Feasibility of variable rate application of diammonium phosphate fertilizer to wheat crop under center pivot irrigation system

2.1 Experimental site

The study was conducted, in the period from November 2018 to March 2019, on a 30 ha field (ID: S1-3) in Al-Kafaa commercial farm of Al-Rajhi International Investment Company, located in the Northern region of Sudan between the latitudes of 18°10′8.503″ and 18°11′37.837″ N and the longitudes of 34°0′26.021″ and 34°11′18.415″ E (Fig. 1). The experimental field was planted with spring wheat (variety Imam) under a solar-powered center pivot irrigation system. The wheat crop was planted on November 10th, 2018 using a Calydon seed drill (T4.8 hybrid model) with a spacing of 11 cm between rows. The soil of the experimental field, which was characterized as a sandy loam soil, had an EC ranging between 0.37 and 0.47 dS cm⁻¹, a pH values between 7.64 and 8.78, soil nitrogen values ranging between 0.15% and 0.20% and CaCo3 ranging between 14.10% and 25.10%. Weather in the farm was very hot in the summer (44 ± 2 °C) and cold to moderate in the spring (16 ± 2 °C) with an annual rainfall of about 37 mm.

Close

Fig. 1

Location map of the experimental site.

Export to PPT

2.2 Description of the equipment

Implementation of variable rate DAP fertilizer was achieved using an automatic control system developed by Ag Leader Technology (https://www.agleader.com). A New Holland tractor (140 hp, T7060 model) was used to pull a granular fertilizer spreader (brand Sulky, XT100 model) of 24 m spreading width. The variable rate fertilizer application system included: (a) RTK guidance system that provided the geographical position and speed of the spreader to the system with an accuracy of 2.5 cm, (b) variable-rate module and auto shut-off installed directly into the spreader, (c) 400 MHz RTK- GPS Base station for auto-correction of GPS signal to less than 2.5 cm and (d) AgFiniti for data exchange between the base station and machines in the central office.

The variable rate applicator was designed to primarily operate on the basis of a prescription GIS map (management map) prepared using soil and plant parameters, along with GPS coordinates and desired input rates. The GPS was used by the control system in the applicator to determine, in real time, the location of the applicator in the field, interrogate the GIS management map for the desired application rate at the current location and automatically change the application rate by changing the rotational speed of the conveyor belt of the spreader to apply the desired rate at the current spreader field location.

2.3 Soil parameters

Geo-referenced soil samples, from 36 locations across the experimental field, were collected from the top soil layer at a depth of 0–25 cm using an auto soil sampler (Model: Wintex 2000). The soil samples were collected using a systematic grid sampling approach with grids of 100 m × 100 m following the sampling strategy described by Franzen (2011). The collected soil samples were analyzed in the laboratory for soil phosphorus following the procedure reported by Estefan et al (2013).

2.4 Fertilizer type

The granular fertilizer used in this study was the diammonium phosphate (DAP), which was the only source of phosphorus element that was commonly used in the study farm. Therefore, variable rates of DAP fertilizers were applied based on soil phosphorus content; where, soil nitrogen in the experimental field was found almost homogeneous (ranging from 0.15% to 0.20%). The studied DAP fertilizer rates (120 and 160 kg ha⁻¹) and the conventional rate practiced in the farm (200 kg ha⁻¹) were added at the sowing stage of the wheat crop, based on the prescription map prepared on the basis of the soil phosphorus variations. Mixing the phosphate fertilizer with wheat seed at the time of planting enhances the P element availability in soil, thus increases the total dry matter and grain production, even in soils with medium to high levels of available P (Alessi and Power, 1980).

2.5 Prescription map

Georeferenced grid data points (grid size of 2 m × 2 m) were interpolated and mapped by using geostatistical procedures (i.e. kriging) available in ArcGIS software program (ver. 10.7.1). Based on soil phosphorus concentrations data from the 36 data points, a prescription management map of the study field was developed using the ArcGIS software. The developed prescription map was divided into three equal interval classes based on soil phosphorus content in the experimental field (low: 3.75 to 4.50 ppm, medium: 4.51 to 5.25 ppm and high: 5.26 to 6.00 ppm). Management zones were delineated in the management map based on the soil phosphorus classes (categories).

2.6 Statistical analysis

Split plot design analysis was employed with the three soil phosphorus levels as the main treatments and the three DAP fertilizer application rates as the sub-treatments with three replicates. The ANOVA (analysis of variance) procedure, within the Statistical Analysis System (SAS for Windows v. 9.4), was used to determine the interaction effects of the management zones and the DAP application rates on the yield of wheat crop.

