SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie...
Transcript of SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie...
![Page 1: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/1.jpg)
SWAT modeling for nutrient loading under BMP and
climate scenarios in Lake Erie watersheds
Nate Bosch, Margaret Kalcic, Don Scavia
March 18, 2014
1
![Page 2: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/2.jpg)
Ecofor research team Watershed Team Affiliation
Dave Allan UM
Nate Bosch Grace College
Jinny Han Korean Env. Inst.
Dave Dolan UW
Pete Richards Heidelberg
Tom Croley NOAA
Changsheng He WMU
Hypoxia Team Affiliation
Don Scavia UM
Dmitry Beletsky UM
Dan Rucinski LimnoTech
Joe DePinto LimnoTech
Dave Schwab UM
Ecological
Effects
Affiliation
Tomas Hook Purdue
Kristen Arend Purdue
Ed Rutherford UM
Stuart Ludsin OSU
Doran Mason NOAA
Steve Bartell E2, Inc
Steve Brandt NOAA
Center for Sponsored Coastal Ocean Research (CSCOR) Science Serving Coastal Communities
2
![Page 3: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/3.jpg)
SWAT watersheds
3
![Page 4: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/4.jpg)
Calibration/validation methods
• Model run: 1995-2005 (3 years spin-up, 4 years calibration, 4 years validation)
• Calibration based on daily USGS flow and near daily WQ data from National Center for Water Quality Research, Heidelberg University (except Huron)
• Calibration at daily time step (time series plots) and monthly time step (evaluation statistics after Moriasi et al. 2007)
4
![Page 5: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/5.jpg)
Calibration/validation findings
• Agricultural and forested watersheds lend themselves particularly well to SWAT modeling
• Emphasizes the importance of the availability of observed data with high sampling frequency and long duration
• Indications that over-calibration of hydrology can negatively impact subsequent sediment and nutrient calibration
5
![Page 6: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/6.jpg)
BMP scenarios
Grass filter strips: 10 m width, 25% efficiency
Photo from www.leopold.iastate.edu
Photo from www.oh.nrcs.usda.gov
Photo from www.notilltalk.org
Cover crops: cereal rye planted after soybean harvest
No-till corn and soybean
• Applied randomly to additional row-crop land
• Applied at Moderate (25%) and High (100%) rates
6
![Page 7: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/7.jpg)
Average daily TP loads across watersheds and “feasible” BMP scenarios
Huron
Raisin
GrandSandusky
MaumeeCuyahoga
0
20
40
60
80
100
120
BaseNo-til
CoverFilter
Combo
Dai
ly T
P lo
ad (
kg P
/km
2)
Huron
Raisin
Grand
Sandusky
Maumee
Cuyahoga
7
![Page 8: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/8.jpg)
Effectiveness vs % Implementation
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
0% 20% 40% 60% 80% 100%
Red
uct
ion
Eff
ecti
ven
ess
(%)
Combo BMP Implementation Extent (%)
Sed
TP
SRP
TN
NO3
8
![Page 9: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/9.jpg)
BMP scenario findings
• “Feasible” BMP implementations and source reductions rates are minimally effective
• Implementation of BMPs in specific subwatersheds is much more effective, but may face trade-offs with TP and sediments
• “all-of-above” strategy is needed to substantially reduce nutrient yields and that BMPs should be much more widely implemented
9
![Page 10: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/10.jpg)
Climate scenarios
Moderate Pronounced
Temperature Precipitation Temperature Precipitation
Season ( ̊C) (%) ( ̊C) (%)
Winter +2 +5
Spring +11 +29
Summer +4 +7
Fall -7
Hayhoe et al. 2010
10
![Page 11: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/11.jpg)
0
20
40
60
80
100
120
140
Huron Raisin Maumee Sandusky Cuyahoga Grand
TP lo
ad (
kg P
/ k
m2)
NoneModerateSevere
Amount of climate change
11
Total phosphorus with climate change
![Page 12: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/12.jpg)
Maumee TP: climate and BMPs
0
20
40
60
80
100
120
Moderate Pronounced
TP lo
ad (
kg P
/ k
m2)
Amount of climate change
None Moderate HighBMP implementation
baseline condition
12
![Page 13: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/13.jpg)
Climate scenario findings
• Climate change increases sediment loads more than water flow and nutrient loads
• Individual watershed differ in responsiveness to climate change
• BMPs less effective, but more necessary under climate change conditions
• Stronger BMP implementation and unique management for future watershed load reductions
13
![Page 14: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/14.jpg)
Watershed
Modeling
Affiliation
Don Scavia UM
Nate Bosch Grace College
Lake Modeling Affiliation
Dmitry Beletsky UM
Joe DePinto LimnoTech
WSC research team Climate Affiliation
Anna Michalak Stanford
Allison Steiner UM
Derek Posselt UM
Land Allocation Affiliation
Michael Moore UM
Dan Brown UM
Outreach and Education Affiliation
Elizabeth LaPorte UM
Mary Beth Damm UM
Maria Carmen Lemos UM
Hans Sowder UM
Other Members Affiliation
Tom Bridgeman U of Toledo
Jen Read UM, GLOS
Pete Richards Heidelberg
14
We have many other collaborators as well.
