
COVER
Single crop ET coefficients (Kc)
Estimated crop ET = Kc * PET
Cover
STREAM
NITRATE
STREAMFLOW
(BASEFLOW)
Initial
Midseason
Late
season
trees
0.5
1.10
0.65
pasture
0.4
1.5
0.85
grass
0.85
0.9
0.9
corn
0
1.2
0.35
soybeans
0
1.15
0.5
Food and Agriculture Organization of the United Nations, 1998
3
Effect of land cover on
water table behavior
802
West-No vegetation
East-Grass
799
0
2
4
6
8
10
12
14
0.00
798
797
796
795
0
10
20
0.05
0.10
WNG
0.15
0.20
EG
0.25
0.30
40
30
50
0
1
Elapsed Time (days)
60
2
70
3
4
80
5
Cummulative ET (mm)
800
Change in Water
Table Depth (m)
Water Tab le Elevatio n (ft)
801
6
Elapsed Time (days)
Dinnes et al., Agron. J. 2002
Zhang and Schilling, J. Hydrol. 2006
LONG-TERM WATER BALANCE
FOR THE MR BASIN
P ET
P = Q + ET
Effect of land use change on river flow
A basin covered by
seasonal crops
A basin covered by
perennial grass
PP = Q P + ET P
PS = Q S + ET S
PP = PS
P – Precipitation
Q – River flow
ET – Evapotranspiration
ET P > ET S
Q
QP < QS
Q before 1940 is smaller than Q after 1940.
Effect of land cover on groundwater recharge
(R)
R– Qb = ΔS
↑ more ETP
↑ less ETS
What else has accompanied
changing land use patterns?
• Tile drainage
↓ less RP
↓ more RS
Over long period of time ΔS = 0 and thus R = Qb
Rs > Rp so Qbs > Qbp
Thus, changing land cover from perennial to seasonal
crops would result in an increase in baseflow
4
Relation of land use to baseflow
What else has accompanied
changing land use patterns?
Plot studies
Corn and soybeans
have greater drainage
and less ET than
perennial grasses
• Improved
conservation
practices
Brye et al., Soil
Sci. Am. J. 64:715724 (2000)
Prediction of Baseflow using Multiple
Linear Regression
Ecoregion
Central Irregular
Plains (40)
Mean Mean
Q (in) Qb (in)
9.8
2
Mean
NO3-N
Load
(lb/ac)
3.6
12
12
Q (in)
2
r = 0.78
p<0.01
10
8
Qb (in)
2
r = 0.89
p<0.01
10
8
Northwest Iowa
Loess Prairie (47a)
58
7.8
4.4
14.8
Des Moines Lobe
(47b)
179
11
6.1
20.7
Iowan Surface (47c)
114
11.8
7.7
20.0
6
6
4
4
Predicted Values
n
82
2
2
0
0
0
12
2
4
6
8
2
4
6
8 10 12
80
Qs (in)
2
r = 0.82
p<0.01
10
0
8 10 12
70
60
6
50
4
40
2
%Qb
2
r = 0.89
p<0.01
30
0
Loess Hills and
Rolling Prairies (47e)
54
9.4
5.3
14.7
Southern Iowa
Rolling Loess
Prairies (47f)
190
10.4
4.5
12.5
20
0
2
4
6
8 10 12
20 30 40 50 60 70 80
Loess Hills (47m)
3
7.0
3.8
6.7
Paleozoic Plateau
(52)
30
8.7
4.1
6.7
Measured Values
Q = 0.645(RAIN) + 0.0607(%SAND) + 0.0519(%RC) – 16.3
Qb = 0.247(RAIN) + 0.047(%SAND) + 0.063(%RC) +
0.0142(PERM) + 0.0528(%ALLUV) – 11.2
Schilling and Wolter, JAWRA, 2006
Schilling and Wolter, JAWRA, 2006
Historical evidence – Raccoon River watershed scale
2.0
20
15
10
5
0
15
10
5
0
2.0
Jan
1.0
0.5
0.5
0.0
1900 1920 1940 1960 1980 2000
5
0
80
70
60
2.0
1.5
1.0
1.0
0.5
Jul
1.5
0.0
1.5
1.0
1.0
0.5
0.5
0.0
1900 1920 1940 1960 1980 2000
2.0
30
1.5
Oct
1.0
1.5
1.5
Dec
1.0
0.5
0.5
0.0
0.0
1900 1920 1940 1960 1980 2000
Largest increases
occurring in Apr –
July
1900 1920 1940 1960 1980 2000
2.0
Nov
1.0
0.0
Sep
0.0
1900 1920 1940 1960 1980 2000
2.0
Sig. increases in Q,
Qb in all months
but Feb, Mar
1900 1920 1940 1960 1980 2000
2.0
Aug
0.5
0.0
Jun
0.5
1.0
40
1900 1920 1940 1960 1980 2000
1.5
1900 1920 1940 1960 1980 2000
2.0
Seasonal
Changes
1900 1920 1940 1960 1980 2000
2.0
1.0
0.0
