Transcript Folie 1 - icarda

Modeling water balance and
water productivity in
CropSyst model
M. Glazirina, D. Turner
1
CropSyst model description
2
CropSyst
= “Cropping Systems Simulation Model”
• programmed in C++ (object-oriented)
by Prof. C. Stöckle and R. Nelson
(Visual Basic for Application version available)
• multi-year, multi-crop, daily time step
simulation model
• based on the understanding of plants, soil,
weather and management interactions
–
–
–
–
phenological development
photosynthesis and growth
stress effects (water, N, salt, (K))
root water uptake
• Distributed free of charge via
http://www.bsyse.wsu.edu/cropsyst/
3
CropSyst - some detail
• provides a
– generic crop-growth component, allows adaptation/calibration to any
crop; species and cultivars are characterized by a set of parameters which
determine crop response to the environment
– link to the GIS-software Arc/Info (spatial application)
– report format editor for setting up output style, e.g. MS-Excel
– fast graphics viewer
• is very well documented, maintained and regularly updated!
More specifically
• considers the influence of soil salinity and shallow groundwater table,
• allows using a finite difference solution of Richards equation to simulate
water transport.
• handles conservation agriculture features (to some extent)
4
Input-output fluxes in CropSyst
Evapotranspiration
Volatilization
Rainfall
Management:
irrigation
tillage
Fertilization
harvest
Percolation
Leaching
CROP
Runoff
Soil loss
SOIL
Water rise
Crop processes in CropSyst







development
growth
light interception
net photosynthesis
biomass partitioning
leaf expansion
roots deepening






leaf senescence
water uptake
nitrogen uptake
water stress
nitrogen stress
light stress
Soil processes in CropSyst
 water infiltration
 water






redistribution
runoff
evaporation
percolation
solutes transport
salinization
nitrogen fixation
 residues fate
 O.M. mineralization
 nitrogen
transformations
 water erosion
 ammonia
volatilization
 ammonium sorption
CropSyst data requirments
Constant:
Changing in time:
Soil:
•
•
•
•
Soil:
Soil moisture
NO3-N and NH4-N
SOM
Salinity
•
•
•
•
•
Weather:
Precipitation
Tmax, Tmin
RHmax, RHmin
Solar radiation
Wind speed
•
•
•
•
Management:
Tillage
Irrigation
Fertilization
Harvest
• Texture
• Hydraulic properties
(bulk density, PWP, FC)
• Chemistry (CEC, pH)
Crop
model
Ground water and
salinity
Used for calibration:
•
•
•
•
Crop:
Phenology
N-uptake
AGB
Yield
•
•
•
•
Soil:
Soil moisture
NO3-N and NH4-N
SOM
Soil salinity
8
Water balance components
in CropSyst model
9
Water balance equation
P + I = ET + Inf + R + DS
Where:
The incoming water balance components:
P - precipitation (including snow)
I - irrigation
The outgoing water balance components are:
ET - Evapotranspiration
Inf - Infiltration of water
R - Surface runoff (natural) or surface drainage (artificial)
DS is the change of water storage
10
Evapotranspiration model
• Penman-Monteith
comprehensive, precise
– data requirements:
precipitation, max. temp., min temp., solar
radiation, wind speed, max relative humidity,
min relative humidity
• Priestley-Taylor
simple, less precise
– data requirements:
precipitation, max. temp., min temp., solar
radiation
11
Penman-Monteith
Original:
Δ(RN-G)
RN = net radiation [W m-2]
G = soil heat flux [W m-2]
f(e) = VPD (vapor pressure deficit) [hPa]
radiation term
aerodynamic term
Small modification in CropSyst:
f(e) = DayFrac x VPDday_mean
Fraction of day in daylight
12
Priestley-Taylor
λET = PTc x
Δ (RN-G)
Δ+γ
Priestley-Taylor "constant"
• compensates for the elimination of the
aerodynamic term (of the Penman or PM-model)
• default 1.26, higher in arid regions
AVOID USING Priestley-Taylor ET in arid regions!