![Page 15: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/15.jpg)
Climate model
Agricultural adaptation
model
Land allocation
model
Lake circulation
model
Water quality model
Watershed model
Governance systems
(scenarios)
Quality-of-life hedonic model
Methodology
15
![Page 16: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/16.jpg)
Model setup and calibration
Changes from Ecofor: • 2006 National Land Cover
Dataset (NLCD) • Soil Survey Geographic
(SSURGO) soils • Setup with HUC-12 subbasins • HRUs: no threshold so all are
represented • Newer version of SWAT
(2012) and newer tile drainage routine
• Model run: 1998-2010 (3 years spin-up, 5 years calibration, 5 years validation)
• Climate and land use change 16
![Page 17: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/17.jpg)
Integrating future climate change
Climate variables (daily, 25 years ~2050): Precipitation Temperature (min/max) Solar radiation
Relative humidity Windspeed
Maumee HUC-8 watersheds
Latitude
Lon
gitu
de
Existing climate stations
Locations of projected climate data (25 km resolution, daily data)
17
![Page 18: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/18.jpg)
Integrating future climate change
Spatial location of climate data:
• Weather stations or grid?
Correct use of climate projections:
• Run Global Climate Model (GCM) pair (present-day and future) and compare the difference?
– Pro: uses the projected climate data and compares apples to apples.
– Con: spatial heterogeneity somewhat lacking.
• Use statistical downscaling or bias correction to create more accurate spatial heterogeneity.
– Pro: more realistic spatial heterogeneity.
– Con: assumes present-day statistics apply to future. 18
![Page 19: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/19.jpg)
Integrating land allocation
Cultivated crops
Forest
Urban
Wetland
Schematic of land use pixels (30 m x 30 m)
19
![Page 20: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/20.jpg)
Integrating land allocation
Cultivated crops
Forest
Urban
Wetland
If two agricultural pixels changed to urban
20
![Page 21: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/21.jpg)
Integrating land allocation HRUs = land use + soil type
Cultivated crops
Forest
Urban
Wetland
Soil A Soil B Soil C
21
![Page 22: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/22.jpg)
Integrating land allocation Each pixel moves to HRU with correct land use and soil
22
Cultivated crops
Forest
Urban
Wetland
Soil A Soil B Soil C
![Page 23: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/23.jpg)
Integrating land allocation
SWAT contains a land use change tool (LUP.dat) where HRUs can shrink and grow
SWAT setup decisions:
• Limitation of LUP.dat approach: need to have all possible soil and land use combinations present in each subbasin at setup stage. Alternatively, create new HRUs after setup stage (Looking at a tool that could create new HRUs called LUPSA (Koch et al., 2012, International Congress on Environmental Modelling and Software))
• Need to have 0% land use and soil lumping threshold (no lumping) to ensure every pixel is in an HRU (e.g. Chiang et al., 2010,
Transactions of the ASABE, 53(5):1569-1584)
23
![Page 24: SWAT modeling for nutrient loading under BMP and climate scenarios in Lake Erie watershedsgraham.umich.edu/media/files/watercenter/Bosch-Kalcic... · 2014. 4. 8. · loading under](https://reader035.fdocuments.net/reader035/viewer/2022071414/610df3d37fec86068e2e380e/html5/thumbnails/24.jpg)
Challenges
Data gaps:
• Lack of long-duration, high-frequency sediment and nutrient data for evaluation
• Lack of quantitative BMP implementation data for model parameterization
Methodology:
• Incorporating future climate projections that are most believable in aggregate at small spatial and temporal scales
• Setting up SWAT for land use change experiments
24