1900 1920 1940 1960 1980 2000
2.0
1.5
May
0.5
0.0
1900 1920 1940 1960 1980 2000
Schilling, 2003
2.0
Mar
0.0
1900 1920 1940 1960 1980 2000
2.5
Apr
1.5
50
20
1.5
0.5
2.5
10
2.0
Feb
1.0
0.5
1900 1920 1940 1960 1980 2000
1.5
1.0
0.0
15
1900 1920 1940 1960 1980 2000
1900 1920 1940 1960 1980 2000
1.5
Baseflow Discharge (in)
Baseflow Discharge (in)
25
Baseflow Percentage
Stormflow Discharge (in)
Total Discharge (in)
Total streamflow and baseflow increased in the Raccoon
River watershed from 1920’s to 2000
1900 1920 1940 1960 1980 2000
Increasing Qb
relative to P in all
months but Feb
and Mar
Schilling, 2003
5
70.00%
Discharge (in)
Row Crop
60.00%
15
50.00%
40.00%
10
30.00%
20.00%
5
10.00%
1930
1940
1950
1960
1970
1980
1990
Baseflow increased in Iowa since mid-20th century
80
Annual Baseflow Percentage
80.00%
Qb
20
0
1920
Historical evidence – regional scale - Iowa
90.00%
Q
% Row Crop in Raccoon River Watershed
25
0.00%
2000
70
60
40
30
NRac
Maq
Win
Turk
50
Engl
Floyd
ENish
Wap
Thom
20
1930
1940
1950 1960
1970
1980
1990
2000 2010
Year
• Streamflow changes in Raccoon River significantly
correlated to changing land use patterns
Schilling, 2003
Schilling and Libra, JAWRA, 2003
Period
of Analysis
1933-2000
1937-2000
1942-2000
1936-2000
1937-2000
1934-2000
1933-2000
1927-2000
1942-2000
1927-2000
1927-2000
Watershed
Winnebago
English
Thompson
Floyd
E. Nishnabotna
Wapsipinicon
Turkey
Maquoketa
N. Raccoon
Cedar River
Raccoon River
Increase in Qb (mm)
140
120
100
LogQb
t-value
p-value
4.08
(<0.001)
1.78
(0.081)
0.33
(0.740)
4.40
(<0.001)
4.20
(<0.001)
5.45
(<0.001)
4.51
(<0.001)
3.60
(<0.001)
2.03
(0.047)
4.42
(<0.001)
4.20
(<0.001)
t-value
5.43
1.21
2.04
4.95
5.14
6.55
4.52
4.10
2.40
2.08
3.88
%Qb
p-value
(<0.001)
(0.231)
(0.046)
(<0.001)
(<0.001)
(<0.001)
(<0.001)
(<0.001)
(0.019)
(0.041)
(<0.001)
a.
r2 = 0.52
p = 0.03
80
60
Q, BF & OF
at
4 major
tributaries in
MR basin
since 1940
40
20
0.0
Increase in %Qb
35
30
25
0.1
0.2
0.3
0.4
0.5
0.6
0.2
0.3
0.4
0.5
0.6
b.
13. UMR at
Keokuk, IA
14. Missouri River at
Boonville,MO
51%
38%
15. Ohio River at
Metropolis, IL
16. Arkansas River at
Murray Dam, AR
r2 = 0.35
p = 0.09
20
15
10
5
0.0
0.1
Increase in Row Crop Fraction
Schilling,, Agriculture Ecosystems & Environment, 2005
15
1960
1980
2000
10. Illinois River at
Kingston Mines, IL
11. Wabash River
at Camel, IL
12. Ohio River at
Luisville, KY
Flow rate (mm)
60
1960
1980
2000
5
1960
1980
2000
1960
1980
Year
2000
2000
1980
2000
1980
0
1940
2000
20
0
1940
1960
1980
Year
2000
1960
1980
Year
2000
60
Sep
40
2000
20
1960
60
20
1980
Year
Nov
40
0
1940
40
Q
BF
SW
1960
60
20
1960
Jun
0
1980
Aug
40
60
20
1960
0
1940
60
0
1940
Mar
40
0
1940
May
Ja
n
Fe
b
M
ar
Ap
r
M
ay
Ju
n
Flow rate (mm)
9.2%
20
20
0
1940
-5
35%
20
0
1940
40
20
0
1940
46%
20
60
10
Oct
40
Feb
40
60
Jul
40
Dec
40
0
1940
20
1960
1980
Year
ct
ov
D
ec
20
0
1940
60
Apr
40
O
Changes in flow rate (mm)
from 1940 to 2003
Flow rate (mm)
Jan
2000
N
60
20
40
Ju
l
Au
g
Se
p
60
60
102%
9.3%
Seasonal changes in Mississippi
River at Keokuk
Q, BF & OF
in
4 tributaries
in the MR
basin since
1940
9. Cedar River at Cedar
Rapids，IA
24%
0
1940
1960
1980
Year
2000
Zhang and Schilling, J. Hydrol., 2006
6
Baseflow (mm)
MR at Keokuk, IA
Relation of Baseflow in Mississippi
River to Increasing Soybean Production
400
b.