2.7 Wheat growth

Wheat growth status was monitored using the Normalized Difference Vegetation Index (NDVI) extracted from periodic sets of Sentinel-2 satellite images (10 m × 10 m pixel size). The NDVI, as an indicator of crop growth performance, was calculated using Equation (1) (Rouse et al., 1974). Four Sentinel -2 satellite images were downloaded at different growth stages of the wheat crop (tillering stage, stem elongation, grain filling and maturity stages). NDVI was selected as an indicator of crop growth performance because it was efficient in crop analysis for crops with unclear flowers and simply “green-up” and then “green-down” after entering the reproductive growth stages, such as wheat and maize (Vallentin et al., 2021). In addition, Ren et al. (2008) reported a considerable linear correlation between spatial accumulation of NDVI and wheat yield at different growth stages with R² values ranging between 0.61 and 0.86. $${\mathit{NDVI}}_{(B4,B8)}=\frac{\mathit{NIR}-Red}{\mathit{NIR}+Red}.(1)$$

2.8 Wheat yield

The wheat crop of the experimental field (S1-3) was harvested on March 24th, 2019 (134 days after sowing) using a New Holland combine harvester (TC5.70 model). To estimate wheat yield, 81 wheat samples were randomly collected from the experimental field (3 samples × 3 phosphorus levels × 3 DAP rates × 3 replicates). The collected samples were weighed and converted into t ha⁻¹. Subsequently, the collected yield data was used to generate wheat yield map for the experimental field.

3.1 System accuracy

The static calibration of the variable rate applicator was carried out by measuring the density of the DAP fertilizer (kg m⁻³), adjusting the feed gate opening (cm) and weighing the actual output fertilizer (kg). The calibration results indicated that setting the gate opening at 2.0 cm, using a DAP fertilizer with a density of 984.1 kg m⁻³, produced a fertilizer flow rate of 4.12 kg rev^-1. The response time of the spreader to change between two fertilizer application rates was determined in the field. Results revealed that the average response time and the signal lag time was ≤ 2.63 s.

3.2 Soil phosphorus

Analysis of soil samples showed significant spatial variations in phosphorus concentrations across the study field that ranged between 3.75 ppm and 6.00 ppm. Using the ArcGIS software program, concentrations of soil phosphorus for the 36 collected samples were tagged to the corresponding coordinates to generate a map of soil phosphorus spatial variability across the experimental field. The generated soil phosphorus map was classified into three zones, namely, (a) a low soil phosphorus zone (3.75 to 4.50 ppm) that occupied an area of 10.2 ha, (b) a medium soil phosphorus zone (4.51 to 5.25 ppm) that occupied an area of 14.7 ha, and (c) a high soil phosphorus zone (5.26 to 6.00 ppm) that occupied an area of 5.1 ha. The classified soil phosphorus map was used as a base for the experimental design utilized for the implementation of the proposed treatments of variable rate DAP fertilizer. Three DAP fertilizer rates (120, 160 and 200 kg ha⁻¹) were applied on the experimental field. Soil phosphorus levels and layout of the applied variable DAP fertilizer rates are shown in Fig. 2.

Close

Fig. 2

Soil phosphorus zones and variable DAP fertilizer rates in the study field.

Export to PPT

3.3 Wheat growth performance

3.3.2

3.3.2 NDVI measured at 75 days after sowing

Table 2 summarizes the recorded NDVI values at 75 days after wheat sowing (stem elongation stage). A maximum mean NDVI value of 0.487 recorded under high soil phosphorus zone was significantly higher than that under both the medium (0.465) and low (0.453) soil phosphorus zones. However, there was no significant differences in the NDVI values between low and medium soil phosphorus zones. The ANOVA results showed significant differences in the NDVI values as a result of variable DAP fertilizer rates. The highest mean NDVI value of 0.478 was recorded at a DAP fertilizer rate of 200 kg ha⁻¹, which was not statistically significant compared to the medium DAP level of 160 kg ha⁻¹ with a maximum NDVI value of 0.477. However, significant differences were recorded in the NDVI mean values between the low (120 kg ha⁻¹) and medium (160 kg ha⁻¹) DAP fertilizer application rates. An NDVI map at 75 days after wheat sowing was generated to illustrate the spatial variations in NDVI values within the experimental field (Fig. 3).

Table 2 NDVI measured at 75 days after sowing for variable DAP rates under different soil phosphorus zones.

Phosphorus zone	DAP fertilizer rate			Mean
	120	160	200
Low	0.428	0.459	0.472	0.453
Medium	0.461	0.475	0.459	0.465
High	0.460	0.497	0.504	0.487
Mean	0.449	0.477	0.478	0.468
Statistical Analysis	P > F		LSD0.05
Phosphorus (P)	0.008		0.017 S
Fertilizer Rate (F)	0.011		0.019 S
F at same P	0.260		NS

Close

Fig. 3

NDVI measured at 75 days after sowing.

