Nitrogen use efficiency by selected NERICA varieties in Burkina Faso

A field experiment was carried out in the 2005 and 2006 wet seasons in Bagré, Burkina Faso, to assess the differences in paddy yield and nitrogen (N) utilization of the New Rice for Africa (NERICA) lowland varieties. The test consists of a split-plot design with four replications. The main plots were represented by four NERICAs (NERICA L 41, NERICA L 60, NERICA L 20, and NERICA L 19) and one control (4418). Subplots were constituted with five rates of nitrogen fertilizer (0, 40, 80, 120 and 160 kg N ha -1 ). Significant differences (P<0.001) among varieties were observed in paddy yield and nitrogen uptake. Without nitrogen application (0N), NERICAs insure a good grain yield compared to the control (4418). The N-use efficient varieties that produced high paddy yield at both low and high levels of N were NERICA L 41, NERICA L 20 and NERICA L 19. NERICA L 60 that is not significantly different to the control (4418) appears less efficient as compared to the three other NERICAs.


INTRODUCTION
Rice is developing as a major staple food crop of Burkina Faso.Demand has grown at an annual rate of 3% between 1973 and 1992 compared with an annual population growth rate of 2.9%, which can be explained by changing consumer preferences (Africa Rice Center, 2008).The annual per capita consumption increased from 18.2 kg in 1999 to 21 kg in 2008 (MAHRH, 2010).It reached 50 kg per person in urban centers in Ouagadougou and Bobo-Dioulasso.Domestic production in paddy rice was 195,102 tons in 2008195,102 tons in , 249,063 tons in 2011195,102 tons in and 319,390 tons in 2013195,102 tons in (CEFCOD, 2013)).Currently, in-country production covers 42% of the demand estimated at 255 176 tons of white rice, and 58% is met from imports (CEFCOD, 2013).While irrigated lowlands comprise only about 23% of the total rice area, this system is characterized by considerably higher yields and contributes about 53% to national rice production (INERA-DGPER, 2010).
The lowland rice cropping system is the traditional shape of most practiced rice cropping system in Burkina Faso.This type of rice cropping system takes the most part of areas and combines the rainfed and irrigated systems.It occupies about 90% of the rice areas of Burkina Faso, and contributes for 95% to the country's rice production (INERA-DGPER, 2010).The traditional cultivars that take the most part of valley bottoms are more and more in way of disappearance because of their cycle less adapted to the reduction of the rainfall (Sié and Dakouo, 1998).
The Africa Rice Center and partners have developed a family of 60 interspecific hybrids adapted to lowland growth conditions by crossing Asian rice (Oryza sativa) and African rice (Oryza glaberrima).These genotypes have been named lowland NERICAs (NERICA-L).Their main breeding objectives were yield potential, grain quality, broad adaptation to diverse lowlands in the region, and resistance against Rice Yellow Mottle Virus and African Gall Midge (Sié et al., 1998;Sié and Dakouo, 1998;Sié et al., 2004;Sié et al., 2008).Rodenburg et al. (2006) provided complete information on pedigrees and backcross parents of the lowland NERICA cultivars.
However, breeding program aiming to improve nitrogen use efficiency is still limited.Soil fertility, fertilizer use and crop response to nutrient inputs may vary widely among regions and/or rice fields within smaller irrigated and rainfed rice environments, and also from season to season in the same field (Adhikari et al., 1999;Olk et al., 1999;Dobermann and Cassman, 2002;Dobermann et al. 1998;2003).
Nitrogen is the most yield-limiting nutrient in rice cropping systems worldwide (Mikkelsen, 1987;Cassman et al., 1996aCassman et al., , 1996b;;Jiang et al., 2004) and because of the many pathways for the nitrogen loss, especially in the alternating wet/dry cycles of rice systems; it is also the most difficult nutrient to manage (Mikkelsen, 1987;Buresh et al., 1989).
Beside the importance of the nitrogen on the irrigated rice productivity, the relation between N supply (by fertilization) and indigenous N (from soil) is in general weak (Cassman et al., 1996c); so a weak efficiency of N use is observed (Olk et al., 1999;Wopereis et al., 1999).The losses of nitrogen coming from fertilizers are estimated from 20 to 70% (Cassman et al., 1993;Wopereis et al., 1999).These authors conclude that the contributions of N fertilizer that do not take into account the soil nutrient capacity contribute to significant losses by denitrification and by ammoniac volatilization.Fertilizers represent about 30% of the total production costs and all practices that limit N losses present economic advantages (Donovan et al., 1999).Since the cost of fertilizer is increasing and resource-poor farmers in the fragile environments cannot afford to buy fertilizer, identification or development of rice genotypes that are adapted to non-fertilized conditions and have high responsiveness to nutrient inputs would be an attractive and cost-effective approach (Ladha et al., 1998;Saito and Futakuchi, 2009).Genotypic difference in response to applied fertilizer has been reported repeatedly in the lowlands in Asia (Ladha et al., 1998;Inthapanya et al., 2000a;Haefele et al., 2008, Saito et al., 2010a).However, similar information for West African, like Burkina Faso conditions is scarce.Because of the high potential for N loss, N use efficiency in rice trends to be lower in comparison with other major crops (Keeney and Sahrawat, 1986).The identification of the factors that determine grain yield and nitrogen utilization in rice production systems is necessary to optimize their productivity and reduce the pollution risk for the environment (Koutroubas and Ntanos, 2003).Reduction of N loss would increase both N in the soil, fertilizer-N use efficiency, and reduce the environmental costs associated with denitrification and leaching of NO 3 (George et al., 1993).
The objectives of the study were to evaluate the response of rice varieties to nitrogen fertilizer.More specifically, the study intend (i) to assess the varietal differences in the efficiency of indigenous nitrogen use, (ii) to assess the varietal response to N fertilizer, (iii) to determine the efficiency, the inefficiency and the low responses of varieties to nitrogen supply and (iv) to determine the paddy yield and nitrogen uptake.

