Embed Size (px)
Citation preview
Operational Agromettools and methodologies
Roger Stern21 October 2008
Summary
Main focusClimate variability and climate changeExample:
The McKnight CCRP projectsExample:
Climate change and risks in Southern ZambiaSome toolsTraining for simple climatic analysesDiscussion – any time!
Main focus
Tools and methodsFor agroclimatic analyses
For the NMS to be centre-stageUses the historical data
With climate change always an important issue
How to study climate change?
Use the global climate modelsSo the physicists from Met Services are keyThis has been the main weapon so far
Use the historical dataSo statisticians would (should, could?) be involved
There are time-scale issues10-20 years includes the time-scale for the MDGs50-100 years is to the limit of people born today
How to study climate change
If the role of the historical data were fully recognised then:
a) National Met Service (NMS) staff might be recognised as key players
b) Making good use of the existing recordswould suddenly become relevant to climate change!
African Climate Report 2004
A report commissioned by the UK Govt to review:African climate science, policy and options
for actionMain Conclusion:
“The optimal management of activities directly influenced by interannual climate variability has the potential to serve as a forerunner to engagement in the wider issue of climate change. “And the authors are all modellers!
World Bank view – 2004
Executive summary
A new international consensus has emerged on the need for adaptation to climate change.…While priority attention for adaptation is indeed needed, we argue that it is wise to look before you leap.…Adaptation is likely to be more successful to the extent that it is incorporated into the sustainable development process, and recognizes that:
response to current climate variability and extremes is a necessary, if not sufficient, part of an effective adaptation strategy.
DFID view2005
Climate Proofing Africa
Key points from DFID report
Climate change will hit developing countries hardest.Africa has an extreme and unpredictable climate. This already obstructs its development. Climate change will increase vulnerability levels in Africa.
Developing capacity to deal with today’s climate variability is the best way to equip Africa to deal with tomorrow’s climate change.
ICRISAT 2006
Same idea!
Even if your interest is climate change
The main (initial) task is to use the historical data to study the existing climate variability
And also the farmers options
McKnight Agricultural research
McKnight foundationResearch in 12 countriesAgriculture – wondering about climatologySimilar to ASARECA!This does not affect many other important parts of agroclimatologyBut is an area where a wide range of tools would be useful
An illustration from before
We use an example from Zambia to show what is possible with a simple rainfall analysisPartly because the ease of adding a simple climatic analysis
Impressed the McKnight project teams most!Though the crop simulation models
Will also be very useful in the long run
ZAMBIA
The challenge
Emigration from Southern ZambiaFarmers are citing climate change as a reasonWhat is the evidence for climate change?
In relation to current farming practice
Date of sowing is the critical eventWhat criteria could farmers use?
Combine early sowing with low risk of replantingCan they use their own gauges to decide for themselves?
What options do they have to minimise risk?Crop variety, crop mix and season length
Existing stations for the project
Livingstone Met
Moorings ( Monze)
Chipepo Met
Kalomo F.T.CChoma Met
Magoye
Nanga
Sinazeze
25.00 25.50 26.00 26.50 27.00 27.50 28.00 28.50
-18.00
-17.50
-17.00
-16.50
-16.00
-15.50
Data from 1921 to 2003
Data from 1950 to 2003
Data from 1950 to 2003
The data
The daily data for two years1999 was La Nina, 1991 was El Nino
Including a possible start day each yearexplained later
Moorings 1999
C:\Documents and Settings\Administrator\My Documents\Roger\Zambia\Results\Moorings1999.emf
Annual total 1033mm
Moorings1991
Annual total 395mm
Longest dry spell in January to March for normal yearsDry day defined as less than 0.85mm.