13
Surface runoff
1
Surface runoff [proportion of
rainfall]
Two options:
1. SCS curve number
(CN) approach
(USDA-SCS, 1988)
2. numerical solution
PAW=0.5
0.75
0.5
Daily rainfall = 70 m m
0.25
40 m m
10 m m
0
40
50
60
70
80
90
100
Curve number (CN)
Erosion
• RUSLE parameters:
- Steepness (a percentage 0-100)
- Slope length (m)
14
Soil water infiltration & redistribution
• CropSyst provides basically two different
models for choice:
1. cascade
2. finite difference
15
Cascade model
• each given soil layer is defined by:
– water content at saturation (SAT)
– water content at drained upper limit (DUL, FC)
– the permanent wilting point (PWP)
• The difference between SAT and the current soil water content
(Theta, Θ) determines the capacity of the layer to hold additional
water
• After infiltration events, a fraction of water in excess of DUL is
drained based on a drainage rate constant
• If Theta for the lowest soil layer exceeds DUL, the excess water is
assumed to drain out of the profile
• If the potential drainage for a layer is very large, the net drainage
may be limited by the saturated hydraulic conductivity (Ks).
16
•
•
Finite difference model
builds on the Richards equation
–
common flow equation for (un)saturated flow in porous media (as a
soil can be considered)
–
is a parabolic non-linear partial differential equation of secondary
order, which is solved numerically by a finite difference approach
requires a parameterization (continuous form) of the soil
hydraulic properties via:
–
–
soil water retention characteristics  pF-curve
soil hydraulic conductivity  SHC-curve
CropSyst uses the so-called Campbell approach
17
Soil hydraulics according to Campbell
ψsl = -a x WCl
–b
whereas
a = e (ln(33) + b x ln(WC-33))
b=
[ln(-1500/-33]
[ln(WC-33/WC-1500)]
• Soil hydraulic conductivity:
0.5
Loam, observed
0.45
Soil moisture [cm³ cm -³]
• Soil water potential
of layer l, ψsl:
van Genuchten
0.4
Campbell
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
1
10
van Genuchten
100
whereas
Campbell
10
K [cm/d]
• air entry potential = (-a x Θs -b)
100000
1000
K = Ks x (Θ/ Θs) c
c = 2b + 3
100
1000
10000
Soil water potential [hPa]
1
0.1
0.01
0.001
0.0001
1
10
100
h [hPa]
1000
18
10000
Output
•
•
•
•
Daily report
Seasonal report
Annual report
specific files:
– cum_water_
depth.xls
– hydraulic_
properties.xls
– water_content.xls
– water_depth.xls
– water_potential.xls
– …
19
Output
•
•
•
•
•
•
•
•
•
•
•
Water entering soil = Precipitation + Irrigation - Interception (crop&residue)
Precipitation
Irrigation
Crop water Interception
Residue water Interception
Evapotranspiration = Soil evaporation + Transpiration + Residue evaporation
Soil evaporation
Transpiration
Potential and actual
Residue evaporation
Infiltration
Soil water depletion
Water entering soil - Evapotranspiration – Infiltration =
Soil water depletion
20
Crop growth in CropSyst
21
Crop Development
• Crop development is the progression of a crop
through phenological stages.
• The proper simulation of crop development
(phenological stages) is crucial
– as it determines the length of time when the crop
interacts with the environment
– as it allows matching specific physiological conditions
of a crop to specific environmental conditions.
