Site characteristics
The study area is characterized by a dry tropical savanna climate, with hot summers and cool winters. Three consecutive field trials, viz., a fertility gradient experiment in 2020, a test crop experiment in 2021 and a validation trial in 2022, were conducted at Zonal Agricultural Research Station, University of Agricultural Sciences, Bangalore, Karnataka, India. The research station is located at a latitude of 13° 04′ 55.2′′ N and a longitude of 77° 34′ 10.0′′ E, at an elevation of 930 m above sea level. The annual precipitation level recorded at the site was 1180.0 mm in 2020, increasing to 1328.40 mm in 2021 and further to 1556.80 mm in 2022. Over the three years, the maximum temperatures varied, ranging from 29.07 to 18.28 °C in 2020, 28.89 to 18.13 °C in 2021, and 28.41 to 17.78 °C in 2022, as depicted in Fig. 1.
Variation in monthly rainfall and maximum and minimum temperatures during the field experiments from 2020–2022 (RF: rainfall; T max: maximum temperature; T min: minimum temperature).
The well-drained red soil at the research site belongs to the taxonomically defined large group known as Typic Kandic Paleustalfs, which is a member of the fine mixed Isohyperthermic family. Prior to the start of the experiment, soil samples were taken randomly from the top 15 cm of the experimental field. These samples were then air-dried in the shade, processed, and analyzed for their chemical properties. The analysis revealed that the soil pH was 5.93, with an electrical conductivity of 0.21 dS m−1. The organic carbon content was found to be 4.95 g kg−1, while the levels of available nitrogen, phosphorus, and potassium were 220.17 kg ha−1, 76.19 kg ha−1, and 149.36 kg ha−1, respectively.
Fertility gradient experiment with fodder maize
In 2020, an experiment was conducted to reset the soil fertility levels and establish a controlled fertility gradient in relation to NPK availability before the test crop experiment. Three uniform plots were established, each receiving different amounts of nitrogen (N), phosphorus (P), and potassium (K) fertilizers: 0–0–0 kg NPK ha−1, 150–135–200 kg NPK ha−1, and 300–270–400 kg NPK ha−1 respectively. Details on the fertilizer quantities and the experimental design can be found in Supplementary Fig. 1. The standard dose of nitrogen fertilizers was fixed on the basis of the recommended dose of fertilizers for fodder maize (150 kg ha−1). The phosphorus and potassium doses were fixed on the basis of the P- and K-fixing capacity (135 and 200 kg ha−1, respectively) of the soil. To equalize and artificially induce fertility levels, the fodder maize variety ‘African tall’ was cultivated. Following a 60-day growth period, the crop was harvested, and surface soil samples (0–15 cm depth) were collected from each plot. These samples were then air-dried, finely ground, and analyzed to measure available nitrogen via the alkaline KMnO4 method12, phosphorus through Bray’s extractant and the ascorbic acid method13, and potassium via ammonium acetate extraction and flame photometry14 to evaluate the progression of the fertility gradient.
