Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
2013 International SWAT Conference & Workshops - Toulouse, France
Paul Sabatier University,
July 17-19, 2013
Sediment load responses to simulated conservation management practices
- towards water and soil protection in a Central Brazilian catchment
Michael Strauch, Jorge E.W. Lima, Martin Volk,
Carsten Lorz, Franz Makeschin
Page 2
Distrito Federal Brasília • Inhabitants: 2.6 Mio (2009)
• Area: 5789 km2
• Population density: ~450/km2
Introduction 1
Page 3
Introduction 1
Journal de Brasília 22.3.2011
Page 4
Population in thousands in the DF
Land use change in the DF
Other Urban
Agri-culture
Cerrado
Mata
Increasing risk of water scarcity due to
population growth & land use change
in the Federal District, Brasilia (DF)
IWAS/Água-DF: Research project to
support IWRM for the DF
Program ‘Produtor de Água’: Support of
sustainable management by Payments
for Environmental Services (PES)
Effect of Best Management Practices (BMPs) on sediment loads and water quantity?
Introduction 1
Source: ANA (2010)
5
Study area 2 Pipiripau River Basin (PRB), 188km²
Storm event in Brasília, 27.03.2011
6
Precipitation uncertainty 3
7
Ensemble of precipitation input data
TAQ Gauge Taquara
TAQM Moving average of TAQ
THIE Thiessen polygons
TRMM Satellite data
uniform uniform spatially distributed spatially distributed
Sensitive Model parameters? (LH-OAT; van Griensven et al., 2006)
Best fit parameters? Flow & sediments? (SUFI-2; Abbaspour et al., 2004)
Precipitation uncertainty 3
Init
ial r
ange
of
par
amet
er
valu
es
Model parameters
A B C D E
…
…
Best solution (auto-calibration)
for rain input model: 1 2 3 4
Calibrated range (Ø 37% of initial range!)
8
…and its influence on parameter uncertainty
Precipitation uncertainty 3
9
Streamflow (daily)
Model calibration 4
NSE: 0.67 NSE: 0.57 R²: 0.68 R²: 0.79
Stream gauge Montante Captação
Sampling gauge:
Montante Captação
10
Sediment load (daily)
Model calibration 4
R²: 0.37 R²: 0.52 NSE: 0.37 NSE: -2.10
R²: 0.73 R²: 0.64 NSE: 0.71 NSE: -2.08
Sampling gauge:
Montante Captação
11
Sediment load (monthly)
Model calibration 4
Terraces (TER)
on pasture and cropland using the lup-file USLE P-Factor: 0.5 => 0.12 Curve Number: calibrated value -5
Source: BRASIL (2010) Source: BRASIL (2010)
Barraginhas (BAR)
simulated as ponds pond parameters derived by GIS and expert knowledge SWAT code modification: only surface runoff is routed through ponds
Multi-diverse crop rotation (ROT)
crops change each year:
all scenarios were run in different quantities of implementation
C
1st
2nd
3rd
3 crops per year
soybean/corn/cotton
corn/beans/sorghum/ /sunflower/canola beans/wheat/bell pepper/ sweet corn/potato
12
BMP scenarios 5
on cropland (using lup-file)
Cumulative distribution of daily model predictions (extreme scenarios) for period 2004-2009…
Results
13
BMP scenarios 5
Cost-benefit analysis
Implementation costs (ANA, 2010) :
Terraces: USD 150/ha (implementation) USD 100/ha (re-establishement) Barraginhas: USD 120/unit
14
BMP scenarios 5
15
Model results plausible, however problems regarding
Sediment loads (reference data, model performance) Rain input (ensemble is advantageous but has limitations) Process representation in general (semi-humid tropics!) and for BMPs
Pilot program ‘Produtor de Água‘ as a chance to study the effects of BMPs (monitoring!) and thus to evaluate BMP representation in SWAT
Recommendations for the Pipiripau basin:
Erosion control constructions (terraces, “Barraginhas“) are promising (up to 40 % less sediment load) Crop rotation with irrigation during dry season is no option!
Conclusions 5
Thank you. Comments / questions?
Page 17
Modelling workflow
Model setup for status quo
Calibration, Plausibility check,
Validation
Scenario simulation,
Impact analysis,
Optimization-based trade-offs
Source code adaptation
Input data,
GIS preprocessing,
Initial parameters
Reference data for
streamflow,
turbidity, LAI, ET
Scenario data / design
Appendix X
Plant growth,
Strauch & Volk (subm.),
Ecol. Mod.
Precipitation uncertainty, Strauch et al. (2012), J. Hydrol.
BMP scenarios,
Strauch et al. (2013),
J. Environ. Manage.
Case studies for PhD within IWAS project, Central Brazil:
18
Appendix X Parameter uncertainty due to rain input ensemble
19
Rating curve to derive daily turbidity
Correlation between TU and CSS: CSS = 1.114 TU + 1.4731 [R² = 0.958]
Appendix X
20
Terraces
implemented to 25, 50, 75, and 100 % on pasture and cropland using the land-use-update file (TER25, TER50, TER75, TER100) reduced USLE P-Factor from 0.5 (terraces in poor condition) to 0.12 reduced SCS Curve Number II by 5 from calibrated value
area = X1 area = X2 area = X3 area = X4 area = X5 area = 0 area = 0 area = 0
area =
X1*0.75
area =
X2*0.75
area =
X3*0.75 area = X4 area = X5
area =
X1*0.25
area =
X2*0.25
area =
X3*0.25
Status quo
Scenario
Land-use update using the lup-file
Source: BRASIL (2010)
Appendix X
21
Sediment basins along roads (‘Barraginhas’)
simulated as ponds in different quantities (BAR25, BAR50, BAR75, BAR100) pond parameters (subbasin-level) derived by GIS and expert knowledge SWAT code modification: only surface runoff is routed through ponds
Source: BRASIL (2010)
Appendix X
22
Multi-diverse crop rotation
8-year crop rotation with 3 crops per year suggested by EMBRAPA implemented on cropland (soybean and corn monocultures, or soybean-corn rotation) to different percentages using the lup-file (5, 10, 25, 50 %) and 8 placeholder HRUs per subbasin, in which the rotation is shifted
Appendix X
23
Combined simulations of TER and BAR
Appendix X
24
Appendix X
Average percentage change of streamflow and sediment load in BMP scenarios