Data from Moorings 1921 to 2003
Years20001990198019701960195019401930
Num
ber o
f day
s
25
20
15
10
5
0
R2 = 0.1 P = 0.06
Number of days
30
25
20
15
10
5
0
Num
ber of days
30
25
20
15
10
5
0
200019901980197019601950194019301920
Num
ber of days
30
25
20
15
10
5
0
Longest dry spell in January to March La Nina years
Normal years
El Nino years
Mean = 10 days
Mean = 13 days
Mean = 16.5 days
The start of the season
Here we use the following definitionThe start is the first occasion
after November 1st - or 15th
with more than 20mmwithin a 2-day period
We also add a further conditionno dry spell exceeding 10 daysin the next 30 days
Sowing dates from 1st November (red) with successful dates (blue) when differentFirst occasion w ith more than 20mm w ithin a 2-day period. Extra condition is dry spell of more than 10 days
Data from Moorings 1921 to 2004
Year2010200520001995199019851980197519701965196019551950194519401935193019251920
Day
num
ber f
rom
1st
Jul
y
200
195
190
185
180
175
170
165
160
155
150
145
140
135
130
125
120
115
110
105
100
Replanting is needed on 19 of the 83 years. This is an overall risk of about 23%
1st Nov
15thNov
1st Dec
Risk of replanting in years with sowing before 15th NovemberSow ing possible in 34 years and replanting needed in 13 (no planting possible in 1 year)
Data from Moorings 1921-2004
Year2010200520001995199019851980197519701965196019551950194519401935193019251920
Day
s of
yea
r fro
m 1
st J
uly
180
175
170
165
160
155
150
145
140
135
130
125
120
November 15
December 1
No successful planting
Early planting possible in 34 of the 83 years. Replanting needed in 13 of these years. Risk of replanting = 40%
Sowing dates from 15th November (red) with successful dates (blue) when differentFirst occasion w ith more than 20mm in a 2-day period. Extra condition is dry spell of more than 10 days in the follow ing 30 days
Data from Moorings 1921 to 2004
Year2010200520001995199019851980197519701965196019551950194519401935193019251920
Day
num
ber f
rom
1st
Jul
y
200
195
190
185
180
175
170
165
160
155
150
145
140
135
130
125
120
Replanting needed in 10 of the 83 years. Risk is about 12%
Risk is now more date independent. So little point in waiting longer
Sowing dates from 15th November with dry spell now 12 days
Data from Moorings 1921 to 2004
Year2010200520001995199019851980197519701965196019551950194519401935193019251920
Day
num
ber f
rom
1st
Jul
y
200
195
190
185
180
175
170
165
160
155
150
145
140
135
130
Replanting needed in 6 of the 83 years. A risk of about 7%
Per10Per12Per15SSSeries5Series6Series7Series8
Percentage of years with dry spell exceeding 10, 12 and 15 days in following 30 days
Data from Moorings 1921-2004
Days from 1 July190185180175170165160155150145140135130125120115110105100
Per
cent
age
90
85
80
75
70
65
60
55
50
45
40
35
30
25
20
15
10
5
0
1st Nov 15th Nov 1st Dec
Risk of a dry spell exceeding 10 days within the next 30 days is about 45%
Risk exceeding 15 days is about 15%
Percentage of years with dry spell exceeding 10, 12 and 15 days in the following 30 days
Data from Livingstone 1950 to 2004
Days from 1st July180175170165160155150145140135130125120115110105100
Per
cent
age
90
85
80
75
70
65
60
55
50
45
40
35
30
25
20
15
10
5
0
1st Nov 15th Nov 1st Dec
Livingstone 1950 to 2003
No point in waiting beyond 15 November. But risk remains higher than Moorings.
Advantage of more drought resistant crop, (or larger rainfall for planting) is greater than for Moorings.
Percentage of years with dry spell exceeding 10, 12 and 15 days in 30 days following plantingUsing data from 1950 to correspond to other stations
Data from Moorings 1950 to 2004
Day number from 1st July190185180175170165160155150145140135130125120115110105100
Per
cent
age
90
80
70
60
50
40
30
20
10
0
Moorings with years from 1950 to 2003 for comparison
During the season
Flowering is the critical periodCan we ensure that the planting date is such that the risk of a long dry spell during flowering is minimised?
The flowering period for maize is about 20 days.
We therefore look at the longest dry spell for 20-day periods
Risk of a dry spell of more than 10, 12 or 15 days in the folloowing 20 days
Data from Moorings 1921 to 2004
Days from 1st July260255250245240235230225220215210205200195190185180175170165160
Per
cent
age
of y
ears
55
50
45
40
35
30
25
20
15
10
5
0
1st Jan 1st Feb 1st March
To minimise the risk, try to ensure that the flowering period starts by the 1st February
Chance of a dry spell exceeding 10, 12 or 15 days in a 20 day period following the day shown
Data from Livingstone 1950 to 2004
Days from 1 July260255250245240235230225220215210205200195190185180175170165160
Per
cent
age
60
55
50
45
40
35
30
25
20
15
10
5
0
For Livingstone the pattern is slightly different. (It would be good to check with a neighbouring station, if a long record exists.)
