0321 The System of Rice Intensification (SRI) : What It Is, and How/Why We Think It Works


Published on

Presenter: Norman Uphoff

Published in: Technology, Business
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide
  • Prepared with information available as of February 1, 2003. These slides can be used or adapted, even translated, however SRI colleagues would be useful for explaining this methodology to others.
  • Picture provided by Gamini Batuwitage, Sri Lanka, of field that yielded 17 t/ha in 2000.
  • The "economist's $100 bill" refers to the joke about an economist and his friend who were walking together down the street one day when the friend saw a $100 bill on the sidewalk. Thinking that his friend, being concerned with money, would surely pick the bill up, he did not reach down himself. But the economist walked right by. The friend asked, didn't you see that $100 bill on the sidewalk? Why didn't you pick it up? The economist replied,It wasn't a real $100 bill. If it had been genuine, since people are rational, someone would have picked it up by now, so I am sure that it was a counterfeit, and I didn't want to waste any effort on it. Agronomists have regarded SRI with similar skepticism, dismissing it by saying if it were indeed as good as reported, it should have been discovered previously, given the many millions of farmers and thousands of scientists who have worked with rice. So, therefore, SRI must not be genuine. SRI contradicts a number of key concepts held by agronomists and economists, giving them reasons to reject it, without giving it an empirical evaluation. However, the evidence in support of SRI is mounting year by year, month by month.
  • Average yields where farmers have learned SRI methods, understand them and use them, are about 8 t/ha. In some countries, the average is not yet at that level, but given experience in Madagascar and Sri Lanka, we feel confident that 8 t/ha is a reasonable average to expect with SRI. Maximum yields reported are very controversial. We report data as accurately and truthfully as we can. Farmers have had harvests -- some whole-field, some sampled -- calculated to be 15-20 t/ha, so we report what we think is correct. Over time this will be substantiated by other or not. Water requirement reductions of 40-60% are often reported. That productivity for all four factors of production can increase at the same time goes against the conventional idea of necessary tradeoffs in factor productivity. We have often seen across-the-board productivity improvements, which are more important than yield. Farmers and countries get richer by raising productivity, not by attaining highest yield (because one has to consider the cost of attaining this). Costs of production have been reported to be reduced by 10-50%, depending on how the cost of labor is figured. Because no purchase of external inputs is necessary, cash costs of production invariably go down with SRI. Whether or not labor costs are reduced depends on various factors.
  • Dr. Janaiah visited Sri Lanka the last week of October, 2002, and talked with 30 farmers in four villages who had been practicing SRI and who could give him detailed data. He had previously done such an evaluation for IRRI of the costs and benefits of adopting hybrid rice, having been on the IRRI staff in Los Banos from 1999 to 2002. He found SRI to be a much more profitable innovation for rice production than adoption of hybrids. We have found that SRI methods give the highest yields with hybrid varieties so there is not necessary contradiction or competition between the two. The SRI results reported from the Philippines, by the Agricultural Training Institute of the Department of Agriculture, from trials with three varieties at its Cotobato center in Mindanao (slide 20), calculated that the cost of production per hectare was 25,000 pesos, while the value of the rice yield with SRI was 96,000 pesos, a return of almost four times. Thus there are other evaluations of net profit from SRI that are even more favorable than Janaiah's calculation.
  • SRI methods have given improved yield and factor productivity with all varieties used so far, though some perform better than others (responding to the different growing environment with more tillering and root growth). Traditional varieties can increase from 1-3 t/ha to 4-10 t/ha with SRI methods, with the highest so far being 13.3 t/ha (Sri Lanka). HYVs and hybrid varieties have given yields in the 15-20 t/ha range. SRI was developed in the 1980s with use of fertilizer, but after subsidies were removed at the end of the decade, Fr. de Laulanie switched to using compost and found that it could give even better yields on most soils, and farmers could be spared out-of-pocket cash costs. We have good reason to believe that compost will give higher yield when using SRI than will chemical fertilizer, seen from factorial trials, because of the impact on soil microbiology. However, farmers can use fertilizer with SRI methods usually cost-effectively. We do not discourage use of fertilizer, but rather encourage use of compost. Similarly, farmers can use agrochemicals (insecticides, fungicides, etc.) with SRI, but most farmers so far have reported that their rice with SRI is resistent enough to pest and disease damage that chemical applications are not necessary, i.e., not cost-effective. We do not know why there is this apparent pest and disease resistance, though we have hypotheses (better and more balanced nutrient uptake with a larger root system reaching lower horizons, better growing conditions when fields are not kept flooded, etc.)
  • This is usually a minor consideration, though for small farmers this can be important. Savings of 100 kg/ha are often reported with SRI, which is equivalent to a yield increase of 0.1 t/ha. No lodging is generally reported by farmers, though we have no systematic data on this. Also, farmers report that when harvesting SRI rice, there is less loss in the field from panicles. Environmental benefits remain to be evaluated systematically. It is known that emissions of methane are substantial from continuously flooded paddies, so SRI methods can be expected to reduce this greenhouse gas. Possibly the emission of nitrous oxide could increase when fields are not kept flooded, but if inorganic N is not being added in any large amounts, this is not likely to be much of a problem. The only capital requirement for SRI is purchase, or rental, of the hand push-weeder (rotating hoe) that controls weeds which are more of a problem when fields are not kept flooded. The weeder is not necessary, as hand weeding can control weeds. It will not, however, aerate the soil as the push-weeder does. (Herbicides can also control weeds with SRI, but they do nothing for soil aeration.) A study of SRI adoption and disadoption in Madagascar by Christine Moser from Cornell University in 2000, with CIIFAD support, found a good number of very poor households in her sample from four villages not adopting SRI, or giving it up, because they could not afford the additional labor required. Such households needed daily income during the crop season to meet their subsistance needs and could not afford to "invest" their labor in a higher SRI harvest. For them, SRI was "not more accessible." However, in countries such as Sri Lanka, Cambodia and Indonesia, farmers are now reporting that SRI is not more labor-intensive and has become even labor-saving for them. This matter is still being sorted out. But the statement here we think is generally true.
  • As noted for Slide 7, we are hearing from farmers in a number of countries, that SRI is not more labor-intensive for them. However, we think it best to acknowledge that SRI can require more labor, at least in the first year or two. Studies in Madagascar have put this increase between 25 and 50%, with first-year farmers sometimes even higher. With a doubling of yield, the returns to labor are higher even so. We do not want to minimize -- indeed, we should emphasize -- that SRI requires more skill and knowedge. Farmers are expected not to adopt SRI methods but to gain an understanding of them, particularly why we recommend wide spacing, young seedlings, no continuous flooding, etc. They should adapt the specific practices to their local conditions. SRI was intended by Fr. de Laulanie and Association Tefy Saina to encourage farmers to become more independent thinkings and active innovators. The most objective limitation on SRI is the need for good water control to get best results. Continuous flooding as seen below, leads to root deterioration. Farmers who are part of a cascade (field-to-field) irrigation system will have a hard time managing water for SRI unless there is cooperation among neighbors. Once the economic profitability of SRI has been well established, farmers, governments and even donor agencies should be willing to make investments in improving irrigation infrastructure to make SRI management possible.
  • This slide gives the most essential understanding of SRI. These ideas come from the work of Fr. de Laulanie and from five years of student thesis research in Madagascar and experimentation by a growing number of scientists in countries outside Madagascar. These generalizations remain to be fully documented and calibrated by additional research, but we are confident that this "core" of SRI is solid.
