Introduction

Rice is unquestionably the most widely grown cereal crop in Asia. Nitrogen fertilizer is an essential input in most rice soils to achieve high yields. Fertilizer N is relatively inexpensive, but deficiencies can result in substantial yield reductions. Farmers are inclined to manage fertilizer N to minimize the risk of deficiency, which can lead to excessive fertilizer application. The current blanket recommendations of split applications of N fertilizer with fixed rates at specific growth stages for large rice-growing areas assume a uniform requirement of rice for N fertilizer across large areas and years.

The requirement of rice for N fertilizer can vary greatly from field to field, season to season, and year to year because of high variability among fields, seasons and years in soil N-supplying capacity (Cassman et al. 1996; Dobermann et al. 2003) and crop growth due to differences in climate factors (Kropff et al. 1993). Nitrogen fertilizer management at the field level must, therefore, be responsive to these large variations to achieve high efficiency of N fertilizer use. Despite considerable research to increase N use efficiency in rice, the recovery efficiency of N fertilizer achieved by rice farmers ranges between 30 and 40% (Cassman et al. 1993; Bijay-Singh and Yadvinder-Singh 2003). Improving efficiency of fertilizer N use is vital to achieve and sustain high crop yields and reduce N loss that can potentially contaminate water resources. Fertilizer N is increasingly recognized as the source of nitrate contamination in the groundwater of the Indo-Gangetic plains (Bijay-Singh and Yadvinder-Singh 2004). Inefficient use of N fertilizer, particularly at high levels of yield and cropping intensity, can also result in crop lodging, increased disease and pest incidence, and reduced profitability for farmers. Soil tests are of limited value for predicting the variability in soil N supply because they do not account for the dynamics of soil N supply and for N inputs from other sources such as irrigation water or biological N fixation (Nayyar et al. 2006; Yadvinder-Singh et al. 2003).

One approach to increasing N fertilizer use efficiency is to achieve synchrony between N needs of the crop and N supply. Since plant growth reflects total N supply from all sources, plant N presumably is the best indicator of N availability to crops at any given time. In irrigated rice systems additional N fertilizer can be top dressed at the time of irrigation so that the applied N can rapidly be taken up by the crop to correct N deficiency. Measurement of leaf N concentration by conventional laboratory procedures has limited use as a diagnostic tool for optimizing N topdressing because of the long delay between sampling and obtaining results. Therefore, to synchronize N supply and crop demand, there is a need for a tool for instantaneously assessing crop N status. The Minolta chlorophyll meter (Model SPAD 502) provides a simple, quick, and nondestructive method of estimating relative leaf chlorophyll content (Watanabe et al. 1980). Leaf chlorophyll content shows a close relationship with leaf N content and thus provides an indirect estimate of leaf N content in rice (Takebe et al. 1990; Turner and Jund 1991; Peng et al. 1995). The chlorophyll meter has been successfully used as a tool in improved N management in rice (Bijay-Singh et al. 2002; Shukla et al. 2004). However, its high cost keeps it out of the hands of many farmers in Asian countries.

To support decision-making on the timing of N application in rice, the International Rice Research Institute (IRRI) introduced the use of a relatively inexpensive leaf color chart (LCC), a simple and a portable tool. Yang et al. (2000) have shown a linear relationship between an LCC reading and leaf N content. The LCC has shown great promise in optimizing N use in rice in different countries in Asia (Balasubramanian et al. 2003; Shukla et al. 2004; Alam et al. 2005). Farmers can easily use the LCC to qualitatively assess foliar N status and adjust N topdressing accordingly. While the LCC may be a valuable tool for on-farm decision-making, additional information is required for its optimal use. Although shade 4 on the LCC has been used in Asia (Bijay-Singh et al. 2002; Balasubramanian et al. 2003; Shukla et al. 2004; Alam et al. 2005) as the critical value for deciding N fertilizer application on the basis of the color of the first fully opened leaf of the rice plant from the top, Bijay-Singh et al. (2002) observed that the critical SPAD value could differ for rice cultivars. There was thus an urgent need to know if the critical LCC shade 4 was valid for different inbred cultivars of transplanted rice grown by farmers in northwestern India. Fertilizer N applications guided by the LCC are initiated 2 weeks after transplanting of rice. While application of a basal dose of N at the transplanting of rice constitutes an integral part of the blanket recommendation in the region, not enough is known regarding applying a basal dose of N to rice when fertilizer N applications during the rice season are guided by the LCC. The objectives of this investigation were to determine the critical color shade on the LCC for guiding fertilizer N application to rice, evaluate LCC as an on-farm decision-making aid for applying N fertilizer to different cultivars of transplanted rice grown in the Indo-Gangetic plains in northwestern India, and to assess the need for a basal N application in LCC-based N management in rice.

