A smallholder farmer level decision support system
for climate resilient farming practices improves
community level resilience to climate change. No3:
The smallholder farmer CRA decision support
system
Submitted by Erna Kruger (Director Mahlathini Development Foundation - MDF)
Ph: 0828732289, Email: info@mahlathini.orgWeb: www.mahlathini.org
Partners: Erna Kruger, Mazwi Dlamini, Samukelisiwe Mkhize, Temakholo Mathebula, Phumzile
Ngcobo, Betty Maimela, Sylvester Selala and Lulama Magenuka (MDF)
Palesa Motaung, Nonkanyiso Zondi, Sandile Madlala, Khetiwe Mthethwa, Andries Maponya,
Nozipho Zwane, Lungelo Buthelezi, and Zoli Gwala (students and interns)
Mr Lawrence Sisitka (Research Associate- Environmental Learning Research Centre- Rhodes
University)
Mr Nqe Dlamini (StratAct)
Mr Chris Stimie (Rural Integrated Engineering)
Mr Jon Mc Cosh, Ms Brigid Letty (Institute of Natural Resources)
Mr Hendrik Smith (CA coordinator for GrainSA)
Ms Sharon Pollard (AWARD)
Ms Lindelwa Ndaba (Lima RDF)
Ms Catherine van den Hoof (formerly of WITS Climate Facility, now the United Nations World
Food Programme)
Introduction
A current Water Research Commission adaptive research process entitled “Collaborative knowledge creation
and mediation strategies for the dissemination of Water and Soil Conservation practices and Climate Smart
Agriculture in smallholder farming systems”is exploring best practice options for climate resilient agriculture
for smallholders and evaluating the impact of implementation of a range of these practices on the resilience of
agriculture based livelihoods. Alongside this, a decision support methodology and system has been designed to
assist smallholders and the facilitators who support them to make informed and appropriate decisions about
choices of a ‘basket of options’ for implementation at a local level.
The research process is broadly divided into three elements for purposes of clarity, although all three elements
are tackled concurrently:
1. Community climate change adaptation process design
2. Climate resilient agricultural practices and
3. A decision support system.
In this article we focus on the design of the farmer level decision support system.
The smallholder climate change adaptation decision support process
The decision support process focusseson a bottom -up approach, where individual farmers in a locality make
decisions regarding the ‘basket’ of CSA/CRA approaches and practices most suited to their specific situation.
To do this in a way that also includes the concepts of social learning, innovation and agency the following
decision support concept has been developed.
The process is designed to also support and assist the facilitator in their decision making, in support of the
smallholder farmers; meaning that the facilitator accesses information such as the basic climate change
predictions for the area, theagroecological characteristics including rainfall, temperature, soil texture etc) and
an initial contextualised basket of CSA practices from which to negotiate prioritized practices with farmers.
Practices are thus chosen by both facilitators and farmers.
Figure 1: The smallholder CSA/CRS decision support model.
Situation and vulnerability assessments
The model for vulnerability assessments used in this process provides for a combination of socio-economic
(livelihood) and socio-ecological (access and utilization of natural capital) indicators, in a climate change
context (wellbeing, adaptive capacity and governance). This is a new process design, built from elements of
existing international best practice options.
The process consists of focus groups discussions, individual interviews (baselines) and household visits, or
walkabouts as we call them – as they include a broad and initial assessment of the “lay of the land”.
This information is pulled together into a database that has been put together to provide for a farmer
segmentation/ farmer typology approach. Farmer typologies allow for differentiation between different levels
of vulnerability in a community to target interventions/ practices more specifically.
PHYSICAL ENVIRONMENT: Climate and geographical
parameters; GPS coordinates, agroecological zones,
soil texture, slope and soil organic carbon content
PRACTICES: Database of CSA practices including; managing available
water, improving access to water, controlling soil movement, improving
soil health and fertility, crop management, integrated crop-livestock
management, veld management and veld rehabilitation
The three typologies developed within this process are shown in the figure below
Figure 2: Smallholder typology for a climate resilient farming decision support system.
A typical participant is thus:
These typologies are one of the input categories into the decision support system.
The web-based platform
The model is designed primarily as a participatory and facilitated process at community level. In support of this
process, a computer-based model can be used alongside this methodology to provide further information and
decisions support to the facilitator. It is also possible for a farmer to access this model independently to derive
an initial basket of CSA practice options for themselves.