Test crop experiment with Kodo millet
During the Kharif season of 2021, a test crop experiment was carried out using kodo millet after confirming the establishment of the fertility gradient to assess the impact of graded doses of FYM and NPK fertilizer combinations on crop yield. Each fertility strip was divided into three blocks of different levels of FYM: none (F0), the recommended dose (F1), and twice the recommended dose (F2). These blocks were further subdivided into eight plots, where seven unique nitrogen (N), phosphorus (P), and potassium (K) fertilizer combinations, along with one control plot, were tested. This resulted in a total of 24 plots per fertility strip, resulting in a total of 72 plots across the entire field. The fertilizer treatments were administered in four varying quantities for N (0, 10, 20, and 30 kg ha−1), P2O5 (0, 10, 20, and 30 kg ha−1), and K2O (0, 2.5, 5, and 7.5 kg ha−1), alongside three levels of FYM (0, 6.25, and 12.50 t ha−1). From these, 21 different NPK combinations were selected by incorporating seven treatment combinations and one control for each FYM level. These combinations were then randomly assigned across the remaining two fertility strips and within the FYM blocks. The experimental design and nutrient levels are depicted in Supplementary Fig. 2. Subsequent to the layout, soil samples were collected from the surface (0–15 cm depth) to ascertain the available N, P, and K contents in each plot before sowing, following the standard soil analysis procedure mentioned in the gradient experiment. Nitrogen was sourced from urea, phosphorus from single super phosphate, and potassium from muriate of potash. FYM was applied two weeks before sowing, and half of the nitrogen and the entire recommended amounts of phosphorus and potassium were used as basal fertilizers. Standard cultivation practices were followed for growing the crop. The remaining nitrogen was distributed evenly across the two applications at 30 and 60 days after sowing. Upon maturity, the crop was harvested, and the yield from each plot was calculated and expressed in kg per hectare. The plant samples from each plot were initially air-dried, followed by oven drying at 65 °C. These samples were then ground via a Willey mill. The nitrogen content was quantified via the micro Kjeldahl technique15. Following Jackson’s 1973 protocol, a diacid solution was prepared with nitric and perchloric acid mixture at a ratio of 9:4. Predigesting involved treating the sample with 10 mL of nitric acid per gram. This solution was then utilized to quantify the phosphorus and potassium levels in the samples. Phosphorus was quantified via the vanadomolybdate method for a yellow phosphoric acid color reaction16, whereas potassium levels were determined via flame photometry16.
$$Nutrient~\;upatke~\left( {{\text{kg~h}}{{\text{a}}^{-1}}} \right)=\frac{{\left[ {Nutrient\;~content~\left( \% \right)~ \times ~Dry~\;weight~\left( {{\text{kg~h}}{{\text{a}}^{{\text{–1}}}}} \right)} \right]}}{{100}}$$
From the data on the soil test values, crop dry matter yield and nutrient uptake, basic parameters such as nutrient requirements (NR), and the contributions of nutrients from soil (CS), fertilizer (CF) and organic manure (C-OM) were calculated via the following formulae10.
$${\text{NR }}\left( {{\text{kg k}}{{\text{g}}^{ – {\text{1}}}}} \right){\text{ }}=\frac{{{\text{Nutrient uptake }}\left( {{\text{NPK}}} \right){\text{ by grain }}+{\text{ straw }}\left( {{\text{kg h}}{{\text{a}}^{ – {\text{1}}}}} \right)}}{{{\text{Grain yield or any economic produce }}\left( {{\text{kg h}}{{\text{a}}^{ – {\text{1}}}}} \right)}}$$
$$\% {\text{ CS}}~~=\frac{{{\text{Nutrient uptake }}\left( {{\text{kg h}}{{\text{a}}^{ – {\text{1}}}}} \right){\text{ by grain }}+{\text{ straw in control plot}}}}{{{\text{Soil test values }}\left( {{\text{Av}}.{\text{ NPK}}} \right){\text{ in control plot }}\left( {{\text{kg h}}{{\text{a}}^{ – {\text{1}}}}} \right)}} \times 100$$
$$\% {\text{ CF}}~~=\frac{{({\text{Nutrient uptake bygrain }}+{\text{ straw in treated plot}}) – {\text{Soil test values in treated plot}} \times \frac{{\% {\text{ Contribution}}~~~\left( {{\text{NPK}}} \right){\text{ from soil}}}}{{100}}}}{{{\text{Fertilizer nutrient dose applied in treated plot }}\left( {{\text{kg h}}{{\text{a}}^{ – {\text{1}}}}} \right)}} \times 100$$
$$\% {\text{ C-OM}}~~~=\frac{{{\text{Total uptake of NPK in organic plot }}\left( {{\text{kg h}}{{\text{a}}^{ – {\text{1}}}}} \right) – {\text{Soil test values in}}~~{\text{OM plot}}~ \times \frac{{\% {\text{ Contribution }}\left( {{\text{NPK}}} \right){\text{ from soil}}}}{{{\text{1}}00}}~}}{{{\text{Quantity of OM added in organic plot }}\left( {{\text{kg h}}{{\text{a}}^{ – {\text{1}}}}} \right)}} \times 100$$
By using basic parameters, the fertilizer nutrient requirements for the targeted productivity of kodo millet were calculated as follows:
-
(i)
Fertilizer N, P and K alone through inorganics (without FYM).