1st Jan 1st Feb
A more modern method
Use Choma for illustrationAnd fit a model to the daily rainfall
Then split the modelLooking separately at El Nino, ordinary and La Nina years
We find a usable resultWhich needs checking with other stations
Days from 1st July
The chances of rain Mean rain per rain day
Data for Choma 1950 - 2003
Probability of rain following a dry spell
Days from 1st July
The other curves:a) mean rainfall per rain dayb) probability of rain after rainc) prob of rain after one dry dayare identical for the 3 types of year
Mid December
This is initial evidence that:
a) El Nino affects risk of dry spells
b) only after mid December
c) and dry spells only!
A full analysis
Would use other tools and methodologiesTrend analysisFAO water-balance indexCrop simulation modelsAnd so on
But this has indicated what is possible With a simple summary of the historical rainfall data
Data and tools
Data – are they available?Yes – more than many people thinkWhere not made available can they be simulated?
ToolsConsider the types of analysis for agricultural riskAnd mention three types of tool
SkillsSome training may be neededWe also consider some tools for capacity-building
1. Statistical packagesA first one is Instat
A beginners packageWith an extra menu for climatic analysesLike the start of the rains and dry spellsAnd a climatic guide
A powerful package is Genstat*Already useful for agricultureNow with added tools for environmental analysesLike the statistics of extremes and circular dataA free version for Africa is availableAnd there is a climatic guide
In the future there is R!
2. Marksim
A product from CIAT that GeneratesSimulates daily climatic datafor crop modelling and risk assessment
For the whole developing worldAnd up to 99 years
Of rainfallmax and min temperatureand radiation
Ready for DSSAT or APSIMNeeds more testing – but it is a start
3. Crop simulation modelsCrop simulation software – ApSim or DSSAT
To explore agricultural scenarios
Needs 3 types of inputCrop, Soil, Climatic
But the whole variability (in the model) is justbecause of inter-annual climatic variability!
So using a crop simulation modelTranslates climatic variability into yield variability!(To the best ability of the model!)
Also an obvious tool to study the effects of:El NinoClimate change scenarios
Capacity development
SIAC = Statistics in Applied ClimatologyE-SIAC is an initial e-learning course
Facilitated, part-timeSome internet access, but also off-line workOver 10-week period
F-SIAC is face-to-face trainingOnly for those who survive e-SIAC!
Both courses made use of:Moodle and CASTAs well as the tools for the previous analysis
E-SIAC
Overall – started in 2005Now have over 200 graduates from over 30 countries, mainly African
Keep in touch on ALUMNI siteFor example:
For the ASARECA project (see Friday):About 25 staff from
Rwanda, Ethiopia, Sudan, Uganda, Kenya , DRCTook e-SIAC from January or April 2008
E-SIAC resources include CAST
CAST = Computer Assisted Statistics teaching
From Massey University – New ZealandElectronic statistics textbookInteractive graphics – so to be “played”, not just read
One example of a tool to support improved teaching of statistics for climatology
F-SIAC workshop –August 2008
Regional workshop 4-week workshopHeld annually in Nairobi in KMD
Excellent cost-effective training resources
For ASARECA – biased towards agricultureAnd included one week on crop simulation modelsUsing ApSim as an example
Recap of some points
Coping and adapting to current climate variability is a first step
Much can be done with simple rainfall analysisAnd other tools can also be used
The crop simulation modelsCan then investigate impact of climate change scenariosAlso may be a useful project tool generally
Thank you
CAST adapted for climatology
CAST – text page (for climatic)
CAST – an interactive page
Other datasets show the concept is general
The Moodle menu for e-learning
Click here to go to teaching session
Information on participantsCaptured on 3 Jan, 4 months after the end of the training.
Slide from one of the sessions
Agricultural Production Systems Simulator (APSIM)
Simulates:yield of crops, pastures, trees, weeds key soil processes (water, N, P, carbon)surface residue dynamics & erosionrange of management options crop rotations + fallowing + mixturesshort or long term effects
BUT, not yet pests nor diseases
Point Source model
Climatic menu
Designed to allow tailored products
Marksim
Find the location, then simulate daily data