Description of the study area
The study was conducted in the rice plain of Bagré village (11°30' N, 0°25' W) located in the eastern part of Burkina Faso, West Africa.The climate is typical for the agro ecological zone of the Sudan savanna with rainy season occurring from July to October, followed by a cold and dry season from November to February, and a hot dry season from March to June.Average annual rainfall is 850 mm year -1 and minimum air temperature below 15°C occur in the cold dry season, and maximum temperatures above 39°C occur in the hot dry season (BEGE, 2008).

Soil types
Soils of the scheme are developed in alluvial sediments of Quaternary age.According to FAO classification (FAO, 1988), soils of the irrigated scheme (600 ha on the left bank of the Nakanbe river) were classified as Gleysols and dystric Fluvisols (62% of total area).Soil depth was on average between 0.4 to 1.2 m.Physical and chemical characteristics of experimental site are presented in Table 1.

Experimental design and crop management
The experimental design is a split-plot with four replications.Five increasing doses of nitrogen (0, 40, 80, 120 and 160 N kg ha -1 ) constitute the sub-plot treatments.The main plots are represented by five rice varieties.Four of them are interspecific varieties (O.glaberrima x O. sativa) selected on the basis of their agronomic performances after participatory varietal selection (PVS) (Sié et al., 2004).These varieties are NERICA L 19, NERICA L 20, NERICA L 41 and NERICA L 60.The control variety ( 4418) is one of the more used by Bagré's farmers.Each variety is transplanted in a plot of 6 m² with a density of 0.20 × 0.20 m.Phosphorus (30 kg P2O5 ha -1 ) and potassium (35 kg K2O ha -1 ) have been applied during rice transplantation.Nitrogen (kg ha -1 ) is applied at active tillering stage and at panicle initiation according timing in Table 2. Three types of mineral fertilizer have been used for the experimentation: the urea (46-0-0), the super phosphate triple (TSP, 0-45-0) and the potassium chloride (KCl, 0-0-60).

Chemical analysis
Five top-soils (0-0.20 m) samples have been collected in each plot before soil tillage.Samples have been dried on air and sieved at 2 mm.Soil analysis included pH H20 (1:2.5 extract), electrical conductivity of the 1:5 soil-extract (Exchangeable Cations), total carbon, total nitrogen, phosphorus-Bray1, and exchangeable bases (extraction with ammonium chloride) according to Van Reeuwijk (1992) method.Plant samples have been collected according to Witt et al. (1999) method.Paddy yield was estimated on 6 m² (3 m × 2 m) for each plot and grain moisture was determined using "Riceter grain moisture meter, Kett Electric Laboratory, Tokyo, Japan" device and paddy yield corrected to 14% moisture content.The nitrogen concentrations of grains were determined using 12 plants at physiological maturity.Samples were then dried in oven at 70°C until constant moisture of 3%.

