Managing climate risk in the eastern states

1
Managing climate risk in the eastern states

• Peter Hayman
• NSW Agriculture

2
There is more to managing climate risk in the
east than seasonal climate forecasts

• Risk analysis based on historic climate record -
  simulation modelling (APSRU/QDNR)
• Building robust vs twitchy farming systems
• Questions of climate change
• Questions of drought policy
• Weather derivatives
• however, the focus of most questions from
  farmers is climate forecasts

3

• NSW Agriculture does not issue seasonal climate
  forecasts, rather it works with farmers and their
  advisers to improve risk management in farming
  systems

4
A long wait

• the importance to the farmer, the
  horticulturalist, and pastoralist of knowing
  beforehand the probabilities of dry or wet
  seasons, and whether the rains will be early or
  late, or both, has naturally led to a desire for
  seasonal forecasts. They have them, it is said,
  in India why not in Australia? Sir Charles
  Todd (1893)

5
Dubbo Annual Rainfall
6
Prior to end Nov 2002
6 months
3 months
9 months
12 months
7
24 months
18 months
36 months
8
El Nino drought of 2002 a signal event ?

• From language of risk communication, a signal
  event or threshold event becomes symbolically
  charged.
• Examples are Chernobyl or Bhopal
• Willet 2000 suggested that 1997 El Nino in USA
  was a signal event
• In eastern Aust, 2002 will explicitly or
  implicitly be part of the conversation on climate
  risk, the role of climate science in Ag and
  climate change

9
El Nino drought of 2002 a signal event ?

• Impact extensive and dramatic
• Climate science provided a narrative for pictures
  and then become part of the story (Howard
  Anderson)
• Significant policy shifts during this drought
  (Hanrehans lament - Australia and drought -
  people policy and place)
• Farmhand, drought proofing and Wentworth group
• Climate change and this hotter than normal
  drought (wwf Monash University and Nicholls
  2003)

10
A word from 1902

• when more records are available, an accurate
  forecast can probably be made for a considerable
  period in advance. Needless to say, when that
  time arrives, it will be possible to greatly
  reduce, or even entirely prevent, the now
  constantly recurring losses in stock and crops
  for if it be known that a succession of dry
  seasons are due, understocking the country must
  be resorted to, and its reverse when damp seasons
  are to follow

11
Looking beyond cycles in rainfall

• Barling based his forecasts on cycles
• Forecasts of climate based on the interactions
  between the oceans and the atmosphere is one of
  the premiere advances of the atmospheric science
  at the close of the 20th century. AAS (1999)
• Science's gift to the 21st Century. Glantz
• The New Green Revolution. Cited in Hansen 2002

12
Strong El nino signal leads to higher variability
some prediction
13
(No Transcript)
14
(No Transcript)
15
(No Transcript)
16
(No Transcript)
17
(No Transcript)
18
Given the advances in climate science and a
decade of operational forecasts from BoM and QCCA
and range of private forecasters are farmers
using them ?
19
AFFA survey of 2500 farmers in 2001

• Are you aware of seasonal climate forecasting ?
• Do you take seasonal climate forecasts into
  account when making farm management decisions ?

20
(No Transcript)
21
(No Transcript)
22
(No Transcript)
23
Time to reach 25 of US popn

• PC 15yrs
• Cell phone 13yrs
• Microwave 30yrs
• VCR 34 yrs
• Radio 22 yrs
• TV 26 yrs

24
Many studies on what limits farmers use of
seasonal forecasts

• Accuracy 70-75 or 80-90 accurate Accuracy
  limits on-farm use Kondinin
• Before the event Timing of forecast major
  limitation URS (2000).
• Context What does it mean on this paddock this
  year

25
Three phases of SCF

• 1 Sceptical
• 2 Enthusiastic adoption
• 3a Mature - SCF helps us decide which way to lean
  not jump.
• 3b Cynical rejection
• May be intelligent non-adoption or
  diss-adoption.

26
Appropriate confidence in SCF

• From blind rejection to blind belief
• Dr Who.. I love humans, they can find patterns in
  anything
• Humans will read meaning into anything jagged.

27
Fooled by Randomness

• Humans are not good intuitive statisticians
• We are probability blind - we find it hard to
  think of alternative futures much less
  alternative histories.
• Survivorship bias - careful on case studies of
  decision makers that destocked heavily at the
  start of 2002

28
25 years La Nina type
25 years El Nino type
29
If_Then_Else

• IF the season is going to be dry - THEN plant
  wheat chickpeas ELSE - canola
• If the end point is better risk management,
  misunderstanding forecasts as categorical will
  result in poorer risk management than if people
  never heard of the forecast

30
Why we need probabilities

• 1. It is honest to be clear about the
  uncertainties.
• Laplace Probability refers in part to our
  knowledge and in part to our ignorance
• 2. Probabilities encourage risk management
• The belief in, and acceptance of, a range of
  alternative outcomes.

31
Communicating probability

• Farmers have said they want to know whether it
  is likely to be dry, wet or average, not whether
  there is a 60 chance of getting 40 of the
  average rainfall
• Mumbling so that can never be wrong

32
(No Transcript)
33
However ....

• People deal with uncertainty all the time - buy
  shares, get married, live on fault line, plant
  crops, buy cattle
• Is it that people are not used to hearing about
  uncertainty from scientists ?

34
Winter in Tamworth
lt 266 mm
gt 340 mm
35
June - Nov when April May SOI phase is negative
or rapidly rising
36
(No Transcript)
37
(No Transcript)
38
(No Transcript)
39
(No Transcript)
40
Risky Decisions

• Choice Consequence
• IF you use X rate of fertiliser you will get Y
  yield
• Choice chance consequences
• If you use X rate of fertiliser, depending on the
  season, you will get Y(1) Y(2) Y(3)

41
(No Transcript)
42
(No Transcript)
43
(No Transcript)
44
(No Transcript)
45
100 mm
50 mm
150 mm
200 mm
46
Economic response to N
Good yield
Average yield (mid 1/3)
Poor yield
47
Economic response to N
Average of 90 years
48
Economic response to N
49
Economic response to N
50
The flatness of response...

• Anderson (1975) Precision is pretence and great
  accuracy an absurdity

51
Nature of the N fertiliser decisions many others

• Relative flatness of response near the optimum
  rate of N
• Although the outcome changes substantially the
  decision will only be sensitive to large swings
  in probability
• This has implications for the use of SCF