Economic Analysis of Breeding for Improved Cold Tolerance in Rice in Australia

 

Rajinder P. Singh^A, John P. Brennan^B, Tim Farrell^A, Robert Williams^A,  Russell Reinke^A, Laurie Lewin^A and John Mullen^C*

 

A. NSW Agriculture, Yanco Agricultural Institute, Yanco.

B. NSW Agriculture, Wagga Wagga Agricultural Institute, Wagga Wagga.

C. NSW Agriculture, Orange

---

Abstract

The occurrence of low night temperatures during reproductive development is one of the factors most limiting rice yields in southern Australia. Yield losses due to cold temperature are the result of incomplete pollen formation and subsequent floret sterility. Researchers have found that in 75% of years, rice farmers suffer losses between 0.5 and 2.5 t/ha. Research is being undertaken to identify genetic materials that are cold tolerant under the local weather conditions and by using those genotypes as parent material, develop cold tolerance varieties of rice. A yield simulation model was used to measure reduction in losses due to cold at different minimum threshold temperatures, while the SAMBOY-Rice economic model was used to measure the costs and returns of a breeding program for cold tolerance. The results of the economic analysis reveal that incorporating selection for cold tolerance into the breeding program would lead to significant increase in financial benefits through a reduction in losses due to cold and an increase in yield from the better use of nitrogen by the cold tolerant varieties. The returns to investment in the change to the Australian rice breeding program are estimated to be high.

Key words: Rice, cold, yield loss, breeding

1. Introduction

The rice industry is a major contributor to Australia's agricultural production. Although the Riverina region in the South Eastern Australia is among the highest yielding rice-growing regions in the world, rice production is subject to climatic risk. The occurrence of low night temperatures during reproductive development in rice is one of the principal yield-limiting factors of rice growing in the region (L. Lewin, pers. comm.). Yield losses due to low temperatures are the result of incomplete pollen formation and subsequent floret sterility. However, it is difficult to identify the extent of the losses due to cold temperature from the existing information available on crop yield in different years.

Rice physiologists at Yanco Agricultural Institute have tested the adaptability of some overseas cold tolerant rice varieties under the local agro-climatic conditions and have identified some genotypes that have consistently performed better than the local Australian cultivars in withstanding low temperatures during the reproductive stage. The results of the experiments revealed that low temperatures lowered the harvest index[1] of the overseas varieties by an average of 20% compared to 50% for the typical Australian cultivars (Farrell and Williams, 2001). But the yield and quality of those varieties are low compared to the local commercial rice varieties.

Therefore, introducing these varieties to Australia would come at a considerable cost in terms of yield reduction in the absence of cold, and would result in lower quality. Alternatively, incorporating that tolerance through plant breeding would provide the benefits of the tolerance without detrimental yield effects, but would involve some costs and considerable time lags. This study aims to consider that issue. The objective in this paper is to assess whether it is economic to incorporate increased cold tolerance into new varieties by modifying the current rice breeding program.

2. Cold Damage in Rice

In assessing the economics of cold tolerance, the first step is to quantify the current levels of productivity losses due to cold and the potential losses with different levels of tolerance. Research at the Yanco Agricultural Institute, funded by the Cooperative Research Centre for Sustainable Rice Production (CRC Rice) has provided a basis for estimating losses from cold demage.

For most of the rice varieties grown in Australia, the critical temperature for cold damage to occur during the reproductive stage is 18° C (R. Williams, pers. comm.). Analysis of the information on long-term weather trends show that from the end of January to early February the average minimum temperature is 17° C (Russell Reinke, pers. Comm.). Thus in most years some degree of crop damage occurs.

Plant physiologists at the Yanco Agricultural Institute developed a yield simulation model for Australian rice varieties, incorporating weather data from 1955 to 1999 to estimate the extent of yield losses due to cold during this period. Parameters for the model were developed for the variety Amaroo, using the yield and meteorological data from 1987 to 1999. Amaroo is a medium-grain, long duration variety of good grain quality with a high yield potential, and has been widely grown across the rice belt since its release in 1987 (R. Williams, pers. comm.).

