Optimizing application rate of nitrogen, phosphorus and ... - SciELO

17 downloads 0 Views 803KB Size Report
Hussain, Z., Khattak, R.A., Irshad, M., Mahmood,. Q., An, P. 2016. Effect of saline ... Mahmood, F., Khan, I., Ashraf, U., Shahzad, T., Hus- sain, S., Shahid, M., Abid, ...
Journal of Soil Science and Plant Nutrition, 2018, 18 (1), 13-26 RESEARCH ARTICLE

Optimizing application rate of nitrogen, phosphorus and cattle manure in wheat production: An approach to determine optimum scenario using response-surface methodology

Mohsen Jahan1* , Mohammad Behzad Amiri2 Associate Prof. of Agroecology, Ferdowsi University of Mashhad, Faculty of Agriculture, P.O. Box 91775-

1

1163, Mashhad, Iran. 2Assistant Prof. of Agroecology, Dep. Plant Production Engineering, Gonabad University, Khorasan Razavi, Iran. *Corresponding author: [email protected]

Abstract Optimal application rates of inorganic nitrogen (N), phosphorus (P) and cattle manure were estimated using Response Surface Methodology (RSM). A Central Composite Design (CCD) was conducted at field level during 2012-13 and 2013-14 growing seasons. The applied levels of N were 0, 150 and 300 kg N ha-1 in form of urea. The levels used for P fertilizer were 0, 100 and 200 kg ha-1 (P2O5) and for cattle manure were 0, 15 and 30 t ha-1. Both seed yield (SY), biological yield (BY) were measured at harvest time. N loss (NL) and Agronomic N Use Efficiency (ANUE) were calculated based on other measurements. Increasing N and P rates up to 200 kg ha-1 increased SY. Optimization of N, P and manure application amount was based on economic, environmental and eco-environmental scenarios. Under economic scenario, using 145.4 kg ha-1 N, 200 kg ha-1 P and 18.4 t ha-1 manure resulted in 6500 kg ha-1 SY with ANUE of 10.49. For environmental scenario, by N application of 21.2 kg ha-1, no application of P and applying 16.3 t ha-1 manure, SY and ANUE of 3160 kg ha-1 and 9.08 were obtained, respectively. Using eco-environmental scenario, by applying 144.7 N and 34.3 kg ha-1 P, plus 30 t ha-1 manure, about 4031 kg ha-1 SY and a considerable high ANUE of 16.5 were recorded. The results of this study showed that the privilege of eco-environmental scenario compared to the other scenarios was mainly due to higher ANUE. Keywords: ANUE, nitrogen losses, eco-environmental scenario, seed yield, central composite design

13

14

Jahan and Amiri

1. Introduction Higher required production of wheat has intensified

is an important factor, which reduce nitrogen losses

consumption of agrochemicals that in turn have re-

(Gastal & Lemaire, 2002). Thus integrated use of

sulted in negative impacts on environment. Sustain-

slow release organic fertilizers such as manure could

able and sufficient food production is a necessity that

reduce N losses during crops cultivation (Emilsson et

simultaneously should consider social, economic and

al., 2007). In arid and semi-arid regions, N deficiency

environmental aspects. The first step to achieve this

occurs faster than other nutrients and usually is due

goal is optimization and improvement of resources

to low level of soil organic matter. It seems necessary

use efficiencies (Gliessman, 1998). It is reported that

to design cropping systems with high nutrient uptake

up to 50 percent of applied nitrogen would drift from

and utilization efficiencies (Fageria, 2014; Cassidy

agricultural systems as gaseous compounds and other

et al., 2013) considering global high demand for ce-

types of activated nitrogen (Jarvis et al., 2011; We-

real in coming future. It is suggested to increase the

ligama et al., 2010). At the high level application of

mean efficiency of N utilization by 0.1-0.4 % annu-

P (more than 200 kg ha-1), up to 90% of phosphorous

ally (Doberman & Cassman, 2005). Improving N use

fertilizers would be fixed in soil together with metallic

efficiency, increases economic return and reduces the

elements as insoluble forms which finally lead to fur-

potential of environmental pollution (Fageria, 2014).

ther phosphorus pollution (Adesemoye et al., 2009).