Statistical analysis
Statistical analyses of data were conducted using Genstat 5 Release 4.1 (Payne, 1997).Paddy yield and nitrogen uptake data were submitted to a two way analysis of variance (ANOVA) with confidence interval of 95%.

Effect of fertilizers on paddy yield
The results show a non-significant interaction between varieties and fertilizers respectively for 2005 (P=0.113) and 2006 (P=0.601)(Table 3).This result allows the comparison of each treatment per factor (fertilizers and varieties).Analysis of variance showed significant differences between the yield of different nitrogen fertilizers (all varieties taking into account).For the 2005 wet season, five homogeneous groups are differentiated (P<0.001).
Plots fertilized with 160 kg N ha -1 get the best yield followed by plots fertilized with 120 kg N ha -1 . They got more than 3 to 5 t of paddy ha -1 compared to the plots fertilized with 80, 40 and 0 kg N ha -1 .At the 2006 wet season, three homogeneous groups are differentiated (Table 3).Plots fertilized with 160 kg N ha -1 got the best yield.This yield is significantly different with those of plots fertilized with 120 and 80 kg N ha

Relationship between varieties to N supply and absorption
The results showed a significant interaction between varieties and fertilizers respectively for 2005 (P<0.001) and 2006 (P<0.001)wet seasons (Figure 1).This result allowed the comparison of the combination of factors (fertilizers and varieties).For the 2005 wet season, the highest levels of nitrogen uptake are obtained with NERICA L 41, NERICA L 20 and NERICA L 19 fertilized with 120 or 160 kg N ha -1 (Figure 1).Lower levels of nitrogen uptake are obtained with the combination of all varieties with low doses of nitrogen supply and with the combination of 4418 and NERICA L 60 varieties with the same amounts of nitrogen supply having good results with other varieties.With 0N plots, it was still the NERICA L 41, NERICA L 20 and NERICA L 19 varieties which showed a good absorption of soil nitrogen.Like the 2005 wet season, the same trends are observed in 2006 with best nitrogen uptake for the combinations of NERICA L 41, NERICA L 20 and NERICA L 19 varieties (Figure 2).However, the level of absorption was higher than 2005.The soil nitrogen has been more profitable for these three varieties compared to 4418 and NERICA L60 varieties with lower levels of nitrogen absorbed.For all  varieties tested, N accumulation increased significantly with the availability of nitrogen.Analysis of N uptake showed that almost 97% of paddy yield variations are explained by N uptake at maturity (data not shown) for both seasons.For some varieties, most nitrogen was available, best are their response to N absorption.Maximum uptake is situated at about 100 kg N ha -1 for two varieties (4418 and NERICA L 60) while it is up to 170 kg N ha -1 for the three other NERICAs (L 20, L 19 and L 41).
The paddy yield evolved following genotypic behaviour of varieties to nitrogen uptake and use efficiency.Maximum N uptake for 4418 and NERICA L 60 varieties were 99 and 102 kg N ha -1 .However, highest uptake rates are observed with others NERICAs for nitrogen use like NERICA L 20, NERICA L 19 and NERICA L 41.For high rates of N uptake, high yields are observed for all varieties.For NERICA L 41, NERICA L 19 and NERICA L 20 varieties, lower N uptake rates corresponded to high yields.

Effect of fertilizers on paddy yield
The results of our study confirm those obtained by many others authors.Based on the response of rice according to nitrogen supply, the rice genotypes have been classified in " efficient"," not-efficient" and " inferior type" (Gerloff, 1976;Gourley et al., 1993;Shukla et al., 1998).
The genotypes which do not respond to the increasing doses of N supply because of their weak potential of adaptation to high doses or their weak productive potentials are identified as "inferior type"; the yield of these varieties are influenced by other factors than N availability; two of our varieties (4418 and NERICA L 60) could be identified as inferior type according to these criteria.Those which respond to lower N-Soil and Nfertilizers as well as to high level are identified as "efficient" varieties (Singh et al., 1998).Three of our varieties could be identified as efficient according to these criteria (NERICA L 41, NERICA L 19 and NERICA L 20).These varieties are more efficient for yield building.
The varieties which respond to lower N supply but not to increased level are "not efficient".In our case, none of varieties respond to this criterion.