To measure the yield potential and yield losses due to cold damage, the following nine parameters were considered (Farrell 2001): Time of planting, time of physiological maturity, efficiency of conversion of radiation into yield, sensitivity to low temperature, start of cold sensitivity stage, threshold minimum temperature for cold damage, daily solar radiation data, and daily minimum temperature. Productivity losses were calculated as the difference between simulated potential yield with no cold losses and simulated yield given current levels of tolerance.

Using this model, we measured the extent of productivity losses due to cold and the probability of occurrence of such weather conditions during the critical microspore development stage over the period 1955 to 1999. We also estimated the probabilities of yield loss that would occur if rice were to become more tolerant by 1 to 3 degrees C because of the genetic material introduced from cold tolerant varieties. Between 1955 to 1999 with the existing levels of threshold temperature, the rice industry suffered a maximum loss of 2.5 t/ha once in ten years (Table 1), whereas in 22% of the years the industry suffered yield losses of less than 0.5 t/ha, and in a further 29% of years the losses were between 0.5 and 1.0 t/ha. The analysis further revealed that at the current threshold temperature, the probability of a cold year was 0.73. That probability declined to 0.64, 0.40 and 0.27 with the reduction of the threshold temperature by 1°, 2° and 3° C respectively (Table 1).

Table 1: Probability of Rice Yield Losses Due to Cold at Different Threshold Temperatures

Yield Losses	Probability of yield losses at threshold temperatures
(t/ha)	Current	Current-1oC	Current-2oC	Current-3oC
More than 2.5	0.00	0.00	0.00	0.00
Between 2.0 - 2.5	0.11	0.00	0.00	0.00
Between 1.5 - 2.0	0.02	0.02	0.00	0.00
Between 1.0 - 1.5	0.09	0.07	0.00	0.00
Between 0.5 - 1.0	0.29	0.22	0.09	0.02
Less than 0.50	0.22	0.33	0.31	0.24
Total cold years	33	29	18	12
Total years	45	45	45	45
Probability of cold year	0.73	0.64	0.40	0.27

Source: Estimates from data provided by R. Williams

The economic value of these losses is determined by the area, yield and price of rice. Over the five years to 2001, the average area under rice was 154,600 ha, the average price of rice was $208 per tonne, and the average yield was 9.36 t/ha.

The results presented in Table 2 show that on average farmers suffered productivity losses from cold damage of 0.72 t /ha. At recent average price of $208 per tonne, this is valued at $150/hectare per year. Given the total area under rice, on average the rice industry suffers a loss of $ 23.2 million per year due to cold damage, with existing varieties. 

Table 2: Estimated Productivity Losses Due to Cold Damage, from 1955 to 2000

Yield losses	Yield losses at threshold temperatures
	Current	Current-1oC	Current-2oC	Current-3oC
Average losses (t/ha/yr)	0.72	0.34	0.14	0.04
Value of average losses ($/ha/yr)	150	71	29	8
Value of total losses ($million/yr)	23.2	10.9	4.5	1.3
Gain from current threshold
- per ha ($/ha)	-	79	121	142
- Total ($ million)	-	12.3	18.7	21.9

Source: Estimates from data provided by R. Williams

Further, the benefits of improving cold tolerance in the local commercial varieties were estimated by reducing the threshold temperature by 1^o C steps to 3^o C.  The results (Table 2) show that cold tolerant varieties would lead to a reduction in losses from $150/ha to $71, $29, and $8 per hectare per year at 1°, 2° and 3° C lower threshold temperature, respectively. In other words, a cold tolerant variety with 1° C lower threshold temperature would lead to a gain in productivity of $79 per ha, whereas new cold tolerant varieties with 2° and 3° C lower thresholds would lead to $121 and  $142 gains in productivity per ha, respectively.

3. Options for Managing Cold Tolerance in Rice

Different management strategies are being followed to minimise the losses due to cold. The most common and effective practice to minimise these losses is increasing the depth of standing water once the rice has reached the panicle initiation stage (Williams and Angus, 1994). Deep water (between 20–25cm) is usually 6° to 7° C warmer than the night temperatures, and helps to protect rice from damage at the cold sensitive early pollen microspore stage. Under uncertain weather conditions, another strategy to protect the crop from cold damage is to apply nitrogen at less than the optimum level. Although this helps to protect the crop from cold because the plant is smaller and more likely to be protected by water, low nitrogen also leads to low yield. There is little opportunity to escape cold damage by other means such as adjusting sowing time. Therefore, there is a need to incorporate cold tolerance into local Australian rice varieties of all maturity groups.