Materials and methods

The field experiments at 100 on-farm locations were conducted between 2000 and 2003 to study N management in irrigated rice (June–October) using the LCC. Experimental sites during 2000, 2001, and 2002 were located in different villages in the Ludhiana and Gurdaspur districts of the Punjab state in northwestern India. In 2003, the experiments were located in six districts (Amritsar, Patiala, Gurdaspur, Jalandhar, Ferozepur, and Ropar) of the Punjab. Within each village, the fields were located in a radius of 1–2 km, and within each district the villages were located in a radius of about 10 km. The number of experiments varied among villages and years. The overall number of experiments conducted during different years were: 8 in 2000, 8 in 2001, 36 (11 and 25 in two sets) in 2002, and 48 in 2003.

Experimental area

The sites were representative of the loamy sand to clay loam soils commonly used for rice production. Each year, new field sites were selected for experimentation. The climate of the experimental sites in Ludhiana (30°56′ N longitude, 75°52′ E, and altitude 247 m above mean sea level) is subtropical, semi-arid with an annual rainfall of 733 mm, of which about 80% is received during June to September. The mean monthly temperatures during the rice season (June–September) varied from 29.3 to 33.8°C. The climate of Gurdaspur region (32°02′ N longitude, 75°24′ E) is sub-humid and semi-arid. Annual rainfall of the region is 887 mm, about 80% of which occurs from June to September. The mean monthly temperatures during the rice season (June–September) varied from 28.1 to 33.0°C. The mean monthly temperatures during the rice season in five districts of Punjab other than Ludhiana and Gurdaspur where experiments were located during 2003 varied from 29.0 to 33.7°C. Annual rainfall ranged from as low as 427 mm in Ferozepur to as high as 844 mm in Ropar.

Treatments

The LCC developed by the IRRI consisting of six green shades, ranging from yellowish green to dark green, was used in this study. LCC readings were taken at weekly intervals starting at 14 days after transplanting until onset of flowering from the uppermost fully expanded leaf of 10 randomly selected healthy rice hills in each plot. When the LCC value of more than 5 out of 10 leaves fell below the critical level (4 or 5 as per the treatment), 30 kg N ha−1 was topdressed on the same day.

The treatments during 2000 and 2001 included:

  1. 1.

    No N, control

  2. 2.

    Recommended N (120 kg ha−1) in three equal splits

  3. 3.

    LCC with a critical value of 4, LCC 4 (no basal N)

  4. 4.

    LCC 4 (with 20 kg N ha−1 basal dose)

  5. 5.

    LCC with a critical value of 5, LCC 5 (no basal N)

  6. 6.

    LCC 5 (with 20 kg N ha−1 basal dose)

Two sets of experiments were conducted during 2002. In the first set, 11 experiments were conducted with 4 treatments:

  1. 1.

    No N, control

  2. 2.

    Farmers’ N management

  3. 3.

    LCC 4 (no basal N)

  4. 4.

    LCC 4 (with 20 kg N ha−1 basal dose)

In the second set, 25 experiments were conducted and the treatments included:

  1. 1.

    No N, control

  2. 2.

    Recommended N (120 kg ha−1) in three equal splits

  3. 3.

