Effect of Real time nitrogen management on yield attributes of direct seeded rice by using SPAD meter and Greenseeker

A field experiment was conducted during Kharif 2018, laid out in Randomized Block Design with three replications having seven treatments viz . N omission (T 1 ), N applied as basal and AT (T 2 ), N as basal, AT and PI (T 3 ), N as basal and top dressing at NDVI threshold of 0.75 (T 4 ), at NDVI threshold of 0.8 (T 5 ), at SPAD threshold of 35.0 (T 6 ) and SPAD threshold of 37.5 (T 7 ) with Rice variety Sahabhagidhan.The study revealed that application of 30 kg N/ha as basal dose and top dressing of 20 kg N/ha twice at 35 and 63 DAS guided by NDVI threshold value of 0.8 (T 5 ) was found to be superior over other treatments with respect to productivity. T 5 recorded highest grain yield of 4438 kg/ha which was 17.0% higher than that top dressed at NDVI threshold of 0.75 (T 4 ) and 7.1% higher than that top dressed at SPAD threshold value of 37.5 (T 7 ). In case of SPAD meter, nitrogen top dressed at threshold value of 37.5 (T 7 ) produced grain yield of 4143 kg/ha which was 15.0% higher than T 6 . T 5 produced maximum dry matter of 8678 kg/ha with highest grain yield (4438 kg/ha), straw yield (5092 kg/ha) and harvest index 46.0%.


INTRODUCTION
agriculture need to be addressed for better correlation of crop yields and also for their wider acceptability. It is suggested that different SPAD threshold and NDVI threshold values might to be used for different varietal groups and agro climatic conditions. (Balasubramarian et al., 2002 andVarinderpal Singh et al., 2010) Keeping this in view to enhance fertilizer nitrogen use efficiency and reduce fertilizer nitrogen related environmental pollution, the current study on "Real time nitrogen management in direct seeded rainfed rice using GreenSeeker and SPAD meter" has been contemplated to determine the critical thresholds of GreenSeeker (0.75 and 0.8) and SPAD (37 and 37.5) for rice crop.

MATERIALS AND METHODS
The experiment was conducted in the Instructional Farm of the College of Agriculture, Odisha University of Agriculture and Technology, Bhubaneswar during 2018.The field experiment was laid out in Randomized Block Design with three replications and seven treatments , gross plot size 7m×5m, net plot size 6m×4m.The Rice variety Sahabhagidhan

Sampling technique
Two rows of one meter length from 3 different location of each plot were randomly selected and observations on plants height and number of tillers were recorded. Destructive samples were drawn from the third row form the net plot on each side of the plots. Observation on leaf area and dry matter accumulation were recorded from continious clumps from 0.5 m length from four spots. Border rows were not sampled to avoid border effects.

PRE-HARVEST STUDY
Plant height,tiller number,leaf area Index,Crop Growth Rate are carried out according to SES

Measurement of spectral properties of leaves
Spectral properties of rice leaves were measured using SPAD 502 chlorophyll meter and GreenSeeker optical sensor. Spectral properties were measured at 7 days interval starting from 28 DAS to initiation of flowering. GreenSeeker and SPAD meter readings were recorded on 4, 11, 18 and 25 of Aug, 02, 09, 16, 23 of Sept.