• Crop development is governed by growing degree days
(GDDs)
22
GDDs
Temperature
Photoperiod
GDDs
Vernalization
Water stress
23
Key phenological stages in CropSyst
GDDs (°C days) from seeding to
• Emergence
• maximum rooting depth
• Peak LAI (end of vegetative growth)
• begin Flowering
• begin Grain filling
• Maturity
Also expressed in GDDs:
• Leave duration
24
Crop-growth – governing equations
B Rad
= Tlim  RUE  PAR  (1 - e -k  LAI )
Eq. 1
[kg m-2 day-1]
1- e(-k*LAI)
1
Adsorbed radiation [fraction]
BRad = biomass production
(radiation-dependent)
Tlim = temperature-dependent
limiting factor*
RUE = radiation use efficiency
[kg MJ-1]
PAR = photosynthetic active
radiation [MJ m-2 day-1]
k = radiation extinction
coefficient [-]
LAI = leaf area index [m2 m-2]
k = 0.5
0.75
k = 0.6
k = 0.7
0.5
k = 0.8
k = 0.9
0.25
k=1
0
0
1
* in view of optimum mean daily temperature for growth
2
3
LAI [m2 m-2]
4
5
6
25
Crop-growth – governing equations (cont.)
B PT
BTR  Tact
=
VPD
Eq. 2
BPT = biomass production (transpiration dependent)
BTR = aboveground biomass transpiration coefficient [kg m-2 kPa m-1],
often simply called Transpiration Use Efficiency
Tact = actual transpiration [m d-1]
VPD = vapor pressure deficit [kPa]
Assumptions/Preconditions
• Maintenance and growth respiration losses are accounted
for in the experimental determination of BTR
• The difference between leaf and atmospheric vapor density
can be approximated by the atmospheric deficit expressed
as the atmospheric vapor pressure deficit (VPD).
26
Transpiration dependent growth
• CropSyst versions later than 4.12 offer three different modes for
calculating BPT:
1. classical Tanner-Sinclair model
•
BTR is a constant, eq. 2 valid
2. FAO AquaCrop water
productivity
•
BTR is a constant;
VPD is not considered;
equation 2 not used;
unit of water productivity
is g biomass/kg water)
3. Transpiration use
efficiency curve
27
Crop growth
• (optimal) crop growth is governed by the most
limiting condition, either
– radiation (eq. 1) or
– transpiration (eq. 2).
28
Water limited growth, how?
• via reducing transpiration…
29
Crop water uptake, WU (= Tact)
soil layer
n
WU = Σ WUl [mm d-1]
soil water potential
l=1
leaf water potential
WUl = K · Cl/1.5 · (ψsl - ψl)
number of seconds
per day = 86400
root conductance
of soil layer l
30
A range of other "water" factors
•
•
•
•
•
•
•
•
•
•
Act. to pot. transpiration ratio that limits leaf area growth
Act. to pot. transpiration ratio that limits root growth
Maximum daily water uptake
ET crop coefficient at full canopy
Leaf water potential at the onset of stomatal closure
Wilting leaf water potential
Leaf duration sensitivity to water stress
Phenological sensitivity to water stress
Initial leaf area index
fraction of max. LAI at physiological maturity
31
Stress indexes
Stress index is determined as one minus the ratio of actual
to overall potential biomass growth for each day of the
growing season.
Potential growth is defined as the growth calculated from
potential transpiration (Trpot) substituted for Tract.
Actual biomass growth is obtained after growth limitations
have been applied.
This overall stress index is partitioned into light,
temperature, water, and nitrogen stress indices. These
quantities are used as indicators of the plant response
to environmental conditions. All these indices range
from 0 to 1, where 0 is no stress and 1 is maximum
stress.
32
Climate change impact
assessment using CropSyst
(by example of wheat growing in
Central Asia)
33
Objectives
1. Model calibration and evaluation for wheat grown under
currently prevailing climatic conditions in selected agroeco-zones of the study region
a. Crop model selection
b. Site selection (by AEZ), and data collection (surveys)
c. Crop model calibration
2. Definition of business-as-usual management
3. Generation of daily time-step weather data (historic and
future)
4. Modeling the impact of climate change on crop
productivity utilizing developed climate change scenarios
34
Potential biophysical impact of climate
change on crop production in Central Asia
1. Increasing temperature
– warmer winter and early spring (winter crops)
 better early crop growth, less damage by frost
– hotter late-spring, hotter summer
 crop heat stress (lower grain production)
– shorter cropping cycle
 lower biomass production
2. Changes in precipitation (amount and intensity)
3. Increasing CO2
– “carbon fertilization effect”  moderate increase in crop
growth
• Interactions of 1. – 3.