  • We emphasize that these are "starting point" because farmers are expected and encouraged to do some experimentation and adaptation with these practices, based on their understanding of the core concepts. The reasons for transplanting young seedlings are given in Slides 34-39. Quick and careful transplanting is necessary so there is little or no trauma to the young roots, which would set back their subsequent growth. We recommend that the roots be laid gently into the soil, only 1-2 cm deep, with the root straight downward or at least horizontal (L-shaped) rather than being plunged vertically down into the soil which causes the tip of the root to invert back upwards (J-shaped). When there is such inversion, it takes days, even weeks, for the root to reposition itself for resumed downward growth. Single plants spaced widely have room for both the roots and canopy to grow vigorously, as they will with young seedlings and aerated soil. We recommend starting with 25x25 spacing, but with good soil (and the soil usually improves year-to-year with SRI culivation), higher yields will be achieved with 30x30 or 40x40 spacing. The highest yield we know was with 50x50 cm spacing once soil had been improved by plant, soil, water and nutrient management. So farmers should experiment with wider spacing each year to see whether it leads to better crop performance than 25x25. They can experiment with narrower spacing if they like. No continuous flooding at least up to panicle initiation is key. This can be done, however, either by adding small amounts of water each day to keep the soil moist but not saturated, and not watering the field for 3-6 days several times during the growing season to dry out the field, up to the point of surface cracking; or by flooding the field for 3-6 days, and then draining it and leaving it dry for 3-6 days, until there is enough cracking to make reflooding necessary. We are still learning how to manage water for best effect with SRI. The best practices for any particular farmer will surely depend on soil, climatic, topographic and other variables. We recommend that farmers keep 1-2 cm of water on the field after panicle initiation, up to 10 days before harvest when the field should be drained (as done with all irrigated rice cultivation systems). Possibly the soil should be kept more aerated than this after panicle initiation, but no systematic research has been done. Weeding is important, for soil aeration as well as removal of weeds. See Slide 44.
  • The data summarize our observations and measurements. Regarding panicle size, we have had single panicles with as many as 900 grains, but this is so fantastic, few will believe it. The maximum of number of fertile tillers observed so far is 140, with 50x50 spacing. The most important phenotypic difference is the last one, discussed in Slides 21-22.
  • This picture was contributed from Cambodia by Koma Yang Saing (CEDAC). Viewers should try to imagine the very small single young seedling from which this massive plant grew.
  • This helps to explain our problem of "the agronomists' $100 bill." SRI is quite "counterintuitive." Indeed, it even sounds crazy. But we have experience and evidence that this "less is more" dynamic operates, and subsequent slides provide a number of scientific explanations for why fewer or smaller inpouts produce more in the case of irrigated rice
  • This figure is from research from the China National Rice Research Institute reported at the Sanya conference in April 2002 and published in the Proceedings. Two different rice varieties were used with SRI and conventional (CK) methods. The second responded more positively to the methods in terms of leaf area and dry matter as measured at different elevations, but there was a very obvious difference in the phenotypes produced from the first variety's genome by changing cultivation methods from conventional to SRI.
  • Usually we find researchers getting higher yields with new methods and farmers having difficulty "replicating" those results on their own fields. With SRI, we have often the opposite situation: researchers get lower yield on-station than farmers get in their fields. This remains to be thoroughly documented and fully explained. Fortunately, there is growing interest from crop and soil scientists after a number of years of skepticism and even resistance. We want to acknowledge, and express our appreciation for, the work of Fr. Henri de Laulanie, who developed SRI as a labor of love and innovative "lay science" during 34 years of living in Madagascar. He came there from his native France in 1961 and synthesized SRI first in the 1983-84 season. He had been educated in agriculture at the best French agricultural university (ENA) before World War II (1937-39) and before entering a Jesuit seminar (1941-45). So he knew basic agricultural science. But he had not learned about rice in France, so learned about it from and with farmers. The following slide shows Fr. de Laulanie visiting a farmers' field shortly before he died in 1995. In 1990, with his friends Sebastien Rafaralahy and Justin Rabenandrasana, Fr. de Laulanie formed Association Tefy Saina, a Malagasy NGO dedicated to improving rural conditions in Madagascar, including through the dissemination of SRI. Sebastien and Justin, now President and Secretary-General of Tefy Saina, are seen in Slide 17.
  • The first SRI evaluation outside Madagascar was done by Nanjing Agricultural University in China in 1999, getting 9.2-10.5 t/ha yields in three trials. Such yields can be attained in China using hybrid rice varieties and substantial inputs, but they could not be obtained using only half the usual amount of water. This is an important consideration in China where water for agricultural use is in increasingly short supply. Next, results came from AARD's Sukamandi rice research station. The first season gavie 1-1.5 t/ha higher yield than farmer practices, but the second season's yield reached 9.5 t/ha, spurring interest. After three years of evaluation, AARD made SRI part of its new "Integrated Crop and Resource Management" (ICM) strategy to get rice yields rising once again. Other countries learned about SRI through CIIFAD, through the International Institute for Rural Reconstruction, through ILEIA in the Netherlands, through ECHO in Ft. Myers, FL, and other channels. In April 2002, with a grant from the Rockefeller Foundation and CIIFAD support, an international meeting was held in China, hosted by the China National Hybrid Rice Research and Development Center and its director, Prof. L. P. Yuan. The proceedings include reports from 15 countries where SRI had been tried already, often with spectacular success, and only sometimes without seeing "the SRI effect" (Thailand, Nepal, Laos). The next slide gives a summary of results reported from the 10 countries where there had been most experience with SRI as of early 2002.
  • The bolded numbers are arithmetic averages for a number of SRI trials/evaluations in each country. Some of the data sets are from on-farm trials (with the number of farmers involved shown in parentheses) or from on-station trials. The range of results reported is also shown in each cell. The Comparison Yield figures are not national average figures, often lower, but averages for the trials or farmer practice using standard cultivation methods. The Maximum SRI yields are the highest yields reported in each data set.
  • Bruce Ewart, ADRA representative in Indonesia, got 7 farmers in West Timor to try SRI methods in 2002, with the encouragement of Roland Bunch. These are better farmers than their peers, as seen from their yield that season with current methods (4.4 t/ha), more than double the usual yield in the area. Their SRI plots averaged 11.7. Farmers working with ADRA in Lampung, Sumatra, got 8.5 t/ha with SRI methods compared to their usual production of 3 t/ha. Pablo Best reported that when farmers in Pucallpa, a lowland jungle area, tried SRI, they got a yield of 8 t/ha, four times their previous average, and not needing to do 8-10 hours/day of bird scaring at the end of the season because with SRI, the heavy panicles hung downward (but not lodging) so that birds could not get to them. Instead of letting cattle graze on the regrowth after harvest, the rice was allowed to produce a second (ratoon) crop, which was 5.5 t/ha, 70% of the first. Controlled trials in Benin, having read the account of SRI in ECHO Development Notes, found about a 5-fold difference in yield between SRI and conventional practice. The Agricultural Training Institute in the Philippines tried SRI methods with three varieties in Cotabato, Mindanao, and got an average yield of 12 t/ha, three times the usual yield in that area. The economic return averaged 290% as the value of rice produced was almost four times the cost of production.
  • This begins a consideration of the "science" behind SRI. This statement from an article by a number of leading rice scientists, published in a leading agronomy journal, states the standard scientific understanding of how irrigated rice grows: if there are more panicles per plant, there will be fewer grains per panicle, a manifestation of "the law of diminishing returns." This means that it is unpromising to increase rice plant tillering substantially, so rice breeding strategies, and particularly the "super-rice" being developed at IRRI, have aimed to reduce tillering. We find that with SRI methods, as seen in the next slide, there is a POSITIVE correlation, as plants with more panicles also have larger panicles. This is because SRI plants maintain a large and intact root system, as discussed below, making them "open systems" in which there is no necessary tradeoff (partitioning of photosynthate) between tillers and grains. With root degradation under continuously flooded conditions, rice plants are "closed systems" and there is a zero-sum relationship between tillering and grain-filling.
