revisiting n application rates in low-rainfall grain cropping regions of ...
Revisiting N fertilization rates in low-rainfall
grain cropping regions of Australia
A risk analysis
M. Monjardino, T. McBeath, L. Brennan, R. Llewellyn
56 th AARES Conference, 8-10 February 2012, Fremantle
Research problem
Grain production is increasingly risky as
wheat revenue variance has more than
doubled in Australia over the past 15 years
(Kingwell, 2011).
Consequently, decisions surrounding the
most profitable fertiliser rates pose a major
risk management dilemma to farmers.
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Context
• Agriculture the most volatile sector of the Australian economy,
and more volatile than agriculture elsewhere;
• In particular, low-rainfall farmers operate in increasingly
challenging conditions, including:
• High climate and seasonal risk and uncertainty;
• High spatial variability;
• Low nutrient use efficiency (low uptake, losses);
• Intense market volatility;
• Tighter budgets
• 2.2%/yr decline in terms of trade
• fertilizers now 60% of total var. costs of growing a crop
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Index
Context
• Agriculture the most volatile sector of the Australian economy,
and more volatile than agriculture elsewhere;
• In particular, low-rainfall farmers operate in increasingly
challenging conditions, including:
• High climate and seasonal risk and uncertainty;
• High spatial variability;
• Low nutrient use efficiency (low uptake, losses);
• Intense market volatility;
• Tighter budgets
• 2.2%/yr decline in terms of trade
• fertilizers now 60% of total var. costs of growing a crop
250
200
150
100
50
Terms of trade
Prices received
Prices paid
ABARE
Prices paid for fert.
0
1969 1979 1989 1999 2009
Years
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Context – N management
• History of fixed low rates of N in low-rainfall regions as to
minimize downside risk in anticipation of poor seasons;
• N considered risk-increasing input (literature);
• Recommendations based on average season/rules of thumb;
• N fertilizer applications often up front at sowing, but ‘playing the
season’ on the rise with intensification of cropping;
• Slow increase in uptake of precision technologies for fertilization
(e.g. VRT).
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Research question
Since consistently low N inputs lead to N deficiency, which is likely
to be one of the main causes of a gap between actual and
potential yield, especially in the wetter seasons…
Could those farmers in the low-rainfall Australian wheatbelt who
adopt a low-input strategy in the attempt to minimize economic
risk in fact be missing out on greater returns from more intense
cropping in the more favourable seasons
In a nutshell:
Are dryland farmers under-fertilizing with N
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Methodology
We analyse the problem by using a combination of different
tools:
• Case-study climatic and agronomic data
• Crop simulation modelling
• Profit function
• Monte Carlo simulation
• Probability theory
• Finance theory techniques
• Stochastic efficiency analysis
… to evaluate the combined impact of yield and price risk on
long-term performance of N fertiliser strategies on different soil
types.
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Study area – the Mallee
. Karoonda
Map of the SA/Vic Mallee region
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Study area – the Mallee
• Mallee region located across SA, Vic and bit of NSW;
• Approximately 4million ha of cropping land;
• Highly variable soils in dune-swale land system;
• Typical Mediterranean-type climate;
• Mean annual rainfall 250-350 mm;
• Mean annual evaporation around 2000 mm;
• Poor soils with low water-holding capacity;
• Cereals exposed to moisture stress and drought:
• Total crop failure in 10% of seasons;
• High water stress in 30-50% of seasons;
• Crop damage due to unpredictable factors (e.g. locusts last year).
• District average N rate: 15 kg N/ha.
. Karoonda
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Yield simulation response curves
APSIM-simulated wheat crops grown for a range of scenarios
assuming the following conditions:
• Data for 3 representative land management zones in Karoonda;
• Starting mineral soil N of 18, 36, 52 kg N/ha for each zone (0-
110 cm);
• Soil water reset each year on Jan 1 to harvest lower limit
determined by prior simulation;
• Surface residue reset to 1.5 t/ha at the same time;
• Soil N and OM reset on April 1 each year;
• Crops sown soon after break, based on rainfall and timing.
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Land management zones
Zone Location Av. % of
typical
farm
Main soil
type
Level of
subsoil
constr.
