National Agriculture and Food Research Organization
SWAT APPLICATION TO ASSESS EFFECTS OF DIFFERENT FERTILIZATIONS ON WATER QUALITY IN AN AGRICULTURAL WATERSHED
Seiko Yoshikawa, Kazunori Kohyama, Yuta Shimizu, Saeko Yada, Kei Asada,Sunao Itahashi, Yasuhiro Nakajima, Sadao Eguchi
Seiko YoshikawaSenior Principal Researcher
Inst. Agro-Environ. Sci., NARO
1
National Agriculture and Food Research Organization
茨城県
http://blogs.yahoo.co.jp/stockfan21/13842604.html
霞ケ浦
SWAT (Soil and Water Assessment Tool) by USDA
Many physical models and empirical parameters are included in the tool.
★Estimation of water, sediment(SED) and nutrient movement★Scenario analysis・・・Assessing effects of different fertilizations on water
quality in an agricultural watershed
Human activityAgricultureIndustryDomestic
EnvironmentWaterSoilAir
STUDY OUTLINE
Objective : To study the desirable balance of agriculture and water conservation
Background
Methods
2
Google map
Lake
Kasumigaura
Outline for Sakura River Watershed Area 335km2,⊿H 853m,subbasin 35,HRU 424 (Threshold: LandUse/Soil/Slope=5% /10% /20%)LundUse:Forest 34%,Paddy (RICE) 29%,Upland fileds (AGRL) 20%,Urban (residence,road etc.) 14%,
Water 2%,Pasture 1%,others 1%Soil:Andisol 32%, Glay lowland soil 20%, Brown forest soil 29%, Gley soil 9%, Andic gley soil 5%, Black soil 2%,
others 3%
Study area: Sakura River Watershed
3
Other Input data for model configurationWeather data ・・・precipitation, temperature, radiation etc. at 3 meteorological stations Soil-profile physical properties data (Solphy-J, NARO)Irrigation water supplying data General crop management (fertilization, tillage etc.)Domestic discharge (cf. Ibaraki Pref. data)
Constant flow out from each sub basin in proportion to its urban area.water 0.2m3, SED 1g, TN 2g, TP 0.2g /capita/day
SWAT Run
CalibrationComparing modeled flow, sediment, Org-N, NO3-N, Org-P, Min-P with observed data from Ministry of the Environment and Ibaraki pref.
ParameterizationFine adjustment of parameters using SWAT-CUP(calibration/uncertainty or sensitivity program interface for SWAT)
ValidationComparing modeled values with observed data Evaluation by R2 and NS (Nash-Sutcliffe model efficiency coefficient )
Warm-up (2000~2002), Calibration (2003~2005), Validation (2006~2008)
Analytical methods“Daily Rain/CN (curve number) /Daily Route” method for surface runoff“Penman/Monteith ” method for evapotranspiration
4
0
20
40
60
80
100
120
140
1600
50
100
150
200
250
Pre
cip
itat
ion
(m
m)
Flo
w r
ate
(m
3s-1
)
2003 2004 2005 2006 2007 2008
Precip_obs
FLOW_sim
Flow_obs
R2= 0.64 NS= 0.62
R2=0.62NS=0.61
R2=0.66NS=0.73
Calibration Validation
mod
Mo
del
led
Flo
w(m
3 s-1
)
Mo
del
led
Flo
w(m
3 s-1
)
Mo
del
led
Flo
w(m
3s-1
)
5
FlowRESULTS
0
20
40
60
80
100
120
140
1600
50
100
150
200
250
300
350
400
450
Pre
cip
itat
ion
(m
m)
SED
(to
ns
day
-1)
2003 2004 2005 2006 2007 2008
Precip_obs
SED_sim
SS_obsCalibration Validationmod
6
R2=0.85NS=0.66
R2=0.64NS=-6.15 R2= 0.82
NS= 0.62
SedimentM
od
elle
dSE
D (
ton
s d
ay-1
)
Mo
del
led
SED
(to
ns
day
-1)
Mo
del
led
SED
(to
ns
day
-1)
R2=0.59NS=0.27
R2=0.62NS=0.45
R2= 0.59 NS= 0.39
0
20
40
60
80
100
120
140
1600
1,000
2,000
3,000
4,000
5,000
6,000
7,000
8,000
9,000
14
79
31
39
18
52
31
27
73
23
36
94
15
46
15
07
55
35
99
64
56
91
73
77
83
82
98
75
92
19
67
10
13
10
59
11
05
11
51
11
97
12
43
12
89
13
35
13
81
14
27
14
73
15
19
15
65
16
11
16
57
17
03
17
49
17
95
18
41
18
87
19
33
19
79
20
25
20
71
21
17
21
63
Pre
cip
itat
ion
(m
m)
OR
GN
(kg
day
-1)
2003 2004 2005 2006 2007 2008
Precip_obs
ORGN_sim
ORGN_obsCalibration Validationmod
7
Org-NM
od
elle
dO
RG
N (
kg d
ay-1
)
Mo
del
led
OR
GN
(kg
day
-1)
Mo
del
led
OR
GN
(kg
day
-1)
Mo
del
led
OR
GN
(kg
day
-1)
ORGN (2003~2008) y=0.8612x-7.7396R2=0.5898
0
20
40
60
80
100
120
140
1600
1,000
2,000
3,000
4,000
5,000
6,000
7,000
8,000
9,000
10,000
Pre
cip
itat
ion
(m
m)
NO
3-N
(kg