  • This was one of the first data sets that began laying a scientific foundation for SRI. Data were gathered from 76 farmers around Ambatovaky, a town on the western side of the peripheral zone around Ranomafana National Park in Madagascar, during the 1996-97 season. We had confidence in the field worker who collected the data, Simon Pierre, who had worked with Fr. de Laulanie before his death. The correlation between number of tillers per plant and number of grains per panicle was +.65, rather than the negative one expected from the literature. We have seen this positive relationship many times since this first analysis was done.
  • To get a systematic understanding of how the different factors brought together in SRI contribute to greater yield, two top students in the faculty of agriculture at the University of Antananarivo did baccalaureate thesis research involving large-scale factorial trials in 2000 and 2001. Jean de Dieu Rajaonarison did his study near Morondava on the west coast of Madagascar at the Centre de Baobab. He did not vary soil quality (the soil there was poor sandy soil, sable roux) but evaluated SRI effects with different varieties (half of the 288 plots were planted with a high-yielding variety and half with a local variety). Andry Andriankaja did his study in the village of Anjomakely 18 south of the capital Antananarivo, on the high plateau and on farmers' fields. He used the same variety (riz rouge) on two different kinds of soil (better clay and poorer loam). These were contrasting locations agroecologically as the Morondava site was near sea level, with tropical climate and poor soil, while the Anjomakely site was about 1200 m elevation, with temperate climate, and better soils. Both varied water management (aeration according to SRI principles vs. continuously saturated soil), seedling age (8 days vs. 16 days at Morondava and 20 days at Anjomakely, these latter ages being equivalent given differences in ambient temperature), plants per hill (1 vs. 3), and fertilization (compost vs. NPK vs. none as a control with the Morondava soils). Spacing was 25x25 or 30x30, but both were within the SRI range and gave no significant difference (absolutely no difference in Morondava and only 80 kg/ha at Anjomakely, other factors being equal). This was in a way a "mistake" in research design, as we should have compared a SRI and non-SRI spacing. But because there was no significant difference according to the spacing factor, we could combine those trials, having thus SIX replications for all of the averages calculated for the 48 combinations at Morondava and 40 combinations at Anjomakely.
  • Note that conventional practices -- 20-day seedlings, 3/hill, in saturated soil, with NPK amendments -- give 2-3 t/ha, while all-SRI practices -- 8-day seedlings, 1/hill, aerated soil, with compost -- give more than 3 times greater yield. (Note also that weeding and spacing were not varied in these factorial trials, which would have doubled the number of plots needed to 480, assuming just 3 rather than 6 replications.) The individual practice adding most to yield was young seedlings, then aerated soil, then single seedling, then compost. Pooling results from better and poorer soil (last column), we see that going from 75% to 100% SRI adds more to yield (almost 2 t/ha) than going from 0 to 25%, 25 to 50%, or 50 to 75%, an indication of synergistic effects among practices. The pattern of increase was similar for the trials at Morondava on poorer soil, with SRI practices increasing the HYV yield from 2.84 to 6.83 t/ha for the HYV and from 2.11 to 5.96 for the traditional variety, by 2.4 and 2.8 times respectively.
  • Here we look just at the effect of young seedlings, on better and poorer soil, at Anjomakely. The synergistic effect of compost with aerated soil is seen in the bottom three lines. Compost with saturated soil does less well (7.7 t/ha) than NPK with aerated soil (8.77 t/ha), but compost with aerated soil does by far the best (10.35 t/ha) on better soil. The same relationship is seen on poorer soil (right-hand column).
  • These are comparisons of resuls for the two sets of trials, looking at "other things being equal" effects. The averages shown are calculated for sub-samples (N= 144 in the Morondava study and N = 120 in the Anjomakely study) where all of the other practices than the one singled out for evaluation were an EQUAL NUMBER of SRI and non-SRI practices. These comments apply for the next slide as well.
  • This slide speaks for itself. The Kirk and Solivas statement was for flooded rice 29 DAT. This exact number can vary according to variety, soil type and irrigation practices, but it is agreed that the roots of continuously flooded rice remain in a "mat" near the surface, because the roots are trying to capture dissolved oxygen in the irrigation water. In such a situation, the roots begin degenerating after about 2 weeks of continuous flooding, as documented by Kar et al. There is practically no research on this process, since it has been characterized as "senescence" and thus as a natural (and an unavoidable) biological process. In fact, as the research by Puard et al. shows, the formation of aerenchyma (air pockets) in rice roots under flooding is a man-made process, leading to root degeneration.
  • This is the abstract for an article that documented this process of root degeneration. Unfortunately, it was published in a little-read journal. The exact number (78%) can vary according to variety, soil type, etc., but this phenomenon is well known. Only it is not regarded as a serious impediment for rice production.
  • These pictures of cross-sections of rice root, from French research. The upper-left cross-section is of an 'upland' variety grown under upland, i.e., unflooded, and the lower-left, the same variety grown under flooded conditions. The upper-right cross-section is of an 'irrigated' variety grown under flooded conditions -- note the larger, more regular air pockets formed by degeneration of the root's cortex -- and the lower-right, of the same variety grown under unflooded conditions. There is every reason to believe that the upper-left and lower-right cross-sections are the more 'natural' or 'normal' condition, and that the lower-left and upper-right represent adaptations, for the plant root tissues to survive longer as oxygen can diffuse through these air pockets. But this will not keep the tissues alive throughout the growth cycle, as seen from Slide 29.
  • This is a figure also from research reported by the China National Rice Research Institute to the Sanya conference and published in its proceedings. It shows how the roots of the same variety (two varieties shown) grow deeper into the soil with SRI methods compared to conventional ones (CK).
  • This figure from report by Nanjing Agricultural University researchers to the Sanya conference, and reproduced from those proceedings, shows that the oxygenation ability of rice roots growing under SRI conditions are about double the ability, throughout the growth cycle, compared to the same variety grown under conventional conditions.
  • This picture from Sri Lanka shows two fields having the same soil, climate and irrigation access, during a drought period. On the left, the rice grown with conventional practices, with continuous flooding from the time of transplanting, has a shallower root system that cannot withstand water stress. On the right, SRI rice receiving less water during its growth has deeper rooting, and thus it can continue to thrive during the drought. Farmers in Sri Lanka are coming to accept SRI in part because it reduces their risk of crop failure during drought.