Initial
fert.
status
Resp. to
fertiliser
Average
wheat
yld (t/ha)
Dune Hill-top 30 Sand Low Low Good 2.1
Slope
Mid-
slope
50 Sandy
to clay
loam
Med Med Average 1.7
Flat
Plain or
swale
20 Medheavy
clay
High High Poor 1.1
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Dune
Slope
Flat
Wheat yield map reflecting the typical paddock elevation of the Mallee
(Source: Davoren, CSIRO 2010) (0-0.5 t/ha; 0.5-0.8 t/ha; 0.8-1.5t/ha)
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
N fertilization scenarios
• Eight N fertilizer application rates:
• 0, 7.5, 15, 30, 60, 90, 120, 150 kg N/ha
• applied up front at sowing (rate X + 0) and in combination with an inseason
application at GS 31-40 (tactical or split) in each zone.
• Tactical N application triggered in APSIM by the simultaneous
occurrence of threshold values of:
• soil water (> crop lower limit);
• soil N (< 100 kg N ha -1 );
• rainfall (≥ 10 mm over 3 days) within GS31-40.
(Tactical application of N was not triggered in 21 seasons (35%)).
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Wheat yield (t/ha)
Wheat yield (t/ha)
N fertilization scenarios
• Eight N fertilizer application rates:
• 0, 7.5, 15, 30, 60, 90, 120, 150 kg N/ha
• applied up front at sowing (rate X + 0) and in combination with an inseason
application at GS 31-40 (tactical or split) in each zone.
• Tactical N application triggered in APSIM by the simultaneous
occurrence of threshold values of:
• soil water (> crop lower limit);
• soil N (< 100 kg N ha -1 );
• rainfall (≥ 10 mm over 3 days) within GS31-40.
(Tactical application of N was not triggered in 21 seasons (35%)).
0.5
0.4
0.3
0.2
0.1
0
2002
0 20 40 60 80 100 120 140 160
N rate (kg/ha)
5
4
3
2
1
0
1958
0 20 40 60 80 100120140160
N rate (kg/ha)
Dune
Slope
Flat
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Data sets
• APSIM-generated wheat yields based on 60-year rainfall timeseries
(1950-2010);
• Historical prices (1970-2010) for wheat grain and urea fertiliser
(46% N) (adjusted to 1998 consumer price index);
• Wheat x N price correlation coefficient of 0.12.
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Yield variability
• Frequency distributions of wheat yields generated for each of
the N treatments;
• PDF with the best fit (as measured by the AD stats test) selected
for use in Monte Carlo simulation of economic net returns.
• Mean yields, SD, CV, as well as Deciles, used for comparison of
yield production and variance.
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Profit function
NR nz = (Y nz x P w ) – ((R n1z +(R n2z x f))*P n ) – C o – (C t * f)
NR nz
Y nz
P w
R n1z
R n2z
P n
C o
C t
f
net returns by total N rate n on management zone z ($/ha)
crop yield by total N rate n on management zone z (kg/ha)
price of ASW wheat grain ($/kg)
rate of N applied at sowing on management zone z (kg N/ha)
rate of N applied in-season on management zone z (kg N/ha)
price of N (i.e. price of urea/0.46) ($ kg/N)
other costs ($/ha)
operational cost of applying extra fertilizer in-season ($/ha)
frequency of seasons with tactical N application in-season
C o = C v + C f + I v + D m
C v variable costs of growing wheat, incl seed purchase and treatment, herbicides, fuel and oil,
and fertilizers other than N
C f
fixed costs of production apportioned on a $/ha basis such as repairs and maintenance,
labour, insurance and levies
I v interest on variable costs (half at 8%)
depreciation of machinery investment (10% of $200/ha in machinery investment)
D m
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Variability of net returns
• @RISK used to generate 1000 Monte Carlo simulations of
the profit function with random samples for yield and prices
drawn from their PDFs;
• PDF of economic net returns developed for all scenarios;
• Quantification of NR for alternative scenarios using a
combination of 8 measures of risk around the mean...