$$FN=\left( {\frac{{NR}}{{CF}} \times 100 \times T} \right) – \left( {\frac{{CS}}{{CF}} \times SN} \right)$$
$$FP=\left( {\frac{{NR}}{{CF}} \times 100 \times T} \right) – \left( {\frac{{CS}}{{CF}} \times SP} \right)$$
$$FK=\left( {\frac{{NR}}{{CF}} \times 100 \times T} \right) – \left( {\frac{{CS}}{{CF}} \times SK} \right)$$
-
(ii)
Fertilizer N, P and K along with FYM for integrated plantrient supply (with FYM).
$$FN=\left( {\frac{{NR}}{{CF}} \times 100 \times T} \right) – \left( {\frac{{CS}}{{CF}} \times SN} \right) – \left( {\frac{{C – OM}}{{CF}} \times OM} \right)$$
$$FP=\left( {\frac{{NR}}{{CF}} \times 100 \times T} \right) – \left( {\frac{{CS}}{{CF}} \times SP} \right) – \left( {\frac{{C – OM}}{{CF}} \times OM} \right)$$
$$FK=\left( {\frac{{NR}}{{CF}} \times 100 \times T} \right) – \left( {\frac{{CS}}{{CF}} \times SK} \right) – \left( {\frac{{C – OM}}{{CF}} \times OM} \right)$$
where FN, fertilizer N (kg ha−1); FP, fertilizer P (kg ha−1); FK, fertilizer K (kg ha−1); SN, soil test value for available N (kg ha−1); SP, soil test value for available P (kg ha−1); SK, soil test value for available K (kg ha−1); C-OM, contribution of nutrients from the FYM; and OM, quantity of the FYM applied (t ha−1).
Validation trial/follow-up trial
During Kharif 2022, a field study was carried out on kodo millet to assess the effectiveness of soil test-based crop response-based equations aimed at achieving targeted yields. These equations were measured against other fertilizer recommendation methods, such as the soil fertility rating approach and the General Recommended Dose (GRD), using a randomized block design with three repetitions. The GRD is derived from extensive trials across multiple locations in which various levels of nitrogen, phosphorus, and potassium fertilizers are tested to determine the most cost-effective dosage for the maximum yield of a specific crop. The soil fertility rating method aligns medium fertility levels with the GRD. In situations of extremely low or very high soil fertility, the fertilizer amount is adjusted up or down by 25–30% from the GRD17. The treatments included T1—STCR NPK TY 17 kg ha–1, T2—STCR NPK + FYM TY 17 kg ha–1, T3—STCR NPK TY 15 kg ha–1, T4—STCR NPK + FYM TY 15 kg ha–1, T5—general recommended dose (GRD), T6—soil fertility rating (SFR) and T7—absolute control. A composite soil sample was collected at 0–20 cm depth from each plot after the plan was laid out and before the start of the experiment. On the basis of the soil test values, NPK fertilizer nutrients were applied for specific yield targets in the STCR and SFR approaches. The crop was grown according to standard agronomic practices and harvested at the full maturity stage, and the grain and straw yields were computed from the net plot and expressed in kg ha−1. Grain and straw samples were collected from each treatment and analyzed for total NPK content via the same procedure, and uptake was determined as described in the test crop experiment. The response yard stick (RYS), percent deviation and value cost ratio (VCR) were computed via the standard formulae shown below10.