  • This gets into the most complicated part of the explanation for SRI success, drawing on knowledge that is available in the literature but seldom known by scientists who do not read Japanese or who did not have teachers educated in Japan. T. Katayama studied tillering in rice, wheat and barley during the 1920s and 1930s, but never published his results until after the war (1951), and his book has not been translated into English. Fortunately, Fr. de Laulanie happened to read a book in French which reports on Katayama's concept of "phyllochrons," regular intervals of plant growth (emergence of phytomers -- units of a tiller, a leaf and a root -- from the apical meristem of grass family (gramineae) species. In rice, phyllochrons can be as short as 5 days with good growing conditions, but as long as 8-10 days with adverse conditions. The number of phyllochrons emerging in consecutive periods increases according to a regular pattern known in biological science and mathematics as "a Fibonacci series," where the number emerging in each period is equal to the total the numbers that emerged in the preceding two periods. [Look at numbers in bottom line of the slide, and then at the numbers for rice tillering in the next two slides]
  • In the first phyllochron (about 5 days, but it can be longer), the main tiller (and main root) emerge. Then no more tillers or roots emerge during the next two phyllochrons. This is the best time for transplanting because it minimizes root trauma. In the 4th phyllochron, a first primary tiller emerges from the base of the main tiller, and a second primary tiller emerges during the 5th phyllochron. In the 6th phyllochon, two tillers emerge, one from the main tiller (the third primary tiller) and one from the first primary tiller, because from the 4th phyllochron on, each tiller begins producing tillers, one per phyllochron, after a lag of two phyllochrons (periods). This can be seen from the next slide. The numbers start adding up quickly, as seen from the bottom row -- PROVIDED THAT THE PLANT HAS AN INTACT ROOT SYSTEM AND IS ABLE TO SUPPORT SUCH PROFUSE GROWTH. If the root system is not fully functioning, the phyllochrons lengthen and fewer are completed before panicle initiation (PI) when the rice plant switches from vegetative growth to reproduction and grain filling. This table shows tillering if the plant can complete 12 phyllochrons (periods) of growth before PI. Some SRI plants have have over 100 tillers, meaning that they got into the 13th phyllochron before PI. Under unfavorable growing conditions, phyllochrons become longer and few are completed before PI. A plant going through only8 phyllochrons of growth would have only 13 tillers, not 84. Note the geometric progression of groups of three phyllochrons, exhibiting a mathematical power relationship. The reason there are not 64 in the last period (phyllochrons 10-12) is that there is no room in the first row of tillers for a seventh tiller. Remember also that the growth of roots mirrors that of tillers, so the root structure multiplies similarly, provided that growing conditions are favorable. This table was worked out by Fr. de Laulanie, following the lines of observation and analysis laid out by Katayama.
  • This shows graphically what happens according to the numbers shown in the preceding table (Slide 35). Note that with SRI practices, which create a large and actively functioning root system for the plant, there is no fall-off in tillering before PI. With conventional rice growing practices, a period known as "maximum tillering" PRECEDES panicle initiation, as the earlier growth in tillering rate 'peaks' and subsides. We need more systematic monitoring of tillering rates in SRI and conventionally grown rice to put some parameters on this difference. But we know that this kind of "explosion" in tillering does occur with proper use of SRI practices, which support a large root system, and we know that "maximum tillering" occurs before PI with conventional plants.
  • This slide summarizes the plant physiological (but not the soil microbiological) elements of our 'theory' of SRI. It is informed by the article by Nemoto et al. in CROP SCIENCE (1995), the most explicit article we have found in English on phyllochrons in rice. The left-hand column are conditions that accelerate rice plant growth, i.e., shorten the phyllochron so that more of these periods are completed before PI. The right-hand column lists opposite conditions that slow plant growth. We refer to this, metaphorically, as 'slowing down vs. speeding up the biological clock,' a way of speaking about the rate of cell division and growth. The last two conditions, moisture and oxygen, are inversely correlated in that too much moisture means too little oxygen. Water management with SRI seeks to optimize moisture and oxygen in the root zone, ensuring enough of either, steering between drought and hypoxia. When plants are to close together, their roots (and canopies?) sense that there will not be enough nutrients, moisture, etc. to support maximum plant growth through the entire plant cycle to where the plant can produce mature seeds to ensure a next generation. Plant growth slows in response to the right-hand conditions in order to make more likely that at least some seeds will be produced before the plant senesces. When all of these conditions are favorable, growth speeds up. Raising yield requires providing ideal or optimal conditions. There is not much research done on phyllochrons in rice with regard to their manipulation or use for increasing production. Fr. de Laulanie thought that once we fully understand and utilize phyllochron dynamics, rice yields can be moved into the 20-30 t/ha range, given the inherent genetic potential of rice plants when ideal growing conditions are provided.
  • Young seedlings, transplanted before the 4th phyllochron, can respond more prolificly to these various favorable conditions than older seedlings, for reasons not well understood. It is important that there be no trauma to the rice plant root during transplanting, followed then by favorable soil, temperature, water, nutrient and other conditions. The logic of this analysis suggests that direct seeding could be made as or more productive as transplanting. We have some initial evidence that there is no loss of yield with direct seeding. This is an area where farmers and researchers should start doing systematic evaluation.
  • This summarizes what we know about phyllochrons and their effect on rice plant performance. It reiterates the need to consider what is happening below ground, with the roots, even though they are not seen. They are the basis for increased productivity, in conjunction with a more microbiologically active rhizosphere.
  • This picture was provided by Koma Yang Saing (CEDAC) of a pleased Cambodian farmer, showing the size of a massive root ball with a SRI rice plant.
  • This is known by plant scientists, but it has not been integrated into agronomic research strategies. There is a saying, "you can lead a horse to water, but you cannot make it drink." We want to emphasize that "you can provide large amounts of N to rice plants' roots, but you cannot make the roots take N up unless the plant needs it." This probably applies for other nutrients, but we have not seen research on this. The statements cited here are from IRRI and Cornell agronomists. The usual reaction we get hen pointing this relationship out is that this is already known, nothing new. It helps to explain why average uptake (efficiency) of N application is in the range of 20-30%.
  • SRI experience, buttressed by the preceding slide, suggests that we take a "demand-driven" view of N uptake and plant growth, rather than the prevailing "supply-side" approach. The latter is based on a misconception about how and why rice takes up N. The rapid explosion of tillering in the latter stages preceding PI creates a greater demand for N, and the larger canopy and roots during reproduction and grain filling also increase demand for N. It sounds incorrect to suggest "grow it and the N will come," because this suggests that demand will create its own supply. (This proposition is rejected because economists have correctly rejected Say's Law). However, as discussed below, with SRI, once we understand better the neglected role of soil microorganisms, we can see how subsurface processes can indeed provide N and other nutrients to the rapidly growing plant. There is not enough systematic evidence on rates, amounts, conditions, etc., to speak precisely about this matter, but we have the results of SRI practice to show that the needed nutrients must be coming from somewhere -- and they are not coming from inorganic additions.
  • This is a table from research done in 2001 in Madagascar by Barison for his MS thesis in agronomy at Cornell. He studied a sample of 108 farmers who used both SRI and conventional practices on their farms, so that the comparison of rice plant performance controlled for both farmer and farm differences. He used the QUEFTS model for analyzing yield response to inputs of nutrients (N, P and K) and found that SRI plants plateaued at a yield double that of conventional plants, 10 vs. 5 t/ha. The response of SRI vs. conventional plants was the same for both P and K. This showed that the soil-nutrient-root-shoot interactions were markedly different for plants grown with SRI methods.
  • This figure shows the yields associated with different numbers of mechanical hand weedings (with rotating hoe) for 76 farmers around Ambatovaky in 1996-97 (same as Slide 22). Two farmers did only manual weeding. They got about 6 t/ha yield, more than double the typical yield in the area. The 35 farmers who did 1-2 weedings, the minimum recommended, got 7.5 t/ha, triple the typical yield. The 24 who did 3 weedings got 9.2 t/ha, four times more, while the 15 who did 4 weedings, got 11.7 t/ha, about 5 times. Beyond 2 weedings, we think that the benefit is not really from the weeding but from the active soil aeration during the latter part of the vegetative growth phase. This is an area where controlled studies should be done. So far, all we have is data from farmers' fields. In other soils and other conditions, the absolute numbers will surely be different, but we think the pattern will hold up. On the very poor soils around Morondava, Frederic Bonlieu, doing research with 72 farmers practicing SRI for his thesis from Angers University in France, found that additional weedings added about 0.5 t/ha to yield, other things being equal. The same pattern we seen, but it was more linear and with less increment per weeding.
  • These last slides get into an area of SRI explanation that is more tentative, but probably more important for highest SRI yields. There is a lot of country-to-country variation in SRI results, and also within countries, much larger variations than can be explained by differences in practices or by differences in soil chemical and physical properties. We cite an observation by S. K. DeDatta in his well-known text on rice. We add our own emphasis to underscore our conclusion that there needs to be much more consideration of soil microbes and their contributions to rice yield. There is, however, little research on this subject, so DeDatta devoted very few pages to this compared to genetic, soil and other factors.