    LCC 4 (with 20 kg N ha−1 basal dose)

The experimental plots were laid out in a randomized complete block design with two replicates for each treatment. Each experimental plot was 100 to 200 m2 in size.

In 2003, LCC-based N fertilizer management was evaluated for seven inbred rice cultivars (PR 108, PR 114, PR 116, PR 111, PR 118, Pusa 44, and Indrasan) commonly grown by farmers in the Indian Punjab. Forty-eight field experiments with two treatments (farmers’ N management practice and need-based N management with LCC using a critical value of 4 (with 20 kg N ha−1 basal N) were conducted in different villages throughout the state of Punjab. Each experimental plot was 1,500–2,000 m2 in size.

Crop management

A basal dose of 26 kg P ha−1, 50 kg K ha−1, and 10 kg Zn ha−1 was applied to all the plots at the time of transplanting. Crop management from land preparation to harvesting was identical in all treatments and were based on recommendations from Punjab Agricultural University, Ludhiana (Anonymous 2005). Land was prepared by two ploughings and two harrowings. Fields were leveled and puddled for transplanting, and 25- to 30-day-old seedlings were randomly transplanted at 25–30 seedlings m−2 during the second to fourth week of June each year. Cultivar PR 114 was planted during 2000 and 2001 and PR 116 during 2002 in field experiments at Ludhiana and PR 108 in field experiments at Gurdaspur. Rice was harvested between the second and fourth week of October every year. Grain and straw yields were determined by harvesting net areas of 15 m2 in the center of each plot.

Composite soil samples were collected from depths of 0 to 15 cm in each field before rice transplanting in each year. Samples from each field were mixed, air-dried, crushed, sieved, and analyzed for pH, electrical conductivity, organic carbon, Olsen’s available P, and ammonium acetate extractable available K. Grain moisture content was measured in all samples after drying to a constant weight at 60°C in an oven. Grain yields are reported as having 14% moisture content. Grain and straw samples were analyzed for total N content to calculate total plant N uptake in each treatment.

Agronomic efficiency (AEN) and apparent recovery efficiency (REN) of added fertilizer N were calculated as:

$$ {\hbox{AE}}_{\hbox{N}} {\hbox{(k}}{\hbox{g}} \,{\hbox{grain}}\,{\hbox{kg}}^{{\hbox{ - 1}}} \,{\hbox{N}}\,{\hbox{applied) = }}\frac{{{\hbox{Grain}}\,{\hbox{yield}}\,{\hbox{(kg ha}}^{{\hbox{- 1}}} {\hbox{)}}\,{\hbox{in}}\,{\hbox{N}}\,{\hbox{fertilized}}\,{\hbox{plot}}\,{\hbox{ - }}\,{\hbox{Grain}}\,{\hbox{yield}}\,{\hbox{(kg ha}}^{{\hbox{ - 1}}} {\hbox{)}}\,{\hbox{in}}\,{\hbox{no N}}\,{\hbox{plot}}}} {{{\hbox{Quantity}}\,{\hbox{of}}\,{\hbox{fertilized}}\,{\hbox{N}}\,{\hbox{applied}}\,{\hbox{(kg}}\,{\hbox{ha}}^{{\hbox{ - 1}}} {\hbox{)}}\,{\hbox{in}}\,{\hbox{N}}\,{\hbox{ fertilized}}\,{\hbox{plot}}}} $$
$${\hbox{AE}}_{\hbox{N}} {\hbox{(k}}{\hbox{g}}\,{\hbox{grain}}\,{\hbox{kg}}^{{\hbox{ - 1}}}\,{\hbox{N}}\,{\hbox{applied) =}}\frac{{{\hbox{Grain}}\,{\hbox{yield}}\,{\hbox{(kg ha}}^{{\hbox{- 1}}} {\hbox{)}}\,{\hbox{in}}\,{\hbox{N}}\,{\hbox{fertilized}}\,{\hbox{plot}}\,{\hbox{ -}}\,{\hbox{Grain}}\,{\hbox{yield}}\,{\hbox{(kg ha}}^{{\hbox{ - 1}}}{\hbox{)}}\,{\hbox{in}}\,{\hbox{no N}}\,{\hbox{plot}}}}{{{\hbox{Quantity}}\,{\hbox{of}}\,{\hbox{fertilized}}\,{\hbox{N}}\,{\hbox{applied}}\,{\hbox{(kg}}\,{\hbox{ha}}^{{\hbox{- 1}}} {\hbox{)}}\,{\hbox{in}}\,{\hbox{N}}\,{\hbox{fertilized}}\,{\hbox{plot}}}} $$