The computer model information flow is designed as shown in the figure below –and follows the same basic
steps as the facilitated model shown in Figure 3 below.
Typology A (2,5 million)
•Female headed,
•Farm for food only,
•Very low incomes –mostly
unemployed,
•Access to small plots of land
(<0,1ha),
•No household level access to
water,
•Lower education levels (Primary
school)
•No access to formal markets,
•Belong to village savings and loan
associations and
•Engage in other livelihood
activities
Typology B (250 000)
•Male and female headed,
•Farm for food andsell surplus,
•Slightly higher incomes,
•Access to larger plots of land
(0,1-1ha)
•Some access to hh level
water,
•Somewhat higher education
levels (High school),
•No access to formal markets
and
•Belong to village savings and
loan associations
Typology C (10 000)
•Male headed, Ffarm mainly for
income,
•Much higher incomes from
employment in the household,
•Good access to water at
household and field level,
•Higher education levels (Matric
nad post scholl qualifications),
•Acess to formal markets.
•Belong to cooperatives or farm
individually
A 51 year old woman, who is
the head of her household, has
Grade 9-11 level of education, is
unemployed, has an average
monthly income of R2170,
engages in field cropping,
gardening and livestock
husbandry, has no access to
water in her household,
engages in local markets only
and belongs to a savings group
Figure 3: The computer-based model for the smallholder DSS.
In our case the set of criteria(proxies used asindicators for the complex reality) that helps to make informed
decisions on management practices are:
➢The current farming systems; gardening, field cropping, livestock production and natural resource
management (NRM) (including trees),
➢The physical environment: agroecological zone, soil texture, slope and organic soil carbon and
➢The socio-economic background of the farmer; demographic information (gender HH head, age,
dependency ratio), level of education, sources of income (unemployment vs. external employment,
own business, grants, farm, etc.), total income, access to services, infrastructure, technology
(Electricity, water (tap, borehole, rainwater harvesting, etc.), irrigation (buckets, standpipes, etc.),
fencing and farming tools (hand vs traction/other), social organisation, market access (formal vs.
informal), farm size andfarming purpose (food vs. selling).
Besides this, the resources and related management strategies as well as a list of practices need to be
provided as input to the system. All information, except the physical environment; i.e. climate, soil and
topography, and the resources and management strategies, are derived through the use of a range of
participatory processes. Data on the physical environmental conditions have been taken from datasets freely
available online. This information can however be customised by the DSS user, in case more appropriate
information is available for the specific farmer concerned.
For the Facilitator-Farmer DSS the resources and related management strategies are discussed and negotiated
in the participatory process. For the computer based or Individual Farmer DSS these are provided as an input
into the model using the following framework:
FARMING SYSTEMFARMER SOCIO-ECONOMIC
BACKGROUND
RESOURCES TO MANAGE
SUGGESTED PRACTICES
CONSTRAINED BY
TYPOLOGY, SYSTEM
AND ENVIRONMENT
RANKED PRACTICES
BASED ON FACILITATORRANKED PRACTICES
BASED ON FARMER
FARMER BASED
PRIORITIES
FACILITATOR
BASED PRIORITIES
PHYSICAL ENVIRONMENT
DSS PROCESS FLOW
Figure 4: Resources to manage and their associated management strategies.
Once all the information is inputted into the model an initial list of practices is suggested for each individual
farmer. The model has been tested and refined, through comparison of this computed based process with the
participatory process and assessing how closely these two processes are aligned.
Below is an example for 1 farmer in each of the three provinces where the model has been tested.