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Economic-risk measures
1. Mean of expected NR (magnitude);
2. Mode of expected NR (frequency);
3. Standard deviation of NR, SD (dispersion);
4. Coefficient of variation, CV (variance);
5. Probability of break-even, P(NR ≥ 0) (positive NR);
6. Conditional value at risk of the lowest 10% of possible
outcomes, CVaR0.1, (risk of extreme financial loss associated
with unfavourable events);
7. Return on total fertilizer investment at risk (R N );
8. Return on tactical fertilizer investment at risk (R NT ).
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
CVaR0.1
Prob. of break even
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Farmers’ preferences
• Fertilization preferences under risk revealed through a
Stochastic Efficiency with Respect to a Function (SERF)
analysis (Hardaker et al., 2004);
• SERF ranks a set of alternative (N) strategies in terms of
Certainty Equivalence (CE) for a specified range of risk
attitudes, which is measured by a Constant Absolute Risk
Aversion coefficient (0.0 < CARA < 1.0);
• Constant aversion to risk implies a utility function, e.g. the
negative exponential utility function:
U(W) = 1- Є -cW
W
wealth or income expressed as a wealth equivalent
c constant absolute risk aversion (CARA) coefficient (c > 0)
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Results
1. Yield variance analysis
2. Economic-risk performance
3. Impact of risk aversion
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Yield variance (Slope, 0 – 30 kg N/ha up front )
Mean
SD
CV
% years
Kg N/ha
(kg/ha)
(kg/ha)
< 0.25 t/ha < 0.5 t/ha < 1.0 t/ha < 2.0 t/ha > 5.0 t/ha
0
15
30
+ 0 343 143 0.42 20 90 100 100 0
+ 15 779 462 0.59 20 47 63 100 0
+ 30 735 569 0.77 20 47 63 100 0
+ 60 741 578 0.78 20 47 63 100 0
+ 90 744 584 0.78 20 47 63 100 0
+ 150 746 584 0.78 20 47 63 100 0
+ 0 579 250 0.43 18 27 100 100 0
+ 15 1153 1087 0.94 18 27 57 78 0
+ 30 1261 1378 1.08 18 27 57 75 0
+ 60 1322 1647 1.24 18 27 57 70 0
+ 90 1322 1603 1.20 18 27 58 70 0
+ 150 1339 1780 1.32 18 27 58 70 0
+ 0 883 414 0.47 15 23 45 100 0
+ 15 1319 1028 0.78 17 23 43 75 0
+ 30 1389 1218 0.88 17 23 43 70 0
+ 60 1465 1367 0.93 17 23 43 70 0
+ 90 1477 1411 0.95 17 23 43 70 0
+ 150 1485 1467 0.99 17 23 43 70 0
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Yield variance (Slope, 0 – 30 kg N/ha up front )
Mean
SD
CV
% years
Kg N/ha
(kg/ha)
(kg/ha)
< 0.25 t/ha < 0.5 t/ha < 1.0 t/ha < 2.0 t/ha > 5.0 t/ha
0
15
30
+ 0 343 143 0.42 20 90 100 100 0
+ 15 779 462 0.59 20 47 63 100 0
+ 30 735 569 0.77 20 47 63 100 0
+ 60 741 578 0.78 20 47 63 100 0
+ 90 744 584 0.78 20 47 63 100 0
+ 150 746 584 0.78 20 47 63 100 0
+ 0 579 250 0.43 18 27 100 100 0
+ 15 1153 1087 0.94 18 27 57 78 0
+ 30 1261 1378 1.08 18 27 57 75 0
+ 60 1322 1647 1.24 18 27 57 70 0
+ 90 1322 1603 1.20 18 27 58 70 0
+ 150 1339 1780 1.32 18 27 58 70 0
+ 0 883 414 0.47 15 23 45 100 0
+ 15 1319 1028 0.78 17 23 43 75 0
+ 30 1389 1218 0.88 17 23 43 70 0
+ 60 1465 1367 0.93 17 23 43 70 0
+ 90 1477 1411 0.95 17 23 43 70 0
+ 150 1485 1467 0.99 17 23 43 70 0
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Yield variance (Slope, 0 – 30 kg N/ha up front )
Mean
SD
CV
% years
Kg N/ha
(kg/ha)
(kg/ha)
< 0.25 t/ha < 0.5 t/ha < 1.0 t/ha < 2.0 t/ha > 5.0 t/ha
0
15
30
+ 0 343 143 0.42 20 90 100 100 0
+ 15 779 462 0.59 20 47 63 100 0