  • These are just the most obvious contributions. Our understanding of this netherworld is limited, though fortunately there are a growing number of microbiologists using very advanced modern techniques, such as DNA analysis, to map and track what is going on in the soil. The discussion that follows is can be viewed as introductory or superficial, or both.
  • Most people know that leguminous plants "fix" N in their roots through nodules on the roots inhabited by certain bacteria, rhizobia. And by implication, most thinks that non-leguminous plants "do not fix nitrogen." This is correct in terms of locus, but it misleads. All of the gramineae species (rice, wheat, sugar cane, etc.) have free-living bacteria in their root zones (referred to as 'associated' microbes) that fix N. Even in fertilized crops, a majority of the N taken up by the roots is from organic sources. And there is evidence that adding inorganic N to the root zone inhibits or suppresses the roots' and microbes' production of nitrogenase, the enzyme needed to fix N. So there is a tradeoff, in that adding inorganic N fertilizer reduces the N that is produced by natural biological processes. Or most relevance to SRI is research published more than 30 years ago reporting that when aerobic and anaerobic horizons of soil are mixed, BNF increases greatly compared to that originating from either aerobic or anaerobic soil. This suggests that the water management and weeding practices of SRI could be actively promoting N production in the soil. We have no research results to support this inference (though see data in Slide 49), but the yield increases with SRI practices require large amounts of N. BNF is the most plausible explanation.
  • Johanna Dobereiner has spend almost 40 years working on BNF particularly in sugar cane in Brazil, but also looking at BNF in other non-leguminous crops. Her 1987 book is the most extensive consideration of this subject, though there are a number of symposia also providing scientific information. In Brazil, it is well documented that BNF provides 150-200 kg/ha of N to the crop -- provided that soil has not been previously fertilized with inorganic N for some years, and provided (this is a little hard to understand) the sugar cane cultivars have not been fertilized for several generations. Applying N to the soil or to cultivars inhibits production of nitrogenase needed for BNF. Dobereiner's work is regarded as "controversial" within the agronomy profession because many efforts to replicate her results have failed. But this could be due to a mechanistic (non-biological) concept of the process, not appreciating how prior use of inorganic N can affect BNF potential, an effect documented in the literature (see van Berkum and Sloger).
  • These data from a study done by Fide Raobelison under the supervision of Prof. Robert Randriamiharisoa at Beforona station in Madagascar, and reported in Prof. Robert's paper in the Sanya conference proceedings, give the first direct evidence to support our thinking about the contribution of soil microbes to the super-yields achieved with SRI methods. The bacterium Azospirillum was studied as an "indicator species" presumably reflecting overall levels of microbial populations and activity in and around the plant roots. Somewhat surprisingly, there was no significant difference in Azospirillum populations in the rhizosphere. But there were huge differences in the counts of Azospirillum in the roots themselves according to soil types (clay vs. loam) and cultivation practices (traditional vs. SRI) and nutrient amendments (none vs. NPK vs. compost). NPK amendments with SRI produce very good results, a yield on clay soil five times higher than traditional methods with no amendments. But compost used with SRI gives a six times higher yield. The NPK increases Azospirillum (and other) populations, but most/much of the N that produced a 9 t/ha yield is coming from inorganic sources compared to the higher 10.5 t/ha yield with compost that depends entirely on organic N. On poorer soil, SRI methods do not have much effect, but when enriched with compost, even this poor soil can give a huge increase in production, attributable to the largest of the increases in microbial activity in the roots. At least, this is how we interpret these findings. Similar research should be repeated many times, with different soils, varieties and climates. We consider these findings significant because they mirror results we have seen in other carefully measured SRI results such as the Anjomakely factorial trials (Slide 24) and the previous season's trials with SRI at Beforona (10.2 t/ha).
  • These data are taken from the article by Ladha et al. (1998) but they did not draw any implication from their finding that optimum N fertilizer application was LOWER for late-maturing varieties than for early-maturing varieties -- and that the late-maturing varieties had higher yield with less fertilizer application than did early-maturing varieties. If volatilization and leaching of nutrients, particularl N, is as big a problem as stated in the article, these numbers should have been reversed. If, on the other hand, the N applied can "prime" soil microbiological processes that contribute to plant nutrition, a smaller amount over time could give higher yield. This is speculation, but it is consistent with relationships observed with SRI. It seems worth exploring.
  • The increase in yields around Ranomafana National Park during 1994/95-1998/99 from 2 t/ha to 8 t/ha ith SRI were quite inexplicable given that soil analyses by North Carolina State found on average that available P was only 3-4 ppm, which is less than half the minimum usually considered necessary for an acceptable yield. SRI farmers got twice as much as an acceptable yield without adding any P to the soil. Where did the P come from? The research reported by Turner and Haygarth in NATURE (May 17, 2001) could explain this since SRI methods involved alternate wetting and drying of the soil which the authors showed greatly increased levels of P in the soil solution, almost all from organic sources. They suggested that this effect probably applies for other nutrients too, but they were measuring only P.
  • Mycorrhizal associations have been largely ignored in rice because most is grown under continuously flooded conditions, which are inhospitable to growth of funguses. Yet we now know that 80-90% of plants depend in small to large part on the nutrient acquisition of funguses that "infect" their roots and provide access to a much larger volume of soil through the network of hyphae (filaments) that spread out in all directions. These hyphae acquire water and nutrients that ar shared with the plant, particularly P but also many others. Mycorrhizal hyphae are thinner even than hair roots so can access places in the soil that the root system cannot. One study found that "infected" plants could grow as well with 1/60th as much P in the soil as could "uninfected" plants, reported in the review on mycorrhizae by M. Habte and N.W. Osorio, Mycorrhizas: Producing and applying arbuscular mycorrhizal inoculum (2002). This is available on the web in The Overstory, #102 <http://agroforester.com/overstory/ovbook.html>
  • Research conducted in Egypt where farmers have for centuries alternated growing rice and berseem (clover) has shown that free-associated rhizobia in the root zone of rice (not living in nodules as they do on a legume) are abundant in this soil and have many measurable beneficial effects on rice growth. Surprisingly, the rhizobia do not contribute BNF for rice. Instead they stimulate nutrient uptake and plant growth in other ways. More research should be done on this particular microbe to understand what it could contribute to plant growth more generally. See Y. G. Yanni, et al., The beneficial plant growth-promoting association of Rhizobium leguminosarum bv. trifolii with rice roots, AUSTRALIAN JOURNAL OF PLANT PHYSIOLOGY, 28 (2001), pp. 845-870.
  • Research on rhizosphere function and dynamics is increasing. See the literature review by Robert Pinton et al., THE RHIZOSPHERE: BIOCHEMICAL AND ORGANIC SUBSTANCES AT THE SOIL-PLANT INTERFACE (Marcel Dekker, 2000) which gives an up-to-date review of what is currently known about this domain, with chapters on root exudation,rhizo- deposition, the contributions of mycorrhizae, etc.
  • This review of what is known, and what we think we know, about SRI is not a conclusive or final discussion of the subject. We expect that in 3-5 years' time, much more will be known as scientists become engaged on these topics and as farmers and NGOs continue producing new data that begs for explanations. The two main areas for research that have emerged from our SRI experence is (a) the growth and performance of roots, and (b) the dynamics and contributions of soil microbes. Both of these areas of research should be useful for improving the production of other crops. The explanations for greater root growth with SRI are quite straightforward: young seedlings, wide spacing, aerated soil. What results from this remains to be better documented and explained.