However, progress in plant breeding depends on the number of lines and the probabilities of identifying a line with superior yield (or other characteristics) compared to the current varieties. Any reduction in numbers caused by selecting for an additional characteristic will reduce the probability of finding such a superior line for release, unless additional resources are made available to expand the number of lines in the earlier stages of the program. As the lines are brought through the stages of the program, any reduction in numbers at one stage of the program impacts on the likely numbers and the characteristics of the materials flowing through to the subsequent stages of the program in the following years. Thus, the impact of adding another selection character will be felt through the impact on the flow of new varieties released over a number of years.

The impact of slowing the rate of varietal yield improvement in non-cold years is illustrated in Figure 1. Without the additional selection for cold tolerance, the value of rice production would progress along line A. With the additional selection, the slower rate of progress in yield potential (line B) is followed. The cost to the industry in terms of the loss of progress in yield potential is the difference between the two lines. 

Figure 1: Impact of Slowing in the Rate of Varietal Improvement

 

 

The benefits of selection for increased cold tolerance in rice need to be compared with those costs. As selection is made for increased cold tolerance, costs from susceptibility to cold fall because of the saving in yield losses in cold years. The key question addressed in this analysis is the extent to which the benefits of improved cold tolerance are greater than the losses imposed by the reduced selection for, and thus slower progress with, yield potential.

4. Economics of Breeding Cold Tolerant Varieties

In this study, we are analysing a rice breeding program that aims to develop cold tolerant varieties by using the cold tolerance genes of the overseas varieties with well-adapted local varieties to develop new varieties that combine cold tolerance with high yield and grain quality.

The economic analysis was undertaken to measure the returns from breeding for cold tolerant rice varieties, and to provide a basis for establishing priorities for research in rice. Benefit-cost analyses were used to compare the potential benefits arising from new cold tolerant varieties with the costs of the research to develop such varieties. The criteria used were the Net Present Value of the project (NPV), the Benefit-Cost Ratio (BCR), and the Internal Rate of Return (IRR).

To undertake benefit-cost analyses of the project, the benefits estimated above were scaled up to reflect the rate and extent of adoption of the technology, lags in the development and adoption of the technology. The impact of new technologies were spread over many years. Since it takes a long time to develop a new variety, in this analysis, the period over which benefits and costs of the proposal were accounted for was 30 years; that is, from 2002 to 2031. After 2031 it is anticipated that the technology used to develop such varieties would be replaced by new technology from future research and development.

In discounting, all benefits and costs were expressed in 2002 dollars, using a real discount rate of 7%. We assumed that Australia is a price taker in the world rice market and hence changes in production in Australia as a result of this new technology will have no effect on world price.

The analysis of the breeding for improved cold tolerance was made using the SAMBOY-Rice model (Brennan, Singh and Lewin, 1997). The model follows the processes of crossing and selecting within a breeding program, with the costs being estimated for each stage of the process. When the full cycle of selection has been completed, usually taking nine years from crossing to potential release of a new variety, the expected genetic gain from the selection pressure applied to the different characteristics is estimated in the model. As a result, the costs and expected gains from the new varieties developed can be estimated from the model, and when combined with the expected adoption of new varieties by farmers can determine the aggregate benefits from the breeding program in relation to its costs. More significantly in this case, the model allows analysis of changes in selection methods within the breeding program, so that the costs of a change in selection procedures can be related to the expected benefits from such a change.

The present study has focused on measuring costs and returns of developing cold tolerant medium grain varieties of rice. About 70% of the total rice area is under medium grain varieties. The original SAMBOY-Rice model was modified to focus only on medium-grain rice, and the costs were adjusted by the Consumer Price Index to convert them from 1994 to 2002 values. Specific data used in the analysis of medium-grain rice are given in Table 3. These parameter values were used to measure the value of the breeding programs. The model was initially run using these industry parameters for the current selection program, without any specific selection for cold tolerance to work out the cost of the breeding program for yield improvement only (the “baseline” run).