+ 30 735 569 0.77 20 47 63 100 0
+ 60 741 578 0.78 20 47 63 100 0
+ 90 744 584 0.78 20 47 63 100 0
+ 150 746 584 0.78 20 47 63 100 0
+ 0 579 250 0.43 18 27 100 100 0
+ 15 1153 1087 0.94 18 27 57 78 0
+ 30 1261 1378 1.08 18 27 57 75 0
+ 60 1322 1647 1.24 18 27 57 70 0
+ 90 1322 1603 1.20 18 27 58 70 0
+ 150 1339 1780 1.32 18 27 58 70 0
+ 0 883 414 0.47 15 23 45 100 0
+ 15 1319 1028 0.78 17 23 43 75 0
+ 30 1389 1218 0.88 17 23 43 70 0
+ 60 1465 1367 0.93 17 23 43 70 0
+ 90 1477 1411 0.95 17 23 43 70 0
+ 150 1485 1467 0.99 17 23 43 70 0
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Yield variance - summary
• Lowest CV of the mean yield around 15 kg N/ha at sowing , i.e.
farmers currently target lower yield variance;
• However, up front N rates up to 90 kg N/ha in the dune, 60 kg
N/ha in the slope and 30 kg N/ha in the flat, could be beneficial
in terms of increasing yields significantly, while still managing
yield variance.
• 60 kg N/ha in-season the point at which the highest yield
variance occurred in most cases.
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Economic-risk performance (Dune)
Kg N/ha
Mean
Mode
SD
CV P (NR ≥ 0)
CVaR 0.1
R N
R NT
($/ha)
($/ha)
($/ha)
(%)
($/ha)
($NR/$N)
($NR/N T )
+ 0 -30 -41 24 0.81 11 -68 1.9
15
30
60
90
+ 7.5 27 -42 79 2.92 57 -78 3.7 5.1
+ 15 27 -61 87 3.21 54 -84 2.8 2.6
+ 30 24 -91 93 3.84 53 -96 1.8 1.2
+ 60 10 -96 98 9.92 47 -117 0.9 0.4
+ 0 22 31 49 2.22 69 -67 2.6
+ 7.5 103 25 123 1.19 79 -78 4.2 7.2
+ 15 117 40 146 1.24 77 -84 3.8 4.2
+ 30 118 123 157 1.33 76 -96 2.9 2.1
+ 60 102 -33 159 1.57 70 -114 1.7 0.9
+ 0 106 -2 109 1.02 82 -84 2.7
+ 7.5 185 182 203 1.10 82 -149 3.5 6.8
+ 15 200 162 222 1.11 82 -128 3.4 4.3
+ 30 206 42 236 1.14 81 -139 2.9 2.3
+ 60 198 78 260 1.31 77 -150 2.1 1.0
+ 0 183 273 171 0.93 83 -125 2.6
+ 7.5 208 186 243 1.17 81 -195 2.7 2.2
+ 15 225 187 274 1.22 80 -167 2.6 1.8
+ 30 228 57 290 1.27 79 -147 2.3 1.0
+ 60 215 376 310 1.44 75 -199 1.8 0.3
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Economic-risk performance (Slope)
Kg N/ha
Mean
Mode
SD
CV P (NR ≥ 0)
CVaR 0.1
R N
R NT
($/ha)
($/ha)
($/ha)
(%)
($/ha)
($NR/$N)
($NR/N T )
15
30
60
90
+ 0 7 55 57 8.46 57 -94 2.2
+ 7.5 84 23 153 1.83 69 -121 4.8 6.9
+ 15 113 -38 242 2.14 67 -126 4.6 4.8
+ 30 131 29 400 3.06 64 -131 3.3 2.7
+ 60 117 -52 342 2.92 59 -151 1.9 1.3
+ 0 55 133 92 1.69 72 -111 2.7
+ 7.5 114 166 188 1.64 71 -144 3.7 5.4
+ 15 134 88 234 1.75 69 -149 3.4 3.4
+ 30 137 -23 263 1.93 65 -146 2.7 1.8
+ 60 133 -45 300 2.25 60 -168 1.7 0.9
+ 0 105 144 165 1.58 68 -143 2.1
+ 7.5 142 9 250 1.76 69 -188 2.4 3.2
+ 15 161 -43 306 1.91 67 -178 2.5 2.6
+ 30 158 -181 363 2.30 56 -180 2.0 1.1
+ 60 143 -102 367 2.63 55 -201 1.4 0.4
+ 0 128 63 253 1.98 65 -198 1.7
+ 7.5 149 16 330 2.21 63 -205 1.8 2.1
+ 15 142 -159 358 2.52 55 -199 1.6 0.8
+ 30 142 -187 371 2.61 53 -210 1.4 0.3
+ 60 123 -142 376 3.05 51 -234 1.0 0.0
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Economic-risk performance (Flat)
Kg N/ha
Mean
($/ha)
Mode
($/ha)
SD
($/ha)
CV P (NR ≥ 0)
(%)
CVaR 0.1
($/ha)
R N
($NR/$N)
R NT
($NR/N T )
+ 0 55 86 133 2.39 67 -175 2.1
+ 7.5 79 38 193 2.45 67 -261 2.6 2.1
7.5
+ 15 95 -7 198 2.09 64 -163 2.4 1.8
+ 30 134 -76 487 3.63 54 -140 2.4 1.7
+ 60 122 -143 483 3.96 48 -160 1.2 0.8
+ 0 66 23 154 2.34 68 -200 1.7
+ 7.5 95 -4 197 2.07 64 -165 2.4 2.7
15
+ 15 135 -83 437 3.24 56 -138 3.0 3.0