SPAD 502 chlorophyll meter measurement
Chlorophyll meter are reliable alternatives to traditional tissue analysis as plant N nutritional diagnostic tools. These instantly provide an estimate of leaf N status as chlorophyll content by clamping the unplucked leafy tissue in the meter. It has two LEDs (light emitting diodes) which emitted light with a peak wavelength of 650 nm and an infrared radiation peak wavelength of 940 nm. The red and infrared radiation are made to pass through the leaf. A portion of light is absorbed and the remainder is transmitted through the leaf and silicon photodiode detector converts it into an electrical signal. The amount of light reaching the photodiode detector is inversely proportional to the amount of chlorophyll in the path of the light. Leaf chlorophyll content is displayed in arbitrary units (0-99.9). The SPAD meter reading are unit less and need to be calibrated with chlorophyll or N content and leaf greenness. The SPAD meter is the one tool that enables to determine the relative amount of chlorophyll content by measuring leaf greenness (Peterson et al., 1993) and the linear relationship of SPAD values and N status in crops varies depending on the growth stages and cultivars. Since chlorophyll content is usually strongly related to N concentration, SPAD meters can be used as indicator of need for N application (Schepers et al.,1992, Blackmer andSchepers, 1995). The top most fully expanded leaf is usually considered as the index leaf to reflect N status of the plant. The SPAD meter uses a silicon photodiode to derive the ratio of transmittance through the leaf tissue at 650 nm compared with transmittance at 940 nm and a value is given based on that ratio.
SPAD meter reading were taken with a Minolta 502 Chlorophyll meter starting from 28 DAS.
Before taking readings, the meter was calibrated by switching on the meter with no leaf sample slot, followed by complete closing the measuring head until a beep sound was heard. The meter was thus calibrated and ready for measurements.
SPAD meter reading were recorded from the uppermost fully expanded leaf of 5 randomly selected plants at 7 days interval till the flowering. The receiving window of the chlorophyll meter was cleaned with a soft cloth before taking measurements. The extremely thick, disease affected and insect bitten leaves were avoided for measurement. The data was recorded when leaves were dry. The selected leaf was inserted into sample slot of measuring head of chlorophyll meter to make sure that the sample was completely covered by the receiving window. The measuring head was completely closed until a beep sound was heard which indicated that measurement was over and the value appeared on the display. The chlorophyll reading was automatically saved in the meter. The head was opened and the reading for next leaf was taken on the same way. The average of all the SPAD readings for a plot was the noted.
After noting SPAD value of the plot, the meter was cleared for the next set of reading by pressing the "all data clear" button. The old readings were deleted by pressing the button "one data clear" before taking average. bands while reflecting light in the green band (Campbell 2002). The amount of light reflected in the visible region is defined by the chlorophyll content in the cell and the amount of light reflected in the near infrared (NIR) region is defined by living vegetation or biomass. The amount of blue and red light absorbed by the leaf is proportional to the chlorophyll density of the leaf. Reflectance of the NIR portion of the electromagnetic spectrum (720-1300 nm) is predominantly influenced by the mesophyll cells. The upper layers of the leaf are nearly transparent to NIR energy. Mesophyll tissue scatters and reflects as much as 60 per cent of all incident NIR radiation. The degree to which near infrared energy is reflected depends on the structure of the mesophyll cells and cavities between these cells (Campbell 2002). Several researchers have assessed N status and other physiological parameters of field crops using optical sensors (Zhao et al., 2017). The GreenSeeker TM hand held optical sensor unit Model 505 was used to measure NDVI from the crop canopy. Before taking the readings, battery was charged properly. Then shoulder strap was put around the body and sensor angle were such adjusted that it was parallel to sensing area at a height of about 70 cm above the crop canopy.

GreenSeeker optical sensor
The trigger of GreenSeeker optical sensor was pressed continuously. while moving in the two middle crop rows and trigger was released after completing one plot. After that note down the value which was displayed on the machine. Normalized Difference Vegetation Index (NDVI) measurement made by GreenSeeker computed by the following equation.

Number of effective tillers
The data pertaining to tiller production is presented in the Table 2. The production of effective tillers was influenced by different nitrogen application schedules. It ranged from 183/m 2 to 315/m 2 in different treatments.
The highest number of effective tillers was recorded in T5 (315/m 2 ) where nitrogen top dressing was guided by NDVI threshold of 0.8 which was at par with T7 (312/m 2 ) where it was guided by SPAD threshold of 37.5, but significantly higher than T4 (303/m 2 ) and T6 (229/m 2 ) where it was guided by NDVI threshold of 0.75 and SPAD threshold of 35.0, respectively. It was also higher than the treatments with blanket recommendation. However, no significant difference was obtained between T3 (280/m 2 ) where nitrogen top dressed twice at AT and PI and T2 (271/m 2 ) where entire nitrogen was top dressed only at AT, respectively. The number of effective tiller was the lowest in T1 (183/m 2 ) where no nitrogen was applied.

Grain production
The data on production of grains per panicle as influenced by nitrogen application schedules are presented in Table 2 (103.7) where it was guided by SPAD threshold of 37.5, NDVI threshold of 0.75 and SPAD threshold of 35.0, respectively, but higher than that under blanket top dressings. However, no significant difference was obtained between T3 (86.0) where nitrogen was top dressed at AT and PI and T2 (82.6) entire nitrogen was top dressed only once at AT. But the grain production was markedly reduced in T1 (72.1) which was grown without nitrogen application.