day
-1)
2003 2004 2005 2006 2007 2008
Precip_obs
NO3_sim
NO3-N_obsCalibration Validation
mod
8
R2= 0.86 NS= 0.13
R2=0.89NS=0.15
R2=0.69NS=-0.50
Mo
del
led
NO
3-N
(kg
day
-1)
Mo
del
led
NO
3-N
(kg
day
-1)
Mo
del
led
NO
3-N
(kg
day
-1)
NO3-N
R2= 0.39 NS= 0.37
R2=0.36NS=0.36
R2=0.46NS=0.37
0
20
40
60
80
100
120
140
1600
200
400
600
800
1,000
1,200
1,400
19
00/1
/1
19
00/2
/9
19
00/3
/19
19
00/4
/27
19
00/6
/5
19
00/7
/14
19
00/8
/22
19
00/9
/30
19
00/1
1/8
19
00/1
2/1
7
19
01/1
/25
19
01/3
/5
19
01/4
/13
19
01/5
/22
19
01/6
/30
19
01/8
/8
19
01/9
/16
19
01/1
0/2
5
19
01/1
2/3
19
02/1
/11
19
02/2
/19
19
02/3
/30
19
02/5
/8
19
02/6
/16
19
02/7
/25
19
02/9
/2
19
02/1
0/1
1
19
02/1
1/1
9
19
02/1
2/2
8
19
03/2
/51
903
/3/1
6
19
03/4
/24
19
03/6
/2
19
03/7
/11
19
03/8
/19
19
03/9
/27
19
03/1
1/5
19
03/1
2/1
4
19
04/1
/22
19
04/3
/1
19
04/4
/91
904
/5/1
8
19
04/6
/26
19
04/8
/4
19
04/9
/12
19
04/1
0/2
1
19
04/1
1/2
9
19
05/1
/7
19
05/2
/15
19
05/3
/26
19
05/5
/4
19
05/6
/12
19
05/7
/21
19
05/8
/29
19
05/1
0/7
19
05/1
1/1
5
19
05/1
2/2
4
Pre
cip
itat
ion
(m
m)
OR
GP
(kg
day
-1)
2003 2004 2005 2006 2007 2008
Precip_obs
ORGP_sim
ORGP_obsCalibration Validation
mod
9
Org-PM
od
elle
dO
RG
P (
kg d
ay-1
)
Mo
del
led
OR
GP
(kg
day
-1)
Mo
del
led
OR
GP
(kg
day
-1)
0
20
40
60
80
100
120
140
1600
20
40
60
80
100
120
140
160
180
Pre
cip
itat
ion
(m
m)
Min
-P (
kg d
ay-1
)
2003 2004 2005 2006 2007 2008
Precip_obs
MINP_sim
PO4-P_obsCalibration Validationmod
Simulated Min-P (mainly Ortho-P) overestimated measured Ortho-P. It might be caused fromhigh phosphoric acid absorptivity of Andisols which spread widely in the watershed.
R2= 0.33 NS= -2.55
R2=0.37NS=-0.89
R2=0.29NS=-7.89
Min-PM
od
elle
dM
INP
(kg
day
-1)
Mo
del
led
MIN
P (
kg d
ay-1
)
Mo
del
led
MIN
P (
kg d
ay-1
)
10
CURRENT MANAGEMENT SETTING Basal and additional fertilization
・Paddy ; basal fertilization at transplanting as chemical fertilizer (25-05-00) N80kg/ha, P16kg/ha
・Upland ; basal fertilization as chemical fertilizerN100kg/ha, P45kg/haadditional fertilization as chemical fertilizer N100kg/ha
・Pasture; leaping 4 times/yearauto fertilization as chemical fertilizer (25-05-00)
(each N 10kg/ha,max N 20kg/ha×4)
REVISED MANAGEMENT SETTINGSmall and frequent fertilization
・Paddy ; auto fertilization as chemical fertilizer(25-05-00) (each N 10kg/ha &P2kg/ha, max N80kg/ha & P16kg/ha)
・Upland ; auto fertilization as chemical fertilizer(25-05-00) (each N 10kg/ha & P 2kg/ha, max N100kg/ha & P 20kg/ha)
・Pasture;leaping 4 times/yearauto fertilization as chemical fertilizer(25-05-00)
(each N 2kg/ha,max N18kg/ha×4)
Fertilization pattern image Fertilization pattern image t t
q q
Acceptable estimation for flow, SED, N → Scenario analysis
11
NO3-N decrease by 21%
0
2000
4000
6000
8000
10000
12000
Yld Biomass Yld Biomass Biomass
RICE UPLAND FIELD CROPS PASTURE
DR
Y M
AT
TER
(kg
/ha/
Year
)
Fig. Comparison of Crop Yields
CURRENT(Basal&Additional
fertilization)
REVISED (Small & Frequent
fertilization)
By introducing revisedfertilization (smaller and more frequent fertilization) to paddy, upland and pasture fields, equal or greater crop yields were estimated with smaller amount of N application and mitigated NO3-N discharge.
0
20
40
60
80
100
120
140
1600
1,000
2,000
3,000
4,000
5,000
6,000
7,000
8,000
9,000
10,000
Pre
cip
itat
ion
(m
m)
NO
3-N
(kg
day
-1)
2003 2004 2005 2006 2007 2008
Precip_obs
NO3_sim
NO3rev_sim
mod
mod
12
Thank you very much for your attention.
We thank Ibaraki Prefecture for providing useful data and help to attend WLC16 conference.
SWAT was applied to assess effects of different fertilizations on water quality in an agricultural watershed. The results of scenario analysis for agricultural management changes showed that smaller and more frequent fertilization was effective for crop production and decrease in fertilizer application and water pollution.
CONCLUSION
13