+ 30 145 -128 498 3.44 53 -148 2.4 1.9
+ 60 154 -82 845 5.50 50 -167 1.4 0.9
+ 0 74 26 188 2.55 66 -260 1.2
+ 7.5 129 -95 387 2.99 54 -149 2.4 4.9
30
+ 15 177 -88 1018 5.74 52 -151 2.7 3.9
+ 30 173 -92 738 4.26 50 -164 2.3 2.3
+ 60 168 -139 808 4.79 45 -182 1.5 1.1
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Economic-risk performance
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Economic-risk performance
• Highest mean NR with mid to high N rates (> 30 kg N/ha)
applied up front and/or tactically on all three zones, especially
the dune;
• The standout strategies include 30 - 90 kg N/ha in the dune, 15-
60 kg N/ha in the slope, but only two scenarios (15 + 7.5 and
7.5 + 15 kg N/ha) offer a benefit relative to the standard rate in
the flat.
• The least attractive economic-risk scenarios result both from
under-fertilizing with zero/low N inputs and from over-fertilizing
with excessively high inputs (especially in-season).
• The results indicate that there is much scope to use more N
within a paddock, especially in the dune and the slope zones,
as well as potential for a small practice change in the flat zone.
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Best economic-risk scenarios
• In summary, the best scenarios overall in terms of economics
and risk performance are assumed to meet all of the following
conditions:
• Mean NR ≥ Mean NR 15 kgN/ha in that zone
• CV≤ 3.0
• P(NR≥0) > 50% and P(NR≥0) > P(NR≥0) 15 kgN/ha in that zone
• CVaR 0.1 ≤ -$150/ha
• R N ≥ R N 15 kgN/ha and R N >0
• R NT ≥ $1.0 per dollar of N tactical
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Dune
Slope
Flat
Best economic-risk scenarios
Kg N ha -1
Mean
CV P (NR ≥ 0)
CVaR 0.1
R N
R NT
($/ ha)
(%)
($/ha)
($NR/$N)
($NR/$N T )
+ 7.5 103 1.19 79 -78 4.2 7.2
30 + 15 117 1.24 77 -84 3.8 4.2
+ 30 118 1.33 76 -96 2.9 2.1
+ 0 106 1.02 82 -84 2.7
+ 7.5 185 1.10 82 -149 3.5 6.8
60 + 15 200 1.11 82 -128 3.4 4.3
+ 30 206 1.14 81 -139 2.9 2.3
+ 60 198 1.31 77 -150 2.1 1.0
90
+ 0 183 0.93 83 -125 2.6
+ 30 228 1.27 79 -147 2.3 1.0
+ 7.5 58 2.22 64 -110 5.5 6.2
7.5 + 15 84 1.56 68 -108 4.8 4.2
+ 30 87 2.80 59 -124 3.0 2.2
15
+ 7.5 84 1.83 69 -121 4.8 6.9
+ 15 113 2.14 67 -126 4.6 4.8
+ 0 55 1.69 72 -111 2.7
30
+ 7.5 114 1.64 71 -144 3.7 5.4
+ 15 134 1.75 69 -149 3.4 3.4
+ 30 137 1.93 65 -146 2.7 1.8
60 + 0 105 1.58 68 -143 2.1
7.5 +15 95 2.09 64 -163 2.4 1.8
15 +7.5 95 2.07 64 -165 2.4 2.7
CSIRO. Insert presentation title, do not remove CSIRO from start of footer
Selection of top economic-risk scenarios
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Selection of top economic-risk scenarios
80%
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Selection of top economic-risk scenarios
80%
20%
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Selection of top economic-risk scenarios
60%
15%
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Selection of top economic-risk scenarios
40%
10%
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Impact of risk aversion
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Impact of risk aversion
Upfront overtake tactical
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Impact of risk aversion
Upfront overtake tactical
Typical rate overtakes top scenarios
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Impact of risk aversion