35
Model selection criteria
• Capacity to handle the impact of climate change on crop
growth:
– CO2 response
– temperature response (cold & hot)
– water stress (rainfall variability)
• Capacity for reasonable prediction of
– impact of shallow groundwater (GW-module; upward
movement of water in the soil)
– salinity response (saline soils)
– evapotranspiration in arid environments
– response to soil conservation measures (zero-tillage, surface
residue retention)
• Availability of further, useful modeling tools, such as
– automatic irrigation
36
CO2 fertilization effect in CropSyst
• increase in radiation use efficiency (ε) by a G-ratio factor
• decrease in canopy conductance, increase of WUE
Tubiello et al., 2000
37
Generation of weather data
• Historic: data bases of national met-services, ICARDA
and www
• Future: using greenhouse gas emission scenarios of
IPCC, 2007
– A2: pessimistic; assumes a continuous population growth,
increasing divergence between regions, less transfer of
technological innovations
– A1b: neither optimistic nor pessimistic; assumes population
stabilization, continued globalized world, balance between
fossil-intensive and non-fossil energy sources
• Future periods:
– immediate future: 2011-2040
– mid-term future: 2041-2070
– long-term future: 2071-2100
38
38
Climate change – CO2 concentrations
Increase of the atmospheric CO2
concentration as predicted by
SRES A1B and A2 (redrawn from
IPCC, 2000)
39
Projections of climate change
• Underlying data base: seven IPCC GCMs
No
Name
Country
Year
Resolution
(degrees)
01
BCCR-BCM2.0
Norway
2005
2.8 x 2.8
02
CSIRO-MK3.0
Australia
2001
1.9 x 1.9
04
MIROC3.2
Japan
2004
2.8 x 2.8
08
CGCM3.1(T63)
Canada
2005
2.8 x 2.8
09
CNRM-CM3
France
2005
2.8 x 2.8
10
ECHAM5/MPI-OM
Germany
2003
1.9 x 1.9
12
GFDL-CM2.0
USA
2005
2 x 2.5
 average deviation (delta) from historic climate
(temperature and precipitation) of the seven models
40
40
Business-as-usual (BAU)
• Definition of agronomic management
scenarios based on the usual farmer’s practice
 Model simulations should reflect reality
41
41
Business-as-usual (BAU)
Information about BAU:
From socio-economists team:
1. Fertilizer type
2. Fertilizer amount
3. Week of planting
4. First week of irrigation
5. Last week of irrigation
6. Number of irrigation events
7. Week of harvest
BAU
Planting date
optimal
average
Depending on
location
sub-optimal
+
National recommendations:
1. Dates of fertilizer
application
2. Dates of irrigation
3. Irrigation rates
Fertilizer
application
highest
average/median
Irrigation
lowest
water stressed
recommended
average
42
Weather generators
LARS-WG
stochastic weather generator
Developed by M. Semenov
(Rothamsted Research of
BBSRC)
ClimGen
Available at:
Available at:
modified version of WGEN developed
by Gaylon S. Campbell, Washington
State University
http://www.bsyse.wsu.edu/CS_Suite/ClimGe
http://www.rothamsted.bbsrc.ac.uk/mas43
n/index.html
models/larswg.php
43
Climate change simulations
Historic daily meteorological data:
precipitation, solar radiation, Tmax,
Tmin, RHmax, RHmin, wind speed
Soil:
•Soil physical
properties
•Nmin and SOM
•Soil salinity
•Groundwater
Crop:
•Crop
physiology
•Crop
phenology