Table 3: Data and Assumptions Used in Analysis of Breeding Programs

Parameter Value Average area under medium grain rice (ha) 108400 Average yield of medium grain rice (t/ha) 9.36 Average production of medium grain rice (million t) 1.02 Average price of rice ($/t) 208 Accounting period (years) 30 Discount rate (%) 7

A key issue in the analysis is the time lags involved. Given that the program is on-going, the issue being analysed here is whether or not to change the crossing that takes place in the breeding program in 2002 and thereafter. Existing material in the program would not be affected by that change, only the material that results from the 2002 crosses. In a conventional program, a complete selection cycle takes nine years (including the year of crossing), so that the best line from the progeny of the cross carried out in 2002 would be identified in 2010. Given the selection program in place for the “base” program, expected yield progress from the program is estimated at 0.83% per year. On the basis that varieties are only released when there is a 5% yield advantage over existing varieties, then 6 (= 5.0/0.83) complete cycles are needed before a new variety is released. In that case, the first new variety resulting from crosses made in 2002 (and beyond) would be released in 2015. Subsequently, varieties with an additional 5% yield advantage would be released every 6 years.

Given the current area and yield of medium-grain rice varieties (Table 3), the value of a 5% productivity gain from the breeding program is $98 per hectare.

The pattern of adoption of the improved varieties adapted from Singh and Brennan (1997). In adapting that analysis to this study of medium-grain varieties, adoption was assumed to rise steadily to a peak of 40% of the total rice area in the sixth year after release of the first variety, and the cumulative area to all subsequent varieties rises to 60% of the total area in the tenth year after release (Table 4). In all subsequent years, adoption of the improved varieties remained at 60% of the total area.

 

Table 4: Adoption of Conventional and Cold Tolerant Rice Varieties

	Conventional Program Adoption %					Cold Tolerance Program Adoption %
Year	1	2	3	Total		1	2	3	Total
2002				0%					0%
2003				0%					0%
2004				0%					0%
2005				0%					0%
2006				0%					0%
2007				0%					0%
2008				0%					0%
2009				0%					0%
2010				0%					0%
2011				0%					0%
2012				0%					0%
2013				0%					0%
2014				0%					0%
2015	5%			5%					0%
2016	10%			10%					0%
2017	25%			25%					0%
2018	35%			35%		5%			5%
2019	35%			35%		10%			10%
2020	40%			40%		25%			25%
2021	40%	5%		45%		35%			35%
2022	35%	10%		45%		35%			35%
2023	30%	25%		55%		40%			40%
2024	25%	35%		60%		40%	5%		45%
2025	25%	35%		60%		35%	10%		45%
2026	20%	40%		60%		30%	25%		55%
2027	15%	40%	5%	60%		25%	35%		60%
2028	15%	35%	10%	60%		25%	35%		60%
2029	5%	30%	25%	60%		20%	40%		60%
2030	0%	25%	35%	60%		15%	40%	5%	60%
2031	0%	25%	35%	60%		15%	35%	10%	60%

The results for the base run of the model are shown in Table 5. The Net Present Value of a conventional breeding program over the next 30 years is $25.6 million, the benefit-cost ratio 11.3, and the internal rate of return 25.6%.

Table 5: Economic Analysis of Conventional Breeding Program

Measure	Results
Total costs of each cycle ($’000)	187
Average annual yield increase from selection (%)	0.83%
Present value of benefits ($ million)	28.1
Present value of costs ($ million)	2.5
Net Present Value ($ million)	25.6
Benefit-cost ratio	11.3
IRR (%)	25.6%

The model was then run with selection for cold tolerance incorporated into the program, to assess the costs and benefits of breeding for yield improvement and selection for cold tolerance. In defining the model with selection for cold tolerance incorporated, the following assumptions were made:

• Testing for cold tolerance is undertaken in a glasshouse at F₃ stage
• The best 50% of lines for cold tolerance were selected following that evaluation
• The cost of testing 3,000 lines per year for cold tolerance is $61,000
• As 50% of lines were discarded (because of cold tolerance) after F₃, fewer lines progressed through the later stages of the program
• As a result, it takes longer to develop a variety with the same yield advantage (5%) in non-cold years as in the baseline program.
• This selection program for cold tolerance would lead to 1° C reduction[2] in the minimum threshold temperature.
• From the results of the yield simulation model, the potential benefits from developing a variety that could withstand cold temperature 1° C below the current minimum threshold temperature for the existing varieties (Table 3) are $79 per ha per year.
• Many rice farmers apply less than optimal levels of nitrogen to minimise cold damage. The new cold tolerant varieties would help growers to increase the use of nitrogen to optimum levels which would further increase crop yields by a further 5% (L. Lewin, pers. comm.). Benefits from the increase in yield from the additional use of 25 kg / ha of nitrogen are estimated to be $76 per hectare (Table 6).