  • More complicated and difficult to examine and understand than roots will be soil microbial dynamics, though these two subjects should be looked at jointly, not totally separately. The contributions of exudates to microbial growth are well documented in Pinton et al. (2001), but we do not know much about this process for irrigated (SRI) rice.
  • This is a SRI rice nursery in Sri Lanka, showing one way (but only one of many ways) to grow young seedlings. The soil in this raised bed was a mixture of one-third soil, one-third compost, and one-third chicken manure. (The flooding around it is because the surrounding field is being readied for transplanting; normally there would not be so much water standing around the nursery.)
  • Here the seedlings are being removed. We would recommend that they be lifted with a trowel, to have minimum disturbance of the roots, but these seedlings are so vigorous that this manual method is successful. This farmer has found that his seedlings, when transplanted with two leaves at time of transplanting, already put out a third leave the next day after transplanting, indicating that there was no transplant 'shock.'
  • Here the field is being 'marked' for transplanting with a simple wooden 'rake.' If the soil is too wet, these lines will not remain long enough for transplanting. There are drains within the field to carry excess water away from the root zone.
  • Here are seedlings being removed from a clump for transplanting. Note that the yellow color comes from the sunlight reflecting off the plant. The plant's color is a rich green, indicating no N deficiency.
  • Here the seedlings are being set into the soil, very shallow (only 1-2 cm deep). The transplanted seedlings are barely visible at the intersections of the lines. This operation proceeds very quickly once the transplanters have gained some skill and confidence in the method. As noted already, these seedling set out with two leaves can already have a third leaf by the next day.
  • The SRI field looks rather sparse and unproductive at first. Up to the 5th or 6th week, SRI fields look rather miserable, and farmers can wonder why they ever tried this method and 'wasted' their precious land with such a crazy scheme. But the SRI plot here will yield twice as much rice as the surrounding ones once the rapid tillering (and root growth) begins between 35 and 45 days.
  • This is one of many happy Sri Lankan farmers with his SRI field nearing harvest time. Some young farmers have taken up growing "eco-rice," i.e., traditional varieties grown organically to be sold for a much higher price than conventional HYV rice, because of better texture, taste, smell and aroma and more assurance of healthy food. SRI in this way is starting to contribute to the preservation of rice biodiversity. As noted above, SRI methods work well with hybrid varieties and HYVs. These give the highest yields with SRI methods. But as SRI methods can double or triple traditional-variety yields, these old varieties become economically more advantageous with SRI. Much more remains to be learned about and from SRI. But we have now enough accumulated evidence, based on experience in farmers' fields, not just on experiment stations, and consistent with what is known in the literature (though often not previously connected up to promote increased rice productivity), to have confidence that this methodology will contribute to greater food security and a better environment. SRI, developed by Fr. de Laulanie and promoted by his friends in Association Tefy Saina, and by a growing number of colleagues in many countries around the world, could help to improve other crop production. The world does not need a doubling of rice production, but it does need increased productivity in the rice sector, as this is the largest single agricultural sector in the world in terms of the resources devoted to it. By raising the productivity of land, labor, water and capital in the rice sector, we should be able to meet our staple food needs with less of these resources, which have significant opportunity costs. We hope that SRI methods will enable farmers to redeploy some of their land, labor, water and capital to producing other, higher-value and more nutritious crops, thereby enhancing the well-being of rural households and urban populations. The urban poor should benefit from lower prices for rice that will follow from higher productivity. SRI is not a labor-intensive method that will 'keep rice production backward,' as was alleged by its critics in Madagascar for many years, but a strategy for achieving diversification and modernization in the agricultural sector.
  • 0321 The System of Rice Intensification (SRI) : What It Is, and How/Why We Think It Works

    1. 1. The System of Rice Intensification (SRI) -- What It Is, and How/Why We Think It Works Norman Uphoff Cornell International Institute for Food, Agriculture and Development
    2. 2. More tillers and more than 400 grains per panicle
    3. 3. SRI appears ‘too good to be true’ (like the economist’s $100 bill?) <ul><li>It goes against many concepts and beliefs of agronomists and economists: yield ceiling, soil depletion, tradeoffs, diminishing returns, etc. </li></ul><ul><li>However, there is increasing evidence that SRI greatly raises rice productivity </li></ul><ul><li>SRI is being used successfully by </li></ul><ul><ul><li>a growing number of farmers in </li></ul></ul><ul><ul><li>a growing number of countries (16+) </li></ul></ul>
    4. 4. OBSERVABLE BENEFITS <ul><li>Average yields about 8 t/ha -- </li></ul><ul><li>twice present world average of 3.8 t/ha </li></ul><ul><li>Maximum yields can be twice this -- 15-16 t/ha, with some over 20 t/ha </li></ul><ul><li>Water required reducible by about 50% </li></ul><ul><li>Increased factor productivity from land, labor, capital and water ( > yield) </li></ul><ul><li>Lower costs of production -- this is most important for farmers </li></ul>
    5. 5. SRI Data from Sri Lanka <ul><li> SRI Standard </li></ul><ul><li>Yields (tons/ha) 8 4 +88% </li></ul><ul><li>Market price (Rs/ton) 1,500 1,300 +15% </li></ul><ul><li>Total cash cost (Rs/ha) 18,000 22,000 -18% </li></ul><ul><li>Gross returns (Rs/ha) 120,000 58,500 +74% </li></ul><ul><li>Net profit (Rs/ha) 102,000 36,500 +180% </li></ul><ul><li>Family labor earnings Increased with SRI </li></ul><ul><li>Water savings 40-50% </li></ul><ul><li>Data from Dr. Janaiah Aldas, formerly economist at IRRI, now at Indira Gandhi Development Studies Institute, Mumbai, based on visit to Sri Lanka and interviews with SRI farmers, October, 2002 </li></ul>
    6. 6. LESS OR NO NEED FOR: <ul><li>Changing varieties , though best yields from high-yielding varieties and hybrids -- traditional varieties produce 4-10 t/ha </li></ul><ul><li>Chemical fertilizers -- these will give a positive yield response with SRI, but best results are obtained with compost </li></ul><ul><li>Agrochemicals – plants more resistant to pests and diseases with SRI methods </li></ul>
    7. 7. ADDITIONAL BENEFITS <ul><li>Seeding rate reduced as much as 90%, 5-10 kg/ha gives more than 50-100 kg </li></ul><ul><li>No lodging because of stronger roots, and stronger panicles (pedicules) </li></ul><ul><li>Environmentally friendly production due to water saving, no/fewer chemicals </li></ul><ul><li>More accessible to poor households because few capital requirements </li></ul>
