Farming in Response

to the Weather:

A Guide for Extension

Sue Walker

Agrometeorology Professor, University of the Free StateDept Soil, Crop and Climate Sciences,

Bloemfontein, South Africa

Introduction

� Weather parameters affect crop production• Temperature affects plant growth

• Very high or very low give less growth according to crop species

• Water essential for growth & development• If low water available give low production

• Climate variability high in semi-arid tropics

Gl e n Longt e r m Annua l Ra i nf a l l

547

0

200

400

600

800

1000

1200

1920 1930 1940 1950 1960 1970 1980 1990 2000Time

Rai

nfal

l (m

m)

Introduction cont.

� Farmers’ decisions• When to plant?• What crop to plant?• Where & how much to plant?

� Use local indigenous knowledge

Introduction cont.

Need to answer following:

� Which on-farm routine operations are dependent on the weather?• e.g. hay making, irrigation, planting, spraying,

� What information can be provided to reduce risk or to assist in planning these decisions?• daily weather forecasts• weekly forecasts• seasonal outlooks

� Can one farm in response to weather?

Original “Response Farming” Concept

� Developed by Dr. J. Ian Stewart in 1980s

� WHARF (World Hunger Alleviation through Response Farming)

� Use interaction of rainfall and farming system to optimize crop production

� Need following:• start date and amount of rainfall• yield for corresponding amount of rain• use to construct a “rainfall flag”

“rainfall flag”

graph

� Y-axis is rainfall (mm)� X-axis is yield and rain

onset date

Example Davis CA from Stewart, 1988.

To show probabilities:

• Y-axis is rainfall• X-axis is specific groups of

ranges of onset dates

Example Davis CA from Stewart, 1988

Assumptions for response farming

� Assume • Early onset of rain means more rain will be received• Onset date is proportional to amount• Onset is related to production

� Can only be true if ‘end’ of rain season is stable each year but not strictly true everywhere

� Need to modify some definitions

� Expand response farming concept

From

data and knowledge

to useful information

From data

� Use data• Climate - need long-term daily rainfall• Soil – depth & type & water holding capacity• Crop – type & agronomic management variations• Socio-economic – farmers’ aims & markets etc

� Via calculation & analysis & manipulation� To useful information

• Rain onset date & seasonal amount• Potential crop yield for certain rainfall• Information applied according to farmers’ needs

From indigenous knowledge

All farmers have valuable information as inputs

� Local farmer information about practical farming systems• Local seeds & landraces characteristics & availability• Location of shallow &/ poor soil• Microclimate variations (e.g. wind, temperature variation)

• Pests & diseases occurrences• etc.

Integrate into

information sources

All role-players have valuable inputs

� Agrometeorologist - climate data analysis� Extension - experience in area� Local farmer - information about practical

farming systems

Use participatory needs assessmentVenda village meeting

For farmers’ study groups

Each bring information:� Farmers

• aim of farming • household food or for

market

• local knowledge• availability of resources

• stored seed

• manure / mulch

• Labour

� Extension • seed & inputs availability• market location• communication skills

� Agromet• long-term trends• Current season outlook• Monitor current season rain

Examples from farmers :

a) Resource status informationb) On-farm decision-making

i. Planting calendarii. Crops problem areasiii. Management options

� Collect data using participatory methods

a) Resource status information

Farmers access

resources

� soil types� access to water� transport� pests� co-op� markets Map of Hoxane

Irrigation Scheme

b) On-farm decision-making –

i. Timing of field operations

� soil tillage / preparation / planting / weeding

Time line of maize production

at Veerplaats

b) On-farm decision-making –

ii. According to current weather

� Which are most dependent on weather?• e.g.

• frost can destroy young sensitive plants• high temperature causes heat stress & wilting• rain soften soil crust for seedling emergence• heat stress reduce milk produced• vegetables need frequent rains• etc.

b) On-farm decision-making –

iii. Ranking of problems encountered

Matrix ranking -allow each farmer to vote for problems encountered with various crops

crops

b) On-farm decision-making –

iv. Farmers decision options

� Below normal rainfall:• Plant animal fodder crops• Less maize• More sorghum• Lower density• Plan to try adding water• Sell animals• Take animals to grazing

� Above normal rain: (good rains)

• plant earlier• Grow more vegetables• Grow more cash crops• Increase sharecropping• Watch for pests & diseases

(crops & stock)

• Winter breeding for sheep & goats

From participatory survey in Lesotho by Dr G Ziervogel

Steps for Agromet calculations

� Compile dataset• Daily rainfall amount• Crop yield

� Analyze data• Onset of rain• Length of season• Seasonal total rainfall

� Prepare discussion materials

Agromet discussion materials

i. Seasonal rainfall versus onset dates

0

100

200

300

400

500

600

700

800

900

110 120 130 140 150 160 170 180 190 200 210

Number of day(1-365) when season starts

Tot

al s

easo

nal r

ainf

all (

mm

)

Niamey, 1954-83

Agromet discussion materials

ii. Typical yield production function

02468

101214

0 1 2 3 4

Water Use

Tot

al D

ry M

atte

r

Agromet discussion materials

iii. Simulated yield from crop models

0 300 600 900 1200 1500 1800 2100 2400 2700 3000 3300 36000

0.2

0.4

0.6

0.8

1

Grain yield (kg/ha/year)

Pro

babi

lity

of n

on-e

xcee

danc

e

Full

3/4

1/2

1/4

Empty

Long term maize yields using the Putu crop model under conventional tillage on Glen/Bonheim ecotope, starting with 5 initial soil water content regimes planted in mid-December Climate data, with effective rainfall, from Glen College, 1922-2001.