Table 6: Productivity Gains from Additional Use of Nitrogen in Cold Tolerant Rice Varieties

Measure	Value
Increase in use of nitrogen (kg/ha)	25
Cost of nitrogen used ($/t)	$22
5 % increase in yield (t/ha)	0.47
Value of increased yield ($/ha)	$98
Net value of increased benefits from additional nitrogen use ($/ha)	$76

As with the conventional program, a complete selection cycle takes nine years, so that the best line from the progeny of the cross carried out in 2002 would be identified in 2010. On the basis of the selection program in place for the program incorporating selection for cold tolerance, expected yield progress is estimated at 0.56% per year. If a new variety is released when there is a 5% yield advantage over existing varieties, then 9 (= 5.0/0.56) cycles are needed before a new variety is released with the improved cold tolerance. In that case, the first new variety resulting from crosses made in 2002 (and beyond) would be released in 2018. Subsequently, varieties with an additional 5% yield advantage and that same level of cold tolerance would be released every 6 years.

The total annual benefits from reduced cold losses for medium grain rice depend on adoption of the new cold tolerant variety. As low temperature prior to flowering is one of the most serious issues in sustainable growth of rice, adoption of the new cold-tolerant rice varieties would be faster and higher compared to the non cold-tolerant varieties of rice. We assume initially that a cold tolerant variety's adoption pattern is the same as for a conventional variety. (see Table 4 above). 

The results for the run of the model incorporating cold tolerance are shown in Table 7. The Net Present Value of a cold tolerance breeding program over the next 30 years is $34.6 million, the benefit-cost ratio 14.5, and the internal rate of return 25.9%.

Table 7: Economic analysis of breeding program with selection for cold tolerance

Measure Results Total costs of each cycle ($’000) 193 Average annual yield increase from selection (%) 0.56% Present value of benefits ($ million) 37.2 Present value of costs ($ million) 2.6 Net Present Value ($ million) 34.6 Benefit-cost ratio 14.5 IRR (%) 25.9%

 

 

5. Analysis of Breeding Cold Tolerant Rather than Conventional Varieties

These results were compared to those of the “baseline” run to find out the additional costs and benefits of the breeding program with selection for cold tolerance. Using the SAMBOY-Rice model, the costs involved in developing a medium grain variety both for existing breeding program and the breeding program with selection of cold tolerance have been estimated. The annual costs are $6,414 higher where there is selection for cold tolerance.

The stream of benefits and costs over time from the two programs is illustrated in Table 8. The benefits from the conventional program begin sooner, and the benefits from yield improvement are larger than for the cold tolerance program. However, the additional benefits from the reduction of losses due to cold and productivity gains from additional use of nitrogen more than compensate for the lower and slower rate of yield improvement, and provide almost half of the total benefits from the development of cold tolerant varieties. While in the early years of adoption there is a net reduction in benefits to growers from the cold tolerance program, net benefits increase after the fifth year, and lead to substantial increases in benefits by 2031.