To reach such goals, research priorities might be shift-

In many crops, low absorption efficiency of fertil-

ed from maximizing yield to developing methods for

izers, is the main reason of leaching, volatilization

optimization of nutrients application.

and diffusion of soluble chemical fertilizers easily

Response Surface Methodology (RSM) was first intro-

released to soil and air (Akiyama et al., 2000). It has

duced as optimization method for industrial use (My-

been reported that between 18-41 percent of applied

ers and Montgomery, 1995) however; this methodol-

nitrogen retain in soil after crop harvesting (Fageria,

ogy could also be used in fertilizer optimization. Central

2014). Nitrogen losses occur in various ways as am-

Composite Design (CCD), which is the most popular

monium volatilization in lime soils (10-70%), denitri-

RSM design, was implemented to design a series of tests

fication (9-22%) and leaching (14-40%) (Doberman

with least number of experiments. This approach tries to

& Cassman, 2004).

investigate the effect of parameters involved (i.e. contact

Wheat crop shows a strong positive correlation be-

time, dosage, pH, initial concentration) on responses in a

tween productivity and NPK fertilizers (Hawkesford

cost- and time-effective way. The CCD makes it feasible

& Barraclough, 2011; Osborne, 2007). On the other

to observe the possible interaction of the parameters and

hand, applying organic fertilizers in integration with

their influences used RSM methodology to optimize N,

chemical ones could be an efficient management

water and plant density in canola cultivation. reported

practice to reduce application rate of chemical fertil-

that the application of 93.48 kg N ha-1 based on eco-envi-

izers and subsequently reduce their negative impacts

ronmental scenario was an optimized N use, was able to

on environment. Moreover, improved nutrient bal-

reduce environmental hazards and produced acceptable

ance in soil and plant, enhance crop productivity and

onion yield. It has been claimed that oxalic acid and lactic

yield stability in intensified cultivation. Synchrony

acid are the major acids responsible for enhancing the P

between application time and crop nutrient demand,

solubilization .

Journal of Soil Science and Plant Nutrition, 2018, 18 (1), 13-26

Optimizing application rate of nitrogen

15

This study was aimed to optimize the chemical and

Soil samples were taken from 0-15 and 15-30 cm

organic fertilizer use in winter wheat production and

depths and analyzed for some physiochemical prop-

determine the best applicable scenario in Kashaf-rood

erties before beginning the experiment (Table 1). To

watershed in northeast of Iran. It has also been stud-

determine soil nitrogen content, soil sampling was

ied the application trend of different N, P and cattle

repeated at the end of growing season.

manure levels and their effects on wheat production. Furthermore, the effectiveness of manure compared

2.2. Experimental design

to chemical fertilizer was studied based on NUE and CCD (sometimes called Box-Behnken design) with

wheat yield improvement.

two replicates was used for fitting response surface 2. Material and Methods

to experimental data (Tawfik et al., 2017). Two years data where subjected to combined analysis after en-

2.1. Site description

suring uniformity of the error mean squares. The results of combined analysis indicated that year × treat-

Field studies conducted during 2012-13 and 2013-14

ment interaction was not significant (P >0.05), there-

growing seasons at the Research Station of Agricul-

fore the two years data were joined before exposing to

ture Faculty, Ferdowsi University of Mashhad, Iran

response surface analysis.

(latitude: 36° 15¢ N; longitude: 59° 28¢ E; elevation:

The experimental factors were the combination of

985 m above sea level). Experiment station was lo-

different amounts of nitrogen, phosphorus and cattle

cated in Kashaf-rood watershed in northeast of the

manure. The total number of experimental runs for a

country in a semi-arid region with mean annual pre-

3-factor CCD is 15 including 12 factorial points and 3

cipitation of 252 mm and temperature of 15 °C.

replications for center points.

Table 1. Soil properties of the experimental field (mean of two years). Soil depth (cm) Soil properties Total N (%) Available P (ppm) Available K (ppm) C/N pH (saturation extract) EC (dS m-1) Water storage capacity (%) Bulk density (g cm-3) Texture grade

0-15

15-30

0.078 19 380 12.8 7.7 1.3 22.3 1.34 Loamy-silt

0.065 15 372 12.3 7.5 1.3 20 1.41 Loamy-silt

Journal of Soil Science and Plant Nutrition, 2018, 18 (1), 13-26

16

Jahan and Amiri

By conducting CCD, it is possible to obtain all in-

kg ha-1) and manure (0, 30 t ha-1) (Table 2). The N, P

formation from the least operational practices due to

and K content of manure were determined as 1.18%,

distribution of experimental points through treatments

0.29% and 1.04%, respectively. The high and low

confined. The design points were defined based on the

levels of manure were determined based on nutrient

low and high levels of N (0, 300 kg ha-1), P (0, 200

content and local recommendations.