Export to PPT

3.4 Wheat yield

Wheat yield (t ha⁻¹) of the experimental study field was determined for the different soil phosphorus zones and variable DAP fertilizer application rates (Fig. 4) and evaluated using the statistical analysis presented in Table 5. Results revealed significant differences in wheat yield as a result of soil phosphorus levels and variable DAP fertilizer application rates. Optimum wheat yield was observed at high soil phosphorus level with a mean value of 2.541 t ha⁻¹, which was statistically significant compared to that under low soil phosphorus zone (2.332 t ha⁻¹). However, no significant differences were observed in wheat yield between the high (2.541 t ha⁻¹) and the medium (2.425 t ha⁻¹) phosphorus zones. Wheat yield results under variable DAP fertilizer application rates showed that the highest yield was recorded at the high DAP fertilizer application rate (200 kg ha⁻¹) with a mean value of 2.449 t ha⁻¹, followed by that at the medium application rate (160 kg ha⁻¹) of 2.441 t ha⁻¹, with no significant differences between them. Significant differences were observed between wheat yields under the low (2.407 t ha⁻¹) and the high (2.449 t ha⁻¹) fertilizer application rates. The results were at the same trend of the NDVI values of wheat at the tilliering, stem elongation and maturity stages.

Close

Fig. 4

Wheat yield map of the experimental field.

Export to PPT

These results are in agreement with the findings of Sultana et al. (2014) at the maturity stage. As illustrated in Table 5, the highest wheat yield (2.560 t ha⁻¹) was recorded at the medium application rate in the medium soil phosphorus zone. Therefore, adding more DAP fertilizer would not be useful in improving wheat yield because only 20 to 30% of phosphorus fertilizers was effectively used by crops, while the remaining would leach down to groundwater, or get fixed within the soil (Rowse and Stone, 1981).

Results of this study was in agreement with the findings reported by Rowse and Stone (1981), where the highest yield in the low soil phosphorus zone was recorded at the high DAP fertilizer application rate; however, the highest yield in the high soil phosphorus zone was obtained at the medium DAP fertilizer application rate. However, the application of uniform DAP fertilizer rates practiced in the study area would not improve the yield of wheat crop, but would only increase the cost of production.

Based on the results of this study, wheat yield wasn’t significantly improved, but the decrease in the amount of DAP fertilizer, without affecting the yield, led to reduction in the total cost. Moreover, reduction in DAP would reduce the negative impacts on the environment by applying only the required fertilizer rates on each zone of the field. Excessive use of fertilization causes the accumulation of heavy metals in the soil and plants, which leads to pollution of water, soil and air (Savci, 2012). This considered as one of the advantages of variable rate technology that can provide farmers with a better understanding of soil fertility variables within the field to optimize fertilizer application rates; thus, reducing the negative effects of excessive application, such as environment degradation, crop damage and yield losses (Azhagesan and Ravikumar, 2016). Results of this study can be extended to develop reasonably fast and accurate models for predicting crop yields using machine learning or deep learning techniques, such as that described by Mudunuru et al. (2021).

3.5 Economic feasibility of variable rate application of DAP fertilizer

As indicated in Table 6, the economic feasibility of variable rate application was studied based on the results of statistical analysis. The analysis showed that there were no significant differences in wheat yield between the DAP fertilizer rates of 160 and 200 kg ha⁻¹ in high and medium phosphorus zones. Therefore, the DAP fertilizer rate of 160 kg ha⁻¹ was considered as the optimum rate for wheat production in the high and medium phosphorus zones of the study area. However, the DAP fertilizer of 200 kg ha⁻¹, which was significant with other rates, was considered as the optimum rate for the low soil phosphorus zones. The total DAP amount for the conventional rate (200 kg/ha), practiced in the study area, for the total area of the experimental field (30 ha) was 6000 kg, at a cost of 3600 $. Likewise, the total amount DAP fertilizer for the rate of 160 kg ha⁻¹ in the medium and high phosphorus areas (19.8 ha) was 3168 kg, at a cost of 1900.8 $. Also, the total DAP amount for the rate of 200 kg ha⁻¹ in the low soil phosphorus zone (10.2 ha) was 2040 kg, at a cost of 1224 $. The total amount of 6000 kg of DAP fertilizer was required for the uniform application rate in the experimental field compared to 5208 kg when practicing variable rate application. This results in savings of 792 kg of DAP fertilizer at a cost of 475.2 $ for the experimental 30 ha field (i.e. 15.84 $ ha⁻¹).