Filled grains per panicle
The data on production of filled grains per panicle as influenced by nitrogen application schedules are presented in than that under blanket recommendation. No significant difference between T3 (78.9) when nitrogen was top dressed either twice at AT and PI and T2 (74.9) where it was top dressed only once at AT. However, T1 (60.6) recorded the lowest filled grain per panicle, where no nitrogen was supplied.

Sterility
Sterility of grains differed with different nitrogen application schedules. The data related to sterility per cent are presented in Table 2. It ranged from 6.7% to 15.9% in different nitrogen application schedules.
The lowest sterility was recorded in T5 (6.7%) where nitrogen top dressing was guided by NDVI threshold of 0.8 which was significantly lower than rest of the treatments. Sterility obtained where T7 (7.4%) where nitrogen top dressing was guided by SPAD threshold of 37.5, T4 (7.5%) where it was guided by NDVI threshold of 0.75 and T6 (7.8%) where it was guided by SPAD threshold of 35.0, were at par. The sterility in T3 (13%) where nitrogen was top dressed in two splits at AT and PI was significantly lower than T2 (14.8%) where it was top dressed only at AT. However maximum sterility was recorded in T1 (15.9 %) where no nitrogen was applied.

Test weight
Data recorded on 1000-grain weight are presented in Table 2.. It differed significantly among different nitrogen application schedules. Test weight of grains ranged from 23.8 g to 18.1 g in different treatments.
Highest test weight was recorded in T5 (23.8 g) where nitrogen top dressing was guided by NDVI threshold of 0.8 which was at par with T7 (23.5 g) where nitrogen top dressing was guided by SPAD threshold 37.5, but was higher than T4 (23.1 g) and T6 (23.1 g) where it was guided by NDVI threshold of 0.75 and SPAD threshold of 35.0, respectively. These were also significantly higher than T3 (22.6 g) where nitrogen was top dressed at AT and PI and T2 (21.6 g) where entire nitrogen top dressed only at AT, who were at par with each other. However, the test weight was the lowest in T1 (18.1 g) with nitrogen omission.

Grain yield
The data pertaining to grain yield as influenced by different nitrogen application schedules are presented in Table 2.. Different nitrogen application schedules caused significant variation in grain yield in all the treatments.
Grain yield ranged from 1509 kg/ha to 4438 kg/ha in different treatments with an average yield of 3358 kg/ha. Maximum grain yield of 4438 kg/ha was obtained in T5 where nitrogen top dressing was guided by NDVI threshold of 0.8. This was significantly higher than T4 (4712 kg/ha) with NDVI threshold of 0.75 and also higher than T7 and T6 where nitrogen top dressing was guided by SPAD threshold of 37.5 and 35.0, respectively. Yield obtained under SPAD threshold of 37.5 (4143 kg/ha) was also significantly higher than SPAD threshold of 35.0 (3385 kg/ha).
The grain yield obtained by optical sensor guided nitrogen application was higher than that fixed applications. A grain yield of 4155 kg/ha was recorded with blanket top dressing at AT and PI (T3) which was significantly higher than T2 (3428 kg/ha) with only one top dressing at AT. However, the yield obtained with nitrogen omission (T1) was the lowest (1887 kg/ha).

Straw yield
The data pertaining to straw yield as influenced by different nitrogen application schedules are presented in Table 2. Different nitrogen application schedules caused significant variation in straw yield in all the treatments.
Straw yield ranged from 2000 kg/ha to 5092 kg/ha in different treatments with an average yield of 3950 kg/ha. Maximum straw yield was obtained in T5 (5092 kg/ha) where nitrogen top dressing was guided by NDVI threshold of 0.8 which was at par with T7 (5900 kg/ha) where it was guided by SPAD threshold of 37.5, but was significantly higher than T4 (5433 kg/ha) with NDVI threshold of 0.75 and T6 (5100 kg/ha) with SPAD threshold of 35.0. Straw yield obtained under SPAD threshold of 37.5 was also significantly higher than SPAD threshold of 35.0.
The straw yield obtained by optical sensor guided nitrogen application was higher than that under fixed applications. A straw yield of 4800 kg/ha was recorded with blanket top dressing of nitrogen twice at AT and PI (T3) which was significantly higher than T2 (4266 kg/ha) with only one top dressing at AT. However, the yield obtained with nitrogen omission (T1) was the lowest (2500 kg/ha).