    8. 8. DISADVANTAGES / COSTS <ul><li>SRI is more labor-intensive , at least initially -- but may become labor-saving </li></ul><ul><li>SRI requires greater knowledge/skill from farmers to become better decision-makers and managers -- but this contri-butes to human resource development </li></ul><ul><li>SRI requires good water control to get best results, making regular applications of smaller amounts of water -- this can be obtained through investments </li></ul>
    9. 9. The basic idea of SRI is that RICE PLANTS DO BEST when <ul><li>Their ROOTS grow large and deep since they were transplanted carefully and the seedlings experienced little trauma, and with wider spacing between plants ; and </li></ul><ul><li>They can grow in SOIL that is kept </li></ul><ul><li>well aerated , never continuously saturated, </li></ul><ul><li>with abundant and diverse populations of </li></ul><ul><li>soil microbes that aid in plant nutrition </li></ul>
    10. 10. ‘ Starting Points’ for SRI: <ul><li>Transplant young seedlings , 8-15 days (2 leaves) -- quickly and very carefully </li></ul><ul><li>Single plants per hill with wide spacing in a square pattern -- 25x25 cm or wider </li></ul><ul><li>No continuous flooding of field during the vegetative growth phase (AWD ok) </li></ul><ul><li>Weeding with rotating hoe early (10 DAT) and often -- 2 to 4 times </li></ul><ul><li>Application of compost is recommended </li></ul><ul><li>These are adapted to local situations </li></ul>
    11. 11. SRI practices produce a different RICE PHENOTYPE: <ul><li>Profuse TILLERING -- 30 to 50/plant, 80-100 possible, sometimes 100+ </li></ul><ul><li>Greater ROOT GROWTH -- 5-6x more resistance (kg/plant) for uprooting </li></ul><ul><li>Larger PANICLES -- 150-250+ grains </li></ul><ul><li>Higher GRAIN WEIGHT -- often 5-10% </li></ul><ul><li>A POSITIVE CORRELATION between tillers/plant and grains/panicle </li></ul>
    12. 13. SRI goes against LOGIC <ul><li>LESS PRODUCES MORE -- by utilizing the potentials and dynamics of biology </li></ul><ul><li>Smaller, younger seedlings will give larger, more productive mature plants </li></ul><ul><li>Fewer plants per hill and per m 2 can give more yield </li></ul><ul><li>Half as much water gives higher yield </li></ul><ul><li>Fewer or no external inputs are associated with greater output </li></ul><ul><li>New plant types from existing genomes </li></ul>
    13. 14. Plant Physical Structure and Light Intensity Distribution at Heading Stage (CNRRI Research: Tao et al. 2002)
    14. 15. These results more often come from farms than experiment stations <ul><li>But increasing number of scientists working on SRI -- in China, Indonesia, India, Bangladesh, Cuba, etc. </li></ul><ul><li>SRI is the due entirely to the work of Fr. Henri de Laulanié, S.J . (1920-1995), trained in agriculture at INA (1937-1939) </li></ul><ul><li>He lived and worked with farmers in Madagascar (1961-1995), SRI from 1983 </li></ul><ul><li>SRI being promoted by Malagasy NGO, Association Tefy Saina , assisted by CIIFAD </li></ul>
    15. 18. Spread beyond Madagascar <ul><li>Nanjing Agricultural University - 1999 </li></ul><ul><li>Agency for Agricultural Research and Development, Indonesia - 1999-2000 </li></ul><ul><li>Philippines, Cambodia, Sri Lanka, etc. </li></ul><ul><li>International conference, Sanya, China, April 2001 -- 15 countries reported on experience with SRI </li></ul>
    16. 19. Data from Sanya Conference
    17. 20. Average Yields Impressive: Certain Cases Hard to Explain <ul><li>Indonesia -- West Timor (ADRA) </li></ul><ul><li>Yield with current methods -- 4.4 t/ha </li></ul><ul><li>Yield with SRI methods -- 11.7 t/ha </li></ul><ul><li>Peru -- Pucallpa, jungle area </li></ul><ul><li>Previous yields -- 2 t/ha, with more labor </li></ul><ul><li>SRI yield -- 8 t/ha, with less labor </li></ul><ul><li>+ Ratoon crop = 70% of first crop (5.5 t/ha) </li></ul><ul><li>Benin -- controlled trial: 1.6 vs. 7.5 t/ha </li></ul><ul><li> WHAT IS GOING ON? </li></ul>
    18. 24. Factorial Trials Evaluating 6 Factors <ul><li>Variety : HYV (2798) vs. local ( riz rouge ) or Soil quality : better (clay) vs. poorer (loam) </li></ul><ul><li>Water mgmt : aerated vs. saturated soil </li></ul><ul><li>Seedling age : 8 days vs. 16 or 20 days </li></ul><ul><li>Plants per hill : 1/hill vs. 3/hill </li></ul><ul><li>Fertilization : compost vs. NPK vs. none </li></ul><ul><li>Spacing : 25x25cm vs. 30x30cm (NS diff.) </li></ul><ul><li>6 replications : 2.5x2.5m plots (N=288, 240) </li></ul>
    19. 26. Effect of Young Seedlings <ul><li>Better Poorer </li></ul><ul><li>SS/20/3/NPK 3.00 2.04 </li></ul><ul><li>SS/ 8 /3/NPK 7.16 3.89 </li></ul><ul><li>SS/ 8 / 1 /NPK 8.13 4.36 </li></ul><ul><li>AS / 8 /3/NPK 8.15 4.44 </li></ul><ul><li>AS / 8 /3/ Comp 6.86 3.61 </li></ul><ul><li>SS/ 8 / 1 / Comp 7.70 4.07 </li></ul><ul><li>AS / 8 / 1 /NPK 8.77 5.00 </li></ul><ul><li>AS / 8 / 1 / Comp 10.35 6.39 </li></ul>
    20. 27. Results of Factorial Trials ( ceteris paribus effects) <ul><li>Seedling Age 16/20 8 t/ha </li></ul><ul><li>Morondava (288) 2.61 3.96 +1.35 </li></ul><ul><li>Anjomakely (240) 3.80 6.78 +2.48 </li></ul><ul><li>Water Management Flood Control t/ha </li></ul><ul><li>Morondava (288) 2.86 3.71 +0.85 </li></ul><ul><li>Anjomakely (240) 4.34 5.75 +1.41 </li></ul>
    21. 28. More Factorial Results <ul><li>Plants per Hill 3 1 t/ha </li></ul><ul><li>Morondava (288) 3.05 3.51 +0.46 </li></ul><ul><li>Anjomakely (240) 4.65 5.43 +0.78 </li></ul><ul><li> Com- </li></ul><ul><li>Nutrient Amendments NPK post t/ha </li></ul><ul><li>Morondava (half HYV) 3.69 3.96 +0.27 </li></ul><ul><li>Anjomakely (all trad’l.) 4.48 5.49 +1.01 </li></ul>
    22. 29. Critical Factor is Root Growth <ul><li>3/4 of rice roots in continuously flooded soil remain in top 6 cm (Kirk and Solivas 1997) </li></ul><ul><li>3/4 of rice roots in continuously flooded soil degenerate by time of flowering (Kar 1974) </li></ul><ul><li>Air pockets (aerenchyma) form in roots of rice plants when continuously flooded. </li></ul><ul><li>These air pockets enable rice plants to survive under submerged conditions. </li></ul><ul><li>But submerged plants do not thrive -- </li></ul><ul><li>Lacking oxygen, their roots degenerate </li></ul>
    23. 32. Dry Matter Distribution of Roots in SRI and Conventionally-Grown Plants at Heading Stage (CNRRI research: Tao et al. 2002) Root dry weight (g)
    24. 33. Root Activity in SRI and Conventionally-Grown Rice (Nanjing Agr. Univ. research: Wang et al. 2002) (Wuxianggeng 9 variety)
    25. 35. With young transplants and vigorous root growth , <ul><li>TILLERING can be much greater </li></ul><ul><li>This can be explained in terms of phyllochrons -- interval of plant growth found in all “grass” species </li></ul><ul><li>Discovered by Japanese scientist Katayama in 1920s-30s </li></ul><ul><li>Tillering pattern follows sequence of ‘ Fibonacci series ’ -- 1, 1, 2, 3, 5, 8, 13... </li></ul>
    26. 38. What speeds up the biological clock? (adapted from Nemoto et al. 1995) <ul><li>Shorter phyllochrons Longer phyllochrons </li></ul><ul><li>Higher temperature > cold temperature </li></ul><ul><li>Wider spacing > crowding of roots/canopy </li></ul><ul><li>More illumination > shading of plants </li></ul><ul><li>Ample nutrients in soil > nutrient deficits </li></ul><ul><li>Soil penetrability > compaction of soil </li></ul><ul><li>Sufficient moisture > drought conditions </li></ul><ul><li>Sufficient oxygen > hypoxic soil </li></ul>