Table 8: Flow of Benefits from Alternative Breeding Options

	Conventional		Cold Tolerance Program
	Benefits:		Benefits from:
Year	% Adoption	Yield increase ($,000)	% Adoption	Yield increase ($,000)	Cold tolerance ($,000)	Nitrogen     use ($,000)	Total ($,000)
2002	0%	0	0%	0	0	0	0
2003	0%	0	0%	0	0	0	0
2004	0%	0	0%	0	0	0	0
2005	0%	0	0%	0	0	0	0
2006	0%	0	0%	0	0	0	0
2007	0%	0	0%	0	0	0	0
2008	0%	0	0%	0	0	0	0
2009	0%	0	0%	0	0	0	0
2010	0%	0	0%	0	0	0	0
2011	0%	0	0%	0	0	0	0
2012	0%	0	0%	0	0	0	0
2013	0%	0	0%	0	0	0	0
2014	0%	0	0%	0	0	0	0
2015	5%	528	0%	0	0	0	0
2016	10%	1056	0%	0	0	0	0
2017	25%	2639	0%	0	0	0	0
2018	35%	3694	5%	528	429	411	1367
2019	35%	3694	10%	1056	857	821	2734
2020	40%	4222	25%	2639	2143	2054	6835
2021	45%	5278	35%	3694	3000	2875	9569
2022	45%	5805	35%	3694	3000	2875	9569
2023	55%	8444	40%	4222	3428	3286	10936
2024	60%	10027	45%	5278	3857	3697	12831
2025	60%	10027	45%	5805	3857	3697	13358
2026	60%	10555	55%	8444	4714	4518	17676
2027	60%	11611	60%	10027	5142	4929	20098
2028	60%	12138	60%	10027	5142	4929	20098
2029	60%	14777	60%	10555	5142	4929	20626
2030	60%	16360	60%	11611	5142	4929	21681
2031	60%	16360	60%	12138	5142	4929	22209

The results of the analysis of the two breeding options are shown in Tables 5 and 7, respectively. Where there is no selection for cold tolerance (table 5), the Net present Value of the program is $25.6 million, while where selection for cold tolerance is incorporated (table 7), the Net Present Value is $34.6 million. The benefit-cost ratio and the internal rate of return are both also higher for the program incorporating selection for cold tolerance. Therefore, the results indicate that the economic returns from the Australian rice breeding program will be increased if selection for cold tolerance is incorporated.

Therefore, even at conservative assumptions about the likely adoption of cold-tolerant varieties, the project has a very high economic return. The incorporation of selection for cold tolerance, which can now be undertaken as a result of the availability of the cold tolerant genotypes, is a profitable change for the breeding program. Selection for cold tolerance in the rice breeding programs would increase the net returns to the breeding program, and lead to substantial benefits to the industry.

6. Conclusions

Low temperature prior to flowering is one of the most serious issues for sustainable growth of the rice industry. To protect the rice crop from cold damage, different farmers use different management strategies like applying N at less than optimal level, increasing depth of standing water at PI stage, and early sowing of rice. In 75% of the years, rice farmers suffer losses ranging from 0.5 to 2.5 t/ha. Overseas genotypes have been tested under local weather conditions to identify varieties that are more cold tolerant than the local varieties. Using those genotypes as parent material, initially the rice breeders would develop cold-tolerant varieties of medium grain rice. The new cold-tolerant varieties would not only lead to reduction in yield losses due to cold but would also help to increase yield from better use of nitrogen.

A simulation yield model was used to measure reduction in losses due to cold at different minimum threshold temperatures, while the SAMBOY-Rice economic model was used to measure the costs and returns of breeding programs both with and without the selection for cold program. The results of the economic analysis show that a 1° C increase in cold tolerance expected from incorporating the parental material with improved cold tolerance into the rice breeding program leads to a significant increase in financial benefits through reduction in losses due to cold, and an increase in yield from the better use of nitrogen.  If the breeders were able to develop varieties that could withstand cold conditions 2° to 3° C below the current threshold, the returns to investment on the breeding program could be expected to be even higher.

In addition, if adoption of the cold tolerant varieties by farmers was faster and/or higher than for conventional varieties, the benefits of the change to selection for cold tolerance would be even greater.

Thus, the incorporation of selection for cold tolerance into the Australian rice breeding program is likely to lead to higher returns on the funds invested in rice breeding.

References

Brennan, J.P., Singh, I.P. and Lewin, L.G. (1997), Maximising the Future Pay-Off from Rice Breeding, Project DAN 108A, A Final Report prepared for the Rural Industries Research and Development Corporation

Farrell, T., Williams, R. and Fukai, S. (2001), "Cost of Low Temperature to the NSW Rice", Contributed paper presented at the 10^th Australian Agronomy Conference, Hobart.

Singh, I.P. and Brennan, J.P. (1997), “Analysing adoption and replacement of rice varieties in NSW”, Contributed paper presented at the 41st Annual Conference of the Australian Agricultural and Resource Economics Society, Gold Coast.

Williams, R. and Angus, J. (1994), "Deep floodwater protects high- nitrogen rice crops from low-temperature damage", Australian Journal of Experimental Agriculture, 34, 927-33.