Harvest index
Data on harvest index (HI) of rice grown under different nitrogen application schedules are presented in Table 2.

DISCUSSION
Grain yield of cereals is highly dependent upon the number of effective tillers produced by each plant as yield is function of number of ear bearing tillers per unit area (Power andAlessi, 1978 andNerson, 1980). Lee et al. (2010) stated that effective panicle number is considered to be the main factor for determining yield because the effect of increase in panicle number per unit area is similar to that of increasing rice yield. When top dressing of nitrogen was guided by NDVI threshold value T5 (threshold of 0.8) more number of effective tillers (315/m 2 ) were produced which were 3.9% higher than T4 (threshold of 0.75). When nitrogen top dressing was guided by SPAD threshold value T7 (threshold of 37.5) more number of effective tillers (312/m 2 ) were produced which were 4.1% higher than T6 (threshold of 35.0). But when fixed time schedules were followed, top dressing twice at AT and PI (T3) produced 280 effective tillers/m 2 which were 3.3% higher than that top dressed once at AT (T2). However, the maximum number of effective tillers were produced under T5 which was 12% higher than T3. However, the lowest effective tillers were obtained in T1 (193/m 2 ) with nitrogen omission. Since SPAD readings are affected by leaf water content and irradiance, SPAD measurements might change throughout the day in parallel with the daily changes in photon flux density and leaf water status. The progressive improvement in the in formation of tillers might be due to increase in nitrogen use efficiency which enhanced tillering. Adequacy of nitrogen probably favoured the cellular activities during panicle formation and development which led to increased number of productive tillers. The reason for more number of effective tillers might be due to higher LAI resulting increased photosynthetic ability and dry matter accumulation and reduced tiller mortality thereby increasing the proportion of effective tillers to total tillers.
The maximum lengths of panicles were recorded under T5 which were 3.2% higher than that of T7 and 11% higher than T3. The lowest sterility was recorded in T5 (6.7%) where nitrogen top dressing was guided by NDVI threshold of 0.8. It was 8.0% lower than T4 (threshold of 0.75). In T7 (threshold of 37.5) 7.7% sterility was noticed which was 7.8% lower than T6 (threshold of 35.0). But when fixed time schedules were followed, topdressing twice at AT and PI (T3) recorded 13.0% sterility which was 12.1 % lower than that top dressed once at AT (T2).
The total number of grains per panicle (filled grains) is an important parameter which contributes towards grain yield. Hasegawa et al. (1994) indicated that increased number of spikelets and vigorous growth of rice due to high rates of N fertilizer application induced competition for carbohydrate available for grain filling and spikelet formation. This was in agreement with the findings of Channabasavanna and Setty (1994). Blanket recommendation of nitrogen top dressing at AT and PI (T3) produced 74.9 filled grain per panicle, which was 6.5% higher than that under T2 where it was top dressed only once AT. The overall number of filled grains obtained under T5 was 4.5% higher than T7 and 26% higher than T3. This might be due to adequate nitrogen supply which resulted in higher photosynthate accumulation and translocation of these photosynthates to the reproductive part there by increasing the number of Fageria et al. (2010) noticed that spikelet sterility accounted for 7% variation in yield and 3% variation in test weight.
Grain weight, an important yield determining attribute, is a genetic character least influenced by environment. However favourable environment at grain filling stage ensured higher grain weight. Higher test weight was recorded in T5 (23.8g) followed by T7 (23.5 g), T4 (23.1 g) and T6 (23.1 g) which was higher than blanket recommendation T3 (22.6 g) and T2 (21.6 g). The increase in test weight in optical sensor guided treatments might be due to proper time of nitrogen application which increased photosynthetically active period and translocated the photosynthates more efficiently to the grains for a longer period extended with duration resulting in well filled grains.