Partial factor productivity (PFPN) from applied N fertilizer was calculated as (Cassman et al. 1998):

$$ {\hbox{PFP}}_{\hbox{N}} \,{\hbox{(kg}}\,{\hbox{grain}}\,{\hbox{kg}}^{{\hbox{ - 1}}} \,{\hbox{N}}\,{\hbox{applied) = }}\frac{{{\hbox{Grain}}\,{\hbox{yield}}\,{\hbox{(kg}}\,{\hbox{ha}}^{{\hbox{ - 1}}} {\hbox{)}}\,{\hbox{in}}\,{\hbox{N}}\,{\hbox{fertilized}}\,{\hbox{plot}}}} {{{\hbox{Quantity}}\,{\hbox{of}}\,{\hbox{fertilizer}}\,{\hbox{N}}\,{\hbox{applied}}\,{\hbox{N}}\,\,{\hbox{(kg}}\,{\hbox{N}}\,{\hbox{ha}}^{{\hbox{ - 1}}} {\hbox{)}}}} $$

Analysis of variance (ANOVA) was performed using IRRISTAT version 5.0. The fields within each village were considered replications. Least significance difference (LSD) at a 0.05 level of probability was used to test the significance of differences among treatment means. Descriptive statistics such as means and standard deviation were used to determine the variability of different parameters.

Results and discussion

Initial soil fertility status of farmers’ fields

The soils of the experimental fields varied from loamy sand to clay loam in texture (Table 1). Soil pH ranged from 6.0 to 9.1, the lower values being for fields from Gurdaspur. There was a wide range of organic carbon content of between 0.21 and 0.76%, suggesting a large variation in indigenous soil N supply (Nayyar et al. 2006). Available P (Olsen-P) ranged from 2.7 to 37.0 mg kg−1 soil, classified as the medium to high range. Available K ranged between 28.1 and 135.3 mg kg−1, classified as the low to high range. Total N uptake at maturity in grain and straw by rice on control plots ranged from 41 to 150 kg ha−1 across all villages and years (data not presented). These data suggest a large field-to-field variation in indigenous N supply (INS) to rice. INS represents N available from all sources (soil organic matter, organic materials, crop residues, rhizosphere N fixation, irrigation water, rainfall, etc.) to rice during the growing season (Dobermann et al. 2003).

Table 1 Some important properties of soil at depths of 0 to 15 cm in farmers’ fields at different on-farm locations in Punjab, India
Full size table

Critical LCC value

Critical or threshold value of the LCC is defined as the intensity of green color that must be maintained in the uppermost fully opened leaf of the rice plant and fertilizer N needs to be applied whenever leaf greenness is below the critical LCC value. Leaf greenness or leaf N content is closely related to photosynthesis rate and biomass production and is a sensitive indicator of changes in crop N demand during the growing season. Thus, maintaining the leaf greenness just above the LCC critical value ensures high yields with need-based N applications thereby leading to high fertilizer N use efficiency. Grain yields of rice were generally similar under LCC 4 and LCC 5, except at Dakha village in the Ludhiana district during 2000 when rice yield was significantly higher for LCC 5 than LCC 4. Total N applied using the critical LCC value of 5 was always higher (ranging from 15 to 45 kg N ha−1) than using LCC 4. Nitrogen use efficiency (NUE; AEN, REN, and PFPN) was always higher for LCC 4 than for LCC 5 (see Table 6). It suggests that a critical LCC value of 4 is better than LCC 5 for need-based N management in transplanted rice. Shukla et al. (2004) reported a threshold LCC value of 4 for an inbred line (Saket 4) for an optimal yield and NUE in the western Indo-Gangetic plains of India, while Alam et al. (2005) reported a further increase in NUE in rice in Bangladesh when the critical value for LCC was reduced from 4 to 3.5.

