1. Identifying the target environments
in breeding for rainfed ecosystems
B.P. Mallikarjuna Swamy
August 3, 2012
Rice Breeding Course 2011
2. Outline
• The concept of the target population of environments(TPE)
• Complexity of rainfed ecosystem
• Factors to be considered in delineating the TPE
• Methods of grouping locations into TPE
• Relationship between selection environments and TPE
• Setting goals and prioritizing traits
• Appropriate breeding strategy for a TPE
• A case study
3. Target population of environments(TPE)
Comstock (1977) defined the concept of a target
population of environments (TPE) associated with a
breeding program as the complete set of "types" of
environments in which cultivars can be grown within the
geographical area targeted by a breeding program.
4. Concept of the TPE
• TPE is the set of all environments, fields and seasons in which
an improved variety is targeted to perform well.
• TPE consists of – Not a single environment but a set of present
and future production environments.
• Environmental variability changes the performance of varieties.
• Breeders wish to develop cultivars that are superior to currently
used varieties in most years and on most farms within the TPE.
5. Concept of the TPE
• Seasonal variation can result in very different conditions
in the same field in different years
• Weather records for a certain location shows-
Favorable years - 10
Years with drought at seedling stage - 4
Years with drought at flowering - 6
Years with submergence - 0
Blast - 10
BPH - 0
BLB - 3
6. Concept of the TPE
• Development of appropriate research strategy and
prioritization of rice research activities requires -
1. In depth understanding of TPE
(More relevant with the Global climate
change scenario)
2. Farmers traditional knowledge and
farming practices
3. Socio-economic environments
4. Farmers perception of the new
technologies being introduced
8. Complexity of rainfed environments
• Large variations in rain fed ecosystems – rainfed upland,
rainfed lowland
• rainfall
• soil type
• topography
• occurrence of abiotic stresses- drought, submergence
• prevalent-insect-pests
• low input use
• quality preference
• socio-economic conditions of the farmers
9. Factors in delineating the TPE
• Environmental factors
Rainfall, topography, Soil type, Insect-pest
Many breeding programs considers this and have
separate breeding program
• Socioeconomic factors
Resource capacity, Input use availability
Very few breeding programs takes this into account
10. Methods of Grouping sites in TPE
• Crop Models
• Environmental parameters
• Currently used varieties
• Statistical analysis of multi environment data
- Principal component analysis
-Tests of fixed environmental effects
- Examination of correlations of cultivar means across sites
• Geographical information system (GIS)
11. Crop modeling
• The need to manage and predict crop's behavior over a
wide range of planting dates, geographies and situations
has become increasingly important.
• Use of crop simulation models incorporating local climatic
conditions with management operations increases our
ability to make more timely and educated decisions.
• The time has finally arrived in which crop modeling tools
are increasingly being deployed to help address questions
and problems on a larger, farm scale size. The full
potential and value of crop models have not yet been
realized in production agriculture.
12. Grouping sites in TPE: Crop Models
• Use of crop models to identify environments in terms of
water stress has been suggested
• Historical weather data required
• Information can be obtained from farmers if environmental
information is not available:
-Rapid rural appraisal
-Agro system analysis
13. Grouping sites in TPE: Environmental
parameters
• Sites with similar rainfall
pattern, soil types and
depths of standing
water accumulation
within each region may
be grouped together for
breeding purposes
14. Sites Target environment Soil type Field topography Drought Presently grown
varieties
Indira Gandhi Krishi Rainfed low land Clay, clay loam, low Bunded shallow Reproductive, early MTU 1010, IR 64,
Vishwavidyalaya ecosystem organic carbon lowland to mid stage Swarna, Mahamaya
(IGKV), lowland
NDUAT, Faizabad Rainfed shallow low Clay, clay loam, low Bunded Early season drought Sarjoo 52, Swarna,