LAND
COVER
2
r = 0.18
p = 0.001
300
200
100
STREAM
NITRATE
0
0.00
0.05
0.10
0.15
0.20
STREAMFLOW
(BASEFLOW)
Soybean Fractional Area
Zhang and Schilling, J. Hydrol., 2006
Groundwater discharge as baseflow provides
main source of nitrate to streams
Baseflow contribution to N-loads
in Raccoon River
For example, in two central Iowa watersheds, export of
nitrate occurred primarily with baseflow (61-68%)
Baseflow Fraction
0.8
0.6
0.8
0.6
0.4
0.2
0.0
1.40
1.20
1.00
0.80
0.60
Discharge
NO3-N Load
0.40
0.20
1972 1976 1980 1984 1988 1992 1996 2000
0.00
Water Jan
Year Feb Mar Apr May Jun
WNT2
SQW2
0.4
Generalized Annual Crop Soil
1.60
Baseflow Fraction
Nitrate N
1.0
Baseflow Fraction
Discharge
Baseflow Enrichment Ratio
1.80
1.0
1.0
Nitrogen Utilization Rate
0.9
0.8
0.7
0.6
0.5
0.4
Discharge
0.3
NO3-N Load
0.2
J F M A M J
Jul
J
A S O N D
Aug Sep Oct Nov Dec
0.2
J F M A M J J A S O N D
J F M A M J J A S O N D
66.7% of nitrate delivered by baseflow
Schilling, JEQ, 2002
Schilling and Zhang, J. Hydrol, 2004
Tile Drainage Contributes to Baseflow and
Nitrate Concentrations and Loads
LAND
COVER
STREAM
NITRATE
STREAMFLOW
(BASEFLOW)
7
Relation of Row Crop Land Use to
Nitrate Concentrations
W6
5m
1 Camp Creek Member (post-settlement) 6 Pre-Illinoian till and paleosol
Organics
2 Roberts Creek Member
Well Screen (typ.)
3 Gunder Member
4 Pre-Holocene alluvium
C Well Depth Identifier (typ.)
5 Wisconsinan loess
A
NH4
Concentrations
A’
50m
E7
Treatment
6
W5
W3
W2
Control
E6
TW-3 TW-4
W1 E1
E2
E3
5
E5
Mean Annual NO3-N
Conc. (mg/L)
1
6
2
1-5
mg/l
12
10
8
6
4
2
0
y = 0.1077x - 0.812
0.2-0.5 mg/l
3
4
>10 mg/l
2
R = 0.6501
W6
5m
1 Camp Creek Member (post-settlement) 6 Pre-Illinoian till and paleosol
Organics
2 Roberts Creek Member
Well Screen (typ.)
3 Gunder Member
4 Pre-Holocene alluvium
C Well Depth Identifier (typ.)
5 Wisconsinan loess
A
NO3-N
Concentrations
20
40
60
80
E7
Treatment
Control
20+ mg/l
6
W5
W3
0
A’
50m
100
W2
E6
TW-3 TW-4
W1 E1
1-3 mg/l
<0.5 mg/l
Percent Row Crop
E2
E3
5
E5
1
6
2
3
4
Schilling and Libra, JEQ, 2000
Schilling and Jacobson, In prep.
No Grass Cover
Historical Perspective
45
10,000
40
30
25
1,000
20
NO3-N (mg/L)
35
15
10
NO3-N Concentration (mg/l)
50
Cedar River at Cedar Rapids
8
7
Cedar River
Raccoon River
1990-2000
6
1990-2000
5
4
3
2
rR
da
Ce
1
0
STREAM
0
NITRATE
Ra
cc
oo
20
b
rQ
ive
1946
1944-52
b
nQ
1907
40
60
80
220
200
180
160
140
120
100
80
60
40
20
0
Baseflow (Qb) (mm)
LAND
COVER
Nitrate concentrations in Iowa’s streams have
increased since the mid-20th century
Average Annual Discharge (cubic feet per second)
Grass Cover
STREAMFLOW
100
(BASEFLOW)
Row Crop Percentage
5
100
0
45
50
55
60
65
70
75
80
85
90
95
Water Year
Schilling,, Agriculture Ecosystems & Environment, 2005
Conclusion
Changing land cover from perennial mixed cropping systems to
row crops of corn and soybeans increased baseflow and has
likely contributed to increasing nitrate-N losses from the
agricultural Midwest.
8