Basal nitrogen application with leaf color chart-guided N management

In most of the experiments between 2000 and 2002, when need-based N management was followed using either 4 or 5 as the critical LCC value combined with a basal application of 20 kg N ha−1, grain yields were similar to those of treatments with no basal application (Tables 2–5). Only in 2 out of 27 experiments, grain yields in LCC 4 with basal N were significantly higher than LCC 4 without basal N. Since mean yield of rice in plots receiving no N fertilizer in the on-farm experiments conducted between 2000 and 2002 was more than 4 t ha−1, no response to basal N application was expected. Balasubramanian et al. (1999) and Bijay-Singh et al. (2002) reported that soils producing a grain yield of ≥3 t ha−1 without fertilizer application did not need basal N, as basal N application was not efficiently used by the crop and possibly prone to losses. According to Meelu and Gupta (1980), N uptake by rice within 2 weeks of transplanting is likely to be very small, because rice seedlings need about 7–8 days to recover from transplanting shock. Shukla et al. (2004) found that NUE can be increased using LCC-based N management without basal N application, provided indigenous soil N supply is sufficiently high (50–60 kg N ha−1).

Table 2 Grain yield and amount of fertilizer N applied in leaf color chart (LCC)-guided fertilizer N management and recommended N fertilizer management (RN) practices during 2000 at on-farm locations in Punjab, India
Full size table
Table 3 Grain yield and amount of fertilizer N applied in LCC-guided fertilizer N management and RN practices during 2001 at on-farm locations in Punjab, India
Full size table
Table 4 Grain yield and amount of fertilizer N applied in LCC-guided fertilizer N management and farmers’ N fertilizer management (FP) practices during 2002 at on-farm locations in Punjab, India
Full size table
Table 5 Grain yield and amount of fertilizer N applied in LCC-guided fertilizer N management and RN practices during 2002 at on-farm locations in Punjab, India
Full size table

Comparison of LCC-guided nitrogen management with recommended three splits and farmers’ practice

Application of N fertilizer using LCC with a critical value of 4 produced rice yields equal to those obtained with the recommended fixed-time N schedule or farmers’ fixed-time schedule from 2000 to 2002 (Tables 2–5). Mean grain yields (pooled across all sites) of 6.53, 7.10, and 6.93 t ha−1 under the recommended fixed-time N schedule and 6.61, 7.04, and 7.12 t ha−1 in the LCC 4 (irrespective of basal N) treatment were almost similar during 2000, 2001, and 2002 respectively. Rice yield at the village of Hissowal under LCC 4 with basal N applied was 17.4% higher in 2000, 7.8% in 2001, and 6.9% in 2002 than the recommended N or farmers’ practice, while at Gurdaspur in 2002 yields were 7.2% and 22.9% higher than farmers’ practice and recommended N fertilizer respectively (Tables 4, 5). In 2001, LCC-based N management and the recommended N fixed-time schedule resulted in similar grain yields in 3 out of 4 locations (Table 3).

Using LCC-based N fertilizer management, higher rice yields at lower N rates were achieved than with the recommended N fertilizer treatment or farmers’ practice. The amount of N used was much lower in LCC plots (mean 93 kg N ha−1 in 2000, 95 kg N ha−1 in 2001, and 80 kg N ha−1 in 2002), less than 20.8 to 33.3% than the 120 kg N ha−1 in the recommended N fertilizer practice (Tables 2–4). The higher grain yields resulted from improved synchrony of fertilizer N supply and the crop’s needs. The use of LCC with a critical value of 5, with or without basal N application resulted in a significantly higher yield than the recommended N treatment at different villages in Ludhiana over the years, similar to the grain yield obtained using LCC with a critical value of 4 with basal N application or that with LCC 5 with or without basal N application in the majority of the locations (Tables 2–4). Thus, there was hardly any yield benefit of applying a basal dose of N along with LCC-based N management. Earlier studies by Shukla et al. (2004) and Alam et al. (2005) showed that N applied based on crop need as determined by LCC was used more efficiently. Alam et al. (2005, 2006), Bijay-Singh et al. (2002) and Shukla et al. (2004) proved that the current recommendation of three split applications at specified growth stages is not adequate to synchronize N supply with crop N demand.