Relationship between varieties to N supply and absorption
A study carried out during three successive years at the IRRI in Philippines (Singh et al., 1993) showed that when nitrogen does not constitute a constraint, the available nitrogen is the first factor limiting the paddy yield.The solar radiation accounts for less 5% in the observed yield gap without nitrogen.Without looking if available N comes from soil or fertilizer, the genotypic variation remains consistent.The yield gap without nitrogen supply is due to the genotypic variations in the capacity of the different varieties to absorb and to use the nitrogen.
The most efficient varieties are researched by the most part of rice breeders in Africa.Given the weak use of fertilizers by farmers (Nébié, 1995;N'Diaye et al., 1997;Wopereis et al., 1999;Haefele et al., 2002;Segda et al., 2004) such varieties appear to be most suitable.Besides, as these same varieties respond to variable levels of nitrogen, this criterion could interest those of the farmers who afford to use high levels of fertilizers.Three lowland NERICA genotypes (NERICA L 41, NERICA L 19 and NERICA L 20) were identified for high yield in both nonfertilized and fertilized conditions.This result indicates that interspecific breeding between O. sativa and O. glaberrima appears to offer an effective approach to increasing lowland rice productivity.This finding is consistent with Heuer et al. (2003), Oikeh et al. (2008), Rodenburg et al. (2009), andSaito et al. (2010a).Similar results were also reported from previous studies in flooded lowlands rice in Asia, Latin America and West Africa (Kawano, 1990;Peng et al., 2000;Saito et al., 2010b).During the last decades, many investigations have been reported, related to nitrogen efficiency of different plant genotypes (Fang and Wu, 2001;Fang et al., 2004, Zhao et al., 2012).Matsunami et al. (2013) indicated that substantial genotypic variation in N uptake ability under water deficient conditions exists among diverse rice genetic resources.Other field experiments have shown that genetic variability for N use efficiency exists in rice (Tirol-Padre et al., 1996;Singh et al., 1998;Inthapanya et al., 2000b).Therefore, plant breeders need to develop cultivars that can exploit N more efficiently, in order to minimize loss of N from the soil and make more economic use of the absorbed N, which could increase rice yield and improve environments.The main goals for rice production systems are to get more grain yield, to reduce the production cost and to minimize the pollution risk for the environment (Zhao et al., 2012).

Conclusions
The paddy yield increases significantly according to the quantity of nitrogen fertilizer applied.It evolves according to the genotypic behavior of the varieties in the absorption and the efficiency of nitrogen use.Three NERICA varieties (NERICA L 19, NERICA L 20 and especially NERICA L 41) can be considered as "efficient varieties" according to the use of nitrogen.
The NERICA varieties can create a major impact in improving incomes and food security of farmers.This may be the way to a real "green revolution" in Africa.However, it would be necessary to carry out complementary activities: (i) To assess the economic profitability of each input option, and (ii) To extend the study to all NERICA varieties diffused.
results show significant differences between varieties (all fertilizers taking into account) at 2005 (P<0.001) then at 2006 (P<0.001)wet seasons.For the 2005 wet season, two homogeneous groups are differentiated; the best varieties were NERICA L 41, NERICA L 20 and NERICA L 19.This group differs significantly with the group of 4418 and NERICA L 60 varieties.For the 2006 wet season, results show two heterogeneous groups; NERICA L 41 has got the best paddy yield.It differs significantly from 4418 and NERICA L 60 varieties.But it forms with NERICA L 20 and NERICA L 19, a heterogeneous group.

Figure 1 .
Figure1.Effect of fertilizers*variety interaction on nitrogen uptake in 2005 wet season.Average nitrogen uptakes followed by the same letter in the graph do not differ significantly by student Newman Keuls test at 5% level of significance.

Figure 2 .
Figure 2. Effect of fertilizers*variety interaction on nitrogen uptake in 2006 wet season.Average nitrogen uptakes followed by the same letter in the graph do not differ significantly by student Newman Keuls test at 5% level of significance.

Table 1 .
Application timing of fertilizers.

Table 2 .
Physical and chemical characteristics of soil.

Table 3 .
Effect of fertilizers and varieties on paddy yield in 2005 and 2006 wet seasons.
Average yields followed by the same letter in each column do not differ significantly by student Newman Keuls test at 5% level of significance.