land drought prone organic carbon and reproductive NDR 97, NDR 359,
and submergence stage Baranideep
prone
CRURRS, Hazaribag Rainfed shallow low acidic in nature, very Highly undulating Drought at all stages IR 36, IR 64, MTU
land and bunded poor in fertility, low in of growth 1010, Hazaridhan,
uplands available N and Sadabahar, Birsa 201,
organic carbon Some other land races
CRRI, Cuttack Costal region, rainfed sandy loam and clay Bunded low lands seedling stage, and Lalat, Swarna,
upland and lowland loam soil vegetative stage Varshadhan Naveen,
MTU 1010,
Khandagiri, Vandana,
TNAU, Coimbatore Irrigated Lowland Clay Bunded lowland Reproductive stage IR 64, Co-47
TNAU, Paramakudi Rainfed Upland Clay loam Bunded Reproductive stage PMK-3, ADT 38, Local
races
UAS, Bangalore Eastern Dry Zone red, loamy and light Unbunded Reproductive stage MTU 1001, IR 64,
drought Jyoti, MTU 1010, BPT
5204, Rasi,
BF, Hydrabad Low land Clay loam Bunded Reproductive stage Samba Masuri,
Swarna, MTU 1010
BAU, Ranchi Rainfed shallow low acidic in nature, very Highly undulating Drought at all stages Lalat, IR 36, IR 64,
land and bunded poor in fertility, low in of growth MTU 1010, Birsa 201,
uplands available N and Some other land races
organic carbon
15. Grouping sites in TPE: Currently used varieties
• Predominant farmers variety can also be used to
characterize the target environment
• Variety Brown Gora grown in Chhotonagpur plateau region
of eastern India can be classified as upland TPE
• Variety Swarna grown in much of south Asia can be
classified as shallow rainfed lowland
• Most reliable and simplest way to define TPE
• Adverse situations that these cultivars face in these
ecosystems, and performance under such situation has to
be taken into account
16. Sites categorization into favorable (F) or unfavorable
(U) environments based on the ranking of mean yield
in different years
SITE Site mean yield (t/ha) Mean Ranking Category
2003 2004 2005 2003 2004 2005
Santhapur 2.35 2.39 2.37 1 1 F
Semiligud
a 2 2.24 1.2 1.82 1 2 5 F
Faizabad 1.75 1.4 1.58 3 4 F
Jagdalpur 1.86 1.02 1.76 1.54 2 4 3 F
Ambikapur 1.36 0.62 1.77 1.25 6 7 2 F
Ranchi 1.59 0.54 0.79 0.97 5 8 7 U
Almora 0.94 0.76 0.9 0.86 7 6 6 U
Hazaribag 1.65 0.36 0.39 0.8 3 10 9 U
Rewa 1.64 0.46 0.26 0.78 4 9 10 U
Derol 0.59 0.77 0.68 8 5 U
Banswara 0.54 0.23 0.74 0.5 9 11 8 U
18. Grouping sites in TPE: Principal component analysis
• Pattern analysis to group sites on the basis of minimum G
x E within groups and maximum G x E among groups
• Used when TPE is very large and diverse
• Used when researchers do not have a good hypothesis
about the causes of G x E
• Use of probe and reference genotypes (known for their
adaptation in each environment) is very helpful
21. Grouping the sites in TPE: Correlation of cultivar means
• An easy and effective way of assessing the G x E across
environments within target region
• If correlations are above 0.3 for a single 3 replicate trial, G x E
is unlikely to be large
• If lower than 0.3:
- either G x E is large
or
- trial have high error and little genetic variations among
cultivars – F value for cultivars is not significant
23. Results of correlations
• Shallow sites tend to be correlated
• Deeper sites (Cuttack and Chinsurah are correlated)
• Correlation between shallow and deep sites is poor
24. Grouping sites in TPE: GIS
• Geographical information system (GIS) and crop
modeling is used to predict the performance of a variety
for a range of environmental conditions.
• These performance ranges can be combined with known
spatial variation of key environmental variables
contained in GIS to generate performance domain over
space.
• These tools also help to identify areas that are relatively
homogeneous in terms of key constraints to productivity
26. Relationship between the selection
environment (SE) and the TPE
• SE is the nursery in which breeder makes selections
• The chosen SE should predict performance in the TPE
• May need more than one SE if TPE is highly variable
• SE is not the same as TPE, so the relationship must be
monitored
27. Requirements for the SE
• The SE must predict performance in seasons and
locations within the TPE - genetic correlation (rG)
between TPE and SE must be high
• The SE must clearly and repeatably differentiate among
genotypes under evaluation.