    27. 39. What is essential to speed up the biological clock? <ul><li>Use of YOUNG SEEDLINGS -- best to transplant during ‘window of opportunity’ before 4th phyllochron </li></ul><ul><li>Rice plants can respond best to other favorable conditions if minimally disturbed during transplantation </li></ul><ul><li>This suggests that DIRECT SEEDING may work as well as transplanting and lower labor needs -- should evaluate </li></ul>
    28. 40. With best growing conditions, the phyllochrons are shorter <ul><li>So more periods of growth can be completed before the rice plant switches from </li></ul><ul><li>its vegetative growth phase </li></ul><ul><li>to its reproductive phase </li></ul><ul><li>With more tillering , there is also more root development </li></ul>
    29. 42. SRI capitalizes on the fact that the uptake of N is a demand-led process
    30. 43. Alternative Models for Nitrogen Uptake <ul><li>Supply-Side Model to Increase Growth </li></ul><ul><li>Apply N to the soil to raise N availability </li></ul><ul><li>This assumes that rice plants will take up more N if it is easier for them to access N because of higher concentrations of N in the root zone </li></ul><ul><li>Demand-Driven Model for Promoting Growth </li></ul><ul><li>Manage plants so as to accelerate their rate of growth </li></ul><ul><li>This reflects an under-standing that increased plant demand for N is what induces the roots to take up more N </li></ul>
    31. 45. Benefits are observed from soil aeration during the vegetative growth period
    32. 46. Soil microbial activity is critical for plant nutrition and SRI performance <ul><li>“ The microbial flora causes a large number of biochemical changes in the soil that largely determine the fertility of the soil.” (DeDatta, 1981, p. 60, emphasis added) </li></ul>
    33. 47. Bacteria, funguses, protozoa, amoeba, actinomycetes, etc. <ul><li>Decompose organic matter , making nutrients available </li></ul><ul><li>Acquire nutrients that are unavailable to plant roots </li></ul><ul><li>Improve soil structure and health (water retention, pathogen control) </li></ul>
    34. 48. Biological Nitrogen Fixation <ul><li>Microorganisms -- particularly bacteria, both aerobic and anaerobic -- can fix nitrogen (N) from air into forms available to plant roots </li></ul><ul><li>Research has shown that when aerobic soil and anaerobic soil are mixed , rather than having only aerobic soil or only anaerobic soil, BNF increases greatly (Magdoff and Bouldin, 1970) </li></ul>
    35. 49. Biological Nitrogen Fixation <ul><li>BNF can occur with all gramineae species, including rice (Döbereiner 1987, and others) </li></ul><ul><li>In flooded paddies, BNF is limited to anaerobic processes ; SRI provides aerobic conditions as well ; BNF must be occurring for the higher yields observed; not enough N measured in the soil </li></ul><ul><li>The use of chemical fertilizers inhibits the production by roots and microbes of nitrogenase , the enzyme needed for BNF (van Berkum and Sloger 1983) </li></ul>
    36. 51. This helps to solve puzzle <ul><li>Why were many Madagascar farmers putting their compost for SRI on their contra-saison crop -- not on rice crop? </li></ul><ul><li>Both crops reportedly gave better yield </li></ul><ul><li>This makes no sense if LEACHING and VOLATILIZATION are big problems, or if nutrients are ‘used up’ by plants </li></ul><ul><li>It makes sense, however, for BNF </li></ul>
    37. 53. Phosphorus Solubilization <ul><li>Aerobic bacteria can acquire phosphorus from unflooded soil for their own use </li></ul><ul><li>When the soil is flooded, these bacteria die (lyse) and release their contents into the water that permeates the soil </li></ul><ul><li>When the soil dries again, surviving bacteria begin their growth again </li></ul><ul><li>Soluble P can increase by 185-1,900% by such ‘mining’ of the soil that increases nutrient supply (Turner & Haygarth, 2001) </li></ul>
    38. 54. Microbiological ‘Weathering’ of Soil? <ul><li>Soluble P can increase by 185-1,900% by microbiological ‘mining’ of the soil (Turner & Haygarth, 2001) </li></ul><ul><li>Speculation that this process operates increase supply of other nutrients too </li></ul><ul><li>Under ‘natural’ conditions, ‘depletion’ of soil is very rare occurrence -- due to microbiological processes </li></ul>
    39. 55. Mycorrhizal Associations <ul><li>Mycorrhizal funguses ‘infect’ plant roots </li></ul><ul><li>They send out hyphae (filaments/threads) in all directions and expand the volume of soil that the plant can extract nutrients from by 10-100 times </li></ul><ul><li>Mycorrhizae are very good at harvesting phosphorus -- increased efficiency by 60x </li></ul><ul><li>Mycorrhizae cannot grow in anaerobic soil conditions, so cannot benefit irrigated rice </li></ul>
    40. 56. Benefits from Rhizobia in rice now being explored <ul><li>Studied where rice and clover grown in rotation in Egypt, for many centuries </li></ul><ul><li>These endophytic bacteria induce more efficient acquisition of N, P, K, Mg, Ca, Zn, etc. in rice (Yanni et al. 2001) </li></ul><ul><li>Rhizobia increase yield and total protein quantity/ha , by producing auxins and other plant-growth promoting hormones -- however, no BNF demonstrated </li></ul>
    41. 57. Root Exudation <ul><li>Plant stems & roots are ‘two-way’ streets </li></ul><ul><li>30-60% of the energy (sugars, proteins) made in the canopy is sent to the roots (Pinton et al., 2000) </li></ul><ul><li>20-40% of this energy supply is exuded by the roots into the soil -- feeding the bacteria, funguses, etc. in the root zone </li></ul><ul><li>Root cells also die and provide energy to microbes through rhizodeposition </li></ul><ul><li>Plants gain more than they lose from this </li></ul>
    42. 58. SRI Supports the Motto of Organic Farmers <ul><li>Don’t try to feed the plant -- </li></ul><ul><li>Feed the soil -- and the soil will feed the plant </li></ul><ul><li>Emphasis on symbiosis between plants and soil microorganisms </li></ul>
    43. 59. SRI can be seen as an agronomic system for: <ul><li>Plant management -- young seedlings, careful transplanting, wide spacing </li></ul><ul><li>Soil and water management -- leveling, ‘minimum of water’ for soil aeration </li></ul><ul><li>Nutrient management -- increase SOM </li></ul><ul><li>Microorganism management -- result of the above, promoted by root exudation </li></ul>
    44. 60. SRI Raises More Questions than we have answers for <ul><li>Many of the answers will be found in the growth and functioning of ROOTS, which grow better from: </li></ul><ul><li>YOUNG SEEDLINGS, with </li></ul><ul><li>WIDE SPACING, and in </li></ul><ul><li>AERATED SOIL </li></ul>
    45. 61. <ul><li>Answers will also be found in SOIL MICROBIAL DYNAMICS -- in the abundance & diversity of soil microbes (bacteria, fungi) </li></ul><ul><li>Microbes grow better in: </li></ul><ul><li>SOIL not continuously flooded , </li></ul><ul><li>with more soil organic matter </li></ul><ul><li>Microbes benefit from exudation resulting from more root growth </li></ul>
    46. 67. .
    47. 69. THANK YOU <ul><li>More information is available </li></ul><ul><li>on the SRI WEB PAGE : </li></ul><ul><li>http://ciifad.cornell.edu/sri/ </li></ul><ul><li>including Sanya conference proceedings, </li></ul><ul><li>available on CD ROM discs </li></ul><ul><li>E-MAIL ADDRESSES : </li></ul><ul><li>[email_address] </li></ul><ul><li>[email_address] </li></ul><ul><li>[email_address] </li></ul>