Table 1: Basket/list of practices recommended for version 2 of the DSS
Province
KZN
Limpopo
EC
Village
Ezibomvini
Sekororo
Mxumbu
Name and Surname
Phumelele Hlongwane
Chenne Mailula
Xolisa Dwane
Drip irrigation
0
0
0
Bucket drip kits
0
0
0
Furrows and ridges/ furrow irrigation
0
0
0
Greywater management
1
1
0
Shade cloth tunnels
1
1
0
Mulching
1
1
0
Improved organic matter (manure and crop
residues)
1
1
1
Diversion ditches
1
0
0
Grass water ways
0
0
0
Infiltration pits / banana circles
1
1
0
Zai pits
1
0
0
Rain water harvesting storage
1
1
1
Tied ridges
0
0
0
Half- moon basins
0
0
1
Small dams
0
0
0
Contours; ploughing and planting
1
0
0
Gabions
0
0
1
Stone bunds
0
0
0
Check dams
0
0
1
Cut off drains / swales
0
0
1
Terraces
0
0
0
Stone packs
1
0
0
Strip cropping
1
0
0
Pitting
1
1
0
Woodlots for soil reclamation
1
0
0
Targeted application of small quantities of
fertilizer, lime etc
1
0
0
Liquid manures
1
1
0
Woody hedgerows for browse, mulch, green
manure, soil conservation
1
0
0
Conservation Agriculture
1
0
0
Planting legumes, manure, green manures
1
0
0
Mixed cropping
1
0
0
Planting herbs and multifunctional plants
1
0
0
Agroforestry (trees + agriculture)
1
0
0
Trench beds/ eco circles
1
1
0
push-pull technology
1
0
0
Natural pest and disease control
1
0
0
Integrated weed management
1
1
1
Breeding improved varieties (early maturing,
drought tolerant, improved nutrient
utilization),
1
1
1
Seed production / saving / storing
1
1
1
Crop rotation
1
1
1
Stall feeding and haymaking
0
0
0
Creep feeding and supplementation
1
0
0
Rotational grazing
1
0
1
De-bushing and over sowing
1
0
1
Rangeland reinforcement
1
0
1
Bioturbation
1
1
1
Tower garden
1
1
0
Keyhole beds
1
1
0
No of practices recommended
35
16
14
For the KZN participant, this means that around 88% of the full list of practices have been recommended for
her. She has a wide range of recommendations being a farmer in Typology B (fewer restrictions) and engaging
in gardening, cropping and livestock production. Although this is quite high, it is understood thatthe farmer
level ranking is still to take place and these practices can then be prioritized and narrowed down further. For
the Limpopo and EC participants, around 1/3 of practices have been recommended in their basket of options.
Ranking can be undertaken first by the facilitator, or can be done directly by the farmer depending on the
circumstances. Below is the ranking exercise undertaken for Phumelele Hlongwane (Ezibomvini, KZN).The
practices shown in green are those that Phumelele are already implementing. This ranked list then provides
options for inclusion of further ideas and practices
Table 2: Ranking of CRA practices recommended for Phumelele Hlongwane
(KZN; Bergville)Phumelele Hlongwane: List of practices scored by facilitator
Practices
Field
cropping
Vegetable
gardening
Livestock
Natural
resources
and trees
Shade cloth tunnels
8
Mulching
9
Improved organic matter
11
11
11
Diversion ditches
9
9
9
Infiltration pits
10
Zai pits
10
10
RWH storage
9
9
9
9
Stone packs
9
9
9
Strip cropping
11
Pitting
11
11
11
Woodlots for soil reclamation
9
9
9
Targeted fertilizer application
8
Liquid manure
7
Woody hedge rows
10
10
10
Conservation agriculture
11
11
11
11
Planting legumes, manure, green manures
8
8
8
Mixed cropping
9
9
Planting herbs and multifunctional plants
9
9
Agroforestry (trees + agriculture)
11
11
11
11
Trench beds/ eco circles
9
push-pull technology
7
Natural pest and disease control
7
7
7
Integrated weed management
7
7
7
Breeding improved varieties (early maturing,
drought tolerant, improved nutrients),
7
7
7
7
Seed production / saving / storing
6
6
6
Crop rotation
9
9
Stall feeding and haymaking
Creep feeding and supplementation
7
Rotational grazing
9
De-bushing and over sowing
9
Rangeland reinforcement
9
Bioturbation
9
9
9
9
Tower garden
10
Keyhole beds
10
Below are a few indicative photographs of Phumelele’s CRA practices.
Above clockwise from top left: A view of Phumelele Hlongwane’s vegetable garden, a newly
constructed tower garden, trench beds planted to a mixture of vegetables in her shade cloth
tunnel, a plot of Dolichos in her CA field and a plot of summer cover crops- sunnhemp and millet.
Conclusion
The decision support system for climate resilient agriculture implementation by smallholder farmers is an
important new innovation in the field of community-based climate change adaptation and can be scaled up as
a framework in research, learning and implementation in this field.