$${\text{RYS~=}}\frac{{{\text{Yield~response~}}\left( {{\text{kg~h}}{{\text{a}}^{ – {\text{1}}}}} \right)}}{{{\text{Total~nutrient~applied~~~}}\left( {{\text{kg~h}}{{\text{a}}^{ – {\text{1}}}}} \right)}}$$
$${\text{Percent~deviation~=}}\frac{{\left[ {{\text{Actual~yield~obtained~~}}\left( {{\text{kg~h}}{{\text{a}}^{ – {\text{1}}}}} \right) – {\text{Targeted~yield~~}}\left( {{\text{kg~h}}{{\text{a}}^{ – {\text{1}}}}} \right)} \right]}}{{{\text{~~Targeted~yield~}}\left( {{\text{kg~h}}{{\text{a}}^{ – {\text{1}}}}} \right)}}~~ \times 100$$
$${\text{VCR~ = }}\frac{{\left[ {{\text{Yield~in~treated~plot~}}\left( {{\rm kg}{\text{~ha}}^{{ – {\text{1}}}} } \right){\text{~}} – {\text{Yield~in~control~plot~}}\left( {{\rm kg}{\text{~ha}}^{{ – {\text{1}}}} } \right)} \right]}}{{{\text{~Cost~of~fertilizers~and~FYM~applied~to~treated~plot}}}} \times {\text{Cost }}{\rm kg}^{{ – 1{\text{~}}}} {\text{of~grain}}$$
The nutrient (N/P/K) use efficiency parameters, viz., apparent recovery efficiency (RE), agronomic nutrient use efficiency (AE), and reciprocal internal utilization efficacy (RIUE), were calculated via the following formulae18 to determine the crop response to the added fertilizers and to compare the use efficiency of added nutrients and economics under different fertilizer recommendation approaches.
$${\text{RE}}\left( {{\rm kg}~{\rm k}{{\rm g}^{ – 1}}} \right){\text{~=~}}\frac{{\left[ {{\text{Total~nutrient~uptake~in~treated~plot~}}\left( {{\text{kg~h}}{{\text{a}}^{ – {\text{1}}}}} \right) – {\text{Total~nutrient~uptake~in~control~plot~}}\left( {{\text{kg~h}}{{\text{a}}^{ – {\text{1}}}}} \right)} \right]}}{{{\text{Fertilizer~nutrient~~applied~~}}\left( {{\text{kg~h}}{{\text{a}}^{ – {\text{1}}}}} \right)}}$$
$${\text{AE}}\left( {{\rm kg}~{\rm k}{{\rm g}^{ – 1}}} \right){\text{~=~}}\frac{{\left[ {{\text{Grain~~yield~in~treated~plot~}}\left( {{\text{kg~h}}{{\text{a}}^{ – {\text{1}}}}} \right) – {\text{Grain~~~yield~in~control~plot~}}\left( {{\text{kg~h}}{{\text{a}}^{ – {\text{1}}}}} \right)} \right]}}{{{\text{Fertilizer~nutrient~~applied~~}}\left( {{\text{kg~h}}{{\text{a}}^{ – {\text{1}}}}} \right)}}$$
$${\text{RIUE~=}}\frac{{{\text{Total~uptake~~}}\left( {{\text{kg~h}}{{\text{a}}^{ – {\text{1}}}}} \right)}}{{{\text{Grain~~yield~}}\left( {{\text{kg~h}}{{\text{a}}^{ – {\text{1}}}}} \right)}}$$
Statistical analysis
The basic data from the test-crop experiments were used for developing fertilizer prescription equations. The fertility gradient created is displayed in a box-and-whisker plot, and test crop experimental data were summarized via descriptive statistics with SPSS 16.0 software. A standard regression procedure was used to relate the soil test values and fertilizer dose with the crop yield response19. The data from the validation trial was subjected to analysis of variance (ANOVA) with a randomized block design20. The online statistical program OPSTAT was used for the data analysis (validation trial), and least significance difference (LSD) values at α = 0.05 were used to compare treatment means. Pearson correlation coefficient (r) values were calculated via Microsoft Excel 2016.