Management
:
•Planting date
• Irrigation,
fertilization
•Tillage
Weather
generator
Generated
daily
meteo
data
CropSyst Simulations
Location
Regional
downscaling
Scenario 4
Scenario 3
Current
conditions
Scenario
outputs
GCM - СС
Scenario 2
Scenario 1
44
44
Climate change crop model simulation
results – major governing factors
• higher temperatures:
→
→
→
→
faster growth, shorter growing season
less time for biomass accumulation
higher evaporative demand
increase in crop water requirements
“warmer” (less cold) winters and springs
less frost damage, faster early growth in spring
hotter late spring and summer
increased risk of sterility of flowers
• higher precipitation:
→
→
more water for the crop
increased risk of nitrate leaching
• higher concentration of CO2:
→
carbon fertilization effect
45
Selected sites
Country
Kazakhstan
Uzbekistan
Site name
Shieli
Vozdvizhenka
Petropavlovsk
Kostanay
Khorezm
Syrdarya
Kushmanata
Kuva
Akkavak (2 experiments)
AEZ
310
521
821
521
310
510
510
310
510
Country
Tajikistan
Kyrgyzstan
Site name
Faizabad
Shahristan
Khorasan
Bakht
Spitamen
Uchkhoz
Zhany pakhta
Daniyar
KyrNIIZ
AEZ
1032
532
510
510
510
510
510
510
510
46
46
Climate change projections for the selected sites
47
CC simulation results: Grain yield
Akkavak – Mars, (UZ, irrigated)
Yield (Mg/ha)
6
4
* *
* *
I M L
I M L
A1b
A2
LSD (0.42 Mg/ha)
2
0
H
I M L
A1b
I M L H
A2
Suboptimal Mgmt.
I M L
A1b
I M L H
A2
Average Mgmt.
Optimal Mgmt.
48
Days from emergence until maturity
Days from emergence till maturity
Example: Kushmanata (UZ)
240
-10 days
-12 days
230
Avg. ( ±SD)
Min
220
Max
210
H
I
M
A1b
L
I
M
A2
L
49
(Minimum) temperatures during
vegetative growth
Change in average temperature across all sites
and scenarios
Immediate Mid-term
future
future
Avg.
Range
+ 0.8
0.6-1.0
+ 1.7
1.4-2.4
Long-term
future
+ 2.9
2.2-4.1
50
Maximum temperatures during flowering
Astana
40
50-year Max.
T (°C)
35
95%-Perc.
30
Avg. (±SD)
25
20
I
Hist.
M
A1b
L
I
M
L
A2
51
Irrigation
Irrigation (mm)
(mm) (mm)
Irrigation
Irrigation (mm)
Kuva
Irrigation water requirements
400
Subopt. Mgmt. A1b
Subopt. Mgmt. A2
Shieli
300
Avg. Mgmt. A1b
Avg. Mgmt. A2
250
200
Optimal Mgmt. A1b
Kuva
200
Optimal Mgmt. A2
100
400
Subopt. Mgmt. A1b
150
0
300
100
Subopt. Mgmt. A2
H
I
M
Kuva
Avg. Mgmt. A1b
L
Avg. Mgmt. A2
200
50
400
OptimalMgmt.
Mgmt. A1b
A1b
Subopt.
0
100
OptimalMgmt.
Mgmt. A2
A2
Subopt.
300
0
200
Overall:
H
I
M
L
Avg. Mgmt. A1b
Avg. Mgmt. A2
•
no noteworthy
change
Optimal Mgmt. A1b
H
I
M
L
Optimal Mgmt. A2
• Some
sites:
reduction
in
irrigation
water
requirement
52
100
Water use efficiency
Grain yield vs. actual transpiration, all Uzbek sites
Historic
10
7.5
5
7.5
5
2.5
2.5
0
0
0
100
200
300
400
500
Actual transpiration (mm)
600
Slope (kg/ha/mm):
25.8
22.5
20.3
18.3
Long-term
Mid-term
Immediate
Historic
y = 18.3x – 405.6
R² = 0.758
Yield (Mg/ha)
Yield (Mg/ha)
10
0
100
200
300
400
500
Actual transpiration (mm)
Transpiration Use efficiency increased from 18.3 kg/ha/mm under historic (CO2)
conditions to 25.8 kg/ha/mm in the long-term future
53
Thank you for your attention!
12th CGIAR Steering Committee Meeting for Central Asia and the Caucasus, September 12-14, 2009, Tbilisi, Georgia
54