The interactive effect of villages and treatments on NUE was not significant; pooled data are therefore provided in Table 6. Nitrogen use efficiencies were markedly higher in the LCC 4 treatment, with or without basal N application than recommended N (120 kg ha−1) or farmers’ practice in all the years (Table 6). The mean AEN across villages and years in LCC 4 ranged from 16.4 to 28.1 kg grain/kg N compared with 11.3 to 20.8 kg grain/kg N in recommended N fertilizer or farmers’ practice (Table 6). Mean recovery efficiency of N under LCC (with 20 kg N ha−1 as basal dose) was 42.1, 45.4, and 52.7% compared with 30.9, 29.1, and 39.8% obtained in the recommended N treatment or farmers’ practice in 2000, 2001, and 2002 respectively. Mean PFPN was also significantly higher under LCC-based N management than under recommended N or farmers’ practice over all years, except under LCC 5 in 2001 (Table 6). Mean PFPN with LCC (20 kg N ha−1 as basal dose) was 86.3, 90.5, and 73.8 kg grain/kg N compared with 57.2, 59.9, and 52.1 kg grain/kg N obtained in the recommended N or farmers’ practice in 2000, 2001, and 2002 respectively (Tables 5, 6). The increase in NUE through the use of LCC resulted from better synchronization of timing of fertilizer N applications and the crop’s needs for N fertilizer. N fertilizer was reduced through the use of LCC by 10 to 45 kg ha−1 in 2000, 0 to 60 in 2001 (with LCC 4), and 40 to 82 in 2002. Shukla et al. (2004) and Alam et al. (2005) have observed not only higher NUE, but also higher grain yields through LCC-based N management.

Table 6 Agronomic efficiency (AEN), recovery efficiency (REN), and partial factor of production (PEPN) of applied fertilizer N in rice as influenced by N application method at on-farm locations in Punjab, India
Full size table

Leaf Color Chart for N management for different rice cultivars

Lead color chart-guided N management with no other change in farmers’ crop management practices in different rice cultivars resulted in a total N fertilizer application ranging from 80 to 106 kg N ha−1 compared with 108 to 123 kg N ha−1 in farmers’ practice of applying N in fixed-time three split doses (Table 7). Grain yields of all rice cultivars were similar (mean yield of 6.52 t ha−1 versus 6.50 t ha−1) under the two practices of applying N fertilizer. The LCC-guided N management thus resulted in savings of 16 to 43 kg N ha−1 (24 kg N ha−1, averaged across 48 experiments) compared with farmers’ practice without any reduction in grain yield. PFPN was markedly higher (71 kg grain/kg N applied, averaged across different cultivars) with LCC-guided N management than the farmers’ practice (56.5 kg grain/kg N).

Table 7 Evaluation of LCC-based N management for different rice varieties at on-farm locations in Punjab, India in 2003
Full size table

Conclusions

Nitrogen use efficiency can be significantly improved by reducing the amount of fertilizer N applied without sacrificing grain yield of irrigated rice through need-based N management using LCC in the Indo-Gangetic plains of northwestern India. LCC-based N management assures optimal rice yields consistent with efficient N use and enhanced farmers’ profits due to the saving in the use of N fertilizers. LCC is a simple, cheap, and easy-to-use tool that can help farmers manage N judiciously. The critical LCC value of 4 with 20 kg N ha−1 as basal dose at transplanting can be used for all the inbred lines presently popular with the farmers in the region. Basal application of fertilizer N can be avoided if yields without applying fertilizer N in a field are more than 3 t ha−1.