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Impact of risk aversion
Upfront overtakes tactical
Typical rate overtakes top scenarios
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Impact of risk aversion
Mallee farmers
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Conclusions
• Mallee dryland farmers relatively risk-averse and potentially
missing out on the returns available from more intense
cropping in the good years on most of their farm (80%);
• In other words, when long-term yield and price risks are
factored in, it is evident that farmers are better off if they
reduce the probability of under-fertilizing in the dry seasons
(i.e. making a loss), while increasing the probability of
sufficiently fertilizing when the season develops well (by
providing enough N up front or ‘playing the season’);
• Some tactical fertilization can increase farmers’ mean NR
while in some cases reducing income variance, although full
potential of tactical fert is likely to be realised when factors like
grain protein, crop rotation and whole-farm farm budget are
factored in the analysis;
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Conclusions
• Use of variable fertiliser rates based on soil-specific
management zones has the potential to not only increase profit
but reduce risk, based on a number of risk metrics;
• High riskiness and complexity of dryland cropping presents
opportunity to manage nutrients more efficiently and explore
sustainable intensification;
• If Mallee farmers are able to establish the relevance of this
analysis to their own farm and soil conditions by e.g. defining
the % of a certain soil type on their farm, and relating the
analysis to the initial N fertility on their given management unit,
then these results could usefully inform individual farmers’
decisions.
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Study limitations
• Variance of yield potentials in APSIM lower than actual crops
since the model cannot accurately capture all phenomena,
including unpredictable damage by weather, pests, diseases
and weeds, or occasional crop failures caused by ‘haying off’ in
extremely dry seasons;
• The model has been fixed to have the same starting N
conditions for every season in a given soil type, which may in
fact vary considerably with prior management.
• Need to place study in whole-farm context as mixed-enterprise
farms likely to cope better with risk.
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
Future directions
• Validate Mallee case study against data attained from growers;
• Follow-up analysis to test the effect of higher intensity of
cropping on increasing initial soil test N values;
• Conduct similar analysis for P to test value of replacement
strategies;
• Apply risk framework to other regionally relevant nutrient
management case studies across Australia to compare risk
profiles and management options.
CSIRO. Revisiting N fertilisation rates in low-rainfall grain cropping regions. A risk analysis
CSIRO Ecosystem Sciences
Marta Monjardino
Social and Economic Sciences
Phone: 61 (0)8 8303 8413
Email: marta.monjardino@csiro.au
Web: www.csiro.au/ces
Acknowledgements
A. Whitbread
M. Robertson
J. Kandulu
E. Qureshi
J. Ouzman
D. Gobbett
I. Fillery
(all from CSIRO)
Thank you
Contact Us
Phone: 1300 363 400 or +61 3 9545 2176
Email: enquiries@csiro.au Web: www.csiro.au
Change language