• Heritability (H) for screening in the SE must be high
• The SE must permit relatively large numbers of
genotypes to be screened at low cost.
• SE must permit a high selection intensity (i) to be
achieved.
28. Examples of SE for specific requirements
• In addition to the yield potential, local quality
preferences, SE may require to screen for
– Insect-disease (Bacterial blight) resistance
– Submergence tolerance
– Drought tolerance
– Salinity tolerance
29. How to make SE more close to TPE
• Multi location testing under diverse set of
environments – Many national programs follow this
approach
• Testing under managed screens- Identify few
promising lines and carry multi location testing with
them
30. Setting goals and prioritizing traits
• Determining farmers preferences:
- Focus group discussions. Farmers are asked about
positive and negative features of present cultivars
- Preference analysis. Farmers are asked to rate
experimental lines in a trial
- In some cases, future farmer preferences must be
predicted (i.e., they may not be aware of new
options/technologies)
31. Main objectives of a breeding program
• Generally, to develop a cultivar that is superior to
farmers’ varieties in a particular target population of
environments
• Specific objectives
– Replace a specific cultivar
– Develop a new product
– Change a single trait
32. Breeding goals: specific traits and strategies
• Deficiencies of currently grown varieties
• List of required traits with parents that are sources of
these traits
• Strategy for generating populations and selection for
the desired traits
33. Breeding goals: specific traits and strategies
• A high-quality locally-preferred variety should be used as
a parent in most crosses. Because it can be difficult to
recover quality characteristics in a single cross, the high-
quality parent may be used as the recurrent parent in
generating a BC-derived population
• Quality parameters should be the focus of early-
generation selection, because they are highly heritable,
whereas yield is not.
• The program should be structured to generate a large
population of breeding lines with acceptable quality,
which can then be evaluated for yield under farmer
management.
34. Appropriate breeding strategy: Broad
adaptations vs. specific adaptation
• Irrigated ecosystem is more uniform as compared to
the rainfed ecosystem
• Broad adaptation has been very successful
• Rainfed ecosystem is highly diverse
• Broad adaptation through reduced G x E interaction
effect not clearly visible in rainfed ecosystem?
35. Appropriate breeding strategy
• Among G, G x E and E variance – G is comparatively high
in irrigated situation but G x E is high in rainfed ecosystem.
• G x E interactions have been earlier considered a
hindrance to the crop improvement.
• Nevertheless, G x E offer opportunity in selection and
adoption of genotypes showing positive interactions with
the location*.
• However, if G x E is small, possible advantage of breeding
for specific adaptation is reduced.
36. TPE classification: A case study from Thailand
• A water balance model was used to estimate the level of
standing water in paddy using weekly rainfall data from 1987
to 2001 for Nong Khai and Nakhon Ratchasima provinces of
Thailand
• Time and duration of standing water in the paddy
(Boonrat et al. 2006)
37. A successful breeder must:
• Be in close touch with the farmers
• Know the constraints in the target environment
• Know the market requirements
• Know the plant traits considered important by the
farmers
• Imply all these information in his breeding program
38. A case study from Thailand
Parameter Nong Khai Nakhon Ratchasima
Total Rainfall 2000 mm 1000 mm
Growing season Earlier Later (High chance of
beginning early drought)
Rainfall withdrawl 8-14 Oct. 8-14 Oct.
Chances of late High High
drought
Varietal requirement Late season drought Cultivars with tolerance
tolerant cultivars to drought at early as
well as at late season
39. A superior cultivar is one that:
• Will be grown by the farmer because it performs
better (or obtains a better price) than the existing
cultivar
• Under management practices currently used or
available to the farmer
Economic benefit
40. A Case study from Thailand: Village level
surveys
• Earlier, the target domain was classified using G x E
interaction and cluster analysis of multi location trials
• Because of year to year variation, the TPE changed
from year to year.
41. A Case study from Thailand: Village level
surveys
• Surveys at village and house hold level have
been used to define the TPE
• Hydrology of rice paddies even at local level
in farmers fields is utilized.
• Four terrace paddy levels are identified-
Upper – drought prone
Middle - drought prone
Middle - Favorable
Lower - Flooded