Table 2. Actual and coded values of expeimental factors for CCD. Treatment values* Runs 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Nitrogen (kg ha-1) 300 150 300 300 0 150 150 0 150 150 150 150 0 300 150

Phosphorus (kg ha-1) 0 0 100 100 100 100 0 100 0 200 200 100 200 200 100

Coefficients Manure (t ha-1) 15 15 30 0 30 15 30 0 0 0 30 15 15 15 15

Nitrogen (X1) +1 0 +1 +1 -1 0 0 -1 0 0 0 0 -1 +1 0

Phosphorus (X2) -1 -1 0 0 0 0 -1 0 -1 +1 +1 0 +1 +1 0

Manure (X3) 0 0 +1 -1 +1 0 +1 -1 -1 -1 +1 0 0 0 0

* +1, -1, and 0 indicates up, down and medium level of each factor

A normality test was already performed and transfor-

were arranged to sow wheat seeds on double side

mation was also performed for numerical data where

of rows. Manure was well mixed with soil one

needed. To ensure uniformity of treatment variances,

month before planting. The first N installment and

the Bartlett's test was used. Since there was no statis-

the whole P amount were applied at sowing. The

tical difference between both years of data, the mean

sowing dates (October 10, 2013-14) were the same

value of each trait was reported. Data analysis and

for both years of experiment. Plots were immedi-

graph plotting were performed using Minitab® Statis-

ately irrigated after sowing and after that irrigation

tical Software Ver. 16.1.1, and Microsoft Excel Ver. 14

continued at 10-day intervals. An eco-climate appropriate cultivar (Gascogne) was employed in this

2.3. Crop management

study. The plant density was 350 plants per square meter. Weeds were controlled two times during

Plots of 3 × 4 m with a distance of 1 meter be-

the growth period by hand. No agrochemical

tween, to avoid nutrients mixing due to irrigation,

was used during soil cultivation, planting and

were prepared. Each plot consists of 6 rows that

growing season.

Journal of Soil Science and Plant Nutrition, 2018, 18 (1), 13-26

Optimizing application rate of nitrogen

2.4. Measurements

ANUE =

Y Grain N initial + N fertilizer

17

Eq.2

Each plot was partitioned into two sections, one for seed yield and its components and the second sec-

YGrain: seed yield (kg m-2)

tion used for destructive time series sampling during

To determine relative water content (RWC) of leaves,

the crop growth season. Before the final harvest, 5

the samples were prepared between 9:00 to 10:00 AM

plants from each plot were randomly selected and

at flowering stage. The samples were submerged in

the number of fertile tillers was recorded. Seed yield

distilled water for 6 hours and then turgor weight was

was determined from 4 m2 of each plot which was

measured and their dry weight were also recorded in

kept untouched, considering marginal effect. The air

after drying samples in 75 oC in oven. Finally RWC

dried plants were weighed and biological yield (dry

was calculated using Equation 3 (Kramer, 1988):

matter yield), seed yield and harvest index were also measured. The N contents in plant tissue were measured using

RW C =

AOAC Official Method by a Kajehldal Semi Auto-

(FW − DW ) (TW − DW )

Eq.3

mated Distillation Unit (Horwitz & Latimer, 2005). The total N was determined for each soil plot (AOAC

2.5. Statistical analysis

official method 968.06 (4.2.04)). Nitrogen loss was calculated using Equation 1 (Jarvis et al., 2011):

Response of measured variables (y) to experimental factors (X) was estimated using second order polynomials including the interaction (Equation 4):

Eq.1

Nloss= Ninitial+ Nfertilizer – (Nplant+ Nsoil)

Where, Nloss: nitrogen loss (kg m ), Ninitial: soil avail-2

m

m

m

i =1

i< j

i =1

y = β 0 + ∑ βi X i + ∑ β i j X i X j + ∑ β ii X i 2

Eq.4

able nitrogen content at the early season (kg m-2) which was calculated by: NTotal End ×0.03 0.03= availability coefficient of total N in soil (as

Where b0 is constant and bi, bij and bii are coefficients

mineral for crop use) (Fageria, 2014).

for linear, interaction and quadratic terms, respec-

Nfertilizer: applied nitrogen (kg m-2), Nplant: plant nitrogen

tively.

content at the end of season, and Nsoil: the available ni-

The result was a second order polynomial which de-

trogen content in soil after final harvest (kg m ) which

scribes the estimated of response (yield) as a func-

was calculated as:

tion of inputs variables. Finally, after optimizing

NTotal End ×0.03.

the resulted function and eliminating the low effect

Agronomical nitrogen use efficiency (ANUE, kg

terms using statistical tests and criteria such as, F

Grain/kg Nfertilizer) was calculated using Equation 2

test, lack of fit test, coefficient of determination (R2),

(Rathke et al., 2006):

a final function was calculated to predict yield and

-2

other expected variables as below (Equation 5):

Journal of Soil Science and Plant Nutrition, 2018, 18 (1), 13-26

18

Jahan and Amiri

Y = a0 + a1X 1 + a2 X 2 + a3 X 3 + a4 X 12 + a5 X 2 2 + a6 X 3 2 + a7 X 1X 2 + a8 X 1X 3 + a9 X 2 X 3

Eq.5

In this function, Y is dependent variable, X1 is N fer-

The RMSE percentage states the different be-

tilizer, X2 is P fertilizer, X3 is manure, and a0 to a9 are

tween predicted versus observed values. When

coefficients. The equation is only functional in the de-

RMSE