Agromet discussion materials

iv. Rainfall probabilities for each siteCumulative Distribution Function of Rainfall

0

0.33

0.66

0.99

0 200 400 600 800 1000 1200 1400 1600

Rainfall (mm)

Pro

babi

lity

of n

on-e

xcee

danc

e

Pietermaritzburg

Bethlehem

Bloemfontein

Upington

Agromet

discussion materials

v. Seasonal rainfall forecasts for region

Develop decision tables

“What if” discussion with all parties concerned� Long-term graphs� 3-month seasonal rainfall outlook� Current rainfall situation� Discuss local available options and outcomes

� Integrate model & local information

Local options for decision tables

Farmers questions� Which crop?

� What area to plant?

� What plant density?

� What inputs?

Agromet model options

� Maize / sorghum / sunflower / beans

� Deep / shallow soil� Hi / lo potential soil

� Hi / medium / lo

� Manure / mulch / pest control

Example - tillage optionsConventional full tillage versus

in-field water harvesting

runoff areaCollection

& infiltration

in-field water harvesting on clay soil Conventional

Water harvesting

Example with tillage options

• Compare simulated maize yields for conventional tillage (CT) and in-field water harvesting (WH) for range of farmer options:

• When to plant? Nov / Dec / Jan

• How much seed to use? Low / medium / high plant populations

• What cultivar to use? short / medium / long growth period

• How much water to start? empty / half / full soil profile

Different Initial Soil Water

0.00

0.25

0.50

0.75

1.00

0 1000 2000 3000 4000 5000 6000

Yield (kg/ha)

Cum

ulat

ive

prob

abilit

y

EmptyHalfFull

Conventional tillage

0.00

0.25

0.50

0.75

1.00

0 1000 2000 3000 4000 5000 6000

Yield (kg/ha)

Cum

ulat

ive

prob

abilit

y

EmptyHalfFull

Water harvesting

For example, probability of 50% (i.e. half years) ofproducing less than 1.38, 2.23 and 2.90 t ha-1 for CT and less than 3.27, 3.52 and 3.63 t ha-1 for WH with empty, half and full initial soil water, respectively

Different Cultivars (time to maturity)

0.00

0.25

0.50

0.75

1.00

0 2000 4000 6000

Yield (kg/ha)

Cum

ulat

ive

prob

abilit

y

EarlyMediumLate

Conventional tillage

0.00

0.25

0.50

0.75

1.00

0 1000 2000 3000 4000 5000

EarlyMediumLate

Water harvesting

For example, probability of 50% producing less than 2.18, 2.17 and 2.15 t ha-1 for CT and less than 3.58, 3.50 and 3.34 t ha-1 for WH with cultivars of short, medium and long time to maturity

short

long

short

long

Different Plant Densities

0.00

0.25

0.50

0.75

1.00

0 1000 2000 3000 4000 5000 6000

Yield (kg/ha)

Cum

ulat

ive

prob

abilit

y

LowOptimumHigh

Conventional tillage

0.00

0.25

0.50

0.75

1.00

0 1000 2000 3000 4000 5000 6000

LowOptimumHigh

Water harvesting

For example, probability of 50% producingless than 1.80, 2.39 and 2.30 t ha-1 for CT andless than 2.01, 3.77 and 4.64 t ha-1 for WHwith low, optimum and high plant densities

Different Planting Dates

0.00

0.25

0.50

0.75

1.00

0 1000 2000 3000 4000 5000 6000

Yield (kg/ha)

Cum

ulat

ive

prob

abilit

y

NovDecJan

Conventional tillage

0.00

0.25

0.50

0.75

1.00

0 1000 2000 3000 4000 5000 6000

Yield (kg/ha)

Cum

ulat

ive

prob

abilit

y

NovDecJan

Water harvesting

For example, probability of 50% of producing less than 2.22, 2.49 and 1.80 t ha-1 for CT and less than 3.97, 4.00 and 2.45 t ha-1 for WHwith November, December and January sowing dates

Develop decision tables

farmeragromet

extension

Pre-season cropping decisions

Simulated yield a/c to management

Potential Crop Yield

Examples of options

Monitoring weather data

TransportLabourrequired

Monitoring weather data

Labouravailable

3-6month seasonal outlook

Commod-ity prices

Land preparation

Daily forecasts of rain & temp.

Post-harvest storage

Fair, dry weather to harvest

Late season

Weekly /dekadal forecast of rain & temperature

Availability of inputs

Cash for inputs

Mid-season

Long-term means & probabilities

MarketsSeed available

Pre-planting

AgrometExtensionFarmerTime in season

Conclusions

� Should be study group with farmers , extension& agrometeorologist

� Use local knowledge & model outputs to simulate potential variation according to management practices

� Farming more viable if done in response to long-term climate info and seasonal forecast together with current weather information

Publication

“Farming in Response to the Weather: A Guide for Extension”

by S Walker and H Pfeiffer

Chapters to include:1. Stepwise Data Analysis for Response Farming2. From ‘data’ and ‘knowledge’ to ‘information’3. Towards Use of Decision-making Tools

Lets help the farmers make a success

under variable weather conditions