The grain yield expression is basically a function of absolute infrastructure and seed development activity of plant (Pandey et al., 1991). It is also a cumulative effect of several growth regulating factors consisting mainly of genetic, environmental and management aspects dovetailed one into other to meet the optimum need of the crop at different stages of growth. According to Dobbemann and Fairhurst (2000) top dressing with split application of nitrogen is needed when the crop has a great need for it and when the rate of nitrogen uptake is large. When nitrogen top dressing was guided by NDVI threshold of 0.8, maximum yield was obtained in T5 (4438 kg/ha) which was 17% higher than T4 (NDVI threshold of 0.75). When nitrogen top dressing was guided by SPAD threshold of 37.5, it recorded a higher yield of 5179 kg/ha (T7) which was 15% higher than T6 (SPAD threshold of 35). But when fixed time schedules were followed top dressing twice at AT and PI (T3) recorded yield of 3324 kg/ha which was 20% higher than that top dressed only once at AT (T2). However, the highest yield recorded under T5 was 7.1% higher than that of T7 and 30% higher than T3. The lowest yield was obtained under nitrogen omission was (1509 kg/ha). Hirzel et al. (2011) confirmed the low grain yield record from the basal application of the entire recommended dose at planting. This could be due to low available nitrogen from loss by denitrification, leaching and volatilization.
When NDVI value 0.75 and 0.8 were used as critical values for applying 20 kgN/ha to rice irrespective of basal application, accordingly grain yield of these treatments was significantly higher than blanket application. Yield obtained under NDVI threshold of 0.8 was also higher than NDVI threshold of 0.75. where he witnessed that the yield at NDVI threshold 0.8 was significantly higher than SPAD 40 and LCC 4 and 5. This is because NDVI considers only crop reflectance, but SPAD considers both chlorophyll content and soil characterstics.
That is why in SPAD thresholds fertilizer application was 7-10 days later than NDVI thresholds. When SPAD value of 35.0 and 37.5 were used as critical values for applying 20 kgN/ha to rice irrespective of basal application, accordingly grain yield of these treatments were significantly higher than blanket application. Yield obtained under SPAD threshold of 37.5 was also higher than SPAD threshold of 35.0. These results are in contrasts to findings of Peng et al. (1996)  However, the highest HI obtained under T5 was 17% higher than that of T7 and 6% higher than T3. Maximum harvest index was found in T5, because of higher grain yield comparative to corresponding straw yield that might have resulted due to proper availability of nutrients, which ultimately lead to higher LAI, high dry matter accumulation and translocation, production of higher panicle bearing tillers per unit area and more number of filled grains per panicle Sidhu.
Better chlorophyll development that might have improved the vegetative and reproductive growth of the crop influenced directly or indirectly for higher production under higher fertilizer uptake. More over sustaining higher leaf area due to balanced plant food available in post flowering phase might have encouraged dry matter partitioning leading to better grain filling and higher test weight realizing higher grain yield. Artacho (2009) believed that HI, which is the ratio of grain yield to total biomass, was not affected by N fertilization. But present experiment was in contrast to that. Fageria et al. (2010) explained that the HI represents partitioning of photosynthate between grain and vegetative plant parts.

CONCLUSION
Real time approach to nitrogen management is cost effective and reliable where the timing of nitrogen application is determined through periodic monitoring of crop nitrogen status. The synchronization between demand and supply from all available sources including fertilizers can help to achieve higher nitrogen use efficiency. The present investigation was carried out to study the threshold colour of rice leaves in terms of NDVI and SPAD values at which fertilizer N needs to be applied to achieve higher yield and fertilizer N use efficiency.
The study revealed that application of 30 kg N/ha as basal dose and top dressing of 20 kg N/ha twice at 35 and 63 DAS guided by NDVI threshold value of 0.8 (T5) was found to be superior over other treatments with respect to productivity, profitability, nutrient and energy use efficiency. It recorded highest grain yield of 4438 kg/ha which was 30% higher than fixed schedules (T3) and 7.1% higher than that top dressed at SPAD threshold value of 37.5 (T7). T5 produced maximum dry matter of 8678 kg/ha with highest grain yield (4438 kg/ha), straw yield (5092 kg/ha) and harvest index 46.0%.
As GreenSeeker and SPAD thresholds vary for different agro climatic conditions, study revealed that GreenSeeker threshold of 0.8 and SPAD threshold of 37.5 may be suitable for Odisha condition